Standard Handbook of Broadcast Engineering

ii
Additional updates relating to broadcast engineering in general, and this book in particular, can be found at the
Standard Handbook of Broadcast Engineering web site:
www.tvhandbook.com
The tvhandbook.com web site supports the professional audio/video community with news, updates, and
product information relating to the broadcast, post production, and business/industrial applications of digital
video.
Check the site regularly for news, updated chapters, and special events related to audio/video engineering.
The technologies encompassed by the Standard Handbook of Broadcast Engineering are changing rapidly,
with new standards proposed and adopted each month. Changing market conditions and regulatory issues are
adding to the rapid flow of news and information in this area.
Specific services found at www.tvhandbook.com include:
• Audio/Video Technology News. News reports and technical articles on the latest developments in digital
radio and television, both in the U.S. and around the world. Check in at least once a month to see what's
happening in the fast-moving area of digital broadcasting.
• Resource Center. Check for the latest information on professional and broadcast audio/video systems. The
Resource Center provides updates on implementation and standardization efforts, plus links to related web
sites.
• tvhandbook.com Update Port. Updated material for the Standard Handbook of Broadcast Engineering is
posted on the site each month. Material available includes updated sections and chapters in areas of rapidly
advancing technologies.
In addition to the resources outlined above, detailed information is available on other books in the
McGraw-Hill Video/Audio Series.
iii
Standard Handbook of Broadcast Engineering

Jerry C. Whitaker and K. Blair Benson
Editors
McGraw-Hill
New York San Francisco Washington D.C. Auckland Bogotá Caracas Lisbon London
Madrid Mexico City Milan Montreal New Delhi San Juan Singapore Sydney Tokyo Toronto
iv
Copyright © 2005 by The McGraw-Hill Companies, Inc.

All rights reserved. Printed in the United States of America. Except as permitted under the United States Copyright Act of
1976, no part of this publication may be reproduced or distributed in any form or by any means, or stored in a data base or
retrieval system, without the prior written permission of the publisher.
1 2 3 4 5 6 7 8 9 0 DOC/DOC 0 1 0 9 8 7 6 5
ISBN 0-07-145100-5
The sponsoring editor for this book was Steve Chapman. The production supervisor was Sherri Souffrance. This book
was set in Times New Roman and Helvetica by Technical Press, Morgan Hill, CA. The art director for the cover was
Anthony Landi.
Printed and bound by RR Donnelley.
McGraw-Hill books are available at special quantity discounts to use as premiums and sales promotions, or for use in
corporate training programs. For more information, please write to the Director of Special Sales, McGraw-Hill
Professional, 2 Penn Plaza, New York, NY 10121-2298. Or contact your local bookstore.
Information contained in this work has been obtained by The McGraw-Hill Companies, Inc. (“McGraw-Hill”) from
sources believed to be reliable. However, neither McGraw-Hill nor its authors guarantee the accuracy or completeness of
any information published herein, and neither McGraw-Hill nor its authors shall be responsible for any errors, omissions,
or damages arising out of use of this information. This work is published with the understanding that McGraw-Hill and its
authors are supplying information but are not attempting to render engineering or other professional services. If such
services are required, the assistance of an appropriate professional should be sought.
This book is printed on recycled, acid-free paper containing a minimum of 50% recycled, de-inked fiber.
v
For
Christy and Tara
You never really leave the broadcast business…

vi
Other books in the McGraw-Hill Video/Audio Series

The McGraw-Hill Video/Audio series continues to expand with new books on a variety of topics of
importance to engineers and technicians involved in professional media. Current offerings include:
• Standard Handbook of Video and Television Engineering, 4th edition, Jerry C. Whitaker and K. Blair
Benson, editors, 2003
• Standard Handbook of Audio and Radio Engineering, 2nd edition, Jerry C. Whitaker and K. Blair Benson,
editors, 2001
• Master Handbook of Audio Production, Jerry. C. Whitaker, 2002
• Master Handbook of Video Production, Jerry C. Whitaker, 2002
• Audio/Video Protocol Handbook, Jerry C. Whitaker, 2002
• Audio/Video Professional's Field Manual, Jerry C. Whitaker, editor, 2001
• DTV Handbook: The Revolution in Digital Video, 3rd edition, Jerry C. Whitaker, 2001
• Television Receivers, Jerry C. Whitaker, 2001
• Video Display Engineering, Jerry C. Whitaker, 2001
• Digital Television Fundamentals, 2nd edition, Michael Robin and Michel Poulin, 2000
• Video Compression Demystified, Peter D. Symes, 2001
• DTV Survival Guide, Jim Boston, 2000
• Data Broadcasting: Understanding the ATSC Data Broadcast Standard; Richard Chernock, Regis Crinon,
Michael Dolan, and John Mick; 2001
• PSIP: Program and System Information Protocol, Mark K. Eyer, 2003
vii
Contents
Contributors ix
Preface xi
Section 1: Frequency Bands, Propagation, and Modulation 1-1

The Electromagnetic Spectrum 1-9
Propagation 1-19
Frequency Sources and References 1-45
Modulation Systems and Characteristics 1-67
Section 2: Radio Transmission Systems 2-1

Radio Broadcast Systems 2-5
Radio STL Systems 2-21
Digital Radio Systems 2-57
IBOC AM Digital Radio System 2-69
IBOC FM Digital Radio System 2-79
Section 3: Television Transmission Systems 3-1

Television Transmission Standards 3-9
Ghost Canceling Reference Signal 3-35
The ATSC DTV System 3-41
Longley-Rice Propagation Model 3-67
Television Transmitters 3-75
Multiple Transmitter Networks 3-97
DTV Satellite Transmission 3-111
The DVB Standard 3-121
Section 4: RF Interconnection Devices and Systems 4-1

Transmission Line 4-5
RF Combiner and Diplexer Systems 4-17
Section 5: Transmitting Antennas and Systems 5-1

Radio Antenna Principles 5-11
Television Antenna Principles 5-33
Television Transmitting Antennas 5-65
Tower Construction and Maintenance 5-81
Tower Grounding 5-87
Lightning Effects 5-115
Transmitter Building Grounding Practices 5-133
Section 6: Radio Receivers 6-1

Receiver Characteristics 6-5
The Radio Channel 6-27
AM and FM Receivers 6-39
viii
Section 7: Television Receivers and Cable/Satellite Distribution Systems 7-1

Television Reception Principles 7-11
ATSC DTV Receiver Systems 7-61
Receiver Antenna Systems 7-85
Cable Television Systems 7-105
Satellite Delivery Systems 7-137
Content Distribution 7-151
Section 8: RF System Maintenance 8-1

RF System Reliability Considerations 8-7
Preventing RF System Failures 8-23
Troubleshooting RF Equipment 8-39
Testing Coaxial Transmission Line 8-57
The Smith Chart 8-63
Standby Power Systems 8-71
Section 9: Test Equipment 9-1

Troubleshooting Digital Systems 9-9
Digital Test Instruments 9-19
Oscilloscopes 9-35
Spectrum Analysis 9-45
Reliability Engineering 9-59
Section 10: Audio Test and Measurement 10-1

Audio Measurement and Analysis 10-5
Audio Phase and Frequency Measurement 10-19
Nonlinear Audio Distortion 10-33
Time Domain Audio Measurements 10-49
Section 11: Video Signal Measurement and Analysis 11-1

Video Information Concepts 11-7
Measurement of Color Displays 11-19
Camera Performance Verification 11-35
Conventional Video Measurements 11-43
Application of the Zone Plate Signal 11-69
Picture Quality Measurement 11-79
DTV Transmission Performance Issues 11-87
ix
Contributors
Edward W. Allen D. E. Manners

William F. Bailey D. Stevens McVoy
Oded Ben-Dov Robert Metzler
Carl Bentz W. G. Miller
Richard C. Cabot R. J. Peffer
William Daniel Skip Pizzi
L. E. Donovan Ken Pohlmann
Donald G. Fink Krishna Praba
J. J. Gibson Dalton H. Pritchard
Peter Gloeggler Wilbur L. Pritchard
Scott Johnson Kishore Manghnani
Cecil Harrison Chandy Nilakantan
Bruce Hofer John Norgard
L. H. Hoke, Jr. Ulrich L. Rohde
Charles A. Kase Joseph L. Stern
J. D. Knox Robert A. Surette
Donald C. Livingston John T. Wilner
Bernard D. Loughlin J. G. Zahnen
Some of the chapters authored by these contributors have been adapted from the McGraw-Hill Standard
Handbook of Video and Television Engineering, 3rd edition, and the Standard Handbook of Audio/Radio
Engineering, 2nd edition. Used with permission. All rights reserved.
xi
Preface
The broadcast industry has embarked on the most significant transition of technologies and business models in
the history of radio and television. Dramatic advancements in computer systems, imaging, display, and
compression technologies have all vastly reshaped the technical landscape, and theses have all affected
broadcasting. For television, the transition to digital television (DTV) is already well underway; and for radio,
digital audio broadcasting (DAB) is beginning to take hold. These changes give rise to a new book in the
McGraw-Hill Video/Audio Series, the Standard Handbook of Broadcast Engineering. This handbook
continues the rich tradition of past offerings, examining analog and digital broadcast technologies, including
AM radio, FM radio, DAB, NTSC, and DTV.
This book is a companion volume to the other landmark handbooks in the McGraw-Hill Video/Audio
Series:
• Standard Handbook of Video and Television Engineering, 4rd ed., which focuses on video information
theory, production standards and equipment, and digital coding.
• Standard Handbook of Audio and Radio Engineering, 2nd ed., which focuses on audio capture, storage, and
reproduction systems.
The Standard Handbook of Broadcast Engineering picks up where these books leave off—covering in
considerable detail the transmission/reception aspects of broadcasting. In earlier editions of these cornerstone
books, the transmission elements of radio and television were included along with the production elements.
However, as the handbooks have grown, and as the scope of broadcasting has expanded to encompass both
analog and digital transmission, the practical limitations of page count for a printed book come into play.
Therefore, to maximize the amount of information that can be made available to readers, the radio and
television broadcast/reception elements have been separated into this stand-alone publication.
About the Book

The Standard Handbook of Broadcast Engineering examines the technologies of radio and television
broadcasting. The underlying technologies of both conventional (analog) and digital television systems are
examined, and extensive examples are provided to explain the fundamental concepts. New developmental
efforts also are explained and the benefits of the work are outlined.
This publication is directed toward technical and engineering personnel involved in the design,
specification, installation, and maintenance of broadcast transmission systems. The basic principles of analog
and digital transmission are discussed, with emphasis on how the underlying technologies influence the
ultimate applications. Extensive references are provided to aid in understanding the material presented.
Because of the rapidly changing nature of broadcast technologies in general, and digital radio/ television in
particular, McGraw-Hill and the Editor-in-Chief have established an Internet site to support this book
(http://www.tvhandbook.com). Visitors can find news articles, revised chapters, new chapters, background
information, and links to audio/video standards organizations.
xii
Structure
In any large handbook, finding the information that a reader needs can be a major undertaking. The sheer size
of an 1000-plus page book makes finding a specific table, reference, or tutorial section a challenge. For this
reason, the Standard Handbook of Broadcast Engineering has been organized into—essentially—eleven
separate “books.” The section titles listed in the Table of Contents outline the scope of the handbook and
each section is self-contained. The section introductions include a detailed table of contents and a complete
listing of references cited in the section. It is the goal of this approach to make the book easier to use and
more useful on the job. In addition, a master subject index is provided at the end of the book.
Sources of Information
The editor has made every effort to cover the subject of broadcast technology in a comprehensive manner.
Extensive references are provided at the end of each chapter to direct readers to sources of additional
information.
Considerable detail has been included on the Advanced Television Systems Committee (ATSC) digital
television (DTV) system. These chapters are based on documents published by the ATSC, and the editor
gratefully acknowledges this contribution.
Within the limits of a practical page count, there are always more items that could be examined in greater
detail. Excellent books on the subject of broadcast engineering are available that cover areas that may not be
addressed in this handbook, notably the National Association of Broadcaster’s NAB Engineering Handbook,
9th edition, and the Proceedings of the NAB Broadcast Engineering Conference, published annually by NAB.
For more information, see the NAB Web site at http://www.nab.org.
The field of science encompassed by radio and television engineering is broad and exciting. It is an area of
growing importance to market segments of all types and—of course—to the public. It is the intent of the
Standard Handbook of Broadcast Engineering to bring these diverse concepts and technologies together in an
understandable form.
Jerry C. Whitaker
Editor-in-Chief
1_1
Section 1
Frequency Bands, Propagation, and Modulation
The usable spectrum of electromagnetic radiation frequencies extends over a range from below 100 Hz for
power distribution to 1020 Hz for the shortest X rays. The lower frequencies are used primarily for terrestrial
broadcasting and communications. The higher frequencies include visible and near-visible infrared and
ultraviolet light, and X rays. The frequencies typically of interest to RF engineers range from 30 kHz to 30
GHz.
Low Frequency (LF): 30 to 300 kHz

The LF band is used for around-the-clock communications services over long distances and where adequate
power is available to overcome high levels of atmospheric noise. Applications include:
• Radionavigation
• Fixed/maritime communications and navigation
• Aeronautical radionavigation
• Low-frequency broadcasting (Europe)
• Underwater submarine communications (up to about 30 kHz)
Medium Frequency (MF): 300 kHz to 3 MHz

The low-frequency portion of this band is used for around-the-clock communication services over moderately
long distances. The upper portion of the MF band is used principally for moderate-distance voice
communications. Applications in this band include:
• AM radio broadcasting (535.5 to 1605.5 kHz)
• Radionavigation
• Fixed/maritime communications
1_2
• Fixed and mobile commercial communications

• Amateur radio
• Standard time and frequency services
High Frequency (HF): 3 to 30 MHz

This band provides reliable medium-range coverage during daylight and, when the transmission path is in total
darkness, worldwide long-distance service. The reliability and signal quality of long-distance service depends
to a large degree upon ionospheric conditions and related long-term variations in sunspot activity affecting
skywave propagation. Applications include:
• Shortwave broadcasting
• Fixed and mobile service
• Telemetry
• Amateur radio
• Fixed/maritime mobile
• Radio astronomy
• Aeronautical fixed and mobile
Very High Frequency (VHF): 30 to 300 MHz

The VHF band is characterized by reliable transmission over medium distances. At the higher portion of the
VHF band, communication is limited by the horizon. Applications include:
• FM radio broadcasting (88 to 108 MHz)
• Low-band VHF-TV broadcasting (54 to 72 MHz and 76 to 88 MHz)
• High-band VHF-TV broadcasting (174 to 216 MHz)
• Commercial fixed and mobile radio
• Space research
• Amateur radio
• Radiolocation
Ultrahigh Frequency (UHF): 300 MHz to 3 GHz

Transmissions in this band are typically line of sight. Short wavelengths at the upper end of the band permit
the use of highly directional parabolic or multielement antennas. Applications include:
• UHF terrestrial television (470 to 806 MHz)
1_3
• Fixed and mobile communications

• Telemetry
• Meteorological aids
• Space operations
• Radio astronomy
• Radionavigation
• Satellite communications
• Point-to-point microwave relay
Superhigh Frequency (SHF): 3 to 30 GHz

Communication in this band is strictly line of sight. Very short wavelengths permit the use of parabolic
transmit and receive antennas of exceptional gain. Applications include:
• Point-to-point wideband relay
• Radar
• Specialized wideband communications
• Developmental research
• Military support systems
• Radiolocation
• Radionavigation
• Space research
In This Section:
Chapter 1.1: The Electromagnetic Spectrum 1-9

Introduction 1-9
Spectral Sub-Regions 1-10
Optical Spectrum 1-10
Visible Light Band 1-10
IR Band 1-11
UV Band 1-12
DC to Light 1-12
Microwave Band 1-12
Radio Frequency (RF) Band 1-13
Power Frequency (PF)/Telephone Band 1-13
Frequency Band Designations 1-13
Light to Gamma Rays 1-14
X Ray Band 1-15
Gamma Ray Band 1-17
1_4
Bibliography 1-18
Chapter 1.2: Propagation 1-19
Introduction 1-19
Propagation in Free Space 1-19
Transmission Loss Between Antennas in Free Space 1-21
Propagation Over Plane Earth 1-23
Field Strengths Over Plane Earth 1-23
Transmission Loss Between Antennas Over Plane Earth 1-27
Propagation Over Smooth Spherical Earth 1-28
Propagation Beyond the Line of Sight 1-30
Effects of Hills, Buildings, Vegetation, and the Atmosphere 1-33
Effects of Hills 1-33
Effects of Buildings 1-37
Effects of Trees and Other Vegetation 1-39
Effects of the Lower Atmosphere (Troposphere) 1-39
Stratification and Ducts 1-40
Tropospheric Scatter 1-41
Atmospheric Fading 1-41
Effects of the Upper Atmosphere (Ionosphere) 1-41
References 1-42
Chapter 1.3: Frequency Sources and References 1-45
Introduction 1-45
Characteristics of Crystal Devices 1-45
Frequency Stabilization 1-46
Equivalent Circuit of a Quartz Resonator 1-47
Temperature Compensation 1-48
Stress Compensation 1-49
Aging Effects 1-50
Oscillators 1-50
Key Terms 1-51
Phase-Locked Loop Synthesizers 1-53
Practical PLL Circuits 1-54
Fractional-Division Synthesizers 1-56
Multiloop Synthesizers 1-59
Direct Digital Synthesis 1-61
References 1-64
Bibliography 1-65
Chapter 1.4: Modulation Systems and Characteristics 1-67
Introduction 1-67
Amplitude Modulation 1-68
Vestigial-Sideband Amplitude Modulation 1-71
Single-Sideband Amplitude Modulation 1-71
Quadrature Amplitude Modulation (QAM) 1-74
Frequency Modulation 1-74
1_5
Modulation Index 1-74

Phase Modulation 1-78
Modifying FM Waves 1-78
Preemphasis and Deemphasis 1-78
Pulse Modulation 1-79
Digital Modulation Systems 1-79
Pulse Amplitude Modulation 1-79
Pulse Time Modulation (PTM) 1-80
Pulse Code Modulation 1-80
Delta Modulation 1-82
Digital Coding Systems 1-84
Spread Spectrum Systems 1-86
References 1-86
Bibliography 1-87
Reference Documents for this Section

Bean, B. R., and E. J. Dutton: “Radio Meteorology,” National Bureau of Standards Monograph 92, March 1, 1966.
Benson, K. B., and Jerry. C. Whitaker: Television and Audio Handbook for Technicians and Engineers, McGraw-Hill,
New York, N.Y., 1989.
Bullington, K.: “Radio Propagation at Frequencies above 30 Mc,” Proc. IRE, pg. 1122, October 1947.
Bullington, K.: “Radio Propagation Variations at VHF and UHF,” Proc. IRE, pg. 27, January 1950.
Burrows, C. R., and M. C. Gray: “The Effect of the Earth’s Curvature on Groundwave Propagation,” Proc. IRE, pg. 16,
January 1941.
Collocott, T. C., A. B. Dobson, and W. R. Chambers (eds.): Dictionary of Science & Technology.
Crutchfield, E. B. (ed.): NAB Engineering Handbook, 8th ed., National Association of Broadcasters, Washington, D.C.,
1991.
de Lisle, E. W.: “Computations of VHF and UHF Propagation for Radio Relay Applications,” RCA, Report by
International Division, New York, N.Y.
Dickson, F. H., J. J. Egli, J. W. Herbstreit, and G. S. Wickizer: “Large Reductions of VHF Transmission Loss and Fading
by the Presence of a Mountain Obstacle in Beyond-Line-of-Sight Paths,” Proc. IRE, vol. 41, no. 8, pg. 96, August 1953.
“Documents of the XVth Plenary Assembly,” CCIR Report 238, vol. 5, Geneva, 1982.
Dougherty, H. T., and E. J. Dutton: “The Role of Elevated Ducting for Radio Service and Interference Fields,” NTIA
Report 81–69, March 1981.
Dye, D. W.: Proc. Phys. Soc., Vol. 38, pp. 399–457, 1926.
1_6
Eckersley, T. L.: “Ultra-Short-Wave Refraction and Diffraction,” J. Inst. Elec. Engrs., pg. 286, March 1937.
Epstein, J., and D. Peterson: “An Experimental Study of Wave Propagation at 850 Mc,” Proc. IRE, pg. 595, May 1953.
Fink, D. G., (ed.): Television Engineering Handbook, McGraw-Hill, New York, N.Y., 1957.
Fink, D., and D. Christiansen (eds.): Electronics Engineers’ Handbook, 3rd ed., McGraw-Hill, New York, N.Y., 1989.
Frerking, M. E.: Crystal Oscillator Design and Temperature Compensation, Van Nostrand Reinhold, New York, N. Y.,
1978.
Handbook of Physics, McGraw-Hill, New York, N.Y., 1958.
Hietala, Alexander W., and Duane C. Rabe: “Latched Accumulator Fractional-N Synthesis With Residual Error
Reduction,” United States Patent, Patent No. 5,093,632, March 3, 1992.
Jordan, Edward C. (ed.): Reference Data for Engineers: Radio, Electronics, Computer and Communications, 7th ed.,
Howard W. Sams, Indianapolis, IN, 1985.
Judd, D. B., and G. Wyszecki: Color in Business, Science and Industry, 3rd ed., John Wiley and Sons, New York, N.Y.
Kaufman, Ed: IES Illumination Handbook, Illumination Engineering Society.
King, Nigel J. R.: “Phase Locked Loop Variable Frequency Generator,” United States Patent, Patent No. 4,204,174, May
20, 1980.
Kubichek, Robert: “Amplitude Modulation,” in The Electronics Handbook, Jerry C. Whitaker (ed.), CRC Press, Boca
Raton, Fla., pp. 1175–1187, 1996.
Lapedes, D. N. (ed.): The McGraw-Hill Encyclopedia of Science & Technology, 2nd ed., McGraw-Hill, New York, N.Y.
Longley, A. G., and P. L. Rice: “Prediction of Tropospheric Radio Transmission over Irregular Terrain—A Computer
Method,” ESSA (Environmental Science Services Administration), U.S. Dept. of Commerce, Report ERL (Environment
Research Laboratories) 79-ITS 67, July 1968.
McPetrie, J. S., and L. H. Ford: “An Experimental Investigation on the Propagation of Radio Waves over Bare Ridges in
the Wavelength Range 10 cm to 10 m,” J. Inst. Elec. Engrs., pt. 3, vol. 93, pg. 527, 1946.
Megaw, E. C. S.: “Some Effects of Obstacles on the Propagation of Very Short Radio Waves,” J. Inst. Elec. Engrs., pt. 3,
vol. 95, no. 34, pg. 97, March 1948.
National Bureau of Standards Circular 462, “Ionospheric Radio Propagation,” June 1948.
NIST: Manual of Regulations and Procedures for Federal Radio Frequency Management, September 1995 edition,
revisions for September 1996, January and May 1997, NTIA, Washington, D.C., 1997.
Norgard, John: “Electromagnetic Spectrum,” NAB Engineering Handbook, 9th ed., Jerry C. Whitaker (ed.), National
Association of Broadcasters, Washington, D.C., 1999.
1_7
Norgard, John: “Electromagnetic Spectrum,” The Electronics Handbook, Jerry C. Whitaker (ed.), CRC Press, Boca
Raton, Fla., 1996.
Norton, K. A.: “Ground Wave Intensity over a Finitely Conducting Spherical Earth,” Proc. IRE, pg. 622, December 1941.
Norton, K. A.: “The Propagation of Radio Waves over a Finitely Conducting Spherical Earth,” Phil. Mag., June 1938.
“Radio Wave Propagation,” Summary Technical Report of the Committee on Propagation of the National Defense
Research Committee, Academic Press, New York, N.Y., 1949.
Riley, Thomas A. D.: “Frequency Synthesizers Having Dividing Ratio Controlled Sigma-Delta Modulator,” United States
Patent, Patent No. 4,965,531, October 23, 1990.
Rohde, Ulrich L.: Digital PLL Frequency Synthesizers, Prentice-Hall, Englewood Cliffs, N.J., 1983.
Rohde, Ulrich L.: Microwave and Wireless Synthesizers: Theory and Design, John Wiley & Sons, New York, N.Y., pg.
209, 1997.
Selvidge, H.:“Diffraction Measurements at Ultra High Frequencies,” Proc. IRE, pg. 10, January 1941.
Seymour, Ken: “Frequency Modulation,” in The Electronics Handbook, Jerry C. Whitaker (ed.), CRC Press, Boca Raton,
Fla., pp. 1188–1200, 1996.
Smith, E. E., and E. W. Davis: “Wind-induced Ions Thwart TV Reception,” IEEE Spectrum, pp. 52—55, February 1981.
Stemson, A: Photometry and Radiometry for Engineers, John Wiley and Sons, New York, N.Y.
Tate, Jeffrey P., and Patricia F. Mead: “Crystal Oscillators,” in The Electronics Handbook, Jerry C. Whitaker (ed.), CRC
Press, Boca Raton, Fla., pp. 185–199, 1996.
Terman, F. E.: Radio Engineering, 3rd ed., McGraw-Hill, New York, N.Y., pg. 468, 1947
The Cambridge Encyclopedia, Cambridge University Press, 1990.
The Columbia Encyclopedia, Columbia University Press, 1993.
“The Propagation of Radio Waves through the Standard Atmosphere,” Summary Technical Report of the Committee on
Propagation, vol. 3, National Defense Research Council, Washington, D.C., 1946, published by Academic Press, New
York, N.Y.
van der Pol, Balth, and H. Bremmer: “The Diffraction of Electromagnetic Waves from an Electrical Point Source Round a
Finitely Conducting Sphere, with Applications to Radiotelegraphy and to Theory of the Rainbow,” pt. 1, Phil. Mag.,
July, 1937; pt. 2, Phil. Mag., November 1937.
Webster’s New World Encyclopedia, Prentice Hall, 1992.
Wells, John Norman: “Frequency Synthesizers,” European Patent, Patent No. 012579OB2, July 5, 1995.
Whitaker, Jerry. C.: Radio Frequency Transmission Systems: Design and Operation, McGraw-Hill, New York, N.Y.,
1991.
1_8
Wyszecki, G., and W. S. Stiles: Color Science, Concepts and Methods, Quantitative Data and Formulae, 2nd ed., John
Wiley and Sons, New York, N.Y.
Ziemer, Rodger E.: “Pulse Modulation,” in The Electronics Handbook, Jerry C. Whitaker (ed.), CRC Press, Boca Raton,
Fla., pp. 1201–1212, 1996.
g g
Section
Frequency Bands, Propagation, and

1
Modulation
The usable spectrum of electromagnetic radiation frequencies extends over a range from below
100 Hz for power distribution to 1020 Hz for the shortest X rays. The lower frequencies are used
primarily for terrestrial broadcasting and communications. The higher frequencies include visi-
ble and near-visible infrared and ultraviolet light, and X rays. The frequencies typically of inter-
est to RF engineers range from 30 kHz to 30 GHz.
Low Frequency (LF): 30 to 300 kHz

The LF band is used for around-the-clock communications services over long distances and
where adequate power is available to overcome high levels of atmospheric noise. Applications
include:
• Radionavigation
• Fixed/maritime communications and navigation
• Low-frequency broadcasting (Europe)
• Underwater submarine communications (up to about 30 kHz)
Medium Frequency (MF): 300 kHz to 3 MHz

The low-frequency portion of this band is used for around-the-clock communication services
over moderately long distances. The upper portion of the MF band is used principally for moder-
ate-distance voice communications. Applications in this band include:
• AM radio broadcasting (535.5 to 1605.5 kHz)
• Radionavigation
• Fixed/maritime communications
1-1
q y p g
1-2 Section One
• Fixed and mobile commercial communications

• Amateur radio
High Frequency (HF): 3 to 30 MHz

This band provides reliable medium-range coverage during daylight and, when the transmission
path is in total darkness, worldwide long-distance service. The reliability and signal quality of
long-distance service depends to a large degree upon ionospheric conditions and related long-
term variations in sunspot activity affecting skywave propagation. Applications include:
• Shortwave broadcasting
• Fixed and mobile service
• Telemetry
• Amateur radio
• Radio astronomy
• Aeronautical fixed and mobile
Very High Frequency (VHF): 30 to 300 MHz

The VHF band is characterized by reliable transmission over medium distances. At the higher
portion of the VHF band, communication is limited by the horizon. Applications include:
• FM radio broadcasting (88 to 108 MHz)
• Low-band VHF-TV broadcasting (54 to 72 MHz and 76 to 88 MHz)
• High-band VHF-TV broadcasting (174 to 216 MHz)
• Commercial fixed and mobile radio
• Space research
• Amateur radio
• Radiolocation
Ultrahigh Frequency (UHF): 300 MHz to 3 GHz

Transmissions in this band are typically line of sight. Short wavelengths at the upper end of the
band permit the use of highly directional parabolic or multielement antennas. Applications
include:
• UHF terrestrial television (470 to 806 MHz)
q y p g
Frequency Bands, Propagation, and Modulation 1-3
• Fixed and mobile communications

• Telemetry
• Meteorological aids
• Space operations
• Radio astronomy
• Radionavigation
• Point-to-point microwave relay
Superhigh Frequency (SHF): 3 to 30 GHz

Communication in this band is strictly line of sight. Very short wavelengths permit the use of
parabolic transmit and receive antennas of exceptional gain. Applications include:
• Point-to-point wideband relay
• Radar
• Specialized wideband communications
• Developmental research
• Military support systems
• Radiolocation
• Radionavigation
• Space research
In This Section:
Chapter 1.1: The Electromagnetic Spectrum 1-9

Introduction 1-9
Spectral Sub-Regions 1-10
Optical Spectrum 1-10
Visible Light Band 1-10
IR Band 1-11
UV Band 1-12
DC to Light 1-12
Microwave Band 1-12
Radio Frequency (RF) Band 1-13
Power Frequency (PF)/Telephone Band 1-13
Frequency Band Designations 1-13
Light to Gamma Rays 1-14
X Ray Band 1-15
Gamma Ray Band 1-17
q y p g
1-4 Section One
Bibliography 1-18
Chapter 1.2: Propagation 1-19

Introduction 1-19
Propagation in Free Space 1-19
Transmission Loss Between Antennas in Free Space 1-21
Propagation Over Plane Earth 1-23
Field Strengths Over Plane Earth 1-23
Transmission Loss Between Antennas Over Plane Earth 1-27
Propagation Over Smooth Spherical Earth 1-28
Propagation Beyond the Line of Sight 1-30
Effects of Hills, Buildings, Vegetation, and the Atmosphere 1-33
Effects of Hills 1-33
Effects of Buildings 1-37
Effects of Trees and Other Vegetation 1-39
Effects of the Lower Atmosphere (Troposphere) 1-39
Stratification and Ducts 1-40
Tropospheric Scatter 1-41
Atmospheric Fading 1-41
Effects of the Upper Atmosphere (Ionosphere) 1-41
References 1-42
Chapter 1.3: Frequency Sources and References 1-45

Introduction 1-45
Characteristics of Crystal Devices 1-45
Frequency Stabilization 1-46
Equivalent Circuit of a Quartz Resonator 1-47
Temperature Compensation 1-48
Stress Compensation 1-49
Aging Effects 1-50
Oscillators 1-50
Key Terms 1-51
Phase-Locked Loop Synthesizers 1-53
Practical PLL Circuits 1-54
Fractional-Division Synthesizers 1-56
Multiloop Synthesizers 1-59
Direct Digital Synthesis 1-61
References 1-64
Bibliography 1-65
Chapter 1.4: Modulation Systems and Characteristics 1-67

Introduction 1-67
Amplitude Modulation 1-68
Vestigial-Sideband Amplitude Modulation 1-71
Single-Sideband Amplitude Modulation 1-71
Quadrature Amplitude Modulation (QAM) 1-74
q y p g
Modulation Index 1-74

Phase Modulation 1-78
Modifying FM Waves 1-78
Preemphasis and Deemphasis 1-78
Pulse Modulation 1-79
Digital Modulation Systems 1-79
Pulse Amplitude Modulation 1-79
Pulse Time Modulation (PTM) 1-80
Pulse Code Modulation 1-80
Delta Modulation 1-82
Digital Coding Systems 1-84
Spread Spectrum Systems 1-86
References 1-86
Bibliography 1-87

Bean, B. R., and E. J. Dutton: “Radio Meteorology,” National Bureau of Standards Monograph
92, March 1, 1966.
Benson, K. B., and Jerry. C. Whitaker: Television and Audio Handbook for Technicians and
Engineers, McGraw-Hill, New York, N.Y., 1989.
Bullington, K.: “Radio Propagation at Frequencies above 30 Mc,” Proc. IRE, pg. 1122, October
1947.
Bullington, K.: “Radio Propagation Variations at VHF and UHF,” Proc. IRE, pg. 27, January
1950.
Burrows, C. R., and M. C. Gray: “The Effect of the Earth’s Curvature on Groundwave Propaga-
tion,” Proc. IRE, pg. 16, January 1941.
Crutchfield, E. B. (ed.): NAB Engineering Handbook, 8th ed., National Association of Broad-
casters, Washington, D.C., 1991.
de Lisle, E. W.: “Computations of VHF and UHF Propagation for Radio Relay Applications,”
RCA, Report by International Division, New York, N.Y.
Dickson, F. H., J. J. Egli, J. W. Herbstreit, and G. S. Wickizer: “Large Reductions of VHF Trans-
mission Loss and Fading by the Presence of a Mountain Obstacle in Beyond-Line-of-Sight
Paths,” Proc. IRE, vol. 41, no. 8, pg. 96, August 1953.
Dougherty, H. T., and E. J. Dutton: “The Role of Elevated Ducting for Radio Service and Inter-
ference Fields,” NTIA Report 81–69, March 1981.
Dye, D. W.: Proc. Phys. Soc., Vol. 38, pp. 399–457, 1926.
q y p g
1-6 Section One
Eckersley, T. L.: “Ultra-Short-Wave Refraction and Diffraction,” J. Inst. Elec. Engrs., pg. 286,
March 1937.
Epstein, J., and D. Peterson: “An Experimental Study of Wave Propagation at 850 Mc,” Proc.
IRE, pg. 595, May 1953.
Fink, D. G., (ed.): Television Engineering Handbook, McGraw-Hill, New York, N.Y., 1957.
Fink, D., and D. Christiansen (eds.): Electronics Engineers’ Handbook, 3rd ed., McGraw-Hill,
Frerking, M. E.: Crystal Oscillator Design and Temperature Compensation, Van Nostrand Rein-
hold, New York, N. Y., 1978.
Hietala, Alexander W., and Duane C. Rabe: “Latched Accumulator Fractional-N Synthesis With
Residual Error Reduction,” United States Patent, Patent No. 5,093,632, March 3, 1992.
Jordan, Edward C. (ed.): Reference Data for Engineers: Radio, Electronics, Computer and Com-
munications, 7th ed., Howard W. Sams, Indianapolis, IN, 1985.
Judd, D. B., and G. Wyszecki: Color in Business, Science and Industry, 3rd ed., John Wiley and
Sons, New York, N.Y.
King, Nigel J. R.: “Phase Locked Loop Variable Frequency Generator,” United States Patent,
Patent No. 4,204,174, May 20, 1980.
Kubichek, Robert: “Amplitude Modulation,” in The Electronics Handbook, Jerry C. Whitaker
(ed.), CRC Press, Boca Raton, Fla., pp. 1175–1187, 1996.
Lapedes, D. N. (ed.): The McGraw-Hill Encyclopedia of Science & Technology, 2nd ed.,
McGraw-Hill, New York, N.Y.
Longley, A. G., and P. L. Rice: “Prediction of Tropospheric Radio Transmission over Irregular
Terrain—A Computer Method,” ESSA (Environmental Science Services Administration),
U.S. Dept. of Commerce, Report ERL (Environment Research Laboratories) 79-ITS 67,
July 1968.
McPetrie, J. S., and L. H. Ford: “An Experimental Investigation on the Propagation of Radio
Waves over Bare Ridges in the Wavelength Range 10 cm to 10 m,” J. Inst. Elec. Engrs., pt.
3, vol. 93, pg. 527, 1946.
Megaw, E. C. S.: “Some Effects of Obstacles on the Propagation of Very Short Radio Waves,” J.
Inst. Elec. Engrs., pt. 3, vol. 95, no. 34, pg. 97, March 1948.
National Bureau of Standards Circular 462, “Ionospheric Radio Propagation,” June 1948.
NIST: Manual of Regulations and Procedures for Federal Radio Frequency Management, Sep-
tember 1995 edition, revisions for September 1996, January and May 1997, NTIA, Wash-
ington, D.C., 1997.
Norgard, John: “Electromagnetic Spectrum,” NAB Engineering Handbook, 9th ed., Jerry C.
Whitaker (ed.), National Association of Broadcasters, Washington, D.C., 1999.
q y p g
Norgard, John: “Electromagnetic Spectrum,” The Electronics Handbook, Jerry C. Whitaker

(ed.), CRC Press, Boca Raton, Fla., 1996.
Norton, K. A.: “Ground Wave Intensity over a Finitely Conducting Spherical Earth,” Proc. IRE,
pg. 622, December 1941.
Norton, K. A.: “The Propagation of Radio Waves over a Finitely Conducting Spherical Earth,”
Phil. Mag., June 1938.
“Radio Wave Propagation,” Summary Technical Report of the Committee on Propagation of the
National Defense Research Committee, Academic Press, New York, N.Y., 1949.
Riley, Thomas A. D.: “Frequency Synthesizers Having Dividing Ratio Controlled Sigma-Delta
Modulator,” United States Patent, Patent No. 4,965,531, October 23, 1990.
Rohde, Ulrich L.: Digital PLL Frequency Synthesizers, Prentice-Hall, Englewood Cliffs, N.J.,
1983.
Rohde, Ulrich L.: Microwave and Wireless Synthesizers: Theory and Design, John Wiley &
Sons, New York, N.Y., pg. 209, 1997.
Selvidge, H.:“Diffraction Measurements at Ultra High Frequencies,” Proc. IRE, pg. 10, January
1941.
Seymour, Ken: “Frequency Modulation,” in The Electronics Handbook, Jerry C. Whitaker (ed.),
CRC Press, Boca Raton, Fla., pp. 1188–1200, 1996.
Smith, E. E., and E. W. Davis: “Wind-induced Ions Thwart TV Reception,” IEEE Spectrum, pp.
52—55, February 1981.
Tate, Jeffrey P., and Patricia F. Mead: “Crystal Oscillators,” in The Electronics Handbook, Jerry
C. Whitaker (ed.), CRC Press, Boca Raton, Fla., pp. 185–199, 1996.
Terman, F. E.: Radio Engineering, 3rd ed., McGraw-Hill, New York, N.Y., pg. 468, 1947
“The Propagation of Radio Waves through the Standard Atmosphere,” Summary Technical
Report of the Committee on Propagation, vol. 3, National Defense Research Council,
Washington, D.C., 1946, published by Academic Press, New York, N.Y.
van der Pol, Balth, and H. Bremmer: “The Diffraction of Electromagnetic Waves from an Electri-
cal Point Source Round a Finitely Conducting Sphere, with Applications to Radiotelegra-
phy and to Theory of the Rainbow,” pt. 1, Phil. Mag., July, 1937; pt. 2, Phil. Mag.,
November 1937.
Wells, John Norman: “Frequency Synthesizers,” European Patent, Patent No. 012579OB2, July
5, 1995.
Whitaker, Jerry. C.: Radio Frequency Transmission Systems: Design and Operation, McGraw-
Hill, New York, N.Y., 1991.
q y p g
1-8 Section One
Wyszecki, G., and W. S. Stiles: Color Science, Concepts and Methods, Quantitative Data and
Formulae, 2nd ed., John Wiley and Sons, New York, N.Y.
Ziemer, Rodger E.: “Pulse Modulation,” in The Electronics Handbook, Jerry C. Whitaker (ed.),
g g
Chapter
1.1
The Electromagnetic Spectrum
John Norgard
1.1.1 Introduction
The electromagnetic (EM) spectrum consists of all forms of EM radiation—EM waves (radiant
energy) propagating through space, from dc to light to gamma rays. The EM spectrum can be
arranged in order of frequency and/or wavelength into a number of regions, usually wide in
extent, within which the EM waves have some specified common characteristics, such as charac-
teristics relating to the production or detection of the radiation. A common example is the spec-
trum of the radiant energy in white light, as dispersed by a prism, to produce a “rainbow” of its
constituent colors. Specific frequency ranges are often called bands; several contiguous fre-
quency bands are usually called spectrums; and sub-frequency ranges within a band are some-
times called segments.
The EM spectrum can be displayed as a function of frequency (or wavelength). In air, fre-
quency and wavelength are inversely proportional, f = c/λ (where c ≈ 3 × 108 m/s, the speed of
light in a vacuum). The MKS unit of frequency is the Hertz and the MKS unit of wavelength is
the meter. Frequency is also measured in the following sub-units:
• Kilohertz, 1 kHz = 103 Hz
• Megahertz, 1 MHz = 106 Hz
• Gigahertz, 1 GHz = 109 Hz
• Terahertz, 1 THz = 1012 Hz
• Petahertz, 1 PHz = 1015 Hz
• Exahertz, 1 EHz = 1018 Hz
Or for very high frequencies, electron volts, 1 ev ~ 2.41 × 1014 Hz. Wavelength is also measured
in the following sub-units:
• Centimeters, 1 cm = 10–2 m
• Millimeters, 1 mm = 10–3 m
• Micrometers, 1 μm = 10–6 m (microns)
• Nanometers, 1 nm = 10–9 m
1-9
g p
1-10 Frequency Bands, Propagation, and Modulation
• Ångstroms, 1 Å = 10–10 m
• Picometers, 1 pm = 10–12 m
• Femtometers, 1 fm = 10–15 m
• Attometers, 1 am = 10–18 m
1.1.2 Spectral Sub-Regions

For convenience, the overall EM spectrum can be divided into three main sub-regions:
• Optical spectrum
• DC to light spectrum
• Light to gamma ray spectrum
These main sub-regions of the EM spectrum are next discussed. Note that the boundaries
between some of the spectral regions are somewhat arbitrary. Certain spectral bands have no
sharp edges and merge into each other, while other spectral segments overlap each other slightly.
1.1.2a Optical Spectrum

The optical spectrum is the “middle” frequency/wavelength region of the EM spectrum. It is
defined here as the visible and near-visible regions of the EM spectrum and includes:
• The infrared (IR) band, circa 300 μm–0.7 μm (circa 1 THz–429 THz)
• The visible light band, 0.7 μm–0.4 μm (429 THz–750 THz)
• The ultraviolet (UV) band, 0.4 μm–circa 10 nm (750 THz–circa 30 PHz), approximately 100
ev
These regions of the EM spectrum are usually described in terms of their wavelengths.
Atomic and molecular radiation produce radiant light energy. Molecular radiation and radia-
tion from hot bodies produce EM waves in the IR band. Atomic radiation (outer shell electrons)
and radiation from arcs and sparks produce EM waves in the UV band.
Visible Light Band

In the “middle” of the optical spectrum is the visible light band, extending approximately from
0.4 μm (violet) up to 0.7 μm (red), i.e. from 750 THz (violet) down to 429 THz (red). EM radia-
tion in this region of the EM spectrum, when entering the eye, gives rise to visual sensations
(colors), according to the spectral response of the eye, which responds only to radiant energy in
the visible light band extending from the extreme long wavelength edge of red to the extreme
short wavelength edge of violet. (The spectral response of the eye is sometimes quoted as extend-
ing from 0.38 μm (violet) up to 0.75 or 0.78 μm (red); i.e., from 789 THz down to 400 or 385
THz.) This visible light band is further subdivided into the various colors of the rainbow, in
decreasing wavelength/increasing frequency:
• Red, a primary color, peak intensity at 700.0 nm (429 THz)
g p
• Orange
• Yellow
• Green, a primary color, peak intensity at 546.1 nm (549 THz)
• Cyan
• Blue, a primary color, peak intensity at 435.8 nm (688 THz)
• Indigo
• Violet
IR Band
The IR band is the region of the EM spectrum lying immediately below the visible light band.
The IR band consists of EM radiation with wavelengths extending between the longest visible
red (circa 0.7 μm) and the shortest microwaves (300 μm–1 mm); i.e., from circa 429 THz down
to 1 THz–300 GHz.
The IR band is further subdivided into the “near” (shortwave), “intermediate” (midwave), and
“far” (longwave) IR segments as follows 1:
• Near IR segment, 0.7 μm up to 3 μm (429 THz down to 100 THz)
• Intermediate IR segment, 3 μm up to 7 μm (100 THz down to 42.9 THz)
• Far IR segment, 7 μm up to 300 μm (42.9 THz down to 1 THz)
• Sub-millimeter band, 100 μm up to 1 mm (3 THz down to 300 GHz). Note that the sub-milli-
meter region of wavelengths is sometimes included in the very far region of the IR band.
EM radiation is produced by oscillating and rotating molecules and atoms. Therefore, all
objects at temperatures above absolute zero emit EM radiation by virtue of their thermal motion
(warmth) alone. Objects near room temperature emit most of their radiation in the IR band. How-
ever, even relatively cool objects emit some IR radiation; hot objects, such as incandescent fila-
ments, emit strong IR radiation.
IR radiation is sometimes incorrectly called “radiant heat” because warm bodies emit IR radi-
ation and bodies that absorb IR radiation are warmed. However, IR radiation is not itself “heat”.
This radiant energy is called “black body” radiation. Such waves are emitted by all material
objects. For example, the background cosmic radiation (2.7K) emits microwaves; room tempera-
ture objects (293K) emit IR rays; the Sun (6000K) emits yellow light; the Solar Corona (1 mil-
lion K) emits X rays.
IR astronomy uses the 1 μm to 1 mm part of the IR band to study celestial objects by their IR
emissions. IR detectors are used in night vision systems, intruder alarm systems, weather fore-
casting, and missile guidance systems. IR photography uses multilayered color film, with an IR
sensitive emulsion in the wavelengths between 700–900 nm, for medical and forensic applica-
tions, and for aerial surveying.
1. Some reference texts use 2.5 mm (120 THz) as the breakpoint between the near and the inter-
mediate IR bands, and 10 mm (30 THz) as the breakpoint between the intermediate and the
far IR bands. Also, 15 mm (20 Thz) is sometimes considered as the long wavelength end of
the far IR band.
g p
UV Band
The UV band is the region of the EM spectrum lying immediately above the visible light band.
The UV band consists of EM radiation with wavelengths extending between the shortest visible
violet (circa 0.4 μm) and the longest X rays (circa 10 nm); i.e., from 750 THz—approximately 3
ev—up to circa 30 PHz—approximately 100 ev.2
The UV band is further subdivided into the “near” and the “far” UV segments as follows:
• Near UV segment, circa 0.4 μm down to 100 nm (circa 750 THz up to 3 PHz, approximately
3 ev up to 10 ev)
• Far UV segment, 100 nm down to circa 10 nm, (3 PHz up to circa 30 PHz, approximately 10
ev up to 100 ev)
The far UV band is also referred to as the vacuum UV band, since air is opaque to all UV radia-
tion in this region.
UV radiation is produced by electron transitions in atoms and molecules, as in a mercury dis-
charge lamp. Radiation in the UV range is easily detected and can cause florescence in some
substances, and can produce photographic and ionizing effects.
In UV astronomy, the emissions of celestial bodies in the wavelength band between 50–320
nm are detected and analyzed to study the heavens. The hottest stars emit most of their radiation
in the UV band.
1.1.2b DC to Light
Below the IR band are the lower frequency (longer wavelength) regions of the EM spectrum,
subdivided generally into the following spectral bands (by frequency/wavelength):
• Microwave band, 300 GHz down to 300 MHz (1 mm up to 1 m). Some reference works define
the lower edge of the microwave spectrum at 1 GHz.
• Radio frequency band, 300 MHz down to 10 kHz (1 m up to 30 Km)
• Power/telephony band, 10 kHz down to dc (30 Km up to ∞ )
These regions of the EM spectrum are usually described in terms of their frequencies.
Radiations whose wavelengths are of the order of millimeters and centimeters are called
microwaves, and those still longer are called radio frequency (RF) waves (or Hertzian waves).
Radiation from electronic devices produces EM waves in both the microwave and RF bands.
Power frequency energy is generated by rotating machinery. Direct current (dc) is produced by
batteries or rectified alternating current (ac).
Microwave Band
The microwave band is the region of wavelengths lying between the far IR/sub-millimeter region
and the conventional RF region. The boundaries of the microwave band have not been definitely
fixed, but it is commonly regarded as the region of the EM spectrum extending from about 1 mm
up to 1 m in wavelengths, i.e. from 300 GHz down to 300 MHz. The microwave band is further
sub-divided into the following segments:
2. Some references use 4, 5, or 6 nm as the upper edge of the UV band.

g p
• Millimeter waves, 300 GHz down to 30 GHz (1 mm up to 1 cm); the Extremely High Fre-
quency band. (Some references consider the top edge of the millimeter region to stop at 100
GHz.)
• Centimeter waves, 30 GHz down to 3 GHz (1 cm up to 10 cm); the Super High Frequency
band.
The microwave band usually includes the Ultra High Frequency band from 3 GHz down to 300
MHz (from 10 cm up to 1 m). Microwaves are used in radar, space communication, terrestrial
links spanning moderate distances, as radio carrier waves in television broadcasting, for mechan-
ical heating, and cooking in microwave ovens.
Radio Frequency (RF) Band

The RF range of the EM spectrum is the wavelength band suitable for utilization in radio com-
munications extending from 10 kHz up to 300 MHz (from 30 Km down to 1 m). (Some refer-
ences consider the RF band as extending from 10 kHz to 300 GHz, with the microwave band as a
subset of the RF band from 300 MHz to 300 GHz.)
Some of the radio waves in this band serve as the carriers of low-frequency audio signals;
other radio waves are modulated by video and digital information. The amplitude modulated
(AM) broadcasting band uses waves with frequencies between 535–1705 kHz; the frequency
modulated (FM) broadcasting band uses waves with frequencies between 88–108 MHz.
In the U.S., the Federal Communications Commission (FCC) is responsible for assigning a
range of frequencies to specific services. The International Telecommunications Union (ITU)
coordinates frequency band allocation and cooperation on a worldwide basis.
Radio astronomy uses radio telescopes to receive and study radio waves naturally emitted by
objects in space. Radio waves are emitted from hot gases (thermal radiation), from charged par-
ticles spiraling in magnetic fields (synchrotron radiation), and from excited atoms and molecules
in space (spectral lines), such as the 21 cm line emitted by hydrogen gas.
Power Frequency (PF)/Telephone Band

The PF range of the EM spectrum is the wavelength band suitable for generating, transmitting,
and consuming low frequency power, extending from 10 kHz down to dc (zero frequency); i.e.,
from 30 Km up in wavelength. In the US, most power is generated at 60 Hz (some military and
computer applications use 400 Hz); in other countries, including Europe, power is generated at
50 Hz.
Frequency Band Designations

The combined microwave, RF (Hertzian Waves), and power/telephone spectra are subdivided
into the specific bands given in Table 1.1.1, which lists the international radio frequency band
designations and the numerical designations. Note that the band designated (12) has no com-
monly used name or abbreviation.
The radar band often is considered to extend from the middle of the High Frequency (7) band
to the end of the EHF (11) band. The current U.S. Tri-Service radar band designations are listed
in Table 1.1.2. An alternate and more detailed sub-division of the UHF (9), SHF (10), and EHF
(11) bands is given in Table 1.1.3. Several other frequency bands of interest (not exclusive) are
listed in Tables 1.1.4–1.1.6.
g p
Table 1.1.1 Frequency Band Designations
Description Band Designation Frequency Wavelength

Extremely Low Frequency ELF (1) Band 3 Hz up to 30 Hz 100 Mm down to 10 Mm
Super Low Frequency SLF (2) Band 30 Hz up to 300 Hz 10 Mm down to 1 Mm
Ultra Low Frequency ULF (3) Band 300 Hz up to 3 kHz 1 Mm down to 100 Km
Very Low Frequency VLF (4) Band 3 kHz up to 30 kHz 100 Km down to 10 Km
Low Frequency LF (5) Band 30 kHz up to 300 kHz 10 Km down to 1 Km
Medium Frequency MF (6) Band 300 kHz up to 3 MHz 1 Km down to 100 m
High Frequency HF (7) Band 3 MHz up to 30 MHz 100 m down to 10 m
Very High Frequency VHF (8) Band 30 MHz up to 300 MHz 10 m down to 1 m
Ultra High Frequency UHF (9) Band 300 MHz up to 3 GHz 1 m down to 10 cm
Super High Frequency SHF (10) Band 3 GHz up to 30 GHz 10 cm down to 1 cm
Extremely High Frequency EHF (11) Band 30 GHz up to 300 GHz 1 cm down to 1 mm
— (12) Band 300 GHz up to 3 THz 1 mm down to 100 μm
1.1.2c Light to Gamma Rays

Above the UV spectrum are the higher frequency (shorter wavelength) regions of the EM spec-
trum, subdivided generally into the following spectral bands (by frequency/wavelength):
• X ray band, approximately 10 ev up to 1 Mev (circa 10 nm down to circa 1 pm), circa 3 PHz
up to circa 300 EHz
• Gamma ray band, approximately 1 Kev up to ∞ (circa 300 pm down to 0 m), circa 1 EHz up
to ∞
These regions of the EM spectrum are usually described in terms of their photon energies in
electron volts. Note that the bottom of the gamma ray band overlaps the top of the X ray band.
It should be pointed out that cosmic “rays” (from astronomical sources) are not EM waves
(rays) and, therefore, are not part of the EM spectrum. Cosmic rays are high energy charged par-
ticles (electrons, protons, and ions) of extraterrestrial origin moving through space, which may
Table 1.1.2 Radar Band Designations
Band Frequency Wavelength

A Band 0 Hz up to 250 MHz ∞ down to 1.2 m
B Band 250 MHz up to 500 MHz 1.2 m down to 60 cm
C Band 500 MHz up to 1 GHz 60 cm down to 30 cm
D Band 1 GHz up to 2 GHz 30 cm down to 15 cm
E Band 2 GHz up to 3 GHz 15 cm down to 10 cm
F Band 3 GHz up to 4 GHz 10 cm down to 7.5 cm
G Band 4 GHz up to 6 GHz 7.5 cm down to 5 cm
H Band 6 GHz up to 8 GHz 5 cm down to 3.75 cm
I Band 8 GHz up to 10 GHz 3.75 cm down to 3 cm
J Band 10 GHz up to 20 GHz 3 cm down to 1.5 cm
K Band 20 GHz up to 40 GHz 1.5 cm down to 7.5 mm
L Band 40 GHz up to 60 GHz 7.5 mm down to 5 mm)
M Band 60 GHz up to 100 GHz 5 mm down to 3 mm
N Band 100 GHz up to 200 GHz 3 mm down to 1.5 mm
O Band 200 GHz up to 300 GHz 1.5 mm down to 1 mm
g p
Table 1.1.3 Detail of UHF, SHF, and EHF Band Designations
Band Frequency Wavelength

L Band 1.12 GHz up to 1.7 GHz 26.8 cm down to 17.6 cm
LS Band 1.7 GHz up to 2.6 GHz 17.6 cm down to 11.5 cm
S Band 2.6 GHz up to 3.95 GHz 11.5 cm down to 7.59 cm
C(G) Band 3.95 GHz up to 5.85 GHz 7.59 cm down to 5.13 cm
XN(J, XC) Band 5.85 GHz up to 8.2 GHz 5.13 cm down to 3.66 cm
XB(H, BL) Band 7.05 GHz up to 10 GHz 4.26 cm down to 3 cm
X Band 8.2 GHz up to 12.4 GHz 3.66 cm down to 2.42 cm
Ku(P) Band 12.4 GHz up to 18 GHz 2.42 cm down to 1.67 cm
K Band 18 GHz up to 26.5 GHz 1.67 cm down to 1.13 cm
V(R, Ka) Band 26.5 GHz up to 40 GHz 1.13 cm down to 7.5 mm
Q(V) Band 33 GHz up to 50 GHz 9.09 mm down to 6 mm
M(W) Band 50 GHz up to 75 GHz 6 mm down to 4 mm
E(Y) Band 60 GHz up to 90 GHz 5 mm down to 3.33 mm
F(N) Band 90 GHz up to 140 GHz 3.33 mm down to 2.14 mm
G(A) 140 GHz p to 220 GHz 2.14 mm down to 1.36 mm
R Band 220 GHz up to 325 GHz 1.36 mm down to 0.923 mm
have energies as high as 1020 ev. Cosmic rays have been traced to cataclysmic astrophysical/cos-
mological events, such as exploding stars and black holes. Cosmic rays are emitted by supernova
remnants, pulsars, quasars, and radio galaxies. Comic rays that collide with molecules in the
Earth’s upper atmosphere produce secondary cosmic rays and gamma rays of high energy that
also contribute to natural background radiation. These gamma rays are sometimes called cosmic
or secondary gamma rays. Cosmic rays are a useful source of high-energy particles for certain
scientific experiments.
Radiation from atomic inner shell excitations produces EM waves in the X ray band. Radia-
tion from naturally radioactive nuclei produces EM waves in the gamma ray band.
X Ray Band Table 1.1.4 Low Frequency Bands

of Interest
The X ray band is further sub-divided into the following
segments: Band Frequency
Sub-sonic band 0 Hz–10 Hz
• Soft X rays, approximately 10 ev up to 10 Kev (circa Audio band 10 Hz–10 kHz
10 nm down to 100 pm), circa 3 PHz up to 3 EHz Ultra-sonic band 10 kHz and up
Table 1.1.5 Applications of Interest in the RF Band
Band Frequency
Longwave broadcasting band 150–290 kHz
AM broadcasting band 535–1705 kHz (1.640 MHz), 107 channels, 10 kHz separation
International broadcasting band 3–30 MHz
Shortwave broadcasting band 5.95–26.1 MHz (8 bands)
VHF TV (Channels 2 - 4) 54–72 MHz
FM broadcasting band 88–108 MHz
UHF TV (Channels 14 - 69) 512–806 MHz
g p
Table 1.1.6 Applications of Interest in the Microwave Band
Application Frequency
Aero Navigation 0.96–1.215 GHz
GPS Down Link 1.2276 GHz
Military COM/Radar 1.35–1.40 GHz
Miscellaneous COM/Radar 1.40–1.71 GHz
L-Band Telemetry 1.435–1.535 GHz
GPS Down Link 1.57 GHz
Military COM (Troposcatter/Telemetry) 1.71–1.85 GHz
Commercial COM & Private LOS 1.85–2.20 GHz
Microwave Ovens 2.45 GHz
Commercial COM/Radar 2.45–2.69 GHz
Instructional TV 2.50–2.69 GHz
Military Radar (Airport Surveillance) 2.70–2.90 GHz
Maritime Navigation Radar 2.90–3.10 GHz
Miscellaneous Radars 2.90–3.70 GHz
Commercial C-Band SAT COM Down Link 3.70–4.20 GHz
Radar Altimeter 4.20–4.40 GHz
Military COM (Troposcatter) 4.40–4.99 GHz
Commercial Microwave Landing System 5.00–5.25 GHz
C-Band Weather Radar 5.35–5.47 GHz
Commercial C-Band SAT COM Up Link 5.925–6.425 GHz
Commercial COM 6.425–7.125 GHz
Mobile TV Links 6.875–7.125 GHz
Military LOS COM 7.125–7.25 GHz
Military SAT COM Down Link 7.25–7.75 GHz
Military LOS COM 7.75–7.9 GHz
Military SAT COM Up Link 7.90–8.40 GHz
Precision Approach Radar 9.00–9.20 GHz
X-Band Weather Radar (& Maritime Navigation Radar) 9.30–9.50 GHz
Police Radar 10.525 GHz
Commercial Mobile COM (LOS & ENG) 10.55–10.68 GHz
Common Carrier LOS COM 10.70–11.70 GHz
Commercial Ku-Band SAT COM Down Link 11.70–12.20 GHz
DBS Down Link & Private LOS COM 12.20–12.70 GHz
ENG & LOS COM 12.75–13.25 GHz
Miscellaneous Radars & SAT COM 13.25–14.00 GHz
Commercial Ku-Band SAT COM Up Link 14.00–14.50 GHz
Military COM (LOS, Mobile, &Tactical) 14.50–15.35 GHz
Aero Navigation 15.40–15.70 GHz
DBS Up Link 17.30–17.80 GHz
Common Carrier LOS COM 17.70–19.70 GHz
Commercial COM (SAT COM & LOS) 17.70–20.20 GHz
Private LOS COM 18.36–19.04 GHz
Military SAT COM 20.20–21.20 GHz
Miscellaneous COM 21.20–24.00 GHz
Police Radar 24.15 GHz
Navigation Radar 24.25–25.25 GHz
Military COM 25.25–27.50 GHz
Military SAT COM 30.00–31.00 GHz
g p
• Hard X rays, approximately 10 Kev up to 1Mev (100 pm down to circa 1 pm), 3 EHz up to
circa 300 EHz
Because the physical nature of these rays was at first unknown, this radiation was called “X
rays.” The designation continues to this day. The more powerful X rays are called hard X rays and
are of high frequencies and, therefore, are more energetic; less powerful X rays are called soft X
rays and have lower energies.
X rays are produced by transitions of electrons in the inner levels of excited atoms or by rapid
deceleration of charged particles (Brehmsstrahlung or breaking radiation). An important source
of X rays is synchrotron radiation. X rays can also be produced when high energy electrons from
a heated filament cathode strike the surface of a target anode (usually tungsten) between which a
high alternating voltage (approximately 100 kV) is applied.
X rays are a highly penetrating form of EM radiation and applications of X rays are based on
their short wavelengths and their ability to easily pass through matter. X rays are very useful in
crystallography for determining crystalline structure and in medicine for photographing the
body. Because different parts of the body absorb X rays to a different extent, X rays passing
through the body provide a visual image of its interior structure when striking a photographic
plate. X rays are dangerous and can destroy living tissue. They can also cause severe skin burns.
X rays are useful in the diagnosis and non-destructive testing of products for defects.
Gamma Ray Band

The gamma ray band is sub-divided into the following segments:
• Primary gamma rays, approximately 1 Kev up to 1 Mev (circa 300 pm down to 300 fm), circa
1 EHz up to 1000 EHz
• Secondary gamma rays, approximately 1 Mev up to ∞ (300 fm down to 0 m), 1000 EHz up to
∞
Secondary gamma rays are created from collisions of high energy cosmic rays with particles in
the Earth’s upper atmosphere.
The primary gamma rays are further sub-divided into the following segments:
• Soft gamma rays, approximately 1 Kev up to circa 300 Kev (circa 300 pm down to circa 3
pm), circa 1 EHz up to circa 100 EHz
• Hard gamma rays, approximately 300 Kev up to 1 Mev (circa 3 pm down to 300 fm), circa
100 EHz up to 1000 EHz
Gamma rays are essentially very energetic X rays. The distinction between the two is based on
their origin. X rays are emitted during atomic processes involving energetic electrons; gamma
rays are emitted by excited nuclei or other processes involving sub-atomic particles.
Gamma rays are emitted by the nucleus of radioactive material during the process of natural
radioactive decay as a result of transitions from high energy excited states to low energy states in
atomic nuclei. Cobalt 90 is a common gamma ray source (with a half-life of 5.26 years). Gamma
rays are also produced by the interaction of high energy electrons with matter. “Cosmic” gamma
rays cannot penetrate the Earth’s atmosphere.
Applications of gamma rays are found both in medicine and in industry. In medicine, gamma
rays are used for cancer treatment and diagnoses. Gamma ray emitting radioisotopes are used as
tracers. In industry, gamma rays are used in the inspection of castings, seams, and welds.
g p
1.1.3 Bibliography
Judd, D. B., and G. Wyszecki: Color in Business, Science and Industry, 3rd ed., John Wiley and
Sons, New York, N.Y.
Lapedes, D. N. (ed.): The McGraw-Hill Encyclopedia of Science & Technology, 2nd ed.,
McGraw-Hill, New York, N.Y.
Norgard, John: “Electromagnetic Spectrum,” NAB Engineering Handbook, 9th ed., Jerry C. Whi-
taker (ed.), National Association of Broadcasters, Washington, D.C., 1999.
Norgard, John: “Electromagnetic Spectrum,” The Electronics Handbook, Jerry C. Whitaker
(ed.), CRC Press, Boca Raton, Fla., 1996.
Wyszecki, G., and W. S. Stiles: Color Science, Concepts and Methods, Quantitative Data and
Formulae, 2nd ed., John Wiley and Sons, New York, N.Y.
g g
Chapter
1.2
Propagation
William Daniel, Edward W. Allen, Donald G. Fink
1.2.1 Introduction
The portion of the electromagnetic spectrum commonly used for radio transmissions lies
between approximately 10 kHz and 40 GHz. The influence on radio waves of the medium
through which they propagate is frequency-dependent. The lower frequencies are greatly influ-
enced by the characteristics of the earth’s surface and the ionosphere, while the highest frequen-
cies are greatly affected by the atmosphere, especially rain. There are no clear-cut boundaries
between frequency ranges but instead considerable overlap in propagation modes and effects of
the path medium.
In the U.S., those frequencies allocated for broadcast-related use include the following:
• 550–1640 kHz: AM radio
• 54–72 MHz: TV channels 2–4
• 88–108 MHz: FM radio
• 0.9–12.2 GHz: nonexclusive TV terrestrial and satellite ancillary services
• 12.2–12.7 GHz: direct satellite broadcasting
• 12.7–40 GHz: nonexclusive direct satellite broadcasting
1.2.2 Propagation in Free Space

For simplicity and ease of explanation, propagation in space and under certain conditions involv-
ing simple geometry, in which the wave fronts remain coherent, may be treated as ray propaga-
tion. It. should be kept in mind that this assumption may not hold in the presence of obstructions,
surface roughness, and other conditions which are often encountered in practice.
1-19
p g
For the simplest case of propagation in space, namely that of uniform radiation in all direc-
tions from a point source, or isotropic radiator, it is useful to consider the analogy to a point
source of light, The radiant energy passes with uniform intensity through all portions of an imag-
inary spherical surface located at a radius r from the source. The area of such a surface is 4 π r 2
and the power flow per unit area W = Pt /4 π r 2, where Pt is the total power radiated by the
source and W is represented in W/m2. In the engineering of broadcasting and of some other radio
services, it is conventional to measure the intensity of radiation in terms of the strength of the
electric field Eo rather than in terms of power density W. The power density is equal to the square
of the field strength divided by the impedance of the medium, so for free space
E 2o
W = ------------ (1.2.1)
120 π
and
4π r2E2 o
Pt = ------------------ (1.2.2)
120 π
or
r 2 E o2
Pt = ----------- (1.2.3)
30
Where:
Pt = watts radiated
E o = the free space field in volts per meter
r = the radius in meters
A more conventional and useful form of this equation, which applies also to antennas other
than isotropic radiators, is
30 gP t t
Eo = -------------------- (1.2.4)
r
where gt is the power gain of the antenna in the pertinent direction compared to an isotropic radi-
ator.
An isotropic antenna is useful as a reference for specifying the radiation patterns for more
complex antennas but does not in fact exist. The simplest forms of practical antennas are the
electric doublet and the magnetic doublet, the former a straight conductor that is short compared
with the wavelength and the latter a conducting loop of short radius compared with the wave-
length. For the doublet radiator, the gain is 1.5 and the field strength in the equatorial plane is
p g
Propagation 1-21
45 P t
Eo = --------------- (1.2.5)
r
For a half-wave dipole, namely, a straight conductor one-half wave in length, the power gain is
1.64 and
7 P t
Eo = ------------ (1.2.6)
r
From the foregoing equations it can be seen that for free space:
• The radiation intensity in watts per square meter is proportional to the radiated power and
inversely proportional to the square of the radius or distance from the radiator.
• The electric field strength is proportional to the square root of the radiated power and
inversely proportional to the distance from the radiator.
1.2.2a Transmission Loss Between Antennas in Free Space

The maximum useful power P r that can be delivered to a matched receiver is given by [1]
Eλ 2 g
P r = ⎛⎝ -------⎞⎠ ------r-- W (1.2.7)
2π 120
Where:
E = received field strength in volts per meter
λ = wavelength in meters, 300/F
F = frequency in MHz
gr = receiving antenna power gain over an isotropic radiator
This relationship between received power and the received field strength is shown by scales 2,
3, and 4 in Figure 1.2.1 for a half-wave dipole. For example, the maximum useful power at 100
MHz that can be delivered by a half-wave dipole in a field of 50 dB above 1 μV/m is 95 dB
below 1 W.
A general relation for the ratio of the received power to the radiated power obtained from
Equations (1.2.4) and (1.2.7) is
Pr ⎛ λ ⎞ 2 ⎛ E ⎞ 2
----- = -------- g g ----- (1.2.8)
P t ⎝ 4 π r⎠ t r ⎝ Eo ⎠
When both antennas are half-wave dipoles, the power-transfer ratio is
P r ⎛ 1.64 λ⎞ 2 ⎛ E ⎞ 2 ⎛ 0.13 λ⎞ 2 ⎛ E ⎞ 2
----- = ------------- ----- = ------------- ----- (1.2.9)
P t ⎝ 4 πr ⎠ ⎝ Eo ⎠ ⎝ r ⎠ ⎝ Eo ⎠
p g
Figure 1.2.1 Free-space field intensity and received power between half-wave dipoles. (From [2].
Used with permission.)
p g
Propagation 1-23
and is shown on scales 1 to 4 of Figure 1.2.2. For free-space transmission, E/Eo = 1.

When the antennas are horns, paraboloids, or multielement arrays, a more convenient expres-
sion for the ratio of the received power to the radiated power is given by
P r
B B ⎛ E ⎞2
t r
----- = ------------- ----- (1.2.10)
( λ r) E
P 2⎝ ⎠
t o
where Bt and Br are the effective areas of the transmitting and receiving antennas, respectively.
This relation is obtained from Equation (1.2.8) by substituting as follows
g = -4---π--2-B--- (1.2.11)
λ
This is shown in Figure 1.2.2 for free-space transmission when Bt = B r. For example, the free-
space loss at 4000 MHz between two antennas of 10 ft 2 (0.93 m2) effective area is about 72 dB
for a distance of 30 mi (48 km).
1.2.3 Propagation Over Plane Earth

The presence of the ground modifies the generation and propagation of radio waves so that the
received field strength is ordinarily different than would be expected in free space [3, 4]. The
ground acts as a partial reflector and as a partial absorber, and both of these properties affect the
distribution of energy in the region above the earth.
1.2.3a Field Strengths Over Plane Earth

The geometry of the simple case of propagation between two antennas each placed several wave-
lengths above a plane earth is shown in Figure 1.2.3. For isotropic antennas, for simple magnetic-
doublet antennas with vertical polarization, or for simple electric-doublet antennas with horizon-
tal polarization the resultant received field is [4, 5]
E d E d Re Δ j
E = --------- + --------------------- = E ( cos θ1 + R cos θ2 e Δ)

o o j
(1.2.12)
r1 r2 o
For simple magnetic-doublet antennas with horizontal polarization or electric-doublet anten-

nas with vertical polarization at both the transmitter and receiver, it is necessary to correct for the
cosine radiation and absorption patterns in the plane of propagation. The received field is
E = E ( cos3θ1 + R cos3θ2 e Δ)
o
j
(1.2.13)
Where:
p g
Figure 1.2.2 Received power in free space between two antennas of equal effective areas. (From
[2]. Used with permission.)
E o = the free-space field at distance d in the equatorial plane of the doublet

R = the complex reflection coefficient of the earth
j = the square root of –1
e jΔ = cos Δ + j sin Δ
Δ = the phase difference between the direct wave received over path r1 and the ground-reflected
wave received over path r2, which is due to the difference in path lengths
p g
Propagation 1-25
Figure 1.2.3 Ray paths for antennas above plane earth. (From [2]. Used with permission.)
For distances such that θ is small and the differences between d and r1 and r2 can be
neglected, Equations (1.2.12) and (1.2.13) become
E = E ( 1 + Re Δ)
o
j
(1.2.14)
When the angle θ is very small, R is approximately equal to –1. For the case of two antennas,
one or both of which may be relatively close to the earth, a surface-wave term must be added and
Equation (1.2.14) becomes [3, 6]
E = E [ 1 + Re Δ + ( 1 – R ) Ae Δ]
o
j j
(1.2.15)
The quantity A is the surface-wave attenuation factor, which depends upon the frequency,
ground constants, and type of polarization. It is never greater than unity and decreases with
increasing distance and frequency, as indicated by the following approximate equation [1]
–1
A≅ ------------------------------------------------------- (1.2.16)
1+ j⎝ d
⎛ -2---π-----⎞ ( sin θ + ) 2 z
λ ⎠
This approximate expression is sufficiently accurate as long as A < 0.1, and it gives the magni-
tude of A within about 2 dB for all values of A. However, as A approaches unity, the error in
phase approaches 180°. More accurate values are given by Norton [3] where, in his nomencla-
ture, A = f (P,B) eiφ.
The equation (1.2.15) for the absolute value of field strength has been developed from the
successive consideration of the various components that make up the ground wave, but the fol-
lowing equivalent expressions may be found more convenient for rapid calculation
p g
⎧ Δ ⎫
E = Eo ⎨2 sin + j ⎡⎣(1 + R) + (1 − R) A⎤⎦ e jΔ 2 ⎬
⎩ 2 ⎭
(1.2.17)
When the distance d between antennas is greater than about five times the sum of the two
antenna heights ht and hr, the phase difference angle Δ (rad) is
4 π h t hr
Δ = (1.2.18)
λd
-------------
---
Also, when the angle Δ is greater than about 0.5 rad, the terms inside the brackets of Equation
(1.2.17)—which include the surface wave—are usually negligible, and a sufficiently accurate
expression is given by
⎛ 2πht hr ⎞
E = Eo ⎜ 2 sin
⎝ λd ⎟⎠
(1.2.19)
In this case, the principal effect of the ground is to produce interference fringes or lobes, so that
the field strength oscillates about the free-space field as the distance between antennas or the
height of either antenna is varied.
When the angle Δ is less than about 0.5 rad, there is a region in which the surface wave may
be important but not controlling. In this region, sin Δ/2 is approximately equal to Δ/2 and
4 π h t hr' '
E = Eo (1.2.20)
λd
---
--------------
-
In this equation h′ = h + jho, where h is the actual antenna height and ho = λ/2πz has been desig-
nated as the minimum effective antenna height. The magnitude of the minimum effective height
ho is shown in Figure 1.2.4 for seawater and for “good” and “poor” soil. “Good” soil corresponds
roughly to clay, loam, marsh, or swamp, while “poor” soil means rocky or sandy ground [1].
The surface wave is controlling for antenna heights less than the minimum effective height,
and in this region the received field or power is not affected appreciably by changes in the
antenna height. For antenna heights that are greater than the minimum effective height, the
received field or power is increased approximately 6 dB every time the antenna height is dou-
bled, until free-space transmission is reached. It is ordinarily sufficiently accurate to assume that
h′ is equal to the actual antenna height or the minimum effective antenna height, whichever is
the larger.
When translated into terms of antenna heights in feet, distance in miles, effective power in
kilowatts radiated from a half-wave dipole, and frequency F in megahertz, Equation (1.2.20)
becomes the following very useful formula for the rapid calculation of approximate values of
field strength for purposes of prediction or for comparison with measured values
p g
Propagation 1-27
Figure 1.2.4 Minimum effective antenna height. (From [2]. Used with permission.)
ht′ hr′ Pt
E≅F
3d2
(1.2.21)
1.2.3b Transmission Loss Between Antennas Over Plane Earth

The ratio of the received power to the radiated power for transmission over plane earth is
obtained by substituting Equation (1.2.20) into (1.2.8), resulting in
p g
2 2
Pr ⎛ λ ⎞ ⎛ 4 π ht′ hr′ ⎞ ⎛ ht′ hr′ ⎞
= gt g r ⎜ = gt g r
Pt ⎜⎝ 4 π d ⎟⎠ ⎝ λ d ⎟⎠ ⎜⎝ d 2 ⎟⎠
(1.2.22)
This relationship is independent of frequency, and is shown on Figure 1.2.5 for half-wave
dipoles (gt = gr = 1.64). A line through the two scales of antenna height determines a point on
the unlabeled scale between them, and a second line through this point and the distance scale
determines the received power for 1 W radiated. When the received field strength is desired, the
power indicated on Figure 1.2.5 can be transferred to scale 4 of Figure 1.2.1, and a line through
the frequency on scale 3 indicates the received field strength on scale 2. The results shown on
Figure 1.2.5 are valid as long as the value of received power indicated is lower than that shown
on Figure 1.2.3 for free-space transmission. When this condition is not met, it means that the
angle Δ is too large for Equation (1.2.20) to be accurate and that the received field strength or
power oscillates around the free-space value as indicated by Equation (1.2.19) [1].
1.2.3c Propagation Over Smooth Spherical Earth

The curvature of the earth has three effects on the propagation of radio waves at points within the
line of sight:
• The reflection coefficient of the ground-reflected wave differs for the curved surface of the
earth from that for a plane surface. This effect is of little importance, however, under the cir-
cumstances normally encountered in practice.
• Because the ground-reflected wave is reflected against the curved surface of the earth, its
energy diverges more than would be indicated by the inverse distance-squared law, and the
ground-reflected wave must be multiplied by a divergence factor D.
• The heights of the transmitting and receiving antennas ht′ and hr′ , above the plane that is tan-
gent to the surface of the earth at the point of reflection of the ground-reflected wave, are less
than the antenna heights ht and hr above the surface of the earth, as shown in Figure 1.2.6.
Under these conditions, Equation (1.2.14), which applies to larger distances within the line of
sight and to antennas of sufficient height that the surface component may be neglected, becomes
E = E ( + DR
o
1 ' e Δ)
j
(1.2.23)
Similar substitutions of the values that correspond in Figures 1.2.3 and 1.2.6 can be made in
Equations (1.2.15 through (1.2.22). However, under practical conditions, it is generally satisfac-
tory to use the plane-earth formulas for the purpose of calculating smooth-earth values. An
exception to this is usually made in the preparation of standard reference curves, which are gen-
erally calculated by the use of the more exact formulas [1, 4–9].
p g
Propagation 1-29
Figure 1.2.5 Received power over plane earth between half-wave dipoles. Notes: (1) This chart is
not valid when the indicated received power is greater than the free space power shown in Figure
1.2.1. (2) Use the actual antenna height or the minimum effective height shown in Figure 1.2.4,
whichever is the larger. (From [2]. Used with permission.)
p g
Figure 1.2.6 Ray paths for antennas above spherical earth. (From [2]. Used with permission.)
1.2.3d Propagation Beyond the Line of Sight

Radio waves are bent around the earth by the phenomenon of diffraction, with the ease of bend-
ing decreasing as the frequency increases. Diffraction is a fundamental property of wave motion,
and in optics it is the correction to apply to geometrical optics (ray theory) to obtain the more
accurate wave optics. In wave optics, each point on the wave front is considered to act as a radi-
ating source. When the wave front is coherent or undisturbed, the resultant is a progression of the
front in a direction perpendicular thereto, along a path that constitutes the ray. When the front is
disturbed, the resultant front can be changed in both magnitude and direction with resulting
attenuation and bending of the ray. Thus, all shadows are somewhat “fuzzy” on the edges and the
transition from “light” to “dark” areas is gradual, rather than infinitely sharp.
The effect of diffraction around the earth’s curvature is to make possible transmission beyond
the line of sight, with somewhat greater loss than is incurred in free space or over plane earth.
The magnitude of this loss increases as either the distance or the frequency is increased and it
depends to some extent on the antenna height.
The calculation of the field strength to be expected at any particular point in space beyond the
line of sight around a spherical earth is rather complex, so that individual calculations are seldom
made except with specially designed software. Rather, nomograms or families of curves are usu-
ally prepared for general application to large numbers of cases. The original wave equations of
Van der Pol and Bremmer [6] have been modified by Burrows [7] and by Norton [3, 5] so as to
make them more readily usable and particularly adaptable to the production of families of
curves. Such curves have been prepared by a variety of organizations. These curves have not
been included herein, in view of the large number of curves that are required to satisfy the possi-
p g
Propagation 1-31
Figure 1.2.7 Loss beyond line of sight in decibels. (From [2]. Used with permission.)
ble variations in frequency, electrical characteristics of the earth, polarization, and antenna
height. Also, the values of field strength indicated by smooth-earth curves are subject to consid-
erable modification under actual conditions found in practice. For VHF and UHF broadcast pur-
poses, the smooth-earth curves have been to a great extent superseded by curves modified to
reflect average conditions of terrain.
Figure 1.2.7 is a nomogram to determine the additional loss caused by the curvature of the
earth [1]. This loss must be added to the free-space loss found from Figure 1.2.1. A scale is
included to provide for the effect of changes in the effective radius of the earth, caused by atmo-
spheric refraction. Figure 1.2.7 gives the loss relative to free space as a function of three dis-
p g
Figure 1.2.8 Distance to the horizon. (From [2]. Used with permission.)
tances; d1 is the distance to the horizon from the lower antenna, d2 is the distance to the horizon
from the higher antenna, and d3 is the distance between the horizons. The total distance between
antennas is d = d1 + d2 + d3.
p g
Propagation 1-33
Figure 1.2.9 Ray paths for antennas over rough terrain. (From [2]. Used with permission.)
The horizon distances d1 and d2 for the respective antenna heights h1 and h2 and for any
assumed value of the earth’s radius factor k can be determined from Figure 1.2.8 [1].
1.2.3e Effects of Hills, Buildings, Vegetation, and the Atmosphere

The preceding discussion assumes that the earth is a perfectly smooth sphere with a uniform or a
simple atmosphere, for which condition calculations of expected field strengths or transmission
losses can be computed for the regions within the line of sight and regions well beyond the line
of sight, and interpolations can be made for intermediate distances. The presence of hills, build-
ings, and trees has such complex effects on propagation that it is impossible to compute in detail
the field strengths to be expected at discrete points in the immediate vicinity of such obstructions
or even the median values over very small areas. However, by the examination of the earth pro-
file over the path of propagation and by the use of certain simplifying assumptions, predictions
that are more accurate than smooth-earth calculations can be made of the median values to be
expected over areas representative of the gross features of terrain.
Effects of Hills
The profile of the earth between the transmitting and receiving points is taken from available
topographic maps and is plotted on a chart that provides for average air refraction by the use of a
4/3 earth radius, as shown in Figure 1.2.9. The vertical scale is greatly exaggerated for conve-
nience in displaying significant angles and path differences. Under these conditions, vertical
dimensions are measured along vertical parallel lines rather than along radii normal to the curved
surface, and the propagation paths appear as straight lines. The field to be expected at a low
receiving antenna at A from a high transmitting antenna at B can be predicted by plane-earth
methods, by drawing a tangent to the profile at the point at which reflection appears to occur
with equal incident and reflection angles. The heights of the transmitting and receiving antennas
above the tangent are used in conjunction with Figure 1.2.5 to compute the transmission loss, or
with Equation (1.2.21) to compute the field strength. A similar procedure can be used for more
p g
(a)
(b)
(c)
(d)
Figure 1.2.10 Ray paths for antennas behind hills: (a–d), see text. (From [2]. Used with permis-
sion.)
distantly spaced high antennas when the line of sight does not clear the profile by at least the first
Fresnel zone [10].
Propagation over a sharp ridge, or over a hill when both the transmitting and receiving
antenna locations are distant from the hill, may be treated as diffraction over a knife edge, shown
schematically in Figure 1.2.10a [1, 9–14]. The height of the obstruction H is measured from the
line joining the centers of the two antennas to the top of the ridge. As shown in Figure 1.2.11, the
shadow loss approaches 6 dB as H approaches 0—grazing incidence—and it increases with
increasing positive values of H. When the direct ray clears the obstruction, H is negative, and the
shadow loss approaches 0 dB in an oscillatory manner as the clearance is increased. Thus, a sub-
stantial clearance is required over line-of-sight paths in order to obtain free-space transmission.
There is an optimum clearance, called the first Fresnel-zone clearance, for which the transmis-
sion is theoretically 1.2 dB better than in free space. Physically, this clearance is of such magni-
p g
Propagation 1-35
Note: When accuracy greater than ±1.5

dB is required, values on
the d1 scale should be:
2
d1
1+ d1 d2
Figure 1.2.11 Shadow loss relative to free space. (From [2]. Used with permission.)
tude that the phase shift along a line from the antenna to the top of the obstruction and from there
to the second antenna is about one-half wavelength greater than the phase shift of the direct path
between antennas.
The locations of the first three Fresnel zones are indicated on the right-hand scale on Figure
1.2.11, and by means of this chart the required clearances can be obtained. At 3000 MHz, for
example, the direct ray should clear all obstructions in the center of a 40 mi (64 km) path by
about 120 ft (36 m) to obtain full first-zone clearance, as shown at “C” in Figure 1.2.9. The cor-
responding clearance for a ridge 100 ft (30 m) in front of either antenna is 4 ft (1.2 m). The locus
p g
of all points that satisfy this condition for all distances is an ellipsoid of revolution with foci at
the two antennas.
When there are two or more knife-edge obstructions or hills between the transmitting and
receiving antennas, an equivalent knife edge can be represented by drawing a line from each
antenna through the top of the peak that blocks the line of sight, as in Figure 1.2.10b.
Alternatively, the transmission loss can be computed by adding the losses incurred when
passing over each of the successive hills, as in Figure 1.2.10c. The height H1 is measured from
the top of hill 1 to the line connecting antenna 1 and the top of hill 2. Similarly, H2 is measured
from the top of hill 2 to the line connecting antenna 2 and the top of hill 1. The nomogram given
in Figure 1.2.11 is used for calculating the losses for terrain conditions represented by Figure
1.2.10a–c.
This procedure applies to conditions for which the earth-reflected wave can be neglected,
such as the presence of rough earth, trees, or structures at locations along the profile at points
where earth reflection would otherwise take place at the frequency under consideration; or where
first Fresnel-zone clearance is obtained in the foreground of each antenna and the geometry is
such that reflected components do not contribute to the field within the first Fresnel zone above
the obstruction. If conditions are favorable to earth reflection, the base line of the diffraction tri-
angle should not be drawn through the antennas, but through the points of earth reflection, as in
Figure 1.2.10d. H is measured vertically from this base line to the top of the hill, while d1 and d2
are measured to the antennas as before. In this case, Figure 1.2.12 is used to estimate the shadow
loss to be added to the plane-earth attenuation [1].
Under conditions where the earth-reflected components reinforce the direct components at
the transmitting and receiving antenna locations, paths may be found for which the transmission
loss over an obstacle is less than the loss over spherical earth. This effect may be useful in estab-
lishing VHF relay circuits where line-of-sight operation is not practical. Little utility, however,
can be expected for mobile or broadcast services [14].
An alternative method for predicting the median value for all measurements in a completely
shadowed area is as follows [15]:
1. The roughness of the terrain is assumed to be represented by height H, shown on the profile
at the top of Figure 1.2.13.
2. This height is the difference in elevation between the bottom of the valley and the elevation
necessary to obtain line of sight with the transmitting antenna.
3. The difference between the measured value of field intensity and the value to be expected
over plane earth is computed for each point of measurement within the shadowed area.
4. The median value for each of several such locations is plotted as a function of sq. rt. (H/λ).
These empirical relationships are summarized in the nomogram shown in Figure 1.2.13. The
scales on the right-hand line indicate the median value of shadow loss, compared with plane-
earth values, and the difference in shadow loss to be expected between the median and the 90
percent values. For example, with variations in terrain of 500 ft (150 m), the estimated median
shadow loss at 4500 MHz is about 20 dB and the shadow loss exceeded in 90 percent of the pos-
sible locations is about 20 + 15 = 35 dB. This analysis is based on large-scale variations in field
intensity, and does not include the standing-wave effects that sometimes cause the field intensity
to vary considerably within a matter of a few feet.
p g
Propagation 1-37
Note: When accuracy greater than ±1.5

dB is required, values on
the d1 scale should be: 2
d1
1+ d1 d2
Figure 1.2.12 Shadow loss relative to plane earth. (From [2]. Used with permission.)
Effects of Buildings
Built-up areas have little effect on radio transmission at frequencies below a few megahertz,
since the size of any obstruction is usually small compared with the wavelength, and the shadows
caused by steel buildings and bridges are not noticeable except immediately behind these
obstructions. However, at 30 MHz and above, the absorption of a radio wave in going through an
obstruction and the shadow loss in going over it are not negligible, and both types of losses tend
to increase as the frequency increases. The attenuation through a brick wall, for example, can
p g
Figure 1.2.13 Estimated distribution of shadow loss for random locations (referred to plane-earth
values). (From [2]. Used with permission.)
vary from 2 to 5 dB at 30 MHz and from 10 to 40 dB at 3000 MHz, depending on whether the
wall is dry or wet. Consequently, most buildings are rather opaque at frequencies of the order of
thousands of megahertz.
For radio-relay purposes, it is the usual practice to select clear sites; but where this is not fea-
sible the expected fields behind large buildings can be predicted by the preceding diffraction
methods. In the engineering of mobile- and broadcast-radio systems it has not been found practi-
cal in general to relate measurements made in built-up areas to the particular geometry of build-
ings, so that it is conventional to treat them statistically. However, measurements have been
divided according to general categories into which buildings can readily be classified, namely,
the tall buildings typical of the centers of cities on the one hand, and typical two-story residential
areas on the other.
p g
Propagation 1-39
Buildings are more transparent to radio waves than the solid earth, and there is ordinarily
much more backscatter in the city than in the open country. Both of these factors tend to reduce
the shadow losses caused by the buildings. On the other hand, the angles of diffraction over or
around the buildings are usually greater than for natural terrain, and this factor tends to increase
the loss resulting from the presence of buildings. Quantitative data on the effects of buildings
indicate that in the range of 40 to 450 MHz there is no significant change with frequency, or at
least the variation with frequency is somewhat less than the square-root relationship noted in the
case of hills. The median field strength at street level for random locations in New York City is
about 25 dB below the corresponding plane-earth value. The corresponding values for the 10
percent and 90 percent points are about –15 and –35 dB, respectively [1, 15]. Measurements in
congested residential areas indicate somewhat less attenuation than among large buildings.
Effects of Trees and Other Vegetation

When an antenna is surrounded by moderately thick trees and below treetop level, the average
loss at 30 MHz resulting from the trees is usually 2 or 3 dB for vertical polarization and negligi-
ble with horizontal polarization. However, large and rapid variations in the received field
strength can exist within a small area, resulting from the standing-wave pattern set up by reflec-
tions from trees located at a distance of as much as 100 ft (30 m) or more from the antenna. Con-
sequently, several nearby locations should be investigated for best results. At 100 MHz, the
average loss from surrounding trees may be 5 to 10 dB for vertical polarization and 2 or 3 dB for
horizontal polarization. The tree losses continue to increase as the frequency increases, and
above 300 to 500 MHz they tend to be independent of the type of polarization. Above 1000
MHz, trees that are thick enough to block vision present an almost solid obstruction, and the dif-
fraction loss over or around these obstructions can be obtained from Figures 1.2.9 or 1.2.11.
There is a pronounced seasonal effect in the case of deciduous trees, with less shadowing and
absorption in the winter months when the leaves have fallen. However, when the path of travel
through the trees is sufficiently long that it is obscured, losses of the above magnitudes can be
incurred, and the principal mode of propagation may be by diffraction over the trees.
When the antenna is raised above trees and other forms of vegetation, the prediction of field
strengths again depends upon the proper estimation of the height of the antenna above the areas
of reflection and of the applicable reflection coefficients. For growth of fairly uniform height and
for angles near grazing incidence, reflection coefficients will approach –1 at frequencies near 30
MHz. As indicated by Rayleigh’s criterion of roughness, the apparent roughness for given condi-
tions of geometry increases with frequency so that near 1000 MHz even such low and relatively
uniform growth as farm crops or tall grass may have reflection coefficients of about –0.3 for
small angles of reflection [17].
The distribution of losses in the immediate vicinity of trees does not follow normal probabil-
ity law but is more accurately represented by Rayleigh’s law, which is the distribution of the sum
of a large number of equal vectors having random phases.
1.2.3f Effects of the Lower Atmosphere (Troposphere)

Radio waves propagating through the lower atmosphere, or troposphere, are subject to absorp-
tion, scattering, and bending. Absorption is negligible in the VHF–UHF frequency range but
becomes significant at frequencies above 10 GHz. The index of refraction of the atmosphere, n,
is slightly greater than 1 and varies with temperature, pressure, and water vapor pressure, and
p g
therefore with height, climate, and local meteorological conditions. An exponential model show-
ing a decrease with height to 37 to 43 mi (60 to 70 kin) is generally accepted [18, 19]. For this
model, variation of n is approximately linear for the first kilometer above the surface in which
most of the effect on radio waves traveling horizontally occurs. For average conditions, the effect
of the atmosphere can be included in the expression of earth diffraction around the smooth earth
without discarding the useful concept of straight-line propagation by multiplying the actual
earth’s radius by k to obtain an effective earth’s radius, where
1
k=
1+ a ( dn dh)
(1.2.24)
Where:
a = the actual radius of the earth
dn/dh = the rate of change of the refractive index with height
Through the use of average annual values of the refractive index gradient, k is found to be 4/3 for
temperate climates.
Stratification and Ducts

As a result of climatological and weather processes such as subsidence, advection, and surface
heating and radiative cooling, the lower atmosphere tends to be stratified in layers with contrast-
ing refractivity gradients [20]. For convenience in evaluating the effect of this stratification,
radio refractivity N is defined as N = (n – 1) × 106 and can be derived from
P e
N = 77.6 + 3.73×105 2
T T
(1.2.25)
Where:
P = atmospheric pressure, mbar
T = absolute temperature, K
e = water vapor pressure, mbar
When the gradient of N is equal to –39 N-units per kilometer, normal propagation takes place,
corresponding to the effective earth’s radius ka, where k = 4/3.
When dN/dh is less than –39 N-units per kilometer, subrefraction occurs and the radio wave is
bent strongly downward.
When dN/dh is less than –157 N-units per kilometer, the radio energy may be bent downward
sufficiently to be reflected from the earth, after which the ray is again bent toward the earth, and
so on. The radio energy thus is trapped in a duct or waveguide. The wave also may be trapped
between two elevated layers, in which case energy is not lost at the ground reflection points and
even greater enhancement occurs. Radio waves thus trapped or ducted can produce fields
exceeding those for free-space propagation because the spread of energy in the vertical direction
is eliminated as opposed to the free-space case, where the energy spreads out in two directions
p g
Propagation 1-41
orthogonal to the direction of propagation. Ducting is responsible for abnormally high fields
beyond the radio horizon. These enhanced fields occur for significant periods of time on overwa-
ter paths in areas where meteorological conditions are favorable. Such conditions exist for signif-
icant periods of time and over significant horizontal extent in the coastal areas of southern
California and around the Gulf of Mexico. Over land, the effect is less pronounced because sur-
face features of the earth tend to limit the horizontal dimension of ducting layers [20].
Tropospheric Scatter
The most consistent long-term mode of propagation beyond the radio horizon is that of scatter-
ing by small-scale fluctuations in the refractive index resulting from turbulence. Energy is scat-
tered from multitudinous irregularities in the common volume which consists of that portion of
troposphere visible to both the transmitting and receiving sites. There are some empirical data
that show a correlation between the variations in the field beyond the horizon and ΔN, the differ-
ence between the reflectivity on the ground and at a height of 1 km [21]. Procedures have been
developed for calculating scatter fields for beyond-the-horizon radio relay systems as a function
of frequency and distance [22, 23]. These procedures, however, require detailed knowledge of
path configuration and climate.
The effect of scatter propagation is incorporated in the statistical evaluation of propagation
(considered previously in this chapter), where the attenuation of fields beyond the diffraction
zone is based on empirical data and shows a linear decrease with distance of approximately 0.2
dB/mi (0.1 dB/km) for the VHF–UHF frequency band.
1.2.3g Atmospheric Fading

Variations in the received field strengths around the median values are caused by changes in
atmospheric conditions. Field strengths tend to be higher in summer than in winter, and higher at
night than during the day, for paths over land beyond the line of sight. As a first approximation,
the distribution of long-term variations in field strength in decibels follows a normal probability
law.
Measurements indicate that the fading range reaches a maximum somewhat beyond the hori-
zon and then decreases slowly with distance out to several hundred miles. Also, the fading range
at the distance of maximum fading increases with frequency, while at the greater distances where
the fading range decreases, the range is also less dependent on frequency. Thus, the slope of the
graph N must be adjusted for both distance and frequency. This behavior does not lend itself to
treatment as a function of the earth’s radius factor k, since calculations based on the same range
of k produce families of curves in which the fading range increases systematically with increas-
ing distance and with increasing frequency.
Effects of the Upper Atmosphere (Ionosphere)

Four principal recognized layers or regions in the ionosphere are the E layer, the F1 layer, the F2
layer (centered at heights of about 100, 200, and 300 km, respectively), and the D region, which
is less clearly defined but lies below the E layer. These regular layers are produced by radiation
from the sun, so that the ion density—and hence the frequency of the radio waves that can be
reflected thereby—is higher during the day than at night. The characteristics of the layers are dif-
ferent for different geographic locations and the geographic effects are not the same for all lay-
p g
ers. The characteristics also differ with the seasons and with the intensity of the sun’s radiation,
as evidenced by the sunspot numbers, and the differences are generally more pronounced upon
the F2 than upon the F1 and E layers. There are also certain random effects that are associated
with solar and magnetic disturbances. Other effects that occur at or just below the E layer have
been established as being caused by meteors [24].
The greatest potential for television interference by way of the ionosphere is from sporadic E
ionization, which consists of occasional patches of intense ionization occurring 62 to 75 mi (100
to 120 km) above the earth’s surface and apparently formed by the interaction of winds in the
neutral atmosphere with the earth’s magnetic field. Sporadic E ionization can reflect VHF sig-
nals back to earth at levels capable of causing interference to analog television reception for peri-
ods lasting from 1 h or more, and in some cases totaling more than 100 h per year. In the U.S.,
VHF sporadic E propagation occurs a greater percentage of the time in the southern half of the
country and during the May to August period [25].
1.2.4 References
1. Bullington, K.: “Radio Propagation at Frequencies above 30 Mc,” Proc. IRE, pg. 1122,
October 1947.
2. Fink, D. G., (ed.): Television Engineering Handbook, McGraw-Hill, New York, N.Y., 1957.
3. Eckersley, T. L.: “Ultra-Short-Wave Refraction and Diffraction,” J. Inst. Elec. Engrs., pg.
286, March 1937.
4. Norton, K. A.: “Ground Wave Intensity over a Finitely Conducting Spherical Earth,” Proc.
IRE, pg. 622, December 1941.
5. Norton, K. A.: “The Propagation of Radio Waves over a Finitely Conducting Spherical
Earth,” Phil. Mag., June 1938.
6. van der Pol, Balth, and H. Bremmer: “The Diffraction of Electromagnetic Waves from an
Electrical Point Source Round a Finitely Conducting Sphere, with Applications to Radio-
telegraphy and to Theory of the Rainbow,” pt. 1, Phil. Mag., July, 1937; pt. 2, Phil. Mag.,
November 1937.
7. Burrows, C. R., and M. C. Gray: “The Effect of the Earth’s Curvature on Groundwave
Propagation,” Proc. IRE, pg. 16, January 1941.
8. “The Propagation of Radio Waves through the Standard Atmosphere,” Summary Technical
Report of the Committee on Propagation, vol. 3, National Defense Research Council,
Washington, D.C., 1946, published by Academic Press, New York, N.Y.
9. “Radio Wave Propagation,” Summary Technical Report of the Committee on Propagation
of the National Defense Research Committee, Academic Press, New York, N.Y., 1949.
10. de Lisle, E. W.: “Computations of VHF and UHF Propagation for Radio Relay Applica-
tions,” RCA, Report by International Division, New York, N.Y.
11. Selvidge, H.: “Diffraction Measurements at Ultra High Frequencies,” Proc. IRE, pg. 10,
January 1941.
p g
Propagation 1-43
12. McPetrie, J. S., and L. H. Ford: “An Experimental Investigation on the Propagation of
Radio Waves over Bare Ridges in the Wavelength Range 10 cm to 10 m,” J. Inst. Elec.
Engrs., pt. 3, vol. 93, pg. 527, 1946.
13. Megaw, E. C. S.: “Some Effects of Obstacles on the Propagation of Very Short Radio
Waves,” J. Inst. Elec. Engrs., pt. 3, vol. 95, no. 34, pg. 97, March 1948.
14. Dickson, F. H., J. J. Egli, J. W. Herbstreit, and G. S. Wickizer: “Large Reductions of VHF
Transmission Loss and Fading by the Presence of a Mountain Obstacle in Beyond-Line-of-
Sight Paths,” Proc. IRE, vol. 41, no. 8, pg. 96, August 1953.
15. Bullington, K.: “Radio Propagation Variations at VHF and UHF,” Proc. IRE, pg. 27, Janu-
ary 1950.
16. “Report of the Ad Hoc Committee, Federal Communications Commission,” vol. 1, May
1949; vol. 2, July 1950.
17. Epstein, J., and D. Peterson: “An Experimental Study of Wave Propagation at 850 Mc,”
Proc. IRE, pg. 595, May 1953.
18. “Documents of the XVth Plenary Assembly,” CCIR Report 563, vol. 5, Geneva, 1982.
19. Bean, B. R., and E. J. Dutton: “Radio Meteorology,” National Bureau of Standards Mono-
graph 92, March 1, 1966.
20. Dougherty, H. T., and E. J. Dutton: “The Role of Elevated Ducting for Radio Service and
Interference Fields,” NTIA Report 81–69, March 1981.
23. Longley, A. G., and P. L. Rice: “Prediction of Tropospheric Radio Transmission over Irreg-
ular Terrain—A Computer Method,” ESSA (Environmental Science Services Administra-
tion), U.S. Dept. of Commerce, Report ERL (Environment Research Laboratories) 79-ITS
67, July 1968.
24. National Bureau of Standards Circular 462, “Ionospheric Radio Propagation,” June 1948.
25. Smith, E. E., and E. W. Davis: “Wind-induced Ions Thwart TV Reception,” IEEE Spec-
trum, pp. 52—55, February 1981.
p g
g g
Chapter
1.3
Frequency Sources and References
Ulrich L. Rohde
Jerry C. Whitaker, Editor-in-Chief
1.3.1 Introduction1
A stable frequency reference is an important component of any transmission/reception system.
The quartz crystal is the classic device used for frequency generation and control. Quartz acts as
a stable high Q mechanical resonator. Crystal resonators are available for operation at frequen-
cies ranging from 1 kHz to 300 MHz and beyond.
1.3.2 Characteristics of Crystal Devices

The operating characteristics of a crystal are determined by the cut of the device from a bulk
“mother” crystal. The behavior of the device depends heavily on the size and shape of the crys-
tal, and the angle of the cut. To provide for operation at a wide range of frequencies, different
cuts, vibrating in one or more selected modes, are used.
The properties of a piezoelectric quartz crystal can be expressed in terms of three sets of axes.
(See Figure 1.3.1.) The axis joining the points at the ends of the crystal is known as the optical
axis, and electrical stresses applied in this direction exhibit no piezoelectric properties. The three
axes X ′ , X ′ ′ , and X ′ ′ ′ , passing through the corners of the hexagon that forms the section
perpendicular to the optical axes, are known as the electrical axes. The three axes Y ′ , Y ′ ′ , and
Y ′ ′ ′ , which are perpendicular to the faces of the crystal, are the mechanical axes.
If a flat section is cut from a quartz crystal in such a way that the flat sides are perpendicular
to an electrical axis, as shown in Figure 1.3.1b (X-cut), the mechanical stresses along the Y-axis
of such a section produce electric charges on the flat sides of the crystal. If the direction of these
stresses is changed from tension to compression, or vice versa, the polarity of the charges on the
1. Portions of this chapter are based on: Rohde, Ulrich L., and Jerry C. Whitaker: Communica-
tions Receivers: Principles and Design, 3rd ed., McGraw-Hill, New York, N.Y., 2000. Used
with permission.
1-45
(a ) X Y'' X'' (b) X Y'' X''

0˚
Y 30˚ Y' Y Y'
X' X' X' X'
Normal Normal
Y' Y Y' Y
X'' Y'' X X'' Y'' X
Figure 1.3.1 Cross section of a quartz crystal taken in the plane perpendicular to the optical axis:
(a) Y-cut plate, (b) X-cut plate.
crystal surfaces is reversed. Conversely, if electric charges are placed on the flat sides of the crys-
tal by applying a voltage across the faces, a mechanical stress is produced in the direction of the
Y-axis. This property by which mechanical and electrical properties are interconnected in a crys-
tal is known as the piezoelectric effect and is exhibited by all sections cut from a piezoelectric
crystal. Thus, if mechanical forces are applied across the faces of a crystal section having its flat
sides perpendicular to a Y-axis, as shown in Figure 1.3.1a, piezoelectric charges will be produced
because forces and potentials developed in such a crystal have components across at least one of
the Y- or X-axes, respectively.
An alternating voltage applied across a quartz crystal will cause the crystal to vibrate and, if
the frequency of the applied alternating voltage approximates a frequency at which mechanical
resonance can exist in the crystal, the amplitude vibrations will be large. Any crystal has a num-
ber of such resonant frequencies that depend upon the crystal dimensions, the type of mechanical
oscillation involved, and the orientation of the plate cut from the natural crystal.
Crystals are temperature-sensitive, as shown in Figure 1.3.2. The extent to which a device is
affected by changes in temperature is determined by its cut and packaging. Crystals also exhibit
changes in frequency with time. Such aging is caused by one or both of the following:
• Mass transfer to or from the resonator surface
• Stress relief within the device itself
Crystal aging is most pronounced when the device is new. As stress within the internal structure
is relieved, the aging process slows.
1.3.2a Frequency Stabilization

The stability of a quartz crystal is inadequate for most commercial and industrial applications.
Two common methods are used to provide the required long-term frequency stability:
• Oven-controlled crystal oscillator—The crystal is installed in a temperature-controlled box.
Because the temperature in the box is constant, controlled by a thermostat, the crystal remains
on-frequency. The temperature of the enclosure usually is set to the turnover temperature of
the crystal. (The turnover point is illustrated in Figure 1.3.2.)
q y
Curve 1
20
15
10
5
Δ f/f (ppm)
-5
Ti
-10 Curve 2
-15
-20
0 10 20 30 40 50 60 70 80 90 100
Temperature (deg C)
Figure 1.3.2 The effects of temperature on two types of AT-cut crystals.
• Temperature-compensated crystal oscillator (TCXO)—The frequency vs. temperature

changes of the crystal are compensated by varying a load capacitor. A thermistor network
typically is used to generate a correction voltage that feeds a voltage-controlled capacitor
(varactor) to retune the crystal to the desired on-frequency value.
1.3.2b Equivalent Circuit of a Quartz Resonator

Because of the piezoelectric effect, the quartz resonator behaves physically as a vibrating object
when driven by a periodic electric signal near a resonant frequency of the cavity [1, 2]. This res-
onant behavior may be characterized by equivalent electrical characteristics, which may then be
used to accurately determine the response and performance of the electromechanical system.
Dye and Butterworth (1926) [3] developed the equivalent lump electrical network that describes
an electrostatic signal driving a mechanically vibrating system (an electromechanical trans-
ducer). The equivalent electric network that models the piezo resonator was later developed by
Van Dyke (1928), Butterworth, and others, and is shown in Figure 1.3.3. In the equivalent net-
work, the circuit parameters R1, L1, and C1 represent parameters associated with the quartz crys-
tal, while C0 represents the shunt capacitance of the resonator electrodes in parallel with the
container that packages the crystal.
q y
The inductance L1 is associated with the

mass of the quartz slab, and the capacitance
C1 is associated with the stiffness of the slab.
Finally, the resistance, R1 is determined by
the loss processes associated with the opera-
tion of the crystal. Each of the equivalent
parameters can be measured using standard
network measurement techniques. Typical
parameters are shown in Table 1.3.1. Figure 1.3.3 Equivalent network of a piezoelec-
Application of an electric field across the tric resonator. (From [1]. Used with permission.)
electrodes of the crystal results in both face
and thickness shear waves developing within the crystal resonator. For a Y-cut crystal, face shear
modes occur in the X-Z plane, while thickness shear modes develop in the X-Y plane.
The vibration is enhanced for electric signals near a resonant frequency of the resonator. Low
frequency resonances tend to occur in the face shear modes, while high frequency resonances
occur in thickness shear modes. In either case, when the applied electric signal is near a resonant
frequency, the energy required to maintain the vibrational motion is small.
1.3.2c Temperature Compensation

Compensation for temperature-related frequency shifts can be accomplished by connecting a
reactance, such as a variable voltage capacitor (a varactor) in series with the resonator [1]. The
effect of this reactance is to pull the frequency to cancel the effect of frequency drift caused by
temperature changes. An example of a frequency versus temperature characteristic both before
and after compensation is shown in Figure 1.3.4.
Several techniques for temperature compensation are possible. They include:
• Analog temperature compensation
• Hybrid analog-digital compensation
• Digital temperature compensation
For analog compensation, a varactor is controlled by a voltage divider network that contains
both thermistors and resistors to manage the compensation. Stability values are in the range 0.5
to 10 ppm. The lowest value requires that individual components in the voltage divider network
be adjusted under controlled-temperature conditions. This is difficult over large temperature
ranges for analog networks because of the tight component value tolerances required and
because of interactions between different analog segments. Analog networks can be segmented
Table 1.3.1 Typical Equivalent Circuit Parameter Values for Quartz Crystals (After [1].)
Parameter 200 kHz 2 Mhz 30 Mhz, 3rd 50 Mhz, 5th

Fundamental Overtone Overtone
R1 2 kΩ 100 Ω 40 Ω 20 Ω
L1 27 H 520 mH 11 mH 6 mH
C1 0.024 pF 0.012 pF 0.0026 pF 0.0005 pF
Co 9 pF 4 pF 6 pF 4 pF
Q 18 × 103 18 × 103 18 × 103 18 × 103
q y
Figure 1.3.4 Frequency vs. temperature characteristics for a typical temperature-compensated

crystal oscillator. (After [2].)
in order to increase the independence of the component adjustments required to improve stabil-
ity.
These problems can be alleviated through the use of some form of hybrid or digital compen-
sation. In this scheme, the crystal oscillator is connected to a varactor in series, as in the analog
case. This capacitor provides the coarse compensation (about ±4 parts in 107). The fine compen-
sation is provided by the digital network to further improve stability. The hybrid analog-digital
method is then capable of producing stability values of ±5 parts in 10 8 over the range –40°C to
+80°C.
1.3.2d Stress Compensation

Stress compensation is employed in order to stabilize the crystal resonator frequency against
shifts caused by mechanical stresses [1]. These result from temperature changes or from elec-
trode changes, due to differences in the expansion coefficients, which can occur over time. The
frequency variations are caused by alteration of the crystal elastic coefficients with stress bias.
The method developed for stress compensation is a special crystal cut called the stress-compen-
sated or SC-cut. This cut was developed to obtain a resonator whose frequency would be inde-
pendent of specified stresses.
The SC-cut has some advantages over the more common AT- and BT-cuts. These include a
flatter frequency-temperature response, especially at the turning points, and less susceptibility to
aging because of changes in the mounting stresses. The lower sensitivity to frequency changes
means that SC-cut resonators can be frequency-temperature tested more rapidly than their AT-
and BT-cut counterparts. A disadvantage of the SC-cut is that it is sensitive to air-damping and
must be operated in a vacuum environment to obtain optimum Q factors. The frequency-temper-
ature curve is also roughly 70° higher for the SC-cut over the more common AT-cut, which could
also be a disadvantage in some applications.
q y
1.3.2e Aging Effects

Aging in crystal oscillators generally refers to any change over time that will affect the frequency
characteristics of the oscillator, or the physical parameters that describe the device; such as
motional time constant and equivalent circuit parameters [1]. Some factors that influence aging
include the following:
• Surface deterioration
• Surface contamination
• Electrode composition
• Environmental conditions
Deterioration on the surface contributes to strain in the crystal, and this type of strain can
impact the operating frequency of the oscillator. More importantly, the surface strain is sensitive
to humidity and temperature early in the life of the crystal. One method of combating this sur-
face degradation mechanism is to etch away the strained surface layers. The initial slope of the
frequency change versus etch time curve is exponential. However, after the strained surface lay-
ers have been removed, the slope becomes linear.
Another surface issue is contamination from foreign materials. The presence of foreign sub-
stances will load the oscillator with an additional mass, and this can be reflected in the equivalent
circuit parameters. The contaminants can be introduced during fabrication processes, from out-
gassing of other materials after the device has been packaged, or from contaminants in the envi-
ronment during storage. The contamination process is often increased in higher temperature
environments, and temperature cycling will lead to stability problems for the device.
Aging problems associated with the device electrodes includes delamination, leading to
increased resistance or corrosion effects. Typical electrode materials include gold, silver, and alu-
minum. Of these, silver is the most popular material.
1.3.3 Oscillators
Receivers are seldom single-channel devices, but more often cover wide frequency ranges. In the
superheterodyne receiver, this is accomplished by mixing the input signal with an LO signal. The
LO source must meet a variety of requirements:
• It must have high spectral purity.
• It must be agile so that it can move rapidly (jump) between frequencies in a time frame that
can be as short as a few microseconds.
• The increments in which frequencies can be selected must be small.
Frequency resolution between 1 and 100 Hz is generally adequate below 30 MHz; however, there
are a number of systems that provide 0.001-Hz steps. At higher frequencies, resolution is gener-
ally greater than 1 kHz.
In most modern receivers, such a frequency source is typically a synthesizer that generates all
individual frequencies as needed over the required frequency band. The modern synthesizer pro-
vides stable phase-coherent outputs. The frequencies are derived from a master standard, which
can be a high-precision crystal oscillator, a secondary atomic standard (such as a rubidium gas
q y
cell), or a primary standard using a cesium atomic beam. The following characteristics must be
specified for the synthesizer:
• Frequency range
• Frequency resolution
• Frequency indication
• Maximum frequency error
• Settling time
• Reference frequency
• Output power
• Harmonic distortion
• SSB phase noise
• Discrete spurs (spurious frequencies)
• Wide-band noise
• Control interface
• Power consumption
• Mechanical size
• Environmental conditions
Free-running tunable oscillators, once used in radio receivers, have generally been replaced in
modern receivers because of their lack of precision and stability. Fixed-tuned crystal-controlled
oscillators are still used in second- and third-oscillator applications in multiconversion superhet-
erodyne receivers that do not require single-reference precision. Oscillators used in synthesizers
have variable tuning capability, which may be voltage-controlled, generally by varactor diodes.
Synthesizer designs have used mixing from multiple crystal sources and mixing of signals
derived from a single source through frequency multiplication and division. Synthesizers may be
“direct” and use the product of multiple mixing and filtering or “indirect” and use a phase-
locked-loop (PLL) slaved to the direct output to provide reduced spurious signals. There are a
number of publications describing these and other techniques, the classic texts being [4] through
[6]. Most modern receivers operating below 3 GHz use single- or multiple-loop digital PLL syn-
thesizers, although for some applications, direct digital waveform synthesis may be used.
1.3.3a Key Terms

The following characteristics are commonly used to describe oscillator performance:
• Frequency pushing. Frequency pushing characterizes the degree to which an oscillator’s fre-
quency is affected by its supply voltage. For example, a sudden current surge caused by acti-
vating a transceiver’s RF power amplifier may produce a spike on the VCO’s dc power supply
and a consequent frequency jump. Frequency pushing is specified in frequency/voltage form,
and is tested by varying the VCO’s dc supply voltage (typically ±1 V) with its tuning voltage
held constant.
q y
• Harmonic output power. Harmonic output power is measured relative to the output power of
the oscillator. Typical values are 20 dB or more suppression relative to the fundamental. This
suppression can be improved by additional filtering.
• Output power. The output power of the oscillator, typically expressed in dBm, is measured
into a 50-Ω load. The output power is always combined with a specification for flatness or
variation. A typical spec would be 0 dBm ±1 dB.
• Output power as a function of temperature. All active circuits vary in performance as a
function of temperature. The output power of an oscillator over a temperature range should
vary less than a specified value, such as 1 dB.
• Post-tuning drift. After a voltage step is applied to the tuning diode input, the oscillator fre-
quency may continue to change until it settles to a final value. This post-tuning drift is one of
the parameters that limits the bandwidth of the VCO input.
• Power consumption. This characteristic conveys the dc power, usually specified in milliwatts
and sometimes qualified by operating voltage, required by the oscillator to function properly.
• Sensitivity to load changes. To keep manufacturing costs down, many wireless applications
use a VCO alone, without the buffering action of a high reverse-isolation amplifier stage. In
such applications, frequency pulling, the change of frequency resulting from partially reactive
loads, is an important oscillator characteristic. Pulling is commonly specified in terms of the
frequency shift that occurs when the oscillator is connected to a load that exhibits a non-unity
VSWR (such as 1.75:1, usually referenced to 50 Ω ), compared to the frequency that results
with a unity-VSWR load (usually 50 Ω ). Frequency pulling must be minimized, especially in
cases where power stages are close to the VCO unit and short pulses may affect the output fre-
quency. Such feedback can make-phase locking impossible.
• Spurious outputs. The spurious output specification of a VCO, expressed in decibels, char-
acterizes the strength of unwanted and nonharmonically related components relative to the
oscillator fundamental. Because a stable, properly designed oscillator is inherently clean,
such spurs are typically introduced only by external sources in the form of radiated or con-
ducted interference.
• Temperature drift. Although the synthesizer is responsible for locking and maintaining the
oscillator frequency, the VCO frequency change as a function of temperature is a critical
parameter and must be specified. This value varies between 10 kHz/ºC to several hundred
kHz/ºC depending on the center frequency and tuning range.
• Tuning characteristic. This specification shows the relationship, depicted as a graph,
between the VCO operating frequency and the tuning voltage applied. Ideally, the correspon-
dence between operating frequency and tuning voltage is linear.
• Tuning linearity. For stable synthesizers, a constant deviation of frequency versus tuning
voltage is desirable. It is also important to make sure that there are no breaks in the tuning
range—for example, that the oscillator does not stop operating with a tuning voltage of 0 V.
• Tuning sensitivity, tuning performance. This datum, typically expressed in megahertz per
volt (MHz/V), characterizes how much the frequency of a VCO changes per unit of tuning-
voltage change.
q y
• Tuning speed. This characteristic is defined as the time necessary for the VCO to reach 90
percent of its final frequency upon the application of a tuning-voltage step. Tuning speed
depends on the internal components between the input pin and tuning diode—including,
among other things, the capacitance present at the input port. The input port’s parasitic ele-
ments determine the VCO’s maximum possible modulation bandwidth.
1.3.4 Phase-Locked Loop Synthesizers

The traditional synthesizer consists of a single loop and the step size of the output frequency is
equal to the reference frequency at the phase detector. Figure 1.3.5 shows this classic approach.
Wireless applications with a step size of 200 kHz have made the life of the designer somewhat
easier, since such a wide step size reduces the division ratio. As can be seen in Figure 1.3.5, the
simplest form of a digital synthesizer consists of a voltage-controlled oscillator (VCO). For PLL
applications, the oscillator sensitivity, typically expressed in megahertz per volt (MHz/V), needs
to be stated. For high-performance test equipment applications, the VCO is frequently provided
in the form of a YIG oscillator. These oscillators operate at extremely high Q and are fairly
expensive. The VCO needs to be separated from any load by a post amplifier, which drives a
sine-wave-to-logic-waveform translator.
Typically, an ECL line receiver or its equivalent would serve for this function. This stage, in
turn, drives a programmable divider and divides the oscillator frequency down to a reference fre-
quency, such as 200 kHz. Assuming an oscillator frequency of 1 GHz, the division ratio would be
1 GHz/200 kHz = 5000. In Figure 1.3.5, we are looking at a 3 GHz output frequency and a step
size determined by the reference source resolution, assuming a fixed division ratio.
The phase detector, actually the phase/frequency detector (PFD), is driven by the reference
frequency on one side and the oscillator frequency, divided down, on the other side. The PFD is
typically designed to operate from a few kilohertz to several tens of megahertz, 50 MHz for
example. In our case, this would mean a step size of 50 MHz. Most of these types of phase detec-
tors use MOS technology to keep the levels for on/off voltage high, and their noise contribution
needs to be carefully evaluated. While the synthesizer is not locked, the output of the PFD is a dc
control voltage with a superimposed ac voltage equal to the difference between the two frequen-
cies prior to lock. Most PFDs have a flip-flop-based architecture, and their output consists of a
train of pulses that must to be integrated to produce the control voltage necessary for driving the
VCO into the locked condition. This integrator also serves as a loop filter. Its purpose is to sup-
press the reference frequency and provide the necessary phase/frequency response for stable
locking. The basic loop really is a nonlinear control loop that, for the purpose of analysis, is
always assumed to be linear or piecewise-linear. The most linear phase detector is a diode ring,
but it has a low-level dc output that requires an operational amplifier to shift the level of typically
±0.3 V to the high voltage required for the tuning diode. These values are usually somewhere
between 5 and 30 V. The tuning diode itself needs to have the appropriate breakdown voltage and
voltage-dependent linearity to provide constant loop gain. In most cases, this is not possible,
especially if the division ratio changes by a factor of 2 or more; in this case, it is a wise decision
to use coarse steering so that fine tuning shows a reasonably linear performance. These loops are
also called Type 2 second-order loops. This is because the loop has two integrators, one being
the loop filter and other one being the tuning diode. The order of the loop is determined by the
filter. Table 1.3.2 shows circuit and transfer characteristics of several PLL filters.
q y
Figure 1.3.5 Block diagram of a PLL synthesizer driven by a frequency standard, direct digital
synthesis (DDS), or fractional-N synthesizer for high resolution at the output. The last two stan-
dards allow a relatively low division ratio and provide quasi-arbitrary resolution.
The filters shown in the table are single-ended, which means that they are driven from the out-
put of a CMOS switch in the phase/frequency detector. This type of configuration exhibits a
problem under lock conditions: If we assume that the CMOS switches are identical and have no
leakage, initially the output current charges the integrator and the system will go into lock. If it is
locked and stays locked, there is no need for a correction voltage, and therefore the CMOS
switches will not supply any output. The very moment a tiny drift occurs, a correction voltage is
required, and therefore there is a drastic change from no loop gain (closed condition) to a loop
gain (necessary for acquiring lock). This transition causes a number of nonlinear phenomena,
and therefore it is a better choice to either use a passive filter in a symmetrical configuration or a
symmetrical loop filter with an operational amplifier instead of the CMOS switches. Many of
the modern PFDs have different outputs to accommodate this. An ill-conditioned filter or selec-
tion of values for the filter frequently leads to a very low-frequency type of oscillation, often
referred to as motorboating. This can only be corrected by adjusting the phase/frequency behav-
ior of the filter.
The Bode diagram is a useful tool in designing the appropriate loop filter. The Bode diagram
shows the open-loop performance, both magnitude and phase, of the phase-locked loop.
1.3.4a Practical PLL Circuits

Figure 1.3.6 shows a passive filter that is used for a synthesizer chip (National LMX). This chip
has a charge-pump output, which explains the need for the first capacitor.
Figure 1.3.7 shows an active integrator operating at a reference frequency of several mega-
hertz. The notch filter at the output reduces the reference frequency considerably. The notch is at
about 4.5 MHz.
q y
Table 1.3.2 PLL Filter Characteristics
Circuit and Transfer Characteristics of Several PLL Filters
Implementation of Different Loop Filters
Recommended Passive Filters for Charge Pumps

q y
Figure 1.3.8 shows the combination of a

phase/frequency discriminator and a higher-
order loop filter as used in more complicated
systems, such as fractional-division synthe-
sizers.
Figure 1.3.9 shows a custom-built phase
detector with a nominal noise floor of better
than –168 dBc/Hz.
1.3.4b Fractional-Division Synthesizers

Figure 1.3.6 Type 1 high-order loop filter used
In conventional synthesizers, the minimum for passive filter evaluation. The 1-nF capacitor
step size is equal to the reference frequency. is used for spike suppression. The filter consists
In order to achieve a finer resolution, we can of a lag portion and an additional low pass sec-
tion.
either manipulate the reference as outlined in
Figure 1.3.5, or we can use fractional divi-
sion. The principle of the fractional-N-division synthesizer has been around for some time. In the
past, implementation of this technique has been done in an analog system. The Figure 1.3.5 sin-
gle loop uses a frequency divider where the division ratio is an integer value between 1 and some
very large number, hopefully not as high as 50,000. It would be ideal to be able to build a synthe-
sizer with the 1.25 MHz reference or 50 MHz reference and yet obtain the desired step size reso-
lution, such as 25 kHz. This would lead to a much smaller division ratio and better phase noise
performance.
An alternative would be for N to take on fractional values. The output frequency could then be
changed in fractional increments of the reference frequency. Although a digital divider cannot
provide a fractional division ratio, ways can be found to accomplish the same task effectively.
The most frequently used method is to divide the output frequency by N + 1 every M cycles and
to divide by N the rest of the time. The effective division ratio is then N + 1/M, and the average
output frequency is given by
Figure 1.3.7 A Type 2 high-order filter with a notch to suppress the discrete reference spurs.
q y
Figure 1.3.8 Phase/frequency discriminator including an active loop filter capable of operating up
to 100 MHz.
Figure 1.3.9 Custom-built phase detector with a noise floor of better than –168 dBc/Hz. This
phase detector shows extremely low phase jitter.
q y
⎛ 1 ⎞
f0 =⎜N + ⎟fr
⎝ M⎠
(1.3.1)
This expression shows that f0 can be varied in fractional increments of the reference frequency
by varying M. The technique is equivalent to constructing a fractional divider, but the fractional
part of the division is actually implemented using a phase accumulator. This method can be
expanded to frequencies much higher than 6 GHz using the appropriate synchronous dividers.
The phase accumulator approach is illustrated by the following example.
Consider the problem of generating 899.8 MHz using a fractional-N loop with a 50-MHz ref-
erence frequency; 899.8 MHz = 50 MHz (N – K/F). The integral part of the division N is set to
17 and the fractional part K/F is 996/1000 (the fractional part K/F is not a integer). The VCO
output is divided by 996 × every 1,000 cycles. This can easily be implemented by adding the
number 0.996 to the contents of an accumulator every cycle. Every time the accumulator over-
flows, the divider divides by 18 rather than by 17. Only the fractional value of the addition is
retained in the phase accumulator. If we move to the lower band or try to generate 850.2 MHz, N
remains 17 and K/F becomes 4/1000. This method of fractional division was first introduced by
using analog implementation and noise cancellation, but today it is implemented as a totally dig-
ital approach. The necessary resolution is obtained from dual-modulus prescaling, which allows
for a well-established method for achieving a high-performance frequency synthesizer operating
at UHF and higher frequencies. Dual-modulus prescaling avoids the loss of resolution in a sys-
tem compared to a simple prescaler. It allows a VCO step equal to the value of the reference fre-
quency to be obtained. This method needs an additional counter and the dual-modulus prescaler
then divides one or two values depending upon the state of its control. The only drawback of
prescalers is the minimum division ratio of the prescaler for approximately N2.
The dual modulus divider is the key to implementing the fractional-N synthesizer principle.
Although the fractional-N technique appears to have a good potential of solving the resolution
limitation, it is not free of complications. Typically, an overflow from the phase accumulator,
which is the adder with output feedback to the input after being latched, is used to change the
instantaneous division ratio. Each overflow produces a jitter at the output frequency, caused by
the fractional division, and is limited to the fractional portion of the desired division ratio.
In our example, we had chosen a step size of 200 kHz, and yet the discrete side bands vary
from 200 kHz for K/F = 4/1000 to 49.8 MHz for K/F = 996/1000. It will become the task of the
loop filter to remove those discrete spurious elements. While in the past the removal of discrete
spurs was accomplished by analog techniques, various digital methods are now available. The
microprocessor has to solve the following equation:
⎛ K⎞
N * = ⎜ N + ⎟ = [N ( F − K ) + (N +1)K ]
⎝ F⎠
(1.3.2)
For example, for F0 = 850.2 MHz, we obtain

q y
850.2 MHz
N* = = 17.004
50 MHz
(1.3.3)
Following the formula above
⎛ K ⎞ [17(1000 − 4) + (17 + 1) × 4] [16932_ 72]

N* = ⎜ N + ⎟ = = = 17.004
⎝ F⎠ 1000 1000
[16932 − 72]
Fout = 50 MHz × = 846.6 MHz + 36
. MHz = 850.2 MHz
1000
(1.3.4)
By increasing the number of accumulators, frequency resolution considerably below a step

size of 1 Hz is possible with the same switching speed.
1.3.5 Multiloop Synthesizers

To avoid the limitations of the single-loop synthesizer, synthesizers are often designed to employ
more than one loop. Figure 1.3.10 shows a block diagram of a multiloop synthesizer. The first
LO, operating from 81.4 to 111.4 MHz, is a two-loop synthesizer using a frequency translation
stage. It comprises a 70- to 80-MHz loop, a divider, two frequency translators, and an output
loop at the final LO frequency. Two single-loop synthesizers are also used later in the receiver,
but our discussion will be confined to the multiloop unit.
A 10-MHz crystal oscillator is used as the standard to which all of the internal oscillator fre-
quencies are locked. A divide-by-100 circuit reduces this to a 100-kHz reference used in both
loops of the synthesizer. The 100-kHz reference is further divided by 100 to provide the 1-kHz
reference for the 70- to 80-MHz loop. The output of this loop is then further divided by 100 to
provide 10-Hz steps between 0.7 and 0.8 MHz. This division improves the noise sidebands and
spurious signal suppression of the loop by 40 dB.
The 0.7- to 0.8-MHz band is converted to 10.7 to 10.8 MHz by mixing it with the 10-MHz
reference. A crystal filter is used to provide the necessary suppression of the two inputs to the
mixer. The resultant signal is translated to 69.2 to 69.3 MHz by further mixing with a signal of
80 MHz, the eighth harmonic of the 10-MHz frequency standard. The 80-MHz signal can be
generated either by a frequency multiplier and crystal filter or by using an 80-MHz crystal oscil-
lator under phase-lock control of the standard. In the former case, the noise sideband perfor-
mance of the standard is degraded by 18 dB over the standard. Another possibility is the use of
an 80-MHz crystal oscillator standard, followed by a divide-by-8 circuit to produce the 10-MHz
internal reference. However, it is not possible to build crystal oscillators at 80 MHz with as good
long- and short-term frequency stability as at 10 MHz. Hence, the phase-locked crystal oscillator
approach was used to achieve high stability.
q y
Figure 1.3.10 Block diagram of a multiloop synthesizer. (Courtesy of Rohde & Schwarz.)
The 69.2- to 69.3-MHz output frequency from the mixer, after filtering, is mixed with the
final VCO output frequency to produce a signal of 12.2 to 42.1 MHz, which after division by M
is used for comparison in the final PLL with the 100-kHz reference. The M value is used to
select the 0.1-MHz steps, while the value of N shifts the 70- to 80-MHz oscillator to provide 10-
Hz resolution over the 0.1-MHz band resulting from its division by 100. This synthesizer pro-
vides the first oscillator frequency for a receiver with an 81.4-MHz first IF, for a band of input
frequencies up to 30 MHz, with 10-Hz step resolution.
This multiloop synthesizer illustrates the most important principles found in general-purpose
synthesizers. A different auxiliary loop could be used to provide further resolution. For example,
by replacing the 10.7- to 10.8-MHz loop by a digital direct frequency synthesizer, quasi-infinite
resolution could be obtained by a microprocessor-controlled low-frequency synthesizer. Such a
synthesizer has a very fast lock-up time, comparable to the speed of the 100-kHz loop. In the
design indicated, the switching speed is determined by the 1-kHz fine resolution loop. For a loop
bandwidth of 50 Hz for this loop, we will obtain a settling time in the vicinity of 40 ms. The
longest time is needed when the resolution synthesizer is required to jump from 80 to 70 MHz.
During this frequency jump, the loop will go out of both phase and frequency lock and will need
complete reacquisition. Thus, each time a 100-kHz segment is passed through, there is such a
q y
disturbance. The same occurs when the output frequency loop jumps over a large frequency seg-
ment. The VCO, operating from 81.4 to 111.4 MHz, is coarse-tuned by switching diodes, and
some of the jumps are objectionable. The noise sideband performance of this synthesizer is
determined inside the loop bandwidth by the reference noise multiplied to the output frequency,
and outside the loop bandwidth by the VCO noise. The latter noise can be kept low by building a
high-quality oscillator.
Another interesting multiloop synthesizer is shown in Figure 1.3.11. A 13- to 14-MHz sin-
gleloop synthesizer with 1-kHz steps is divided by 10 to provide 100-Hz steps and a 20-dB noise
improvement. A synthesizer in this frequency range can be built with one oscillator in a single
chip. The output loop, operating from 69.5 to 98 MHz, is generated in a synthesizer loop with
50-kHz reference and 66.6- to 66.7-MHz frequency offset generated by translation of the fine
loop frequency. The reference frequencies for the two loops are generated from a 10.7 MHz tem-
perature-compensated crystal oscillator standard. An additional 57.3-MHz TCXO is used to
drive a second oscillator in the receiver as well as to offset the fine frequency loop of the synthe-
sizer. An increase in frequency of this oscillator changes the output frequency of the first LO
(synthesizer) in the opposite direction to the second LO, thereby canceling the drift error. This
drift-canceling technique is sometimes referred to as the Barlow-Wadley principle.
1.3.6 Direct Digital Synthesis

Direct digital frequency synthesis (DDFS) is an oscillator scheme wherein a digital representa-
tion of the desired signal is generated and then applied to a D/A converter to reform the digital
representation to an analog waveform. Advances in high-speed microelectronics, particularly the
microprocessor, make DDFS practical at frequencies in the very-high-frequency band and below.
Systems can be compact, use low power, and provide fine frequency resolution with virtually
instantaneous switching of frequencies. DDFS is finding increasing application, particularly in
conjunction with PLL synthesizers.
DDFS uses a single-frequency source (clock) as a time reference. One method of digitally
generating the values of a sine wave is to solve the digital recursion relation as follows
Y n = [ 2 cos ( 2 π f t)] Y ( n – 1 ) – Y ( n – 2 )
(1.3.5)
This is solved by Yn = cos (2fnt). There are at least two problems with this method, however. The
noise can increase until a limit cycle (nonlinear oscillation) occurs. Also, the finite word length
used to represent 2 cos (2ft) places a limitation on the frequency resolution. Another method of
DDFS, direct table lookup, consists of storing the sinusoidal amplitude coefficients for succes-
sive phase increments in memory. The continuing miniaturization in size and cost of ROM make
this the most frequently used technique.
One method of direct table lookup outputs the same N points for each cycle of the sine wave,
and changes the output frequency by adjusting the rate at which the points are computed. It is rel-
atively difficult to obtain fine frequency resolution with this approach, so a modified table look-
up method is generally used. It is this method that we describe here. The function cos (2ft) is
approximated by outputting the function cos (2fnT) for n = 1, 2, 3, ... , where T is the interval
q y
Figure 1.3.11 Block diagram of a multiloop frequency synthesizer incorporating a drift-canceling

technique. (Courtesy of Rohde & Schwarz.)
between conversions of digital words in the D/A converter and n represents the successive sam-
ple numbers. The sampling frequency, or rate, of the system is 1/T. The lowest output frequency
waveform contains N distinct points in its waveform, as illustrated in Figure 1.3.12. A waveform
of twice the frequency can be generated, using the same sampling rate, but outputting every other
data point. A waveform k times as fast is
obtained by outputting every kth point at the
same rate 1/T. The frequency resolution, then,
is the same as the lowest frequency fL.
The maximum output frequency is selected
so that it is an integral multiple of fL, that is, fU
= kF L. If P points are used in the waveform of
the highest frequency, then N (= kP) points are
used in the lowest frequency waveform. The
number N is limited by the available memory
size. The minimum value that P can assume is
Figure 1.3.12 Synthesized waveform gener-
usually taken to be four. With this small value
ated by direct digital synthesis.
of P, the output contains many harmonics of the
q y
Figure 1.3.13 Block diagram of a direct digital frequency synthesizer. (From [8]. Reprinted with
permission.)
desired frequency. These can be removed by the use of low-pass filtering at the D/A output. For
P = 4, the period of the highest frequency is 4T, resulting in fU = 4fL. Thus, the highest attainable
frequency is determined by the fastest sampling rate possible.
In the design of this type of DDFS, the following guidelines apply:
• The desired frequency resolution determines the lowest output frequency fL.
• The number of D/A conversions used to generate fL is N = 4k = 4fU/fL provided that four con-
versions are used to generate fU (P = 4).
• The maximum output frequency fU is limited by the maximum sampling rate of the DDFS,
≤1/4T. Conversely, T ≤1/4fU.
The architecture of the complete DDFS is shown in Figure 1.3.13. To generate nfL, the integer
n addresses the register, and each clock cycle kn is added to the content of the accumulator so
that the content of the memory address register is increased by kn. Each knth point of the mem-
ory is addressed, and the content of this memory location is transferred to the D/A converter to
produce the output sampled waveform.
To complete the DDFS, the memory size and the length (number of bits) of the memory word
must be determined. The word length is determined by system noise requirements. The ampli-
tude of the D/A output is that of an exact sinusoid corrupted with the deterministic noise result-
ing from truncation caused by the finite length of the digital words (quantization noise). If an (n
+ l)-bit word length (including one sign bit) is used and the output of the A/D converter varies
between 1, the mean noise from the quantization will be
q y
2n 2 (n + 1 )
1 ⎛ 1 ⎞ 1 ⎛ 1 ⎞
σ2 = ⎜ ⎟ = ⎜ ⎟
12 ⎝ 2 ⎠ 3 ⎝2 ⎠
(1.3.6)
The mean noise is averaged over all possible waveforms. For a worst-case waveform, the noise is
a square wave with amplitude
n
( 1 ⁄ 2 )( 1 ⁄ 2 )
and
2n
σ = ( 1 ⁄ 4 )( 1 ⁄ 2 )
For each bit added to the word length, the spectral purity improves by 6 dB.
The main drawback of the DDFS is that it is limited to relatively low frequencies. The upper
frequency is directly related to the maximum usable clock frequency. DDFS tends to be noisier
than other methods, but adequate spectral purity can be obtained if sufficient low-pass filtering
is used at the output. DDFS systems are easily constructed using readily available microproces-
sors. The combination of DDFS for fine frequency resolution plus other synthesis techniques to
obtain higher-frequency output can provide high resolution with very rapid settling time after a
frequency change. This is especially valuable for frequency-hopping spread-spectrum systems.
1.3.7 References
1. Tate, Jeffrey P., and Patricia F. Mead: “Crystal Oscillators,” in The Electronics Handbook,
Jerry C. Whitaker (ed.), CRC Press, Boca Raton, Fla., pp. 185–199, 1996.
2. Frerking, M. E.: Crystal Oscillator Design and Temperature Compensation, Van Nostrand
Reinhold, New York, N. Y., 1978.
3. Dye, D. W.: Proc. Phys. Soc., Vol. 38, pp. 399–457, 1926.
4. Hietala, Alexander W., and Duane C. Rabe: “Latched Accumulator Fractional-N Synthesis
With Residual Error Reduction,” United States Patent, Patent No. 5,093,632, March 3,
1992.
5. Riley, Thomas A. D.: “Frequency Synthesizers Having Dividing Ratio Controlled Sigma-
Delta Modulator,” United States Patent, Patent No. 4,965,531, October 23, 1990.
6. King, Nigel J. R.: “Phase Locked Loop Variable Frequency Generator,” United States
Patent, Patent No. 4,204,174, May 20, 1980.
7. Wells, John Norman: “Frequency Synthesizers,” European Patent, Patent No. 012579OB2,
July 5, 1995.
8. Rohde, Ulrich L.: Digital PLL Frequency Synthesizers, Prentice-Hall, Englewood Cliffs,
N.J., 1983.
q y
1.3.8 Bibliography
Rohde, Ulrich L.: Microwave and Wireless Synthesizers: Theory and Design, John Wiley &
Sons, New York, N.Y., 1997.
q y
g g
Chapter
1.4
Modulation Systems and Characteristics
1.4.1 Introduction
The primary purpose of most communications and signaling systems is to transfer information
from one location to another. The message signals used in communication and control systems
usually must be limited in frequency to provide for efficient transfer. This frequency may range
from a few hertz for control systems to a few megahertz for video signals to many megahertz for
multiplexed data signals. To facilitate efficient and controlled distribution of these components,
an encoder generally is required between the source and the transmission channel. The encoder
acts to modulate the signal, producing at its output the modulated waveform. Modulation is a
process whereby the characteristics of a wave (the carrier) are varied in accordance with a mes-
sage signal, the modulating waveform. Frequency translation is usually a by-product of this pro-
cess. Modulation may be continuous, where the modulated wave is always present, or pulsed,
where no signal is present between pulses.
There are a number of reasons for producing modulated waves, including:
• Frequency translation. The modulation process provides a vehicle to perform the necessary
frequency translation required for distribution of information. An input signal may be trans-
lated to its assigned frequency band for transmission or radiation.
• Signal processing. It is often easier to amplify or process a signal in one frequency range as
opposed to another.
• Antenna efficiency. Generally speaking, for an antenna to be efficient, it must be large com-
pared with the signal wavelength. Frequency translation provided by modulation allows
antenna gain and beamwidth to become part of the system design considerations. The use of
higher frequencies permits antenna structures of reasonable size and cost.
• Bandwidth modification. The modulation process permits the bandwidth of the input signal to
be increased or decreased as required by the application. Bandwidth reduction permits more
efficient use of the spectrum, at the cost of signal fidelity. Increased bandwidth, on the other
hand, provides increased immunity to transmission channel disturbances.
• Signal multiplexing. In a given transmission system, it may be necessary or desirable to com-
bine several different signals into one baseband waveform for distribution. Modulation pro-
vides the vehicle for such multiplexing. Various modulation schemes allow separate signals to
1-67
y
be combined at the transmission end and separated (demultiplexed) at the receiving end. Mul-
tiplexing may be accomplished by using, among other systems, frequency-domain multiplex-
ing (FDM) or time-domain multiplexing (TDM).
Modulation of a signal does not come without the possible introduction of undesirable
attributes. Bandwidth restriction or the addition of noise or other disturbances are the two pri-
mary problems faced by the transmission system designer.
1.4.2 Amplitude Modulation

In the simplest form of amplitude modulation, an analog carrier is controlled by an analog mod-
ulating signal. The desired result is an RF waveform whose amplitude is varied by the magnitude
of the applied modulating signal and at a rate equal to the frequency of the applied signal. The
resulting waveform consists of a carrier wave plus two additional signals:
• An upper-sideband signal, which is equal in frequency to the carrier plus the frequency of the
modulating signal
• A lower-sideband signal, which is equal in frequency to the carrier minus the frequency of the
modulating signal
This type of modulation system is referred to as double-sideband amplitude modulation
(DSAM).
The radio carrier wave signal onto which the analog amplitude variations are to be impressed
is expressed as
e ( t ) = AE c cos ( ωc t ) (1.4.1)
Where:
e(t) = instantaneous amplitude of carrier wave as a function of time (t)
A = a factor of amplitude modulation of the carrier wave
ω c = angular frequency of carrier wave (radians per second)
E c = peak amplitude of carrier wave
If A is a constant, the peak amplitude of the carrier wave is constant, and no modulation
exists. Periodic modulation of the carrier wave results if the amplitude of A is caused to vary with
respect to time, as in the case of a sinusoidal wave
E m
A = 1 + ⎛ -------⎞ cos ( ωm t ) (1.4.2)
⎝ Ec ⎠
where Em/Ec = the ratio of modulation amplitude to carrier amplitude.

The foregoing relationship leads to
Em
e ( t ) = Ec 1 + ⎛ ------- ⎞ cos ( ωm t ) cos ( ωc t ) (1.4.3)
⎝ Ec ⎠
y
This is the basic equation for periodic (sinusoidal) amplitude modulation. When all multiplica-
tions and a simple trigonometric identity are performed, the result is
e ( t ) = E c cos ( ωc t ) + M c m
M
---- cos ( ω t + ω t ) + ---- cos ( ω t – ω t )
c m (1.4.4)
2 2
where M = the amplitude modulation factor (Em/Ec).

Amplitude modulation is, essentially, a multiplication process in which the time functions that
describe the modulating signal and the carrier are multiplied to produce a modulated wave con-
taining intelligence (information or data of some kind). The frequency components of the modu-
lating signal are translated in this process to occupy a different position in the spectrum.
The bandwidth of an AM transmission is determined by the modulating frequency. The band-
width required for full-fidelity reproduction in a receiver is equal to twice the applied modulat-
ing frequency.
The magnitude of the upper sideband and lower sideband will not normally exceed 50 percent
of the carrier amplitude during modulation. This results in an upper-sideband power of one-
fourth the carrier power. The same power exists in the lower sideband. As a result, up to one-half
of the actual carrier power appears additionally in the sum of the sidebands of the modulated sig-
nal. A representation of the AM carrier and its sidebands is shown in Figure 1.4.1. The actual
occupied bandwidth, assuming pure sinusoidal modulating signals and no distortion during the
modulation process, is equal to twice the frequency of the modulating signal.
The extent of the amplitude variations in a modulated wave is expressed in terms of the
degree of modulation or percentage of modulation. For sinusoidal variation, the degree of modu-
lation m is determined from
E – E min
m = ----avg
------------------------- (1.4.5)
E avg
Ec
Where:
Eavg = average envelope amplitude
E min = minimum envelope amplitude
Full (100 percent) modulation occurs when
the peak value of the modulated envelope Ec /2 Ec/ 2
reaches twice the value of the unmodulated

carrier, and the minimum value of the enve-
lope is zero. The envelope of a modulated
AM signal in the time domain is shown in
Figure 1.4.2.
When the envelope variation is not sinu- fc - f m fc f c + fm
soidal, it is necessary to define the degree of
modulation separately for the peaks and Figure 1.4.1 Frequency-domain representation
of an amplitude-modulated signal at 100 per-
troughs of the envelope
cent modulation. Ec = carrier power, fc = fre-
quency of the carrier, and fm = frequency of the
modulating signal.
y
Positive
Carrier amplitude
Time
Negative
Figure 1.4.2 Time-domain representation of an amplitude-modulated signal. Modulation at 100

percent is defined as the point at which the peak of the waveform reaches twice the carrier level,
and the minimum point of the waveform is zero.
E – E avg
mpp = ----max
------------------------- × 100 (1.4.6)
E avg
E – E min
mnp = ----avg
------------------------- × 100 (1.4.7)
E avg
Where:
mpp = positive peak modulation (percent)
Emax = peak value of modulation envelope
mnp = negative peak modulation (percent)
Eavg = average envelope amplitude
E min = minimum envelope amplitude
When modulation exceeds 100 percent on the negative swing of the carrier, spurious signals
are emitted. It is possible to modulate an AM carrier asymmetrically; that is, to restrict modula-
tion in the negative direction to 100 percent, but to allow modulation in the positive direction to
exceed 100 percent without a significant loss of fidelity. In fact, many modulating signals nor-
mally exhibit asymmetry, most notably human speech waveforms.
The carrier wave represents the average amplitude of the envelope and, because it is the same
regardless of the presence or absence of modulation, the carrier transmits no information. The
information is carried by the sideband frequencies. The amplitude of the modulated envelope
may be expressed as [1]
E = E 0 + E 1 sin ( 2π f1 t + Φ1 ) + E 2 sin ( 2π f2 + Φ2 ) (1.4.8)
Where:
E = envelope amplitude
E0 = carrier wave crest value, V
y
E 1 = 2 × first sideband crest amplitude, V

f1 = frequency difference between the carrier and the first upper/lower sidebands
E2 = 2 × second sideband crest amplitude, V
f2 = frequency difference between the carrier and the second upper/lower sidebands
Φ1 = phase of the first sideband component
Φ2 = phase of the second sideband component
The amplitude-modulated signal can be generated in the basic circuit shown in Figure 1.4.3
1.4.2a Vestigial-Sideband Amplitude Modulation

Because the intelligence (modulating signal) of conventional AM transmission is identical in the
upper and lower sidebands, it is possible to eliminate one sideband and still convey the required
information. This scheme is implemented in vestigial-sideband AM (VSBAM). Complete elimi-
nation of one sideband (for example, the lower sideband) requires an ideal high-pass filter with
infinitely sharp cutoff. Such a filter is quite difficult to implement in any practical design.
VSBAM is a compromise technique wherein one sideband (typically the lower sideband) is
attenuated significantly. The result is a savings in occupied bandwidth and transmitter power.
VSBAM is used for analog television broadcast transmission and other applications. A typi-
cal bandwidth trace for a VSBAM TV transmitter is shown in Figure 1.4.4.
1.4.2b Single-Sideband Amplitude Modulation

The carrier in an AM signal does not convey any intelligence. All of the modulating information
is in the sidebands. It is possible, therefore, to suppress the carrier upon transmission, radiating
only one or both sidebands of the AM signal. The result is much greater efficiency at the trans-
mitter (that is, a reduction in the required transmitter power). Suppression of the carrier may be
accomplished with DSAM and SSBAM signals. Single-sideband suppressed carrier AM (SSB-
SC) is the most spectrum- and energy-efficient mode of AM transmission. Figure 1.4.5 shows
representative waveforms for suppressed carrier transmissions.
A waveform with carrier suppression differs from a modulated wave containing a carrier pri-
marily in that the envelope varies at twice the modulating frequency. In addition, it will be noted
that the SSB-SC wave has an apparent phase that reverses every time the modulating signal
passes through zero. The wave representing a single sideband consists of a number of frequency
components, one for each component in the original signal. Each of these components has an
amplitude proportional to the amplitude of the corresponding modulating component and a fre-
quency differing from that of the carrier by the modulating frequency. The result is that, in gen-
eral, the envelope amplitude of the single sideband signal increases with the degree of
modulation, and the envelope varies in amplitude in accordance with the difference frequencies
formed by the various frequency components of the single sideband interacting with each other.
An SSB-SC system is capable of transmitting a given intelligence within a frequency band
only half as wide as that required by a DSAM waveform. Furthermore, the SSB system saves
more than two-thirds of the transmission power because of the elimination of one sideband and
the carrier.
The drawback to suppressed carrier systems is the requirement for a more complicated
receiver. The carrier must be regenerated at the receiver to permit demodulation of the signal.
Also, in the case of SSBAM transmitters, it is usually necessary to generate the SSB signal in a
y
Plate blocking capacitor Output matching network

Output to
Class C antenna
RF amplifier
RFC
RF drive
input
Filament power
input
Class B
modulator
Audio
drive Modulation
transformer
Modulation
reactor
Audio
drive
Class B
modulator
DC power
supply input (HV)
Figure 1.4.3 Simplified diagram of a high-level amplitude-modulated amplifier.
1.0
Visual carrier
Chrominance
subcarrier
Relative maximum radiated field
strength (picture carrier = 1.0)
0.5 Aural carrier
0
3.579545 MHz
0.75 MHz (min)
4.2 MHz
4.5 MHz
6.0 MHz
Figure 1.4.4 Idealized amplitude characteristics of the FCC standard waveform for monochrome
and color TV transmission. (Adapted from FCC Rules, Sec. 73.699.)
low-power stage and then amplify the signal with a linear power amplifier to drive the antenna.
Linear amplifiers generally exhibit relatively low efficiency.
y
(a ) Positive
Modulating signal amplitude
Time
Negative
(b ) Positive
100% positive modulation
Carrier
Carrier amplitude
Time
100% negative modulation

Negative
(c ) Positive
Carrier amplitude
Time
Negative
(d ) Positive
Carrier amplitude
Time
Negative
Figure 1.4.5 Types of suppressed carried amplitude modulation: (a) the modulating signal, (b)
double-sideband AM, (c) double-sideband suppressed carrier AM, (d) single-sideband sup-
pressed carrier AM.
y
1.4.2c Quadrature Amplitude Modulation (QAM)

Single sideband transmission makes very efficient use of the spectrum; for example, two SSB
signals can be transmitted within the bandwidth normally required for a single DSB signal. How-
ever, DSB signals can achieve the same efficiency by means of quadrature amplitude modula-
tion (QAM), which permits two DSB signals to be transmitted and received simultaneously
using the same carrier frequency.
Two DSB signals coexist separately within the same bandwidth by virtue of the 90° phase
shift between them. The signals are, thus, said to be in quadrature. Demodulation uses two local
oscillator signals that are also in quadrature, i.e., a sine and a cosine signal.
The chief disadvantage of QAM is the need for a coherent local oscillator at the receiver
exactly in phase with the transmitter oscillator signal. Slight errors in phase or frequency can
cause both loss of signal and interference between the two signals (cochannel interference or
crosstalk).
1.4.3 Frequency Modulation

Frequency modulation is a technique whereby the phase angle or phase shift of a carrier is
varied by an applied modulating signal. The magnitude of frequency change of the carrier is
a direct function of the magnitude of the modulating signal. The rate at which the frequency
of the carrier is changed is a direct function of the frequency of the modulating signal. In
FM modulation, multiple pairs of sidebands are produced. The actual number of sidebands
that make up the modulated wave is determined by the modulation index (MI) of the system.
1.4.3a Modulation Index

The modulation index is a function of the frequency deviation of the system and the applied
modulating signal:
Fd
MI = ------ (1.4.9)
Mf
Where:
MI = the modulation index
F d = frequency deviation
Mf = modulating frequency
The higher the MI, the more sidebands produced. It follows that the higher the modulating fre-
quency for a given deviation, the fewer number of sidebands produced, but the greater their spac-
ing.
To determine the frequency spectrum of a transmitted FM waveform, it is necessary to com-
pute a Fourier series or Fourier expansion to show the actual signal components involved. This
work is difficult for a waveform of this type, because the integrals that must be performed in the
Fourier expansion or Fourier series are not easily solved. The result, however, is that the integral
y
produces a particular class of solution that is identified as the Bessel function, illustrated in Fig-
ure 1.4.6.
The carrier amplitude and phase, plus the sidebands, can be expressed mathematically by
making the modulation index the argument of a simplified Bessel function. The general expres-
sion is given from the following equations:
RF output voltage = E t = E c + S 1 u – S1 l + S 2 u – S 2 l + S 3 u – S 3 l + Snu – S nl
Carrier amplitude = E c = A [ J 0 ( M ) sin ω c( t )]
First-order upper sideband = S 1 u = J 1 ( M ) sin (ω c + ω m) t
First-order lower sideband = S1 l = J 1 ( M ) sin (ωc – ω m) t
Second-order upper sideband = S 2 u = J 2 ( M ) sin (ω c + 2 ω m) t
Second-order lower sideband = S 2 l = J 2 ( M ) sin (ω c – 2 ω m) t
Third-order upper sideband = S3 u = J 3 ( M ) sin (ω c + 3ω m) t
Third-order lower sideband = S3 l = J 3 ( M ) sin (ωc – 3 ω m) t
Nth-order upper sideband = Snu = J n ( M ) sin (ω c + nω m) t
Nth-order lower sideband = S nl = J n ( M ) sin (ω c – nω m) t
Where:
A = the unmodulated carrier amplitude constant
J0 = modulated carrier amplitude
J1, J2, J3... Jn = amplitudes of the nth-order sidebands
M = modulation index
ω c = 2 π F c , the carrier frequency
ω m = 2 π F m , the modulating frequency
Further supporting mathematics will show that an FM signal using the modulation indices that
occur in a wideband system will have a multitude of sidebands. From the purist point of view, all
sidebands would have to be transmitted, received, and demodulated to reconstruct the modulating
signal with complete accuracy. In practice, however, the channel bandwidths permitted practical
FM systems usually are sufficient to reconstruct the modulating signal with little discernible loss
in fidelity, or at least an acceptable loss in fidelity.
Figure 1.4.7 illustrates the frequency components present for a modulation index of 5. Figure
1.4.8 shows the components for an index of 15. Note that the number of significant sideband
y
Figure 1.4.6 Plot of Bessel functions of the first kind as a function of modulation index.
components becomes quite large with a high MI. This simple representation of a single-tone fre-
quency-modulated spectrum is useful for understanding the general nature of FM, and for mak-
ing tests and measurements. When typical modulation signals are applied, however, many more
sideband components are generated. These components vary to the extent that sideband energy
becomes distributed over the entire occupied bandwidth, rather than appearing at discrete fre-
quencies.
Although complex modulation of an FM carrier greatly increases the number of frequency
components present in the frequency-modulated wave, it does not, in general, widen the fre-
quency band occupied by the energy of the wave. To a first approximation, this band is still
roughly twice the sum of the maximum frequency deviation at the peak of the modulation cycle
plus the highest modulating frequency involved.
FM is not a simple frequency translation, as with AM, but involves the generation of entirely
new frequency components. In general, the new spectrum is much wider than the original modu-
lating signal. This greater bandwidth may be used to improve the signal-to-noise ratio (S/N) of
the transmission system. FM thereby makes it possible to exchange bandwidth for S/N enhance-
ment.
The power in an FM system is constant throughout the modulation process. The output power
is increased in the amplitude modulation system by the modulation process, but the FM system
simply distributes the power throughout the various frequency components that are produced by
y
Modulating frequency = 5 kHz

Deviation = 25 kHz
Modulation index = 5
Center frequency
0.5
0.4
Relative amplitude
0.3
0.2
0.1
0
fc - 75 kHz fc fc + 75 kHz
Carrier frequency
Figure 1.4.7 RF spectrum of a frequency-modulated signal with a modulation index of 5 and oper-
ating parameters as shown.
Modulating frequency = 5 kHz

Deviation = 75 kHz
Carrier approximately zero amplitude Modulation index = 15
0.5
0.4
Relative amplitude
0.3
0.2
0.1
0
fc - 75 kHz fc fc + 75 kHz
Carrier frequency
Figure 1.4.8 RF spectrum of a frequency-modulated signal with a modulation index of 15 and

operating parameters as shown.
modulation. During modulation, a wideband FM system does not have a high amount of energy
present in the carrier. Most of the energy will be found in the sum of the sidebands.
The constant-amplitude characteristic of FM greatly assists in capitalizing on the low noise
advantage of FM reception. Upon being received and amplified, the FM signal normally is
clipped to eliminate all amplitude variations above a certain threshold. This removes noise
picked up by the receiver as a result of man-made or atmospheric signals. It is not possible (gen-
erally speaking) for these random noise sources to change the frequency of the desired signal;
y
they can affect only its amplitude. The use of hard limiting in the receiver will strip off such
interference.
1.4.3b Phase Modulation

In a phase modulation (PM) system, intelligence is conveyed by varying the phase of the RF
wave. Phase modulation is similar in many respects to frequency modulation, except in the inter-
pretation of the modulation index. In the case of PM, the modulation index depends only on the
amplitude of the modulation; MI is independent of the frequency of the modulating signal. It is
apparent, therefore, that the phase-modulated wave contains the same sideband components as
the FM wave and, if the modulation indices in the two cases are the same, the relative amplitudes
of these different components also will be the same.
The modulation parameters of a PM system relate as follows:
Δ f = mp × f m (1.4.10)
Where:
Δ f = frequency deviation of the carrier
mp = phase shift of the carrier
fm = modulating frequency
In a phase-modulated wave, the phase shift mp is independent of the modulating frequency;
the frequency deviation Δ f is proportional to the modulating frequency. In contrast, with a fre-
quency-modulated wave, the frequency deviation is independent of modulating frequency.
Therefore, a frequency-modulated wave can be obtained from a phase modulator by making the
modulating voltage applied to the phase modulator inversely proportional to frequency. This can
be readily achieved in hardware.
1.4.3c Modifying FM Waves

When a frequency-modulated wave is passed through a harmonic generator, the effect is to
increase the modulation index by a factor equal to the frequency multiplication involved. Simi-
larly, if the frequency-modulated wave is passed through a frequency divider, the effect is to
reduce the modulation index by the factor of frequency division. Thus, the frequency compo-
nents contained in the wave—and, consequently, the bandwidth of the wave—will be increased
or decreased, respectively, by frequency multiplication or division. No distortion in the nature of
the modulation is introduced by the frequency change.
When an FM wave is translated in the frequency spectrum by heterodyne action, the modula-
tion index—hence the relative positions of the sideband frequencies and the bandwidths occu-
pied by them—remains unchanged.
1.4.3d Preemphasis and Deemphasis

The FM transmission/reception system offers significantly better noise-rejection characteristics
than AM. However, FM noise rejection is more favorable at low modulating frequencies than at
high frequencies because of the reduction in the number of sidebands at higher frequencies. To
y
offset this problem, the input signal to the FM transmitter may be preemphasized to increase the
amplitude of higher-frequency signal components in normal program material. FM receivers uti-
lize complementary deemphasis to produce flat overall system frequency response.
1.4.4 Pulse Modulation

The growth of digital processing and communications has led to the development of modulation
systems tailor-made for high-speed, spectrum-efficient transmission. In a pulse modulation sys-
tem, the unmodulated carrier usually consists of a series of recurrent pulses. Information is con-
veyed by modulating some parameter of the pulses, such as amplitude, duration, time of
occurrence, or shape. Pulse modulation is based on the sampling principle, which states that a
message waveform with a spectrum of finite width can be recovered from a set of discrete sam-
ples if the sampling rate is higher than twice the highest sampled frequency (the Nyquist crite-
ria). The samples of the input signal are used to modulate some characteristic of the carrier
pulses.
1.4.4a Digital Modulation Systems

Because of the nature of digital signals (on or off), it follows that the amplitude of the signal in a
pulse modulation system should be one of two heights (present or absent/positive or negative) for
maximum efficiency. Noise immunity is a significant advantage of such a system. It is necessary
for the receiving system to detect only the presence or absence (or polarity) of each transmitted
pulse to allow complete reconstruction of the original intelligence. The pulse shape and noise
level have minimal effect (to a point). Furthermore, if the waveform is to be transmitted over
long distances, it is possible to regenerate the original signal exactly for retransmission to the
next relay point. This feature is in striking contrast to analog modulation systems in which each
modulation step introduces some amount of noise and signal corruption.
In any practical digital data system, some corruption of the intelligence is likely to occur over
a sufficiently large span of time. Data encoding and manipulation schemes have been developed
to detect and correct or conceal such errors. The addition of error-correction features comes at
the expense of increased system overhead and (usually) slightly lower intelligence throughput.
1.4.4b Pulse Amplitude Modulation

Pulse amplitude modulation (PAM) is one of the simplest forms of data modulation. PAM
departs from conventional modulation systems in that the carrier exists as a series of pulses,
rather than as a continuous waveform. The amplitude of the pulse train is modified in accordance
with the applied modulating signal to convey intelligence, as illustrated in Figure 1.4.9. There are
two primary forms of PAM sampling:
• Natural sampling (or top sampling), where the modulated pulses follow the amplitude varia-
tion of the sampled time function during the sampling interval.
• Instantaneous sampling (or square-topped sampling), where the amplitude of the pulses is
determined by the instantaneous value of the sampled time function corresponding to a single
y
instant of the sampling interval. This “single instant” may be the center or edge of the sam-
pling interval.
There are two common methods of generating a PAM signal:
• Variation of the amplitude of a pulse sequence about a fixed nonzero value (or pedestal). This
approach constitutes double-sideband amplitude modulation.
• Double-polarity modulated pulses with no pedestal. This approach constitutes double-side-
band suppressed carrier modulation.
1.4.4c Pulse Time Modulation (PTM)

A number of modulating schemes have been developed to take advantage of the noise immunity
afforded by a constant amplitude modulating system. Pulse time modulation (PTM) is one of
those systems. In a PTM system, instantaneous samples of the intelligence are used to vary the
time of occurrence of some parameter of the pulsed carrier. Subsets of the PTM process include:
• Pulse duration modulation (PDM), where the time of occurrence of either the leading or trail-
ing edge of each pulse (or both pulses) is varied from its unmodulated position by samples of
the input modulating waveform. PDM also may be described as pulse length or pulse width
modulation (PWM).
• Pulse position modulation (PPM), where samples of the modulating input signal are used to
vary the position in time of pulses, relative to the unmodulated waveform. Several types of
pulse time modulation waveforms are shown in Figure 1.4.10.
• Pulse frequency modulation (PFM), where samples of the input signal are used to modulate
the frequency of a series of carrier pulses. The PFM process is illustrated in Figure 1.4.11.
It should be emphasized that all of the pulse modulation systems discussed thus far may be
used with both analog and digital input signals. Conversion is required for either signal into a
form that can be accepted by the pulse modulator.
1.4.4d Pulse Code Modulation

The pulse modulation systems discussed previously are unencoded systems. Pulse code modula-
tion (PCM) is a scheme wherein the input signal is quantized into discrete steps and then sam-
pled at regular intervals (as in conventional pulse modulation). In the quantization process, the
input signal is sampled to produce a code representing the instantaneous value of the input
within a predetermined range of values. Figure 1.4.12 illustrates the concept. Only certain dis-
crete levels are allowed in the quantization process. The code is then transmitted over the com-
munications system as a pattern of pulses.
Quantization inherently introduces an initial error in the amplitude of the samples taken. This
quantization error is reduced as the number of quantization steps is increased. In system design,
tradeoffs must be made regarding low quantization error, hardware complexity, and occupied
bandwidth. The greater the number of quantization steps, the wider the bandwidth required to
transmit the intelligence or, in the case of some signal sources, the slower the intelligence must
be transmitted.
y
(a) Positive
Waveform trace
Time
Negative
(b) Positive
Waveform trace
Time
Negative
Positive
(c)
Waveform trace
Time
Negative
(d) Positive
Waveform trace
0 Time
(e) Positive
Waveform trace
0 Time
Figure 1.4.9 Pulse amplitude modulation waveforms: (a) modulating signal; (b) square-topped
sampling, bipolar pulse train; (c) topped sampling, bipolar pulse train; (d) square-topped sampling,
unipolar pulse train; (e) top sampling, unipolar pulse train.
y
(a ) +
Signal amplitude
Time
(b ) +
Signal amplitude
Time
(c ) +
Signal amplitude
Time
(d ) +
Signal amplitude
Time
Figure 1.4.10 Pulse time modulation waveforms: (a) modulating signal and sample-and-hold (S/
H) waveforms, (b) sawtooth waveform added to S/H, (c) leading-edge PTM, (d) trailing-edge PTM.
In the classic design of a PCM encoder, the quantization steps are equal. The quantization
error (or quantization noise) usually can be reduced, however, through the use of nonuniform
spacing of levels. Smaller quantization steps are provided for weaker signals, and larger steps are
provided near the peak of large signals. Quantization noise is reduced by providing an encoder
that is matched to the level distribution (probability density) of the input signal.
Nonuniform quantization typically is realized in an encoder through processing of the input
(analog) signal to compress it to match the desired nonuniformity. After compression, the signal
is fed to a uniform quantization stage.
1.4.4e Delta Modulation

Delta modulation (DM) is a coding system that measures changes in the direction of the input
waveform, rather than the instantaneous value of the wave itself. Figure 1.4.13 illustrates the
y
Signal amplitude Time
+
Signal amplitude
Time
Figure 1.4.11 Pulse frequency modulation.
Quantization level of sample

5
4 Quantization levels
Amplitude
1
Sample
Time
Figure 1.4.12 The quantization process.
concept. The clock rate is assumed to be constant. Transmitted pulses from the pulse generator
are positive if the signal is changing in a positive direction; they are negative if the signal is
changing in a negative direction.
As with the PCM encoding system, quantization noise is a parameter of concern for DM.
Quantization noise can be reduced by increasing the sampling frequency (the pulse generator fre-
quency). The DM system has no fixed maximum (or minimum) signal amplitude. The limiting
factor is the slope of the sampled signal, which must not change by more than one level or step
during each pulse interval.
y
(a) +
Amplitude
Time
(b) +
Amplitude
Time
(c) +
Amplitude
Time
(d) +
Amplitude
Time
Figure 1.4.13 Delta modulation waveforms: (a) modulating signal, (b) quantized modulating sig-
nal, (c) pulse train, (d) resulting delta modulation waveform.
1.4.4f Digital Coding Systems

A number of methods exist to transmit digital signals over long distances in analog transmission
channels. Some of the more common systems include:
• Binary on-off keying (BOOK), a method by which a high-frequency sinusoidal signal is
switched on and off corresponding to 1 and 0 (on and off) periods in the input digital data
stream. In practice, the transmitted sinusoidal waveform does not start or stop abruptly, but
follows a predefined ramp up or down.
• Binary frequency-shift keying (BFSK), a modulation method in which a continuous wave is
transmitted that is shifted between two frequencies, representing 1s and 0s in the input data
stream. The BFSK signal may be generated by switching between two oscillators (set to dif-
ferent operating frequencies) or by applying a binary baseband signal to the input of a volt-
y
+ f2
Signal amplitude
Time
-
f1
Figure 1.4.14 Binary FSK waveform
0 character
+
Signal amplitude
Time
-
Clock period
T 2T 3T 4T
1 character
Figure 1.4.15 Binary PSK waveform.
age-controlled oscillator (VCO). The transmitted signals often are referred to as a mark
(binary digit 1) or a space (binary digit 0). Figure 1.4.14 illustrates the transmitted waveform
of a BFSK system.
• Binary phase-shift keying (BPSK), a modulating method in which the phase of the transmit-
ted wave is shifted 180° in synchronism with the input digital signal. The phase of the RF car-
rier is shifted by π/2 radians or –π/2 radians, depending upon whether the data bit is a 0 or a 1.
Figure 1.4.15 shows the BPSK transmitted waveform.
• Quadriphase-shift keying (QPSK), a modulation scheme similar to BPSK except that quater-
nary modulation is employed, rather than binary modulation. QPSK requires half the band-
width of BPSK for the same transmitted data rate.
y
1.4.5 Spread Spectrum Systems

As the name implies, a spread spectrum system requires a frequency range substantially greater
than the basic information-bearing signal. Spread spectrum systems have some or all of the fol-
lowing properties:
• Low interference to other communications systems
• Ability to reject high levels of external interference
• Immunity to jamming by hostile forces
• Provision for secure communications paths
• Operability over multiple RF paths
Spread spectrum systems operate with an entirely different set of requirements than transmis-
sion systems discussed previously. Conventional modulation methods are designed to provide for
the easiest possible reception and demodulation of the transmitted intelligence. The goals of
spread spectrum systems, on the other hand, are secure and reliable communications that cannot
be intercepted by unauthorized persons. The most common modulating and encoding techniques
used in spread spectrum communications include:
• Frequency hopping, where a random or pseudorandom number (PN) sequence is used to
change the carrier frequency of the transmitter. This approach has two basic variations: slow
frequency hopping, where the hopping rate is smaller than the data rate, and fast frequency
hopping, where the hopping rate is larger than the data rate. In a fast frequency-hopping sys-
tem, the transmission of a single piece of data occupies more than one frequency. Frequency-
hopping systems permit multiple-access capability to a given band of frequencies because
each transmitted signal occupies only a fraction of the total transmitted bandwidth.
• Time hopping, where a PN sequence is used to switch the position of a message-carrying
pulse within a series of frames.
• Message corruption, where a PN sequence is added to the message before modulation.
• Chirp spread spectrum, where linear frequency modulation of the main carrier is used to
spread the transmitted spectrum. This technique is commonly used in radar and also has been
applied to communications systems.
In a spread spectrum system, the signal power is divided over a large bandwidth. The signal,
therefore, has a small average power in any single narrowband slot. This means that a spread
spectrum system can share a given frequency band with one or more narrowband systems. Fur-
thermore, because of the low energy in any particular band, detection or interception of the trans-
mission is difficult.
1.4.6 References
1. Terman, F. E.: Radio Engineering, 3rd ed., McGraw-Hill, New York, N.Y., pg. 468, 1947.
y
1.4.7 Bibliography
Benson, K. B., and Jerry. C. Whitaker: Television and Audio Handbook for Technicians and
Engineers, McGraw-Hill, New York, N.Y., 1989.
Crutchfield, E. B. (ed.): NAB Engineering Handbook, 8th ed., National Association of Broad-
Jordan, Edward C. (ed.): Reference Data for Engineers: Radio, Electronics, Computer and Com-
munications, 7th ed., Howard W. Sams, Indianapolis, IN, 1985.
Kubichek, Robert: “Amplitude Modulation,” in The Electronics Handbook, Jerry C. Whitaker
(ed.), CRC Press, Boca Raton, Fla., pp. 1175–1187, 1996.
Seymour, Ken: “Frequency Modulation,” in The Electronics Handbook, Jerry C. Whitaker (ed.),
Whitaker, Jerry. C.: Radio Frequency Transmission Systems: Design and Operation, McGraw-
Ziemer, Rodger E.: “Pulse Modulation,” in The Electronics Handbook, Jerry C. Whitaker (ed.),
y
g g
Section
Radio Transmission Systems

2
Radio broadcasting has been around for a long time. Amplitude modulation (AM) was the first
system that permitted voice communications to take place. In this scheme, the magnitude of the
carrier wave is varied in accordance with the amplitude and frequency of the applied audio input.
The magnitude of the incoming signal determines the magnitude of the carrier wave, while the
frequency of the modulating signal determines the rate at which the carrier wave is varied. This
simple modulation system was predominant throughout the 1920s and 1930s. Frequency modu-
lation (FM) came into regular broadcast service during the 1940s.
These two basic approaches to communicating over long distances have served the broadcast
industry well. While the basic modulating schemes still exist today, numerous enhancements
have been made, including stereo operation and subcarrier programming. New technologies have
given radio broadcasters new ways to serve their listeners.
Technology has also changed the rules by which the AM and FM broadcasting systems devel-
oped. AM radio, as a technical system, offered limited audio fidelity but provided design engi-
neers with a system that allowed uncomplicated transmitters and simple, inexpensive receivers.
FM radio, on the other hand, offered excellent audio fidelity but required a more complex and
unstable transmitter (in the early days) and complex, expensive receivers. It is, therefore, no won-
der that AM radio flourished and FM remained relatively unchanged for at least two decades fol-
lowing its introduction.
Advancements in transmitter and receiver technology during the 1960s changed the picture
dramatically. Designers found ways of producing FM transmission equipment that was reliable
and capable of high power output. Receiver manufacturers then began producing compact, high-
quality receivers at prices consumers could afford. These developments began a significant
change in listening habits that is still unfolding today. As late as the mid-1970s, AM radio sta-
tions as a whole commanded the majority of all listeners in the U.S. By 1980, however, the tables
had turned. FM is now king, and AM is still struggling to keep up with its former stepchild.
As new digital radio systems and services emerge, the radio playing field will again be
changed—just how much no one is certain. Stay tuned.
2-1
y
2-2 Section Two
In This Section:
Chapter 2.1: Radio Broadcast Systems 2-5

Introduction 2-5
Standard Broadcasting 2-5
Modulation Techniques 2-5
Modulation of the AM Carrier Wave 2-6
Channel and Station Classifications 2-6
Field Strength 2-7
Propagation 2-7
Transmitters 2-8
High-Level AM Modulation 2-8
Pulse-Width Modulation 2-9
Solid-State Transmitters 2-11
FM Broadcasting 2-12
FM Broadcast Modulation Techniques 2-14
Modulation Circuits 2-14
Subsidiary Communications Authorization (SCA) 2-15
Frequency Allocations 2-16
Transmitters 2-16
FM Power Amplifiers 2-16
Tuning the Cavity 2-17
Transmitter Performance and Maintenance 2-18
Key System Measurements 2-18
Synchronous AM in FM Systems 2-19
Incidental Phase Modulation 2-19
Bibliography 2-20
Chapter 2.2: Radio STL Systems 2-21

Introduction 2-21
Radio vs. Landline Links 2-21
STL System Configuration 2-23
Digital STL vs. Analog STL Link 2-24
Digital STL Performance 2-25
Coding System 2-26
Analog Composite STL Transmitter and Receiver Characteristics 2-28
Components of an STL System 2-30
Transmitter and Receiver Requirements 2-30
Transmission Lines 2-33
Antenna System 2-34
Mounting Structures 2-36
Hardware Considerations 2-37
System Planning 2-37
Frequency Selection 2-37
Spectrum Considerations 2-38
Path Layout 2-38
Line of Sight 2-38
y
Radio Transmission Systems 2-3
STL Site Selection 2-38

Terrain Considerations 2-39
Path Reliability 2-42
Dealing with Problem Paths 2-45
Hot-Standby Redundant Systems 2-45
Pre-Installation Checkout 2-47
Planning Cable Runs 2-52
Installation 2-52
Checking System Operation 2-54
Troubleshooting 2-54
Operating the System 2-54
References 2-55
Bibliography 2-55
Chapter 2.3: Digital Radio Systems 2-57

Introduction 2-57
Technical Considerations 2-57
Data Reduction 2-58
System Planning Issues 2-58
Eureka 147/DAB 2-62
System Details 2-62
Spectrum Issues 2-64
In-Band Digital Radio 2-64
iBiquity Digital Radio 2-66
National Radio Systems Committee Test Results 2-66
FCC Actions 2-67
References 2-67
Chapter 2.4: IBOC AM Digital Radio System 2-69

Introduction 2-69
Glossary of Terms 2-70
iBiquity AM IBOC System 2-72
Test Program Issues 2-72
NRSC Test Program 2-75
Test Conclusions 2-76
Analog Compatibility 2-77
References 2-77
Chapter 2.5: IBOC FM Digital Radio System 2-79

Introduction 2-79
Glossary of Terms 2-80
iBiquity FM IBOC System 2-82
NRSC Test Issues 2-83
NRSC Test Program 2-84
Lab tests 2-85
Field tests 2-85
Test Conclusions and Recommendations 2-85
y
2-4 Section Two
References 2-86

Bordonaro, Dominic: “Reducing IPM in AM Transmitters,” Broadcast Engineering, Intertec
Publishing, Overland Park, Kan. May 1988.
Crutchfield, E. B. (ed): NAB Engineering Handbook, 8th ed., National Association of Broadcast-
ers, Washington, D.C., 1992.
Hauptstuek, Jim: “Interconnecting the Digital Chain,” NAB 1996 Broadcast Engineering Confer-
ence Proceedings, National Association of Broadcasters, Washington, D.C., pp. 360–358,
1996.
McClanahan, M. E.: “Aural Broadcast Auxiliary Links,” in NAB Engineering Handbook, 8th ed.,
E. B. Cructhfield (ed.), National Association of Broadcasters, Washington, D.C, pp. 671–
678, 1992.
NRSC: “DAB Subcommittee Evaluation of the iBiquity Digital Corporation IBOC System, Part
1—FM IBOC,” National Radio Systems Committee, Washington, D.C., November 29,
2001.
NRSC: “DAB Subcommittee Evaluation of the iBiquity Digital Corporation IBOC System, Part
2—AM IBOC,” National Radio Systems Committee, Washington, D.C., April 6, 2002.
Parker, Darryl: “TFT DMM92 Meets STL Requirements,” Radio World, Falls Church, VA, Octo-
ber 21, 1992.
Pohlmann, Ken: “Digital Radio and Television Broadcasting,” Principles of Digital Audio,
McGraw-Hill, New York, N.Y., 2000.
Rollins, William W., and Robert L. Band: “T1 Digital STL: Discrete vs. Composite Transmis-
sion,” NAB 1996 Broadcast Engineering Conference Proceedings, National Association of
Broadcasters, Washington, D.C., pp. 356–359, 1996.
Salek, Stanley: “Analysis of FM Booster System Configurations,” Proceedings of the 1992 NAB
Broadcast Engineering Conference, National Association of Broadcasters, Washington,
DC, April 1992.
Whitaker, Jerry C., (ed.): A Primer: Digital Aural Studio to Transmitter Links, TFT, Santa Clara,
CA, 1994.
Whitaker, Jerry C., and Skip. Pizzi: “Radio Electronic News Gathering and Field Production,” in
NAB Engineering Handbook, 8th ed., E. B. Cructhfield (ed.), National Association of
Broadcasters, Washington, D.C, pp. 1051–1072, 1992.
g g
Chapter
2.1
Radio Broadcast Systems
2.1.1 Introduction
Standard broadcasting refers to the transmission of voice and music received by the general pub-
lic in the 535–1705 kHz frequency band. Amplitude modulation is used to provide service rang-
ing from that needed for small communities to higher powered broadcast stations needed for
larger regional areas. The primary service area is defined as the area in which the ground or sur-
face-wave signal is not subject to objectionable interference or objectionable fading. The second-
ary service area refers to an area serviced by skywaves and not subject to objectionable
interference. Intermittent service area refers to an area receiving service from either a surface
wave or a skywave but beyond the primary service area and subject to some interference and fad-
ing.
2.1.2 Standard Broadcasting

The carrier frequencies for standard broadcasting in the U.S. (referred to internationally as
medium wave broadcasting) are designated in the Federal Communications Commission (FCC)
Rules and Regulations. A total of 117 carrier frequencies are allocated from 540 to 1700 kHz in
10-kHz intervals. Each carrier frequency is required by the FCC rules to deviate no more than
± 20 Hz from the allocated frequency to minimize heterodyning from two or more interfering
stations.
2.1.2a Modulation Techniques

Double-sideband full-carrier modulation, commonly called amplitude modulation (AM), is used
in standard broadcasting for sound transmission. Basically, the amplitude of an analog radio fre-
quency (RF) carrier is controlled by an analog audio frequency (AF) modulating signal. The
resulting RF waveform consists of a carrier wave plus two additional signals that are the result of
the modulation process: an upper-sideband signal, which is equal in frequency to the carrier plus
the audio modulating frequency, and a lower-sideband signal, which equals the carrier frequency
minus the audio modulating frequency.
2-5
y
2-6 Radio Transmission Systems
Modulation of the AM Carrier Wave

Figure 2.1.1 shows a fully amplitude modulated
signal. When the peak value of the modulated car-
rier wave (produced by a positive excursion of the
audio modulating signal) is twice the value of the
unmodulated carrier, and the minimum value of
the RF carrier reaches zero (produced by a nega-
tive excursion of the modulating signal), 100 per-
cent modulation occurs. If the transmitter is well
designed, with sufficient headroom, it is possible
Figure 2.1.1 Double-sideband amplitude
to exceed 100 percent positive modulation with
modulation.
an asymmetrical waveform, but the negative mod-
ulation may never exceed 100 percent (the RF
carrier at zero or completely cut off) without
causing distortion of the demodulated signal. The
power contained in the audio sidebands created in
the process of modulating the carrier 100 percent
is equal to 25 percent of the carrier power in both
the upper and lower sidebands. As a result, the
overall power transmitted during periods of 100
percent modulation is 150 percent of the unmodu- Figure 2.1.2 An amplitude-modulated
lated carrier power. (See Figure 2.1.2.) waveform.
The overall bandwidth is equal to twice the
highest audio modulating frequency. Although full fidelity is possible with amplitude modula-
tion, the FCC requires standard broadcast stations to limit the fidelity, thus restricting occupied
bandwidth of the transmitted signal. Typical modulation frequencies for voice and music range
from 50 Hz to 10 kHz. Each channel is generally thought of as 10 kHz in width, and thus the fre-
quency band is designated from 535 to 1705 kHz; however, when the modulation frequency
exceeds 5 kHz, the radio frequency bandwidth of the channel exceeds 10 kHz and adjacent chan-
nel interference may occur.
2.1.2b Channel and Station Classifications

In standard broadcast (AM), stations in the U.S. are classified by the FCC according to their
operating power, protection from interference, and hours of operation. A Class A station operates
with 10 to 50 kW of power servicing a large area with primary, secondary, and intermittent cov-
erage and is protected from interference both day and night. These stations are called clear chan-
nel stations because the channel is cleared of nighttime interference over a major portion of the
country. Class B stations operate full time with transmitter powers of 0.25 to 50 kW and are
designed to render primary service only over a principal center of population and contiguous
rural area. Whereas nearly all Class A stations operate with 50 kW, most Class B stations must
restrict their power to 5 kW or less to avoid interfering with other stations. Class B stations oper-
ating in the 1605 to 1705 kHz band are restricted to a power level of 10 kW daytime and 1 kW
nighttime. Class C stations operate on six designated channels (1230, 1240, 1340, 1400, 1450,
and 1490) with a maximum power of 1 kW or less full time and render primarily local service to
smaller communities. Class D stations operate on Class A or B frequencies with Class B trans-
y
mitter powers during daytime, but nighttime operation—if permitted at all—must be at low
power (less than 0.25 kW) with no protection from interference.
When a station is already limited by interference from other stations to a contour of higher
value than that normally protected for its class, this higher-value contour is the established pro-
tection standard for the station.
Although Class A stations cover large areas at night (approximately a 1000 km radius), the
nighttime coverage of Class B, C, and D stations is limited by interference from other stations,
electrical devices, and atmospheric conditions to a relatively small area. Class C stations, for
example, have an interference-free nighttime coverage radius of approximately 8 to 16 km. As a
result there may be large differences in the area that the station covers daytime vs. nighttime.
With thousands of AM stations licensed for operation by the FCC, interference—both day and
night—is a factor that significantly limits the service stations can provide. In the absence of
interference, a daytime signal strength of 2 mV/m is required for reception in populated towns
and cities, whereas a signal of 0.5 mV/m is generally acceptable in rural areas without large
amounts of man-made interference present. Secondary nighttime service is provided in areas
receiving a 0.5-mV/m signal 50 percent or more of the time without objectionable interference.
However, it should be noted that these limits apply to new stations and modifications to existing
stations. Nearly every station on the air was allocated prior to the implementation of these rules
with interference criteria that were less restrictive.
2.1.2c Field Strength

The field strength produced by a standard broadcast station is a key factor in determining the pri-
mary and secondary service areas and interference limitations of possible future radio stations.
The FCC and other organizations have conducted significant testing to determine what levels of
interference are acceptable from adjacent and co-channel stations and have used that data in
determining how close stations can be located to one another. If the transmitted power, antenna
radiation characteristics, and ground conductivity are known, the extent of coverage (or interfer-
ence) in a given direction for a particular station can be calculated with a high degree of accu-
racy. These calculations form the basis of the FCC's station allocation system and are based on
field intensities with the unit of measure as volts per meter (the voltage induced in an antenna 1
m in length).
2.1.2d Propagation
One of the major factors in the determination of field strength is the propagation characteristic,
described as the change in electric field intensity with an increase in distance from the broadcast
station antenna. This variation depends on a number of factors including frequency, distance,
surface dielectric constant, surface loss tangent, polarization, local topography, and time of day.
Generally speaking, surface-wave propagation occurs over shorter ranges both during day and
night periods. Skywave propagation in the AM broadcast band permits longer ranges and occurs
during night periods, and thus some stations must either reduce power or cease to operate at night
to avoid causing interference. Skywave propagation is much less predictable than surface-wave
propagation, because it necessarily involves some fading and less predictable field intensities,
and is most appropriately described in terms of statistics or the percentage of time a particular
field strength level is found.
y
Figure 2.1.3 Basic configuration for high-level amplitude modulation in the standard broadcast
band.
2.1.2e Transmitters
Standard AM broadcast transmitters range in power output from 5 W up to 50 kW units. Modern
transmitters utilize low-voltage, high-current metal-oxide-semiconductor field-effect transistor
(MOSFET) devices to generate the RF power. However, in a high-power transmitter, it may take
hundreds of transistors to generate the rated power. These transistors are combined into modules,
the outputs of which are combined to produce the output signal. This design approach has the
added benefit that in the event of a module failure, the transmitter usually continues to operate,
but at slightly reduced power.
High-Level AM Modulation
High-level anode modulation is the oldest and simplest way of generating a high power AM sig-
nal. In this system, the modulating signal is amplified and combined with the de supply source to
the anode of the final RF amplifier stage. The RF amplifier is normally operated class C. The
final stage of the modulator usually consists of a pair of tubes operating class B. A basic modula-
tor of this type is shown in Figure 2.1.3.
The RF signal is normally generated in a low-level oscillator. It is then amplified by one or
more solid-state or vacuum-tube stages to provide final RF drive at the appropriate frequency to
the grid of the final class C amplifier. The audio input is applied to an intermediate power ampli-
fier (usually solid state) and used to drive two class B (or class AB) push-pull output stages. The
final amplifiers provide the necessary modulating power to drive the final RF stage. For 100 per-
cent modulation, this modulating power is 50 percent of the actual carrier power.
y
Figure 2.1.4 Basic configuration of the PDM system.
The modulation transformer shown in Figure 2.1.3 does not usually carry the dc supply cur-
rent for the final RF amplifier. The modulation reactor and capacitor shown provide a means to
combine the audio signal voltage from the modulator with the dc supply to the final RF ampli-
fier. This arrangement eliminates the necessity of having dc current flow through the secondary
of the modulation transformer, which would result in magnetic losses and saturation effects. In
some transmitter designs, the modulation reactor is eliminated from the system, thanks to
improvements in transformer technology.
The RF amplifier normally operates class C with grid current drawn during positive peaks of
the cycle. Typical stage efficiency is 75 to 83 percent.
This type of system was popular in AM broadcasting for many years, primarily because of its
simplicity. The primary drawback is low overall system efficiency. The class B modulator tubes
cannot operate with greater than 50 percent efficiency. Still, with inexpensive electricity, this was
not considered to be a significant problem. As energy costs increased, however, more efficient
methods of generating high-power AM signals were developed. Increased efficiency normally
came at the expense of added technical complexity.
Pulse-Width Modulation
Pulse-width (also known as pulse-duration) modulation is one of the most popular systems
developed for modern vacuum-tube AM transmitters. Figure 2.1.4 shows a scheme for pulse-
width modulation identified as PDM (as patented by the Harris Corporation). The PDM system
works by utilizing a square wave switching system, illustrated in Fig. 2.1.5.
The PDM process begins with a signal generator (Figure 2.1.6). A 75-kHz sine wave is pro-
duced by an oscillator and used to drive a square wave generator, resulting in a simple 75-kHz
square wave. The square wave is then integrated, resulting in a triangular waveform, which is
mixed with the input audio in a summing circuit. The result is a triangular waveform that effec-
tively rides on the incoming audio (as shown in Figure 2.1.5). This composite signal is then
y
Figure 2.1.5 Principle waveforms of the PDM system.
Figure 2.1.6 Simplified block diagram of the PDM system.
applied to a threshold amplifier, which functions as a switch that is turned on whenever the value
of the input signal exceeds a certain limit. The result is a string of pulses in which the width of
pulse is proportional to the period of time the triangular waveform exceeds the threshold. The
pulse output is applied to an amplifier to obtain the necessary power to drive subsequent stages.
A filter eliminates whatever transients exist after the switching process is complete.
The PDM scheme is—in effect—a digital modulation system with the audio information
being sampled at a 75-kHz rate. The width of the pulses contain all the audio information. The
pulse-width-modulated signal is applied to a switch or modulator tube. The tube is simply turned
on, to a fully saturated state, or off in accordance with the instantaneous value of the pulse. When
the pulse goes positive, the modulator tube is turned on and the voltage across the tube drops to a
minimum. When the pulse returns to its minimum value, the modulator tube turns off.
y
This PDM signal becomes the power supply to the final RF amplifier tube. When the modula-
tor is switched on, the final amplifier will have current flow and RF will be generated. When the
switch or modulator tube goes off, the final amplifier current will cease. This system, in effect,
causes the final amplifier to operate in a highly efficient class D switching mode. A dc offset
voltage to the summing amplifier of Figure 2.1.6 is used to set the carrier (no modulation) level
of the transmitter.
A high degree of third-harmonic energy will exist at the output of the final amplifier because
of the switching-mode operation. This energy is eliminated by a third-harmonic trap. The result
is a stable amplifier system that normally operates in excess of 90 percent efficiency. The power
consumed by the modulator and its driver is usually a fraction of a full class B amplifier stage.
The damping diode shown in Figure 2.1.4 is used to prevent potentially damaging transient
overvoltages during the switching process. When the switching tube turns off the supply current
during a period when the final amplifier is conducting, the high current through the inductors
contained in the PDM filters could cause a large transient voltage to be generated. The energy in
the PDM filter is returned to the power supply by the damping diode. If no alternative route were
established, the energy would return by arcing through the modulator tube itself.
The pulse-width-modulation system makes it possible to completely eliminate audio fre-
quency transformers in the transmitter. The result is wide frequency response and low distortion.
It should be noted that variations on this amplifier and modulation scheme have been used by
other manufacturers for both standard broadcast and short-wave service.
Solid-State Transmitters
Solid-state transmitters make use of various schemes employing pulse modulation techniques.
One system is shown in simplified form in Figure 2.1.7a. This method of generating a modulated
RF signal employs a solid-state switch to swing back and forth between two voltage levels at the
carrier frequency. The result is a square-wave signal that is filtered to eliminate all components
except the fundamental frequency itself. A push-pull version of the circuit is shown in Figure
2.1.7b.
Figure 2.1.8 illustrates a class D switching system utilizing bipolar transistors that permits the
generation of a modulated carrier. Modern solid-state transmitters use field-effect-transistor
(FET) power devices as RF amplifiers. Basically, the dc supply to the RF amplifier stages is
switched on and off by an electronic switch in series with a filter. Operating in the class D mode
results in a composite signal similar to that generated by the vacuum-tube class D amplifier in
the PDM system (at much higher power). The individual solid-state amplifiers are typically com-
bined through a toroidal filter/combiner. The result is a group of low-power amplifiers operating
in parallel and combined to generate the required energy.
Transmitters well above 50 kW have been constructed using this design philosophy. They
have proved to be reliable and efficient. The parallel design provides users with automatic redun-
dancy and has ushered in the era of “graceful degradation” failures. In almost all solid-state
transmitters, the failure of a single power device will reduce the overall power output or modula-
tion capability of the system but will not take the transmitter off the air. This feature allows repair
of defective modules or components at times convenient to the service engineer. The negative
effects on system performance are usually negligible.
The audio performance of current technology solid-state transmitters is usually far superior to
vacuum-tube designs of comparable power levels. Frequency response is typically within ±1 dB
from 50 Hz to 10 kHz. Distortion is usually less than 1 percent at 95 percent modulation.
y
(a)
(b)
Figure 2.1.7 Solid-state switching RF amplifier: (a) basic system, (b) push-pull configuration.
2.1.3 FM Broadcasting
Frequency modulation utilizes the audio modulating signal to vary the frequency of the RF car-
rier. The greater the amplitude of the modulating frequency, the greater the frequency deviation
from the center carrier frequency. The rate of the frequency variation is a direct function of the
frequency of the audio modulating signal. In FM modulation, multiple pairs of sidebands are
produced. The actual number of sidebands that make up the modulated wave is determined by
the modulation index (MI) of the system. The modulation index is a function of the frequency
deviation of the system and the applied modulating signal.
As the MI increases there are more sidebands produced. As the modulating frequency
increases for a given maximum deviation, there will be a smaller number of sidebands spaced at
wider intervals. Unlike amplitude modulation, which has a percentage of modulation directly
proportional to the carrier power, the percentage of modulation in FM is generally referenced to
the maximum allowable occupied bandwidth set by regulation. For example, U.S. FM broadcast
stations are required to restrict frequency deviation to ±75 kHz from the main carrier. This is
referred to as 100 percent modulation for FM broadcast stations.
To determine the frequency spectrum of a transmitted FM waveform, it is necessary to com-
pute a Fourier series or Fourier expansion to show the actual signal components involved. This
work is difficult for a waveform of this type, as the integrals that must be performed in the Fou-
rier expansion or Fourier series are difficult to solve. The actual result is that the integral pro-
duces a particular class of solution that is identified as the Bessel function.
Supporting mathematics will show that an FM signal using the modulation indices that occur
in a broadcast system will have a multitude of sidebands. To achieve zero distortion, all side-
y
Figure 2.1.8 Bipolar transistor RF stage operating in a class D switching mode for AM service.
bands would have to be transmitted, received, and demodulated. However, in reality, the ±75-kHz
maximum deviation for FM broadcasting allows the transmission and reception of audio with
negligible distortion.
The power emitted by an FM transmitter is virtually constant, regardless of the modulating
signal. Additional noise or distortion of the amplitude of the waveform has virtually no effect on
the quality of the received audio. Thus, FM transmitters may utilize Class C type amplifiers,
which cause amplitude distortion but are inherently more efficient than Class A or Class B
amplifiers. In addition, atmospheric and man-made noise has little affect, since the receiver clips
all amplitude variations off the signal prior to demodulation.
2.1.3a Frequency Modulation

The monophonic system of FM broadcasting was developed to allow sound transmission of
voice and music for reception by the general public for audio frequencies from 50 to 15,000 Hz,
all to be contained within a ±75-kHz RF bandwidth. This technique provided higher fidelity
reception than was available with standard broadcast AM, along with less received noise and
interference. FM broadcasting in the U.S. is allocated the 88 to 108 MHz frequency band.
Pre-emphasis is employed in an FM broadcast transmitter to improve the received signal-to-
noise ratio. The pre-emphasis upper-frequency limit is based on a time constant of 75 μs, as
required by the FCC. Audio frequencies from 50 to 2120 Hz are transmitted with normal FM;
whereas audio frequencies from 2120 Hz to 15 kHz are emphasized with a larger modulation
y
Figure 2.1.9 Composite stereo FM signal.
index. There is significant signal-to-noise improvement at the receiver, which is equipped with a
matching de-emphasis circuit.
2.1.3b FM Broadcast Modulation Techniques

FM stereo was developed in 1961 to provide transmission capability for a left- and right-stereo
audio signal. Stereophonic transmission is accomplished by adding the left- and right-channel
stereo information together in the baseband signal. In addition, a left-minus-right channel is
added and frequency multiplexed on a subcarrier of 38 kHz using double sideband suppressed
carrier (DSSC) modulation. An unmodulated 19-kHz subcarrier is derived from the 38-kHz sub-
carrier to provide a synchronous demodulation reference for the stereophonic receiver. In the
receiver, a synchronous detector at 38 kHz recovers the left-minus-right channel information,
which is then combined with the left-plus-right channel information in sum and difference com-
biners to produce the original left-channel and right-channel signals. The system works well and,
because the baseband mono signal is a combination of both left and right audio, provides full
compatibility with monophonic FM receivers.
Modulation Circuits
Early FM transmitters used reactance modulators that operated at a low frequency. The output of
the modulator was then multiplied to reach the desired output frequency. This approach was
acceptable for monaural FM transmission but not for modern stereo systems or other applica-
tions that utilize subcarriers on the FM broadcast signal. Modern FM systems utilize direct mod-
ulation. That is, the frequency modulation occurs in a modulated oscillator that operates on a
center frequency equal to the desired transmitter output frequency. In stereo broadcast systems, a
composite FM signal is applied to the FM modulator. The basic parameters of this composite
signal are shown in Figure 2.1.9.
Various techniques have been developed to generate the direct-FM signal. One of the most
popular uses a variable-capacity diode as the reactive element in the oscillator. The modulating
signal is applied to the diode, which causes the capacitance of the device to vary as a function of
the magnitude of the modulating signal. Variations in the capacitance cause the frequency of the
oscillator to vary. Again, the magnitude of the frequency shift is proportional to the amplitude of
the modulating signal, and the rate of frequency shift is equal to the frequency of the modulating
signal.
The direct-FM modulator is one element of an FM transmitter exciter, which generates the
composite FM waveform. A block diagram of a complete FM exciter is shown in Figure 2.1.10.
Audio inputs of various types (stereo left and right signals, plus subcarrier programming, if used)
y
Figure 2.1.10 Block diagram of an FM exciter.
are buffered, filtered, and preemphasized before being summed to feed the modulated oscillator.
It should be noted that the oscillator is not normally coupled directly to a crystal but is a free-run-
ning oscillator adjusted as near as practical to the carrier frequency of the transmitter. The final
operating frequency is carefully maintained by an automatic frequency control system employ-
ing a phase-locked loop (PLL) tied to a reference crystal oscillator or frequency synthesizer.
A solid-state class C amplifier follows the modulated oscillator and increases the operating
power of the FM signal to 20–30 W. One or more subsequent amplifiers in the transmitter raise
the signal power to several hundred watts for application to the final power amplifier (PA) stage.
Nearly all current high-power FM transmitters utilize solid-state amplifiers up to the final RF
stage, which is generally a vacuum tube for operating powers of 15 kW and above. All stages
operate in the class C mode.
2.1.3c Subsidiary Communications Authorization (SCA)

In addition to its main channel and 38-kHz subcarrier stereo broadcasting, an FM station may
utilize additional subcarriers for private commercial use. These SCAs may be used in a variety of
ways, such as paging, data transmission, specialized foreign language programs, radio reading
services, utility load management, and background music. An FM stereo station may utilize mul-
tiplexed subcarriers within the range of 53–99 kHz with up to 20 percent total SCA modulation
of the main carrier using any form of modulation, including techniques for the transmission of
digital information. The only requirement is that the station does not exceed its occupied band-
width limitations or cause interference to itself or other stations.
A number of advanced modulation techniques have successfully been applied to transmission
of a high-speed digital data stream. These digital audio broadcasting (DAB) systems utilize in-
band on channel technique to provide high-quality digital audio without interfering with the cur-
rent analog FM broadcast signals.
y
2.1.3d Frequency Allocations

The 100 carrier frequencies for FM broadcast range from 88.1 to 107.9 MHz and are equally
spaced every 200 kHz. The channels from 88.1 to 91.9 MHz are reserved for educational and
noncommercial broadcasting, and those from 92.1 to 107.9 MHz for commercial broadcasting.
Each channel has a 200-kHz bandwidth. The maximum frequency swing under normal condi-
tions is ±75 kHz. Stations operating with an SCA may, under certain conditions, exceed this
level, but in no case may exceed a frequency swing of ±82.5 kHz. The carrier center frequency is
required to be maintained within ±2000 Hz. The frequencies used for FM broadcasting limit the
coverage to essentially line-of-sight distances. As a result, consistent FM coverage is limited to a
maximum receivable range of a few hundred kilometers depending on the antenna height above
average terrain (HAAT) and effective radiated power (ERP). The actual coverage area for a given
station can be reliably predicted after the power and the antenna height are known. Generally
speaking, either increasing the power or raising the antenna will increase the coverage area.
In FM broadcast, stations are classified by the FCC according to their maximum allowable
ERP and the transmitting antenna HAAT in their service area. Class A stations provide primary
service to a radius of about 28 km with 6000 W of ERP at a maximum HAAT of 100 m. The
most powerful class, Class C, operates with maximums of 100,000 W of ERP and heights up to
600 m with a primary coverage radius of over 92 km. All classes may operate at antenna heights
above those specified but must reduce the ERP accordingly. Stations may not exceed the maxi-
mum power specified, even if the antenna height is reduced. The classification of the station
determines the allowable distance to other cochannel and adjacent channel stations.
2.1.3e Transmitters
FM broadcast transmitters typically range in power output from 10 W to 50 kW. Digital technol-
ogy is being used extensively in the processing and exciter stages, allowing for precision modula-
tion control, whereas solid-state devices are predominately used for RF drivers and amplifiers up
to several kilowatts. High-power transmitters still rely on tubes in the final stages for the genera-
tion of RF power, but manufacturers are developing new devices and technologies that have
made high-power solid-state transmitters cost effective and reliable.
FM Power Amplifiers
The majority of high-power FM transmitters employ cavity designs. The 1/4-wavelength cavity
is the most common. The design is simple and straightforward. A number of variations can be
found in different transmitters, but the underlying theory of operation is the same.
The goal of any cavity amplifier is to simulate a resonant tank circuit at the operating fre-
quency and provide a means to couple the energy in the cavity to the transmission line. Because
of the operating frequencies involved (88 to 108 MHz), the elements of the “tank” take on unfa-
miliar forms.
A typical 1/4-wave cavity is shown in Figure 2.1.11. The plate of the tube connects directly to
the inner section (tube) of the plate-blocking capacitor. The blocking capacitor can be formed in
one of several ways. In at least one design, it is made by wrapping the outside surface of the inner
tube conductor with multiple layers of 8-in-wide and 0.005-in-thick polymide (Kapton) film.
The exhaust chimney-inner conductor forms the other element of the blocking capacitor. The
cavity walls form the outer conductor of the 1/4-wave transmission line circuit. The dc plate volt-
y
Figure 2.1.11 Physical layout of a common type of 1/4-wave PA cavity for FM broadcast service.
age is applied to the PA tube by a cable routed inside the exhaust chimney and inner tube conduc-
tor.
In the design shown in Figure 2.1.11, the screen-contact fingerstock ring mounts on a metal
plate that is insulated from the grounded-cavity deck by a Kapton blocker. This hardware makes
up the screen-blocker assembly. The dc screen voltage feeds to the fingerstock ring from under-
neath the cavity deck through an insulated feed through.
Some transmitters that employ the 1/4-wave cavity design use a grounded screen configura-
tion in which the screen contact fingerstock ring is connected directly to the grounded cavity
deck. The PA cathode then operates at below ground potential (in other words, at a negative volt-
age), establishing the required screen voltage for the tube.
The cavity design shown in Figure 2.1.11 is set up to be slightly shorter than a full 1/4-wave-
length at the operating frequency. This makes the load inductive and resonates the tube's output
capacity. Thus, the physically foreshortened shorted transmission line is resonated and electri-
cally lengthened to 1/4-wavelength.
Tuning the Cavity

Coarse tuning of the cavity is accomplished by adjusting the cavity length. The top of the cavity
(the cavity shorting deck) is fastened by screws or clamps and can be raised or lowered to set the
length of the assembly for the particular operating frequency.
Fine-tuning is accomplished by a variable-capacity plate-tuning control built into the cavity.
In the example, one plate of this capacitor, the stationary plate, is fastened to the inner conductor
just above the plate-blocking capacitor. The movable tuning plate is fastened to the cavity box,
the outer conductor, and is mechanically linked to the front-panel tuning control. This capacity
y
shunts the inner conductor to the outer conductor and can vary the electrical length and resonant
frequency of the cavity.
2.1.4 Transmitter Performance and Maintenance

Assessing the performance of an AM or FM transmitter has been greatly simplified by recent
advancements in RF system design. Most transmitters can be checked by measuring the audio
performance of the overall system. If the audio tests indicate good performance, chances are
good that all the RF parameters also meet specifications.
Both AM and FM broadcasting have certain limitations that prevent them from ever being a
completely transparent medium. Top on the list for FM is multipath distortion. In many loca-
tions, some degree of multipath is unavoidable. Running a close second is the practical audio
bandwidth limitation of the FM stereo multiplex system. The theoretical limit is 19 kHz; how-
ever, real-world filter designs result in a high-end passband of between 15 and 17 kHz.
For AM radio, the primary limitation is restricted bandwidth. Modern transmitters are capable
of faithfully transmitting wideband audio, but most receivers are not capable of reproducing it.
The problem is not just poor-quality AM radios but rather the high-noise environment of the AM
band. To reduce adjacent channel interference and atmospheric and artificial noise, receiver
designers have restricted the IF and audio bandwidth of their products. The result is a typical AM
receiver in which the audio response falls off rapidly beyond 3 kHz.
2.1.4a Key System Measurements

The procedures for measuring AM and FM transmitter performance vary widely depending on
the type of equipment being used. Some generalizations, however, can be made with respect to
equipment performance measurements that apply to most systems.
Frequency response is the actual deviation from a constant amplitude across a given span of
frequencies. Research has shown that frequency-response accuracy over the audio passband does
make an audible difference. Researchers exploring subtle differences in audio amplifier design
have found that errors as small as 0.2 dB can be heard. Therefore, flat response (strict adherence
to the 75-μs preemphasis curve) is an important goal in FM broadcasting. Typical performance
targets for an FM station are ±1 dB, 50 Hz to 15 kHz. Typical targets for an AM station are ±1
dB, 50 Hz to 10 kHz.
Total harmonic distortion (THD) is the creation by a nonlinear device of spurious signals har-
monically related to the applied audio waveform. Research has shown that although THD levels
greater than 1 percent are detectable during sinewave testing, listeners will tolerate somewhat
higher levels of THD on musical material. It must be noted that depending upon the setup of the
station's audio processor, distortion targets of 1 percent or less may be impossible to meet with
the processor on-line. Because of the highly competitive nature of broadcasting today, audio pro-
cessors cannot always be adjusted to provide for the purest reproduction of the incoming signal.
Audio processing aside, most FM transmitters will yield THD readings well below 1 percent
at frequencies up to 7.5 kHz. AM transmitters can typically produce distortion figures below 1.5
percent at frequencies up to 5 kHz at 95 percent modulation.
y
The THD test is sensitive to the noise floor of the system under test. If the system has a sig-
nal-to-noise ratio of 60 dB, the distortion analyzer's best possible reading will be greater than 0.1
percent (60 dB = 0.001 = 0.1 percent).
Intermodulation distortion (IMD) is the creation by a nonlinear device of spurious signals not
harmonically related to the audio waveform. These distortion components are sum-and-differ-
ence (beat notes) mixing products that research has shown are more objectionable to listeners
than even-harmonic distortion products. The IMD measurement is relatively impervious to the
noise floor of the system under test. IMD performance targets for AM and FM transmitters are
the same as the TH D targets mentioned previously.
Signal-to-noise ratio (S/N) is the amplitude difference, expressed in decibels, between a refer-
ence level audio signal and the system's residual noise and hum. In many cases, system noise is
the most difficult parameter to bring under control. An FM performance target of –70 dB per ste-
reo channel reflects reasonable exciter-transmitter performance. Most AM transmitters are capa-
ble of –60 dB S/N or better.
Separation is a specialized definition for signal crosstalk between the left and right channels
of a stereo system. The separation test is performed by feeding a test tone into one channel while
measuring leakage into the other channel (whose input is terminated with a 600-Ω wirewound
resistor, or other appropriate value). Typical performance targets for an FM station are –40 dB or
better from 50 Hz to 15 kHz.
Synchronous AM in FM Systems
The output spectrum of an FM transmitter contains many sideband frequency components, theo-
retically an infinite number. The practical considerations of transmitter design and frequency
allocation make it necessary to restrict the bandwidth of all FM broadcast signals. Bandwidth
restriction brings with it the undesirable side-effects of phase shifts through the transmission
chain, the generation of synchronous AM components, and distortion in the demodulated output
of a receiver.
In most medium- and high-power FM transmitters, the primary out-of-band filtering is per-
formed in the output cavity of the final stage. Previous stages in the transmitter (the exciter and
IPA) are designed to be broadband, or at least more broadband than the PA.
As the bandwidth of an FM transmission system is reduced, synchronous amplitude modula-
tion increases for a given carrier deviation (modulation). Synchronous AM is generated as tuned
circuits with finite bandwidth are swept by the frequency of modulation. The amount of synchro-
nous AM generated is dependent on tuning, which determines (to a large extent) the bandwidth
of the system. Figure 2.1.12 illustrates how synchronous AM is generated through modulation of
the FM carrier.
Incidental Phase Modulation

Incidental phase modulation (IPM) is defined as phase modulation produced by an AM trans-
mitter as a result of amplitude modulation. In other words, as an AM transmitter develops the
amplitude-modulated signal, it also produces a phase modulated, or PM, version of the audio as
well.
In theory, IPM is of little consequence for monophonic AM because FM and PM do not affect
the carrier amplitude. An envelope detector will, in theory at least, ignore IPM.
y
Figure 2.1.12 Synchronous AM is generated by one or more narrow-band stages in an FM trans-

mitter. The amount of AM produced is a function of the flatness of the composite RF system.
As a general rule, because IPM is a direct result of the modulation process, it can be gener-
ated in any stage that is influenced by modulation. The most common cause of IPM in plate-
modulated and pulse-modulated transmitters is improper neutralization of the final RF amplifier.
Adjusting the transmitter for minimum IPM is an accurate way of achieving proper neutraliza-
tion. The reverse is not always true, however, because some neutralization methods will not nec-
essarily result in the lowest amount of IPM.
Improper tuning of the IPA stage is the second most common cause of IPM. As modulation
changes the driver loading to the PA grid, the driver output also may change. The circuits that
feed the driver stage usually are isolated enough from the PA that they do not produce IPM. An
exception can be found when the power supply for the IPA is influenced by modulation. Such a
problem could be caused by a loss of capacitance in the high-voltage power supply itself.
2.1.5 Bibliography
Bordonaro, Dominic: “Reducing IPM in AM Transmitters,” Broadcast Engineering, Intertec
Publishing, Overland Park, Kan. May 1988.
Crutchfield, E. B. (ed): NAB Engineering Handbook, 8th ed., National Association of Broadcast-
ers, Washington, D.C., 1992.
g g
Chapter
2.2
Radio STL Systems
2.2.1 Introduction
One of the major concerns in the design and operation of a radio broadcasting facility is the
means by which the program audio from the studio is conveyed to the transmitter site. As illus-
trated in Figure 2.2.1, this link represents an important element in the overall reliability of the
transmission chain. Furthermore, as digital technology continues to move into daily radio station
operation, the studio to transmitter link (STL) must become as transparent as possible. An infe-
rior link will impose an unacceptable limit on overall audio quality. The requirements for reli-
ability and transparent program relay have led to the development of new STL systems based on
digital technology.
Changes in FCC broadcast ownership rules and the popularity of local marketing agreements
(LMAs) have reshaped radio broadcasting. The need for high-quality audio programming is one
outgrowth of new competition and new alliances. STL systems are an important component of
these audio improvement efforts. Furthermore, increasing numbers of stations are using intercity
relay (ICR) facilities to share programming. Unfortunately, in many areas of the United States,
the demand for 950 MHz STL channels has far outstripped the available supply. The Part 74
bandwidth allocations for STL systems, therefore, necessitate highly efficient designs to meet
the needs of radio broadcasters today.
Market demand for STL systems is not limited to North America. The commercialization of
radio broadcasting in Europe and elsewhere has led to increased use of radio links to relay pro-
gramming from one studio to the next, and from the studio to the transmitter site. In some areas,
repeated use of the same frequency places stringent demands on system design.
2.2.1a Radio vs. Landline Links

The arguments over which approach—radio STL or landline (telco)—is the better way to convey
program audio from the studio to the transmitter is as old as radio broadcasting itself. Now, how-
ever, a new element has been added to the equation that makes the landline option considerably
more attractive. That new element is digital technology.
For years, broadcasters were limited in their choice of STL equipment to two main options:
950 MHz radio, and equalized analog telephone lines. Today, a third approach, digital transmis-
2-21
y
Studio site Transmitter site
Figure 2.2.1 The basic function of a studio-to-transmitter link. (After [2].)
Figure 2.2.2 Discrete channel landline digital audio STL system. (After [1].)
sion, is rapidly becoming the system of choice for many radio stations [1]. The decision to use
some form of digital landline technology may be based on several factors, including:
• Necessity. The station has no line-of-sight to the transmitter, or suitable frequencies are
unavailable.
• Sound quality. A digital landline STL can sound better than even the best analog systems.
• Cost. A single leased data line can cost less than multiple leased analog lines.
Whatever the reasons, a broadcaster who decides to use digital transmission must then choose
what type of system to implement. Digital STL systems can be designed either to transmit dis-
crete left and right channel stereo, as shown in Figure 2.2.2, or to transmit a composite stereo sig-
nal, illustrated in Figure 2.2.3.
One of the advantages to a broadcaster in using a digital landline STL is the ability to multi-
plex several signals together. SCAs, SAP channels, transmitter remote control signals, data, and
voice can all be combined with the broadcast audio signal for transmission on a single circuit.
The duplex nature of data links permits use of the same system for both STL and transmitter-to-
studio (TSL) functions as well.
y
Figure 2.2.3 Composite landline digital audio STL system. (After [1].)
2.2.2 STL System Configuration

The frequency range used by aural STL systems in the U.S. is 944.5 to 951.5 MHz. Frequencies
ranging from 200 to 940 MHz are used in other parts of the world. STL systems are also operated
above these traditional bands, however, such links often face difficulties associated with
increased susceptibility to signal loss as a result of precipitation, foliage, and fog, which limits
usage to relatively short path lengths. Frequency modulation is used for analog STL systems.
The vast majority of radio STL systems in operation today carry a composite stereo baseband
signal from the studio to the transmitter. An alternative approach involves the use of one or two
narrowband STL systems. The two approaches are compared in Figure 2.2.4.
The monaural STL typically has an audio bandwidth of up to 15 kHz for program signals, and
usually can accommodate a single FM subcarrier at approximately 39 kHz. For stereo or dual
mono installations, a common antenna can be used at the transmitter and at the receiver. At the
transmit site, a combiner/isolator adds the signals and prevents the generation of intermodulation
products. At the receive site, a low-noise antenna splitter feeds the individual receivers.
The composite STL has sufficient bandwidth (220 kHz) to accommodate the output of a ste-
reo FM generator and usually several separate subcarriers. A low-pass filter generally is used at
the output of the receiver to separate the signals to be broadcast over-the-air from the non-broad-
cast (or closed-circuit) signals. A typical configuration is illustrated in Figure 2.2.5. The com-
posite STL provides superior stereo performance compared to dual monaural radio links in
several respects, including:
• Elimination of interchannel phase and amplitude errors that can arise in a dual channel sys-
tem.
• Elimination of audio headroom considerations because the STL input signal has already been
passed through the station's audio processing system and the stereo generator, which are
located at the studio.
For FM applications, the composite signal from the STL receiver is fed directly into the FM
transmitter exciter.
Because AM stereo exciters require discrete right and left channel audio inputs, a comple-
mentary stereo decoder is used to recover the left and right channel audio if a composite STL
(fed by a stereo generator) is used.
y
(a) Program Studio site Transmitter site

audio
input RF
L STL STL output
Stereo RF link Exciter
generator transmitter receiver
R
(b) Studio site Transmitter site

Left Left
audio audio
input Automatic STL STL Limiting output
gain transmitter receiver amplifier
control
RF link RF power
divider
Right Right
audio audio
input Automatic STL STL Limiting output
gain transmitter receiver amplifier
control
Figure 2.2.4 Comparison of STL systems: (a) composite transmitter-receiver system, (b) dual
monaural transmitter-receiver system. (After [2].)
2.2.2a Digital STL vs. Analog STL Link

Digital signal encoding of audio frequency information provides a number of advantages over
conventional analog systems. The principal strengths of digital coding for transmission and stor-
age vs. analog methods include the following:
• Greater immunity to noise and interference in the transmission path.
• Elimination of transmission-path-dependent distortion mechanisms, such as harmonic distor-
tion, intermodulation distortion, and crosstalk.
• Efficient use of baseband and RF spectrum.
• Efficient and predictable regeneration of the digital signal.
• Easy and effective encryption for security and coding purposes.
The use of digital coding and modulation in an STL system provides the user a number of
benefits in terms of performance and path reliability. By its nature, digital coding offers a more
robust link. As long as the received carrier-to-noise ratio (C/N) exceed a defined minimum
value, the bit error rate remains relatively constant, and the decoded baseband signal is an exact
duplicate of the input signal at the studio. (See Figure 2.2.6.) With adequate error correction, sig-
nals can be received error-free. With a digital STL, multiple links are practical because signal
degradations are not additive, as in an analog system.
y
(a )
ed
itul Program channel (15 kHz high-frequency limit)
p
m
a
eiv Subcarrier (39 kHz center frequency)
ta
elR
0 20 40 60 80 100 120 140 160 180 200

Baseband frequency
(b ) L+R main channel

L-R subchannel
ed Subcarriers
ulit
p Broadcast signals Non-broadcast signals
m
a
eiv 67 kHz center 152 kHz center
ta 92 kHz center 185 kHz center
le
R
0 20 40 60 80 100 120 140 160 180 200

38 kHz center Baseband frequency
19 kHz pilot
Figure 2.2.5 Baseband spectrum of STL systems: (a) monaural, (b) composite. (After [2].)
It is fair to point out that digitization of the input audio signal always brings with it a measure
of degradation (quantization errors), but with the high sampling rates typically used for profes-
sional audio applications, such degradation is minor and almost always inaudible.
The process of quantization is illustrated in Figure 2.2.7. The sampling rate and quality of the
sampling circuit determine, in large part, the overall quality of the digital system. A properly
operating transmission channel can be assumed to provide error-free throughput. This being the
case, the digital signal can be regenerated at the receiving point as an exact duplicate of the input
waveform. Figure 2.2.8 shows a general representation of a digital communications channel. In
the case of a radio link, such as an STL, the transmission medium is analog in nature (FM). The
circuits used to excite the FM modulator, however, are essentially identical to those used for an
all-digital link, such as fiber optic cable.
The functions of the encoder and decoder, shown in Figure 2.2.8a, usually are formed into a
single device, or set of devices (a chip set), known as a codec (coding and decoding device). At
the transmission end, the codec provides the necessary filtering to band-limit the analog signal to
avoid aliasing, thereby preventing analog-to-digital (A/D) conversion errors. At the receiver, the
codec performs the reciprocal digital-to-analog (D/A) conversion and interpolates (smooths) the
resulting analog waveform.
Digital STL Performance

The benefits of a digital STL for radio broadcasters can best be appreciated by comparing the
performance of a digital system and an analog system. A digital STL typically permits broad-
y
Digital system Analog system

90
80
)
70
B
d
60
(
o
it
a
r
e
is
o
50
n
-
o
t
-
l
a
40
n
ig
S 30
20
10
0
1.0 2.0 3.0 5.0 10 20 30 50 100 200 300 500 1E3
Received RF signal ( V)
Figure 2.2.6 The benefits of digital vs. analog STL systems in terms of S/N and received RF level.
(After [2].)
casters to extend the fade margin of an existing analog link by 20 dB or more. Furthermore,
audio signal-to-noise (S/N) improvements of at least 10 dB can be expected for a given RF signal
strength. Alternatively, for the same S/N, the maximum possible path distance of a given com-
posite STL transmitter and receiver can be extended. These features could, in some cases, make
the difference between a one-hop system and a two-hop system.
The spectrum-efficiency of a digital STL is of great importance today in highly congested
markets. The system may, for example, be capable of relaying four program channels and two
voice-grade channels. The use of digital coding also makes the signals more tolerant of co-chan-
nel interference than a comparable analog STL.
Coding System
Several approaches may be used to digitize or encode the input audio signals. The complexity of
the method used is a function of the availability of processing power and encoder memory, and of
the resulting delay incurred during the encoding/decoding process. For a real-time function such
as an STL, significant encoding/decoding delay is unacceptable. Pulse code modulation (PCM)
is a common scheme that meets the requirements for speed and accuracy. In the PCM process,
the sampled analog values of the input waveform are coded into unique and discrete values. This
quantization may be uniform, as illustrated in Figure 2.2.7, or nonuniform. With nonuniform
quantization, compression at the coder and subsequent expansion at the decoder is performed.
By using larger quantization steps for high energy signals and smaller steps for low energy sig-
nals, efficient use is made of the data bits, while maintaining a specified signal-to-quantization
noise level. This process is known as companding (compression and expansion).
y
(a) Q uantized output (Y)
0111
0110
0101
0100
0011
0010
0000 0001
1111 Analog input signal (X)

1110
1101
1100
1011
1010
1001
1000
(b)
Analog input
Q uantized output
ed
uit
lp
m
al Time
an
igS
ro
rr
E
Figure 2.2.7 Quantization of an input signal: (a) quantization steps, (b) quantization error signal.
(After [2].)
PCM encoding, in a simple real-time system, provides a high-speed string of discrete digital
values that represent the input audio waveform. Each value is independent of all previous sam-
ples. No encoder memory is required. This approach, while simple and fast, is not particularly
efficient insofar as the transmission channel is concerned. There are many redundancies in any
given input signal. By eliminating the redundancies, and taking advantage of the masking effects
of human hearing, greater transmission efficiency can be realized. Viewed from another perspec-
tive, for a given radio transmission bandwidth, more information can be transferred by using a
compression system that removes nonessential data bits.
(a )
Transmitter
Analog Anti- A/D
input aliasing converter Encoder
filter
Digital
channel
Receiver
Analog Smoothing and D/A
output interpolation Decoder
filter converter
(b )
Transcoding Analog
transmitter Modulator communications Demodulator
channel
Input signal Output signal Transcoding

receiver
Figure 2.2.8 Digital transmission system: (a) coding/decoding functions, (b) overall communica-
tions link. (After [2].)
2.2.2b Analog Composite STL Transmitter and Receiver Characteristics

As mentioned previously, the composite aural STL is the workhorse of the radio industry. A
number of hardware variations can be found in transmitter and receiver systems, but all are
designed to accomplish basically the same goal: relay of a baseband signal from the studio site to
the transmitter. Figure 2.2.9 illustrates the basic STL configurations. A block diagram of a repre-
sentative composite STL is given in Figure 2.2.10.
The RF carrier is generated by a voltage-controlled oscillator (VCO) that is phase-locked to a
modulating VCO operating at an intermediate frequency of approximately 70 MHz. Direct com-
posite FM modulation takes place at the IF VCO, which is also phase-locked to an oven-con-
trolled crystal oscillator that serves as the master timebase for the transmitter. Because frequency
multiplication is not used, the RF carrier is an exact reproduction of the IF VCO. Therefore, a
number of potential compromises resulting from frequency-multiplication and post-heterodyne
filtering are eliminated, including:
• Degradation of S/N
• Generation of spurious signals
• Degraded stereo performance
y
Figure 2.2.9 Analog composite STL link for FM radio applications. (After [2].)
M ix er/ RF power
up-converter amplifier Antenna
Metering Status Local

oscillator
circuitry inputs and multiplier
Input signals
Composite Synthesized
MUX 1 Baseband direct-FM
MUX 2 processor oscillator
Figure 2.2.10 Block diagram of a composite STL transmitter. (After [2].)
IF modulation provides excellent overall stereo performance, as documented in Table 2.2.1.

In this design, the operating frequency of the transmitter is programmable by DIP switches in
12.5 kHz steps, so that frequency changes can be made in the field. The composite baseband sig-
nal accommodates the full stereo baseband, including two SCA and/or MUX channels. These
features permit all processing equipment for analog operation to be located at the studio. An
illustration of the composite baseband of such a system is shown in Figure 2.2.11.
y
Table 2.2.1 Specifications for a Typical Composite STL System (After [2].)
Parameter Specification
Power output 6 to 8 W
Frequency stability ± 0.0002%, 0 to 50° C
Spurious emissions 60 dB below maximum carrier power
Baseband frequency response ± 0.1 dB or less, 30 Hz to 75 kHz
Stereo separation Greater than 55 dB at 1 kHz
Total harmonic distortion 0.02%, 75 μs deemphasis
S/N 85 dB below ± 75 kHz deviation, 75 μs deemphasis
Nonlinear crosstalk 50 dB or less
Subchannel-to-main crosstalk 60 dB or less
L+R main channel

L-R subchannel
ed Subcarriers
uit
lp
m Broadcast signals Non-broadcast signals
a 67 kHz center 152 kHz center
ev
it 92 kHz center 185 kHz center
lae
R
0 20 40 60 80 100 120 140 160 180 200

Baseband frequency
38 kHz center
19 kHz pilot
Figure 2.2.11 The baseband of a composite STL transmitter. (After [2].)
A block diagram of the companion composite STL receiver is shown in Figure 2.2.12. Like
the transmitter, the receiver is user-programmable in 12.5 kHz steps through the use of internal
DIP switches. The front-end uses cascaded high-Q cavity filters and surface acoustic wave
(SAW) IF filters to provide high selectivity and phase linearity. Triple conversion IF is used to
feed a pulse-counting discriminator for linear baseband demodulation.
2.2.2c Components of an STL System

Any STL installation is only as good as the hardware used to interconnect the link. All compo-
nents, from the transmitter to the output connectors, must be carefully chosen and properly
installed. A well-designed system will provide years of trouble-free service. A poorly-designed
system will cause problems on a regular basis.
Transmitter and Receiver Requirements

When choosing an STL transmitter, it will be necessary to select a unit that will deliver sufficient
power to overcome the losses determined by path gain/loss calculations. Figure 2.2.13 provides a
starting point for system planning. In cases where additional power is needed, a separate power
y
RF input
Preselector/ First 70 MHz Second

preamplifier mixer IF amplifier mixer
First local Second local Third local Third 10.7 MHz

oscillator oscillator oscillator mixer IF amplifier
Output
Composite signals
3 MHz FM baseband
IF amplifier demodulator processor
Status input signals Mute

circuitry
Metering
circuitry
Figure 2.2.12 Block diagram of a composite STL receiver. (After [2].)
amplifier can be used. For convenience in manipulating figures, the transmitter power output
should be converted to gain in decibels above a 1 mW reference (dBm). (See Table 2.2.2.)
When choosing an STL receiver, specifications should be carefully analyzed, particularly
receiver sensitivity. This figure, necessary in STL path calculations, is usually specified as a sig-
nal level required for a specified S/N. This value should be converted to dBm. (See Table 2.2.3.)
For example, a receiver may require 100 μV for 60 dB S/N. This is equivalent to –66.9 dBm. In
receiver design, sensitivity, S/N, selectivity, and the method of demodulation are determining
factors of receiver quality. The use of SAW filters provide sharper selectivity and more linear
phase response. These attributes yield better stereo separation and lower distortion. A pulse-
counting discriminator also provides low distortion and accurate demodulation of the received
signal. Phase-linear lowpass filtering is critical for best stereo separation.
A low-noise RF preamplifier may be added to the system when the received signal level is
low. For best performance, the preamplifier should be mounted directly at the receive antenna.
Care must be taken, however, to prevent overloading the receiver front-end by unwanted, and
often strong, interfering signals.
In areas of frequency congestion, narrowband receivers are important in preventing interfer-
ence from other transmitters. STL manufacturers have responded to the needs of broadcasters in
congested RF environments by providing suitable narrowband STL systems. Such receivers typ-
ically incorporate bandpass cavity filters, helical resonators, or mechanical coaxial filters. SAW
filters and ceramic filters in IF stages also may be included to improve selectivity.
y
Figure 2.2.13 Nomogram to quickly determine if a path is viable. (After [2].)
Choose a monaural, composite, or digital STL as needed. Compare performance specifica-

tions and carefully analyze the long-term benefits of each system. Consider what the STL will
need to handle two or five years from now.
y
Table2.2.2 Conversion of Watts to dBm (50 Ω Impedance)
Power in Watts dBm Power in Watts dBm

5.0 37.0 18.0 42.6
5.5 37.42 19.0 42.8
6.0 37.79 20.0 43.0
6.5 38.12 25.0 43.9
7.0 38.46 30.0 44.8
7.5 38.75 35.0 45.4
8.0 39.0 40.0 46.0
8.5 39.3 45.0 46.5
9.0 39.55 50.0 47.0
9.5 39.78 55.0 47.4
10.0 40.0 60.0 47.8
10.5 40.2 65.0 48.1
11.0 40.4 70.0 48.5
12.0 40.8 75.0 48.8
13.0 41.1 80.0 49.0
14.0 41.5 85.0 49.3
15.0 41.8 90.0 49.5
16.0 42.0 95.0 49.9
17.0 42.3 100.0 50
Transmission Lines
Figure 2.2.14 shows the primary hardware elements required for an aural STL. Transmission line
sections, connections, and strain-relief provisions are important for long-term reliability. The
main criteria in the selection of transmission line include the following:
• Amount of signal attenuation
• Physical parameters (dielectric material and size)
• Purchase and installation cost
In general, the larger the diameter of the transmission line, the lower the loss, and the greater
the cost of the line. Loss is also affected by the type of dielectric material used. The most com-
mon types of dielectric are air and foam. Air dielectric cable typically requires pressurization and
is, therefore, seldom used for 950 MHz installations. For the purpose of gain/loss calculations,
cable loss for a particular installation can be determined from the transmission line manufac-
turer's specification table, given in decibels per unit of measure. Attenuation for common types
of coaxial line are given in Table 2.2.4
Other electrical and mechanical specifications (mainly cable size and impedance) must be
compatible with the transmitter, receiver, and antennas to be used. Connector loss must also be
considered. It is important to minimize each potential source of signal loss when assembling the
system. There are no “minor details” in the installation of an STL.
Stain relief must be provided on each end of the cable run (at the transmitter and at the
receiver). So-called pigtail or jumper cables are commonly used for this purpose. They permit
movement without straining cable and chassis connections. Because the pigtails commonly are
terminated with N-type male connectors on both ends, the main transmission line must be con-
figured with female N-type connectors on both ends if a pair of pigtails are used.
y
Table 2.2.3 Conversion of Microvolts to dBm (50 Ω Impedance)
Microvolts dBm Microvolts dBm

0.10 –127 40 –74.9
0.12 –125.25 45 –73.9
0.14 –124 50 –72.9
0.16 –122.9 60 –71.25
0.18 –121.9 70 –70
0.20 –120.9 80 –68.9
0.25 –119 90 –67.9
0.30 –117.25 100 –66.9
0.35 –116 120 –65.25
0.40 –115 140 –64
0.60 –111.25 160 –62.9
0.70 –110 180 –61.9
0.80 –108.9 200 –60.9
0.90 –107.9 250 –59
1.0 –106.9 300 –57.5
1.2 –105.25 350 –56
1.4 –104 400 –54.9
1.6 –102.9 450 –53.9
1.8 –101.9 500 –52.9
2.0 –100.9 600 –51.25
2.5 –99 700 –50
3.0 –97.5 800 –49
3.5 –96 900 –48
4.0 –95 1000 –46.9
4.5 –93.9 1200 –45.25
5.0 –92.9 1400 –44
6.0 –91.25 1600 –42.9
7.0 –90 1800 –41.9
8.0 –88.9 2000 –40.9
9.0 –87.9 2500 –39
10 –86.9 3000 –37.25
1.1 –86 3500 –36
12 –85.25 4000 –34.9
14 –84 4500 –33.9
16 –82.9 5000 –33
18 –81.9 6000 –31.25
20 –80.9 7000 –30
25 –79 8000 –28.9
30 –77.25 9000 –27.9
35 –76 10,000 –26.9
Antenna System
At the frequencies commonly used for STL operation, antennas can readily focus the signal into
a narrow beam. This focusing, required for point-to-point communications, provides high gain at
both the transmitter and the receiver. Several types of antennas are available, including parabolic
and parabolic section. Parabolic antennas are available in solid or grid styles, while parabolic
section antennas usually are grid-type in design. Antenna models differ in a number of respects,
including:
y
Pigtail Pigtail
Antennas
Transmission line Transmission line
STL STL
transmitter receiver
Pigtail Pigtail
Figure 2.2.14 Principle physical components of an STL system. (After [2].)
Table 2.2.4 Typical Coaxial Transmission Line Loss at 950 MHz
Cable Type Loss (dB/100 Feet)

RG-8/U 8.5
RG-218/U 3.8
1/2-in foam dielectric 3.0
7/8-in foam dielectiic 2.0
1/2-in low-loss foam dielectric 2.4
7/8-in low-loss foam dielectric 1.4
1-5/8-in foam dielectric 1.4
1/2-in air dielectiic 2.7
7/8-in air dielectric 1.4
1-5/8-in air dielectric 0.7
• Gain (directly proportional to size)

• Operating frequency range
• Polarization (most antennas can be set for either horizontal or vertical polarization using uni-
versal mounting hardware kits)
• Beamwidth
• Front-to-back ratio
• Windloading
• Structural strength
Environmental factors may also need to be considered so that suitable antenna accessories,
such as deicers and/or radomes, are specified.
Antenna gain is specified in decibels referenced to an isotropic antenna (dBi) or decibels ref-
erenced to a dipole antenna (dBd). Conversion from dBd to dBi is as follows:
dBi = 2.2 + dBd (2.2.1)
For path analysis calculations of system gains and losses, dBi is used.
Figure 2.2.15 plots the response of a common type of STL antenna. Note the high directivity
provided by the device and the difference in gain between operation in the 450 MHz band as
y
(a) (b)
Figure 2.2.15 Radiating plots for a parabolic section antenna at 450 MHz and 950 MHz: (a) hori-
zontal response, (b) vertical response. (After [2].)
opposed to the 950 MHz band. Table 2.2.5 lists the

Table 2.2.5 Typical Isotropic Gain for
typical gain provided by antennas of different physical Full Parabolic Antennas at 950 MHz
size.
Antenna Diameter Isotropic Gain
Mounting Structures 4 feet 18.9 dBi
5 feet 21.0 dBi
Each antenna requires a mounting structure, such as 6 feet 22.0 dBi
the side of a roof or building, or an existing or custom- 8 f eet 25.0 dBi
built tower. The structure must be sufficiently high to 10 feet 27.0 dBi
mount the antenna at the minimum elevation deter-
mined by STL path calculations. The structure must also be capable of supporting the weight of
the antenna and transmission line, as well as withstanding any forces that affect the antenna
tower, such as wind and the weight of snow and ice. Be certain to take into account the mounting
hardware used with the antenna, which typically accommodates pipe up to a diameter of 3.5 in.
Provisions must be made to secure the transmission line. The structure may also be subject to
local building and zoning codes. Strict adherence to these codes is mandatory. Adequate physical
clearance must be considered to allow for proper horizontal and vertical alignment during instal-
lation.
y
Hardware Considerations
Depending on the complexity of the STL system, additional hardware may be required. Where
more than one transmitter or receiver is used with a single antenna, a combiner, splitter, or
duplexer will be needed. These items contribute additional loss and must be accounted for in
path calculations.
Certain installations may require the installation of an external power amplifier in cases
where the output of a standard STL transmitter is insufficient for a particular path. In path calcu-
lations, the power output of the external power amplifier (converted to dBm) is substituted for
the output of the STL transmitter. In general practice, most engineers choose to use an external
power amplifier only as a last resort. Higher gain antennas usually are a more attractive alterna-
tive.
2.2.3 System Planning

The ultimate goal in selecting STL equipment is to choose a combination of STL transmitter,
transmission line, antennas, and STL receiver that will give adequate quieting (S/N) when trans-
mitting over the path between the studio and the station transmitter. Allowance in the form of
fade margin must also be made for the uncertainties that the path imposes on the received signal.
The STL transmitter supplies the operating power that drives the system. The transmit and
receive antennas add gain to the link. All remaining components add loss. By far the greatest loss
results from the path itself. Passive components of the system, including the transmission line,
connectors, and pigtail cables, also add loss. All gains and losses must be taken into account to
ensure that the RF level at the receiver input is adequate. Losses should be minimized wherever
possible.
2.2.3a Frequency Selection

Frequency selection would be a relatively simple procedure were it not for other broadcasters,
who also may be using STL frequencies, and perhaps sharing the same transmitter site. In prac-
tice, the choice of operating frequency is governed by the availability of unused STL channels in
the area, and is further dictated by the need to avoid interference with other STL users. The local
frequency coordinator (under the auspices of the Society of Broadcast Engineers in most areas of
the U.S.) usually can provide information on available frequencies and potential frequency con-
flicts.
STL engineering can be divided into three broad categories:
• Frequency selection
• Path layout
• Path gain/loss calculations
Path layout concerns the factors affecting the physical location of the STL transmitter and
receiver, as well as the intended path. The goal in path layout is to ensure a correctly-oriented and
unobstructed route for unhampered propagation of the radio wave. Path gain/loss calculations
involve analyzing RF power levels from the transmitter output to the receiver input so that an
adequate receive level is provided.
y
Spectrum Considerations
In view of the serious spectrum congestion problems that exist today in many areas of the U.S.
and elsewhere, an STL system should be designed to be as spectrum-efficient as possible and—
equally important—to be as immune to undesired transmissions as possible. Even if the system
will be operated in an area that currently does not have a spectrum congestion problem, there is
no guarantee that such a situation will not surface in the near future. In any event, a well-engi-
neered system is also a spectrum-efficient system.
The first rule of spectrum-efficiency is to use only the effective radiated power (ERP) neces-
sary to do the job. There is no justification for putting 15 W into the air when 5 W will provide
the required receiver quieting and fade margin.
A simple and sometimes effective spectrum coordination tool is cross-polarization. Two sta-
tions on nearby frequencies may achieve as much as 25 dB RF isolation through the use of differ-
ent polarizations of transmit antennas, matched by like polarization at their respective receive
antennas. Cross-polarization results in varying degrees of success, depending upon the frequency
of operation and the surrounding terrain.
2.2.3b Path Layout

Careful path engineering should be performed prior to any licensing work to determine if the
proposed locations of the STL transmitter and receiver will be able to achieve the desired results.
There is much more to path engineering than simply pointing the transmitting and receiving
antennas at each other and turning the equipment on. Antennas cannot always be located in the
best possible geographic points because of space availability problems, excessive construction or
site rental costs, or local/federal licensing difficulties. In such cases, the required path is not the
ideal path and the link will have to be engineered around these fixed points.
Planning for any STL system should begin with an accurate, detailed U.S. Geological Survey
(USGS) map covering the proposed path. Note should be made of any natural obstructions (such
as mountains, hills, or vegetation) or man-made obstructions (such as buildings, water tanks, or
transmitting towers) in the path.
Line of Sight
Because microwave frequencies are used for STL systems, the signal path is theoretically limited
to the line-of-sight between the studio and transmitter locations. In reality, the radio horizon is
frequently situated beyond the visual horizon. This is the result of the gradual decrease in the
refractive index of the atmosphere with increasing altitude above the earth. This effect bends
radio waves downward. The degree of bending is characterized by the K factor, which is the ratio
of the effective earth radius to the true earth radius. A typical value for K is 4/3, or 1.33, valid
over 90 percent of the time in most parts of the world. For long paths, consult a map showing the
minimum K factor occurring in the specific area so that proper antenna heights can be planned.
Figure 2.2.16 plots an example path on 4/3 earth graph paper.
STL Site Selection

Usually, there is little choice for the STL transmitter and receiver locations, because most STL
systems will be required to provide a path between already determined studio and transmitter
points. If, however, a new studio location is being planned, the possible STL path is an important
y
Figure 2.2.16 An aural STL path profile drawn on true-earth radius graph paper. (After [2].)
consideration. Figure 2.2.17a illustrates a poor path. Generally speaking, a good path is literally
line-of-sight, with no obstructions to block the signal, and no other terrestrial or atmospheric
conditions that would compromise the path (Figure 2.2.17b.
Terrain Considerations
One of the major tasks required to engineer an STL system is the path analysis between the STL
transmitter at the studio and the STL receiver location. To determine what constitutes a clear
path, the concept of Fresnel zones for optical theory is applied to radio waves. Most of the elec-
tromagnetic energy at a receiving point is concentrated in an elliptical volume that is a function
of the distance between the transmit and receive points and the wavelength. The energy outside
this volume either cancels or reinforces the energy within the volume, depending on whether the
distance that the energy travels to the receive point is longer by an even or odd number of 1/4-
wavelengths. Even distances result in radio wave cancellations; odd distances result in radio
wave reinforcement. (See Figure 2.2.18.) The radius of the first Fresnel zone, which defines the
boundary of the elliptical volume, is given by
d1 d 2
F1 = (2.2.2)
fD
72.1 -----------
y
(a)
Unacceptable path Transmitter
site
Studio
site
(b) Transmitter
Studio site
site
Clear path
Figure 2.2.17 Path considerations in planning an STL: (a) unacceptable path because of obstruc-
tions, (b) usable path. (After [2].)
Where:
F 1 = first Fresnel zone radius in feet
d1 = distance from the transmitting antenna to the obstruction in miles
D = total path length in miles
d2 = D – d1 in miles
f = frequency in GHz
H = distance from the top of the obstruction to the radio path
For reliable operation, obstructions should not project into the area thus defined. Empirical
studies, however, have shown that performance is substantially the same as long as H is greater
than 0.6 F1.
The first step in evaluating the path is to make a subjective check of the planned route. First,
determine that a reasonable possibility of success exists. Next, draw a best-case route for the
path. If obstructions are found, consider accomplishing the path by using a repeater system, or by
shifting the location of the STL transmit antenna. Although a detailed path analysis may not be
required in cases where line-of-sight is clearly present and the distance is less than about 10
miles, it is still good engineering practice to review some of the basic elements of the evaluation.
Obtain an accurate map showing the topography between the STL transmitter and receiver
locations. After determining the transmitter and receiver sites on the map, connect them with a
straight line showing the proposed path. After the path has been drawn, a protractor can be used
y
STL
Fresnel zone receiver
STL
transmitter Obstruction
Figure 2.2.18 Fresnel zone clearance for an STL path. (After [2].)
to determine the direction (azimuth) of the path in degrees from true North. This data will later
assist in antenna setup, and is necessary for filling out the appropriate FCC application.
Using the scale of miles on the map, measure the distance between the transmitter and
receiver sites. Determine the altitude of the proposed transmit antenna location from the contour
lines on the map, and add to that the height of the building or other structure on which the
antenna will be mounted. Make a similar determination for the receive antenna location. Adjust
the heights to ensure that all obstructions are cleared. Depending on the path length and the
height of the antennas, it may be necessary to take the curvature of the earth into account, and
use earth radius graph paper to plot a cross-section of the path.
Study the map to see what terrain features are between the path points. Prior to making any
other evaluations, conduct a visual field survey. Check for any structures or features not listed on
the map. Anticipate any possible new construction or tree growth that may cause problems in the
future.
The terrain from the transmitting antenna to the receiving antenna must be examined not only
for obstructions, but for reflection possibilities as well. A large body of water will often cause
problems for an STL system. If the water is an even number of Fresnel zones from the direct
path, signal attenuation will likely occur at the receiver. Temperature changes and tidal condi-
tions will also have an effect. Likewise, thick vegetation or forested areas can be reflective to RF
signals when wet, creating a similar (but not so troublesome) problem. Generally, the solution to
reflection conditions is to change either the transmitting or receiving antenna height. In extreme
cases, a diversity reception system may also be used.
y
Transmit Receive
Transmission line antenna antenna Transmission line
RF path
Transmitter Receiver
Transmitter
power
)
m
Bd 0 dBm
(l
ev
el
F
R
Fade margin
Receiver
quieting
Figure 2.2.19 Path analysis plotting gains vs. losses for an STL. (After [2].)
Path Reliability
The long-term reliability of an STL path is determined in large part by the initial engineering
work done before the system is installed. Figure 2.2.19 charts gains and losses through a typical
system. The most important factors are free space loss and allowance for fade margin. Thereaf-
ter, effects such has diffraction, reflection, refraction, absorption, scattering, and terrain loss
must be considered.
A gain and loss balance sheet should be computed to determine the fade margin of the
planned STL system. An adequate fade margin is vital to reliable performance because a link
that is operating on the edge of the minimum acceptable receiver quieting will encounter prob-
lems later down the road. Normal component aging in the receiver or transmitter can cause a loss
in received signal level and, thus, degrade system performance. Atmospheric conditions, such as
severe weather in the area or ice on the transmitting or receiving antennas, can also cause sharp
fading and even a complete loss of signal if an adequate fade margin above minimum receiver
quieting is not provided. The STL fade margin can be computed using the following equations:
Gs = Gt + Gta + Gra (2.2.3)
Where:
Gs = total system gain in decibels
Gt = transmitter power output in dBm
Gta = transmit antenna gain in dBi
Gra = receive antenna gain in dBi
The values for Gta and Gra are gathered from the antenna manufacturer's literature.
The value for Gt is given by
Gt = 30 + 10 log Po (2.2.4)
y
10
5 .0
)
dB
1 .0
(t
f
00 0 .5
1r
ep
no
it
au
ne
t 0 .1
A
0 .0 5
0 .0 1
1 .0 10 1 00 1E3 1E4
O p e ra tin g fre q u e n c y (M H z )
Figure 2.2.20 Loss vs. frequency for 1/2-inch foam dielectric transmission line. (After [2].)
Where:
Gt = transmitter power output in dBm
Po = transmitter power output in watts
The total system losses are then computed per
Ls = Lp + Ll + Lc + Lm (2.2.5)
Where:
Ls = total system losses in decibels
Lp = path loss in dB
Ll = transmission line loss in dB
Lc = connector losses in dB
Lm = miscellaneous losses in dB
The values for Lt and Lc can be determined from manufacturer's literature. Figure 2.2.20
shows typical loss values for 1/2-in foam-filled transmission line. A reasonable value for connec-
tor loss with components normally used in 1/2-in coax installations is 0.5 dB. The value for Lp
can be determined by using the formula
Lp = 36.6 + 20 log F + 20 log D (2.2.6)

y
Table 2.2.6 Path Attenuation at 950 MHz
Distance in Miles Loss in dB Distance in Miles Loss in dB

1 –96.2 16 –120.2
2 –102.2 17 –120.7
3 –105.7 18 –121.3
4 –108.2 19 –121.7
5 –110.1 20 –122.2
6 –111.7 21 –122.6
7 –113.1 22 –123.0
8 –114.2 23 –123.4
9 –115.3 24 –123.8
10 –116.2 25 –124.1
11 –116.9 26 –124.5
12 –117.7 27 –124.8
13 –118.4 28 –125.1
14 –119.1 29 –125.4
15 –119.7 30 –125.7
Where:
Lp = free space attenuation loss between two isotropic radiators (in dB)
F = frequency of operation in megahertz
D = distance between the antennas in statute miles
Free space loss can also be found using a table of approximate values, as given in Table 2.2.6.
With the foregoing information, the fade margin can be calculated per
M f = G s – L s – Rm (2.2.7)
Where:
Mf = fade margin (in dB)
Gs = total system gain (dB)
Ls = total system losses (dB)
Rm = minimum signal strength required for the target S/N in dBm (a negative number)
Gs and Ls are determined by the equations given previously. Rm (receiver sensitivity) is deter-
mined from the receiver manufacturer's specifications. If the manufacturer gives a receiver sensi-
tivity figure in microvolts, the following formula can be used to convert to dBm:
–6
⎛ V × 10 ⎞
R m = 20 log ⎜ ----r-------------------⎟ (2.2.8)
⎝ 0.7746 ⎠
Where:
Rm = minimum required signal strength (in dBm)
Vr = receiver sensitivity (in μV)
In order to predict accurately the performance of the STL radio link, the value of Rm must be
determined carefully. For maximum system performance and reliability, the fade margin deter-
y
mination should be made based upon the signal level required Table 2.2.7 Recommended
to provide the minimum acceptable receiver S/N perfor- Fade Margin as a Function of
mance. Longer paths require greater margins for fade. Path Length
The primary cause of signal fade in an STL system below Path Length Fade Margin
1.0 GHz is changes in the refractive indexes of the atmo- 5 mi 5 dB
sphere along the signal path. These fluctuations affect the 10 mi 7 dB
path of reflected or refracted signals differently from the 15 mi 15 dB
direct, line-of-sight signal. When the interfering signals com- 20 mi 22 dB
bine with the direct signal, the level at the receiver increases 25 mi 27 dB
30 mi 30 dB
or decreases depending upon the degree of reinforcement or
cancellation. Because atmospheric conditions are seldom sta-
ble, some fade margin is always necessary.
Another cause of signal fade is earth bulge (or inverse beam) fading, where the overall refrac-
tive index in the vicinity of the signal path decreases, thus hindering the full signal from reaching
the receive antenna. Again, allowance for signal fade will minimize degradation. Precipitation is
another potential cause of signal fading, although it is not generally considered significant at fre-
quencies below 1.0 GHz.
Fade margin also can be determined approximately from Table 2.2.7. The relationship
between system reliability and fade margin is detailed in Table 2.2.8. A sample STL path, show-
ing gain and loss elements, is given in Figure 2.2.21.
2.2.3c Dealing with Problem Paths

The STL systems described thus far have been simple installations with a transmitter at the stu-
dio and a receiver at the broadcast transmitter site. There are cases, however, where a simple, sin-
gle-hop STL is not be practical. Reasons include the following:
• The distance is too great to be reasonably covered by a single-hop system.
• A direct path cannot be used because of obstructions of some type.
• An unusually large fade margin is required for the application.
Figure 2.2.22 shows three instances where a two-hop system is required. A two-hop repeater uses
two transmitters and two receivers, with one pair located at the intermediate site. As shown in
Figure 2.2.22c, interference avoidance is another practical application of a multi-hop system.
With a conventional active repeater, the received signal is demodulated down to the baseband
at the intermediate site before being retransmitted. This approach is shown in Figure 2.2.23.
Unfortunately, this method results in inevitable degradation of the signal. For an analog system, a
better approach, an IF repeater, is shown in Figure 2.2.24. Instead of being demodulated to the
baseband level, the received signal is brought down to the IF level before being injected into the
repeater transmitter.
Hot-Standby Redundant Systems

The importance of continuous, uninterrupted service from an STL system cannot be overstated.
Modern STL equipment is stable and reliable. Failures rarely occur. However, a station may elect
to install a hot-standby redundant system to ensure zero downtime from the STL. Figure 2.2.25
shows a block diagram of a common hot-standby system. The installation includes a second
y
Table 2.2.8 Relationship Between Fade Margin, Reliability, and Outage Time for Rayleigh Distributed Paths
Fade Margin Path Reliability/ Outage Hours Outage Minutes per Outage Seconds
(dB) Availability per Year Month per Day
10 90.4837 834.20 4170.98 8222.05
20 99.0050 87.22 436.12 859.69
21 99.2088 69-35 346.77 683.58
22 99.3710 55.14 275.68 543.43
23 99.5001 43.82 219.12 431.94
24 99.6027 34.83 174.14 343.28
25 99-6843 27.68 138.38 272.79
26 99.7491 21.99 109.96 216.75
27 99.9007 17.47 87.37 172.22
28 99.8416 13.88 69.41 136.83
29 99.8742 11.03 55.14 108.70
30 99.9000 8.76 43.81 86.36
31 99.9206 6.96 34.80 68.60
32 99.9369 5.53 27.65 54.50
33 99.9499 4.39 21.96 43.29
34 99.9602 3.49 17.45 34.39
35 99.9684 2.77 13.86 27.32
36 99.9749 2.20 11.01 21.70
37 99-9800 1.75 8.74 17.24
38 99.9842 1.39 6.95 13.69
39 99.9874 1.10 5.52 10.88
40 99.9900 0.88 4.38 8.64
41 99.9921 0.70 3.48 6.86
42 99-9937 0.55 2.77 5.45
43 99.9950 0.44 2.20 4.33
44 99.9960 0.35 1.74 3.44
45 99.9968 0.28 1.39 2.73
50 99.9990 0.09 0.44 0.86
55 99.9997 0.03 0.14 0.27
60 99.9999 0.01 0.04 0.09
transmitter and a second receiver. An automatic changeover unit is included for switching pur-
poses. The changeover system monitors each transmitter and receiver pair to sense critical
parameters. If a failure occurs, the changeover controller switches from the faulty transmitter to
the standby transmitter, or from the faulty receiver to the standby receiver. The changeover units
work independently.
This system can be simplified to include only hot-standby provisions at the transmitter site, or
at the receiver site. If this approach is taken, a case can be made for either backup system. One
argument states that the transmitter is more likely to fail because it contains high-power stages,
which typically are more prone to failures than low-power circuits. On the other hand, the trans-
mitter is almost always more accessible for repair—being at the studio site—than the receiver.
y
25 miles (-125 dB)
Transmitter
site
20.15 dBi gain 250 ft line
Clear path
20.15 dBi gain STL receiver
Studio 30 V = -77.25 dBm

site
STL transmitter Total transmission line length = 300 ft (8 dB loss)

+39 dBm output
50 ft line
Figure 2.2.21 Analysis of an example STL path.
2.2.3d Pre-Installation Checkout

The best planned STL system, well engineered and comprised of the finest equipment, will not
perform properly if incorrectly installed. Quality control steps taken before installation actually
begins is an important part of the installation process. Delivery schedules for STL hardware vary
widely. After all the equipment is ordered, components will arrive, typically, over a period of
some weeks. Good engineering practice dictates that each component is checked or inspected as
it arrives. This process is preferable to waiting for all of the components to arrive on site because
any problems discovered at that point will delay the project.
The STL transmitter can be operated into a dummy load to confirm proper operation. With an
ideal load, the transmitter front panel readings should correspond closely with the final test sheet
supplied by the manufacturer. Modulation can be applied to the transmitter to ensure proper
operation.
The STL receiver is more difficult to test effectively with the instruments typically available
at a radio station. The engineer can, however, bench test the receiver by operating the transmitter
into a dummy load and attaching a short wire to the receiver antenna input. Be sure to use the
proper connector on the receiver to avoid possible damage to the center pin. Maintain sufficient
separation of the transmitter and receiver to prevent overloading the receiver front-end. Arranged
in this manner, the STL system can be checked from end-to-end with a signal generator (or pro-
gram material) on the bench.
For a digital STL, the transmitter, receiver, and modem/multiplexer can be tested in a similar
manner. Bench tests before installation usually resolve hidden or unexpected problems associ-
ated with equipment operation and/or interfacing. Functionality of the transmitter and receiver
(a ) Transmitter
site
Studio
Insufficient clearance
Transmitter
site
Studio
site
Repeater
site
(b ) Transmitter Transmitter
site site
Studio Studio
site site
Obstructed path
Clear path Clear path
(c ) Repeater
site
Transmitter
site
Transmitter
Interference condition site
Studio
site #1 Studio
site #1
948.5 MHz
948.5 MHz
945.0 MHz
948.5 MHz 948.5 MHz
Studio
site #2 Studio
site #2
Repeater
site
Figure 2.2.22 Applications requiring the use of a multi-hop STL system: (a) excessively long path,
(b) path obstructions, (c) interference possibility. (After [2].)
STL Program input

transmitter
Studio site
Radio path
STL Repeater site

receiver
Front end, Demodulator, Composite output

mixer, filtering
IF
STL
transmitter
Filtering,
IPA, PA modulator
Radio path
Transmitter site
STL
Program output receiver
Figure 2.2.23 Repeater link using conventional demodulation and remodulation of the composite
signal. (After [2].)
should be checked before adding the encoder to the system. In this way, any problems detected
can be readily pinpointed.
An STL using one or more subcarriers for program relay, closed-circuit communications, or
remote control functions can be bench tested in a similar manner. Having all hardware on one
bench and easily accessible makes adjustment of levels and other parameters much easier than
when the units are separated by many miles.
Nearly all STL systems checked in this manner pass specification testing with no problem.
Still, there may be instances where a unit was damaged during shipment. It is far easier to solve
such problems before the hardware is installed.
Antennas should be given a close visual inspection. Many antenna models used for STL work
are shipped in a partially disassembled state. Final assembly of the antennas should be completed
before the installation process is begun. Pay particular attention to mounting brackets; make sure
y
STL Program input

transmitter
Studio site
Radio path
STL Repeater site

receiver
Front end, Demodulator,

mixer, filtering IF output
IF
STL
transmitter
IPA, PA Filtering,
modulator
Radio path
Transmitter site
Program output STL

receiver
Figure 2.2.24 Repeater link using IF transfer of the composite signal. (After [2].)
the hardware will mate properly with the tower or structure on which the antennas will be
mounted. Check carefully for missing parts.
The transmission line and connectors require no pre-installation quality control, however, it is
suggested that as many connectors as possible be installed in the shop, rather than at the site.
Type N connectors require patience and skill to install. If possible, have the cable supplier attach
the proper connectors to both ends of the cable reel. In this way, only two connectors will need to
be installed in the field. Always place the manufacturer-installed connector ends at the antennas.
This provides the best assurance of trouble free operation in an exposed environment. Consider
ordering a couple extra connectors just in case a part is lost or damaged during construction. The
engineer may also want to use one connector as a test case to become more familiar with the
required installation technique.
(a)
Transmitter RF
#1 output
#1
Fault
sense
Switched output
Program Changeover Antenna
and MUX switch
inputs
Fault
sense
Transmitter RF
#2 output
#2
(b)
Receiver
#1 Receiver
output #1
Fault
sense
Program
Antenna Signal Changeover and MUX
splitter switch outputs
Fault
sense Receiver
Receiver output #2
#2
Figure 2.2.25 Equipment configuration for a hot-standby system: (a) studio site, (b) broadcast
transmitter site. (After [2].)
The test equipment required for pre-installation checkout is usually readily available in the
radio station shop. Basic items include the following:
• A high-quality 50 Ω dummy load capable of dissipating approximately 25 W.
• An in-line RF power output meter capable of reading forward and reverse power at 1.0 GHz.
• Audio frequency signal generator.
• Audio frequency distortion analyzer.
• Frequency counter accurate to 1.0 GHz.
If possible, a spectrum analyzer also should be available. Figure 2.2.26 shows a typical bench
test setup for an STL system.
y
Load
Input signal RF sample
Signal generator STL STL receiver Distortion

transmitter Analyzer
Frequency
counter
Power
meter
RF signal input
Spectrum analyzer
Baseband signal input
Figure 2.2.26 Equipment setup for bench testing an STL system prior to installation. (After [2].)
Planning Cable Runs

Routing of the transmission line from the transmitter to the antenna is an often overlooked ele-
ment of STL installation. Confirm that there is a practical route for the coax from the transmitter
to the rooftop or tower location. Double-check the length of the transmission line path at both the
studio site and the station transmitter site. If the actual run is longer than predicted during fade
margin calculations, run the numbers again to be certain that a problem will not develop during
system commissioning. Confirm also that the pigtail cables to be used are sufficiently long and
have the proper connector terminations.
2.2.3e Installation
The locations commonly used for broadcast transmitter sites are seldom ideal from an environ-
mental standpoint. The site may be difficult to reach during certain parts of the year, very hot
during the summer, and very cold during the winter. For these reasons, rugged equipment should
be chosen and properly installed. Temperature extremes can cause problems for frequency-deter-
mining elements, as well as accessories such as cavity filters, preselectors, and preamplifiers.
Environmental control at the broadcast transmitter site, therefore, is highly recommended.
The STL transmitter and receiver should be mounted in an equipment rack in a protected
location adjacent to the stereo generator at the studio site, and adjacent to the exciter at the
broadcast transmitter site. Keep all cable runs as short and direct as possible. Excessive cable
lengths between the stereo generator and the STL transmitter, or between the STL receiver and
y
Transmit antenna Bond at antenna
STL antenna
STL receiver
Transmitter building Bond at antenna

Bond at
bulkhead
Bond to tower
at the point the
line begins its
horizontal run
Figure 2.2.27 Grounding practice for an STL transmission line on a tower. (After [2].)
the exciter, can degrade stereo separation and frequency response. Follow good grounding prac-
tices at all times.
The antenna presents probably the greatest installation challenge. Because of its directional
nature, the antenna must be properly oriented. Compass bearings are desirable along with what-
ever visual sightings are practical. The received signal strength at the broadcast transmitter site
can be used as an indication of proper orientation of the STL receive antenna. Ensure also that
the antenna is set to the proper angle relative to the horizon. Because the STL antennas may be
located high on a tower or building, a professional tower crew may be required to mount the
devices. Make certain that all cables on the tower are securely fastened to prevent stress and
wear. Seal the external connections with sealant made for that purpose and completely wrap the
connection joints with tape.
While it is desirable to mount the receive antenna high on a tower to ensure a good path, it is
also good engineering practice to keep the antenna as far as possible from radiating elements on
the structure, such as the station's main broadcast antenna. Other potential sources of RF prob-
lems include 2-way radio, cellular radio, and TSL systems. If the STL receive antenna is located
in a strong RF field, the front-end may be overloaded or desensitized. Placing the STL antenna
close to other antennas on the structure also can detune the antennas, degrading the performance
of both systems.
One of the most common problems encountered during installation is damage to the transmis-
sion line. Lines must not be twisted or bent beyond their minimum bending radius. To retain its
characteristic impedance, the line must be treated with care. A line damaged by carelessness can
cause poor performance of the system.
The transmission line must be properly grounded. As illustrated in Figure 2.2.27, grounding
typically is performed at the point where the line leaves the tower and at the point where it enters
the building. This procedure, obviously, applies only to grounded tower systems. Grounding kits
are available for this purpose.
y
In situations where the STL antenna is to be mounted on an AM antenna tower, which typi-
cally is insulated from ground, an isocoupler will need to be installed where the line begins its
run up the tower. The isocoupler will pass the STL signal efficiently, while providing a high
impedance to ground for the tower at the AM frequency. When this is done, the base impedance
of the tower can be excepted to change slightly.
2.2.3f Checking System Operation

After all of the components of the STL system have been installed, checkout can begin. Start by
powering-up the STL transmitter. Confirm that all meter readings are within normal parameters.
If any problems are detected, shut the system down immediately and troubleshoot the failure.
After proper operation has been confirmed, adjust the modulation levels into the STL trans-
mitter according to the instructions given with the unit. If a digital modem/multiplexer is used,
make the necessary adjustments on that unit.
Let the transmitter run for a period of time—several hours or so—and check the operating
parameters on a regular basis to be certain that no problems develop. Pay particular attention to
VSWR and internal high-temperature alarms. If abnormal operation is suspected, consult the
manufacturer's equipment manual.
After checkout of the STL transmitter is complete, power up the STL receiver and confirm
correct parameters on the front-panel meter. Check and adjust program levels as needed. If a dig-
ital modem/multiplexer is used in the installation, check the status monitors on the unit. If dis-
played, observe the bit error rate monitor to be sure the link is operating properly.
Using an accurate modulation monitor at the broadcast transmitter site, adjust the modulation
level of the transmitter. Check operation of the STL system carefully for an hour or so before
committing the system to the air.
After the link is operating properly, run a complete proof-of-performance on the broadcast
chain. Document the measured parameters so they can be used as a benchmark for later refer-
ence.
2.2.3g Troubleshooting
The most common problem associated with STL commissioning is high VSWR. If the indicated
VSWR is outside normal limits, the transmission line, a connector, or an antenna usually is at
fault. A dummy load may be used to determine where the fault lies. Begin at the first pigtail and
substitute the dummy load for the transmission line. At each point the load is substituted for the
line and antenna, check the displayed VSWR on the transmitter. It is necessary, of course, to
power-down the transmitter while the dummy load is moved from one point to the next.
2.2.3h Operating the System

Regular performance tests should be made on the STL system. Routine inspections allow the
engineer to spot potential problems that could cause a system failure if left unattended.
The STL transmitter and receiver should provide trouble-free performance for many years. If
problems develop, the likely trouble areas will be the antennas and transmission lines—the com-
ponents exposed to the elements. The best check of system performance at the studio is the STL
transmitter multimeter. A periodic check of the meter will help to spot oncoming problems.
y
Gradually changing readings can provide clues that will help prevent a failure. Many users will
wish to monitor forward power continuously. Reverse power also should be checked regularly to
monitor the condition of the line and antennas, as discussed previously. Compare the test data
taken after system commissioning with the operating parameters observed during routine checks.
They should closely agree.
Operation of the receiver is best checked with its built-in multimeter. The RF level indication
should be carefully noted during initial installation, and subsequent observations should be com-
pared with this reference. Remember there will be some change in signal strength readings
because of weather conditions and temperature variations. Consider the fade margin conditions
used in the path analysis calculations when making judgments about the observed readings. Be
aware of unusual variations resulting from temperature inversions, if they are common in the
area. If trouble is experienced with the receiver, the possibility of interference from another STL
or another service should not be overlooked. A spectrum analyzer is invaluable for such trouble-
shooting.
For digital STL systems, error rate monitors are provided to assess overall performance. As
long as the status indicators on the front panel indicate proper operation, the system usually can
be considered to be transparent to the program material.
The most definitive overall check of the system will be the audio proof of performance. First
attention should be given to the noise measurement. If this is not satisfactory, it will be impossi-
ble to achieve meaningful distortion measurements, because noise will be indicated as distortion
by the analyzer. For dual monaural systems, the engineer will need to carefully check left and
right cannel balance.
Any broadcast system is only as strong as its weakest component. Before placing blame for
poor performance on the STL, start at the beginning of the broadcast chain and follow the signal
step-by-step to the point where it deteriorates.
2.2.4 References
1. Rollins, William W., and Robert L. Band: “T1 Digital STL: Discrete vs. Composite Trans-
mission,” NAB 1996 Broadcast Engineering Conference Proceedings, National Association
of Broadcasters, Washington, D.C., pp. 356–359, 1996.
2. Whitaker, Jerry C., (ed.): A Primer: Digital Aural Studio to Transmitter Links, TFT, Santa
Clara, CA, 1994.
2.2.5 Bibliography
Hauptstuek, Jim: “Interconnecting the Digital Chain,” NAB 1996 Broadcast Engineering Confer-
ence Proceedings, National Association of Broadcasters, Washington, D.C., pp. 360–358,
1996.
McClanahan, M. E.: “Aural Broadcast Auxiliary Links,” in NAB Engineering Handbook, 8th ed.,
E. B. Cructhfield (ed.), National Association of Broadcasters, Washington, D.C, pp. 671–
678, 1992.
y
Parker, Darryl: “TFT DMM92 Meets STL Requirements,” Radio World, Falls Church, VA, Octo-
ber 21, 1992.
Salek, Stanley: “Analysis of FM Booster System Configurations,” Proceedings of the 1992 NAB
DC, April 1992.
Whitaker, Jerry C., and Skip. Pizzi: “Radio Electronic News Gathering and Field Production,” in
NAB Engineering Handbook, 8th ed., E. B. Cructhfield (ed.), National Association of
Broadcasters, Washington, D.C, pp. 1051–1072, 1992.
g g
Chapter
2.3
Digital Radio Systems
Ken Pohlmann
2.3.1 Introduction
Following many years of work, digital audio radio (DAR), also known as digital audio broadcast-
ing (DAB), is a reality. Instead of using analog modulation methods such as AM or FM, DAR
transmits audio signals digitally. DAR is designed to eventually replace analog AM and FM
broadcasting, providing a signal that is robust against reception problems such as multipath inter-
ference, with fidelity comparable to that of the compact disc. In addition to audio data, DAR
supports auxiliary data transmission; for example, text, graphics, or still video images.
The evolution of a practical DAR standard was a complicated process because broadcasting is
regulated by governments, and swayed by economic concerns. Two principal DAR technologies
evolved: Eureka 147 DAB, and in-band on channel (IBOC) broadcasting. The route to DAR was
complex, with each country choosing one method or another. Unfortunately, there will not be a
single worldwide standard.
2.3.2 Technical Considerations

A significant complication for any new broadcast system is where to locate the band (perhaps
100 MHz wide) in the electromagnetic spectrum1. Spectral space is a limited resource that has
been substantially allocated. Furthermore, the frequency of the DAR transmission band will
impact the technology's quality, cost, and worldwide compatibility. Any band from 100 to 1700
MHz could be used for terrestrial DAR, but the spectrum is already crowded with applications.
In general, lower bands are preferable (because RF attenuation increases with frequency) but are
difficult to obtain. The S-Band is not suitable for terrestrial DAR because it is prone to interfer-
1. Portions of this chapter were adapted from: Pohlmann, Ken: Principles of Digital Audio,
McGraw-Hill, New York, N.Y., 2000. Used with permission.
2-57
g y
ence. However, the S-band is suitable for satellite delivery. Portions of the VHF and UHF bands
have been allocated to DTV applications.
A worldwide allocation would assist manufacturers, and would ultimately lower costs for all
concerned, but such a consensus was impossible to obtain. The World Administrative Radio
Conference (WARC) allocated 40 MHz at 1500 MHz (L-band) for digital audio broadcasting via
satellite, but ultimately deferred selection for regional solutions. Similarly, the CCIR (Interna-
tional Radio Consultative Committee) proposed a worldwide 60-MHz band at 1500 MHz for
both terrestrial and satellite DAR; however; terrestrial broadcasting at 1500 MHz is prone to
absorption and obstruction, and satellite broadcasting requires repeaters. In the U.S., in 1995, the
FCC allocated the S-band (2310–2360 MHz) spectrum to establish satellite-delivered digital
audio broadcasting services. Canada and Mexico allocated space at 1500 MHz. In Europe, both
1500 MHz and 2600 MHz regions have been developed. Ideally, whether using adjacent or sepa-
rated bands, DAR would permit compatibility between terrestrial and satellite channels. In prac-
tice, there is not a mutually ideal band space, and any allocation involves compromises.
Alternatively, new DAR systems could cohabit spectral space with existing applications. Spe-
cifically, DAR could use a shared-spectrum technique to locate the digital signal in the FM and
AM bands. By using an “in-band” approach, power multiplexing can provide compatibility with
analog transmissions, with the digital broadcast signal coexisting with the analog carriers.
Because of its greater efficiency, the DAR signal transmits at lower power relative to the analog
station. An analog receiver rejects the weaker digital signal as noise, and DAR receivers pick up
both DAR and analog broadcasts. No matter how DAR is implemented, the eventual disposition
of AM and FM broadcasting is a concern. A transition period is required, lasting until conven-
tional AM and FM gradually disappear.
2.3.2a Data Reduction

Digital audio signals cannot be practically transmitted in a linear PCM format because the band-
width requirements would be extreme. A stereo DAB signal might occupy 2 MHz of bandwidth,
compared to the approximately 200 kHz required by an analog FM broadcast. Thus, DAR must
use data reduction to reduce the spectral requirement. For example, instead of a digital signal
transmitted at a 2-Mbits/s rate, a data-reduced signal might be transmitted at 256 kbits/s. There
are numerous perceptual coding methods suitable for broadcasting. For example, the ISO/MPEG
algorithms use subband and transform coding with numerous data rates. Similarly, the Dolby
AC-3 system uses transform coding. Such systems can reduce an audio signal to a 256-, 128-,
96-, or 64-kbits/s rate. Although telephony has used data reduction for years, this technology has
only recently entered the domain of commercial radio broadcasting where transmission demands
are more challenging. In addition, an audio signal passing through a broadcast chain could
undergo multiple data compression/decompression stages, potentially increasing distortion and
artifacts.
2.3.2b System Planning Issues

The performance of a digital audio broadcasting system can be evaluated using a number of cri-
teria, including:
• Delivered sound quality
g y
• Coverage range for reliable reception

• Interference between analog and digital signals at the same or adjacent frequencies
• Signal loss in mountains or tunnels
• Deep “stoplight” fades
• Signal “flutter” produced by passing aircraft
• Data errors in the presence of man-made and atmospheric noise
• Interference from power lines and overhead signs
• Attenuation by buildings
• Multipath distortion during fixed and mobile reception
• Receiver complexity
• Capacity for auxiliary data services
The design of a DAR network must balance many variables to produce a system with low
error rate, moderate transmitter power levels, and sufficient data rate—all within the smallest
possible bandwidth. As with any digital data system, a broadcasting system must minimize
errors; the bit error rate (BER) can be reduced through error correction data accompanying the
audio (payload) data.
Most digital communications systems use pulse-shaping techniques prior to modulation to
limit bandwidth requirements. Pulse shaping performs lowpass filtering to reduce high-fre-
quency content of the digital signal. Because this spreads the bit width, resulting in intersymbol
interference, raised cosine filters are used so that the interference from each bit is nulled at the
center of other bit intervals, eliminating interference.
Multipath interference occurs when a direct signal and one or more strongly reflected and
delayed signals destructively combine at the receiver. The delay might be on the order of 5-μs. In
addition, other weak reflected signals might persist for up to 20 μs. The result at the receiver is a
comb filter with 10- to 50-dB dips in signal strength, as shown in Figure 2.3.1. This type of RF
multipath is a frequency selective problem, and short wavelengths—for example, in FM broad-
casting—are more vulnerable. When the receiver is moving, multipath interference results in the
amplitude modulation “picket fence” effect familiar to mobile FM listeners. Even worse, in a
stoplight fade, when the receiver is stopped in a signal null, the signal is continuously degraded;
a single, strong specular reflection completely cancels the transmitted signal. Increasing power is
not a viable remedy because both the direct and reflected signal will increase proportionally, pre-
serving the interference nulls.
Another effect of multipath interference, caused by short delays, occurs in the demodulated
bit stream. Known as delay spread, multiple reflections arrive at the receiver over a time interval
of perhaps 15 μs The result is intersymbol interference in the received data because specific bits
arrive at multiple times. This degradation can be overcome with bit periods longer than the
spread time; however; with conventional modulation techniques, this would limit bit rate to less
than 100 kbits/s. Therefore, data reduction also must be used.
Frequency diversity techniques can be used to combat multipath interference by placing the
data signal on multiple carriers. With such an approach, interference on one carrier frequency
can often be overcome.
g y
Figure 2.3.1 The effects of multipath on signal reception in a mobile environment. (From [1]. Used
with permission.)
Two types of multiplexing are used. The most common method is time division multiplexing
(TDM) in which multiple channels share a single carrier by time-interleaving their data streams
on a bit or word basis. Frequency-division multiplexing (FDM) divides a band into subbands,
and individual channels modulate individual carriers within the available bandwidth. A single
channel can be frequency multiplexed; this lowers the bit rate on each carrier and lowers bit
errors as well. Because different carriers are used, multipath interference is reduced because only
one carrier frequency is affected.
Phase-shift keying (PSK) modulation
methods are commonly used because they (a) (c)
yield the lowest BER for a given signal
strength. In binary phase-shift keying
(BPSK) two phase shifts represent two
binary states. For example, a binary 0 places (b)
the carrier in phase, and a binary 1 places it
180° out of phase, as shown in Figure 2.3.2a.
This phase change codes the binary signal, as
shown in Figure 2.3.2b. The symbol rate
equals the data rate.
In quadrature phase-shift keying (QPSK)
four phase shifts are used; thus, two bits per
Figure 2.3.2 Phase-shift keying modulation: (a)
symbol are represented (Figure 2.3.2c). The binary phase-shift keying phasor diagram, (b)
symbol rate is half the transmission rate. BPSK phase change coding, (c) quadrature
QPSK is the most widely used method, espe- phase-shift keying phasor diagram. (From [1].
cially for data rates above 100 Mbits/s. Other Used with permission.)
modulation methods include amplitude shift
g y
keying (ASK) in which different carrier powers represent binary values, and frequency shift key-
ing (FSK) in which the carrier frequency is varied.
The bandwidth (BW) for an M-PSK signal is given by
D
--------------- < BW 2 D
> --------------- (2.3.1)
log
2M log
2M
where D is the data rate in bits/s. For example, a QPSK signal transmitting a 400-kbits/s signal
would require a bandwidth of between 200 to 400 kHz. A 16-PSK signal could transmit the same
data rate in half the bandwidth, but would require 8 dB more power for a satisfactory BER.
Given the inherently high bandwidth of digital audio signals, data reduction is mandatory to con-
serve spectral space and provide low BER for a reasonable transmission power level.
One of the great strengths of a digital radio system is its transmission power efficiency. This
can be seen by relating coverage area to the carrier-to-noise ratio (C/N) at the receiver. A digital
system might need a C/N of only 6 dB, but a FM receiver needs a C/N of 30 dB—a difference of
24 dB—to provide the same coverage area. The field strength for a DAR system can be esti-
mated from
E = V i + NF + C / N – --96.5
----------- (2.3.2)
F MHz
where E = minimum acceptable field strength at the receiver in dBu, Vi = thermal noise of the
receiver into 300 Ω in dBu, and
Vi = ⎛ kTRB
20 log --------------
⎞
1⁄ 2
(2.3.3)
⎝ –6 ⎠
10
Where:
k = 1.38 × 10-23 W/kHz
T = temperature in degrees Kelvin (290 at room temperature)
R = input impedance of the receiver
B = bandwidth of the digital signal
NF = noise figure of the receiver
C/N = carrier-to-noise ratio for a given BER
F MHZ = transmission frequency
For example, if a DAR signal is broadcast at 100 MHz, with 200-kHz bandwidth, into a
receiver with 6-dB noise figure, and with a C/N of 6 dB, then E = 5.5 dBu. In contrast, an FM
receiver might require a field strength of 60 dBu for good reception, and about 30 dBu for noisy
reception.
g y
2.3.3 Eureka 147/DAB

The Eureka 147 digital audio broadcasting system was selected as the European standard in 1995
for broadcasting to mobile, portable, and fixed receivers. The Eureka 147 technology is suitable
for use in terrestrial, satellite, hybrid (satellite and terrestrial), and cable applications. Eureka is a
research and development consortium of European governments, corporations, and universities,
established in 1985 to develop new technologies. By 1990 over 7.5 billion dollars had been
invested in over 300 projects ranging from biotechnology to transportation. Project number 147,
begun in 1986, aimed to develop a wideband digital audio broadcasting system (formally known
as DAB, as opposed to DAR). A prototype Eureka 147/DAB system was first demonstrated in a
moving vehicle in Geneva in September 1988, and many improvements followed. System speci-
fications were finalized at the end of 1994.
In traditional radio broadcasting, a single carrier frequency is used to transmit a mono or ste-
reo audio program, with one carrier per radio station. This method allows complete indepen-
dence of stations, but poses a number of problems. For example, reception conditions at the
receiver might produce multipath interference at the desired carrier frequency, in part because
the station's bandwidth is narrow (e.g., approximately 200 kHz for analog FM radio). In addition,
wide guardbands must be placed around each carrier to prevent adjacent interference. In short,
independent carrier transmission methods are not particularly robust, and are relatively ineffi-
cient from a spectral standpoint. Eureka 147 employs a different method of transmission coding
that overcomes many problems incurred by traditional broadcast methods.
2.3.3a System Details

Eureka 147 digitally combines multiple audio channels, and the combined signal is interleaved in
both frequency and time across a wide broadcast band. This approach provides spectrum- and
power-efficient transmission and reliable reception, even over a multipath fading channel. A
block diagram of a Eureka 147 transmitter is shown in Figure 2.3.3a. Audio data (as well as other
data) are individually encoded with channel coders and interleaved. A multiplexer combines
many different services to create a main service channel (MSC). The multiplexer output is fre-
quency interleaved and synchronization symbols are added. Channel coding is applied: coded
orthogonal frequency division multiplexing (COFDM) with quadrature phase shift keying modu-
lation is employed for each carrier to create an ensemble DAB signal.
The COFDM processing divides the transmitted information into many bit streams, each with
a low bit rate, which modulate individual orthogonal carriers so that the symbol duration is
longer than the delay spread of the transmission channels. The carriers in this ensemble signal
may be generated by an FFT. Depending on the transmission mode employed, there may be 1536,
384, or 192 carriers. In the presence of multipath interference, some of the carriers undergo
destructive interference while others undergo constructive interference. Because of frequency
interleaving among the carriers, successive samples from the same service are not affected by a
selective fade, even in a fixed receiver. Time interleaving further assists reception, particularly in
mobile receivers.
Carrier frequency centers are separated by the inverse of the time interval between bits and
separately modulated within their fractional spectral space with a portion of the overall signal.
This reduces the data rate on any one carrier, which promotes long bit periods. This frequency
diversity yields great immunity to intersymbol interference and multipath interference. Convolu-
tional coding adds redundancy using, for example, a code with a constraint length of 7. A collec-
g y
(a)
(b)
Figure 2.3.3 Eureka 147 DAB system: (a) transmitter, (b) receiver. (From [1]. Used with permis-
sion.)
tion of error correction profiles is used, optimized for the error characteristics of MPEG layer II
encoded data. The coders aim to provide graceful degradation as opposed to brick-wall failure.
Thus, stronger protection is given to data for which an error would yield muting or other obvious
effects, and weaker protection where errors would be less audible. Specifically, three levels of
protection are used within an MPEG frame; the frame header and bit allocation data are given the
strongest protection, followed by the scale factors, and subband audio samples respectively.
A block diagram of a Eureka 147 receiver is shown in Figure 2.3.3b. The DAB receiver uses
an analog tuner to select the desired DAB ensemble; it also performs down-conversion and filter-
ing. The signal is quadrature-demodulated and converted into digital form. FFT and differential
demodulation is performed, followed by time and frequency de-interleaving and error correction.
Final audio decoding completes the signal chain. A receiver may be designed to simultaneously
recover more than one service component from an ensemble signal.
g y
The DAB standard defines three basic transmission mode options, allowing a range of trans-
mitting frequencies up to 3 GHz:
• Mode I with a frame duration of 96 ms, 1536 carriers, and nominal frequency range of less
than 375 MHz is suited for a terrestrial VHF network because it allows the greatest transmit-
ter separations.
• Mode II with a frame duration of 24 ms, 384 carriers, and nominal frequency range of less
than 1.5 GHz is suited for UHF and local radio applications.
• Mode III with a frame duration of 24 ms (as in Mode II), 192 carriers, and nominal fre-
quency range of less than 3 GHz is suited for cable, satellite, and hybrid (terrestrial gap filler)
applications.
• Other modes can and have been defined.
In all modes, the transmitted signal uses a frame structure with a fixed sequence of symbols. The
gross capacity of the main service channel is about 2.3 Mbits/s within a 1.54-MHz bandwidth
DAB signal. The net bit rate ranges from approximately 0.6 to 1.7 Mbits/s depending on the
error correction redundancy used.
Eureka 147 uses ISO/MPEG-1 Layer II bit rate reduction in its source coding to minimize
spectrum requirements. Bit rates may range from 32 to 384 kbits/s in 14 steps; nominally, a rate
of 128 kbps per channel is used. Stereo or surround-sound signals can be conveyed. Nominally, a
sampling frequency of 48 kHz is used; however; a 24-kHz sampling rate is optional.
2.3.3b Spectrum Issues

Eureka 147 is inherently a wideband system, and it can operate at any frequency range up to 3
GHz for mobile reception and higher frequencies for fixed reception. Practical implementation
requires a spectrum allocation outside the existing commercial broadcast bands. The narrowest
Eureka 147 configuration uses 1.5 MHz to transmit six stereo channels. In practice, a much
wider band would be required for most applications. A fully implemented Eureka 147 system
might occupy an entire radio band.
Other drawbacks exist. In particular, the need to combine stations leads to practical problems
in some implementations. Eureka 147's designers, taking a European community bias, envi-
sioned a satellite delivery system that would blanket Europe with a single footprint. Terrestrial
transmitters operating on the same frequencies would be used mainly as gap fillers, operating on
the same frequency. This monolithic approach is opposed in the U.S. where independent local
programming is preferred. With satellite delivery of Eureka 147, the concept of local markets
becomes more difficult to implement, while national stations become easier to implement. This
would redefine the existing U.S. broadcast industry.
2.3.4 In-Band Digital Radio

In the search for a practical U.S. DAR system, several manufacturers developed in-band schemes
that convey digital audio signals in existing FM (88 to 108 MHz) and AM (510 to 1710 kHz)
bands along with analog radio signals. These systems are hybrids because the analog and digital
signals are broadcast simultaneously so that analog radios can continue to receive analog signals
g y
(a) (b)
Figure 2.3.4 RF emissions mask per FCC Rules: (a) FM broadcasting, (b) AM broadcasting.
(From [1]. Used with permission.)
while new digital radios can receive digital signals. At the end of a transition period, broadcasters
would turn off their analog transmitters and continue to broadcast in an all-digital mode. Such in-
band systems offer commercial advantages over a wideband system because broadcasters can
retain their existing listener base during a transition period, much of their current equipment can
be reused, existing spectral allocation can be used, and no new spectral space is needed.
In-band systems permit broadcasters to simultaneously transmit analog and digital programs.
Digital signals are inherently immune to interference, thus a digital receiver is able to reject the
analog signals. However, it is more difficult for an analog receiver to reject the digital signal's
interference. Coexistence can be achieved if the digital signal is broadcast at much lower power;
because of the broadcast efficiency of DAR, a low-power signal can maintain existing coverage
areas for digital receivers and allow analog receivers to reject the interfering signal.
With an in-band on-channel (IBOC) system, DAR signals are superimposed on current FM
and AM transmission frequencies. In the U.S., FM radio stations have a nominal bandwidth of
400 kHz with approximately a 200-kHz signal spectrum. FM radio stations are spaced 200 kHz
apart, and there is a guard band of 400 kHz between co-located stations to minimize interference.
In-band systems fit within the same bandwidth constraints, and furthermore, efficiently use
the FCC RF mask in which the channel's spectrum widens as power decreases. Specifically, if a
DAR signal is 25 dB below the FM signal, it could occupy a 480-kHz bandwidth, as shown in
Figure 2.3.4a. In the case of AM, if the DAR signal is 25 dB below the AM signal, the band can
be 40 kHz wide, as shown in Figure 2.3.4b. Because the digital signal's power can be lower, it can
thus efficiently use the entire frequency mask area. In practice, different IBOC systems use the
mask in different ways, seeking to optimize data throughput and robustness. Clearly, because of
the wider FM bandwidth, an FM in-band system is considerably easier to implement; rates of
256 kbits/s can be achieved. The narrow AM channels can limit DAR data rates to no more than
half that rate. In addition, existing AM radios are not as amenable to DAR signals as FM receiv-
ers. On the other hand, AM broadcast is not hampered by multipath problems.
An in-band FM system might require spatial diversity antennas on cars to combat multipath
interference. Any DAR system must rely on perceptual coding to reduce the channel data rate to
128 kbits/s or so, to allow the high fidelity signal (along with nonaudio data) to be transmitted in
the narrow bands available.
g y
The IBOC method is highly attractive because it fits within much of the existing regulatory
statutes and commercial interests. No modifications of existing analog AM and FM receivers are
required, and DAR sets receive both analog and digital signals. Moreover, because digital signals
are simply simulcast over existing equipment, start-up costs are low. An in-band system provides
improved frequency response, and lower noise and distortion within existing coverage areas.
Receivers can be designed so that if the digital signal is lost, the radio will automatically switch
to the analog signal.
2.3.4a iBiquity Digital Radio

iBiquity Digital was created by the merger of USA Digital Radio and Lucent Digital Radio, two
early proponents of IBOC technology. The iBiquity IBOC system provides a method of transmit-
ting compact-disc quality audio signals to radio receivers along with data services, such as sta-
tion, song and artist identification, stock and news information, and local traffic and weather.
The system allows existing radio stations to use their current AM and FM spectrum to transmit
analog signals simultaneously with new higher quality digital signals.
National Radio Systems Committee Test Results

iBiquity and its predecessor companies, USA Digital Radio and Lucent Digital Radio, actively
supported the National Radio Systems Committee (NRSC) and its evaluation of IBOC technol-
ogy. The NRSC is an industry standards setting body sponsored by the National Association of
Broadcasters (NAB) and the Consumer Electronics Association (CEA). In 1999, the NRSC
developed guidelines for testing prototype IBOC systems. On December 15, 1999, USA Digital
Radio filed its report on laboratory and field testing with the NRSC. On January 24, 2000,
Lucent Digital Radio filed its test report, which demonstrated system performance using percep-
tual audio coding (PAC). These reports confirmed the ability of IBOC technology to provide
upgraded audio quality and robustness in both the AM and FM bands without causing harmful
interference to existing analog broadcasts.
During 2000 and early 2001, the NRSC continued its analysis of IBOC technology. In
December 2000, the NRSC adopted FM laboratory and field test procedures for final IBOC val-
idation testing. In April 2001, the NRSC adopted new laboratory and field test procedures, and
in early August 2001, iBiquity Digital submitted its digital FM radio system tests to the NRSC.
The results of the tests, which were conducted in accordance with the NRSC's test procedures,
confirm that iBiquity Digital's digital FM radio system significantly outperformed conventional
analog FM radio in terms of audio quality and reception, while providing robust coverage
throughout the markets served. The test results also confirm that the company's digital FM sys-
tem preserves the integrity of the current analog FM broadcasting system. As part of the NRSC
test program, iBiquity Digital conducted field tests at eight FM stations across the country. The
NRSC test program also included laboratory tests and subjective sound quality evaluations by
480 listeners who evaluated thousands of audio samples recorded during the field tests. All field
tests and evaluations conducted as part of the NRSC FM test program were monitored or per-
formed by independent laboratories in accordance with procedures developed by the NRSC.
g y
FCC Actions
In a First Report and Order issued October 10, 2002, the FCC selected in-band, on-channel
(IBOC) as the technology to bring the benefits of digital audio broadcasting to AM and FM radio
broadcasters efficiently and rapidly. Also, the Commission announced notification procedures
that would allow AM (daytime operations only) and FM stations on a voluntary basis to begin
interim digital transmissions immediately using the IBOC systems developed by iBiquity Digital
Corporation. During the interim IBOC operations, stations would broadcast the same main chan-
nel program material in both analog and digital modes.
The Commission also announced its support for a public and open process to develop formal
IBOC AM and FM standards. It also deferred consideration of IBOC licensing and service rule
changes to a future Further Notice of Proposed Rulemaking.
2.3.5 References
1. Pohlmann, Ken: “Digital Radio and Television Broadcasting,” Principles of Digital Audio,
g y
g g
Chapter
2.4
IBOC AM Digital Radio System
2.4.1 Introduction1
The principal system analysis work on the in-band on-channel (IBOC) digital radio system for
AM broadcasting was performed by the DAB Subcommittee of the National Radio Systems
Committee (NRSC). The goals and objectives of the subcommittee were [1]:
• To study IBOC DAB systems and determine if they provide broadcasters and users with: 1) a
digital signal with significantly greater quality and durability than available from the analog
systems that presently exist in the U.S.; 2) a digital service area that is at least equivalent to
the host station's analog service area while simultaneously providing suitable protection in co-
channel and adjacent channel situations; 3) a smooth transition from analog to digital ser-
vices.
• To provide broadcasters and receiver manufacturers with the information they need to make
an informed decision on the future of digital audio broadcasting in the U.S., and if appropriate
to foster its implementation.
To meet its objectives, the subcommittee resolved to work towards achieving the following goals:
• To develop a technical record and, where applicable, draw conclusions that will be useful to
the NRSC in the evaluation of IBOC systems.
• Provide a direct comparison between IBOC DAB and existing analog broadcasting systems,
and between an IBOC signal and its host analog signal, over a wide variation of terrain and
under adverse propagation conditions that could be expected to be found throughout the U.S.
• Fully assess the impact of the IBOC DAB signal upon the existing analog broadcast signals
with which they must co-exist.
1. This chapter is based on the following document: NRSC, “DAB Subcommittee Evaluation of
the iBiquity Digital Corporation IBOC System, Part 2—AM IBOC,” National Radio Systems
Committee, Washington, D.C., April 6, 2002.
2-69
g y
• Develop a testing process and measurement criteria that would produce conclusive, believ-
able, and acceptable results, and be of a streamlined nature so as not to impede rapid develop-
ment of this new technology.
• Work closely with IBOC system proponents in the development of their laboratory and field
test plans, which would be used to provide the basis for the comparisons.
• Indirectly participate in the test process, by assisting in the selection of (one or more) inde-
pendent testing agencies, or by closely observing proponent-conducted tests to insure that the
testing is executed in a thorough, fair, and impartial manner.
2.4.1a Glossary of Terms

The following terms are used to describe the AM IBOC system [1].
ACR-MOS (absolute category rating mean opinion score) A methodology for subjectively test-
ing audio quality where participants are presented with sound samples, one at a time, and
are asked to grade them on a 5 point scale. For the NRSC AM IBOC tests, the MOS scale
used was 5 = excellent, 4 = good, 3 = fair, 2 = poor, and 1 = bad.
aftermarket receiver A radio designed for purchase and installation some time after purchasing
an automobile.
all-digital IBOC The “final” mode of the iBiquity AM IBOC system that increases data capac-
ity by increasing signal power and adjusting the bandwidth of the digital sidebands to min-
imize adjacent channel interference. All-digital AM IBOC uses four frequency partitions
and no analog carrier. In this mode, the digital audio data rate can range from 40 kbits/s to
60 kbits/s, and the corresponding ancillary data rate will remain at 0.4 kbits/s.
ATTC The Advance Television Technology Center, the primary lab test contractor for the IBOC
tests.
AWGN Additive white gaussian noise, also known as white noise, which contains equal energy
per frequency across the spectrum of the noise employed. In the context of the AM IBOC
system tests, AWGN at radio frequencies was utilized in the laboratory tests to simulate the
background noise present in the AM spectrum, which affects the quality of radio reception.
blend to analog The point at which the BLER of an AM IBOC receiver falls below some pre-
defined threshold and the digital audio is faded out while the analog audio is simulta-
neously faded in. This prevents the received audio from simply muting when the digital
signal is lost. The receiver audio will also “blend to digital” upon re-acquisition of the dig-
ital signal.
blend to mono The process of progressively attenuating the L–R component of a stereo decoded
signal as the received RF signal decreases. The net result is a lowering of audible noise.
BLER (block error rate) A ratio of the number of data blocks received with at least one errone-
ous bit to the number of blocks received.
compatibility When one system has little to no negative impact on another system, it can gener-
ally be considered compatible. In the context of digital radio, compatibility testing refers to
the extent to which the addition of an AM IBOC signal will impact current analog perfor-
mance.
g y
D/U The ratio of desired to undesired signals.

hybrid IBOC The initial mode of the iBiquity AM IBOC system that adds digital audio capacity
to an AM signal by inserting digital sidebands in the spectrum above, below, and within the
analog AM signal. The digital audio data rate can range from 36 kbits/s to 56 kbits/s, and
the corresponding ancillary data rate is 0.4 kbits/s in both cases.
IBOC An in-band/on-channel system of digital radio where the digital signals are placed within
the current AM and FM bands and within the FCC-assigned channel of a radio station.
Longley-Rice A model used to predict the long-term median transmission loss over irregular ter-
rain that is applied to predicting signal strength at one or more locations.
MPEG-2 AAC The advanced audio coder, a high-quality, low bit rate perceptual audio coding
system developed jointly by AT&T, Dolby Laboratories, Fraunhofer IIG, and Sony.
multipath An RF reception condition in which a radio signal reaching a receiving antenna
arrives by multiple paths due to reflections of the signal off of various surfaces in the envi-
ronment. By traveling different distances to the receiver, the reflections arrive with differ-
ent time delays and signal strengths. When multipath conditions are great enough, such as
in the area where the first reflection from the ionosphere arrives back at Earth and meets
the station’s groundwave signal, analog reception of AM radio broadcasts can become dis-
torted.
NRSC National Radio Systems Committee, a technical standards setting body of the radio
broadcasting industry, co-sponsored by the Consumer Electronics Association (CEA) and
the National Association of Broadcasters (NAB).
objective testing Using test equipment to directly measure the performance of a system under
test. For example, the power output of a transmitter can be objectively measured using a
wattmeter.
OEM (original equipment manufacturer) Generally describes the “factory” radio installed in a
car before purchase.
PAC A flexible high-quality perceptual audio coding system originally developed by Lucent
Technologies and later refined by iBiquity. The system can operate over a wide range of bit
rates and is capable of supporting multichannel audio.
perceptual audio coding Also known as audio compression or audio bit rate reduction, this is
the process of representing an audio signal with fewer bits while still preserving audio
quality. The coding schemes are based on the perceptual characteristics of the human ear.
Some examples of these coders are PAC, AAC, MPEG-2, and AC-3.
RMS (root mean square) The root mean square value of a periodic function, like a sine wave
used for audio measurements, is the square root of the average of the square of the value of
the function taken throughout one period.
RSS (root sum square) A method for combining the power of multiple signals by taking the
square root of the sum of the squares of all of the signals.
SDARS (satellite digital audio radio service) A satellite-delivered digital audio systems such as
those from XM Radio and Sirius. The digital audio data rate in these systems is specified
as being 64 kbits/s.
g y
subjective testing Using human subjects to judge the performance of a system. Subjective test-
ing is especially useful when testing systems that include components such as perceptual
audio coders. Traditional audio measurement techniques, such as signal-to-noise and dis-
tortion measurements, are often not sensitive to the way perceptual audio coders work and
cannot characterize their performance in a manner that can be compared with other coders,
or with traditional analog systems.
WQP (weighted quasi-peak) Refers to a fast attack, slow-decay detector circuit that approxi-
mately responds to signal peaks, and that has varying attenuation as a function of frequency
so as to produce a measurement that approximates the human hearing system.
2.4.2 iBiquity AM IBOC System

The iBiquity AM IBOC system supports transmission of digital audio and auxiliary digital data
within an existing AM channel allocation by placing six groups of digitally modulated carrier
signals within and adjacent to an analog AM signal (Figure 2.4.1) [1]. Because digitally modu-
lated carriers are inserted within the same spectrum occupied by the analog AM signal, the AM
IBOC system is not compatible with analog AM stereo signals. Corresponding sideband groups
on either side of the carrier (i.e., upper primary and lower primary) are independent in that only
one of them is needed for an IBOC receiver to be able to generate digital audio. However, in
order to generate stereo (or enhanced fidelity) digital audio, the secondary and tertiary sideband
groups are needed.
Orthogonal frequency division multiplexing (OFDM) modulation is utilized in the AM IBOC
system. The digital audio modulated onto these OFDM carriers is perceptually coded, allowing
for high-quality digital audio using a relatively low bit rate. At the transmission site, the audio
coder creates two audio streams—a “core” stream and an “enhanced” stream—and the system
assigns the streams to different parts of the spectrum. The core stream carries monaural audio
and the enhanced stream, at the broadcaster’s option, carries enhanced fidelity stereo audio.
The audio codec can be set to provide stereo audio in the enhanced portion of the audio
stream. At the receive site, then, the listener will hear either enhanced or core digital audio or
analog audio depending upon the reception conditions. The audio bandwidth for the digital audio
(both enhanced and core modes) is approximately 15 kHz.
The AM IBOC signal incorporates a 4-1/2 s delay between the analog and digital (simulcast)
audio signals to improve performance in the presence of certain types of interference, which may
affect how broadcasters monitor off-air signals.
2.4.2a Test Program Issues

Some of the specific attributes of the AM IBOC system that influenced the design of the NRSC
test program include the following [1]:
• Proximity of digital sidebands to first-adjacent channel signals. The digital sidebands of the
AM IBOC signal are located such that they could potentially interfere with (and receive inter-
ference from) a first-adjacent analog AM signal (Figure 2.4.2). The NRSC test procedures
included tests that characterized this behavior, including tests of IBOC performance when
g y
dBc
(-10 kHz) (+10 kHz)
FCC AM 0
MASK ANALOG
AM
SIGNAL
(-20 kHz) (+20 kHz)

-25
OFDM
TERTIARY SUBCARRIERS
-35
SECONDARY
PRIMARY
-20 -15 -10 -5 5 10 15 20 f, kHz

Channel
center
frequency
Figure 2.4.1 iBiquity AM IBOC system signal spectral power density. (From [1]. Used with permis-
sion.)
there were two first-adjacent channel signals, one on either side of the desired signal (hence
digital sidebands on both sides of the carrier were experiencing interference).
• Proximity of digital sidebands to second-adjacent channel signals. The digital sidebands of
the AM IBOC signal are located such that they could potentially interfere with (and receive
interference from) a second-adjacent AM signal’s digital sidebands (Figure 2.4.3). The NRSC
test procedures included tests that characterized this behavior, including tests of IBOC perfor-
mance when there were two second-adjacent channel signals, one on either side of the desired
signal (hence digital sidebands on both sides of the carrier were experiencing interference).
• Proximity of digital sidebands to third-adjacent channel signals. The digital sidebands of the
AM IBOC signal are located such that they could potentially interfere with (and receive inter-
ference from) a third-adjacent AM IBOC signal’s digital sidebands (Figure 2.4.4). The NRSC
test procedures included tests that characterized this behavior.
• Blend from enhanced to core. The audio coder in the iBiquity system creates two digital
audio streams—enhanced and core. When all sideband groups (primary, secondary, and ter-
tiary) are receivable, an AM IBOC receiver will decode these streams and provide enhanced
digital audio quality to the listener. As reception conditions degrade such that the secondary
and tertiary sideband groups experience errors but the primary sideband groups are still of
acceptable quality, the receiver audio will blend from enhanced quality to core quality.
• Blend-to-analog. The iBiquity AM IBOC system simulcasts a radio station’s main channel
audio signal using the analog AM carrier and IBOC digital sidebands, and under certain cir-
cumstances, the IBOC receiver will blend back and forth between these two signals. Conse-
quently, depending upon the reception environment, the listener will either hear digital audio
(transported over the IBOC digital sidebands) or analog audio (delivered on the AM-modu-
lated analog carrier), with the digital audio being the default condition.
g y
DESIRED 10 kHz
CHANNEL
1s t A DJA CENT
CHANNEL
PRIMARY
DIGITAL SIDEBAND
OVERLAPS 1ST ADJ
ANALOG
Figure 2.4.2 Illustration of potential interference to/from first-adjacent analog signals by AM IBOC
digital sidebands. (From [1]. Used with permission.)
20 kHz
DESIRED PRIMARY
CHANNEL DIGITAL
2nd
SIDEBAND IN ADJACENT
WIDEBAND RX
FILTER
CHANNEL WIDEBAND
RX FILTER
MASK (TYP)
Figure 2.4.3 Illustration of potential interference between second-adjacent channel AM IBOC sig-
nals. (From [1]. Used with permission.)
The two main circumstances under which an IBOC receiver reverts to analog audio output are
during acquisition (i.e., when a radio station is first tuned in, an IBOC receiver acquires the ana-
log signal in milliseconds but takes a few seconds to begin decoding the audio on the digital side-
bands) or, when reception conditions deteriorate to the point where approximately 10 percent of
the data blocks sent in the digital sidebands are corrupted during transmission. Many of the tests
in the NRSC procedures were designed to determine the conditions that would cause blend-to-
analog to occur in this second circumstance, since at this point the IBOC system essentially
reverts to analog AM.
g y
DESIRED 30 kHz 3r d
CHANNEL ADJACENT
500 Hz
GUARD BAND CH ANNEL
BETWEEN
SIGNALS
Figure 2.4.4 Illustration of potential interference between third-adjacent AM IBOC signals. (From
2.4.3 NRSC Test Program

To properly evaluate the AM IBOC radio system, two basic types of tests were required—done in
both the laboratory and the field [1]:
• Performance tests. In the context of the NRSC’s test procedures and evaluation reports, per-
formance tests (sometimes called digital performance tests) were those used to establish the
performance of the IBOC digital radio system itself. Performance test results were obtained
using an IBOC receiver or through direct observation of the received signal.
• Compatibility tests. In the context of the NRSC’s IBOC evaluation, compatibility tests
(sometimes referred to as analog compatibility tests) were designed to determine the effect
that the IBOC digital radio signal had on existing analog signals. Compatibility testing
involved observing the performance of analog receivers with IBOC digital sidebands alter-
nately turned on and off; test results were obtained using either analog AM receivers or
through direct observation of the received signal.
For each type of test, two basic types of measurements were made:
• Objective measurements, where a parameter such as signal power, signal-to-noise ratio, or
error rate was measured, typically by using test equipment designed specifically for that par-
ticular measurement.
• Subjective measurements, where human interpretation or opinion was used to measure sys-
tem performance or compatibility. In the NRSC test program, subjective measurements
involved determining the quality of audio recordings by having people listen to them and rate
them according to a pre-defined quality scale.
Subjective evaluation was especially important when trying to assess the quality of IBOC dig-
ital audio because the IBOC radio system relies upon perceptual audio coding for audio trans-
mission. The listening experience of audio that has passed through a perceptually coded system
is not best characterized by many of the normal objective audio quality measures such as signal-
to-noise, distortion, or bandwidth. The instruments used to make such measurements do not ade-
g y
quately respond to the perceptual aspects of the system. This is one of the reasons why the
NRSC’s test program included a comprehensive subjective evaluation component.
The NRSC’s initial AM IBOC test program did not include any tests specifically designed to
evaluate the digital performance or host compatibility of the IBOC system under skywave propa-
gation conditions, for example, as distant listeners of a clear-channel AM station might experi-
ence at night. In addition, there were no provisions in the initial test program for evaluating the
impact of a skywave IBOC interferer on either analog or IBOC desired signals. Such testing was
beyond the scope of the NRSC’s accelerated test program, given the highly variable and statisti-
cal nature of skywave propagation conditions that make the collection of statistically significant
test data extremely difficult and time-consuming.
2.4.3a Test Conclusions

The NRSC's evaluation of AM IBOC confirmed that the iBiquity AM IBOC system would allow
AM broadcasters to provide listeners with two-channel stereo audio rivaling existing analog FM
stereo in quality, and that under daytime propagation conditions the digital performance of the
AM IBOC system would represent a significant improvement over existing analog services [1].
The NRSC tests furthermore found that AM IBOC would make it possible for AM broadcasters
to provide data services.
Test results showed that, compared to analog AM, AM IBOC offered improved immunity
from reception problems due to overhead power lines, grounded conductive structures (such as
highway overpasses), and other forms of electromagnetic interference that plague existing ser-
vices.
In order to enjoy the improvements that AM IBOC has to offer, however, AM broadcasters
must consider a system-specific trade-off. AM IBOC places digital carriers up to 15 kHz on
either side of an AM station’s main carrier. NRSC tests confirmed that a station transmitting an
IBOC signal encountered very little, if any, interference to its own received signal. Although the
IBOC digital carriers operate at very low power levels, in some cases stations on first adjacent
channels may receive noticeable interference under certain listening conditions.
In order to justify this tradeoff, the NRSC report explained, it must be recognized that the AM
band has been plagued for decades with high levels of natural and man-made interference. Addi-
tionally, station allocations over the years have in many cases worsened the interference problem.
The FCC made a valiant effort to improve the AM band in the 1980s. Rules were changed and
the expanded band was opened allowing stations to move to this new spectrum with the hope of
alleviating at least some of the interference on the original AM channels. AM stereo operation
was approved but never fully implemented. These efforts have in reality accomplished very little
toward improving AM listening. Moreover, the audio bandwidth of most AM receivers has been
reduced to less than 3 kHz—as shown by the receiver characterization tests performed during the
NRSC testing program. Therefore, many AM listeners today are faced with listening to their
favorite AM stations with less than telephone-quality audio bandwidth on these very narrow-
band receivers.
No test results were obtained by the NRSC, nor were they requested from iBiquity, on sky-
wave reception. However, due to the propagation mechanisms that support skywave reception,
the NRSC expected that first adjacent interference may pose potential problems for listeners dur-
ing nighttime hours. Additional testing was needed before the NRSC could further comment on
g y
the nighttime compatibility of hybrid AM IBOC. The NRSC therefore recommended that sta-
tions desiring to operate with AM IBOC do so during daytime hours only.
Analog Compatibility
For the conditions tested, the AM IBOC system was found to have little effect on the host analog
signal; the amount of interference to the host analog signal was receiver-dependent [1]. Narrow
bandwidth automobile receivers were found to be the least sensitive to the digital signal. Wider
bandwidth hi-fi and portable radios were found to be more sensitive to the digital signal. Each
receiver’s frequency and phase response symmetry plays a part in its host compatibility.
The test results suggested that, although the introduction of AM IBOC would be noticeable to
some listeners of the host analog station using certain analog receivers, these listeners are not
expected to find their audio quality sufficiently degraded to impact listening.
Other findings included the following:
• Co-channel compatibility. The introduction of AM IBOC was not expected to have any
impact on the level of co-channel interference due to the design of the AM IBOC system. Co-
channel compatibility was not tested by the NRSC.
• First adjacent compatibility. Overall conclusions about first-adjacent compatibility of the
AM IBOC system were that the interference caused by the introduction of the IBOC signal
was predominantly determined by the D/U ratio. FCC allocation rules permit 6 dB D/U ratios
at an AM station’s daytime protected contour. At the 10 dB D/U point, all AM radios tested,
when receiving speech programming, were unable to provide audio quality that would satisfy
at least half of all listeners whether or not an interfering first-adjacent station was broadcast-
ing the AM IBOC signal. At the 15 dB D/U point, automobile radios provided listenable
audio, and were not significantly affected by the introduction of IBOC; however, hi-fi receiv-
ers did provide listenable audio that would become unlistenable with the introduction of
IBOC. At the same 15 dB D/U point, portable radios appeared to provide unlistenable audio
with or without IBOC. At the 30 dB D/U point, all radios appeared to provide listenable audio
with or without IBOC.
• Second adjacent compatibility. The data indicated that second-adjacent interference from
AM IBOC would be receiver- and D/U-dependent. At the D/U ratios tested, narrowband (typ-
ically automobile) receivers were not sensitive to AM IBOC interference, although hi fi and
portable receivers (i.e., wideband receivers) experienced interference at the 0 dB D/U ratio,
and at negative D/U ratios. FCC allocation rules permit 0 dB D/U ratios at an AM station’s 5
mV/m groundwave contour.
• Third adjacent compatibility. AM IBOC was not expected to have an impact on the amount
of third-adjacent channel interference in the AM band, and the test results confirmed this.
2.4.4 References
1. NRSC: “DAB Subcommittee Evaluation of the iBiquity Digital Corporation IBOC System,
Part 2—AM IBOC,” National Radio Systems Committee, Washington, D.C., April 6, 2002.
g y
g g
Chapter
2.5
IBOC FM Digital Radio System
2.5.1 Introduction1
The principal system analysis work on the in-band on-channel (IBOC) digital radio system for
FM broadcasting was performed by the DAB Subcommittee of the National Radio Systems
Committee. The goals and objectives of the subcommittee were [1]:
• To study IBOC DAB systems and determine if they provide broadcasters and users with: 1) a
digital signal with significantly greater quality and durability than available from the analog
system that presently exists in the U.S.; 2) a digital service area that is at least equivalent to
the host station's analog service area while simultaneously providing suitable protection in co-
channel and adjacent channel situations; 3) a smooth transition from analog to digital ser-
vices.
• To provide broadcasters and receiver manufacturers with the information they need to make
an informed decision on the future of digital audio broadcasting in the U.S., and if appropriate
to foster its implementation.
To meet its objectives, the subcommittee resolved to work towards achieving the following goals:
• To develop a technical record and, where applicable, draw conclusions that will be useful to
the NRSC in the evaluation of IBOC systems.
• Provide a direct comparison between FM IBOC DAB and the existing analog broadcasting
system, and between an IBOC signal and its host analog signal, over a wide variation of ter-
rain and under adverse propagation conditions that could be expected to be found throughout
the U.S.
• Fully assess the impact of the IBOC DAB signal upon the existing analog broadcast signals
with which they must co-exist.
1. This chapter is based on the following document: NRSC, “DAB Subcommittee Evaluation of
the iBiquity Digital Corporation IBOC System, Part 1—FM IBOC,” National Radio Systems
Committee, Washington, D.C., November 29, 2001.
2-79
g y
• Develop a testing process and measurement criteria that would produce conclusive, believable
and acceptable results, and be of a streamlined nature so as not to impede rapid development
of this technology.
• Work closely with IBOC system proponents in the development of laboratory and field test
plans, which would be used to provide the basis for future comparisons.
• Indirectly participate in the test process, by assisting in selection of (one or more) indepen-
dent testing agencies, or by closely observing proponent-conducted tests to insure that the
testing is executed in a thorough, fair, and impartial manner.
2.5.1a Glossary of Terms

The following terms are used to describe the FM IBOC system [1].
ACR-MOS (absolute category rating mean opinion score) A methodology for subjectively test-
ing audio quality where participants are presented with sound samples, one at a time, and
are asked to grade them on a 5 point scale. For the NRSC FM IBOC tests, the MOS scale
used was 5 = excellent, 4 = good, 3 = fair, 2 = poor, 1 = bad.
after-market A radio designed for purchase and installation some time after purchasing an auto-
mobile.
all-digital IBOC The third of three modes in the iBiquity FM IBOC system that increases data
capacity by adding additional digital carriers. All-digital FM IBOC uses four frequency
partitions and no analog carrier. In this mode, the digital audio data rate can range from 64
kbits/s to 96 kbits/s, and the corresponding ancillary data rate can range from 213 kbits/s
for 64 kbits/s audio to 181 kbits/s for 96 kbits/s audio.
ATTC The Advance Television Technology Center, the prime lab test contractor for the FM
IBOC tests.
AWGN Additive white Gaussian noise, also known as white noise, which contains equal energy
per frequency across the spectrum of the noise employed. In the context of the FM IBOC
system tests, AWGN at radio frequencies was utilized in the laboratory tests to simulate the
background noise present in the FM spectrum, which affects the quality of radio reception.
blend to analog The point at which the BLER of an FM IBOC receiver falls below some pre-
defined threshold and the digital audio is faded out while simultaneously the analog audio
is faded in. This prevents the received audio from simply muting when the digital signal is
lost. The receiver audio will also “blend to digital” upon re-acquisition of the digital signal.
blend to mono The process of progressively attenuating the L–R component of a stereo decoded
signal as the received RF signal decreases. The net result is a lowering of audible noise.
BLER (block error rate) A ratio of the number of data blocks received with at least one un-cor-
rectable bit to the total number of blocks received.
compatibility When one system has little to no negative impact on another system, it can gener-
ally be considered compatible. In the case of FM IBOC tests, compatibility testing was per-
formed to determine the extent to which the addition of an FM IBOC signal would impact
analog system performance.
g y
D/U Ratio of desired to undesired signals (usually expressed in dB).

EWG Evaluation Working Group of the NRSC DAB Subcommittee.
extended-hybrid IBOC The second of three modes in the iBiquity FM IBOC system that
increases data capacity by adding additional carriers closer to the analog host signal. The
extended-hybrid IBOC mode adds two frequency partitions around the analog carrier. In
this mode, digital audio data rate can range from 64 kbits/s to 96 kbits/s, and the corre-
sponding ancillary data rate will range from 83 kbits/s for 64 kbits/s audio to 51 kbits/s for
96 kbits/s audio.
hybrid IBOC The first of three modes in the iBiquity FM IBOC system that increases data
capacity by adding additional carriers closer to the analog host signal. The hybrid IBOC
mode adds one frequency partition around the analog carrier and is characterized by the
highest possible digital and analog audio quality with a limited amount of ancillary data
available to the broadcaster. Digital audio data rates can range from 64 kbits/s to 96 kbits/s,
and the corresponding ancillary data rate can range from 33 kbits/s for 64 kbits/s audio to 1
kbits/s for 96 kbits/s audio.
IBOC In-band/on-channel system of digital radio where the digital signals are placed within the
current AM and FM bands and within the FCC-assigned channel of a radio station.
Longley-Rice A model used to predict the long-term median transmission loss over irregular ter-
rain that is applied to predicting signal strength at one or more locations. Longley-Rice
computations are employed both by the FCC allocations rules for FM stations to predict
signal strength contours and by propagation modeling software to predict signal strengths
in a two-dimensional grid on a map. The FCC implementation of Longley-Rice computa-
tions employs average terrain computations and an assumed 30-ft receive antenna height.
MPEG-2 AAC Advanced Audio Coder, a high-quality, low bit rate perceptual audio coding sys-
tem developed jointly by AT&T, Dolby Laboratories, Fraunhofer IIG, and Sony.
multipath An RF reception condition in which a radio signal reaching a receiving antenna
arrives by multiple paths due to reflections of the signal off of various surfaces in the envi-
ronment. By traveling different distances to the receiver, the reflections arrive with differ-
ent time delays and signal strengths. When multipath conditions are great enough, analog
reception of FM radio broadcasts is affected in a variety of ways, including stop-light fades,
picket fencing, and distortion of the received audio.
NRSC National Radio Systems Committee, a technical standards setting body of the radio
broadcasting industry, co-sponsored by the Consumer Electronics Association (CEA) and
the National Association of Broadcasters (NAB).
objective testing Using test equipment to directly measure the performance of a system under
test. For example, the power output of a transmitter can be objectively measured using a
wattmeter.
OEM (original equipment manufacturer) Generally describes the “factory” radio installed in a
car before purchase.
PAC A flexible high-quality perceptual audio coding system originally developed by Lucent
Technologies and later refined by iBiquity. The system can operate over a wide range of bit
rates and is capable of supporting multichannel audio.
g y
Perceptual Audio Coding Also known as audio compression or audio bit rate reduction, this is
the process of representing an audio signal with fewer bits while still preserving audio
quality. The coding schemes are based on the perceptual characteristics of the human ear.
Some examples of these coders are PAC, AAC, MPEG-2, and AC-3.
protected contour A representation of the theoretical signal strength of a radio station that
appears on a map as a closed polygon surrounding the station’s transmitter site. The FCC
defines a particular signal strength contour, such as 60 dBuV/m for certain classes of sta-
tion, as the protected contour. In allocating the facilities of other radio stations, the pro-
tected contour of an existing station may not be overlapped by certain interfering contours
of the other stations. The protected contour coarsely represents the primary coverage area
of a station, within which there is little likelihood that the signals of another station will
cause interference with its reception.
RDS (Radio Data System) The RDS signal is a low bit rate data stream transmitted on the 57
kHz subcarrier of an FM radio signal. Radio listeners know RDS mostly through its ability
to permit RDS radios to display call letters and search for stations based on their program-
ming format. Special traffic announcements can be transmitted to RDS radios, as well as
emergency alerts.
SDARS Satellite Digital Audio Radio Service, describes satellite-delivered digital audio systems
such as those from XM Radio and Sirius. The digital audio data rate in these systems is
specified as being 64 kbits/s.
subjective testing Using human subjects to judge the performance of a system. Subjective test-
ing is especially useful when testing systems that include components such as perceptual
audio coders. Traditional audio measurement techniques, such as signal-to-noise and dis-
tortion measurements, are often not compatible with way perceptual audio coders work and
therefore cannot characterize their performance in a manner that can be compared with
other coders, or with traditional analog systems.
WQP (weighted quasi peak) Refers to a fast attack, slow-decay detector circuit that approxi-
mately responds to signal peaks, and that has varying attenuation as a function of frequency
so as to produce a measurement that approximates the human hearing system.
2.5.2 iBiquity FM IBOC System

The iBiquity FM IBOC system supports transmission of digital audio and auxiliary digital data
within an existing FM channel allocation by placing two groups of digitally modulated carrier
signals adjacent to an analog FM signal as shown in Figure 2.5.1. These sideband groups are
independent in that only one group (either USB or LSB in the figure) is needed for an IBOC
receiver to be able to generate digital audio. Orthogonal frequency division multiplexing
(OFDM) modulation is utilized. The digital audio modulated onto these OFDM carriers is per-
ceptually coded, allowing for high-quality digital audio using a relatively low bit rate.
The system incorporates a 4-1/2 second delay between the analog and digital (simulcast)
audio signals to improve performance in the presence of certain types of interference, which may
affect how broadcasters monitor off-air signals.
g y
dBc
(-1 2 0 k H z ) (+ 1 2 0 kH z )
0
FC C F M M A S K
ANALOG FM
-1 3 S IG N A L
(-2 4 0 k H z ) (+ 2 4 0 kH z )
O FDM -4 1 O FD M
CA R R IER S CA R R IER S
(LS B ) (U S B )
C h a n n e l c e n te r f, k H z
-2 0 0 -1 0 0 fre q u e n c y
+ 100 + 200
-19 8 -1 29 + 12 9 + 1 98
Figure 2.5.1 iBiquity FM IBOC system signal spectral power density. (From [1]. Used with permis-
sion.)
2.5.2a NRSC Test Issues

Some of the specific attributes of the iBiquity FM IBOC system which influenced the design of
the NRSC test program include the following [1]:
• Proximity of digital sidebands to first-adjacent channel signals. The digital sidebands of the
FM IBOC signal are located such that they could potentially interfere with (and receive inter-
ference from) a first-adjacent analog FM signal (Figure 2.5.2). The NRSC test procedures
included tests which characterized this behavior, including tests of IBOC performance when
there were two first-adjacent channel signals, one on either side of the desired signal (hence
both digital sidebands were experiencing interference).
• Proximity of digital sidebands to second-adjacent channel signals. The FM IBOC system
design allows for approximately 4 kHz of “guard band” between second-adjacent IBOC digi-
tal sidebands (Figure 2.5.3). Because this relatively close proximity could have an impact on
performance, the NRSC test procedures included tests for characterizing performance with
second-adjacent interference, including dual 2nd-adjacent channel interferers with power lev-
els up to 40 dB greater than the desired signal power (since FCC rules allow a second-adja-
cent signal to be 40 dB stronger than the desired signal at the desired signal’s protected
contour).
• Blend-to-analog. The iBiquity FM IBOC system simulcasts a radio station’s main channel
audio signal using the analog FM carrier and IBOC digital sidebands, and under certain cir-
cumstances, the IBOC receiver will “blend” back and forth between these two signals. Conse-
quently, depending upon the reception environment, the listener will either hear digital audio
(transported over the IBOC digital sidebands) or analog audio (delivered on the FM-modu-
lated analog carrier).
g y
200 kHz
D E S IR E D
D IG IT A L
CHAN N EL S ID E B A N D 1st ADJACENT
O V E R LA P S
1ST ADJ CHANNEL
A N A LO G
Figure 2.5.2 Illustration of potential interference to/from first-adjacent analog signals by FM IBOC
digital sidebands. (From [1]. Used with permission.)
400 kHz
D E SIR E D
CHANNEL 2 nd A D JA C E N T
4 kH z G U A R D
BAND CHANNEL
BETW EEN
S IG N A LS
Figure 2.5.3 Illustration of potential interference between second-adjacent FM IBOC signals.

The two main circumstances under which an IBOC receiver reverts to analog audio output are
during acquisition; i.e. when a radio station is first tuned in (an IBOC receiver acquires the ana-
log signal in milliseconds but takes a few seconds to begin decoding the audio on the digital side-
bands) or when reception conditions deteriorate to the point where approximately 10 percent of
the data blocks sent in the digital sidebands are corrupted during transmission. Many of the tests
in the NRSC procedures were designed to determine the conditions that would cause blend-to-
analog to occur.
2.5.2b NRSC Test Program

To evaluate the IBOC FM radio system, two basic types of tests are required [1]:
g y
• Performance tests. In the context of the NRSC test procedures, performance tests (some-
times called digital performance tests) were those used to establish the performance of the
IBOC digital radio system itself. Performance test results were obtained using an IBOC
receiver or through direct observation of the received signal.
• Compatibility tests. In the context of the NRSC IBOC evaluation, compatibility tests (some-
times referred to as analog compatibility tests) were designed to determine the effect that the
IBOC digital radio signal had on existing analog signals (main channel audio and subcarri-
ers). Compatibility testing involved observing performance with IBOC digital sidebands
alternately turned on and off; test results were obtained using either analog FM receivers or
FM subcarrier receivers (analog or digital) or through direct observation of the received sig-
nal.
For each of these, two basic types of measurements were made:
• Objective measurements, where a parameter such as signal power, signal to noise ratio, or
error rate was measured, typically by using test equipment designed specifically for that par-
ticular measurement.
• Subjective measurements, which involved human interpretation or opinion (not something
that can be simply measured with a device). In the NRSC test program, subjective measure-
ments involved determining the quality of audio recordings by having people listen to them
and rate them according to a pre-defined quality scale.
Subjective evaluation was especially important when trying to assess the quality of IBOC dig-
ital audio because the IBOC radio system relies upon perceptual audio coding for audio trans-
mission. The listening experience of audio which has passed through a perceptually coded
system is not accurately characterized by many of the normal objective audio quality measures,
such as signal-to-noise, distortion, or bandwidth. The instruments used to make such measure-
ments do not adequately respond to the perceptual aspects of the system.
Lab tests
Laboratory tests are fundamental to any characterization of a new broadcast system such as FM
IBOC [1]. The controlled and repeatable environment of a laboratory makes it possible to deter-
mine how the system behaves with respect to individual factors such as the presence or absence
of RF noise, multipath interference, or co- and adjacent-channel signals. These factors all exist in
the real world but because they exist simultaneously and are constantly changing, it is virtually
impossible to determine, in the real world, the effect each has on system operation.
Field tests
Field testing of a new broadcast system is necessary to determine performance in the real world
where all of the various factors which impact propagation and reception of radio signals exist to
varying degrees depending upon time of day, geographic location, and environmental factors [1].
2.5.2c Test Conclusions and Recommendations

The NRSC concluded that the performance of the iBiquity FM IBOC system as tested repre-
sented a significant improvement over the existing analog services [1]. The impact of IBOC dig-
g y
ital sidebands on the performance of existing main channel audio services was found to be
varied. Still, tests showed that listeners should not perceive an impact on the analog host signal,
nor on the analog signals on carriers that are either co-channel or second-adjacent channel with
respect to an IBOC signal. With respect to carriers that are located first-adjacent to an IBOC sig-
nal, listeners within the protected contour should not perceive an impact, but a limited number of
listeners may perceive an impact outside of the protected contour under certain conditions.
The NRSC also concluded that the tradeoffs necessary for the adoption of FM IBOC are rela-
tively minor. With respect to the main channel audio signal, evaluation of test data showed that a
small decrease in audio signal-to-noise ratio will be evident to some listeners in localized areas
where first-adjacent stations, operating with the FM IBOC system, overlap the coverage of a
desired station. However, listeners in these particular areas may also be subject to adjacent-chan-
nel analog interference which will tend to mask the IBOC-related interference, most appropri-
ately characterized as band-limited white noise, rendering it inaudible under normal listening
conditions. Also, the NRSC reported that all present-day mobile receivers include a stereo blend-
to-mono function dynamically active under conditions of varying signal strength and adjacent
channel interference. This characteristic of mobile receivers will also tend to mask IBOC-related
noise. The validity and effectiveness of these masking mechanisms is apparent from the rigorous
subjective evaluations performed on the data obtained during NRSC adjacent-channel testing.
Careful evaluation of test data showed that the digital SCA services tested (RDS and DARC)
should not be adversely impacted by IBOC. For the case of analog SCA services, some questions
remained as to the impact of IBOC on such services
2.5.3 References
1. NRSC: “DAB Subcommittee Evaluation of the iBiquity Digital Corporation IBOC System,
Part 1—FM IBOC,” National Radio Systems Committee, Washington, D.C., November 29,
2001.
g g
Section
Television Transmission Systems

3
Signal transmission standards describe the specific characteristics of the broadcast television sig-
nal radiated within the allocated spectrum. These standards may be summarized as follows:
• Definitions of fundamental functions involved in producing the radiated signal format,
including the relative carrier and subcarrier frequencies and tolerances as well as modulation,
sideband spectrum, and radio-frequency envelope parameters
• Transmission standards describing the salient baseband signal values relating the visual psy-
chophysical properties of luminance and chrominance values, described in either the time or
frequency domains
• Synchronization and timing signal parameters, both absolute and relative
• Specific test and monitoring signals and facilities
• Relevant mathematical relationships describing the individual modulation signal components
The details of these signal transmission standards are contained within standards documents for
the countries and/or regions served. For conventional (analog) broadcasting, the signal formats
are NTSC, PAL, and SECAM. For digital television, the two principal transmission systems are
the ATSC DTV system and the European DVB system.
Having established the requirements for television transmission, the appropriate hardware
issues then come into play. In this section, a number of technologies are examined that are used
to deliver both analog and digital signals to consumers via over-the-air broadcasting. Consider-
able advancements in both solid-state- and vacuum-tube-based devices and systems have pro-
vided television stations with new capabilities, improved efficiency, extended reliability, and
numerous new operational features.
In This Section:
Chapter 3.1: Television Transmission Standards 3-9

Introduction 3-9
Definitions 3-9
Monochrome Compatible Color TV Systems 3-10
3-1
y
3-2 Section Three
NTSC, PAL, and SECAM Systems Overview 3-12

The NTSC Color System 3-13
PAL Color System 3-18
SECAM Color System 3-22
Additional Systems of Historical Interest 3-26
Summary and Comparisons of Systems Standards 3-26
NTSC Stereo System 3-28
References 3-31
Bibliography 3-33
Chapter 3.2: Ghost Canceling Reference Signal 3-35

Introduction 3-35
History and Purpose 3-35
Basic Parameters 3-35
System Overview 3-36
System Architecture 3-36
Specification of the GCR Signal 3-37
Receiver Implementation Considerations 3-39
References 3-39
Chapter 3.3: The ATSC DTV System 3-41

Introduction 3-41
DTV System Overview 3-42
Basic Video Systems Characteristics 3-45
Receiver 3-45
DTV Transmission Characteristics 3-47
Channel Error Protection and Synchronization 3-49
Reed-Solomon Encoder 3-50
Interleaving 3-50
Trellis Coding 3-50
Modulation 3-54
High Data-Rate Mode 3-55
Bit Rate Delivered to a Transport Decoder 3-56
Performance Characteristics of Terrestrial Broadcast Mode 3-58
Transmitter Signal Processing 3-59
Upconverter and RF Carrier Frequency Offsets 3-61
Nominal DTV Pilot Carrier Frequency 3-61
Requirements for Offsets 3-61
Upper DTV Channel into Lower Analog Channel 3-61
Co-Channel DTV into Analog 3-62
Co-channel DTV into DTV 3-63
Performance Characteristics of High Data Rate Mode 3-64
References 3-64
Bibliography 3-65
Chapter 3.4: Longley-Rice Propagation Model 3-67

Introduction 3-67
y
Television Transmission Systems 3-3
Evaluation of Service 3-67

Planning Factors 3-68
Reference Value of ERP for DTV Operation 3-69
DTV Transmitting Antenna Patterns 3-70
Application of the Longley-Rice Methodology 3-71
Evaluation of Interference 3-72
D/U Ratios 3-72
Receiving Antenna Pattern 3-72
References 3-74
Chapter 3.5: Television Transmitters 3-75

Introduction 3-75
System Considerations 3-75
Power Amplifier 3-77
Application Example 3-79
DTV Transmitter Considerations 3-80
Operating Power 3-81
Technology Options 3-83
Solid-State Devices 3-84
Power Grid Devices 3-86
Klystron-Based Devices 3-88
Constant Efficiency Amplifier 3-90
Digital Signal Pre-Correction 3-92
Implementation Issues 3-94
References 3-94
Chapter 3.6: Multiple Transmitter Networks 3-97

Introduction 3-97
Benefits of Multiple Transmitters 3-97
Limitations of SFNs 3-98
Single Frequency Network Concepts 3-100
Digital On-Channel Repeaters 3-100
Distributed Transmission Systems 3-104
Distributed Transmission Architecture 3-105
Simple DTx Network 3-106
Complex DTx Network 3-106
Large Cell System 3-106
Small Cell System 3-107
Micro-Cell System 3-107
DTx Synchronization Mechanisms 3-107
Distributed Transmission Adapter at Source 3-107
Slave Synchronization of Transmitters 3-109
External Time and Frequency Reference 3-109
Multiple Frequency Network Concepts 3-109
Translators 3-110
Implementation Considerations 3-110
References 3-110
y
3-4 Section Three
Chapter 3.7: DTV Satellite Transmission 3-111

Introduction 3-111
ATSC A/80 Standard 3-111
Services and Applications 3-112
System Overview 3-112
ATSC DTH Satellite Broadcast Standard 3-114
Transmission System 3-115
Integrated Receiver Decoder System 3-116
Compression Format Constraints 3-117
PSIP 3-117
Satellite Virtual Channel Table (SVCT) 3-118
References 3-119
Chapter 3.8: The DVB Standard 3-121

Introduction 3-121
Digital Video Broadcasting (DVB) 3-121
Technical Background of the DVB System 3-121
MPEG-2 3-122
MPEG-2 Data Packets 3-122
DVB Services 3-123
DVB-SI 3-123
DVB-S 3-124
DVB-C 3-124
DVB-MC 3-125
DVB-MS 3-125
DVB-T 3-125
DVB and the ATSC DTV System 3-126
COFDM Technical Principles 3-126
Implementation Considerations 3-127
References 3-132
Bibliography 3-133

Aitken, S., D. Carr, G. Clayworth, R. Heppinstall, and A. Wheelhouse: “A New, Higher Power,
IOT System for Analogue and Digital UHF Television Transmission,” Proceedings of the
1997 NAB Broadcast Engineering Conference, National Association of Broadcasters,
Washington, D.C., p. 531, 1997.
ATSC: “ATSC Digital Television Standard,” Advanced Television Systems Committee, Washing-
ton, D.C., Doc. A/53C, May 21, 2004.
ATSC: “Digital Audio Compression (AC-3) Standard,” Advanced Television Systems Commit-
tee, Washington, D.C., Doc. A/52A, August 20, 2001.
ATSC, “ATSC Recommended Practice: Design Of Synchronized Multiple Transmitter Net-
works,” Advanced Television Systems Committee, Washington, D.C., Doc. A/111, Septem-
ber 3, 2004.
y
ATSC: “Guide to the Use of the Digital Television Standard,” Advanced Television Systems
Committee, Washington, D.C., Doc. A/54A, December 4, 2003.
ATSC: “Ghost Canceling Reference Signal for NTSC,” Advanced Television Systems Commit-
tee, Washington, D.C., Doc. A/49, 1993.
ATSC Standard: “Modulation And Coding Requirements For Digital TV (DTV) Applications
Over Satellite,” Doc. A/80, ATSC, Washington, D.C., July, 17, 1999.
ATSC: “Synchronization Standard for Distributed Transmission,” Advanced Television Systems
Committee, Washington, D.C., Doc. A/110, July 14, 2004.
Carnt, P. S., and G. B. Townsend: Colour Television—Volume 1: NTSC; Volume 2: PAL and
SECAM, ILIFFE Bookes Ltd. (Wireless World), London, 1969.
“CCIR Characteristics of Systems for Monochrome and Colour Television—Recommendations
and Reports,” Recommendations 470-1 (1974–1978) of the Fourteenth Plenary Assembly
of CCIR in Kyoto, Japan, 1978.
FCC Report and Order and Notice of Proposed Rule Making on Rules and Policies Affecting the
Conversion to DTV, MM Docket No. 00-39, Federal Communications Commission, Wash-
ington, D.C., January 19, 2001.
Eilers, C.: “The Zenith Multichannel TV Sound System,” Proc. 38th NAB Eng. Conf., National
Association of Broadcasters, Washington, D.C., 1984.
Electronic Industries Association: Multichannel Television Sound—BTSC System Recommended
Practices, EIA, Washington, D.C., ETA Television Systems Bulletin 5, July 1985.
Ericksen, Dane E.: “A Review of IOT Performance,” Broadcast Engineering, Intertec Publish-
ing, Overland Park, Kan., p. 36, July 1996.
ETS-300-421, “Digital Broadcasting Systems for Television, Sound, and Data Services; Framing
Structure, Channel Coding and Modulation for 11–12 GHz Satellite Services,” DVB
Project technical publication.
Structure, Channel Coding and Modulation for Cable Systems,” DVB Project technical
publication.
ETS-300-468, “Digital Broadcasting Systems for Television, Sound, and Data Services; Specifi-
cation for Service Information (SI) in Digital Video Broadcasting (DVB) Systems,” DVB
cation for Carrying ITU-R System B Teletext in Digital Video Broadcasting (DVB) Bit-
streams,” DVB Project technical publication.
ETS-300-473, “Digital Broadcasting Systems for Television, Sound, and Data Services; Satellite
Master Antenna Television (SMATV) Distribution Systems,” DVB Project technical publi-
cation.
European Telecommunications Standards Institute: “Digital Video Broadcasting; Framing Struc-
ture, Channel Coding and Modulation for Digital Terrestrial Television (DVB-T)”, March
1997.
y
3-6 Section Three
FCC: “Memorandum Opinion and Order on Reconsideration of the Sixth Report and
Order,” Federal Communications Commission, Washington, D.C., February 17, 1998.
Federal Communications Commission: Multichannel Television Sound Transmission and Audio
Processing Requirements for the BTSC System, OST Bulletin 60, FCC, Washington, D.C.
Fink, Donald G. (ed.): Color Television Standards, McGraw-Hill, New York, N.Y., 1955.
Gilmore, A. S.: Microwave Tubes, Artech House, Dedham, Mass., pp. 196–200, 1986.
Herbstreit, J. W., and J. Pouliquen: “International Standards for Color Television,” IEEE Spec-
trum, IEEE, New York, N.Y., March 1967.
Hirsch, C. J.: “Color Television Standards for Region 2,” IEEE Spectrum, IEEE, New York, N.Y.,
February 1968.
Hulick, Timothy P.: “60 kW Diacrode UHF TV Transmitter Design, Performance and Field
Report,” Proceedings of the 1996 NAB Broadcast Engineering Conference, National Asso-
ciation of Broadcasters, Washington, D.C., p. 442, 1996.
Hulick, Timothy P.: “Very Simple Out-of-Band IMD Correctors for Adjacent Channel NTSC/
DTV Transmitters,” Proceedings of the Digital Television '98 Conference, Intertec Publish-
ing, Overland Park, Kan., 1998.
Lee, E. A., and D.G. Messerschmitt: Digital Communication, 2nd ed.,. Kluwer, Boston, Mass.,
1994.
Muschallik, C.: “Improving an OFDM Reception Using an Adaptive Nyquist Windowing,” IEEE
Trans. on Consumer Electronics, no. 03, 1996.
Ostroff, Nat S.: “A Unique Solution to the Design of an ATV Transmitter,” Proceedings of the
Plonka, Robert J.: “Planning Your Digital Television Transmission System,” Proceedings of the
Pollet, T., M. van Bladel, and M. Moeneclaey: “BER Sensitivity of OFDM Systems to Carrier
Frequency Offset and Wiener Phase Noise,” IEEE Trans. on Communications, vol. 43,
1995.
Priest, D. H., and M. B. Shrader: “The Klystrode—An Unusual Transmitting Tube with Potential
for UHF-TV,” Proc. IEEE, vol. 70, no. 11, pp. 1318–1325, November 1982.
Pritchard, D. H.: “U.S. Color Television Fundamentals—A Review,” SMPTE Journal, SMPTE,
White Plains, N.Y., vol. 86, pp. 819–828, November 1977.
Roizen, J.: “Universal Color Television: An Electronic Fantasia,” IEEE Spectrum, IEEE, New
York, N.Y., March 1967.
Rhodes, Charles W.: “Terrestrial High-Definition Television,” The Electronics Handbook, Jerry
Robertson, P., and S. Kaiser: “Analysis of the Effects of Phase-Noise in Orthogonal Frequency
Division Multiplex (OFDM) Systems,” ICC 1995, pp. 1652–1657, 1995.
y
Sari, H., G. Karam, and I. Jeanclaude: “Channel Equalization and Carrier Synchronization in
OFDM Systems,” IEEE Proc. 6th. Tirrenia Workshop on Digital Communications, Tirre-
nia, Italy, pp. 191–202, September 1993.
Symons, R., M. Boyle, J. Cipolla, H. Schult, and R. True: “The Constant Efficiency Amplifier—
A Progress Report,” Proceedings of the NAB Broadcast Engineering Conference, National
Association of Broadcasters, Washington, D.C., pp. 77–84, 1998.
Symons, Robert S.: “The Constant Efficiency Amplifier,” Proceedings of the NAB Broadcast
Engineering Conference, National Association of Broadcasters, Washington, D.C., pp.
523–530, 1997.
Tardy, Michel-Pierre: “The Experience of High-Power UHF Tetrodes,” Proceedings of the 1993
NAB Broadcast Engineering Conference, National Association of Broadcasters, Washing-
ton, D.C., p. 261, 1993.
van Klinken, N., and W. Renirie: “Receiving DVB: Technical Challenges,” Proceedings of the
International Broadcasting Convention, IBC, Amsterdam, September 2000.
Whitaker, Jerry C.: “Microwave Power Tubes,” Power Vacuum Tubes Handbook, Van Nostrand
Reinhold, New York, p. 259, 1994.
Whitaker, Jerry C.: “Microwave Power Tubes,” The Electronics Handbook, Jerry C. Whitaker
(ed.), CRC Press, Boca Raton, Fla., p. 413, 1996.
Whitaker, Jerry C.: “Solid State RF Devices,” Radio Frequency Transmission Systems: Design
and Operation, McGraw-Hill, New York, p. 101, 1990.
y
g g
Chapter
3.1
Television Transmission Standards
Dalton H. Pritchard, J. J. Gibson
3.1.1 Introduction
The performance of a motion picture system in one location of the world is generally the same as
in any other location. Thus, international exchange of film programming is relatively straightfor-
ward. This is not the case, however, with the conventional broadcast color television systems.
The lack of compatibility has its origins in many factors, such as constraints in communications
channel allocations and techniques, differences in local power source characteristics, network
requirements, pickup and display technologies, and political considerations relating to interna-
tional telecommunications agreements.
3.1.1a Definitions
Applicable governmental regulations for the various analog television transmission systems in
use around the world provide the basic framework and detailed specifications pertaining to those
standards. Some of the key parameters specified include the following:
• Amplitude modulation (AM). A system of modulation in which the envelope of the transmit-
ted wave contains a component similar to the waveform of the baseband signal to be transmit-
ted.
• Antenna height above average terrain. The average of the antenna heights above the terrain
from about 2 to 10 mi (3.2 to 16 km) from the antenna as determined for eight radial direc-
tions spaced at 450 intervals of azimuth. Where circular or elliptical polarization is employed,
the average antenna height is based upon the height of the radiation center of the antenna that
produces the horizontal component of radiation.
• Antenna power gain. The square of the ratio of the rms free space field intensity produced at
1 mi (1.6 km) in the horizontal plane, expressed in millivolts per meter for 1 kW antenna
input power to 137.6 mV/m. The ratio is expressed in decibels (dB).
• Aspect ratio. The ratio of picture width to picture height as transmitted. The standard is 4:3
for 525-line NTSC and 625 line-PAL and SECAM systems.
3-9
3-10 Television Transmission Systems
Table 3.1.1 IRE Standard Scale
Level IRE Units Modulation, %

Zero carrier 120 0
Reference white 100 12.5
Blanking 0 75
Sync peaks (max. carrier) –40 100
• Chrominance. The colorimetric difference between any color and a reference color of equal
luminance, the reference color having a specific chromaticity.
• Effective radiated power. The product of the antenna input power and the antenna power gain
expressed in kilowatts and in decibels above 1 kW (dBk). The licensed effective radiated
power is based on the average antenna power gain for each direction in the horizontal plane.
Where circular or elliptical polarization is employed, the effective radiated power is applied
separately to the vertical and horizontal components. For assignment purposes, only the effec-
tive radiated power for horizontal polarization is usually considered.
• Field. A scan of the picture area once in a predetermined pattern.
• Frame. One complete image. In the line-interlaced scanning pattern of 2/1, a frame consists of
two interlaced fields.
• Frequency modulation (FM). A system of modulation where the instantaneous radio fre-
quency varies in proportion to the instantaneous amplitude of the modulating signal, and the
instantaneous radio frequency is independent of the frequency of the modulating signal.
• Interlaced scanning. A scanning pattern where successively scanned lines are spaced an inte-
gral number of line widths, and in which the adjacent lines are scanned during successive
periods of the field rate.
• IRE standard scale. A linear scale for measuring, in arbitrary units, the relative amplitudes of
the various components of a television signal. (See Table 3.1.1.)
• Luminance. Luminance flux emitted, reflected, or transmitted per unit solid angle per unit
projected area of the source (the relative light intensity of a point in the scene).
• Negative transmission. Modulation of the radio-frequency visual carrier in such a way as to
cause an increase in the transmitted power with a decrease in light intensity.
• Polarization. The direction of the electric field vector as radiated from the antenna.
• Scanning. The process of analyzing successively, according to a predetermined method or
pattern, the light values of the picture elements constituting the picture area.
• Vestigial sideband transmission. A system of transmission wherein one of the modulation
sidebands is attenuated at the transmitter and radiated only in part.
3.1.2 Monochrome Compatible Color TV Systems

In order to achieve success in the introduction of a color television system, it was essential that
the color system be fully compatible with the existing black-and-white system. That is, mono-
chrome receivers must be able to produce high-quality black-and-white images from a color
broadcast and color receivers must produce high-quality black-and-white images from mono-
chrome broadcasts. The first such color television system to be placed into commercial broad-
cast service was developed in the U.S. On December 17, 1953, the Federal Communications
Commission (FCC) approved transmission standards and authorized broadcasters, as of January
23, 1954, to provide regular service to the public under these standards. This decision was the
culmination of the work of the NTSC (National Television System Committee) upon whose rec-
ommendation the FCC action was based [1]. Subsequently, this system, commonly referred to as
the NTSC system, was adopted by Canada, Japan, Mexico, and others.
That nearly 50 years later, this standard is still providing color television service of good qual-
ity testifies to the validity and applicability of the fundamental principles underlying the choice
of specific techniques and numerical standards.
The previous existence of a monochrome television standard was two-edged in that it pro-
vided a foundation upon which to build the necessary innovative techniques while simulta-
neously imposing the requirement of compatibility. Within this framework, an underlying
theme—that which the eye does not see does not need to be transmitted nor reproduced—set the
stage for a variety of fascinating developments in what has been characterized as an “economy of
representation” [1].
The countries of Europe delayed the adoption of a color television system, and in the years
between 1953 and 1967, a number of alternative systems that were compatible with the 625-line,
50-field existing monochrome systems were devised. The development of these systems was to
some extent influenced by the fact that the technology necessary to implement some of the
NTSC requirements was still in its infancy. Thus, many of the differences between NTSC and the
other systems are the result of technological rather than fundamental theoretical considerations.
Most of the basic techniques of NTSC are incorporated into the other system approaches. For
example, the use of wideband luminance and relatively narrowband chrominance, following the
principle of mixed highs, is involved in all systems. Similarly, the concept of providing horizontal
interlace for reducing the visibility of the color subcarrier(s) is followed in all approaches. This
feature is required to reduce the visibility of signals carrying color information that are contained
within the same frequency range as the coexisting monochrome signal, thus maintaining a high
order of compatibility.
An early system that received approval was one proposed by Henri de France of the Compag-
nie de Television of Paris. It was argued that if color could be relatively band-limited in the hori-
zontal direction, it could also be band-limited in the vertical direction. Thus, the two pieces of
coloring information (hue and saturation) that need to be added to the one piece of monochrome
information (brightness) could be transmitted as subcarrier modulation that is sequentially trans-
mitted on alternate lines—thereby avoiding the possibility of unwanted crosstalk between color
signal components. Thus, at the receiver, a one-line memory, commonly referred to as a 1-H
delay element, must be employed to store one line to then be concurrent with the following line.
Then a linear matrix of the red and blue signal components (R and B) is used to produce the third
green component (G). Of course, this necessitates the addition of a line-switching identification
technique. Such an approach, designated as SECAM (SEquential Couleur Avec Memoire, for
sequential color with memory) was developed and officially adopted by France and the USSR,
and broadcast service began in France in 1967.
The implementation technique of a 1-H delay element led to the development, largely through
the efforts of Walter Bruch of the Telefunken Company, of the Phase Alternation Line (PAL) sys-
tem. This approach was aimed at overcoming an implementation problem of NTSC that requires
a high order of phase and amplitude integrity (skew-symmetry) of the transmission path charac-
teristics about the color subcarrier to prevent color quadrature distortion. The line-by-line alter-
nation of the phase of one of the color signal components averages any colorimetric distortions
to the observer’s eye to that of the correct value. The system in its simplest form (simple PAL),
however, results in line flicker (Hanover bars). The use of a 1-H delay device in the receiver
greatly alleviates this problem (standard PAL). PAL systems also require a line identification
technique.
The standard PAL system was adopted by numerous countries in continental Europe, as well
as in the United Kingdom. Public broadcasting began in 1967 in Germany and the United King-
dom using two slightly different variants of the PAL system.
3.1.2a NTSC, PAL, and SECAM Systems Overview

In order to properly understand the similarities and differences of the conventional television
systems, a familiarization with the basic principles of NTSC, PAL, and SECAM is required. As
previously stated, because many basic techniques of NTSC are involved in PAL and SECAM, a
thorough knowledge of NTSC is necessary in order to understand PAL and SECAM.
The same R, G, and B pickup devices and the same three primary color display devices are
used in all systems. The basic camera function is to analyze the spectral distribution of the light
from the scene in terms of its red, green, and blue components on a point-by-point basis as deter-
mined by the scanning rates. The three resulting electrical signals must then be transmitted over a
band-limited communications channel to control the three-color display device to make the per-
ceived color at the receiver appear essentially the same as the perceived color at the scene.
It is useful to define color as a psycho-physical property of light—specifically, as the combi-
nation of those characteristics of light that produces the sensations of brightness, hue, and satura-
tion. Brightness refers to the relative intensity; hue refers to that attribute of color that allows
separation into spectral groups perceived as red, green, yellow, and so on (in scientific terms, the
dominant wavelength); and saturation is the degree to which a color deviates from a neutral gray
of the same brightness—the degree to which it is “pure,” or “pastel,” or “vivid.” These three
characteristics represent the total information necessary to define and/or recreate a specific color
stimulus.
This concept is useful to communication engineers in developing encoding and decoding
techniques to efficiently compress the required information within a given channel bandwidth
and to subsequently recombine the specific color signal values in the proper proportions at the
reproducer. The NTSC color standard defines the first commercially broadcast process for
achieving this result.
A preferred signal arrangement was developed that resulted in reciprocal compatibility with
monochrome pictures and was transmitted within the existing monochrome channel. Thus, one
signal (luminance) was chosen in all approaches to occupy the wide-band portion of the channel
and to convey the brightness as well as the detail information content. A second signal (chromi-
nance), representative of the chromatic attributes of hue and saturation, was assigned less chan-
nel width in accordance with the principle that, in human vision, full three-color reproduction is
not required over the entire range of resolution—commonly referred to as the mixed-highs prin-
ciple.
Another fundamental principle employed in all systems involves arranging the chrominance
and luminance signals within the same frequency band without excessive mutual interference.
Table 3.1.2 Comparison of NTSC, PAL, and SECAM Systems
All systems:
Use three primary additive colorimetric principles
Use similar camera pickup and display technologies
Employ wide-band luminance and narrow-band chrominance
All are compatible with coexisting monochrome systems First-order differences are therefore:
Line and field rates
Component bandwidths
Frequency allocations
Major differences lie in color-encoding techniques:
NTSC: Simultaneous amplitude and phase quadrature modulation of an interlaced, suppressed
subcarrier
PAL: Similar to NTSC but with line alternation of one color-modulation component
SECAM: Frequency modulation of line-sequential color subcarrier(s)
Recognition that the scanning process, being equivalent to sampled-data techniques, produces
signal components largely concentrated in uniformly spaced groups across the channel width,
led to introduction of the concept of horizontal frequency interlace (dot interlace). The color
subcarrier frequency was so chosen as to be an odd multiple of one-half the line rate (in the case
of NTSC) such that the phase of the subcarrier is exactly opposite on successive scanning lines.
This substantially reduces the subjective visibility of the color signal “dot” pattern components.
Thus, the major differences among the three main systems of NTSC, PAL, and SECAM are in
the specific modulating processes used for encoding and transmitting the chrominance informa-
tion. The similarities and differences are briefly summarized in Table 3.1.2.
3.1.2b The NTSC Color System

The importance of the colorimetric concepts of brightness, hue, and saturation comprising the
three pieces of information necessary to analyze or recreate a specific color value becomes evi-
dent in the formation of the composite color television NTSC format.
The luminance, or monochrome, signal is formed by addition of specific proportions of the
red, green, and blue signals and occupies the total available video bandwidth of 0–4.2 MHz. The
NTSC, PAL, and SECAM systems all use the same luminance (Y) signal formation, differing
only in the available bandwidths.
The Y signal components have relative voltage values representative of the brightness sensa-
tion in the human eye. Therefore, the red, green, and blue voltage components are tailored in pro-
portion to the standard luminosity curve at the particular values of the dominant wavelengths of
the three color primaries chosen for color television. Thus, the luminance signal makeup for all
systems, as normalized to white, is described by
EY′ = 0.299 ER′ + 0.587EG′ + 0.114EB′ (3.1.1)
The signal of Equation (3.1.1) would be exactly equal to the output of a linear monochrome
sensor with ideal spectral sensitivity if the red, green, and blue elements were also linear devices
with theoretically correct spectral sensitivity curves. In actual practice, the red, green, and blue
primary signals are deliberately made nonlinear to accomplish gamma correction (adjustment of
the slope of the input/output transfer characteristic). The prime mark (´) is used to denote a
gamma-corrected signal.
Signals representative of the chromaticity information (hue and saturation) that relate to the
differences between the luminance signal and the basic red, green, and blue signals are generated
in a linear matrix. This set of signals is termed color-difference signals and is designated as R – Y,
G – Y, and B – Y. These signals modulate a subcarrier that is combined with the luminance com-
ponent and passed through a common communications channel. At the receiver, the color differ-
ence signals are detected, separated, and individually added to the luminance signal in three
separate paths to recreate the original R, G, and B signals according to the equations
EY´ + E´(R – Y) = EY´ + ER´ – EY´ = ER´ (3.1.2)
EY´ + E´(G – Y) = EY´ + EG´ – EY´ = EG´ (3.1.3)
EY´ + E´(B – Y) = EY´ + EB´ – EY´ = EB´ (3.1.4)
In the specific case of NTSC, two other color-difference signals, designated as I and Q, are
formed at the encoder and are used to modulate the color subcarrier.
Another reason for the choice of signal values in the NTSC system is that the eye is more
responsive to spatial and temporal variations in luminance than it is to variations in chrominance.
Therefore, the visibility of luminosity changes resulting from random noise and interference
effects can be reduced by properly proportioning the relative chrominance gain and encoding
angle values with respect to the luminance values. Thus, the principle of constant luminance is
incorporated into the system standard [1, 2].
The chrominance signal components are defined by the following equations. For NTSC
E ′I = 0.274 EG′ + 0.596 E ′R − 0.322 E′B

(3.1.5)
′ =−0.522 EG′ + 0.211E′R + 0.311E′B

EQ
(3.1.6)
B −Y = 0.493 ( E ′B − EY′ )
(3.1.7)
R −Y = 0.877 ( E ′R − EY′ )
(3.1.8)
G −Y = 1.413 ( EG′ − EY′ )

(3.1.9)
For PAL
EU′ = 0.493 ( E ′B − EY′ )

(3.1.10)
EV′ =±0.877 ( E′R − EY′ )

(3.1.11)
For SECAM
D′R =−1.9 ( E ′R − EY′ )

(3.1.12)
D′B = 1.5 ( E′B − EY′ )

(3.1.13)
For the NTSC system, the total chroma signal expression is given by
B −Y R −Y
CNTSC = sin ωSC t + cos ωSC t
2.03 1.14
(3.1.14)
The PAL chroma signal expression is
U V
CPAL = sin ωSC t + cos ωSC t
2.03 1.14
(3.1.15)
where U and V are substituted for B – Y and R – Y, respectively, and the V component is alter-
nated 180º on a line-by-line basis.
The voltage outputs from the three camera sensors are adjusted to be equal when a scene ref-
erence white or neutral gray object is being scanned for the color temperature of the scene ambi-
ent. Under this condition, the color subcarrier also automatically becomes zero. The colormetric
values have been formulated by assuming that the reproducer will be adjusted for illuminant C,
representing the color of average daylight.
Figure 3.1.1 is a CIE chromaticity diagram showing the primary color coordinates for NTSC,
PAL, and SECAM. It is interesting to compare the available color gamut relative to that of color
paint, pigment, film, and dye processes.
In the NTSC color standard, the chrominance information is carried as simultaneous ampli-
tude and phase modulation of a subcarrier chosen to be in the high frequency portion of the 0–
Figure 3.1.1 CIE chromaticity diagram comparison of systems.
4.2 MHz video band and specifically related to the scanning rates as an odd multiple of one-half
the horizontal line rate, as shown by the vector diagram in Figure 3.1.2. The hue information is
assigned to the instantaneous phase of the subcarrier. Saturation is determined by the ratio of the
instantaneous amplitude of the subcarrier to that of the corresponding luminance signal ampli-
tude value.
The choice of the I and Q color modulation components relates to the variation of color acuity
characteristics of human color vision as a function of the field of view and spatial dimensions of
objects in the scene. The color acuity of the eye decreases as the size of the viewed object is
Figure 3.1.2 NTSC chrominance signal vector diagram.
decreased and thereby occupies a small part of the field of view. Small objects, represented by
frequencies above about 1.5 to 2.0 MHz, produce no color sensation (mixed-highs). Intermediate
spatial dimensions (approximately in the 0.5 to 1.5 MHz range) are viewed satisfactorily if repro-
duced along a preferred orange-cyan axis. Large objects (0–0.5 MHz) require full three-color
reproduction for subjectively-pleasing results. Thus, the I and Q bandwidths are chosen accord-
ingly, and the preferred colorimetric reproduction axis is obtained when only the I signal exists
by rotating the subcarrier modulation vectors by 33º. In this way, the principles of mixed-highs
and I, Q color-acuity axis operation are exploited.
At the encoder, the Q signal component is band-limited to about 0.6 MHz and is representa-
tive of the green-purple color axis information. The I signal component has a bandwidth of about
1.5 MHz and contains the orange-cyan color axis information. These two signals are then used to
individually modulate the color subcarrier in two balanced modulators operated in phase quadra-
ture. The sum products are selected and added to form the composite chromaticity subcarrier.
This signal—in turn—is added to the luminance signal along with the appropriate horizontal and
vertical synchronizing and blanking waveforms to include the color synchronization burst. The
result is the total composite color video signal.
Quadrature synchronous detection is used at the receiver to identify the individual color sig-
nal components. When individually recombined with the luminance signal, the desired R, G, and
B signals are recreated. The receiver designer is free to demodulate either at I or Q and matrix to
form B – Y, R – Y, and G – Y, or as in nearly all modern receivers, at B – Y and R – Y and maintain
500 kHz equiband color signals.
The chrominance information can be carried without loss of identity provided that the proper
phase relationship is maintained between the encoding and decoding processes. This is accom-
plished by transmitting a reference burst sig-

nal consisting of eight or nine cycles of the
subcarrier frequency at a specific phase [ –
(B – Y)] following each horizontal synchro-
nizing pulse as shown in Figure 3.1.3.
The specific choice of color subcarrier
frequency in NTSC was dictated by at least
two major factors. First, the necessity for
providing horizontal interlace in order to
reduce the visibility of the subcarrier
requires that the frequency of the subcarrier
be precisely an odd multiple of one-half the
horizontal line rate. This interlace provides
line-to-line phase reversal of the color sub- Figure 3.1.3 NTSC color burst synchronizing
carrier, thereby reducing its visibility (and signal.
thus improving compatibility with mono-
chrome reception). Second, it is advanta-
geous to also provide interlace of the beat-frequency (about 920 kHz) occurring between the
color subcarrier and the average value of the sound carrier. For total compatibility reasons, the
sound carrier was left unchanged at 4.5 MHz and the line number remained at 525. Thus, the
resulting line scanning rate and field rate varied slightly from that of the monochrome values, but
stayed within the previously existing tolerances. The difference is exactly one part in one thou-
sand. Specifically, the line rate is 15.734 kHz, the field rate is 59.94 Hz, and the color subcarrier
is 3.578545 MHz.
3.1.2c PAL Color System

Except for some minor details, the color encoding principles for PAL are essentially the same as
those for NTSC. However, the phase of the color signal, EV = R – Y, is reversed by 180º from
line-to-line. This is done for the purpose of averaging, or canceling, certain color errors resulting
from amplitude and phase distortion of the color modulation sidebands. Such distortions might
occur as a result of equipment or transmission path problems.
The NTSC chroma signal expression within the frequency band common to both I and Q is
given by Equation (3.1.14), and the PAL chroma signal expression is given by Equation (3.1.15).
The PAL format employs equal bandwidths for the U and V color-difference signal compo-
nents that are about the same as the NTSC I signal bandwidth (1.3 MHz at 3 dB). There are slight
differences in the U and V bandwidth in different PAL systems because of the differences in
luminance bandwidth and sound carrier frequencies.
The V component was chosen for the line-by-line reversal process because it has a lower gain
factor than U and, therefore, is less susceptible to switching rate (1/2 fH) imbalance. Figure 3.1.4
illustrates the vector diagram for the PAL quadrature modulated and line-alternating color modu-
lation approach.
The result of the switching of the V signal phase at the line rate is that any phase errors pro-
duce complementary errors from V into the U channel. In addition, a corresponding switch of the
decoder V channel results in a constant V component with complementary errors from the U
channel. Thus, any line-to-line averaging process at the decoder, such as the retentivity of the eye
Figure 3.1.4 PAL color modulation vector diagram.
(simple PAL) or an electronic averaging technique such as the use of a 1-H delay element (stan-
dard PAL), produces cancellation of the phase (hue) error and provides the correct hue but with
somewhat reduced saturation; this error being subjectively much less visible.
Obviously, the PAL receiver must be provided with some means by which the V signal switch-
ing sequence can be identified. The technique employed is known as A B sync, PAL sync, or
swinging burst and consists of alternating the phase of the reference burst by ±45° at a line rate,
as shown in Figure 3.1.4. The burst is constituted from a fixed value of U phase and a switched
value of V phase. Because the sign of the V burst component is the same sign as the V picture
content, the necessary switching “sense” or identification information is available. At the same
time, the fixed-U component is used for reference carrier synchronization. Figure 3.1.5 shows
the degree to which horizontal frequency (dot) interlace of the color subcarrier components with
the luminance components is achieved in PAL.
To summarize, in NTSC, the Y components are spaced at fH intervals as a result of the hori-
zontal sampling (blanking) process. Thus, the choice of a color subcarrier whose harmonics are
also separated from each other by fH (as they are odd multiples of 1/2 fH) provides a half-line off-
set and results in a perfect “dot” interlace pattern that moves upward. Four complete field scans
are required to repeat a specific picture element “dot” position.
In PAL, the luminance components are also spaced at fH intervals. Because the V components
are switched symmetrically at half the line rate, only odd harmonics exist, with the result that the
(a)
(b)
Figure 3.1.5 NTSC and PAL frequency-interlace relationship: (a) NTSC 1/2-H interlace (four fields
for a complete color frame), (b) PAL 1/4-H and 3/4-H offset (eight fields for a complete color
frame.)
V components are spaced at intervals of fH. They are spaced at half-line intervals from the U
components which, in turn have fH spacing intervals because of blanking. If half-line offset were
used, the U components would be perfectly interlaced but the V components would coincide with
Y and, thus, not be interlaced, creating vertical, stationary dot patterns.
For this reason, in PAL, a 1/4-line offset for the subcarrier frequency is used as shown in Fig-
ure 3.1.5. The expression for determining the PAL subcarrier specific frequency for 625-line/50-
field systems is given by
1135 1
FSC = fH + fv
4 2
(3.1.16)
The additional factor 1/2 fV = 25 Hz is introduced to provide motion to the color dot pattern,
thereby reducing its visibility. The degree to which interlace is achieved is, therefore, not perfect,
but is acceptable, and eight complete field scans must occur before a specific picture element
“dot” position is repeated.
One additional function must be accomplished in relation to PAL color synchronization. In all
systems, the burst signal is eliminated during the vertical synchronization pulse period. Because
in the case of PAL, the swinging burst phase is alternating line-by-line, some means must be pro-
vided for ensuring that the phase is the same for the first burst following vertical sync on a field-
by-field basis. Therefore, the burst reinsertion time is shifted by one line at the vertical field rate
Figure 3.1.6 Meander burst-blanking gate timing diagram for systems B, G, H, and PAL/I. (Source:
CCIR.)
by a pulse referred to as the meander gate. The timing of this pulse relative to the A versus B
burst phase is shown in Figure 3.1.6.
The transmitted signal specifications for PAL systems include the basic features discussed
previously. Although a description of the great variety of receiver decoding techniques in com-
mon use is outside the scope and intent of this chapter, it is appropriate to review—at least
briefly—the following major features:
• “Simple” PAL relies upon the eye to average the line-by-line color switching process and can
be plagued with line beats known as Hanover bars caused by the system nonlinearities intro-
ducing visible luminance changes at the line rate.
• “Standard” PAL employs a 1-H delay line element to separate U color signal components
from V color signal components in an averaging technique, coupled with summation and sub-
traction functions. Hanover bars can also occur in this approach if an imbalance of amplitude
or phase occurs between the delayed and direct paths.
• In a PAL system, vertical resolution in chrominance is reduced as a result of the line averag-
ing processes. The visibility of the reduced vertical color resolution as well as the vertical
time coincidence of luminance and chrominance transitions differs depending upon whether
the total system, transmitter through receiver, includes one or more averaging (comb filter)
processes.
Thus, PAL provides a similar system to NTSC and has gained favor in many areas of the world,
particularly for 625-line/50-field systems.
Figure 3.1.7 SECAM FM color modulation system.
3.1.2d SECAM Color System

The “optimized” SECAM system, known as SECAM III, was the system adopted by France and
the USSR in 1967. The SECAM method has several features in common with NTSC, such as the
same EY´ signal and the same EB´ – E Y´ and ER´ – E Y´ color-difference signals. However, the
SECAM approach differs considerably from NTSC and PAL in the manner in which the color
information is modulated onto the subcarrier(s).
First, the R – Y and B – Y color difference signals are transmitted alternately in time sequence
from one successive line to the next; the luminance signal being common to every line. Because
there is an odd number of lines, any given line carries R – Y information on one field and B – Y
information on the next field. Second, the R – Y and B – Y color information is conveyed by fre-
quency modulation of different subcarriers. Thus, at the decoder, a 1-H delay element, switched
in time synchronization with the line switching process at the encoder, is required in order to
have simultaneous existence of the B – Y and R – Y signals in a linear matrix to form the G – Y
component.
The R – Y signal is designated as DR´ and the B – Y signal as DB´. The undeviated frequency
for the two subcarriers, respectively, is determined by
FOB = 272 fH = 4.250000 MHz (3.1.17)
FOR = 282 fH = 4.406250 MHz (3.1.18)
These frequencies represent zero color difference information (zero output from the FM dis-
criminator), or a neutral gray object in the televised scene.
As shown in Figure 3.1.7, the accepted convention for direction of frequency change with
respect to the polarity of the color difference signal is opposite for the DOB and DOR signals. A
Figure 3.1.8 SECAM color signal low-frequency preemphasis. (CCIR Rep. 624-2.)
positive value of DOR means a decrease in frequency, whereas a positive value of DOB indicates
an increase in frequency. This choice relates to the idea of keeping the frequencies representative
of the most critical color away from the upper edge of the available bandwidth to minimize dis-
tortions.
The deviation for DR´ is 280 kHz and DB´ is 230 kHz. The maximum allowable deviation,
including preemphasis, for DR´ = –506 kHz and +350 kHz, while the values for DB´ = –350 kHz
and +506 kHz.
Two types of preemphasis are employed simultaneously in SECAM. First, as shown in Figure
3.1.8, a conventional type of preemphasis of the low-frequency color difference signals is intro-
duced. The characteristic is specified to have a reference level break-point at 85 kHz (f1) and a
maximum emphasis of 2.56 dB. The expression for the characteristic is given as
1+ j ( f / f1 )
A=
1+ j ( f / 3 f1 )
(3.1.19)
where f = signal frequency in kHz.

A second form of preemphasis, shown in Figure 3.1.9, is introduced at the subcarrier level
where the amplitude of the subcarrier is changed as a function of the frequency deviation. The
expression for this inverted “bell” shaped characteristic is given as
Figure 3.1.9 SECAM high-frequency subcarrier preemphasis. (CCIR Rep. 624-2.)
Figure 3.1.10 Color versus line-and-field timing relationship for SECAM. Notes: (1) Two frames
(four fields) for picture completion; (2) subcarrier interlace is field-to-field and line-to-line of same
color.
⎛ f f ⎞
1× j16⎜ − c ⎟
⎝ fc f ⎠
G = MO
⎛ f f ⎞
1+ j1.26⎜ − c ⎟
⎝ fc f ⎠
(3.1.20)
where f = 4.286 MHz and 2M0 = 23 percent of the luminance amplitude.

(a )
(b )
Figure 3.1.11 SECAM line-identification technique: (a) line identification (bottle) signal, (b) hori-
zontal blanking interval. (Source CCIR.)
This type of preemphasis is intended to further reduce the visibility of the frequency modu-
lated subcarriers in low luminance level color values and to improve the signal-to-noise ratio (S/
N) in high luminance and highly saturated colors. Thus, monochrome compatibility is better for
pastel average picture level objects but sacrificed somewhat in favor of S/R in saturated color
areas.
Of course, precise interlace of frequency modulated subcarriers for all values of color modu-
lation cannot occur. Nevertheless, the visibility of the interference represented by the existence
of the subcarriers can be reduced somewhat by the use of two separate carriers, as is done in
SECAM. Figure 3.1.10 illustrates the line switching sequence. In the undeviated “resting” fre-
quency mode, the 2:1 vertical interlace in relation to the continuous color difference line switch-
ing sequence produces adjacent line pairs of fOB and fOR. In order to further reduce the subcarrier
“dot” visibility, the phase of the subcarriers (phase carries no picture information in this case) is
reversed 180º on every third line and between each field. This, coupled with the “bell” preem-
phasis, produces a degree of monochrome compatibility considered subjectively adequate.
As in PAL, the SECAM system must provide some means for identifying the line switching
sequence between the encoding and decoding processes. This is accomplished, as shown in Fig-
ure 3.1.11, by introducing alternate DR and DB color identifying signals for nine lines during the
vertical blanking interval following the equalizing pulses after vertical sync. These “bottle”
shaped signals occupy a full line each and represent the frequency deviation in each time
sequence of DB and DR at zero luminance value. These signals can be thought of as fictitious
green color that is used at the decoder to determine the line switching sequence.
During horizontal blanking, the subcarriers are blanked and a burst of fOB / fOR is inserted and
used as a gray level reference for the FM discriminators to establish their proper operation at the
beginning of each line.
Thus, the SECAM system is a line sequential color approach using frequency modulated sub-
carriers. A special identification signal is provided to identify the line switch sequence and is
especially adapted to the 625-line/50-field wideband systems available in France and the USSR.
It should be noted that SECAM, as practiced, employs amplitude modulation of the sound
carrier as opposed to the FM sound modulation in other systems.
3.1.2e Additional Systems of Historical Interest

Of the numerous system variations proposed over the intervening years since the development of
the NTSC system, at least two others, in addition to PAL and SECAM, should be mentioned
briefly. The first of the these was ART (additional reference transmission), which involved the
transmission of a continuous reference pilot carrier in conjunction with a conventional NTSC
color subcarrier quadrature modulation signal. A modification of this scheme involved a “multi-
burst” approach that utilized three color bursts—one at black level, one at intermediate gray
level, and one at white level—to be used for correcting differential phase distortion.
Another system, perhaps better known, was referred to as NIR or SECAM IV. Developed in
the USSR (NIR stands for Nauchni Issledovatelskaia Rabota or scientific discriminating work),
this system consists of alternating lines of (1) an NTSC-like signal using an amplitude and phase
modulated subcarrier and (2) a reference signal having U phase to demodulate the NTSC-like
signal. In the linear version, the reference is unmodulated, and in the nonlinear version, the
amplitude of the reference signal is modulated with chrominance information.
3.1.2f Summary and Comparisons of Systems Standards

History has shown that it is exceedingly difficult to obtain total international agreement on “uni-
versal” television broadcasting standards. Even with the first scheduled broadcasting of mono-
chrome television in 1936 in England, the actual telecasting started using two different systems
on alternate days from the same transmitter. The Baird system was 250 lines (noninterlaced) with
a 50 Hz frame rate while the EMI (Electric and Musical Industries) system was 405 lines (inter-
laced) with a 25 Hz frame rate.
These efforts were followed in 1939 in the U.S. by broadcasting a 441 line interlaced system
at 60 fields per second (the Radio Manufacturers Association, RMA, system). In 1941, the first
(original) NTSC initiated the present basic monochrome standards in the U.S. of 525 lines (inter-
laced) at 60 fields per second, designated as System M by the CCIR. In those early days, the dif-
ferences in power line frequency were considered as important factors and were largely
responsible for the proliferation of different line rates versus field rates, as well as the wide vari-
ety of video bandwidths. However, the existence and extensive use of monochrome standards
over a period of years soon made it a top priority matter to assume reciprocal compatibility of
any developing color system.
CCIR documents [3] define recommended standards for worldwide color television systems
in terms of the three basic color approaches—NTSC, PAL, and SECAM. The variations—at
least 13 of them—are given alphabetical letter designations; some represent major differences
while others signify only very minor frequency allocation differences in channel spacings or the
differences between the VHF and UHF bands. The key to understanding the CCIR designations
lies in recognizing that the letters refer primarily to local monochrome standards for line and
field rates, video channel bandwidth, and audio carrier relative frequency. Further classification
in terms of the particular color system is then added for NTSC, PAL, or SECAM as appropriate.
For example, the letter “M” designates a 525-line/60-field, 4.2 MHz bandwidth, 4.5 MHz sound
carrier monochrome system. Thus, M(NTSC) describes a color system employing the NTSC
technique for introducing the chrominance information within the constraints of the basic mono-
chrome signal values. Likewise, M(PAL) would indicate the same line/field rates and band-
widths but employing the PAL color subcarrier modulation approach.
In another example, the letters “I” and “G” relate to specific 625-line/50-field, 5.0 or 5.5
MHz bandwidth, 5.5 or 6.0 MHz sound carrier monochrome standards. Thus, G(PAL) would
describe a 625-line/50-field, 5.5 MHz bandwidth, color system utilizing the PAL color subcarrier
modulation approach. The letter “L” refers to a 625-line/50-field, 6.0 MHz bandwidth system to
which the SECAM color modulation method has been added (often referred to as SECAM III).
System E is an 819-line/50-field, 10 MHz bandwidth, monochrome system. This channel was
used in France for early SECAM tests and for system E transmissions.
Some general comparison statements can be made about the underlying monochrome systems
and existing color standards:
• There are three different scanning standards: 525-lines/60-fields, 625-lines/50-fields, and
819-lines/50-fields.
• There are six different spacings of video-to-sound carriers, namely 3.5, 4.5, 5.5, 6.0, 6.5, and
11.15 MHz.
• Some systems use FM and others use AM for the sound modulation.
• Some systems use positive polarity (luminance proportional to voltage) modulation of the
video carrier while others, such as the U.S. (M)NTSC system, use negative modulation.
• There are differences in the techniques of color subcarrier encoding represented by NTSC,
PAL, and SECAM, and of course, in each case there are many differences in the details of
various pulse widths, timing, and tolerance standards.
The signal in the M(NTSC) system occupies the least total channel width, which when the
vestigial sideband plus guard bands are included, requires a minimum radio frequency channel
spacing of 6 MHz. (See Figure 3.1.12.) The L(III) SECAM system signal occupies the greatest
channel space with a full 6 MHz luminance bandwidth. Signals from the two versions of PAL lie
in between and vary in vestigial sideband width as well as color and luminance bandwidths.
NTSC is the only system to incorporate the I, Q color acuity bandwidth variation. PAL mini-
mizes the color quadrature phase distortion effects by line-to-line averaging, and SECAM avoids
this problem by only transmitting the color components sequentially at a line-by-line rate.
Figure 3.1.12 Bandwidth comparison among NTSC, PAL, and SECAM systems.
3.1.3 NTSC Stereo System

The TV stereo system is a modification of the standard FM stereo system [4–6]. (See Figure
3.1.13.) The main differences are:
• 25-kHz deviation for main channel (L + R).
• 50-kHz deviation for subchannel (L – R). This is 2 times, or 6 dB, greater than for the main
channel.
• Noise reduction in the subchannel (L – R).
• Pilot frequency equal to fH (15,734 Hz).
Figure 3.1.13 Typical television MTS system block diagram.
(a )
(b )
Figure 3.1.14 Baseband frequency allocations: (a) stereophonic FM, (b) BTSC TV stereophonic
system.
Compatibility with existing monophonic receivers was the reason behind the selection of 25-kHz
deviation for the main channel audio. The increased deviation of the subchannel and the use of
noise reduction was intended to maintain at least 50 dB S/N in outlying areas. With the noise
reduction compandor in the stereo subchannel and the 6-dB increase in level, the L and R S/N
should be dependent only on the main-channel S/N, which was found to be approximately 63 dB
in system tests. The choice fH as the pilot was to minimize buzz-beat interference.
Figure 3.1.14 shows the FM-versus-TV stereo baseband spectrum. Table 3.1.3 shows the
aural carrier-modulation standards for the TV stereo system. Figure 3.1.5 is a block diagram of a
typical TV stereo generator.
Some of the specifications for TV stereo generators are not different from those for other
audio equipment except that they cannot be verified without use of a decoder and that the
decoder contribution to performance must be recognized.
There are at least two operational modes for TV stereo generators: the normal (noise reduc-
tion) operational mode and a test mode called 75-μs equivalent mode. In the 75-μs equivalent
mode, the noise reduction system in L – R is replaced with a 75-μs preemphasis network identi-
cal to that in L + R. This function was included to allow noise, distortion, and separation mea-
surements to be made without the level-dependent degradation caused by noise reduction.
Subchannel filters serve to limit the noise-reduction-control-line bandwidth and to control the
out-of-band energy created by the noise reduction circuit. When there is no audio input to the
noise reduction circuitry, the spectral compressor gain is at maximum, creating a high-level par-
abolic noise spectrum which needs to be bandwidth-limited to audio frequencies to prevent spec-
Table 3.1.3 Modulation Standards for the TV MTS System
Service or Modulating Modulating Audio Subcarrier Subcarrier Subcarrier Aural Carrier

Signal Signal Frequency Processing/ Frequency Modulation Deviation, Peak
Range, kHz Preemphasis (fH = 15.734 Type kHz Deviation,
kHz) kHz
Monophonic L+R 0.05–15 75 μs 25 (note 1)
Pilot fH 5
Stereophonic L–R 0.05–15 BTSC 2fH AM-DSB SC 50 (note 1)
compression
Second audio 0.05–10 BTSC 5fH FM 10 15
program compression
Professional Voice 0.3–3.4 150 μs FM
6.5fH 3 3
channel Data 0–1.5 0 FSK
Note (1): Sum does not exceed 50 kHz
trum spillover. Because the pilot is separated from the subchannel by only 734 Hz, the filter
slope needs to be very sharp to provide pilot protection with flat response to 15,000 Hz.
Any out-of-band information on the noise-reduction-control line (such as fH) after the sub-
channel filter will cause the noise reduction circuitry to misencode, causing degradation of
received separation and frequency response when decoded. To ensure stereo separation greater
than 40 dB, the subchannel filters need to be matched to within 0.08-dB amplitude difference
and 1° phase difference. For good overall frequency response, the filter's response should be less
than ±0.2 dB from 50 to 15,000 Hz.
The stereo-generator frequency response should be less than ±1 dB from 50 to 15,000 Hz
(with noise reduction) and ±0.5 dB without noise reduction to help meet the total-system
response goals.
3.1.4 References
1. Herbstreit, J. W., and J. Pouliquen: “International Standards for Color Television,” IEEE
Spectrum, IEEE, New York, N.Y., March 1967.
2. Fink, Donald G. (ed.): Color Television Standards, McGraw-Hill, New York, N.Y., 1955.
3. “CCIR Characteristics of Systems for Monochrome and Colour Television—Recommen-
dations and Reports,” Recommendations 470-1 (1974–1978) of the Fourteenth Plenary
Assembly of CCIR in Kyoto, Japan, 1978.
4. Eilers, C.: “The Zenith Multichannel TV Sound System,” Proc. 38th NAB Eng. Conf.,
National Association of Broadcasters, Washington, D.C., 1984.
5. Federal Communications Commission: Multichannel Television Sound Transmission and
Audio Processing Requirements for the BTSC System, OST Bulletin 60, FCC, Washington,
D.C.
6. Electronic Industries Association: Multichannel Television Sound—BTSC System Recom-
mended Practices, EIA, Washington, D.C., ETA Television Systems Bulletin 5, July 1985.
Figure 3.1.15 Stereo-generator block diagram.

3.1.5 Bibliography
Carnt, P. S., and G. B. Townsend: Colour Television—Volume 1: NTSC; Volume 2: PAL and
SECAM, ILIFFE Bookes Ltd. (Wireless World), London, 1969.
Hirsch, C. J.: “Color Television Standards for Region 2,” IEEE Spectrum, IEEE, New York, N.Y.,
February 1968.
Pritchard, D. H.: “U.S. Color Television Fundamentals—A Review,” SMPTE Journal, SMPTE,
White Plains, N.Y., vol. 86, pp. 819–828, November 1977.
Roizen, J.: “Universal Color Television: An Electronic Fantasia,” IEEE Spectrum, IEEE, New
York, N.Y., March 1967.
g g
Chapter
3.2
Ghost Canceling Reference Signal
3.2.1 Introduction
The quality of a received analog television image depends on several factors. Some relate to the
transmission system while others relate to the reception equipment. Ghosting, which is caused by
reflections from buildings, aircraft, and other objects in the environment, falls somewhere in
between.
3.2.1a History and Purpose

The ghost canceling reference (GCR) signal forms a key link between the broadcaster and the
viewer. It carries all of the information that the receiver needs to successfully cancel ghosts in
analog television reception. Realizing that commercial ghost cancellation would not be possible
without a standard GCR signal, the Advanced Television Systems Committee (ATSC) began
searching for the most-suitable GCR in 1989. The first step was examination of the GCR
adopted as a national standard by Japan’s Broadcast Television Association (BTA). Between
1989 and 1992, the ATSC Specialist Group on Ghost Cancellation compared the BTA signal
with candidates submitted by AT&T/Zenith, Sarnoff/Thomson, Samsung, and Philips. The initial
comparisons considered theoretical and computer simulation, and laboratory criteria. NAB,
MSTV, Cable Labs, PBS, and other organizations conducted the final and most-extensive field
tests in metropolitan Washington, D.C. during late 1991 and again (in a limited form) in mid-
1992. Based on these test results, the ATSC recommended that the Philips GCR be chosen as the
U.S. standard signal.
3.2.1b Basic Parameters

The GCR signal occupies one line of the vertical blanking interval (VBI); line 19 was chosen as
the most appropriate VBI line into which to insert the GCR. Because FCC regulations already
reserved line 19 for a vertical interval (color) reference (VIR) signal, an FCC waiver was
required to replace the VIR signal with the GCR. (Research proved that only a few TV receivers,
manufactured through 1985 by the General Electric Company, used the VIR signals.) The FCC
approved the substitution of the GCR into line 19 on October 27, 1992, making the Philips GCR
3-35
g g
Figure 3.2.1 Causes of multipath in television broadcasting.
and its VBI location the official U.S. voluntary standard. The voluntary nature of the standard
relies on the cooperation of broadcasters and receiver manufacturers.
3.2.2 System Overview

Television ghosts are multiple images caused by multipath echoes. In theory, TV signals travel
along straight lines. The signal may have a short, direct path between the transmitting antenna
and the receiving antenna. (See Figure 3.2.1.) The same signal also can reach the receiver
antenna over a different, longer echo path because of reflections off nearby buildings and other
objects. Under such conditions, the receiver will display two images—a strong, main image and
a weaker, shifted echo or ghost. These ghosts can seriously degrade received image quality.
Two types of ghosts exist:
• Post-delay ghost, where the ghosted image is a weaker replica of the main image, delayed to
occur later in time.
• Pre-ghost, where the ghost occurs before the main image. Pre-ghosts can occur in cable sys-
tems, where it is possible for the strong, main image to have a propagation delay through the
cable system longer than the time it takes for a weak, direct-broadcast pickup to be received.
These two ghosts require different methods of cancellation.
3.2.2a System Architecture

Ghost cancellation of television signals is a two-step process:
• Characterization of the ghosting channel
• Cancellation of the ghosts through the use of adaptive filters
To facilitate the channel characterization step, the broadcasting station transmits a ghost-can-
cellation reference signal. The receiver compares the received, ghosted version of the GCR sig-
nal with a stored replica of the same signal. The results of such comparisons are computed by a
g g
Figure 3.2.2 Block diagram of the GCR system in a receiver.
digital signal processor (DSP), which uses ghost cancellation algorithms to calculate coefficients
to be fed to digital adaptive filters that cancel the ghosts. This process occurs continuously, with
the coefficients of the filters being constantly updated to follow transient conditions, as mea-
sured from the received GCR.
Figure 3.2.2 shows a simplified block diagram of the GCR system. The ghosted received
GCR is captured, providing the computed settings for the ghost-canceling filter. The corrected
signal, with the ghosts removed, is converted back to analog for display.
The GCR acts as a test signal in characterizing the impulse or frequency response of the
“black box” representation of the ghosting transmission channel. The ATSC GCR standard
resembles a swept-frequency chirp pulse. It is not such a pulse, but is carefully synthesized to
have a flat frequency spectrum over the entire 4.2 MHz band of interest to NTSC applications.
(The spectrum of a linear FM chirp pulse is not flat.) This signal was preferred to the BTA GCR
because it provides more energy, permitting more reliable operation under noisier conditions.
Although the primary purpose of the GCR signal is for ghost cancellation, the resemblance
between the GCR and a linear FM chirp offers other uses. For example, broadcasters can use this
pulse as a qualitative indicator of the frequency response of their system.
3.2.2b Specification of the GCR Signal

The GCR signal permits detection of ghosts from –3 μs to +45 μs [1]. The signal has a flat spec-
trum and high energy up to 4.1 MHz, and has a very low level of energy beyond 4.3 MHz (Figure
3.2.3). The normalized GCR signal as a function of time is shown in Figure 3.2.4.
The GCR signal is placed on line 191of the vertical blanking interval on a 30 IRE pedestal.
The pedestal has a rising transition from 0 IRE to 30 IRE 9.5 μs after the leading edge of hori-
zontal sync (defined at the 50 percent point). The falling transition from 30 IRE to 0 IRE is 58.5
μs after the leading edge of horizontal sync. Both transitions have a 4T transition curve.
The GCR signal has a time duration of 35.5 μs (measured at 1 percent of the maximum value)
and begins 12.0 μs after the leading edge of horizontal sync. The first peak (positive or negative)
1. Subject to reservation of line 19 by the FCC exclusively for the optional placement of the
GCR signal.
g g
Normalized Signal Amplitude

Normalized Energy Content
1.2 1.0
1.0
0.5
0.8
0.6 0.0
0.4
-0.5
0.2
0.0 -1.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 0 100 200 300 400 500 600 700 800 900
Frequency (MHz) Samples at 14.32 MHz
Figure 3.2.3 Spectrum of the GCR signal. (From [1]. Figure 3.2.4 GCR signal amplitude as a function of
Used with permission.) time. (From [1]. Used with permission.)
100 IRE Amplitude 100

IRE Amplitude
50 50
0 0
-50 -50
0 100 200 300 400 500 600 700 800 900 0 100 200 300 400 500 600 700 800 900
Samples at 14.32 MHz Samples at 14.32 MHz
Figure 3.2.5 GCR signal on Line A. ( From [1]. Used Figure 3.2.6 GCR signal on Line B. (From [1]. Used
with permission.) with permission.)
is 16.7 μs after the leading edge of horizontal sync. The GCR signal varies from –10 to +70 IRE
(the pedestal is the average of these extreme values).
Waveforms of the GCR signal on the pedestal are shown in Figures 3.2.5 and Figure 3.2.6,
and represent line A and line B, respectively. Line A and line B have the same 30 IRE pedestal
but the GCR polarity is inverted from line A to line B. The line A and line B signals are con-
tained in an 8-field sequence as follows:
field 1 – line A; field 2 – line B; field 3 – line A; field 4 – line B
field 5 – line B; field 6 – line A; field 7 – line B; field 8 – line A
Numerical values of the GCR signal as a function of time are given ATSC Standard A/49. The
values were calculated from
g g
A⎧ Ω⎡ 0 ⎫
f (t ) = ⎨ ∫ ⎣ cos (bω2 )+ j sin (bω2 )⎤
⎦W ( ω) e d ω+ ∫ ⎡
j ωt
⎣ cos (bω )− j sin (bω )⎤
2 2
⎦W ( ω) e d ω⎬
j ωt
2π ⎩ 0 −Ω ⎭
Where:
A = 9.0
b = 110.0
Ω = 4.3/7.16 π radians/s
W(ω ) is the window function
π
c ⎡⎛ 1 1 ⎞⎛ 1 Ω1 j γt ⎞⎤ − j ωt
W ( ω) = ∫ ⎢⎜ + cos ( ct )⎟⎜ ∫ e d γ ⎟⎟⎥⎥e dt
π ⎢
⎣⎝2 2 ⎠⎜
⎝ 2 π −Ω1 ⎠⎦
−
c
Where:
c = π/40.0 radians/s
Ω1 = 4.15/7.16 π radians/s
3.2.2c Receiver Implementation Considerations

Ridding a signal of ghost impairments is relatively straightforward. The standard GCR signal
inserted by the broadcaster experiences the same degrading influences as the video signal. A
receiver module digitizes incoming video and compares a standard GCR reference signal within
the module to the digitized input. This comparison produces a dynamic control signal for the fil-
ter that removes information from the picture consistent with the amount of delay experienced by
the reference signal.
In a cable system, the operator can use a ghost-cancellation unit at the head-end to remove
impairments resulting from multipath reception at the receive location. This will improve what is
distributed to the subscribers, but it cannot remove any additional ghosting that occurs because
of a viewer’s location, relative to that of the broadcast station and the cable head-end reception
point.
Although the broadcasters’ expense and effort to transmit the GCR signal is low, such a signal
is useless without receivers that are able to use it. Fortunately, the added cost of the ghost cancel-
ler is low. The most-critical parts of this system are the digital filter chip hardware, and the soft-
ware needed to characterize the echo channel and compute the filter coefficients. For modern
highly-integrated receiver systems, the additional signal processing overhead is minimal.
3.2.3 References
1. ATSC: “Ghost Canceling Reference Signal for NTSC,” Advanced Television Systems
Committee, Washington, D.C., document A/49, 1993.
g g
g g
Chapter
3.3
The ATSC DTV System
3.3.1 Introduction1
The Digital Television (DTV) Standard has ushered in a new era in television broadcasting [1].
The impact of DTV is more significant than simply moving from an analog system to a digital
system. Rather, DTV permits a level of flexibility wholly unattainable with analog broadcasting.
An important element of this flexibility is the ability to expand system functions by building
upon the technical foundations specified in ATSC standards A/53, “ATSC Digital Television
Standard,” and A/52, “Digital Audio Compression (AC-3) Standard.”
With NTSC, and its PAL and SECAM counterparts, the video, audio, and some limited data
information are conveyed by modulating an RF carrier in such a way that a receiver of relatively
simple design can decode and reassemble the various elements of the signal to produce a pro-
gram consisting of video and audio, and perhaps related data (e.g., closed captioning). As such, a
complete program is transmitted by the broadcaster that is essentially in finished form. In the
DTV system, however, additional levels of processing are required after the receiver demodu-
lates the RF signal. The receiver processes the digital bit stream extracted from the received sig-
nal to yield a collection of program elements (video, audio, and/or data) that match the service(s)
that the consumer has selected. This selection is made using system and service information that
is also transmitted. Audio and video are delivered in digitally compressed form and must be
decoded for presentation. Audio may be monophonic, stereo, or multi-channel. Data may supple-
ment the main video/audio program (e.g., closed captioning, descriptive text, or commentary) or
it may be a stand-alone service (e.g., a stock or news ticker).
1. This chapter is based on: ATSC, “Guide to the Use of the Digital Television Standard,”
Advanced Television Systems Committee, Washington, D.C., Doc. A/54A, December 4,
2003; and ATSC, “ATSC Digital Television Standard,” Advanced Television Systems Com-
mittee, Washington, D.C., Doc. A/53C, May 21, 2004. Used with permission.
Editor’s note: This chapter provides an overview of the ATSC DTV transmission system based
on ATSC A/53C, and A54A. For full details on this system, readers are encouraged to
download the source documents from the ATSC Web site (http://www.atsc.org). All ATSC
Standards, Recommended Practices, and Information Guides are available at no charge..
3-41
y
The nature of the DTV system is such that it is possible to provide new features that build
upon the infrastructure within the broadcast plant and the receiver. One of the major enabling
developments of digital television, in fact, is the integration of significant processing power in
the receiving device itself. Historically, in the design of any broadcast system—be it radio or
television—the goal has always been to concentrate technical sophistication (when needed) at the
transmission end and thereby facilitate simpler receivers. Because there are far more receivers
than transmitters, this approach has obvious business advantages. While this trend continues to
be true, the complexity of the transmitted bit stream and compression of the audio and video
components require a significant amount of processing power in the receiver, which is practical
because of the enormous advancements made in computing technology. Once a receiver reaches
a certain level of sophistication (and market success) additional processing power is essentially
“free.”
The Digital Television Standard describes a system designed to transmit high quality video
and audio and ancillary data over a single 6 MHz channel. The system can deliver about 19
Mbits/s in a 6 MHz terrestrial broadcasting channel and about 38 Mbits/s in a 6 MHz cable tele-
vision channel. This means that encoding HD video essence at 1.106 Gbits/s 1 (highest rate pro-
gressive input) or 1.244 Gbits/s2 (highest rate interlaced picture input) requires a bit rate
reduction by about a factor of 50 (when the overhead numbers are added, the rates become
closer). To achieve this bit rate reduction, the system is designed to be efficient in utilizing avail-
able channel capacity by exploiting complex video and audio compression technology.
The compression scheme optimizes the throughput of the transmission channel by represent-
ing the video, audio, and data sources with as few bits as possible while preserving the level of
quality required for the given application.
The RF/transmission subsystems described in the Digital Television Standard are designed
specifically for terrestrial and cable applications. The structure is such that the video, audio, and
service multiplex/transport subsystems are useful in other applications.
3.3.2 DTV System Overview

A basic block diagram representation of the ATSC DTV system is shown in Figure 3.3.1 [2]. The
digital television system can be seen to consist of three subsystems [3].
• Source coding and compression
• Service multiplex and transport
• RF/transmission
Source coding and compression refers to the bit-rate reduction methods (data compression)
appropriate for application to the video, audio, and ancillary digital data streams. The term ancil-
lary data encompasses the following functions:
• Control data
1. 720 × 1280 × 60 × 2 × 10 = 1.105920 Gbits/s (the 2 represents the factor needed for 4:2:2
color subsampling, and the 10 is for 10-bit systems).
2. 1080 × 1920 × 30 × 2 × 10 = 1.244160 Gbits/s (the 2 represents the factor needed for 4:2:2
color subsampling, and the 10 is for 10-bit systems).
y
Service Multiplex and Transport RF/Transmission System

Video Subsystem
Video
Video Source Coding
and Compression
Channel
Transport
Audio Subsystem Coding
Audio
Service Multiplex
Audio Source Coding
and Compression
Modulation
Ancillary Data
Control Data
Receiver Characteristics
Figure 3.3.1 Digital terrestrial television broadcasting model. (From [2]. Used with permission.)
• Conditional-access control data

• Data associated with the program audio and video services, such as closed captioning
Ancillary data also can refer to independent program services. The purpose of the coder is to
minimize the number of bits needed to represent the audio and video information.
Service multiplex and transport refers to the means of dividing the digital data stream into
packets of information, the means of uniquely identifying each packet or packet type, and the
appropriate methods of multiplexing video data-stream packets, audio data-stream packets, and
ancillary data-stream packets into a single data stream. In developing the transport mechanism,
interoperability among digital media—such as terrestrial broadcasting, cable distribution, satel-
lite distribution, recording media, and computer interfaces—was a prime consideration. The
DTV system employs the MPEG-2 transport-stream syntax for the packetization and multiplex-
ing of video, audio, and data signals for digital broadcasting systems [4]. The MPEG-2 trans-
port-stream syntax was developed for applications where channel bandwidth or recording media
capacity is limited, and the requirement for an efficient transport mechanism is paramount.
RF/Transmission refers to channel coding and modulation. The channel coder takes the data
bit stream and adds additional information that can be used by the receiver to reconstruct the data
from the received signal which, because of transmission impairments, may not accurately repre-
y
f27 MHz Program Clock

Reference
33 9 program_clock_reference_base
Frequency program_clock_reference_extension
Divider
Network Adaptation
Header
fv fa Encoder
Video In
Video Transport fTP FEC and fsym RF Out
A/D VSB
Encoder Encoder Sync
Modulator
Insertion
Audio In
Audio
A/D Encoder
Figure 3.3.2 High-level view of the DTV encoding system. (From [2]. Used with permission.)
sent the transmitted signal. The modulation (or physical layer) uses the digital data-stream infor-
mation to modulate the transmitted signal. The modulation subsystem offers two modes:
• Terrestrial broadcast mode (8-VSB)
• High-data-rate mode (16-VSB).
Figure 3.3.2 gives a high-level view of the encoding equipment. This view is not intended to
be complete, but is used to illustrate the relationship of various clock frequencies within the
encoder. There are two domains within the encoder where a set of frequencies are related: the
source-coding domain and the channel-coding domain.
The source-coding domain, represented schematically by the video, audio, and transport
encoders, uses a family of frequencies that are based on a 27 MHz clock (f27MHz ). This clock is
used to generate a 42-bit sample of the frequency, which is partitioned into two elements defined
by the MPEG-2 specification:
• The 33-bit program clock reference base
• The 9-bit program clock reference extension
The 33-bit program clock reference base is equivalent to a sample of a 90 kHz clock that is
locked in frequency to the 27 MHz clock, and is used by the audio and video source encoders
when encoding the presentation time stamp (PTS) and the decode time stamp (DTS). The audio
and video sampling clocks, fa and fv , respectively, must be frequency-locked to the 27 MHz
clock. This condition can be expressed as the requirement that there exist two pairs of integers,
(na, ma) and (nV, mV), such that
na
fa = × 27
------ MHz (3.3.1)
ma
y
and
n
fv = ----v-- × 27 MHz (3.3.2)
mv
The channel-coding domain is represented by the forward error correction/sync insertion sub-
system and the vestigial sideband (VSB) modulator. The relevant frequencies in this domain are
the VSB symbol frequency ( fsym) and the frequency of the transport stream ( fTP), which is the
frequency of transmission of the encoded transport stream. These two frequencies must be
locked, having the relation
188 312
fTP = 2 × -------- -------- fsym (3.3.3)
208 313
The signals in the two domains are not required to be frequency-locked to each other and, in
many implementations, will operate asynchronously. In such systems, the frequency drift can
necessitate the occasional insertion or deletion of a null packet from within the transport stream,
thereby accommodating the frequency disparity.
A basic block diagram representation of the overall ATSC DTV system is shown in Figure
3.3.3.
3.3.2a Basic Video Systems Characteristics

Table 3.3.1 lists the primary television production standards that define video formats relating to
compression techniques applicable to the ATSC DTV standard. These picture formats may be
derived from one or more appropriate video input formats.
The DTV video-compression algorithm conforms to the Main Profile syntax of ISO/IEC
13818-2 (MPEG-2). The allowable parameters are bounded by the upper limits specified for the
Main Profile/High Level. Table 3.3.2 lists the allowed compression formats under the ATSC
DTV standard.
3.3.2b Receiver
The ATSC receiver must recover the bits representing the original video, audio and other data
from the modulated signal [1]. In particular, the receiver must:
• Tune the selected 6 MHz channel
• Reject adjacent channels and other sources of interference
• Demodulate (equalize as necessary) the received signal, applying error correction to produce
a transport bit stream
• Identify the elements of the bit stream using a transport layer processor
• Select each desired element and send it to its appropriate processor
• Decode and synchronize each element
y
Transmitter
Sources for encoding
(video, audio, data)
packetization and
multiplexing
Application
Transport
Encoders
* Mod-
* ulator
*
* Transport
Stream
elementary
streams,PES or
private_sections
Sy stem Time
Clock
RF Transmission
PSI/PSIP
System Time Clock

Receiver clock
control
depacketization and
demultiplexing
Presentation
Application
Transport
Decoders
Demod-
*
* ulator
*
* Transport Note: PSIP ref ers to ATSC
elementary Stream with Standard (A/65) which
error was dev eloped
streams, PES or
signaling subsequent to ATSC
private_sections,
with error signaling Standards (A/52), (A/53)
and to this Guide. For
details regarding PSIP
PSI/PSIP ref er to ATSC Standard
(A/65).
Figure 3.3.3 Sample organization of functionality in a transmitter-receiver pair for a single DTV
program. (From [1]. Used with permission.)
Table 3.3.1 Standardized Video Input Formats
Video Standard Active Lines Active Samples/Line

SMPTE 274M, SMPTE 295M (50 Hz) 1080 1920
SMPTE 296M 720 1280
ITU-R Rec. 601-4 483 720
SMPTE 293M (59.94, p)
SMPTE 294M (59.94, p)
• Present the programming

Noise, interference, and multipath are elements of the terrestrial transmission path, and the
receiver circuits are expected to deal with these impairments.
y
Table 3.3.2 ATSC DTV Compression Format Constraints (After [2].)
Vertical Size Value Horizontal Size Value Aspect Ratio Frame-Rate Code Scanning Sequence
1 1920 16:9, square pixels 1,2,4,5 Progressive
1080
4,5 Interlaced
720 1280 16:9, square pixels 1,2,4,5,7,8 Progressive
480 704 4:3, 16:9 1,2,4,5,7,8 Progressive
4,5 Interlaced
640 4:3, square pixels 1,2,4,5,7,8 Progressive
4,5 Interlaced
Frame-rate code: 1 = 23.976 Hz, 2 = 24 Hz, 4 = 29.97 Hz, 5 = 30 Hz, 7 = 59.94 Hz, 8 = 60 Hz
1
Note that 1088 lines actually are coded in order to satisfy the MPEG-2 requirement that the coded vertical size be
a multiple of 16 (progressive scan) or 32 (interlaced scan).
Optional
Reed- Data RF
Data Trellis Pilot Pre-equalizer VSB Up-
Solomon Inter- MUX
Randomizer Encoder Insertion Filter Modulator Converter
Encoder leaver
Segment Sync
Field Sync
Figure 3.3.4 Simplified block diagram of an example VSB transmitter. (From [2]. Used with per-
mission.)
3.3.3 DTV Transmission Characteristics

The terrestrial broadcast mode (8-VSB) will support a payload data rate of approximately 19.28
Mbits/s in a 6 MHz channel [2]. (In the interest of simplicity, the bit stream data rate is rounded
to two decimal points.) A functional block diagram of a representative 8-VSB terrestrial broad-
cast transmitter is shown in Figure 3.3.4. The basic parameters of the VSB transmission modes
are listed in Table 3.3.3. The input to the transmission subsystem from the transport subsystem is
a serial data stream composed of 188-byte MPEG-compatible data packets (including a sync byte
and 187 bytes of data, which represent a payload data rate of 19.28 Mbits/s).
The incoming data is randomized, then processed for forward error correction (FEC) in the
form of Reed-Solomon (RS) coding (whereby 20 RS parity bytes are added to each packet). This
is followed by 1/6-data-field interleaving and 2/3-rate trellis coding. The randomization and FEC
processes are not applied to the sync byte of the transport packet, which is represented in trans-
mission by a data segment sync signal. Following randomization and FEC processing, the data
y
Table 3.3.3 Parameters for VSB Transmission Modes (After [1].)

Parameter Terrestrial Mode High Data Rate Mode
Channel bandwidth 6 MHz 6 MHz
Guard bandwidth 11.5 percent 11.5 percent
Symbol rate 10.76… Msymbols/s 10.76… Msymbols/s
Bits per symbol 3 4
Trellis FEC 2/3 rate None
Reed-Solomon FEC T = 10 (207,187) T = 10 (207,187)
Segment length 832 symbols 832 symbols
Segment sync 4 symbols per segment 4 symbols per segment
Frame sync 1 per 313 segments 1 per 313 segments
Payload data rate 19.39 Mbps 38.78 Mbps
Analog co-channel rejection Analog rejection filter in receiver N/A
Pilot power contribution 0.3 dB 0.3 dB
C/N threshold ~ 14.9 dB ~ 28.3 dB
4 828 Sym bols
packets are formatted into data frames for Field Sync #1
transmission, and data segment sync and

data field sync are added. S
313
Figure 3.3.5 shows how the data is orga- Segm ents
e Data + FEC 24.2
g ms
nized for transmission. Each data frame con- m
sists of two data fields, each containing 313 e
n
data segments. The first data segment of t
each data field is a unique synchronizing sig- S
Field Sync #2
nal (data field sync) and includes the train- y

n
ing sequence used by the equalizer in the c
313 Data + FEC
receiver. The remaining 312 data segments Segm ents
24.2
ms
each carry the equivalent of the data from
one 188-byte transport packet plus its associ-
ated FEC overhead. The actual data in each Test Segm ent
data segment comes from several transport 1 Segm ent
packets because of data interleaving. Each = 77.3 us
data segment consists of 832 symbols. The

Figure 3.3.5 The basic VSB data frame. ( From
first four symbols are transmitted in binary
form and provide segment synchronization.
This data segment sync signal also represents
the sync byte of the 188-byte MPEG-compatible transport packet. The remaining 828 symbols of
each data segment carry data equivalent to the remaining 187 bytes of a transport packet and its
associated FEC overhead. These 828 symbols are transmitted as 8-level signals and, therefore,
carry 3 bits per symbol. Thus, 2484 (828 × 3) bits of data are carried in each data segment,
which exactly matches the requirement to send a protected transport packet:
• 187 data bytes + 20 RS parity bytes = 207 bytes
• 207 bytes × 8 bits/byte = 1656 bits
• 2/3-rate trellis coding requires 3/2 × 1656 bits = 2484 bits.
y
1.
.7
Pilot
0
Suppressed
Carrier
.31 5.38 MHz .31
6.0 MHz
Figure 3.3.6 Nominal VSB channel occupancy. (From [2]. Used with permission.)
The exact symbol rate Sr is given by
4.5
Sr = -------- × 684 = 10.76 … MHz (3.3.4)
286
The frequency of a data segment fseg is given by
Sr 3
fseg = -------- = 12.94 …× 10 data segments/s (3.3.5)
832
The data frame rate fframe is given by
f
seg 20.66 … frames/s
f
frame = -626
------- = (3.3.6)
The symbol rate Sr and the transport rate Tr are locked to each other in frequency.
The 8-level symbols combined with the binary data segment sync and data field sync signals
are used to suppressed-carrier-modulate a single carrier. Before transmission, however, most of
the lower sideband is removed. The resulting spectrum is flat, except for the band edges where a
nominal square-root raised-cosine response results in 620 kHz transition regions. The nominal
VSB transmission spectrum is shown in Figure 3.3.6. It includes a small pilot signal at the sup-
pressed-carrier frequency, 310 kHz from the lower band edge.
3.3.3a Channel Error Protection and Synchronization

All payload data is carried with the same priority [2]. A data randomizer is used on all input data
to randomize the data payload (except for the data field sync, data segment sync, and RS parity
bytes). The data randomizer exclusive-ORs (XORs) all the incoming data bytes with a 16-bit
maximum-length pseudorandom binary sequence (PRBS), which is initialized at the beginning
of the data field. The PRBS is generated in a 16-bit shift register that has nine feedback taps.
Eight of the shift register outputs are selected as the fixed randomizing byte, where each bit from
y
this byte is used to individually XOR the corresponding input data bit. The randomizer-generator
polynomial and initialization are shown in Figure 3.3.7.
Although a thorough knowledge of the channel error protection and synchronization system
is not required for typical end-users, a familiarity with the basic principles of operation—as out-
lined in the following sections—is useful in understanding the important functions performed.
Reed-Solomon Encoder
The Reed-Solomon (RS) code used in the VSB transmission subsystem is a t = 10 (207,187)
code [2]. The RS data block size is 187 bytes, with 20 RS parity bytes added for error correction.
A total RS block size of 207 bytes is transmitted per data segment. In creating bytes from the
serial bit stream, the most significant bit (MSB) is the first serial bit. The 20 RS parity bytes are
sent at the end of the data segment. The parity-generator polynomial and the primitive-field-gen-
erator polynomial (with the fundamental supporting equations) are shown in Figure 3.3.8.
Reed-Solomon encoding/decoding is expressed as the total number of bytes in a transmitted
packet to the actual application payload bytes, where the overhead is the RS bytes used (i.e.,
207,187).
Interleaving
The interleaver employed in the VSB transmission system is a 52-data-segment (intersegment)
convolutional byte interleaver [2]. Interleaving is provided to a depth of about one-sixth of a data
field (4 ms deep). Only data bytes are interleaved. The interleaver is synchronized to the first
data byte of the data field. Intrasegment interleaving also is performed for the benefit of the trel-
lis coding process. The convolutional interleave stage is shown in Figure 3.3.9.
Trellis Coding
The 8-VSB transmission subsystem employs a 2/3-rate (R = 2/3) trellis code (with one unen-
coded bit that is precoded). Put another way, one input bit is encoded into two output bits using a
1/2-rate convolutional code while the other input bit is precoded [2]. The signaling waveform
used with the trellis code is an 8-level (3-bit), 1-dimensional constellation. Trellis code intraseg-
ment interleaving is used. This requires 12 identical trellis encoders and precoders operating on
interleaved data symbols. The code interleaving is accomplished by encoding symbols (0, 12, 24,
36 ...) as one group, symbols (1, 13, 25, 37 ...) as a second group, symbols (2, 14, 26, 38 ...) as a
third group, and so on for a total of 12 groups.
In creating serial bits from parallel bytes, the MSB is sent out first: (7, 6, 5, 4, 3, 2, 1, 0). The
MSB is precoded (7, 5, 3, 1) and the LSB (least significant bit) is feedback-convolutional
encoded (6, 4, 2, 0). Standard 4-state optimal Ungerboeck codes are used for the encoding. The
trellis code utilizes the 4-state feedback encoder shown in Figure 3.3.10. Also shown in the fig-
ure is the precoder and the symbol mapper. The trellis code and precoder intrasegment inter-
leaver, which feed the mapper shown in Figure 3.3.10, are illustrated in Figure 3.3.11. As shown
in the figure, data bytes are fed from the byte interleaver to the trellis coder and precoder; then
they are processed as whole bytes by each of the 12 encoders. Each byte produces four symbols
from a single encoder.
The output multiplexer shown in Figure 3.3.11 advances by four symbols on each segment
boundary. However, the state of the trellis encoder is not advanced. The data coming out of the
multiplexer follows normal ordering from encoders 0 through 11 for the first segment of the
Generator Polynominal G (16) = X16+X13+X12+X11+X7+X6+X3+X+1
The initalization (pre load) occurs during the field sync interval
Initalization to F180 hex (Load to 1)
X16 X15 X14 X13 X9 X8
X + X2 X3 + X4 X5 X6 + X7 + X8 X9 X10 X11 + X12 + X13 + X14 X15 X16

y
D0 D1 D2 D3 D4 D5 D6 D7
The generator is shifted with the Byte Clock and one 8 bit Byte
of data is extracted per cycle.
Figure 3.3.7 Randomizer polynomial for the DTV transmission subsystem. (From [2]. Used with permission.)
i = 2t- 1
i= 0
y
Figure 3.3.8 Reed-Solomon (207,187) t = 10 parity-generator polynomial. (From [2]. Used with permission.)
y
1
M(=4Bytes)
2
From 2M To
3
Reed-Solomon Pre-Coder and
Encoder Trellis Encoder
(B-2)M
51
(B-1)M
(B=)52
M=4, B=52, N=208, R-S Block =207, BXM=N
Figure 3.3.9 Convolutional interleaving scheme (byte shift register illustration). (From [2]. Used
with permission.)
Interference Filter
Pre-coder Trellis Encoder 8-Level Symbol Mapper
X2 Y2 Z2 MAP
+
Z2 Z1 Z0 R
0 0 0 -7
D 0 0 1 -5
0 1 0 -3 R
X1 Y1 Z1 0 1 1 -1
1 0 0 +1
Z0 1 0 1 +3
D + D 1 1 0 +5
1 1 1 +7
(D = 12 Symbols Delay)
Figure 3.3.10 An 8-VSB trellis encoder, precoder, and symbol mapper. (From [2]. Used with per-
mission.)
frame; on the second segment, the order changes, and symbols are read from encoders 4 through
11, and then 0 through 3. The third segment reads from encoder 8 through 11 and then 0 through
7. This 3-segment pattern repeats through the 312 data segments of the frame. Table 3.3.4 shows
the interleaving sequence for the first three data segments of the frame.
After the data segment sync is inserted, the ordering of the data symbols is such that symbols
from each encoder occur at a spacing of 12 symbols.
A complete conversion of parallel bytes to serial bits needs 828 bytes to produce 6624 bits.
Data symbols are created from 2 bits sent in MSB order, so a complete conversion operation
y
Trellis
Encoder &
Pre-Coder
#0
Trellis
Trellis Encoded
Encoder & and
Pre-Coder Pre-Coded
Interleaved #1
Data Data
In Trellis out to
Encoder & Mapper
Pre-Coder
... #2 ...
... ...
... ...
...
...
... ...
... Trellis
...
Encoder &
Pre-Coder
#10
Trellis
Encoder &
Pre-Coder
#11
Figure 3.3.11 Trellis code interleaver. (From [2]. Used with permission.)
Table 3.3.4 Partial Trellis Coding Interleaving Sequence (After [2].)
Segment Block 0 Block 1 ... Block 68

0 D0 D1 D2 ... D11 D0 D1 D2 ... D11 ... D0 D1 D2 ... D11
1 D4 D5 D6 ... D3 D4 D5 D6 ... D3 ... D4 D5 D6 ... D3
2 D8 D9 D10 ... D7 D8 D9 D10 ... D7 ... D8 D9 D10 ... D7
yields 3312 data symbols, which corresponds to four segments of 828 data symbols. A total of
3312 data symbols divided by 12 trellis encoders gives 276 symbols per trellis encoder, and 276
symbols divided by 4 symbols per byte gives 69 bytes per trellis encoder.
The conversion starts with the first segment of the field and proceeds with groups of four seg-
ments until the end of the field. A total of 312 segments per field divided by 4 gives 78 conver-
sion operations per field.
During segment sync, the input to four encoders is skipped, and the encoders cycle with no
input. The input is held until the next multiplex cycle, then is fed to the correct encoder.
3.3.3b Modulation
You will recall that in Figure 3.3.10 the mapping of the outputs of the trellis decoder to the nom-
inal signal levels (–7, –5, –3, –1, 1, 3, 5, 7) was shown. As detailed in Figure 3.3.12, the nominal
levels of data segment sync and data field sync are –5 and +5. The value of 1.25 is added to all
y
Data Data
Segment Data + FEC Segment
+7 SYNC SYNC
+5
+3
+1
-1
-3
-5
-7
Levels Before 4
828 Symbols 4
Pilot Addition Symbols 207 Bytes Symbols
(Pilot=1.25)
Data Segment
832 Symbols
208 Bytes
Figure 3.3.12 The 8-VSB data segment. (From [2]. Used with permission.)
R = .1 1 5 2
1 .0
.5
0
d d d = .3 1 M H z d d
5 .3 8 M H z
6 MHz
Figure 3.3.13 Nominal VSB system channel response (linear-phase raised-cosine Nyquist filter).
these nominal levels after the bit-to-symbol mapping function for the purpose of creating a small
pilot carrier [2]. The frequency of the pilot is the same as the suppressed-carrier frequency. The
in-phase pilot is 11.3 dB below the average data signal power.
The VSB modulator receives the 10.76 Msymbols/s, 8-level trellis-encoded composite data
signal with pilot and sync added. The DTV system performance is based on a linear-phase
raised-cosine Nyquist filter response in the concatenated transmitter and receiver, as illustrated
in Figure 3.3.13. The system filter response is essentially flat across the entire band, except for
the transition regions at each end of the band. Nominally, the rolloff in the transmitter has the
response of a linear-phase root raised-cosine filter.
3.3.3c High Data-Rate Mode

The high data-rate mode trades off transmission robustness (28.3 dB S/N threshold) for payload
data rate (38.57 Mbits/s) [2]. Most parts of the high data-rate mode VSB system are identical, or
at least similar, to the terrestrial system. A pilot, data segment sync, and data field sync all are
used to provide robust operation. The pilot in the high data-rate mode is 11.3 dB below the data
y
Data Data
Segment Data + FEC Segment
SYNC SYNC
+15
+13
+11
+9
+7
+5
+3
+1
-1
-3
-5
-7
-9
-11
-13
-15
4
828 4
Levels Before
Pilot Addition Symbols Symbols Symbols
(Pilot=2.5)
Data Segment
832 Symbols
Figure 3.3.14 Typical data segment for the 16-VSB mode. (From [2]. Used with permission.)
signal power; and the symbol, segment, and field signals and rates all are the same as well,
allowing either receiver type to lock up on the other’s transmitted signal. Also, the data frame
definitions are identical. The primary difference is the number of transmitted levels (8 vs.16) and
the use of trellis coding and NTSC interference-rejection filtering in the terrestrial system.
The RF spectrum of the high data-rate modem transmitter looks identical to the terrestrial
system. Figure 3.3.14 illustrates a typical data segment, where the number of data levels is seen
to be 16 as a result of the doubled data rate. Each portion of 828 data symbols represents 187
data bytes and 20 Reed-Solomon bytes, followed by a second group of 187 data bytes and 20
Reed-Solomon bytes (before convolutional interleaving).
Figure 3.3.15 shows a functional block diagram of the high data-rate transmitter. It is identical
to the terrestrial VSB system, except that the trellis coding is replaced with a mapper that con-
verts data to multilevel symbols. The interleaver is a 26-data-segment intersegment convolu-
tional byte interleaver. Interleaving is provided to a depth of about one-twelfth of a data field (2
ms deep). Only data bytes are interleaved. Figure 3.3.16 shows the mapping of the outputs of the
interleaver to the nominal signal levels (–15, –13, –11, ..., 11, 13, 15). As shown in Figure 3.3.14,
the nominal levels of data segment sync and data field sync are –9 and +9. The value of 2.5 is
added to all these nominal levels after the bit-to-symbol mapping for the purpose of creating a
small pilot carrier. The frequency of the in-phase pilot is the same as the suppressed-carrier fre-
quency. The modulation method of the high data-rate mode is identical to the terrestrial mode
except that the number of transmitted levels is 16 instead of 8.
3.3.3d Bit Rate Delivered to a Transport Decoder

As noted previously, the symbol rate of the transmission subsystem is given by
y
Reed- Data RF
Data Pilot VSB
Solomon Inter- Mapper MUX Up-
Randomizer Insertion Modulator
Encoder leaver Converter
Segment Sync
Field Sync
Figure 3.3.15 Functional block diagram of the 16-VSB transmitter. (From [2]. Used with permis-
sion.)
Xa Xb Xc Xd O
Byte to 1 1 1 1 +15
Symbol 1 1 1 0 +13
Conversion
1 1 0 1 +11
MSB 1 1 0 0 +9
7 Xa1 Xa
6 Xb1 1 0 1 1 +7
1st Nibble Xb 1 0 1 0 +5
5 Xc1
4 Xd1 1 0 0 1 +3
From 1 0 0 0 +1
3 Xa2 Xc To
Byte 0 1 1 1 -1
Interleaver 2 Xb2 MUX
2nd Nibble 0 1 1 0 -3
1 Xc2 Xd
0 1 0 1 -5
0 Xd2
LSB
0 1 0 0 -7
0 0 1 1 -9
0 0 1 0 -11
0 0 0 1 -13
0 0 0 0 -15
Figure 3.3.16 A 16-VSB mapping table. (From [2]. Used with permission.)
4.5
-------- × 684 = 10.7 ... million symbols/second (megabaud) (3.3.7)
286
The symbol rate must be locked in frequency to the transport rate.

The numbers in the formula for the ATSC symbol rate in 6 MHz systems are related to NTSC
scanning and color frequencies. Because of this relationship, the symbol clock can be used as a
basis for generating an NTSC color subcarrier for analog output from a set top box. The repeti-
tion rates of data segments and data frames are deliberately chosen not to have an integer rela-
tionship to NTSC or PAL scanning rates, to insure that there will be no discernible pattern in co-
channel interference.
The particular numbers used are:
y
• 4.5 MHz = the center frequency of the audio carrier offset in NTSC. This number was tradi-
tionally used in NTSC literature to derive the color subcarrier frequency and scanning rates.
In modern equipment, this may start with a precision 10 MHz reference, which is then multi-
plied by 9/20.
• 4.5 MHz/286 = the horizontal scan rate of NTSC, 15734.2657+...Hz (note that the color sub-
carrier is 455/2 times this, or 3579545 +5/11 Hz).
• 684: this multiplier gives a symbol rate for an efficient use of bandwidth in 6 MHz. It requires
a filter with Nyquist roll-off that is a fairly sharp cutoff (11 percent excess bandwidth), which
is still realizable with a reasonable surface acoustic wave (SAW) filter or digital filter.
In the terrestrial broadcast mode, channel symbols carry three bits/symbol of trellis-coded
data. The trellis code rate is 2/3, providing 2 bits/symbol of gross payload. Therefore the gross
payload is
10.76 …× 2 = 21.52 ... Mbps (3.3.8)
To find the net payload delivered to a decoder it is necessary to adjust Equation (3.3.8) for the
overhead of the data segment sync, data field sync, and Reed-Solomon FEC.
To get the net bit rate for an MPEG-2 stream carried by the system (and supplied to an MPEG
transport decoder), it is first noted that the MPEG sync bytes are removed from the data stream
input to the 8-VSB transmitter and replaced with segment sync, and later reconstituted at the
receiver. For throughput of MPEG packets (the only allowed transport mechanism) segment sync
is simply equivalent to transmitting the MPEG sync byte, and does not reduce the net data rate.
The net bit rate of an MPEG-2 stream carried by the system and delivered to the transport
decoder is accordingly reduced by the data field sync (one segment of every 313) and the Reed-
Solomon coding (20 bytes of every 208)
312 188
21.52 …× -------- × -------- = 19.39 ... Mbps (3.3.9)
313 208
The net bit rate supplied to the transport decoder for the high data rate mode is
19.39 …× 2 = 38.78 ... Mbps (3.3.10)
3.3.3e Performance Characteristics of Terrestrial Broadcast Mode

The terrestrial 8-VSB system can operate in a signal-to-additive-white-Gaussian-noise (S/N)
environment of 14.9 dB [1]. The 8-VSB segment error probability curve including 4-state trellis
decoding and (207,187) Reed-Solomon decoding in Figure 3.3.17 shows a segment error proba-
bility of 1.93 × 10–4. This is equivalent to 2.5 segment errors/second, which was established by
measurement as the threshold of visibility (TOV) of errors in the prototype equipment. Particular
product designs may achieve somewhat better performance for subjective TOV by means of error
masking.
The cumulative distribution function (CDF) of the peak-to-average power ratio, as measured
on a low power transmitted signal with no non-linearities, is plotted in Figure 3.3.18. The plot
y
1E+01
1E+00
Note: Threshold of visibility
Probability of Error 1E-01 (TOV) has been measured to
occur at a S/N of 14.9dB when
there are 2.5 segment errors per
1E-02 second which is a segment
error rate (SER) of 1.93x10 -4.
1E-03
SER at TOV
1E-04
1E-05
1E-06
1E-07
1E-08
8 9 10 11 12 13 14 15 16 17 18
S/N db (RMS)
Figure 3.3.17 Segment-error probability for 8-VSB with 4-state trellis coding; RS (207,187). (From
shows that 99.9 percent of the time the transient peak power is within 6.3 dB of the average
power.
Transmitter Signal Processing

A pre-equalizer filter is recommended for use in over-the-air broadcasts where the high power
transmitter may have significant in-band ripple or significant roll off at band edges [1]. Pre-
equalization is typically required in order to compensate the high-order filter used to meet a
stringent out-of-band emission mask, such as the U.S. FCC required mask [4]. This linear distor-
tion can be measured by an equalizer in a reference demodulator (“ideal” receiver) employed at
the transmitter site. A directional coupler, which is recommended to be located at the sending
end of the antenna feed transmission line, supplies the reference demodulator a small sample of
the antenna signal feed. The equalizer tap weights of the reference demodulator are transferred to
the transmitter pre-equalizer for pre-correction of transmitter linear distortion. This is a one-time
procedure of measurement and transmitter pre-equalizer adjustment. Alternatively, the transmit-
ter pre-equalizer can be made continuously adaptive. In this arrangement, the reference demodu-
lator is provided with a fixed-coefficient equalizer compensating for its own deficiencies in ideal
response.
A pre-equalizer suitable for many applications is an 80 tap, feed-forward transversal filter.
The taps are symbol-spaced (93 ns) with the main tap being approximately at the center, giving
approximately ± 3.7 μs correction range. The pre-equalizer operates on the I channel data signal
(there is no Q channel data signal in the transmitter), and shapes the frequency spectrum of the
IF signal so that there is a flat in-band spectrum at the output of the high power transmitter that
feeds the antenna for transmission. There is no effect on the out-of-band spectrum of the trans-
mitted signal. If desired, complex equalizers or fractional equalizers (with closer-spaced taps)
y
Above Average Power (%)

% Of Time Peak Power is
Peak-To-Average Power Ratio (dB)
Figure 3.3.18 The cumulative distribution function of 8-VSB peak-to-average power ratio in an
ideal linear system. (From [1]. Used with permission.)
can provide independent control of the outer portions of the spectrum (beyond the Nyquist
slopes).
The transmitter vestigial sideband filtering is sometimes implemented by sideband cancella-
tion, using the phasing method. In this method, the baseband data signal is supplied to digital fil-
tering that generates in-phase and quadrature-phase digital modulation signals for application to
respective D/A converters. This filtering process provides the root raised cosine Nyquist filtering
and provides compensation for the (sin x) / x frequency responses of the D/A converters, as well.
The baseband signals are converted to analog form. The in-phase signal modulates the amplitude
of the IF carrier at zero degrees phase, while the quadrature signal modulates a 90-degree shifted
version of the carrier. The amplitude-modulated quadrature IF carriers are added to create the
vestigial sideband IF signal, canceling the unwanted sideband and increasing the desired side-
band by 6 dB. The nominal frequency of the IF carrier (and small in-phase pilot) in the prototype
hardware was 46.69 MHz, which is equal to the IF center frequency (44.000 MHz) plus the sym-
bol rate divided by 4
10.762
---------------- = 2.6905 MHz (3.3.11)
4
Additional adjacent-channel suppression (beyond that achieved by sideband cancellation)

may be performed by a linear phase, flat amplitude response SAW filter. Other implementations
for VSB filtering are possible that may include the prefilter of the previous section.
y
3.3.3f Upconverter and RF Carrier Frequency Offsets

Modern analog TV transmitters use a two-step modulation process [1]. The first step usually is
modulation of the data onto an IF carrier, which is the same frequency for all channels, followed
by translation to the desired RF channel. The digital 8-VSB transmitter applies this same two-
step modulation process. The RF upconverter translates the filtered flat IF data signal spectrum
to the desired RF channel. For the same approximate coverage as an analog transmitter (at the
same frequency), the average power of the DTV signal is on the order of 12 dB less than the ana-
log peak sync power (when operating on the same frequency).
The nominal frequency of the RF upconverter oscillator in DTV terrestrial broadcasts will
typically be the same as that used for analog transmitters, (except for offsets required in particu-
lar situations).
Note that all examples in this section relate to a 6 MHz DTV system. Values may be modified
easily for other channel widths.
Nominal DTV Pilot Carrier Frequency

The nominal DTV pilot carrier frequency is determined by fitting the DTV spectrum symmetri-
cally into the RF channel [1]. This is obtained by taking the bandwidth of the DTV signal—
5,381.1189 kHz (the Nyquist frequency difference or one-half the symbol clock frequency of
10,762.2378 kHz)—and centering it in the 6 MHz TV channel. Subtracting 5,381.1189 kHz
from 6,000 kHz leaves 618.881119 kHz. Half of that is 309.440559 kHz, precisely the standard
pilot offset above the lower channel edge. For example, on channel 45 (656–662 MHz), the nom-
inal pilot frequency is 656.309440559 MHz.
Requirements for Offsets

There are two categories of requirements for pilot frequency offsets [1]:
• Offsets to protect lower adjacent channel analog broadcasts, mandated by FCC rules in the
U.S., and which override other offset considerations.
• Recommended offsets for other considerations such as co-channel interference between DTV
stations or between DTV and analog stations.
Upper DTV Channel into Lower Analog Channel

This is the overriding case mandated by the FCC rules in the U.S.—precision offset with a lower
adjacent analog station, full service or Low Power Television (LPTV) [1]. The FCC Rules, Sec-
tion 73.622(g)(1), states that: “DTV stations operating on a channel allotment designated with a
“c” in paragraph (b) of this section must maintain the pilot carrier frequency of the DTV signal
5.082138 MHz above the visual carrier frequency of any analog TV broadcast station that oper-
ates on the lower adjacent channel and is located within 88 kilometers. This frequency difference
must be maintained within a tolerance of ± 3 Hz.”
This precise offset is necessary to reduce the color beat and high-frequency luminance beat
created by the DTV pilot carrier in some receivers tuned to the lower adjacent analog channel.
The tight tolerance assures that the beat will be visually cancelled, since it will be out of phase on
successive video frames.
y
Note that the frequency is expressed with respect to the lower adjacent analog video carrier,
rather than the nominal channel edge. This is because the beat frequency depends on this rela-
tionship, and therefore the DTV pilot frequency must track any offsets in the analog video carrier
frequency. The offset in the FCC rules is related to the particular horizontal scanning rate of
NTSC, and can easily be modified for PAL. The offset Of was obtained from
F F
Of = 455 × ⎛⎝ ----h--⎞⎠ + 191 × ⎛⎝ ----h- ⎞⎠ – 29.97 = 5082138 Hz (3.3.12)
2 2
Where Fh = NTSC horizontal scanning frequency = 15,734.264 Hz.

The equation indicates that the offset with respect to the lower adjacent chroma is an odd mul-
tiple (191) of one-half the line rate to eliminate the color beat. However, this choice leaves the
possibility of a luma beat. The offset is additionally adjusted by one-half the analog field rate to
eliminate the luma beat. While satisfying the exact adjacent channel criteria, this offset is also as
close as possible to optimal comb filtering of the analog co-channel in the digital receiver. Note
additionally that offsets are to higher frequencies rather than lower, to avoid any possibility of
encroaching on the lower adjacent sound. (It also reduces the likelihood of the automatic fine
tuning (AFT) in the analog receiver experiencing lock-out because the signal energy including
the pilot is moved further from the analog receiver bandpass.
As an example, if a channel 44 NTSC station is operating with a zero offset, the Channel 45
DTV pilot carrier frequency must be 651.250000 MHz plus 5.082138 MHz or 656.332138 MHz;
that is, 22.697 kHz above the nominal frequency. If the lower adjacent NTSC channel is offset
±10 kHz, the DTV frequency will have to be adjusted accordingly.
Co-Channel DTV into Analog

In co-channel cases, DTV interference into analog TV appears noise-like [1]. The pilot carrier is
low on the Nyquist slope of the IF filter in the analog receiver, so no discernable beat is gener-
ated. In this case, offsets to protect the analog channel are not required. Offsets are useful, how-
ever to reduce co-channel interference from analog TV into DTV. The performance of the analog
rejection filter and clock recovery in the DTV receiver will be improved if the DTV carrier is
911.944 kHz below the NTSC visual carrier. In other words, in the case of a 6 MHz NTSC sys-
tem, if the analog TV station is not offset, the DTV pilot carrier frequency will be 338.0556 kHz
above the lower channel edge instead of the nominal 309.44056 kHz. As before, if the NTSC sta-
tion is operating with a ±10 kHz offset, the DTV frequency will have to be adjusted in the same
direction. The formula for calculating this offset is
F pilot = F vis ( n ) – 70.5 × F seg = 338.0556 Hz (for no NTSC analog offset) (3.3.13)
Where:
F pilot = DTV pilot frequency above lower channel edge
Fvis(n) = NTSC visual carrier frequency above lower channel edge
= 1,250 kHz for no NTSC offset (as shown)
= 1,240 kHz for minus offset
= 1,260 kHz for plus offset
Fseg = ATSC data segment rate; = symbol clock frequency / 832= 12,935.381971 Hz
y
1E+01
1E+00
Probability of Error
1E-01
16-VSB
1E-02
Symbol Error Rate
1E-03
1E-04 Segment Error
Rate After Reed-
1E-05 Solomon FEC
1E-06
1E-07
1E-08
8 12 16 20 24 28 32 36
S/N db (RMS)
Figure 3.3.19 The error probability of the 16-VSB signal. (From [1]. Used with permission.)
The factor of 70.5 is chosen to provide the best overall comb filtering of analog color TV co-
channel interference. The use of a value equal to an integer +0.5 results in co-channel analog TV
interference being out-of-phase on successive data segment syncs.
Note that in this case the frequency tolerance is plus or minus one kHz. More precision is not
required. Also note that a different data segment rate would be used for calculating offsets for 7
or 8 MHz systems.
Co-channel DTV into DTV

If two DTV stations share the same channel, interference between the two stations can be
reduced if the pilot is offset by one and a half times the data segment rate [1]. This ensures that
the frame and segment syncs of the interfering signal will each alternate polarity and be averaged
out in the receiver tuned to the desired signal.
The formula for this offset is
F offset = 1.5 × F seg = 19.4031 kHz (3.3.14)
Where:
Foffset = offset to be added to one of the two DTV carriers
F seg = 12,935.381971 Hz (as defined previously)
This results in a pilot carrier 328.84363 kHz above the lower band edge, provided neither
DTV station has any other offset.
Use of the factor 1.5 results in the best co-channel rejection, as determined experimentally
with prototype equipment. The use of an integer +0.5 results in co-channel interference alternat-
ing phase on successive segment syncs.
y
Above Average Power (%)

% Of Time Peak Power is
Peak-To-Average Power Ratio (dB)
Figure 3.3.20 Cumulative distribution function of the 16-VSB peak-to-average power ratio. (From
3.3.3g Performance Characteristics of High Data Rate Mode

The high data rate mode can operate in a signal-to-white-noise environment of 28.3 dB. The
error probability curve is shown in Figure 3.3.19 [1].
The cumulative distribution function (CDF) of the peak-to-average power ratio, as measured
on a low power transmitted signal with no non-linearities, is plotted in Figure 3.3.20 and is
slightly higher than that of the terrestrial mode.
3.3.4 References
1. ATSC, “Guide to the Use of the Digital Television Standard,” Advanced Television Systems
Committee, Washington, D.C., Doc. A/54A, December 4, 2003.
2. ATSC, “ATSC Digital Television Standard,” Advanced Television Systems Committee,
Washington, D.C., Doc. A/53C, May 21, 2004.
3. ITU-R Document TG11/3-2, “Outline of Work for Task Group 11/3, Digital Terrestrial
Television Broadcasting,” June 30, 1992.
4. Chairman, ITU-R Task Group 11/3, “Report of the Second Meeting of ITU-R Task Group
11/3, Geneva, Oct. 13-19, 1993,” p. 40, Jan. 5, 1994.
FCC: “Memorandum Opinion and Order on Reconsideration of the Sixth Report and
Order,” Federal Communications Commission, Washington, D.C., February 17, 1998.
y
3.3.5 Bibliography
Rhodes, Charles W.: “Terrestrial High-Definition Television,” The Electronics Handbook, Jerry
y
g g
Chapter
3.4
Longley-Rice Propagation Model
3.4.1 Introduction1
The Longley-Rice radio propagation model is used to make predictions of radio field strength at
specific geographic points based on the elevation profile of terrain between the transmitter and
each specific reception point [1]. A computer is needed to make these predictions because of the
large number of reception points that must be individually examined. Computer code for the
Longley-Rice point-to-point radio propagation model is published in [2].
This chapter describes the process used by the U.S. Federal Communications Commission
(FCC) in determining the digital television (DTV) channel plan using the Longley-Rice propaga-
tion model.
3.4.2 Evaluation of Service

Under the FCC's rules, computation of service area or coverage using the Longley-Rice method-
ology is limited to the areas within certain specific geographic contours [1].
For analog television service, computations are made inside the conventional Grade B con-
tour defined in Section 73.683 of the FCC rules, with the exception that the defining field for
UHF channels is modified by a dipole factor equal to 20 log [615/(channel mid-frequency)].
Thus, the area subject to calculation for analog TV consists of the geographic points at which the
field strength predicted for 50 percent of locations and 50 percent of time by FCC curves is at
least as great as the values given in Table 3.4.1. The relevant curves for predicting these fields
are the F(50, 50) curves found in Section 73.699 of FCC rules.
For digital television stations, service is evaluated inside contours determined by DTV plan-
ning factors in combination with field strength curves derived for 50 percent of locations and 90
percent of the time from curves which are also found in Section 73.699 of FCC rules. The family
1. This chapter is based on: FCC: “OET Bulletin No. 69—Longley-Rice Methodology for Eval-
uating TV Coverage and Interference,” Federal Communications Commission, Washington,
D.C., July 2, 1997.
3-67
g y p g
Table 3.4.1 Field Strengths Defining the Area Subject to Calculation for Analog Stations (After [1].)
Channels Defining Field Strength, dBu, to be Predicted Using F(50, 50) Curves
2–6 47
7 – 13 56
14 – 69 64 – 20 log[615 / (channel mid-frequency)]
Table 3.4.2 Field Strengths Defining the Area Subject to Calculation for DTV Stations ( After [1].)
Channels Defining Field Strength, dBu, to be Predicted Using F(50, 90) Curves
2–6 28
7 – 13 36
14 – 69 41– 20 log[615 / (channel mid-frequency)]
of FCC propagation curves for predicting field strength at 50 percent of locations 90 percent of
the time is found from
F(50, 90) = F(50, 50) – [F(50, 10) – F(50, 50)] (3.4.1)
That is, the F(50, 90) value is lower than F(50, 50) by the same amount that F(50, 10) exceeds
F(50, 50).
The defining field strengths for DTV service, contained in Section 73.622 of the FCC rules,
are shown in Table 3.4.2. These values are determined from the DTV planning factors identified
in Table 3.4.3. They are used first to determine the area subject to calculation using FCC curves,
and subsequently to determine whether service is present at particular points within this area
using Longley-Rice terrain-dependent prediction.
For digital TV, three different situations arise:
• For DTV stations of the initial allotment plan located at the initial reference coordinates, the
area subject to calculation extends in each direction to the distance at which the field strength
predicted by FCC curves falls to the value identified in Table 3.4.2. The bounding contour is
identical, in most cases, to that of the analog station with which the initial allotment is paired.
The initial allotment plan and reference coordinates are set forth in [3].
• For new DTV stations, the area subject to calculation extends from the transmitter site to the
distance at which the field strength predicted by FCC curves falls to the value identified in
Table 3.4.2.
• In the case where a DTV station of the initial allotment plan has moved, the area subject to
calculation is the combination (logical union) of the area determined for the initial allotment
and the area inside the contour which would apply in the case of a new DTV station.
3.4.2a Planning Factors

The planning factors shown in Table 3.4.3 lead to the values of field strength given in Table 3.4.2
to define the area subject to calculation for DTV stations [1]. These planning factors are
assumed to characterize the equipment, including antenna systems, used for home reception.
g y p g
Table 3.4.3 Planning Factors for DTV Reception (After [1].)

Planning Factor Symbol Low VHF High VHF UHF
Geometric mean frequency (Mhz) F 69 194 615
Dipole factor (dBm-dBu) Kd –111.8 –120.8 –130.8
Dipole factor adjustment Ka none none see text
Thermal noise (dBm) Nt –106.2 –106.2 –106.2
Antenna gain (dB) G 4 6 10
Downlead line loss (dB) L 1 2 4
System noise figure (dB) Ns 10 10 7
Required carrier/noise ratio (dB) C/ N 15 15 15
They determine the minimum field strength for DTV reception as a function of frequency band
and as a function of channel number in the UHF band.
The adjustment, Ka = 20 log[615/(channel mid-frequency)], is added to Kd to account for the
fact that field strength requirements are greater for UHF channels above the geometric mean fre-
quency of the UHF band and smaller for UHF channels below that frequency. The geometric
mean frequency, 615 MHz, is approximately the mid-frequency of channel 38.
The modified Grade B contour of analog UHF stations is determined by applying this same
adjustment factor to the Grade B field strength given in 47 CFR §73.683. With this dipole factor
modification, the field strength defining the Grade B of UHF channels becomes
64 – 20 log[615/(channel mid-frequency)] dBu
in place of simply 64. Thus, the modified Grade B contour for channel 14 is determined by a
median field strength of 61.7 dBu, and the value for channel 51 is 66.3 dBu. The modified values
have been presented in Table 3.4.1. This modified Grade B contour bounds the area subject to
Longley-Rice calculations for analog stations.
The values appearing in Table 3.4.2 follow from the planning factors. They were derived from
Table 3.4.3 by solving the equation
Field + Kd + Ka + G – L – Nt – Ns = C/N (3.4.2)
For a new DTV station with a particular authorized set of facilities, the values given in Table
3.4.2 determine the contour within which the FCC makes all subsequent calculations of service
and interference.
3.4.2b Reference Value of ERP for DTV Operation

The initial allotment plan established a reference value for the effective radiated power (ERP) of
DTV stations [1]. This ERP is the maximum of the values needed to match the service contour of
the paired analog station in each direction supposing that the new station operates at the same
location with the same antenna height. The reference ERP was calculated using the following
methodology.
The distance to the existing analog grade B contour is determined in each of 360 uniformly
spaced compass directions starting from true north using linear interpolation of available data as
g y p g
necessary. This determination is made using information in the FCC Engineering Data Base of
April 3, 1997, including directional antenna data, and from terrain elevation data at points sepa-
rated by 3 arc-seconds of longitude and latitude. FCC curves (Section 73.699 of FCC rules) are
applied in the usual way, as described in Section 73.684 of the rules, to find this grade B contour
distance, with the exception that dipole factor considerations are applied to the field strength
contour for UHF.
Height above average terrain is determined every 45 degrees from terrain elevation data in
combination with the height of the transmitter radiation center above mean sea level, and by lin-
ear interpolation for compass directions in between. In cases where the TV Engineering Data
Base indicates that a directional antenna is employed, the ERP in each specific direction is deter-
mined through linear interpolation of the relative field values describing the directional pattern.
The directional pattern stored in the FCC Directional Antenna Data Base provides relative field
values at 10 degree intervals and may include additional values in special directions. The result
of linear interpolation of these relative field values is squared and multiplied by the overall max-
imum ERP listed for the station in the TV Engineering Data Base to find the ERP in a specific
direction.
The corresponding values of ERP for DTV signals in each direction is then calculated by a
further application of FCC curves, with noise-limited DTV coverage defined as the presence of
the field strengths identified in Table 3.4.2 at 50 percent of locations and 90 percent of the time.
These ERP values are computed for all 360 azimuths using the same radial-specific height above
average terrain as for the analog TV case, but now in conjunction with F(50, 90) curves.
Finally, the ERP for DTV is modified so that it does not exceed 1 megawatt and is not less
than 50 kilowatts. This is done by scaling the azimuthal power, pattern rather than by truncation.
Thus, if replication by FCC curves as described above required an ERP of 2 megawatts, the
power pattern is reduced by a factor of 2 in all directions. The resulting ERP is the reference
value cited in Section 73.622 of the rules.
3.4.2c DTV Transmitting Antenna Patterns

In general, these computations of DTV power to match the distance to the Grade B contour of an
analog station result in ERP values which vary with azimuth [1]. For example, the azimuthal
ERP pattern which replicates in UHF the Grade B contour of an omnidirectional VHF operation
will be somewhat distorted because terrain has a different effect on propagation in the two bands.
In addition, the 90 percent time variability allowance for DTV has an effect on the DTV pattern.
Thus, the procedure described here effectively derives a new directional antenna pattern wher-
ever necessary for a precise match according to FCC curves.
These DTV azimuthal patterns may be calculated using the procedure outlined above. In
addition, they are available from the FCC's Internet site, www.fcc.gov, to supplement the infor-
mation contained in Appendix B of the Sixth Report and Order [3]. The format for describing
DTV transmitting antenna patterns is identical to that of the FCC Directional Antenna Data Base
for analog stations. Relative field values are given at intervals of 10 degrees, and supplemental
values are given at special azimuths. For DTV patterns, special azimuths are included where the
pattern factor is unity, while both bracketing factors at 10-degree azimuths are less than unity.
g y p g
Table 3.4.4 Parameter Values Used in FCC Implementation of the Longley-Rice Fortran Code
(After [1].)
Parameter Value Meaning/Comment
EPS 15.0 Relative permittivity of ground.
SGM 0.005 Ground conductivity, Siemens per meter.
ZSYS 0.0 Coordinated with setting of EN0. See page 72 of NTIA Report.
EN0 301.0 Surface refractivity in N-units (parts per million).
IPOL 0 Denotes horizontal polarization.
MDVAR 3 Code 3 sets broadcast mode of variability calculations.
KLIM 5 Climate code 5 for continental temperate.
HG(1) see text Height of the radiation center above ground.
HG(2) 10 m Height of TV receiving antenna above ground.
3.4.2d Application of the Longley-Rice Methodology

The area subject to calculation is divided into rectangular cells, and the Longley-Rice point-to-
point propagation model Version 1.2.2 is applied to a point in each cell to determine whether the
predicted field strength is above the value found in Table 3.4.1 or Table 3.4.2, as appropriate [1].
The values identified in those tables are considered to be thresholds for reception in the absence
of interference. For cells with population, the point chosen by the FCC computer program is the
population centroid; otherwise it is the geometric center; and the point so determined represents
the cell in all subsequent service and interference calculations. The station's directional transmit-
ting antenna pattern, if any, is taken into account in determining the ERP in the direction of each
cell. Cells 2 kilometers on a side were used to produce the service and interference data appear-
ing in Appendix B of the Sixth Report and Order.
Parameter values set in the Longley-Rice Fortran code as implemented by the FCC are given
in Table 3.4.4. In addition to these parameters, execution of the code requires a specification of
the percent of time and locations at which the predicted fields will be realized or exceeded, and a
third percentage identifying the degree of confidence desired in the results. To predict TV ser-
vice at cells of the area subject to calculation, the FCC set the location variability at 50 percent
and the time variability at 90 percent. The percent confidence is set at 50 percent, indicating that
we are interested in median situations.
HG(1) in Table 3.4.4 is the height of the radiation center above ground. It is determined by
subtracting the ground elevation above mean sea level (AMSL) at the transmitter location from
the height of the radiation center AMSL. The latter is found in the TV Engineering Data Base
while the former is retrieved from the terrain elevation data base as a function of the transmitter
site coordinates also found in the TV Engineering Data Base.
Finally, terrain elevation data at uniformly spaced points between the between transmitter and
receiver must be provided. The FCC computer program is linked to a terrain elevation data base
with values every 3 arc-seconds of latitude and longitude. The program retrieves elevations from
this data base at regular intervals with a spacing increment which is chosen at the time the pro-
gram is compiled. The computer runs which evaluated service and interference for the Sixth
Report and Order used a spacing increment of 1 kilometer. The elevation of a point of interest is
determined by linear interpolation of the values retrieved for the corners of the coordinate rect-
angle in which the point of interest lies.
Evaluations of service coverage and interference using finer spacing increments were
expected to be consistent with those using 1 kilometer. Evaluations using cells smaller than 2 km
g y p g
on a side were also expected to be consistent with the evaluations given in Appendix B of the
Sixth Report and Order.
3.4.3 Evaluation of Interference

The presence or absence of interference in each grid cell of the area subject to calculation is
determined by further application of Longley-Rice. Radio paths between undesired TV transmit-
ters and the point representing each cell are examined [1]. The undesired transmitters included in
the analysis of each cell are those which are possible sources of interference at that cell, consid-
ering their distance from the cell and channel offset relationships. For each such radio path, the
Longley-Rice procedure is applied for median situations (that is, confidence 50 percent), and for
50 percent of locations, 10 percent of the time.
The interference analysis examines only those cells that have already been determined to have
a desired field strength above the threshold for reception given in Table 3.4.1 for analog stations
and Table 3.4.2 for DTV stations. A cell being examined is counted as having interference if the
ratio of the desired field strength to that of any one of the possible interference sources is less
than a certain critical minimum value. The comparison is made after applying the discrimination
effect of the receiving antenna. The critical value is a function of the channel offset relationship.
Cells of the area subject to calculation for an analog station are examined first as to whether
the desired signal is above the threshold for reception, second with regard to whether there is
interference from another analog station, and finally as to whether there is interference from
DTV stations. Thus, a DTV station does not cause interference to analog stations in places where
there is no service because of a weak desired signal, or in places where interference from other
analogue stations already exists.
3.4.3a D/U Ratios

Criteria for the ratio of desired to undesired field strength are specified in Section 73.623 of FCC
rules for interference involving DTV stations as desired or undesired [1]. These criteria are sum-
marized in Tables 3.4.5a and 3.4.5b. The tables also include the criteria for interference between
analog stations used in preparing the service and interference evaluation in Appendix B of the
Sixth Report and Order.
The evaluation of service and interference in Appendix B of the Sixth Report and Order con-
sidered taboo channel relationships for interference into DTV. However, the D/U ratios (approxi-
mately –60 dB) were such that they rarely if ever had an effect on the results, and the FCC rules
adopted in the Sixth Report and Order do not require attention to UHF taboo interference to
DTV stations.
3.4.3b Receiving Antenna Pattern

The receiving antenna is assumed to have a directional gain pattern which tends to discriminate
against off-axis undesired stations [1]. This pattern is a planning factor affecting interference.
The specific form of this pattern was chosen by a working group of the FCC Advisory Commit-
tee for Advanced Television. It is built into the service and interference computer program devel-
oped by the Broadcasters' Caucus and also used in the FCC program.
g y p g
Table 3.4.5a Interference Criteria for Co- and Adjacent Channels (After [1].)
Channel Offset D/U Ratio, dB
Analog into DTV into Analog Analog into DTV DTV into DTV
Analog
–1 (lower adjacent) –3 –17 –48 –42
0 (co-channel) +28 +34 +2 +15
+1 (upper adjacent) –13 –12 –49 –43
Table 3.4.5b Interference Criteria for UHF Taboo Channels (After [1].)
Channel Offset Relative to D/U Ratio, dB
Desired Channel N
Analog into DTV into Analog into DTV into DTV
Analog Analog DTV
N–8 –32 –32 NC NC
N–7 –30 –35 NC NC
N–4 NC –34 NC NC
N-3 –33 –30 NC NC
N–2 –26 –24 NC NC
N+2 –29 –28 NC NC
N+3 –34 –34 NC NC
N+4 –23 –25 NC NC
N+7 –33 –34 NC NC
N+8 –41 –43 NC NC
N+14 –25 –33 NC NC
N+15 –9 –31 NC NC
(NC means not considered)
Table 3.4.6 Front-to-Back Ratios Assumed for Receiving Antennas (After [1].)
TV Service Front-to-Back Ratios, dB
Low VHF High VHF UHF
Analog 6 6 6
DTV 10 12 14
The discrimination, in relative volts, provided by the assumed receiving pattern is a fourth-
power cosine function of the angle between the lines joining the desired and undesired stations to
the reception point. One of these lines goes directly to the desired station, the other goes to the
undesired station. The discrimination is calculated as the fourth power of the cosine of the angle
between these lines but never more than represented by the front-to-back ratios identified in
Table 3.4.6. When both desired and undesired stations are dead ahead, the angle is 0.0 giving a
cosine of unity so that there is no discrimination. When the undesired station is somewhat off-
axis, the cosine will be less than unity bringing discrimination into play; and when the undesired
station is far off axis, the maximum discrimination given by the front-to-back ratio is attained.
g y p g
3.4.4 References
1. FCC: “OET Bulletin No. 69—Longley-Rice Methodology for Evaluating TV Coverage and
Interference,” Federal Communications Commission, Washington, D.C., July 2, 1997
2. Hufford, G. A., A. G. Longley, and W. A. Kissick: A Guide to the Use of the ITS Irregular
Terrain Model in the Area Prediction Mode, U.S. Department of Commerce, Washington,
D.C., NTIA Report 82-100, April 1982. (Note: some modifications to the code were
described by G. A. Hufford in a memorandum to users of the model dated January 30,
1985. With these modifications, the code is referred to as Version 1.2.2 of the Longley-
Rice model.)
3. FCC: “Appendix B, Sixth Report and Order,” MM Docket 87-268, FCC 97-115, Federal
Communications Commission, Washington, D.C., April 3, 1997.
g g
Chapter
3.5
Television Transmitters
John T. Wilner
3.5.1 Introduction
Any analog television transmitter consists of two basic subsystems:
• The visual section, which accepts the video input, amplitude-modulates an RF carrier, and
amplifies the signal to feed the antenna system
• The aural section, which accepts the audio input, frequency-modulates a separate RF carrier,
and amplifies the signal to feed the antenna system
The visual and aural signals usually are combined to feed a single radiating antenna. Different
transmitter manufacturers have different philosophies with regard to the design and construction
of a transmitter. Some generalizations are possible, however, with respect to basic system config-
urations. Transmitters can be divided into categories based on the following criteria:
• Output power
• Final-stage design
• Modulation system
3.5.1a System Considerations

When the power output of a television transmitter is discussed, the visual section is the primary
consideration. (See Figure 3.5.1.) Output power refers to the peak power of the visual stage of the
transmitter (peak-of-sync). The licensed ERP is equal to the transmitter power output times feed-
line efficiency times the power gain of the antenna.
A VHF station can achieve its maximum power output through a wide range of transmitter
and antenna combinations. Reasonable pairings for a high-band NTSC VHF station in the U.S.
range from a transmitter with a power output of 50 kW feeding an antenna with a gain of 8, to a
30 kW transmitter connected to a gain-of-12 antenna. These combinations assume reasonably
3-75
Figure 3.5.1 Idealized picture transmission amplitude characteristics for VHF and UHF analog
systems.
low feedline losses. To reach the exact power level, minor adjustments are made to the power
output of the transmitter, usually by a front-panel power control.
NTSC UHF stations that want to achieve their maximum licensed power output are faced with
installing a very high power transmitter. Typical pairings in the U.S. include a transmitter rated
for 220 kW and an antenna with a gain of 25, or a 110 kW transmitter and a gain-of-50 antenna.
In the latter case, the antenna could pose a significant problem. UHF antennas with gains in the
region of 50 are possible, but not advisable for most installations because of coverage problems
that can result.
The amount of output power required of the transmitter will have a fundamental effect on sys-
tem design. Power levels dictate the following parameters:
• Whether the unit will be of solid-state or vacuum tube design
• Whether air, water, or vapor cooling must be used
• The type of power supply required
• The sophistication of the high-voltage control and supervisory circuitry
• Whether common amplification of the visual and aural signals (rather than separate visual and
aural amplifiers) is practical
Tetrodes generally are used for VHF transmitters above 25 kW, and specialized tetrodes can
be found in UHF transmitters at the 15 kW power level and higher. As solid-state technology
advances, the power levels possible in a reasonable transmitter design steadily increase, making
solid-state systems more attractive options.
In the realm of UHF transmitters, the klystron (and its related devices) reigns supreme.
Klystrons use an electron-bunching technique to generate high power—55 kW from a single tube
is not uncommon—at ultrahigh frequencies. They are currently the first choice for high-power,
high-frequency service. Klystrons, however, are not particularly efficient. A stock klystron with
no special circuitry might be only 40 percent efficient. Various schemes have been devised to
improve klystron efficiency, the best known of which is beam pulsing. Two types of pulsing are
in common use:
• Mod-anode pulsing, a technique designed to reduce power consumption of the device during
the color burst and video portion of the NTSC signal (and thereby improve overall system
efficiency)
• Annular control electrode (ACE) pulsing, which accomplishes basically the same goal by
incorporating the pulsing signal into a low-voltage stage of the transmitter, rather than a high-
voltage stage (as with mod-anode pulsing)
Experience has shown the ACE approach—and other similar designs—to provide greater
improvement in operating efficiency than mod-anode pulsing, and better reliability as well.
Several newer technologies offer additional ways to improve UHF transmitter efficiency,
including:
• The inductive output tube (IOT), also known as the Klystrode. This device essentially com-
bines the cathode/grid structure of the tetrode with the drift tube/collector structure of the
klystron.
• The multistage depressed collector (MSDC) klystron, a device that achieves greater effi-
ciency through a redesign of the collector assembly. A multistage collector is used to recover
energy from the electron stream inside the klystron and return it to the beam power supply.
Improved tetrode devices, featuring higher operating power at UHF and greater efficiency, have
also been developed.
A number of approaches can be taken to amplitude modulation of the visual carrier. Current
technology systems utilize low-level intermediate-frequency (IF) modulation. This approach
allows superior distortion correction, more accurate vestigial sideband shaping, and significant
economic advantages to the transmitter manufacturer.
A TV transmitter can be divided into four major subsystems:
• The exciter
• Intermediate power amplifier (IPA)
• Power amplifier
• High-voltage power supply
Figure 3.5.2 shows the audio, video, and RF paths for a typical design.
3.5.1b Power Amplifier

The power amplifier raises the output energy of the transmitter to the required RF operating
level. Tetrodes in TV service are usually operated in the class B mode to obtain reasonable effi-
ciency while maintaining a linear transfer characteristic. Class B amplifiers, when operated in
tuned circuits, provide linear performance because of the flywheel effect of the resonance circuit.
This allows a single tube to be used instead of two in push-pull fashion. The bias point of the lin-
ear amplifier must be chosen so that the transfer characteristic at low modulation levels matches
that at higher modulation levels. Even so, some nonlinearity is generated in the final stage,
Video subsystem RF processor subsystem
Video
input Video/delay Differential Visual
equalizer phase power
motherboard corrector control
Visual IF
Exciter/modulator subsystem
Visual Power Visual
Frequency IF splitter modulator
synthesizer
Attenuator Differential
gain
corrector
Fv LO mixer
Incidental
Aural phase
IF corrector
Audio Aural
input FMO
Aural Visual
Aural IF output/
output/
mixer mixer
Intermediate power amplifier subsystem
25 W
400 W
Visual 80 W Power visual 400 W
predriver splitter/ amplifier
25 W
attenuator array
Combiner
400 W
Aural visual 400 W
predriver amplifier
array
Power amplifier subsystem 65 W (approx) 800 W (approx)
Visual Notch Harmonic

power filter filter
amplifier
RF
output
Diplexer
Aural Harmonic
power filter
amplifier
Figure 3.5.2 Basic block diagram of a TV transmitter. The three major subassemblies are the
exciter, IPA, and PA. The power supply provides operating voltages to all sections, and high volt-
age to the PA stage.
requiring differential gain correction. The plate (anode) circuit of a tetrode PA usually is built
around a coaxial resonant cavity, providing a stable and reliable tank.
UHF transmitters using a klystron in the final output stage must operate class A, the most lin-
ear but also most inefficient operating mode for a vacuum tube. The basic efficiency of a non-
pulsed klystron is approximately 40 percent. Pulsing, which provides full available beam current
only when it is needed (during peak-of-sync), can improve device efficiency by as much as 25
percent, depending on the type of pulsing used.
Two types of klystrons are presently in service:
• Integral-cavity klystron
• External-cavity klystron
The basic theory of operation is identical for each tube, but the mechanical approach is radi-
cally different. In the integral-cavity klystron, the cavities are built into the klystron to form a
single unit. In the external-cavity klystron, the cavities are outside the vacuum envelope and
bolted around the tube when the klystron is installed in the transmitter.
A number of factors come into play in a discussion of the relative merits of integral- vs. exter-
nal-cavity designs. Primary considerations include operating efficiency, purchase price, and life
expectancy.
The PA stage includes a number of sensors that provide input to supervisory and control cir-
cuits. Because of the power levels present in the PA stage, sophisticated fault-detection circuits
are required to prevent damage to components in the event of a problem inside or outside the
transmitter. An RF sample, obtained from a directional coupler installed at the output of the
transmitter, is used to provide automatic power-level control.
The transmitter system discussed in this section assumes separate visual and aural PA stages.
This configuration is normally used for high-power NTSC transmitters. A combined mode also
may be used, however, in which the aural and visual signals are added prior to the PA. This
approach offers a simplified system, but at the cost of additional precorrection of the input video
signal.
PA stages often are configured so that the circuitry of the visual and aural amplifiers is identi-
cal, providing backup protection in the event of a visual PA failure. The aural PA can then be
reconfigured to amplify both the aural and the visual signals, at reduced power.
The aural output stage of a TV transmitter is similar in basic design to an FM broadcast trans-
mitter. Tetrode output devices generally operate class C, providing good efficiency. Klystron-
based aural PAs are used in UHF transmitters.
3.5.1c Application Example

A 60 kW transmitter is shown in block diagram form in Figure 3.5.3. A single high-power
klystron is used in the visual amplifier, and another is used in the aural amplifier. The tubes are
driven from solid-state intermediate power amplifier modules. The transmitter utilizes ACE-type
beam control, requiring additional predistortion to compensate for the nonlinearities of the final
visual stage. Predistortion is achieved by correction circuitry at an intermediate frequency in the
modulator. Both klystrons are driven from the output of a circulator, which ensures a minimum
of driver-to-load mismatch.
A block diagram of the beam modulator circuit is shown in Figure 3.5.4. The system receives
input signals from the modulator, which synchronizes ACE pulses to the visual tube with the
Sync Beam
modulator
Input Visual Harnonic RF

Visual IPA circulator klystron filter output
Video
input
Exciter RF combiner
Audio
input Input Aural Harmonic
Aural IPA circulator klystron filter
Low voltage supply

to all required sections
Monitoring
system
Low voltage/control
subassemblies
Support subassemblies
Supervisory signals (heater supply, magnet supply,
high voltage supply, ion pump)
Figure 3.5.3 Schematic diagram of a 60 kW klystron-based TV transmitter.
video information. The pulse waveform is developed through a pulse amplifier, rather than a
switch. This permits more accurate adjustments of operating conditions of the visual amplifier.
Although the current demand from the beam modulator is low, the bias is near cathode poten-
tial, which is at a high voltage relative to ground. The modulator, therefore, must be insulated
from the chassis. This is accomplished with optical transmitters and receivers connected via
fiber optic cables. The fiber optic lines carry supervisory, gain control, and modulating signals.
The four-cavity external klystrons will tune to any channel in the UHF-TV band. An adjust-
able beam perveance feature enables the effective electrical length of the device to be varied by
altering the beam voltage as a function of operating frequency. Electromagnetic focusing is used
on both tubes. The cavities, body, and gun areas of the klystrons are air-cooled. The collectors
are vapor-phase-cooled using an external heat exchanger system.
The outputs of the visual and aural klystrons are passed through harmonic filters to an RF
combiner before being applied to the antenna system.
3.5.1 DTV Transmitter Considerations

Two parameters determine the basic design of a DTV transmitter: the operating frequency and
the power level. For DTV, the frequency is spelled out in the FCCs Table of Allotments; the
power parameter, however, deserves—in fact, requires—additional consideration.
Television Transmitters
Transmitter
control
Status control Optical Optical

and interlocks receiver transmitter
Beam modulator control subsystem Beam modulator

drive subsystem
Wideband Wideband
optical optical
Driver transmitter receiver
system
Feeder AGC Optical Optical

monitor controller transmitter receiver
Feeder AGC
monitor controller
AC supply
input Linear Beam
amplifier Linear voltage
driver amplifier
Isolation Power
transformer supply
Visual
output
RF
Visual IPA coupler
Visual input signal Klystron
RF
coupler
Figure 3.5.4 Schematic diagram of the modulator section of a 60 kW TV transmitter.
3.5.1a Operating Power

The FCC allocation table for DTV lists ERP values that are given in watts rms. The use of the
term average power is not always technically correct. The intent was to specify the true heating
power or rms watts of the total DTV signal averaged over a long period of time [1]. The specifi-
cation of transmitter power is further complicated by the DTV system characteristic peak-to-
average ratio, which has a significant impact on the required power output rating of the transmit-
ter. For example, assume a given FCC UHF DTV ERP allocation of 405 kW rms, an antenna
Figure 3.5.5 Comparison of the DTV peak transmitter power rating and NTSC peak-of-sync.
(After [1].)
power gain of 24, and a transmission-line efficiency of 70 percent. The required DTV transmitter
power Tx will equal
405
Tx = -------- ÷ 0.7 = 24.1 kW (3.5.1)
24
Because the DTV peak-to-average ratio is 4 (6 dB), the actual DTV transmitter power rating
must be 96.4 kW (peak). This 4× factor is required to allow sufficient headroom for signal
peaks. Figure 3.5.5 illustrates the situation. The transmitter rating is a peak value because the RF
peak envelope excursions must traverse the linear operating range of the transmitter on a peak
basis to avoid high levels of IMD spectral spreading [1]. In this regard, the DTV peak envelope
power (PEP) is similar to the familiar NTSC peak-of-sync rating for setting the transmitter power
level. Note that NTSC linearizes the PEP envelope from sync tip to zero carrier for best perfor-
mance. Although many UHF transmitters use pulsed sync systems, where the major portion of
envelope linearization extends only from black level to maximum white, the DTV signal has no
peak repetitive portion of the signal to apply a pulsing system and, therefore, must be linearized
from the PEP value to zero carrier. Many analog transmitters also linearize over the full NTSC
amplitude range and, as a result, the comparison between NTSC and DTV peak RF envelope
power applies for setting the transmitter power. The DTV power, however, always is stated as
average (rms) because this is the only consistent parameter of an otherwise pseudorandom sig-
nal.
Figure 3.5.6 shows a DTV signal RF envelope operating through a transmitter IPA stage with
a spectral spread level of approximately –43 dB. Note the large peak circled on the plot at 9 dB.
This is significantly above the previously noted values. The plot also shows the average (rms)
level, the 6 dB peak level, and other sporadic peaks above the 6 dB dotted line. If the modulator
output were measured directly, where it can be assumed to be very linear, peak-to-average ratios
as high as 11 dB could be seen.
Figure 3.5.7 shows another DTV RF envelope, but in this case the power has been increased
to moderately compress the signal. Note that the high peaks above the 6 dB dotted line are nearly
gone. The peak-to-average ratio is 6 dB.
Figure 3.5.6 DTV RF envelope at a spectral spread of –43 dB. (After [1].)
Figure 3.5.7 DTV RF envelope at a spectral spread of –35 dB. (After [1].)
3.5.1 Technology Options

With the operating power and frequency established, the fundamental architecture of the trans-
mitter can be set. Three basic technologies are used for high-power television broadcasting
today:
• Solid-state—bipolar, metal-oxide-semiconductor field-effect transistor (MOSFET), LDMOS,
silicon carbide, and others.
Figure 3.5.8 Schematic diagram of a 600 W VHF amplifier using eight FETs in a parallel device/
parallel module configuration.
• Grid-based power vacuum tubes—tetrode, UHF tetrode, and Diacrode

• Klystron-based UHF devices—conventional klystron, MSDC klystron, and IOT
Each class of device has its strengths and weaknesses. Within each class, additional distinctions
also can be made.
3.5.1a Solid-State Devices

Solid-state devices play an increasingly important role in the generation of RF energy. As design-
ers find new ways to improve operating efficiency and to remove heat generated during use, the
maximum operating power per device continues to increase.
Invariably, parallel amplification is used for solid-state RF systems, as illustrated in Figure
3.5.8. Parallel amplification is attractive from several standpoints. First, redundancy is a part of
the basic design. If one device or one amplifier fails, the remainder of the system will continue to
operate. Second, lower-cost devices can be used. It is often less expensive to put two 150 W tran-
sistors into a circuit than it is to put in one 300 W device. Third, troubleshooting the system is
simplified because an entire RF module can be substituted to return the system to operation.
Solid-state systems are not, however, without their drawbacks. A high-power transmitter
using vacuum tubes is much simpler in design than a comparable solid-state system. The greater
the number of parts, the higher the potential for system failure. It is only fair to point out, how-
ever, that failures in a parallel fault-tolerant transmitter usually will not cause the entire system to
fail. Instead, some parameter, typically peak output power, will drop when one or more amplifier
modules is out of service.
This discussion assumes that the design of the solid-state system is truly fault-tolerant. For a
system to provide the benefits of parallel design, the power supplies, RF divider and combiner
networks, and supervisory/control systems also must be capable of independent operation. Fur-
thermore, hot swapping of defective modules is an important attribute.
The ac-to-RF efficiency of a solid-state transmitter may or may not be any better than a tube
transmitter of the same operating power and frequency. Much depends on the type of modulation
used and the frequency of operation. Fortunately, the lower average power and duty cycle of the
DTV signal suggests that a high-efficiency solid-state solution may be possible [2]. The rela-
tively constant signal power of the DTV waveform eliminates one of the biggest problems in
NTSC applications of class AB solid-state amplifiers: the continual level changes in the video
signal vary the temperature of the class AB amplifier junctions and, thus, their bias points. This,
in turn, varies all of the transistor parameters, including gain and linearity. Sophisticated adaptive
bias circuits are required for reduction or elimination of this limitation to class AB operation.
Solid-state amplifiers operating class A do not suffer from such linearity problems, but the
class A operation imposes a substantial efficiency penalty. Still, many designs use class A
because of its simplicity and excellent linearity.
The two primary frontiers for solid-state devices are 1) power dissipation and 2) improved
materials and processes [3]. With regard to power dissipation, the primary factor in determining
the amount of power a given device can handle is the size of the active junctions on the chip. The
same power output from a device also may be achieved through the use of several smaller chips
in parallel within a single package. This approach, however, can result in unequal currents and
uneven distribution of heat. At high power levels, heat management becomes a significant factor
in chip design. Specialized layout geometries have been developed to ensure even current distri-
bution throughout the device.
The second frontier—improved materials and processes—is being addressed with technolo-
gies such as LDMOS and silicon carbide (SiC). The bottom line with regard to solid-state is that,
from the standpoint of power-handling capabilities, there is a point of diminishing returns for a
given technology. The two basic semiconductor structures, bipolar and FET, have seen numerous
fabrication and implementation enhancements over the years that have steadily increased the
maximum operating power and switching speed. Power MOSFET, LDMOS, and SiC devices are
the by-products of this ongoing effort.
With any new device, or class of devices, economies always come into play. Until a device has
reached a stable production point, at which it can be mass-produced with few rejections, the per-
device cost is usually high, limiting its real-world applications. For example, if an SiC device that
can handle more than 4 times the power of a conventional silicon transistor is to be cost-effective
in a transmitter, the per-device cost must be less than 4 times that of the conventional silicon
product. It is fair to point out in this discussion that the costs of the support circuitry, chassis, and
heat sink are equally important. If, for example, the SiC device—though still at a cost disadvan-
tage relative to a conventional silicon transistor—requires fewer support elements, then a cost
advantage still can be realized.
If increasing the maximum operating power for a given frequency is the primary challenge for
solid-state devices, then using the device more efficiently ranks a close second. The only thing
better than being able to dissipate more power in a transistor is not generating the waste heat in
the first place. The real performance improvements in solid-state transmitter efficiency have not
come as a result of simply swapping out a single tube with 200 transistors, but from using the
transistors in creative ways so that higher efficiency is achieved and fewer devices are required.
This process has been illustrated dramatically in AM broadcast transmitters. Solid-state trans-
mitters have taken over that market at all power levels, not because of their intrinsic feature set
(graceful degradation capability when a module fails, no high voltages used in the system, sim-
plified cooling requirements, and other attributes), but because they lend themselves to enor-
mous improvements in operating efficiency as a result of the waveforms being amplified. For
Figure 3.5.9 Cutaway view of the tetrode (left ) and the Diacrode (right). Note that the RF current
peaks above and below the Diacrode center, but the tetrode has only one peak at the bottom.
(After [6].)
television, most notably UHF, the march to solid-state has been much slower. One of the prom-
ises of DTV is that clever amplifier design will lead to similar, albeit less dramatic, improve-
ments in intrinsic operating efficiency.
3.5.1b Power Grid Devices

Advancements in vacuum tube technology have permitted the construction of numerous high-
power UHF transmitters based on tetrodes [4, 5]. Such devices are attractive for television appli-
cations because they are inherently capable of operating in an efficient class AB mode. UHF tet-
rodes operating at high power levels provide essentially the same specifications, gain, and
efficiency as tubes operating at lower powers. The anode power-supply voltage of the tetrode is
much lower than the collector potential of a klystron- or IOT-based system (8 kV is common).
Also, the tetrode does not require focusing magnets.
Efficient removal of heat is the key to making a UHF tetrode practical at high power levels.
Such devices typically use water or vapor-phase cooling. Air-cooling at such levels is impractical
because of the fin size that would be required. Also, the blower for the tube would have to be
quite large, reducing the overall transmitter ac-to-RF efficiency.
Another drawback inherent in tetrode operation is that the output circuit of the device appears
electrically in series with the input circuit and the load [2]. The parasitic reactance of the tube
elements, therefore, is a part of the input and output tuned circuits. It follows, then, that any
change in the operating parameters of the tube as it ages can affect tuning. More importantly, the
series nature of the tetrode places stringent limitations on internal-element spacings and the
physical size of those elements in order to minimize the electron transit time through the tube
vacuum space. It is also fair to point out, however, that the tetrode’s input-circuit-to-output-cir-
cuit characteristic has at least one advantage: power delivered to the input passes through the
tube and contributes to the total power output of the transmitter. Because tetrodes typically
exhibit low gain compared with klystron-based devices, significant power may be required at the
Figure 3.5.10 The elements of the Diacrode, including the upper cavity. Double current, and con-
sequently, double power is achieved with the device because of the current peaks at the top and
bottom of the tube, as shown. (After [6].)
input circuit. The pass-through effect, therefore, contributes to the overall operating efficiency of
the transmitter.
The expected lifetime of a tetrode in UHF service usually is shorter than that of a klystron of
the same power level. Typical lifetimes of 8000 to 15,000 hours have been reported. Intensive
work, however, has led to products that offer higher output powers and extended operating life-
times, while retaining the benefits inherent in tetrode devices. With regard to DTV application
possibilities, the linearity of the tetrode is excellent, a strong point for DTV consideration [6].
Minimal phase distortion and low intermodulation translate into reduced correction requirements
for the amplifier.
The Diacrode (Thomson) is an adaptation of the high-power UHF tetrode. The operating prin-
ciple of the Diacrode is basically the same as that of the tetrode. The anode current is modulated
by an RF drive voltage applied between the cathode and the power grid. The main difference is in
the position of the active zones of the tube in the resonant coaxial circuits, resulting in improved
reactive current distribution in the electrodes of the device.
Figure 3.5.9 compares the conventional tetrode with the Diacrode. The Diacrode includes an
electrical extension of the output circuit structure to an external cavity [6]. The small dc-blocked
cavity rests on top of the tube, as illustrated in Figure 3.5.10.
The cavity is a quarter-wave transmission line, as measured from the top of the cavity to the
vertical center of the tube. The cavity is short-circuited at the top, reflecting an open circuit (cur-
rent minimum) at the vertical center of the tube and a current maximum at the base of the tube,
like the conventional tetrode, and a second current maximum above the tube at the cavity short-
circuit. (Figure 3.5.9 helps to visualize this action.)
With two current maximums, the Diacrode has an RF power capability twice that of the
equivalent tetrode, while the element voltages remain the same. All other properties and aspects
of the Diacrode are basically identical to those of the TH563 high-power UHF tetrode (Thom-
son), upon which the Diacrode is patterned.
Figure 3.5.11 Mechanical design of the multistage depressed collector assembly. Note the “V”
shape of the 4-element system.
Some of the benefits of such a device, in addition to the robust power output available, are its
low high-voltage requirements (low relative to a klystron/IOT-based system, that is), small size,
and simple replacement procedures.
3.5.1c Klystron-Based Devices

The klystron is a linear-beam device that overcomes the transit-time limitations of a grid-con-
trolled vacuum tube by accelerating an electron stream to a high velocity before it is modulated
[7]. Modulation is accomplished by varying the velocity of the beam, which causes the drifting
of electrons into bunches to produce RF space current. One or more cavities reinforce this action
at the operating frequency. The output cavity acts as a transformer to couple the high-impedance
beam to a low-impedance transmission line. The frequency response of a klystron is limited by
the impedance-bandwidth product of the cavities, which may be extended by stagger tuning or by
the use of multiple-resonance filter-type cavities.
The klystron is one of the primary means of generating high power at UHF frequencies and
above. Output powers for multicavity devices range from a few thousand watts to 10 MW or
more. The klystron provides high gain and requires little external support circuitry. Mechani-
Grid
+ Bias -
power
supply
Anode
Tailpipe
Collector
Electron beam
- HV +
power
supply
Cathode
RF input RF output
Figure 3.5.12 Functional schematic diagram of an IOT.
cally, the klystron is relatively simple. It offers long life and requires a minimum of routine main-
tenance.
The klystron, however, is inefficient in its basic form. Efficiency improvements can be gained
for television applications through the use of beam pulsing; still, a tremendous amount of energy
must be dissipated as waste heat. Years of developmental research have produced two high-effi-
ciency devices for television use: the MSDC klystron and the IOT, also known as the Klystrode.
The MSDC device is essentially identical to a standard klystron, except for the collector
assembly. Beam reconditioning is achieved by including a transition region between the RF
interaction circuit and the collector under the influence of a magnetic field. From an electrical
standpoint, the more stages of a multistage depressed collector klystron, the better. Predictably,
the tradeoff is increased complexity and, therefore, increased cost for the product. Each stage
that is added to the depressed collector system is a step closer to the point of diminishing returns.
As stages are added above four, the resulting improvements in efficiency are proportionally
smaller. Because of these factors, a 4-stage device is common for television service. (See Figure
3.5.11.)
The IOT is a hybrid of a klystron and a tetrode. The high reliability and power-handling capa-
bility of the klystron is due, in part, to the fact that electron-beam dissipation takes place in the
collector electrode, quite separate from the RF circuitry. The electron dissipation in a tetrode is at
the anode and the screen grid, both of which are inherent parts of the RF circuit; therefore, they
must be physically small at UHF frequencies. An advantage of the tetrode, on the other hand, is
that modulation is produced directly at the cathode by a grid so that a long drift space is not
required to produce density modulation. The IOT has a similar advantage over the klystron—
high efficiency in a small package.
The IOT is shown schematically in Figure 3.5.12. The electron beam is formed at the cathode,
density-modulated with the input RF signals by a grid, then accelerated through the anode aper-
ture. In its bunched form, the beam drifts through a field-free region and interacts with the RF
f @ G ( )
Figure 3.5.13 Mechanical configuration of a high-power IOT (EEV). (After [9].)
field in the output cavity. Power is extracted from the beam in the same way it is extracted from a
klystron. The input circuit resembles a typical UHF power grid tube. The output circuit and col-
lector resemble a klystron.
Because the IOT provides beam power variation during sync pulses (as in a pulsed klystron)
as well as over the active modulating waveform, it is capable of high efficiency. The device thus
provides full-time beam modulation as a result of its inherent structure and class B operation.
For DTV service, the IOT is particularly attractive because of its good linearity characteris-
tics. The IOT provides –60 dB or better intermodulation performance in combined 10 dB aural/
visual service [7]. Tube life data varies depending upon the source, but one estimate puts the life
expectancy at more than 35,000 hours [8].
The maximum power output available from an IOT range from 55 kW visual plus 5.5 kW
aural in common amplification [9]. Various improvements have been made over the years,
including changes to the input cavity to improve intermodulation performance of the device (Fig-
ure 3.5.13).
3.5.1d Constant Efficiency Amplifier

Because of the similarity between the spent electron beam in an IOT and that of a klystron, it is
possible to consider the use of a multistage depressed collector on an IOT to improve the operat-
Figure 3.5.14 Schematic overview of the MSDC IOT or constant efficiency amplifier. (After [10].)
ing efficiency [10]. This had been considered by Priest and Shrader [11] and by Gilmore [12],
but the idea was rejected because of the complexity of the multistage depressed collector assem-
bly and because the IOT already exhibited fairly high efficiency. Subsequent development by
Symons [10, 13] led to a working device. An inductive output tube, modified by the addition of a
multistage depressed collector, has the interesting property of providing linear amplification
with (approximately) constant efficiency.
Figure 3.5.14 shows a schematic representation of the constant efficiency amplifier (CEA)
[10]. The cathode, control grid, anode and output gap, and external circuitry are essentially iden-
tical with those of the IOT amplifier. Drive power introduced into the input cavity produces an
electric field between the control grid and cathode, which draws current from the cathode during
positive half-cycles of the input RF signal. For operation as a linear amplifier, the peak value of
the current—or more accurately, the fundamental component of the current—is made (as nearly
as possible) proportional to the square root of the drive power, so that the product of this current
and the voltage it induces in the output cavity will be proportional to the drive power.
Following the output cavity is a multistage depressed collector in which several typical elec-
tron trajectories are shown. These are identified by the letters a through e. The collector elec-
trodes are connected to progressively lower potentials between the anode potential and the
cathode potential so that more energetic electrons penetrate more deeply into the collector struc-
ture and are gathered on electrodes of progressively lower potentials.
In considering the difference between an MSDC IOT and an MSDC klystron, it is important
to recognize that in a class B device, no current flows during the portion of the RF cycle when
the grid voltage is below cutoff and the output gap fields are accelerating. As a result, it is not
necessary to have any collector electrode at a potential equal to or below cathode potential. At
low output powers, when the RF output gap voltage is just equal to the difference in potential
between the lowest-potential collector electrode and the cathode, all the current will flow to that
electrode. Full class B efficiency is thus achieved under these conditions.
As the RF output gap voltage increases with increased drive power, some electrons will have
lost enough energy to the gap fields so they cannot reach the lowest potential collector, and so
current to the next-to-the-lowest potential electrode will start increasing. The efficiency will
drop slightly and then start increasing again until all the current is just barely collected by the
two lowest-potential collectors, and so forth.
Maximum output power is reached when the current delivered to the output gap is sufficient
to build up an electric field or voltage that will just stop a few electrons. At this output power, the
current is divided among all of the collector electrodes and the efficiency will be somewhat
higher than the efficiency of a single collector, class B amplifier [10].
The challenge of developing a multistage depressed collector for an IOT is not quite the same
as that of developing a collector for a conventional klystron [13]. It is different because the dc
component of beam current rises and falls in proportion to the square root of the output power of
the tube. The dc beam current is not constant as it is in a klystron. As a result, the energy spread
is low because the output cavity RF voltage is low at the same time that the RF and dc beam cur-
rents are low. Thus, there will be small space-charge forces, and the beam will not spread as
much as it travels deep into the collector toward electrodes having the lowest potential. For this
reason, the collector must be rather long and thin when compared to the multistage depressed
collector for a conventional klystron, as described previously.
3.5.1e Digital Signal Pre-Correction

Because amplitude modulation contains at least three parent signals (the carrier plus two side-
bands) its amplification has always been fraught with in-band intermodulation distortion [14]. A
common technique to correct for these distortion products is to pre-correct an amplifier by inten-
tionally generating an anti-intermodulation signal. The following process is typically used:
• Mix the three tones together
• Adjust the amplitude of the anti-IM product to match that of the real one
• Invert the phase
• Delay the anti-IM product or the parent signals in time so that it aligns with the parent signals
• Add the anti-IM signal into the main parent signal path to cause cancellation in the output cir-
cuit of the amplifier
This is a common approach that works well in countless applications. Each intermodulation
product may have its own anti-intermodulation product intentionally generated to cause targeted
cancellation.
In a vacuum tube or transistor stage that is amplifying a noise-like signal, such as the 8-VSB
waveform, sideband components arise around the parent signal at the output of the amplifier that
also appear noise-like and are correlated with the parent signal. The amplifying device, thus,
41-47 VSB
Signal. (Low 35-53 MHz
Level. -10 dBm anti-intermodulation
to each mixer) Level
cancellation signals
Main IF Adjust
Signal in X X
(no correction)
Delay
Adjust Main Signal
Path
Summing
Circuit Amplitude Phase
Adjust Adjust
Main IF Signal
out (with correction)
Figure 3.5.15 An example 8-VSB sideband intermodulation distortion corrector functional block
diagram for the broadcast television IF band from 41 to 47 MHz for use with DTV television trans-
mitters. (From [14]. Used with permission.)
behaves precisely like two mixers not driven into switching. As with the analog pre-correction
process described previously, pre-correction for the digital waveform also is possible.
The corrector illustrated in Figure 3.5.15 mimics the amplifying device, but contains phase
shift, delay, and amplitude adjustment circuitry to properly add (subtract), causing cancellation
of the intermodulation components adjacent to the parent signal at the amplifier output. Because
the intentional mixing circuit is done in a controlled way, it may be used at the IF frequency
before the RF conversion process in the transmitter so that an equal and opposite phase signal is
generated and added to the parent signal at IF. It will then go through the same RF conversion
and amplification process as the parent signal that will spawn the real intermodulation products.
The result is intermodulation cancellation at the output of the amplifier.
Figure 3.5.16 shows the relative amplitude content of the correction signal spanning 6 MHz
above and below the desired signal. Because the level of the correction signal must match that of
the out-of-band signal to be suppressed, it must be about 43 dB below the in-band signal accord-
ing to the example. This level of in-band addition is insignificant to the desired signal, but just
enough to cause cancellation of the out-of-band signal.
Using such an approach that intentionally generates the correct anti-intermodulation compo-
nent and causes it to align in time, be opposite phase, and equal in amplitude allows for cancella-
tion of the unwanted component, at least in part. The degree of cancellation has everything to do
with the precise alignment of these three attributes. In practice, it has been demonstrated [19]
that only the left or right shoulder may be optimally canceled by up to about 3–5 dB. This
amount may not seem to be significant, but it must be remembered that 3 dB of improvement is
cutting the power in half.
Pass Band
35-53 MHz anti-intermodulation

distortion signal to be added
(subtracted) to parent signal
Amplitude
35 Frequency 53
Figure 3.5.16 After amplification, the out-of-band signal is corrected or suppressed by an amount
determined by the ability of the correction circuit to precisely match the delay, phase, and ampli-
tude of the cancellation signal. (From [14]. Used with permission.)
3.5.1 Implementation Issues

The design and installation of a television transmitter is a complicated process that must take into
consideration a number of variables. Some of the more important issues include:
• The actual cost of the transmitter, both the purchase price and the ongoing maintenance
expenses for tubes and other supplies.
• The actual ac-to-RF efficiency, which relates directly, of course, to the operating costs.
• Maintenance issues, the most important of which is the mean time between failure (MTBF).
Also significant is the mean time to repair (MTTR), which relates directly to the accessibility
of transmitter components and the type of amplifying devices used in the unit.
• Environmental issues, not the least of which is the occupied space needed for the transmitter.
The cooling requirements also are important and may, for example, affect the ongoing main-
tenance costs.
• The availability of sufficient ac power and power of acceptable reliability and regulation at
the site.
3.5.1 References
1. Plonka, Robert J.: “Planning Your Digital Television Transmission System,” Proceedings of
the 1997 NAB Broadcast Engineering Conference, National Association of Broadcasters,
Washington, D.C., pg. 89, 1997.
2. Ostroff, Nat S.: “A Unique Solution to the Design of an ATV Transmitter,” Proceedings of
3. Whitaker, Jerry C.: “Solid State RF Devices,” Radio Frequency Transmission Systems:
Design and Operation, McGraw-Hill, New York, pg. 101, 1990.
4. Whitaker, Jerry C.: “Microwave Power Tubes,” The Electronics Handbook, Jerry C. Whi-
taker (ed.), CRC Press, Boca Raton, Fla., pg. 413, 1996.
5. Tardy, Michel-Pierre: “The Experience of High-Power UHF Tetrodes,” Proceedings of the
6. Hulick, Timothy P.: “60 kW Diacrode UHF TV Transmitter Design, Performance and Field
Report,” Proceedings of the 1996 NAB Broadcast Engineering Conference, National Asso-
ciation of Broadcasters, Washington, D.C., pg. 442, 1996.
7. Whitaker, Jerry C.: “Microwave Power Tubes,” Power Vacuum Tubes Handbook, Van Nos-
trand Reinhold, New York, pg. 259, 1994.
8. Ericksen, Dane E.: “A Review of IOT Performance,” Broadcast Engineering, Intertec Pub-
lishing, Overland Park, Kan., pg. 36, July 1996.
9. Aitken, S., D. Carr, G. Clayworth, R. Heppinstall, and A. Wheelhouse: “A New, Higher
Power, IOT System for Analogue and Digital UHF Television Transmission,” Proceedings
of the 1997 NAB Broadcast Engineering Conference, National Association of Broadcasters,
10. Symons, Robert S.: “The Constant Efficiency Amplifier,” Proceedings of the NAB Broad-
cast Engineering Conference, National Association of Broadcasters, Washington, D.C., pp.
523–530, 1997.
11. Priest, D. H., and M. B. Shrader: “The Klystrode—An Unusual Transmitting Tube with
Potential for UHF-TV,” Proc. IEEE, vol. 70, no. 11, pp. 1318–1325, November 1982.
12. Gilmore, A. S.: Microwave Tubes, Artech House, Dedham, Mass., pp. 196–200, 1986.
13. Symons, R., M. Boyle, J. Cipolla, H. Schult, and R. True: “The Constant Efficiency Ampli-
fier—A Progress Report,” Proceedings of the NAB Broadcast Engineering Conference,
National Association of Broadcasters, Washington, D.C., pp. 77–84, 1998.
14. Hulick, Timothy P.: “Very Simple Out-of-Band IMD Correctors for Adjacent Channel
NTSC/DTV Transmitters,” Proceedings of the Digital Television '98 Conference, Intertec
Publishing, Overland Park, Kan., 1998.
g g
Chapter
3.6
Multiple Transmitter Networks
3.6.1 Introduction1
Many of the challenges of RF transmission, especially as they apply to digital transmission, can
be addressed by using multiple transmitters to cover a service area [1]. Because of the limitations
in the spectrum available, many systems based on the use of multiple transmitters must operate
those transmitters all on the same frequency, hence the name single frequency network (SFN). At
the same time, use of SFNs leads to a range of additional complications that must be addressed
in the design of the network.
SFNs for single-carrier signals such as 8-VSB become possible because of the presence of
adaptive equalizers in receivers. When signals from multiple transmitters arrive at a receiver,
under the right conditions, the adaptive equalizer in that receiver can treat the several signals as
echoes of one another and extract the data they carry. The conditions are controlled by the capa-
bilities of the adaptive equalizer and will become less stringent as the technology of adaptive
equalizers improves over time.
3.6.1a Benefits of Multiple Transmitters

A number of benefits may accrue to the use of multiple transmitters to cover a service area [1].
Among these can be the ability to obtain more uniform signal levels throughout the area being
served and the maintenance of higher average signal levels over that area. These results come
from the fact that the average distance from any point within the service area to a transmitter is
reduced. Reducing the distance also reduces the variability of the signal level with location and
time and thereby reduces the required fade margin needed to maintain any particular level of reli-
1. This chapter is based on: ATSC, “ATSC Recommended Practice: Design Of Synchronized
Multiple Transmitter Networks,” Advanced Television Systems Committee, Washington,
D.C., Doc. A/111, September 3, 2004. Used with permission.
Editor’s note: This chapter provides an overview of multiple transmitter networks based on
ATSC A/111. Readers are encouraged to download the entire Recommended Practice from
the ATSC Web site (http://www.atsc.org). All ATSC Standards, Recommended Practices, and
Information Guides are available at no charge.
3-97
p
ability of service. These reductions, in turn, permit operation with less overall effective radiated
power (ERP) and/or antenna height.
When transmitters can be operated at lower power levels and/or elevations, the interference
they cause to their neighbors is reduced. Using multiple transmitters allows a station to provide
significantly higher signal levels near the edge of its service area without causing the level of
interference to its neighbor that would arise if the same signal levels were delivered from a sin-
gle, central transmitter. The interference reductions come from the significantly smaller interfer-
ence zones that surround transmitters that use relatively lower power and/or antenna heights.
With the use of multiple transmitters comes the ability to overcome terrain limitations by fill-
ing in areas that would otherwise receive insufficient signal level. When the terrain limitations
are caused by obstructions that isolate an area from another (perhaps the main) transmitter,
advantage may be taken of the obstructions in the design of the network. The obstructions can
serve to help isolate signals from different transmitters within the network, making it easier to
control interference between the network’s transmitters. When terrain obstructions are used in
this way, it may be possible to place transmitters farther apart than if such obstructions were not
utilized for isolation.
Where homes are illuminated by sufficiently strong signals from two or more transmitters, it
may be possible to take advantage of the multiple signals to provide more reliable indoor recep-
tion. When a single transmitter is used, standing waves within a home sheathed in metal likely
will result in areas within that home having signal levels too low to use. Signals arriving from
different directions will enter the resonant cavity of the home through different ports (windows)
and set up standing waves in different places. The result often may be that areas within the home
receiving low signal levels from one transmitter will receive adequate signal levels from another
transmitter, thereby making reliable reception possible in many more places within the home.
3.6.1b Limitations of SFNs

While they have the potential to solve or at least ameliorate many of the challenges of RF trans-
mission, SFNs also have limitations of their own [1]. Foremost among these is the fact that there
will be interference between the signals from the several transmitters in a network. This “system-
internal” interference must be managed so as to bring it within the range of capabilities of the
adaptive equalizers of the largest number of receivers possible. Where the interference falls out-
side the range that can be handled by a given adaptive equalizer, other measures, such as the use
of an outdoor directional antenna, must be applied.
The characteristics of adaptive equalizers that are important for single frequency network
design are the ability to deal with multiple signals (echoes) with equal signal levels, the length of
echo delay time before and after the main (strongest) signal that can be handled by the equalizer
(or the total delay spread of echoes that can be handled when an equalizer design does not treat
one of the signals as the main signal), and the interfering signal level below which the adaptive
equalizer is not needed because the interference is too low in level to prevent reception.
The last of these characteristics, which is somewhat lower in level than the co-channel inter-
ference threshold, determines the areas where the performance of the adaptive equalizer matters.
In places with echo interference below the point at which the adaptive equalizer needs to correct
the signal, it is not necessary in design of the SFN to consider the differential arrival times of the
signals from the various transmitters. In places with echo interference above that threshold, the
arrival times of the signals matter if the adaptive equalizer is to be able to correct for the echo
p
interference. Differential delays between signals above the echo interference threshold must fall
within the time window that the adaptive equalizer can correct if the signals are to be received.
Current receiver designs have a fixed time window inside which echoes can be equalized. The
amplitudes of correctable echoes also are a function of their time displacement from the main
signal. The closer together the signals are in time, the closer they can be in amplitude. The fur-
ther apart they are in time, the lower in level the echoes must be for the equalizer to work. These
relationships are improving dramatically in newer receiver front-end designs, and they can be
expected to continue improving at least over the next several generations of designs. As they
improve, limitations on SFN designs will be reduced.
The handling of signals with equal signal level echoes is a missing capability in early receiver
front-end designs, but it is now recognized as necessary for receivers to work in many situations
that occur naturally, without even considering their generation in SFNs. The reason for this is
that any time there is no direct path or line-of-sight from the transmitter to the receiver, the
receiver will receive all of its input energy from reflections of one sort or another. When this hap-
pens, there may be a number of signals (echoes) arriving at the receiver that are about equal in
amplitude, and they may vary over time, with the strongest one changing from time-to-time. This
is called a Rayleigh channel when it occurs, and it is now recognized that Rayleigh channels are
more prevalent than once thought. For example, they often exist in city canyons and mid-rise
areas, they exist behind hills, and so on. They also exist in many indoor situations. If receivers
are to deal with these cases, adaptive equalizers will have to be designed to handle them. Thus,
SFNs will be able to take advantage of receiver capabilities that are needed in natural circum-
stances.
Radio frequency signals travel at a speed of about 3/16 mile per microsecond. Another way to
express the same relationship is that radio frequency signals travel a mile in about 5-1/3 μs. If a
pair of transmitters emits the same signal simultaneously and a receiver is located equidistant
from the two transmitters, the signals will arrive at the receiver simultaneously. If the receiver is
not equidistant from the transmitters, the arrival times at the receiver of the signals from the two
transmitters will differ by 5-1/3 μs for each mile of difference in path length. In designing the
network, the determination of the sizes of the cells and the related spacing of the transmitters will
be dependent on this relationship between time and distance and on the delay spread capability
of the receiver adaptive equalizer.
Because receivers have limited delay-spread capability, there is a corresponding limit on the
sizes of cells and spacing of transmitters in SFNs. As receiver front-end technology improves
over time, this limitation can be expected to be relaxed. As the limitation on cell sizes is relaxed
and cells become larger, it can be helpful to network design to adjust the relative emission times
of the transmitters in the network. This allows putting the locus of equidistant points from vari-
ous transmitters where needed to maximize the audience reached and to minimize internal inter-
ference within the network. When such time offsets are used, it becomes desirable to be able to
measure the arrival times at receiving locations of the signals from the transmitters in the net-
work. Such measurements can be difficult since the transmitters are intentionally transmitting
exactly the same signals in order to allow receivers to treat them as echoes of one another. Some-
how the transmitters have to be differentiated from one another if their respective contributions
to the aggregate signal received at any location are to be determined. To aid in the identification
of individual transmitters in a network, a buried spread spectrum pseudorandom “RF Water-
mark” signal is included in ATSC Standard A/110, “Synchronization Stnadard for Distributed
Transmission” [2].
p
Table 3.6.1 Comparison of Different Distributed Transmitter Networks ( After [1])
SFN Configuration Distributed Transmitter DOCR Distributed Translator

Complexity/cost High Low Low
Requirement for feeder link Yes No No
Delay adjustment capability Yes No Yes
Output power level No limit Low/moderate No limit
DTV RF channels required One One At least two
Recommended implementa- Large area SFN Cover small area: gap filler, Large area SFN, gap filler, and
tion and coverage extender coverage extender
3.6.2 Single Frequency Network Concepts

Table 3.6.1 lists the most common forms of distributed transmitter networks and some of their
primary characteristics.
3.6.2a Digital On-Channel Repeaters

In block diagram form, digital on-channel repeaters (DOCRs) look just like translators or boost-
ers (Figure 3.6.1) [1]. They comprise a receiving antenna, a receiver, some signal processing, a
transmitter, and a transmitting antenna. They are closest in nature to boosters in that they receive
the off-the-air DTV signal, process it, and retransmit it on the same frequency. This greatly sim-
plifies the portion of a single frequency network that employs them, but it also leads to limita-
tions in how they can be applied. For example, due to the internal signal processing and filtering
delays of a DOCR, the signal from the main transmitter will arrive at the DOCR coverage area
first, which acts as a pre-echo, relative to the output signal from the DOCR. This type of multi-
path distortion is very harmful to reception by ATSC legacy receivers. With the anticipated per-
formance improvements of newer receivers, this problem is expected to be reduced or eliminated
in the future. The main advantage for the DOCR is its simplicity and low cost. With a DOCR,
there is no need for a separate studio-to-transmitter link (STL).
There are several configurations that can be used in DOCRs. Shown in Figures 3.6.2 through
3.6.5, they can be designated as:
• RF processing DOCR (Figure 3.6.2)
• IF processing DOCR (Figure 3.6.3)
• Baseband decoding and re-generation DOCR (Figure 3.6.4)
• Baseband demodulation and equalization DOCR (Figure 3.6.5)
Figure 3.6.2 shows the simplest form of DOCR, the RF processing DOCR, in which the
receiver comprises a preselector and low-level amplifier, the signal processing comprises an RF
bandpass filter at the channel of operation, and the transmitter comprises a power amplifier.
There is no frequency translation of any sort in this arrangement. It has the shortest processing
delay of any of the DOCR configurations—usually a fraction of a microsecond. This configura-
tion, however, has very limited first adjacent channel interference rejection capability. Such a
configuration can result in the generation of intermodulation products in the amplifier and in
degradation of the re-transmitted signal quality. Meanwhile, due to limited isolation (i.e., too
much coupling) between transmitting and receiving antennas, the re-transmitted signal could
p
Main Transmitter Digital On-Channel Repeater
Consumer Receiver
Signal Processing
& Power Amplification
Power Amplification
Data
Receiver Signal Processing
Processing
Figure 3.6.1 Digital on-channel repeater (DOCR) generic block diagram. (From [1]. Used with per-
mission.)
loop back and re-enter the receiving antenna. This can also degrade the signal quality, causing
spectrum ripple and other distortions. The only way to limit or avoid the signal loopback in this
type of DOCR is to increase antenna isolation, which is determined by the site environment, or to
limit the DOCR output power. Usually, the RF processing DOCR transmitter output power is less
than 10 W, resulting in an effective radiated power (ERP) on the order of several dozen watts.
Conversion to an intermediate frequency (IF) for signal processing is the principal feature
that differentiates Figure 3.6.3 from the RF processing DOCR. In this arrangement, a local oscil-
lator and mixer are used to convert the incoming signal to the IF frequency, where it can be more
easily amplified and filtered. The same local oscillator used for the downconversion to IF in the
receiver can be used for upconversion in the transmitter, resulting in the signal being returned to
precisely the same frequency at which it originated (with some amount of the local oscillator
phase noise added to the signal). The delay time through the IF processing DOCR will be
decided mostly by the IF filter implemented. A SAW filter can have much sharper passband
edges, better control of envelope delay, greater attenuation in the stopband, and generally more
repeatable characteristics than most other kinds of filters. Transit delay time for the signal can be
on the order of 1–2 μs, so the delay through an IF processing DOCR, Figure 3.6.3, will be from a
fraction of a microsecond to about 2 μs—somewhat longer than the RF processing DOCR. The
IF processing DOCR has better first adjacent channel interference rejection capability than does
p
Figure 3.6.2 RF Processing DOCR configuration. Figure 3.6.3 IF processing DOCR configuration.
(From [1]. Used with permission.) (From [1]. Used with permission.)
the RF processing DOCR, but it retains the signal loopback problem, which limits its output
power.
Figure 3.6.4 shows a receiver that demodulates the incoming signal all the way to a digital
baseband signal in which forward error correction (FEC) can be applied. This restores the bit
stream to perfect condition, correcting all errors and eliminating all effects of the analog channel
through which the signal passed in reaching the DOCR. The bit stream then is transmitted, start-
ing with formation of the bit stream into the symbols in an exciter, just as in a normal transmitter.
If no special steps are taken to set the correct trellis encoder states, the output of the DOCR of
Figure 3.6.4 would be incoherent with respect to its input. This would result in the signal from
such a repeater acting like noise when interfering with the signal from another transmitter in the
network rather than acting like an echo of the signal from that other transmitter. Thus, additional
data processing is required to establish the correct trellis states for retransmission. It should also
be noted that this form of DOCR has a very long delay through it, measured in milliseconds,
mostly caused by the de-interleaving process. This delay is well outside the ATSC receiver equal-
ization range. Although by regenerating the DTV signal it totally eliminates the adjacent channel
interference and signal loopback problems, this type of DOCR has very little practical use,
unless the intended DOCR coverage area is totally isolated from the main DTV transmitter.
p
Figure 3.6.4 Baseband decoding DOCR configu- Figure 3.6.5 Baseband equalization DOCR con-
ration. (From [1]. Used with permission.) figuration. (From [1]. Used with permission.)
A more practical intermediate method is the baseband equalization DOCR (or EDOCR) as
shown in Figure 3.6.5. It fits between the techniques of Figures 3.6.2 or 3.6.3 and 3.6.4. This
type of DOCR demodulates the received signal and applies adaptive equalization in order to
reduce or eliminate adjacent channel interference, signal loopback, and multipath distortion
occurring in the path from the main transmitter to the EDOCR. In determining the correct 3-bit
symbol levels, it also carries out symbol level slicing or trellis decoding, which can achieve sev-
eral dB of noise reduction and minimizes the impact of channel distortions. The baseband output
of the equalizer and slicer is re-modulated, filtered, frequency shifted, and amplified for re-trans-
mission. The delay of the baseband processing is in the order of a few dozen VSB symbol times.
The total EDOCR internal delay is in the order of a few microseconds once the time delays of the
adaptive equalizer and the pulse shaping (root raised cosine) filters (one each for receive and
transmit) are taken into account. This amount of delay could have an impact on ATSC legacy
receivers. The baseband equalization DOCR allows retransmission of a clean signal without the
lengthy delays inherent in the baseband decode/regeneration method of Figure 3.6.4. It can also
transmit at higher power than the RF and IF processing DOCRs. The shortcoming of this method
is that since there is not complete error correction, any errors that occur in interpretation of the
received data will be built into the retransmitted signal. This makes it important in designing the
EDOCR installation to include sufficient receiving antenna gain and received signal level that
p
errors are minimized in the absence of error correction. If a fairly clean received signal cannot be
obtained, it may be better to use the RF or IF processing DOCR and allow viewers’ receivers to
apply full error correction to the relayed signal rather than retransmitting processed signals that
contain data errors.
All of the forms of DOCR described have a number of limitations in common. Most of the
limitations arise from the facts that a DOCR receives and re-transmits on the same frequency and
that obtaining good isolation between transmitting and receiving antennas is a very difficult
proposition. The result is coupling from the DOCR output back to its input. This coupling leads
to feedback around the amplifiers in the DOCR. Such feedback can result in oscillation in the
cases of the RF and IF processing DOCR. Short of oscillation, it can result in signal distortions
in amplitude and group delay ripple similar to those suffered in a propagation channel. These two
designs will also suffer from the accumulation of noise along the cascade of transmitters and
propagation channels from the signal source to the ultimate receiver. The application of adaptive
equalizers to the feedback path around the DOCR holds some promise to mitigate the distor-
tions, but it cannot help with the noise accumulation.
The feedback around a DOCR puts a limitation on the power that can be transmitted by such a
device. A margin must be provided to keep the system well below the level of oscillation, and the
point of oscillation will be determined by the isolation between the transmitting and receiving
antennas. All of this tends to make high power level operation of DOCRs problematic.
Similarly, the time delay through a DOCR is significant for network design. As one goes from
design-to-design from Figure 3.6.2 to 3.6.3, 3.6.5, and then 3.6.4, the time delay gets longer. The
time delay determines over what area the combination of signals from the transmitters in the net-
work stay within the capability of the receiver adaptive equalizer to correct the apparent echoes
caused by receiving signals from multiple transmitters. The geometry between the source trans-
mitter, the DOCR, and the receiver determine the delay spread actually seen by the receiver. To
this must be added the delay of the DOCR. Additional delay in the DOCR can only push the
delay spread in the wrong direction (extending pre-echo), further limiting the area in which a
receiver having a given adaptive equalizer capability will find the delay spread within its ability
to correct.
The relative merits of different DOCR configurations are summarized in Table 3.6.2.
3.6.2b Distributed Transmission Systems

Distributed transmission networks differ from DOCRs in that each transmitter in the network is
fed over a link separate from the on-air signal delivered to viewers (although this may be over a
different broadcast channel in the case of distributed translators) [1]. This separation of the deliv-
ery and emission channels provides complete flexibility in the design of individual transmitters
and the network, subject only to the limitations of consumer receiver adaptive equalizer capabili-
ties. From the standpoint of the network and its transmitters alone, any power level and any rela-
tive emission timing are possible. The important requirement is to assure that all transmitters in
the network emit the same symbols for the same data delivered to their inputs and at the same
time (plus whatever time offsets may be designed into the network).
There are a number of ways in which signals in a distributed transmission network could be
delivered to the several transmitters. These include methods in which fully modulated signals are
delivered to the various transmitter locations and either up-converted or used to control the re-
modulation of the data onto a new signal, a method in which data representative of the symbols
p
Table 3.6.2 Performance Comparison of Different DOCR (After [1].)
DOCR Configurations RF Processing IF Processing Baseband Baseband Decode/

DOCR DOCR Equalization DOCR Regeneration DOCR
Complexity/cost Very low Low High High
DOCR internal delay Few tenths μs Around 1 – 2 μs Several μs mS level, beyond
(extending pre-echo) very small small medium equalizer range
Input adjacent channel Very weak Medium High High
suppression capability
Output power level Low Low Moderate Moderate
Multihop noise/error accu-Yes Yes Maybe No
mulation
to be broadcast are distributed to the transmitters, and a method in which standard MPEG-2
transport stream packets are delivered to the transmitters together with the necessary information
to allow the various transmitters to emit their signals in synchronization with one another. It is
the latter method that has been selected for use in ATSC Standard A/110 [2].
In the method documented in A/110, the data streams delivered to the transmitters are the
same as now delivered over standard STLs for the single transmitters currently in use, with infor-
mation added to those streams to allow synchronizing the transmitters. This method utilizes a
very small portion of the capacity of the channel but allows continued use of the entire existing
infrastructure designed and built around the 19.39 Mbps data rate. This technique permits com-
plete flexibility in setting the power levels and relative emission timing of the transmitters in a
network while assuring that they emit the same symbols for the same data inputs. While origi-
nally intended for use in SFNs, the selected method also permits extension to multiple frequency
networks (MFNs), using a second broadcast channel as an STL to deliver the data stream to mul-
tiple distributed translators that themselves operate in an SFN. Various combinations of distrib-
uted transmitters and distributed translators are possible, and, in some cases, whether a given
configuration constitutes an SFN or an MFN will depend only upon whether there are viewers in
a position to be able to receive the signals that are also relaying the data stream to successive
transmitters in the network.
3.6.3 Distributed Transmission Architecture

At the most basic level, the application of a distributed transmission (DTx) network can be
divided into two categories [2]:
• Simple, where a second transmitter is added to a main system.
• Complex, where multiple (three or more) transmitters are used in the system. Within the
“complex” realm, systems can be further distinguished by “cell size.”
p
3.6.3a Simple DTx Network

A DTx network has many advantages when compared to on-channel repeaters and translators
[2]. In a DTx network, the digital bit stream is distributed over a separate path to each transmitter
in the network. This eliminates both any errors that might have arisen using over-the-air distribu-
tion and the need for isolation between the transmitter and the over-the-air receiver. The result is
that higher output power can be achieved. The transmitters also are synchronized, which allows
control of the time delay between the two transmitters. Moreover, a second broadcast channel is
not required, as it is when translators are used.
The major limitations of DTx networks are the network designs necessary to minimize
mutual interference to consumer receivers (network internal interference) generated by the multi-
ple transmitters and the limitations created by the operating ranges of consumer receiver adaptive
equalizers. The impact of these limitations is in the distance by which the transmitters in a net-
work can be separated from one another while still permitting the best possible improvement in
reception.
The application of DTx networks can be illustrated by two examples related to terrain and
coverage. In the first example, coverage is improved in terrain-shielded areas within a station's
noise limited contour (NLC). This use of a DTx network recognizes known coverage limitations
by locating a distributed transmitter where it best can fill in an area receiving low signal levels
because of terrain shielding. Because the distributed transmitters operate on the same channel,
the spectrum allocations limitations on the use of translators do not preclude such service
improvements.
The second example is the use of a second transmitter to maximize service in areas beyond
the NLC of a single transmitter. In this case, the population to be served cannot receive the signal
from the first transmitter, and the second transmitter is located so as to extend service to areas
that otherwise could not be reached.
3.6.3b Complex DTx Network

Complex DTx Networks consist of multiple distributed transmitters and can use all of the con-
cepts of the simple types discussed previously [2]. Such systems have the ability to distribute
power more uniformly throughout a total service area and to provide reception at any location
from multiple directions. Three concepts are closely related to the use of complex, multiple-
transmitter applications: large cell systems, small cell systems, and micro-cell systems.
Large Cell System

Distributed transmission networks that use powerful transmitters on tall towers to cover large
areas are called large cell systems [2]. Such systems might have, for example, five cells each
covering its own service area with a radius of 20–30 miles (30–50 km). The service areas of the
individual cell transmitters would have large areas of overlap, and the network would require
careful design to minimize the effects of delay spread on receivers in the overlap regions. Con-
sideration of transmitting antenna patterns, antenna elevations, transmitter power levels, and
transmitter timing would be required. As DTV receiver adaptive equalizer designs improve over
time, it can be expected that the constraints required on large cell designs and the difficulty of
their implementation will be eased by the improved ability of receivers to handle the longer delay
spreads that result from large cell networks.
p
Small Cell System

Distributed transmission networks that use lower power transmitters on shorter towers to cover
smaller regions are called small cell systems [2]. Such systems might have, for example, a dozen
cells, each covering its own service area with a radius of 5–10 miles (8–16 km). The service
areas of the individual cell transmitters would have relatively small areas of overlap, and the net-
work would allow a simpler design while still minimizing the effects of delay spread on receivers
in the overlap regions. Transmitting antenna patterns, antenna elevations, transmitter power lev-
els, and transmitter timing would be designed to adjust the service areas to closely match the
populations intended to be served. As DTV receiver adaptive equalizer designs improve over
time, it can be expected that small cell designs can be further simplified due to the improved
ability of receivers to handle the reception of multiple signals that result from small cell net-
works.
Micro-Cell System
Micro-cell systems use very low power transmitters to cover very small areas [2]. They may be
intermixed with either large cell or small cell networks to fill in coverage in places like city can-
yons, tunnels, or small valleys. Cities with tall buildings that emulate the effect of mountains and
valleys with river gorges (e.g., New York and Chicago) can benefit from the use of microcells.
Design of micro-cell systems will require the use of different design tools than typically used for
broadcast applications—for example, those used for design of cellular telephone networks as
opposed to the Longley-Rice methods used for longer range propagation modeling.
Ultimately, the application of distributed transmission networks can use any or all of these
concepts, integrated together in a master plan. This will allow evolution from a single, high
power, tall tower system to a hybrid system based upon combinations of these concepts and other
unique solutions created for each individual broadcast market.
3.6.3c DTx Synchronization Mechanisms

The synchronization requirements outlined in previously lead to the reference top-level system
configuration shown in Figure 3.6.6 [2]. The system comprises three elements:
• An external time and frequency reference (shown as GPS)
• A distributed transmission adapter (DTxA) situated at the source end of the distribution (stu-
dio-to-transmitter link, STL) subsystem
• A slave synchronization subsystem included in each of the transmitters.
The heavy lines in Figure 3.6.6 show the paths taken by synchronization signals generated in
the DTxA, and a bar across the top of the figure shows the area of applicability of ATSC stan-
dard A/110.
Distributed Transmission Adapter at Source

The distributed transmission adapter is used to create a pair of synchronization signals that are
multiplexed into the transport stream prior to distribution over the STL system [2]. The signals
produced by the DTxA are a cadence signal (CS), which establishes the phase of the data frames
relative to the TS packets, and a distributed transmission packet (DTxP), which carries informa-
Application of Distributed Transmission Standard
RF Transmission System RF Transmission System

GPS
Video Subsystem Service Multiplex and Transport

Video
Distributed Transmission Adapter GPS GPS
Video Souce Coding
Freq
and Compression
Data
Processing
p
Model
Audio Subsystem
Consumer Receiver
Audio
Service Time
Audio Souce Coding
Multiplexer
and Compression Transmitter
Freq Freq
Synchronization
Inserter Signal Processing Signal Processing

& &
Power Amplification Power Amplification

Ancillary Data
Digital Time Time
Transport
Control Data Synchronized Synchronized
Data Data
Processing Processing
Figure 3.6.6 Synchronized DTV transmitter block diagram. (From [2]. Used with permission.)
p
tion for slaving the pre-coders and trellis coders in the transmitters and carries command infor-
mation specifying the necessary time offset for each transmitter. In addition, the DTxA indicates
operating mode to the transmitters and provides information to be transmitted in the data field
sync data segment through a field rate side channel, which carries information updated regularly
at a data field rate. To accomplish these functions, the DTxA includes a data processing model
equivalent to the data processing subsection of an A/53 modulator to serve as a master reference
to which the slave synchronizers at the transmitters are slaved.
Slave Synchronization of Transmitters

At each transmitter, a slave synchronizer is employed to capture the cadence signal and the dis-
tributed transmission packet, to slave the data frame phasing to the cadence signal, and to slave
the pre-coder and trellis encoder to the data in the DTxP [2]. The slave synchronizer extracts
mode information from the field rate side channel to set the transmitter to the desired mode. It
also extracts time offset command information addressed to its associated transmitter and uses it
to adjust the emission time of the output symbols.
External Time and Frequency Reference

A common time and frequency reference (such as GPS) is required at several locations in the
system [2]. The time component of the external reference is used by the distributed transmission
adapter to produce the time-offset information to be sent to the slave synchronizers to adjust the
emission times of their associated transmitters. The DTxA uses the frequency component to pre-
cisely maintain its output transportsStream data rate to tight tolerances. The time component also
is used by the slave synchronizers at the transmitters to adjust the emission times of the associ-
ated signals to the time offsets sent from the distributed transmission adapter. The frequency
component also is used by the slave synchronizers at the transmitters to precisely set the frequen-
cies of the transmitters in order to minimize the apparent creation of Doppler shift and the conse-
quent burdening of receiver adaptive equalizers by frequency differences between transmitters.
The slave synchronizers can also use the frequency component to reestablish the precise bit rate,
and thereby stabilize the timing, of the transport stream after its transmission through STLs hav-
ing some amount of time variation in their delivery of the signal, as in satellite relay, some over-
the-air receivers, and some microwave systems.
3.6.4 Multiple Frequency Network Concepts

A multiple frequency network (MFN) uses more than one channel for transmission [1]. In the
purest case, for N transmitters, N channels are used. But where distributed transmission technol-
ogy is being applied, channels may be shared among a number of transmitters. For N transmit-
ters, the number of channels used will be less than N. Some of the transmitters will be
synchronized, operating on the same channel. In this situation, the network is actually a hybrid of
multiple frequency and single frequency techniques. This classification (hybrid, multiple fre-
quency, distributed transmission networks, where some transmitters are synchronized) is the sub-
ject of this section.
p
3.6.4a Translators
A translator is part of a multiple frequency network [1]. It receives an off-air signal on one chan-
nel and retransmits it on another channel.
Even in some relatively unpopulated areas, especially where NTSC translator systems are
already deployed, there are not enough additional channels to accommodate traditional translator
networks for ATSC signals during the transition phase. In these situations, use of distributed
translators allows ATSC translator systems to be built using fewer channels.
A distributed translator system applies distributed transmission technology to create a net-
work of synchronized transmitters on one channel, which retransmits a signal received off-the-air
from a main transmitter, distributed transmitter, another translator, or a translator network. The
advantages of a distributed translator system over conventional translators include signal regen-
eration and conservation of spectrum.
The distributed transmission system initially was designed to use STLs to convey MPEG-2
transport stream streams to slave transmitters in a network. This certainly could be done for dis-
tributed translator systems, but where an off-air signal is available, use of STLs would be costly
and redundant. The advantage of a distributed translator system over a distributed transmission
system is that STLs are not required—the signal may be taken off the air.
3.6.5 Implementation Considerations

A generalized distributed transmission system may include different combinations of DTx and
DOCR elements [1]. Nothing inherently prohibits using one in the presence of another. System
architects may tailor the deployment of the various technologies to the particular system being
designed, considering such issues as power, cost, and terrain shielding.
For example, where low power is adequate, particularly where there is terrain shielding, a
DOCR may be the most cost-effective solution. If higher power is necessary then a distributed
transmission translator might be better.
3.6.6 References
1. ATSC, “ATSC Recommended Practice: Design Of Synchronized Multiple Transmitter Net-
works,” Advanced Television Systems Committee, Washington, D.C., Doc. A/111, Septem-
ber 3, 2004.
2. ATSC: “Synchronization Standard for Distributed Transmission,” Advanced Television
Systems Committee, Washington, D.C., Doc. A/110, July 14, 2004.
g g
Chapter
3.7
DTV Satellite Transmission
3.7.1 Introduction1
In recognition of the importance of satellite transmission in the distribution of digital video and
audio programming, the ATSC developed two standards optimized for use with the ATSC digital
television suite of standards. The first, document A/80, defines the parameters necessary to
transmit DTV signals (transport, video, audio, and data) over satellite to one or more production
and/or transmission centers. Although the ATSC had identified the particulars for terrestrial
transmission of DTV (using FEC-encoded 8VSB modulation), this method is inappropriate for
satellite transmission because of the many differences between the terrestrial and satellite trans-
mission environments. The second standard, document A/81, describes a direct-to-home satellite
transmission system intended for distribution of programming directly to consumers.
3.7.2 ATSC A/80 Standard

The ATSC, in document A/80, defined a standard for modulation and coding of data delivered
over satellite for digital television applications [1]. These data can be a collection of program
material including video, audio, data, multimedia, or other material generated in a digital format.
They include digital multiplex bit streams constructed in accordance with ISO/IEC 13818-1
(MPEG-2 systems), but are not limited to these. The standard includes provisions for arbitrary
types of data as well.
Document A/80 covers the transformation of data using error correction, signal mapping, and
modulation to produce a digital carrier suitable for satellite transmission. In particular, quadra-
ture phase shift modulation (QPSK), eight phase shift modulation (8PSK), and 16 quadrature
amplitude modulation (16QAM) schemes are specified. The main distinction between QPSK,
8PSK, and 16QAM is the amount of bandwidth and power required for transmission. Generally,
1. Editor’s note: This chapter provides an overview of ATSC DTV satellite services based on
ATSC A/80 and A/81. Readers are encouraged to download the entire standards from the
ATSC Web site (http://www.atsc.org). All ATSC Standards, Recommended Practices, and
3-111
for the same data rate, progressively less bandwidth is consumed by QPSK, 8PSK, and 16QAM,
respectively, but the improved bandwidth efficiency is accompanied by an increase in power to
deliver the same level of signal quality.
A second parameter, coding, also influences the amount of bandwidth and power required for
transmission. Coding, or in this instance forward error correction (FEC), adds information to the
data stream that reduces the amount of power required for transmission and improves reconstruc-
tion of the data stream received at the demodulator. While the addition of more correction bits
improves the quality of the received signal, it also consumes more bandwidth in the process. So,
the selection of FEC serves as another tool to balance bandwidth and power in the satellite trans-
mission link. Other parameters exist as well, such as transmit filter shape factor (α), which have
an effect on the overall bandwidth and power efficiency of the system.
System operators optimize the transmission parameters of a satellite link by carefully consid-
ering a number of trade-offs. In a typical scenario for a broadcast network, material is generated
at multiple locations and requires delivery to multiple destinations by transmitting one or more
carriers over satellite, as dictated by the application. Faced with various size antennas, available
satellite bandwidth, satellite power, and a number of other variables, the operator will tailor the
system to efficiently deliver the data payload. The important tools available to the operator for
dealing with this array of system variables include the selection of the modulation, FEC, and α
value for transmission.
3.7.2a Services and Applications

The ATSC satellite transmission standard includes provisions for two distinct types of services
[1]:
• Contribution service—the transmission of programming/data from a programming source to
a broadcast center. Examples include such material as digital satellite news gathering
(DSNG), sports, and special events.
• Distribution service—the transmission of material (programming and/or data) from a broad-
cast center to its affiliate or member stations.
The A/80 standard relies heavily upon previous work done by the Digital Video Broadcasting
(DVB) Project of the European Broadcast Union (EBU) for satellite transmission. Where appli-
cable, the ATSC standard sets forth requirements by reference to those standards, particularly EN
300 421 (QPSK) and prEN 301 210 (QPSK, 8PSK, and 16QAM).
The modulation and coding defined in the standard have mandatory and optional provisions.
QPSK is considered mandatory as a mode of transmission, while 8PSK and 16QAM are
optional. Whether equipment implements optional features is a decision for the manufacturer.
However, when optional features are implemented they must be in accordance with the standard
in order to be compliant with it.
3.7.2b System Overview

A digital satellite transmission system is designed to deliver data from one location to one or
more destinations. A block diagram of a simple system is shown in Figure 3.7.1 [1]. The drawing
depicts a data source and data sink, which might represent a video encoder/multiplexer or
decoder/demultiplexer for ATSC applications, but can also represent a variety of other sources
RF Equipment
Data
Data Stream IF * Up
Modulator HPA
Source Converter
Encoder /
Multiplexer
* Equalization Required
For Some Applications Satellite
Data
Data Stream De- IF *
LNB
Sink modulator
Decoder /
Demultiplexer Down
LNA
Converter
Reference Points
RF Equipment
Figure 3.7.1 Overall system block diagram of a digital satellite system. The ATSC standard
described in document A/80 covers the elements noted by the given reference points. (From [1].
that produce a digital data stream. This particular point, the accommodation of arbitrary data
streams, is a distinguishing feature of standard A/80.
The subject of the A/80 standard is the segment between the dashed lines designated by the
reference points on Figure 3.7.1; it includes the modulator and demodulator. Only the modula-
tion parameters are specified; the receive equipment is designed to recover the transmitted sig-
nal. The ATSC standard does not preclude combining equipment outside the dashed lines with
the modulator or demodulator, but it sets a logical demarcation between functions.
In the figure, the modulator accepts a data stream and operates upon it to generate an interme-
diate frequency (IF) carrier suitable for satellite transmission. The data are acted upon by for-
ward error correction (FEC), interleaving and mapping to QPSK, 8PSK or 16QAM, frequency
conversion, and other operations to generate the IF carrier. The selection of the modulation type
and FEC affects the bandwidth of the IF signal produced by the modulator. Selecting QPSK,
8PSK, or 16QAM consumes successively less bandwidth as the modulation type changes. It is
possible, then, to use less bandwidth for the same data rate or to increase the data rate through
the available bandwidth by altering the modulation type.
Coding or FEC has a similar impact on bandwidth. More powerful coding adds more infor-
mation to the data stream and increases the occupied bandwidth of the IF signal emitted by the
modulator. There are two types of coding applied in the modulator. An outer Reed-Solomon code
is concatenated with an inner convolutional/trellis code to produce error correction capability
exceeding the ability of either coding method used alone. The amount of coding is referred to as
the code rate, quantified by a dimensionless fraction (k/n) where n indicates the number of bits
out of the encoder given k input bits (e.g., rate 1/2 or rate 7/8). The Reed-Solomon code rate is
DATA
STREAM PACKETIZING RS OUTER CONVOLUTIONAL BASEBAND IF
INTERLEAVER
& ENERGY CODER INNER CODER & SHAPING &
(I=12)
DISPERSAL (204,188) MAPPING MODULATION
QPSK (OPTIONAL 8PSK, 16QAM)
Figure 3.7.2 Block diagram of the baseband and modulator subsystem. (From [1]. Used with per-
mission.)
fixed at 204,188 but the inner convolutional/trellis code rate is selectable, offering the opportu-
nity to modify the transmitted IF bandwidth.
One consequence of selecting a more bandwidth-efficient modulation or a higher inner code
rate is an increase in the amount of power required to deliver the same level of performance. The
key measure of power is the Eb /No (energy per useful bit relative to the noise power per Hz), and
the key performance parameter is the bit error rate (BER) delivered at a particular E b /No. For
digital video, a BER of about 10-10 is necessary to produce high-quality video. Thus, noting the
E b /No required to produce a given BER provides a way of comparing modulation and coding
schemes. It also provides a relative measure of the power required from a satellite transponder, at
least for linear transponder operation.
The basic processes applied to the data stream are illustrated in Figure 3.7.2. Specifically,
• Packetizing and energy dispersal
• Reed-Solomon outer coding
• Interleaving
• Convolutional inner coding
• Baseband shaping for modulation
• Modulation
The input to the modulator is a data stream of specified characteristics. The physical and elec-
trical properties of the data interface, however, are outside the scope of this standard. The output
of the modulator is an IF signal that is modulated by the processed input data stream. This is the
signal delivered to RF equipment for transmission to the satellite. Table 3.7.1 lists the primary
system inputs and outputs.
The data stream is the digital input applied to the modulator. There are two types of packet
structures supported by the standard, as given in Table 3.7.2.
3.7.3 ATSC DTH Satellite Broadcast Standard

ATSC Standard A/81 describes the emission system for a direct-to-home (DTH) satellite broad-
cast system [2]. This specification defines extensions to the audio, video, transport, and PSIP
subsystems as defined in ATSC standards A/52, A/53, and A/65. The emission system defined in
this document includes carriage of data broadcasting as defined in ATSC standard A/90, without
Table 3.7.1 System Interfaces Specified in A/80 (After [1].)
Location System Inputs/Outputs Type Connection

Transmit station Input MPEG-2 transport (Note 1) or From MPEG-2 multiplexer or
arbitrary other device
Output 70/140 MHz IF, L-band IF, RF To RF devices
(Note 2)
Receive installation Input 70/140 MHz IF, L-band IF (Note 2) From RF devices
Output MPEG-2 transport (Note 1) or To MPEG-2 de-multiplexer or
arbitrary other device
1 In accordance with ISO/IEC 13838-1
2 The IF bandwidth may impose a limitation on the maximum symbol rate.
Table 3.7.2 Input Data Stream Structures (After [1].)
Type Description
1 The packet structure shall be a constant rate MPEG-2 transport per ISO/IEC 13818-1 (188 or 204 bytes per
packet including 0x47 sync, MSB first).
2 The input shall be a constant rate data stream that is arbitrary. In this case, the modulator takes successive
187 byte portions from this stream and prepends a 0x47 sync byte to each portion, to create a 188 byte
MPEG-2 like packet. (The demodulator will remove this packetization so as to deliver the original, arbitrary
stream at the demodulator output.)
requiring extensions. Transmission and conditional access subsystems are not defined, allowing
service providers to use existing subsystems.
The ATSC DTH satellite broadcast system consists of two major elements:
• The transmission system
• Integrated receiver decoder, commonly referred as a set top box (STB).
3.7.3a Transmission System

The transmission system comprises an emission multiplex, a modulator/encoder, and a transmit-
ter [2]. The emission multiplexer requirements are discussed in this standard. Specifications for
the modulator/encoder and the transmitter are left for service provider to develop.
Figure 3.7.3 shows a functional block diagram of a DTH satellite transmission system. The
emission multiplexer accepts and combines:
• ATSC multi-program transport streams (A/53, A/65, A/70, and A/90 protocols) from differ-
ent sources
• Satellite extensions to PSIP
Additionally, the emission multiplexer may accept:
• MPEG-compliant (non-ATSC) transport streams
• Data streams such as A/90 and DVB data broadcast
Note : These stream s

delivered by ASI, G .703,
AT M, 10BT /100BT A/81
Ethernet. Com pliant
AT SC Com pliant Transport Modulator
Multiprogram
ASI
T ransport Stream
Broadcast Netw ork
Em ission Mux
ATSC Com pliant A/81
M ultiprogram Com pliant
T ransport Stream T ransport
ASI
Modulator
O ptional Transport
Stream s
PSIP w ith
Satellite
Extensions
Data
Figure 3.7.3 Overview of the DTH satellite transmission system. (From [2]. Used with permission.)
Each multi-program transport stream output from the emission multiplexer to a modulator must
conform with:
• The transport, audio, and video format extensions defined for satellite delivery in standard A/
81.
• System information with all the normative elements from the ATSC PSIP standard (A/65)
and satellite extensions, such as the satellite virtual channel table (defined in A/81).
Transport streams at the output of the emission multiplex may also carry additional informa-
tion to support delivery of system-specific data (such as DVB-SI, ATSC A/56, control data, EIA-
608B captions using ANSI/SCTE 20, and MPEG-1 Layer 2 audio).
3.7.3b Integrated Receiver Decoder System

A functional block diagram of an integrated receiver decoder (IRD) system is depicted in Figure
3.7.4 [2]. This system demodulates and decodes audio, video, and data streams compatible with
the transmission system described in A/81.
Tuner Demod Demux Descrambler Video

Audio
Data
Control CA
PSIP
Figure 3.7.4 Functional block diagram of an IRD system. (From [2]. Used with permission.)
3.7.3c Compression Format Constraints

The allowed compression formats are listed in Table 3.7.3, and for 25/50 Hz video, Table 3 in
ATSC A/63 (see [3]).
3.7.3d PSIP
Some Program and System Information Protocol (PSIP) tables used in the DTH satellite broad-
cast standard are common with terrestrial broadcast and/or cable systems defined in ATSC docu-
ment A/65 [2].
The following tables must be included in all ATSC-compliant transport streams to be trans-
mitted via satellite broadcast:
• The satellite virtual channel table (SVCT) defining, at a minimum, the virtual channel struc-
ture for the collection of MPEG-2 programs embedded in the transport stream in which the
SVCT is carried.
• The master guide table (MGT) defining the type, packet identifiers, and versions for all of the
other satellite PSIP tables included in the transport stream, except for the system time table
(STT).
• The rating region table (RRT) defining the TV parental guideline system referenced by any
content advisory descriptor carried within the transport stream.
Table 3.7.3 Allowed Compression Formats Under ATSC DTH Satellite Standard (After [2].)
vertical_size_value horizontal_size_value aspect_ratio_information frame_rate_code Progressive/
Interlaced
1080 1280 3 1,2,4,5,7,8 P
1080 1280 3 4,5,7,8 I
1080 1920 1, 3 1, 2, 4, 5,7,8 P
1080 1920 1, 3 4, 5,7,8 I
1080 1440 3 1, 2, 4, 5,7,8 P
1080 1440 3 4, 5,7,8 I
720 1280 1, 3 1, 2, 4, 5, 7, 8 P
480 720 2, 3 1, 2, 4, 5, 7, 8 P
480 720 2, 3 4, 5 I
480 704 2, 3 1, 2, 4, 5, 7, 8 P
480 704 2, 3 4, 5 I
480 640 1, 2 1, 2, 4, 5, 7, 8 P
480 640 1, 2 4, 5 I
480 544 2 1 P
480 544 2 4 I
480 480 2 4,5 I
480 528 2 1 P
480 528 2 4 I
480 352 2 1 P
480 352 2 4 I
Legend for MPEG-2 Coded Values
aspect_ratio_information: 1 = square samples, 2 = 4:3 display aspect ratio, 3 = 16:9 display aspect ratio
frame_rate_code: 1 = 23.976 Hz, 2 = 24 Hz, 4 = 29.97 Hz, 5 = 30 Hz, 7 = 59.94 Hz, 8 = 60 Hz
Progressive/Interlace: I= interlaced scan, P = progressive scan
• The system time table (STT), defining the current date and time of day and daylight savings
time transition timing.
• The first four aggregate event information tables (AEIT-0, AEIT-1, AEIT-2, and AEIT-3),
which deliver event title and schedule information that may be used to support an electronic
program guide application.
Satellite Virtual Channel Table (SVCT)

The satellite virtual channel table (SVCT), like its cable and terrestrial broadcast counterparts,
contains a list of attributes for virtual channels carried in the transport stream [2]. Any changes
in the virtual channel structure are conveyed with a new “version number.” The basic informa-
tion contained in the body of the SVCT includes:
• Transport stream ID
• Major and minor channel number
• Short channel name
• Carrier frequency
• Program number
• Location field for extended text messages

• Service type
• Modulation parameters
Additional information may be carried by descriptors, which may be placed in the descriptor
loop after the basic information.
Unlike cable and terrestrial applications where just one virtual channel table is present in any
given transport stream, in the satellite application more than one may be present. A receiver is
expected to capture and record one or more SVCTs for use in navigation and service acquisition.
3.7.4 References
1. ATSC Standard: “Modulation And Coding Requirements For Digital TV (DTV) Applica-
tions Over Satellite,” Doc. A/80, Advanced Television Systems Committee, Washington,
D.C., July, 17, 1999.
2. ATSC Standard: “Direct-to-Home Satellite Broadcast Standard,” Doc. A/81, Advanced
Television Systems Committee, Washington, D.C., 2003.
3. ATSC Standard: “Standard for Coding 25/50 Hz Video,” Doc. A/63, Advanced Television
Systems Committee, Washington, D.C., May 2, 1997.
g g
3.8.Chapter
3.8
The DVB Standard
3.8.1 Introduction
The roots of European digital television can be traced back to the early 1980s, when development
work began on multiplexed analog component (MAC) hardware. Designed from the outset as a
direct broadcast satellite (DBS) service, MAC went through a number of infant stages, including
A, B, C, and E; MAC finally reached maturity in its D and D2 form. High-definition versions of
the transmission formats also were developed with an eye toward European-based HDTV pro-
gramming.
3.8.2 Digital Video Broadcasting (DVB)

In the early 1990s, it was becoming clear that the once state-of-the-art MAC systems would have
to give way to all-digital technology. DVB provided a forum for gathering all the major European
television interests into one group to address the issue [1]. The DVB project promised to develop
a complete digital television system based on a unified approach.
As the DVB effort was taking shape, it was clear that digital satellite and cable television
would provide the first broadcast digital services. Fewer technical problems and a simpler regu-
latory climate meant that these new technologies could develop more rapidly than terrestrial sys-
tems. Market priorities dictated that digital satellite and cable broadcasting systems would have
to be developed rapidly. Terrestrial broadcasting would follow later.
From the beginning, the DVB effort was aimed primarily at the delivery of digital video sig-
nals to consumers. Unlike the 1125-line Japanese HDTV system and the European HDTV
efforts that preceded DVB, the system was not envisioned primarily as a production tool. Still,
the role that the DVB effort played—and continues to play—in the production arena is undeni-
able. The various DVB implementations will be examined in this chapter.
3.8.2a Technical Background of the DVB System

From the outset, it was clear that the sound- and picture-coding systems of ISO/IEC MPEG-2
should form the audio- and image-coding foundations of the DVB system [1]. DVB would need
3-121
to add to the MPEG transport stream the necessary elements to bring digital television to the
home through cable, satellite, and terrestrial broadcast systems. Interactive television also was
examined to see how DVB might fit into such a framework for new video services.
MPEG-2
MPEG-2 specifies a data-stream syntax, and the system designer is given a “toolbox” from
which to make up systems incorporating greater or lesser degrees of sophistication [1]. In this
way, services avoid being overengineered, yet are able to respond fully to market requirements
and are capable of evolution.
The sound-coding system specified for all DVB applications is the MPEG audio standard
MPEG Layer II (MUSICAM), which is an audio coding system used for many audio products
and services throughout the world. MPEG Layer II takes advantage of the fact that a given sound
element will have a masking effect on lower-level sounds (or on noise) at nearby frequencies.
This is used to facilitate the coding of audio at low data rates. Sound elements that are present,
but would not be heard even if reproduced faithfully, are not coded. The MPEG Layer II system
can achieve a sound quality that is, subjectively, very close to the compact disc. The system can
be used for mono, stereo, or multilingual sound, and (in later versions) surround sound.
The first users of DVB digital satellite and cable services planned to broadcast signals up to
and including MPEG-2 Main Profile at Main Level, thus forming the basis for first-generation
European DVB receivers. Service providers, thus, were able to offer programs giving up to “625-
line studio quality” (ITU-R Rec. 601), with either a 4:3, 16:9, or 20:9 aspect ratio.
Having chosen a given MPEG-2 compliance point, the service provider also must decide on
the operating bit rates (variable or constant) that will be used. In general, the higher the bit rate,
the greater the proportion of transmitted pictures that are free of coding artifacts. Nevertheless,
the law of diminishing returns applies, so the relationship of bit rate to picture quality was given
careful consideration.
To complicate the choice, MPEG-2 encoder design has a major impact on receiver picture
quality. In effect, the MPEG-2 specification describes only syntax laws, thus leaving room for
technical-quality improvements in the encoder. Early tests by DVB partners established the
approximate relationship between bit rate and picture quality for the Main Profile/Main Level,
on the basis of readily available encoding technology. These tests suggested the following:
• To comply with ITU-R Rec. 601 “studio quality” on all material, a bit rate of up to approxi-
mately 9 Mbits/s was required.
• To match current “NTSC/PAL/SECAM quality” on most television material, a bit rate of 2.5
to 6 Mbits/s was required, depending upon the program material.
• Film material (24/25 pictures/s) is easier to code than scenes shot with a video camera, and it
also looked good at lower bit rates.
MPEG-2 Data Packets

MPEG-2 data packets are the basic building blocks of the DVB system [1]. The data packets are
fixed-length containers with 188 bytes of data each. MPEG includes program-specific informa-
tion (PSI) so that the MPEG-2 decoder can capture and decode the packet structure. This data,
transmitted with the pictures and sound, automatically configures the decoder and provides the
synchronization information necessary for the decoder to produce a complete video signal.
MPEG-2 also allows a separate service information (SI) system to complement the PSI.
3.8.2b DVB Services

DVB has incorporated an open service information system to accompany the DVB signals, which
can be used by the decoder and the user to navigate through an array of services offered. The fol-
lowing sections detail the major offerings.
DVB-SI
As envisioned by the system planners of DVB, the viewer of the future will be capable of receiv-
ing a multitude (perhaps hundreds) of channels via the DVB integrated receiver decoder (IRD)
[1]. These services could range from interactive television to near video-on-demand to special-
ized programming. To sort out the available offerings, the DVB-SI provides the elements neces-
sary for the development of an electronic program guide (EPG)—a feature which, it was
believed, would become an important part of new digital television services.
Key data necessary for the DVB IRD to automatically configure itself is provided for in the
MPEG-2 PSI. DVB-SI adds information that enables DVB IRDs to automatically tune to partic-
ular services and allows services to be grouped into categories with relevant schedule informa-
tion. Other information provided includes:
• Program start time
• Name of the service provider
• Classification of the event (sports, news, entertainment, and so on)
DVB-SI is based on four tables, plus a series of optional tables. Each table contains descriptors
outlining the characteristics of the services/event being described. The tables are:
• Network information table (NIT). The NIT groups together services belonging to a particular
network provider. It contains all of the tuning information that might be used during the setup
of an IRD. It also is used to signal a change in the tuning information.
• Service description table (SDT). The SDT lists the names and other parameters associated
with each service in a particular MPEG multiplex.
• Event information table (EIT). The EIT is used to transmit information relating to all the
events that occur or will occur in the MPEG multiplex. The table contains information about
the current transport and optionally covers other transport streams that the IRD can receive.
• Time and date table (TDT). The TDT is used to update the IRD internal clock.
In addition, there are three optional SI tables:
• Bouquet association table (BAT). The BAT provides a means of grouping services that might
be used as one way an IRD presents the available services to the viewer. A particular service
can belong to one or more “bouquets.”
• Running status table (RST). The sections of the RST are used to rapidly update the running
status of one or more events. The running status sections are sent out only once—at the time
the status of an event changes. The other SI tables normally are repetitively transmitted.
• Stuffing table (ST). The ST may be used to replace or invalidate either a subtable or a com-
plete SI table.
With these tools, DVB-SI covers the range of practical scenarios, facilitating a seamless tran-
sition between satellite and cable networks, near video-on-demand, and other operational config-
urations.
DVB-S
DVB-S is a satellite-based delivery system designed to operate within a range of transponder
bandwidths (26 to 72 MHz) accommodated by European satellites such as the Astra series,
Eutelsat series, Hispasat, Telecom series, Tele-X, Thor, TDF-1 and 2, and DFS [1].
DVB-S is a single carrier system, with the payload at its core. Surrounding this core are a
series of layers intended not only to make the signal less sensitive to errors, but also to arrange
the payload in a form suitable for broadcasting. The video, audio, and other data are inserted into
fixed-length MPEG transport-stream packets. This packetized data constitutes the payload. A
number of processing stages follow:
• The data is formed into a regular structure by inverting synchronization bytes every eighth
packet header.
• The contents are randomized.
• Reed-Solomon forward error correction (FEC) overhead is added to the packet data. This sys-
tem, which adds less than 12 percent overhead to the signal, is known as the outer code. All
delivery systems have a common outer code.
• Convolutional interleaving is applied to the packet contents.
• Another error-correction system, which uses a punctured convolutional code, is added. This
second error-correction system, the inner code, can be adjusted (in the amount of overhead)
to suit the needs of the service provider.
• The signal modulates the satellite broadcast carrier using quadrature phase-shift keying
(QPSK).
In essence, between the multiplexing and the physical transmission, the system is tailored to the
specific channel properties. The system is arranged to adapt to the error characteristics of the
channel. Burst errors are randomized, and two layers of forward error correction are added. The
second level (inner code) can be adjusted to suit the operational circumstances (power, dish size,
bit rate available, and other parameters).
DVB-C
The cable network system, known as DVB-C, has the same core properties as the satellite sys-
tem, but the modulation is based on quadrature amplitude modulation (QAM) rather than QPSK,
and no inner-code forward error correction is used [1]. The system is centered on 64-QAM, but
lower-level systems, such as 16-QAM and 32-QAM, also can be used. In each case, the data
capacity of the system is traded against robustness of the data. In terms of capacity, an 8 MHz
channel can accommodate a payload capacity of 38.5 Mbits/s if 64-QAM is used, without spill-
over into adjacent channels.
DVB-MC
The DVB-MC digital multipoint distribution system uses microwave frequencies below approxi-
mately 10 GHz for direct distribution to viewers' homes [1]. Because DVB-MC is based on the
DVB-C cable delivery system, it enables a common receiver to be used for both cable and micro-
wave transmissions.
DVB-MS
The DVB-MS digital multipoint distribution system uses microwave frequencies above approxi-
mately 10 GHz for direct distribution to viewers' homes [1]. Because this system is based on the
DVB-S satellite delivery system, DVB-MS signals can be received by DVB-S satellite receivers.
The receiver must be equipped with a small microwave multipoint distribution system (MMDS)
frequency converter, rather than a satellite dish.
DVB-T
DVB-T is the system specification for the terrestrial broadcasting of digital television signals [1].
DVB-T was approved by the DVB Steering Board in December 1995. This work was based on a
set of user requirements produced by the Terrestrial Commercial Module of the DVB project.
DVB members contributed to the technical development of DVB-T through the DTTV-SA (Dig-
ital Terrestrial Television—Systems Aspects) of the Technical Module. The European Projects
SPECTRE, STERNE, HD-DIVINE, HDTVT, dTTb, and several other organizations developed
system hardware and produced test results that were fed back to DTTV-SA.
As with the other DVB standards, MPEG-2 audio and video coding forms the payload of
DVB-T. Other elements of the specification include:
• A transmission scheme based on orthogonal frequency-division multiplexing (OFDM), which
allows for the use of either 1705 carriers (usually known as 2k), or 6817 carriers (8k). Concat-
enated error correction is used. The 2k mode is suitable for single-transmitter operation and
for relatively small single-frequency networks with limited transmitter power. The 8k mode
can be used both for single-transmitter operation and for large-area single-frequency net-
works. The guard interval is selectable.
• Reed-Solomon outer coding and outer convolutional interleaving are used, as with the other
DVB standards.
• The inner coding (punctured convolutional code) is the same as that used for DVB-S.
• The data carriers in the coded orthogonal frequency-division multiplexing (COFDM) frame
can use QPSK and different levels of QAM modulation and code rates to trade bits for rug-
gedness.
• Two-level hierarchical channel coding and modulation can be used, but hierarchical source
coding is not used. The latter was deemed unnecessary by the DVB group because its benefits
did not justify the extra receiver complexity that was involved.
• The modulation system combines OFDM with QPSK/QAM. OFDM uses a large number of
carriers that spread the information content of the signal. Used successfully in DAB (digital
audio broadcasting), OFDM’s major advantage is its resistance to multipath.
Improved multipath immunity is obtained through the use of a guard interval, which is a por-
tion of the digital signal given away for echo resistance. This guard interval reduces the transmis-
sion capacity of OFDM systems. However, the greater the number of OFDM carriers provided,
for a given maximum echo time delay, the less transmission capacity is lost. But, certainly, a
tradeoff is involved. Simply increasing the number of carriers has a significant, detrimental
impact on receiver complexity and on phase-noise sensitivity.
Because of the favorable multipath performance of OFDM, it is possible to operate an over-
lapping network of transmitting stations with a single frequency. In the areas of overlap, the
weaker of the two received signals is similar to an echo signal. However, if the two transmitters
are far apart, causing a large time delay between the two signals, the system will require a large
guard interval.
The potential exists for three different operating environments for digital terrestrial television
in Europe:
• Broadcasting on a currently unused channel, such as an adjacent channel, or broadcasting on
a clear channel
• Broadcasting in a small-area single-frequency network (SFN)
• Broadcasting in a large-area SFN
One of the main challenges for DVB-T developers was that different operating environments
lead to somewhat different optimum OFDM systems. The common 2k/8k specification was
developed to offer solutions for most operating environments.
3.8.2c DVB and the ATSC DTV System

As part of the ATSC DTV specification package, a recommended practice was developed for use
of the ATSC standard to ensure interoperability internationally at the transport level with the
European DVB project (as standardized by the European Telecommunications Standards Insti-
tute, ETSI). Guidelines for use of the DTV standard are outlined to prevent conflicts with DVB
transport in the areas of packet identifier (PID) usage and assignment of user private values for
descriptor tags and table identifiers.
Adherence to these recommendations makes possible the simultaneous carriage of system/
service information (SI) conforming to both the ATSC standard (ATSC A/65) and the ETSI
ETS-300-486 standard. Such dual carriage of SI may be necessary when transport streams con-
forming to the ATSC standard are made available to receivers supporting only the DVB service
information standard, or when transport streams conforming to the DVB standard are made
available to receivers supporting only the ATSC SI standard.
3.8.3 COFDM Technical Principles

The theory of orthogonal frequency division multiplexing (OFDM) has been well known for
some time. The concept of using parallel data transmission and frequency division multiplexing
was first published in the mid-1960s, and a U.S. patent was filed and issued for this approach in
19701. Early studies suggested that such systems might provide some advantages over single-
carrier systems. In particular, multicarrier systems would allow operation of single frequency
networks and could provide improved performance under multipath distortion.
k=0 k=N
Figure 3.8.1 Frequency domain representation of orthogonal carriers. (From [2]. Used with per-
mission.)
3.8.3a Implementation Considerations

Orthogonal frequency demodulation multiplexing is a wideband modulation technique that uses
multiple orthogonal carriers, each modulated in phase and amplitude [2]. OFDM transmission
has a slow signalling rate; the long symbol duration is the key property that makes its application
attractive for terrestrial broadcasting.
For DVB-T transmissions, the OFDM modulator collects 211 or 213 code words into a single
OFDM symbol. Depending on the modulation depth, one code word consists of 2, 4, or 6 bits.
The code words are mapped onto orthogonal carriers. (See Figure 3.8.1.)
An inverse fast Fourier transform (IFFT) is used to convert the set of resulting modulated fre-
quencies into a time-domain signal. Collecting multiple codewords for concurrent transmission
on multiple subcarriers can be advantageous over transmitting these codewords serially on a sin-
gle carrier. When comparing OFDM with the equivalent single-carrier system, the concurrent
use of N subcarriers causes the signalling rate to slow-down N-times.
The long OFDM symbol duration protects the transmission against short-lived distortions.
Because only a small part of a symbol is interfered, the distortion of the information bearing car-
riers is minimal. To further protect OFDM transmission from distortions, certain provisions are
included in its application for DVB-T, including:
• The first and last carriers in the frequency domain are not modulated. The free parts in the
spectrum enable filtering to protect the OFDM signal from adjacent channel interference.
• In the time-domain, the OFDM symbol is enlarged by inserting a copy of the last part of the
symbol in front of it. This guard-interval protects the OFDM symbol from inter-symbol inter-
ference (ISI), caused by echoes with a distance smaller than the guard-interval length. Within
the “stretched” time-domain signal there is always a complete OFDM symbol available with-
out distortion from previous or consecutive symbols. The “zero-ISI” component of the
enlarged symbol is only influenced by an echo of itself. This results in fading and amplifica-
tion of its sub-carriers, an effect that can be remedied by the equalizer in the receiver system.
Zero ISI reception by using a guard interval is only possible because of the long OFDM sym-
1. U.S. Patent 3,488,445, issued January 6, 1970.

OFDM symbol 0
OFDM symbol 1
OFDM symbol 2
time
OFDM symbol 3
OFDM symbol 4
OFDM symbol x
frequency
Figure 3.8.2 Positions of pilots in the OFDM symbol. (From [2]. Used with permission.)
bol duration. Only then, the unused time of the guard interval can be a modest fraction of the
useful time.
• Reference signals, called pilots, are inserted in the frequency domain. The position, phase,
and energy level of the pilots are pre-defined, as shown in Figure 3.8.2, enabling the receiver
to reconstruct the shape of the channel. Knowing the fading, amplification, and phase-shift of
all the individual sub-carriers, the receiver is able to equalize each subcarrier. This reverses
effects of the transmission channel.
An echo is a copy of the original signal delayed in time. Problems can occur when one OFDM
symbol overlaps with the next one. There is no correlation between two consecutive OFDM sym-
bols and therefore interference from one symbol with the other will result in a disturbed signal.
Because of the efficient spectrum usage, the interfering signal very much resembles white noise.
The DVB-T standard adds a cyclic copy of the last part of the OFDM symbol in front of the sym-
bol to overcome this problem. This guard-interval protects the OFDM symbol from being dis-
turbed by its predecessor, as illustrated in Figure 3.8.3.
When ISI occurs, the resulting signal to noise ratio can be described as [2]
⎛ A⋅t ⎞
SNR = 10⋅ 10 log⎜ ⎟dB
⎝ A′⋅Tu ⎠
(3.8.1)
Where:
A’ = the power of the echo
A = the power of the original signal
t = the duration of interference in complex samples
Tu = the duration of one OFDM symbol without guard interval
The receiver must find a window of duration Tu within the Tu plus Tg time frame that suffers
from minimal ISI.
The OFDM method uses N carriers. At least N complex discrete time samples are required to
represent the OFDM symbol. The N complex time domain samples (0…N – 1) resulting from a
single sub-carrier k modulated with Ck within an OFDM symbol are
guard 0 OFDM symbol 0 guard 1 OFDM symbol 1

Main signal
Main signal +
delayed Δ in
time (echo)
ISI free window
=
Sum
ISI
Figure 3.8.3 Protection of the OFDM symbol against echoes through the use of a guard-interval.
2π
C j k⋅n
sofdmk [ n] = N
k
⋅e N
(3.8.2)
Where:
N = number of sub-carriers and time-domain samples used
n = time-domain sample index (0…N – 1)
k = sub-carrier index (0…N – 1)
Ck = complex phase and amplitude information to be transmitted
Both k and Ck are constant for a single sub-carrier during the period of an OFDM-symbol.
When examining Equation (3.8.2), it appears that the N complex samples for sub-carrier k rotate
exactly k circles in the complex plane during the useful period of an OFDM-symbol.
The complete time-domain symbol is constructed from the N subcarriers by super-imposing
these waves onto each other
N −1
sofdm [ n] = ∑ sofdmk [ n ]
k=0
(3.8.3)
Inside the DVB-T receiver, the OFDM signal is analyzed by applying an FFT to the time-
domain signal. The originally transmitted information is reconstructed by comparing each sub-
carrier with a reference subcarrier of known phase and amplitude, and equal frequency
2π
j k⋅n
sref k [ n] = 1⋅e N
(3.8.4)
As a result of the orthogonality of the N subcarriers, a zero result is obtained in the FFT for any
other subcarrier than the reference
N −1
sofdml [ n ]
(l ≠ k ) ⇒ ∑ sref k [ n ]
=0
n=0
(3.8.5)
Therefore, to isolate the transmitted information for subcarrier k, the complete OFDM-symbol
can be analyzed without error using an FFT
N −1
s [n]
∑ sofdm [n] = Ck′
n=0 ref k
(3.8.6)
in which Cκ′ is the reconstructed phase and amplitude information.

If the receiver receives a delayed input signal, Equation (3.8.2) can be substituted by
2π
C j k⋅( n−Δ )
sofdmk [ n] = N
k
⋅e N
(3.8.7)
where Δ = delay in units of complex samples. Then, this yields at the output of the FFT
N −1 2π
s [n] j k⋅Δ
∑ sofdm [n] = Ck′ ⋅e N
n=0 ref k
(3.8.8)
Equation (3.8.8) shows that a delay in the input signal causes a rotation over the carriers in the
frequency domain. Adding this delayed signal to the original will result in fading and amplifica-
tion of different parts of the frequency domain. This effect is graphically shown in Figure 3.8.4.
In order to reconstruct this distortion of the channel, the OFDM symbol contains reference
signals (pilots). They are all modulated with known phase and amplitude information (Ck ). After
collecting pilots from four symbols, channel information is available at every third sub-carrier.
The missing information for the two sub-carriers in between the reference signals is obtained
through interpolation. Of course, this interpolation sets a limit to the maximum frequency of the
distortion and therefore to the maximum delay (Δ) of the echo. Figure 3.8.5 shows this limitation
when using a standard interpolation technique.
Delays in the range of –N/6 and +N/6 can be resolved. The OFDM symbol is protected by the
guard interval from echoes up to a delay of N/4 [2]. Therefore, it makes sense to shift the interpo-
lation filter range to cope with delay in the range of 0 to N/3 (Figure 3.8.6).
1.5
4
1 2
2π
j⋅ ⋅k ⋅ Δ
'
C ⋅e N
Channel Gain (dB)

0.5 k 0
-2
Im
0
-4
-0.5
-6
-1
-8
-1.5 -10
0 1000 2000 3000 4000 5000 6000
0.5
1.5
-1.5
-0.5
0
-1
carrier number k
Re
Figure 3.8.4 Fading effect of an echo. (From [2]. Used with permission.)
subsampling by three
with reference signals (pilots)
interpolation of missing carriers

in between the pilots
-N/2 -N/6 N/6 N/2

Δ
Figure 3.8.5 Limitation caused by a small number of reference signals. (From [2]. Used with per-
mission.)
Main signal to which receiver is locked
Echo at N/4 complex samples delay
-N/2 -N/6 N/6 N/2

Δ
Figure 3.8.6 Shift of the interpolation filter towards 0 to N/3 distortion range. (From [2]. Used with
permission.)
Main signal to which receiver is locked
Pre-echo
Pre-echo aliased into visible
region of interpolation filter.
-N/2 -N/6 N/6 N/2

Δ
Figure 3.8.7 Aliasing of pre-echo. (From [2]. Used with permission.)
Main signal
Pre-echo to which the receiver is now locked
-N/2 -N/6 N/6 N/2

Δ
Figure 3.8.8 Delay input signal of FFT to shift pre-echo into interpolation filters range. (From [2].
When a receiver locks to the strongest signal instead of the nearest in case of a pre-echo, it
can be seen that this pre-echo cannot be resolved by the shifted interpolation filter. This pre-echo
at –Δ is aliased into the interpolation filters range to an echo at N/3 – Δ, which in the case of a
strong pre-echo is disastrous. (See Figure 3.8.7.) Additionally, synchronized in this manner, the
pre-echo causes ISI.
For the interpolation filter to reconstruct the distorted channel correctly, the input signal of
the FFT must be delayed in time. The receiver then synchronizes to the pre-echo that will yield
the channel distortion shown Figure 3.8.8. The interpolation filter can resolve this distortion and,
therefore, the originally transmitted information can be retrieved without error.
In order synchronize to a weak pre-echo, the receiver must be able to detect the presence of it.
Echoes can be detected by performing an inverse Fourier transform on the available reference
carriers. This yields the channel impulse response. The impulse response can be scanned for pre-
echoes.
3.8.4 References
1. Based on technical reports and background information provided by the DVB Consortium.
2. van Klinken, N., and W. Renirie: “Receiving DVB: Technical Challenges,” Proceedings of
the International Broadcasting Convention, IBC, Amsterdam, September 2000.
3.8.5 Bibliography
Structure, Channel Coding and Modulation for 11–12 GHz Satellite Services,” DVB
Structure, Channel Coding and Modulation for Cable Systems,” DVB Project technical
publication.
cation for Service Information (SI) in Digital Video Broadcasting (DVB) Systems,” DVB
cation for Carrying ITU-R System B Teletext in Digital Video Broadcasting (DVB) Bit-
streams,” DVB Project technical publication.
ETS-300-473, “Digital Broadcasting Systems for Television, Sound, and Data Services; Satellite
Master Antenna Television (SMATV) Distribution Systems,” DVB Project technical publi-
cation.
European Telecommunications Standards Institute: “Digital Video Broadcasting; Framing Struc-
ture, Channel Coding and Modulation for Digital Terrestrial Television (DVB-T)”, March
1997.
Lee, E. A., and D.G. Messerschmitt: Digital Communication, 2nd ed.,. Kluwer, Boston, Mass.,
1994.
Muschallik, C.: “Improving an OFDM Reception Using an Adaptive Nyquist Windowing,” IEEE
Trans. on Consumer Electronics, no. 03, 1996.
Pollet, T., M. van Bladel, and M. Moeneclaey: “BER Sensitivity of OFDM Systems to Carrier
Frequency Offset and Wiener Phase Noise,” IEEE Trans. on Communications, vol. 43,
1995.
Robertson, P., and S. Kaiser: “Analysis of the Effects of Phase-Noise in Orthogonal Frequency
Division Multiplex (OFDM) Systems,” ICC 1995, pp. 1652–1657, 1995.
Sari, H., G. Karam, and I. Jeanclaude: “Channel Equalization and Carrier Synchronization in
OFDM Systems,” IEEE Proc. 6th. Tirrenia Workshop on Digital Communications, Tirre-
nia, Italy, pp. 191–202, September 1993.
g g
Section
RF Interconnection Devices and Systems

4
The components that connect, interface, and otherwise transfer RF energy within a given sys-
tem—or between systems—are critical elements in the operation of a broadcast station. Such
hardware, usually passive, determines to a large extent the overall performance of the RF plant.
The mechanical and electrical characteristics of the transmission line, waveguide, RF com-
biners, and associated hardware that convey power from the transmitter to the antenna are critical
to proper operation of any RF system. Mechanical considerations determine the ability of the
components to withstand temperature extremes, lightning, rain, and wind. In other words, they
determine the overall reliability of the system.
Skin Effect
The effective resistance offered by a given conductor to radio frequencies is considerably higher
than the ohmic resistance measured with direct current. This is because of an action known as the
skin effect, which causes the currents to be concentrated in certain parts of the conductor and
leaves the remainder of the cross section to contribute little or nothing toward carrying the
applied current.
When a conductor carries an alternating current, a magnetic field is produced that surrounds
the wire. This field continually expands and contracts as the ac wave increases from zero to its
maximum positive value and back to zero, then through its negative half-cycle. The changing
magnetic lines of force cutting the conductor induce a voltage in the conductor in a direction that
tends to retard the normal flow of current in the wire. This effect is more pronounced at the cen-
ter of the conductor. Thus, current within the conductor tends to flow more easily toward the sur-
face of the wire. The higher the frequency, the greater the tendency for current to flow at the
surface. The depth of current flow is a function of frequency, and it is determined from the equa-
tion d = 2.6 ⁄ μf where d = depth of current in mils, μ = permeability (copper = 1, steel =
300), and f = frequency of signal in mHz.
It can be calculated that at a frequency of 100 kHz, current flow penetrates a conductor by 8
mils. At 1 MHz, the skin effect causes current to travel in only the top 2.6 mils in copper, and
even less in almost all other conductors. Therefore, the series impedance of conductors at high
frequencies is significantly higher than at low frequencies.
When a circuit is operating at high frequencies, the skin effect causes the current to be redis-
tributed over the conductor cross section in such a way as to make most of the current flow where
4-1
y
4-2 Section Four
it is encircled by the smallest number of flux lines. This general principle controls the distribu-
tion of current regardless of the shape of the conductor involved. With a flat-strip conductor, the
current flows primarily along the edges, where it is surrounded by the smallest amount of flux.
The skin effect is one of the fundamental physical properties that influences all RF intercon-
nection devices and systems.
In This Section:
Chapter 4.1: Transmission Line 4-5

Introduction 4-5
Electrical Parameters 4-6
Transverse Electromagnetic Mode 4-7
Dielectric 4-7
Impedance 4-7
Resonant Characteristics 4-9
Electrical Considerations 4-9
Waveguide 4-10
Propagation Modes 4-11
Dual-Polarity Waveguide 4-12
Efficiency 4-12
Ridged Waveguide 4-13
Circular Waveguide 4-13
Parasitic Energy 4-14
Doubly Truncated Waveguide 4-15
Bibliography 4-16
Chapter 4.2: RF Combiner and Diplexer Systems 4-17

Introduction 4-17
Passive Filters 4-17
Filter Type 4-18
Filter Alignment 4-18
Filter Order 4-19
Four-Port Hybrid Combiner 4-20
Microwave Combiners 4-24
Hot Switching Combiners 4-26
High-Power Isolators 4-28
Theory of Operation 4-28
Applications 4-30
Hot Switch 4-30
Diplexer 4-31
Multiplexer 4-32
References 4-32
Bibliography 4-33

Andrew Corporation: “Broadcast Transmission Line Systems,” Technical Bulletin 1063H,
Orland Park, Ill., 1982.
y
RF Interconnection Devices and Systems 4-3
Andrew Corporation: “Circular Waveguide: System Planning, Installation and Tuning,” Techni-
cal Bulletin 1061H, Orland Park, Ill., 1980.
Ben-Dov, O., and C. Plummer: “Doubly Truncated Waveguide,” Broadcast Engineering, Intertec
Publishing, Overland Park, Kan., January 1989.
Benson, K. B., and J. C. Whitaker: Television and Audio Handbook for Technicians and Engi-
neers, McGraw-Hill, New York, N.Y., 1989.
Cablewave Systems: “Rigid Coaxial Transmission Lines,” Cablewave Systems Catalog 700,
North Haven, Conn., 1989.
Cablewave Systems: “The Broadcaster’s Guide to Transmission Line Systems,” Technical Bulle-
tin 21A, North Haven, Conn., 1976.
Crutchfield, E. B. (ed.), NAB Engineering Handbook, 8th ed., National Association of Broad-
DeComier, Bill: “Inside FM Multiplexer Systems,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., May 1988.
Fink, D., and D. Christiansen (eds.), Electronics Engineers’ Handbook, 2nd ed., McGraw-Hill,
Harrison, Cecil: “Passive Filters,” in The Electronics Handbook, Jerry C. Whitaker (ed.), CRC
Press, Boca Raton, Fla., pp. 279–290, 1996.
Heymans, Dennis: “Hot Switches and Combiners,” Broadcast Engineering, Overland Park,
Kan., December 1987.
Jordan, Edward C.: Reference Data for Engineers: Radio, Electronics, Computer and Communi-
cations, 7th ed., Howard W. Sams, Indianapolis, IN, 1985.
Krohe, Gary L.: “Using Circular Waveguide,” Broadcast Engineering, Intertec Publishing, Over-
land Park, Kan., May 1986.
Perelman, R., and T. Sullivan: “Selecting Flexible Coaxial Cable,” Broadcast Engineering, Inter-
tec Publishing, Overland Park, Kan., May 1988.
Stenberg, James T.: “Using Super Power Isolators in the Broadcast Plant,” Proceedings of the
Broadcast Engineering Conference, Society of Broadcast Engineers, Indianapolis, IN,
1988.
Surette, Robert A.: “Combiners and Combining Networks,” in The Electronics Handbook, Jerry
Terman, F. E.: Radio Engineering, 3rd ed., McGraw-Hill, New York, N.Y., 1947.
Whitaker, Jerry C., G. DeSantis, and C. Paulson: Interconnecting Electronic Systems, CRC
Press, Boca Raton, Fla., 1993.
Whitaker, Jerry C.: Radio Frequency Transmission Systems: Design and Operation, McGraw-
y
4-4 Section Four
Vaughan, T., and E. Pivit: “High Power Isolator for UHF Television,” Proceedings of the NAB
Engineering Conference, National Association of Broadcasters, Washington, D.C., 1989.
g g
Chapter
4.1
Transmission Line
4.1.1 Introduction
Two types of coaxial transmission line are in common use today: rigid line and corrugated (semi-
flexible) line. Rigid coaxial cable is constructed of heavy-wall copper tubes with Teflon or
ceramic spacers. (Teflon is a registered trademark of DuPont.) Rigid line provides electrical per-
formance approaching an ideal transmission line, including:
• High power-handling capability
• Low loss
• Low VSWR (voltage standing wave ratio)
Rigid transmission line is, however, expensive to purchase and install.
The primary alternative to rigid coax is semiflexible transmission line made of corrugated
outer and inner conductor tubes with a spiral polyethylene (or Teflon) insulator. The internal con-
struction of a semiflexible line is shown in Figure 4.1.1. Semiflexible line has four primary ben-
efits:
• It is manufactured in a continuous length, rather than the 20-ft sections typically used for
rigid line.
• Because of the corrugated construction, the line may be shaped as required for routing from
the transmitter to the antenna.
• The corrugated construction permits differential expansion of the outer and inner conductors.
Each size of line has a minimum bending radius. For most installations, the flexible nature of
corrugated line permits the use of a single piece of cable from the transmitter to the antenna, with
no elbows or other transition elements. This speeds installation and provides for a more reliable
system.
4-5
4-6 RF Interconnection Devices and Systems
(a ) (b)
Figure 4.1.1 Semiflexible coaxial cable: (a) a section of cable showing the basic construction, (b)
cable with various terminations. (Courtesy of Andrew.)
4.1.2 Electrical Parameters

A signal traveling in free space is unimpeded; it has a free-space velocity equal to the speed of
light. In a transmission line, capacitance and inductance slow the signal as it propagates along
the line. The degree to which the signal is slowed is represented as a percentage of the free-space
velocity. This quantity is called the relative velocity of propagation and is described by
1
Vp = ------------------ (4.1.1)
L× C
Where:
L = inductance in henrys per foot
C = capacitance in farads per foot
and
Vp
Vr = ----- × 100 percent (4.1.2)
c
Where:
V p = velocity of propagation
c = 9.842 × 108 feet per second (free-space velocity)
Vr = velocity of propagation as a percentage of free-space velocity
4.1.2a Transverse Electromagnetic Mode

The principal mode of propagation in a coaxial line is the transverse electromagnetic mode
(TEM). This mode will not propagate in a waveguide, and that is why coaxial lines can propagate
a broad band of frequencies efficiently. The cutoff frequency for a coaxial transmission line is
determined by the line dimensions. Above cutoff, modes other than TEM can exist and the trans-
mission properties are no longer defined. The cutoff frequency is equivalent to
7.50 × V r
Fc = ----------------------- (4.1.3)
D i + Do
Where:
F c = cutoff frequency in gigahertz
V r = velocity (percent)
Di = inner diameter of outer conductor in inches
Do = outer diameter of inner conductor in inches
At dc, current in a conductor flows with uniform density over the cross section of the conduc-
tor. At high frequencies, the current is displaced to the conductor surface. The effective cross
section of the conductor decreases, and the conductor resistance increases because of the skin
effect.
Center conductors are made from copper-clad aluminum or high-purity copper and can be
solid, hollow tubular, or corrugated tubular. Solid center conductors are found on semiflexible
cable with 1/2 in or smaller diameter. Tubular conductors are found in 7/8 in or larger-diameter
cables. Although the tubular center conductor is used primarily to maintain flexibility, it also can
be used to pressurize an antenna through the feeder.
4.1.2b Dielectric
Coaxial lines use two types of dielectric construction to isolate the inner conductor from the
outer conductor. The first is an air dielectric, with the inner conductor supported by a dielectric
spacer and the remaining volume filled with air or nitrogen gas. The spacer, which may be con-
structed of spiral or discrete rings, typically is made of Teflon or polyethylene. Air-dielectric
cable offers lower attenuation and higher average power ratings than foam-filled cable but
requires pressurization to prevent moisture entry.
Foam-dielectric cables are ideal for use as feeders with antennas that do not require pressur-
ization. The center conductor is surrounded completely by foam-dielectric material, resulting in
a high dielectric breakdown level. The dielectric materials are polyethylene-based formulations,
which contain antioxidants to reduce dielectric deterioration at high temperatures.
4.1.2c Impedance
The expression transmission line impedance applied to a point on a transmission line signifies
the vector ratio of line voltage to line current at that particular point. This is the impedance that
would be obtained if the transmission line were cut at the point in question, and the impedance
looking toward the receiver were mea- (a)
Voltage distribution
sured.
Because the voltage and current distri-
bution on a line are such that the current
tends to be small when the voltage is large
(and vice versa), as shown in Figure 4.1.2, Distance from load
the impedance will, in general, be oscilla-
tory in the same manner as the voltage
(large when the voltage is high and small
(b)
when the voltage is low). Thus, in the case
Impedance (ohms)
of a short-circuited receiver, the imped-
ance will be high at distances from the
receiving end that are odd multiples of 1/4
wavelength, and it will be low at distances
Distance from load
that are even multiples of 1/4 wavelength.
The extent to which the impedance
fluctuates with distance depends on the (c )
standing wave ratio (ratio of reflected to
incident waves), being less as the reflected
wave is proportionally smaller than the 90 o lag
Impedance phase
incident wave. In the particular case where
the load impedance equals the characteris-
tic impedance, the impedance of the trans- 90 o lead
mission line is equal to the characteristic Distance from load
impedance at all points along the line.
The power factor of the impedance of a Figure 4.1.2 Magnitude and power factor of line
transmission line varies according to the impedance with increasing distance from the load
standing waves present. When the load for the case of a short-circuited receiver and a line
impedance equals the characteristic with moderate attenuation: (a) voltage distribution,
impedance, there is no reflected wave, and (b) impedance magnitude, (c) impedance phase.
the power factor of the impedance is equal
to the power factor of the characteristic
impedance. At radio frequencies, the power factor under these conditions is accordingly resis-
tive. However, when a reflected wave is present, the power factor is unity (resistive) only at the
points on the line where the voltage passes through a maximum or a minimum. At other points
the power factor will be reactive, alternating from leading to lagging at intervals of 1/4 wave-
length. When the line is short-circuited at the receiver, or when it has a resistive load less than the
characteristic impedance so that the voltage distribution is of the short-circuit type, the power
factor is inductive for lengths corresponding to less than the distance to the first voltage maxi-
mum. Thereafter, it alternates between capacitive and inductive at intervals of 1/4 wavelength.
Similarly, with an open-circuited receiver or with a resistive load greater than the characteristic
impedance so that the voltage distribution is of the open-circuit type, the power factor is capaci-
tive for lengths corresponding to less than the distance to the first voltage minimum. Thereafter,
the power factor alternates between capacitive and inductive at intervals of 1/4 wavelength, as in
the short-circuited case.
4.1.2d Resonant Characteristics

A transmission line can be used to perform the functions of a resonant circuit. If the line, for
example, is short-circuited at the receiver, at frequencies in the vicinity of a frequency at which
the line is an odd number of 1/4 wavelengths long, the impedance will be high and will vary with
frequency in the vicinity of resonance. This characteristic is similar in nature to a conventional
parallel resonant circuit. The difference is that with the transmission line, there are a number of
resonant frequencies, one for each of the infinite number of frequencies that make the line an
odd number of 1/4 wavelengths long. At VHF, the parallel impedance at resonance and the cir-
cuit Q obtainable are far higher than can be realized with lumped circuits. Behavior correspond-
ing to that of a series resonant circuit can be obtained from a transmission line that is an odd
number of 1/4 wavelengths long and open-circuited at the receiver.
Transmission lines also can be used to provide low-loss inductances or capacitances if the
proper combination of length, frequency, and termination is employed. Thus, a line short-cir-
cuited at the receiver will offer an inductive reactance when less than 1/4 wavelength, and a
capacitive reactance when between 1/4 and 1/2 wavelength. With an open-circuited receiver, the
conditions for inductive and capacitive reactances are reversed.
4.1.2e Electrical Considerations

VSWR, attenuation, and power-handling capability are key electrical factors in the application of
coaxial cable. High VSWR can cause power loss, voltage breakdown, and thermal degradation of
the line. High attenuation means less power delivered to the antenna, higher power consumption
at the transmitter, and increased heating of the transmission line itself.
VSWR is a common measure of the quality of a coaxial cable. High VSWR indicates nonuni-
formities in the cable that can be caused by one or more of the following conditions:
• Variations in the dielectric core diameter
• Variations in the outer conductor
• Poor concentricity of the inner conductor
• Nonhomogeneous or periodic dielectric core
Although each of these conditions may contribute only a small reflection, they can add up to a
measurable VSWR at a particular frequency.
Rigid transmission line typically is available in a standard length of 20 ft, and in alternative
lengths of 19.5 ft and 19.75 ft. The shorter lines are used to avoid VSWR buildup caused by dis-
continuities resulting from the physical spacing between line section joints. If the section length
selected and the operating frequency have a 1/2-wave correlation, the connector junction discon-
tinuities will add. This effect is known as flange buildup. The result can be excessive VSWR.
The critical frequency at which a 1/2-wave relationship exists is given by
F cf = -4---90.4 × n
------------------- (4.1.4)
L
Where:
F cr = the critical frequency
Table 4.1.1 Representative Specifications for Various Types of Flexible Air-Dielectric Coaxial
Cable
Cable size Maximum Velocity Peak power Average power Attenuation (note 1)
(in.) frequency (percent) 1 MHz (kW)
100 MHz 1 MHz (kW) 100 MHz (dB) 1 Mhz (dB)
(MHz) (kW)
1 2.7 92.1 145 145 14.4 0.020 0.207
3 1.64 93.3 320 320 37 0.013 0.14
4 1.22 92 490 490 56 0.010 0.113
5 0.96 93.1 765 765 73 0.007 0.079
Note 1: Attenuation specified in dB/100 ft.
n = any integer
L = transmission line length in feet
For most applications, the critical frequency for a chosen line length should not fall closer
than ±2 MHz of the passband at the operating frequency.
Attenuation is related to the construction of the cable itself and varies with frequency, product
dimensions, and dielectric constant. Larger-diameter cable exhibits lower attenuation than
smaller-diameter cable of similar construction when operated at the same frequency. It follows,
therefore, that larger-diameter cables should be used for long runs.
Air-dielectric coax exhibits less attenuation than comparable-size foam-dielectric cable. The
attenuation characteristic of a given cable also is affected by standing waves present on the line
resulting from an impedance mismatch. Table 4.1.1 shows a representative sampling of semiflex-
ible coaxial cable specifications for a variety of line sizes.
4.1.3 Waveguide
As the operating frequency of a system reaches into the UHF band, waveguide-based transmis-
sion line systems become practical. From the mechanical standpoint, waveguide is simplicity
itself. There is no inner conductor; RF energy is launched into the structure and propagates to the
load. Several types of waveguide are available, including rectangular, square, circular, and ellipti-
cal. Waveguide offers several advantages over coax. First, unlike coax, waveguide can carry
more power as the operating frequency increases. Second, efficiency is significantly better with
waveguide at higher frequencies.
Rectangular waveguide commonly is used in high-power transmission systems. Circular
waveguide also may be used. The physical dimensions of the guide are selected to provide for
propagation in the dominant (lowest-order) mode.
Waveguide is not without its drawbacks, however. Rectangular or square guide constitutes a
large windload surface, which places significant structural demands on a tower. Because of the
physical configuration of rectangular and square guide, pressurization is limited, depending on
the type of waveguide used (0.5 psi is typical). Excessive pressure can deform the guide shape
and result in increased VSWR. Wind also may cause deformation and ensuing VSWR problems.
These considerations have led to the development of circular and elliptical waveguide.
4.1.3a Propagation Modes

Propagation modes for waveguide fall into two broad categories:
• Transverse-electric (TE) waves
• Transverse-magnetic (TM) waves
With TE waves, the electric vector (E vector) is perpendicular to the direction of propagation.
With TM waves, the magnetic vector (H vector) is perpendicular to the direction of propagation.
These propagation modes take on integers (from 0 or 1 to infinity) that define field configura-
tions. Only a limited number of these modes can be propagated, depending on the dimensions of
the guide and the operating frequency.
Energy cannot propagate in waveguide unless the operating frequency is above the cutoff fre-
quency. The cutoff frequency for rectangular guide is
Fc = C
------------ (4.1.5)
2× a
Where:
Fc = waveguide cutoff frequency
c = 1.179 × 1010 inches per second (the velocity of light)
a = the wide dimension of the guide
The cutoff frequency for circular waveguide is defined by
Fc = c
------------------------ (4.1.6)
3.41 × ′ a
where a′ = the radius of the guide.
There are four common propagation modes in waveguide:
• TE 0,1, the principal mode in rectangular waveguide.
• TE 1,0, also used in rectangular waveguide.
• TE 1,1, the principal mode in circular waveguide. TE 1,1 develops a complex propagation pat-
tern with electric vectors curving inside the guide. This mode exhibits the lowest cutoff fre-
quency of all modes, which allows a smaller guide diameter for a specified operating
frequency.
• TM 0,1, which has a slightly higher cutoff frequency than TE1,1 for the same size guide.
Developed as a result of discontinuities in the waveguide, such as flanges and transitions,
TM 0,1 energy is not coupled out by either dominant or cross-polar transitions. The parasitic
energy must be filtered out, or the waveguide diameter chosen carefully to reduce the
unwanted mode.
The field configuration for the dominant mode in rectangular waveguide is illustrated in Fig-
ure 4.1.3. Note that the electric field is vertical, with intensity maximum at the center of the
guide and dropping off sinusoidally to zero intensity at the edges. The magnetic field is in the
form of loops that lie in planes that are at right angles to the electric field (parallel to the top and
(a) (b)
b b
(c )
Electric field
Magnetic field
Figure 4.1.3 Field configuration of the dominant or TE1,0 mode in a rectangular waveguide: (a)
side view, (b) end view, (c) top view.
bottom of the guide). The magnetic field distribution is the same for all planes perpendicular to
the Y-axis. In the X direction, the intensity of the component of magnetic field that is transverse
to the axis of the waveguide (the component in the direction of X) is at any point in the
waveguide directly proportional to the intensity of the electric field at that point. This entire con-
figuration of fields travels in the direction of the waveguide axis (the Z direction in Figure 4.1.3).
The field configuration for the TE1,1 mode in circular waveguide is illustrated in Figure 4.1.4.
The TE 1,1 mode has the longest cutoff wavelength and is, accordingly, the dominant mode. The
next higher mode is the TM0,1, followed by TE2,1.
4.1.3b Dual-Polarity Waveguide

Waveguide will support dual-polarity transmission within a single run of line. A combining ele-
ment (dual-polarized transition) is used at the beginning of the run, and a splitter (polarized
transition) is used at the end of the line. Theoretically, the TE1,0 and TE0,1 modes are capable of
propagation without cross coupling, at the same frequency, in lossless waveguide of square cross
section. In practice, surface irregularities, manufacturing tolerances, and wall losses give rise to
TE1,0- and TE0,1-mode cross conversion. Because this conversion occurs continuously along the
waveguide, long guide runs usually are avoided in dual-polarity systems.
4.1.3c Efficiency
Waveguide losses result from the following:
Lines of electric force

Lines of magnetic force
Figure 4.1.4 Field configuration of the dominant mode in circular waveguide.
(a) (b )
Figure 4.1.5 Ridged waveguide: (a) single-ridged, (b) double-ridged.
• Power dissipation in the waveguide walls and the dielectric material filling the enclosed space
• Leakage through the walls and transition connections of the guide
• Localized power absorption and heating at the connection points
The operating power of waveguide can be increased through pressurization. Sulfur hexafluo-
ride commonly is used as the pressurizing gas.
4.1.3d Ridged Waveguide

Rectangular waveguide may be ridged to provide a lower cutoff frequency, thereby permitting
use over a wider frequency band. As illustrated in Figure 4.1.5, one- and two-ridged guides are
used. Increased bandwidth comes at the expense of increased attenuation, relative to an equiva-
lent section of rectangular guide.
4.1.3e Circular Waveguide

Circular waveguide offers several mechanical benefits over rectangular or square guide. The
windload of circular guide is two-thirds that of an equivalent run of rectangular waveguide. It
(a ) Output
Rotation
Input Cross polarization

reflection without
load
(b )
Output
Cross polarization rotation
Input Two to four

tuned loads
Required
cross polarization
load
Cross polarization component reflected here
Figure 4.1.6 The effects of parasitic energy in circular waveguide: (a) trapped cross-polarization
energy, (b) delayed transmission of the trapped energy.
also presents lower and more uniform windloading than rectangular waveguide, reducing tower
structural requirements.
The same physical properties of circular waveguide that give it good power handling and low
attenuation also result in electrical complexities. Circular waveguide has two potentially
unwanted modes of propagation: the cross-polarized TE 1,1 and TM 0,1 modes.
Circular waveguide, by definition, has no short or long dimension and, consequently, no
method to prevent the development of cross-polar or orthogonal energy. Cross-polar energy is
formed by small ellipticities in the waveguide. If the cross-polar energy is not trapped out, the
parasitic energy can recombine with the dominant-mode energy.
Parasitic Energy
Hollow circular waveguide works as a high-Q resonant cavity for some energy and as a transmis-
sion medium for the rest. The parasitic energy present in the cavity formed by the guide will
appear as increased VSWR if not disposed of. The polarization in the guide meanders and rotates
as it propagates from the source to the load. The end pieces of the guide, typically circular-to-
rectangular transitions, are polarization-sensitive. See Figure 4.1.6a. If the polarization of the
incidental energy is not matched to the transition, energy will be reflected.
Outer covering
of waveguide
Waveguide as viewed
from flange
Figure 4.1.7 Physical construction of doubly truncated waveguide.
Several factors can cause this undesirable polarization. One cause is out-of-round guides that
result from nonstandard manufacturing tolerances. In Figure 4.1.6, the solid lines depict the situ-
ation at launching: perfectly circular guide with perpendicular polarization. However, certain
ellipticities cause polarization rotation into unwanted states, while others have no effect. A 0.2
percent change in diameter can produce a –40 dB cross-polarization component per wavelength.
This is roughly 0.03 in for 18 in of guide length.
Other sources of cross polarization include twisted and bent guides, out-of-roundness, offset
flanges, and transitions. Various methods are used to dispose of this energy trapped in the cavity,
including absorbing loads placed at the ground and/or antenna level.
4.1.3f Doubly Truncated Waveguide

The design of doubly truncated waveguide (DTW) is intended to overcome the problems that can
result from parasitic energy in a circular waveguide. As shown in Figure 4.1.7, DTW consists of
an almost elliptical guide inside a circular shell. This guide does not support cross polarization;
tuners and absorbing loads are not required. The low windload of hollow circular guide is main-
tained, except for the flange area.
Each length of waveguide is actually two separate pieces: a doubly truncated center section
and a circular outer skin, joined at the flanges on each end. A large hole in the broadwall serves
to pressurize the circular outer skin. Equal pressure inside the DTW and inside the circular skin
ensures that the guide will not “breathe” or buckle as a result of rapid temperature changes.
DTW exhibits about 3 percent higher windloading than an equivalent run of circular
waveguide (because of the transition section at the flange joints), and 32 percent lower loading
than comparable rectangular waveguide.
4.1.4 Bibliography
Andrew Corporation: “Broadcast Transmission Line Systems,” Technical Bulletin 1063H,
Orland Park, Ill., 1982.
Andrew Corporation: “Circular Waveguide: System Planning, Installation and Tuning,” Techni-
cal Bulletin 1061H, Orland Park, Ill., 1980.
Ben-Dov, O., and C. Plummer: “Doubly Truncated Waveguide,” Broadcast Engineering, Intertec
Publishing, Overland Park, Kan., January 1989.
Benson, K. B., and J. C. Whitaker: Television and Audio Handbook for Technicians and Engi-
neers, McGraw-Hill, New York, N.Y., 1989.
Cablewave Systems: “The Broadcaster’s Guide to Transmission Line Systems,” Technical Bulle-
tin 21A, North Haven, Conn., 1976.
Cablewave Systems: “Rigid Coaxial Transmission Lines,” Cablewave Systems Catalog 700,
North Haven, Conn., 1989.
Crutchfield, E. B. (ed.), NAB Engineering Handbook, 8th Ed., National Association of Broad-
Fink, D., and D. Christiansen (eds.), Electronics Engineers’ Handbook, 2nd ed., McGraw-Hill,
Jordan, Edward C.: Reference Data for Engineers: Radio, Electronics, Computer and Communi-
cations, 7th ed., Howard W. Sams, Indianapolis, IN, 1985.
Krohe, Gary L.: “Using Circular Waveguide,” Broadcast Engineering, Intertec Publishing, Over-
land Park, Kan., May 1986.
Perelman, R., and T. Sullivan: “Selecting Flexible Coaxial Cable,” Broadcast Engineering, Inter-
tec Publishing, Overland Park, Kan., May 1988.
Terman, F. E.: Radio Engineering, 3rd ed., McGraw-Hill, New York, N.Y., 1947.
Whitaker, Jerry C., G. DeSantis, and C. Paulson: Interconnecting Electronic Systems, CRC
Press, Boca Raton, Fla., 1993.
Whitaker, Jerry C.: Radio Frequency Transmission Systems: Design and Operation, McGraw-
g g
Chapter
4.2
RF Combiner and Diplexer Systems
Cecil Harrison, Robert A. Surette
4.2.1 Introduction
The basic purpose of an RF combiner is to add two or more signals to produce an output signal
that is a composite of the inputs. The combiner performs this signal addition while providing iso-
lation between inputs. Combiners perform other functions as well, and can be found in a wide
variety of RF transmission equipment. Combiners are valuable devices because they permit mul-
tiple amplifiers to drive a single load. The isolation provided by the combiner permits tuning
adjustments to be made on one amplifier—including turning it on or off—without significantly
affecting the operation of the other amplifier. In a typical application, two amplifiers drive the
hybrid and provide two output signals:
• A combined output representing the sum of the two input signals, typically directed toward
the antenna.
• A difference output representing the difference in amplitude and phase between the two input
signals. The difference output typically is directed toward a dummy (reject) load.
For systems in which more than two amplifiers must be combined, two or more combiners are
cascaded.
Diplexers are similar in nature to combiners but permit the summing of output signals from
two or more amplifiers operating at different frequencies. This allows, for example, the outputs
of several transmitters operating on different frequencies to utilize a single broadband antenna.
4.2.2 Passive Filters

A filter is a multiport-network designed specifically to respond differently to signals of different
frequency [1]. This definition excludes networks, which incidentally behave as filters, sometimes
to the detriment of their main purpose. Passive filters are constructed exclusively with passive
4-17
p y
elements (i.e., resistors, inductors, and capacitors). Filters are generally categorized by the fol-
lowing general parameters:
• Type
• Alignment (or class)
• Order
4.2.2a Filter Type

Filters are categorized by type, according to the magnitude of the frequency response, as one of
the following [1]:
• Low-pass (LP)
• High-pass (HP)
• Band-pass (BP)
• Band-stop (BS).
The terms band-reject or notch are also used as descriptive of the BS filter. The term all-pass
is sometimes applied to a filter whose purpose is to alter the phase angle without affecting the
magnitude of the frequency response. Ideal and practical interpretations of the types of filters
and the associated terminology are summarized in Figure 4.2.1.
In general terms, the voltage gain of a filter in the stop band (or attenuation band) is less than
2 /2 (≈ 0.707) times the maximum voltage gain in the pass band. In logarithmic terms, the
gain in the stop band is at least 3.01 dB less than the maximum gain in the pass band. The cutoff
(break or corner) frequency separates the pass band from the stop band. In BP and BS filters,
there are two cutoff frequencies, sometimes referred to as the lower and upper cutoff frequen-
cies. Another expression for the cutoff frequency is half-power frequency, because the power
delivered to a resistive load at cutoff frequency is one-half the maximum power delivered to the
same load in the pass band. For BP and BS filters, the center frequency is the frequency of max-
imum or minimum response magnitude, respectively, and bandwidth is the difference between
the upper and lower cutoff frequencies. Rolloff is the transition from pass band to stop band and
is specified in gain unit per frequency unit (e.g., gain unit/Hz, dB/decade, dB/octave, etc.)
4.2.2b Filter Alignment

The alignment (or class) of a filter refers to the shape of the frequency response [1]. Fundamen-
tally, filter alignment is determined by the coefficients of the filter network transfer function, so
there are an indefinite number of filter alignments, some of which may not be realizable. The
more common alignments are:
• Butterworth
• Chebyshev
• Bessel
• Inverse Chebyshev
• Elliptic (or Cauer)
p y
(a) (b )
(c) (d)
Figure 4.2.1 Filter characteristics by type: (a) low-pass, (b) high-pass, (c) bandpass, (d) bandstop.
Each filter alignment has a frequency response with a characteristic shape, which provides
some particular advantage. (See Figure 4.2.2.) Filters with Butterworth, Chebyshev, or Bessel
alignment are called all-pole filters because their low-pass transfer functions have no zeros.
Table 4.2.1 summarizes the characteristics of the standard filter alignments.
4.2.2c Filter Order

The order of a filter is equal to the number of poles in the filter network transfer function [1]. For
a lossless LC filter with resistive (nonreactive) termination, the number of reactive elements
(inductors or capacitors) required to realize an LP or HP filter is equal to the order of the filter.
Twice the number of reactive elements are required to realize a BP or a BS filter of the same
order. In general, the order of a filter determines the slope of the rolloff—the higher the order,
the steeper the rolloff. At frequencies greater than approximately one octave above cutoff (i.e., f
>> 2 fc), the rolloff for all-pole filters is 20n dB/decade (or approximately 6n dB/octave), where
p y
(a) (b )
Figure 4.2.2 Filter characteristics by alignment, third-order, all-pole filters: (a) magnitude, (b) mag-
nitude in decibels. (From [1]. Used with permission.)
Table 4.2.1 Summary of Standard Filter Alignments (After [1].)
Alignment Pass Band Description Stop Band Description Comments

Butterworth Monotonic Monotonic All-pole; maximally flat
Chebyshev Rippled Monotonic All-pole
Bessel Monotonic Monotonic All-pole; constant phase shift
Inverse Chebyshev Monotonic Rippled
Elliptic (or Cauer) Rippled Rippled
n is the order of the filter (Figure 4.2.3). In the vicinity of fc, both filter alignment and filter order
determine rolloff.
4.2.3 Four-Port Hybrid Combiner

A hybrid combiner (coupler) is a reciprocal four-port device that can be used for either splitting
or combining RF energy over a wide range of frequencies. An exploded view of a typical 3 dB
90° hybrid is illustrated in Figure 4.2.4. The device consists of two identical parallel transmission
lines coupled over a distance of approximately one-quarter wavelength and enclosed within a
single outer conductor. Ports at the same end of the coupler are in phase, and ports at the opposite
end of the coupler are in quadrature (90° phase shift) with respect to each other.
The phase shift between the two inputs or outputs is always 90° and is virtually independent
of frequency. If the coupler is being used to combine two signals into one output, these two sig-
nals must be fed to the hybrid in phase quadrature. When the coupler is used as a power splitter,
the division is equal (half-power between the two ports). The hybrid presents a constant imped-
ance to match each source.
Operation of the combiner can best be understood through observation of the device in a prac-
tical application. Figure 4.2.5 shows a four-port hybrid combiner used to add the outputs of two
p y
Figure 4.2.3 The effects of filter order on rolloff (Butterworth alignment). (From [1]. Used with per-
mission.)
Port 3
Bars of identical size

oriented 180 o with
respect to each other
Port 2
Port 4
Outer cover
Port 1
Figure 4.2.4 Physical model of a 90° hybrid combiner.
transmitters to feed a single load. The combiner accepts one RF source and splits it equally into
two parts. One part arrives at output port C with 0° phase (no phase delay; it is the reference
phase). The other part is delayed by 90° at port D. A second RF source connected to input port B,
but with a phase delay of 90°, also will split in two, but the signal arriving at port C now will be
in phase with source 1, and the signal arriving at port D will cancel, as shown in the figure.
Output port C, the summing point of the hybrid, is connected to the load. Output port D is
connected to a resistive load to absorb any residual power resulting from slight differences in
amplitude and/or phase between the two input sources. If one of the RF inputs fails, half of the
remaining transmitter output will be absorbed by the resistive load at port D.
The four-port hybrid works only when the two signals being mixed are identical in frequency
and amplitude, and when their relative phase is 90°.
p y
Antenna
T 1 input (0o)
T 1 (-0 o )
A C T 1 +T 2
T 2 (-0 o ) (twice the power
of either source)
Reject load
T 1 (+90 o )
B D T 1 -T 2
T 2 (-90 o ) (zero power)
T2 input (-90o )
Figure 4.2.5 Operating principles of a hybrid combiner. This circuit is used to add two identical sig-
nals at inputs A and B.
Operation of the hybrid can best be described by a scattering matrix in which vectors are used
to show how the device functions. Such a matrix is shown in Table 4.2.2. In a 3 dB hybrid, two
signals are fed to the inputs. An input signal at port 1 with 0° phase will arrive in phase at port 3,
and at port 4 with a 90° lag (–90°) referenced to port 1. If the signal at port 2 already contains a
90° lag (–90° referenced to port 1), both input signals will combine in phase at port 4. The signal
from port 2 also experiences another 90° change in the hybrid as it reaches port 3. Therefore, the
signals from ports 1 and 2 cancel each other at port 3.
If the signal arriving at port 2 leads by 90° (mode 1 in the table), the combined power from
ports 1 and 2 appears at port 4. If the two input signals are matched in phase (mode 4), the output
ports (3 and 4) contain one-half of the power from each of the inputs.
If one of the inputs is removed, which would occur in a transmitter failure, only one hybrid
input receives power (mode 5). Each output port then would receive one-half the input power of
the remaining transmitter, as shown.
The input ports present a predictable load to each amplifier with a VSWR that is lower than
the VSWR at the output port of the combiner. This characteristic results from the action of the
difference port, typically connected to a dummy load. Reflected power coming into the output
port will be directed to the reject load, and only a portion will be fed back to the amplifiers. Fig-
ure 4.2.6 illustrates the effect of output port VSWR on input port VSWR, and on the isolation
between ports.
As noted previously, if the two inputs from the separate amplifiers are not equal in amplitude
and not exactly in phase quadrature, some power will be dissipated in the difference port reject
load. Figure 4.2.7 plots the effect of power imbalance, and Figure 4.2.8 plots the effects of phase
imbalance. The power lost in the reject load can be reduced to a negligible value by trimming the
amplitude and/or phase of one (or both) amplifiers.
p y
Table 4.2.2 Single 90° Hybrid System Operating Modes

p y
(a ) Theoretical VSWR curve
50 1.18
Practical 45 1.155
40 1.13
35 1.105
Input port VSWR

Isolation (dB)
30 1.08
25 1.055
Isolation
20 1.03
15 1.005
1.0 1.05 1.10 1.15 1.20 1.25 1.30 1.35 1.40 1.45 1.50 1.55
Output port VSWR

(b )
4 1
3 2
Figure 4.2.6 The effects of load VSWR on input VSWR and isolation: (a) respective curves, (b)
coupler schematic.
4.2.3a Microwave Combiners

Hybrid combiners typically are used in microwave amplifiers to add the output energy of individ-
ual power modules to provide the necessary output from an RF generator. Quadrature hybrids
effect a VSWR-canceling phenomenon that results in well-matched power amplifier inputs and
outputs that can be broadbanded with proper selection of hybrid tees. Several hybrid configura-
tions are possible, including the following:
• Split-tee
• Branch-line
• Magic-tee
• Backward-wave
Key design parameters include coupling bandwidth, isolation, and ease of fabrication. The
equal-amplitude quadrature-phase reverse-coupled TEM 1/4-wave hybrid is particularly attrac-
p y
Ratio of useful power output to total power output (percent)
100
80
60
40
20
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Transmitter power imbalance factor (K)
Pa
K=
Pb
P a = Output power of transmitter 1 (lower power system)

P b = Output power of transmitter 2
Figure 4.2.7 The effects of power imbalance at the inputs of a hybrid coupler.
Ratio of useful power output to total power output (percent)
100
80
60
40
20
0
0 20 40 60 80 100 120 140 160 180
Deviation from quadrature (deg.)
Figure 4.2.8 Phase sensitivity of a hybrid coupler.

p y
1 2
Dielectric
λ /4
3 4
Figure 4.2.9 Reverse-coupled 1/4-wave hybrid combiner.
tive because of its bandwidth and amenability to various physical implementations. Such a
device is illustrated in Figure 4.2.9.
4.2.3b Hot Switching Combiners

Switching RF is nothing new. Typically, the process involves coaxial switches, coupled with the
necessary logic to ensure that the “switch” takes place with no RF energy on the contacts. This
process usually takes the system off-line for a few seconds while the switch is completed.
Through the use of hybrid combiners, however, it is possible to redirect RF signals without turn-
ing the carrier off. This process is referred to as hot switching. Figure 4.2.10 illustrates two of the
most common switching functions (SPST and DPDT) available from hot switchers.
The unique phase-related properties of an RF hybrid make it possible to use the device as a
switch. The input signals to the hybrid in Figure 4.2.11a are equally powered but differ in phase
by 90°. This phase difference results in the combined signals being routed to the output terminal
at port 4. If the relative phase between the two input signals is changed by 180°, the summed out-
put then appears on port 3, as shown in Figure 4.2.11b. The 3 dB hybrid combiner, thus, func-
tions as a switch.
This configuration permits the switching of two RF generators to either of two loads.
Remember, however, that the switch takes place when the phase difference between the two
inputs is 90°. To perform the switch in a useful way requires adding a high-power phase shifter to
one input leg of the hybrid. The addition of the phase shifter permits the full power to be com-
bined and switched to either output. This configuration of hybrid and phase shifter, however, will
not permit switching a main or standby transmitter to a main or auxiliary load (DPDT function).
To accomplish this additional switch, a second hybrid and phase shifter must be added, as shown
in Figure 4.2.12. This configuration then can perform the following switching functions:
• RF source 1 routed to output B
• RF source 2 routed to output A
• RF source 1 routed to output A
p y
SPST
Load A
Transmitter
Load B
DPDT transfer
Load A
Transmitter 1
Load B
Transmitter 2
Figure 4.2.10 Common RF switching configurations.
(a) (b )
1 3 Load 1 3
P P 2P
3 dB 3 dB Load
P 2P P
2 4 2 4
Phase Q =0 o Phase Q = 180 o
shifter shifter
Figure 4.2.11 Hybrid switching configurations: (a) phase set so that the combined energy is deliv-
ered to port 4, (b) phase set so that the combined energy is delivered to port 3.
• RF source 2 routed to output B

The key element in developing such a switch is a high-power phase shifter that does not
exhibit reflection characteristics. In this application, the phase shifter allows the line between the
hybrids to be electrically lengthened or shortened. The ability to adjust the relative phase
p y
T1 A
Phase
shifter
T2 B
Phase
shifter
Figure 4.2.12 Additional switching and combining functions enabled by adding a second hybrid
and another phase shifter to a hot switching combiner.
between the two input signals to the second hybrid provides the needed control to switch the
input signal between the two output ports.
If a continuous analog phase shifter is used, the transfer switch shown in Figure 4.2.12 also
can act as a hot switchless combiner where RF generators 1 and 2 can be combined and fed to
either output A or B. The switching or combining functions are accomplished by changing the
physical position of the phase shifter.
Note that it does not matter whether the phase shifter is in one or both legs of the system. It is
the phase difference (θ1 – θ2) between the two input legs of the second hybrid that is important.
With 2-phase shifters, dual drives are required. However, the phase shifter needs only two posi-
tions. In a 1-phase shifter design, only a single drive is required, but the phase shifter must have
four fixed operating positions.
4.2.4 High-Power Isolators

The high-power ferrite isolator offers the ability to stabilize impedance, isolate the RF generator
from load discontinuities, eliminate reflections from the load, and absorb harmonic and inter-
modulation products. The isolator also can be used to switch between an antenna or load under
full power, or to combine two or more generators into a common load.
Isolators commonly are used in microwave transmitters at low power to protect the output
stage from reflections. Ferrite isolators are also available that can handle 500 kW or more of for-
ward power with less than 0.1 dB of forward power loss.
4.2.4a Theory of Operation

High-power isolators are three-port versions of a family of devices known as circulators. The cir-
culator derives its name from the fact that a signal applied to one of the input ports can travel in
only one direction, as shown in Figure 4.2.13. The input port is isolated from the output port. A
signal entering port 1 appears only at port 2; it does not appear at port 3 unless reflected from
p y
(a )
(b )
(c )
Figure 4.2.13 Basic characteristics of a circulator: (a) operational schematic, (b) distributed con-
stant circulator, (c) lump constant circulator. (From [2]. Used with permission.)
port 2. An important benefit of this one-way power transfer is that the input VSWR at port 1 is
dependent only on the VSWR of the load placed at port 3. In most applications, this load is a
resistive (dummy) load that presents a perfect load to the transmitter.
The unidirectional property of the isolator results from magnetization of a ferrite alloy inside
the device. Through correct polarization of the magnetic field of the ferrite, RF energy will travel
through the element in only one direction (port 1 to 2, port 2 to 3, and port 3 to 1). Reversing the
polarity of the magnetic field makes it possible for RF flow in the opposite direction.
p y
In the basic design, the ferrite is placed in the center of a Y-junction of three transmission
lines, either waveguide or coax. Sections of the material are bonded together to form a thin cylin-
der perpendicular to the electric field. Even though the insertion loss is low, the resulting power
dissipated in the cylinder can be as high as 2 percent of the forward power. Special provisions
must be made for heat removal. It is efficient heat-removal capability that makes high-power
operation possible.
The insertion loss of the ferrite must be kept low so that minimal heat is dissipated. Values of
ferrite loss on the order of 0.05 dB have been produced. This equates to an efficiency of 98.9
percent. Additional losses from the transmission line and matching structure contribute slightly
to loss. The overall loss is typically less than 0.1 dB, or 98 percent efficiency. The ferrite element
in a high-power system is usually water-cooled in a closed-loop path that uses an external radia-
tor.
The two basic circulator implementations are shown in Figures 4.2.13a and 4.2.13b. These
designs consist of Y-shaped conductors sandwiched between magnetized ferrite discs [2]. The
final shape, dimensions, and type of material varies according to frequency of operation, power
handling requirements, and the method of coupling. The distributed constant circulator is the
older design; it is a broad-band device, not quite as efficient in terms of insertion loss and leg-to-
leg isolation, and considerably more expensive to produce. It is useful, however, in applications
where broad-band isolation is required. More common is the lump constant circulator, a less
expensive and more efficient, but narrow-band, design.
At least one filter is always installed directly after an isolator, because the ferrite material of
the isolator generates harmonic signals. If an ordinary band-pass or band-reject filter is not to be
used, a harmonic filter will be needed.
4.2.4b Applications
The high-power isolator permits a transmitter to operate with high performance and reliability
despite a load that is less than optimum. The problems presented by ice formations on a transmit-
ting antenna provide a convenient example. Ice buildup will detune an antenna, resulting in
reflections back to the transmitter and high VSWR. If the VSWR is severe enough, transmitter
power will have to be reduced to keep the system on the air. An isolator, however, permits contin-
ued operation with no degradation in signal quality. Power output is affected only to the extent of
the reflected energy, which is dissipated in the resistive load.
A high-power isolator also can be used to provide a stable impedance for devices that are sen-
sitive to load variations, such as klystrons. This allows the device to be tuned for optimum per-
formance regardless of the stability of the RF components located after the isolator. Figure
4.2.14 shows the output of a wideband (6 MHz) klystron operating into a resistive load, and into
an antenna system. The power loss is the result of an impedance difference. The periodicity of
the ripple shown in the trace is a function of the distance of the reflections from the source.
Hot Switch
The circulator can be made to perform a switching function if a short circuit is placed at the out-
put port. Under this condition, all input power will be reflected back into the third port. The use
of a high-power stub on port 2, therefore, permits redirecting the output of an RF generator to
port 3.
p y
(a )
Amplitude, 10 dB/div
Frequency, 1 MHz/div
(b )
Amplitude, 10 dB/div
Frequency, 1 MHz/div
Figure 4.2.14 Output of a klystron operating into different loads through a high-power isolator: (a)
resistive load, (b) an antenna system.
At odd 1/4-wave positions, the stub appears as a high impedance and has no effect on the out-
put port. At even 1/4-wave positions, the stub appears as a short circuit. Switching between the
antenna and a test load, for example, can be accomplished by moving the shorting element 1/4
wavelength.
Diplexer
An isolator can be configured to combine the aural and visual outputs of a TV transmitter into a
single output for the antenna. The approach is shown in Figure 4.2.15. A single notch cavity at
the aural frequency is placed on the visual transmitter output (circulator input), and the aural sig-
nal is added (as shown). The aural signal will be routed to the antenna in the same manner as it is
reflected (because of the hybrid action) in a conventional diplexer.
p y
Aural notch
filter
Visual input signal To antenna
Aural input signal
Figure 4.2.15 Use of a circulator as a diplexer in TV applications.
Notch Notch Notch

filter filter filter
Output to antenna
F1
F1+ F2 F1+ F2+ F3 F1+ F2+ F3+ F4
F F F
2 3 4
Figure 4.2.16 Use of a bank of circulators in a multiplexer application.
Multiplexer
A multiplexer can be formed by cascading multiple circulators, as illustrated in Figure 4.2.16.
Filters must be added, as shown. The primary drawback of this approach is the increased power
dissipation that occurs in circulators nearest the antenna.
4.2.5 References
1. Harrison, Cecil: “Passive Filters,” in The Electronics Handbook, Jerry C. Whitaker (ed.),
2. Surette, Robert A.: “Combiners and Combining Networks,” in The Electronics Handbook,
Jerry C. Whitaker (ed.), CRC Press, Boca Raton, Fla., pp. 1368–1381, 1996.
p y
4.2.6 Bibliography
DeComier, Bill: “Inside FM Multiplexer Systems,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., May 1988.
Heymans, Dennis: “Hot Switches and Combiners,” Broadcast Engineering, Overland Park,
Kan., December 1987.
Stenberg, James T.: “Using Super Power Isolators in the Broadcast Plant,” Proceedings of the
Broadcast Engineering Conference, Society of Broadcast Engineers, Indianapolis, IN,
1988.
Vaughan, T., and E. Pivit: “High Power Isolator for UHF Television,” Proceedings of the NAB
Engineering Conference, National Association of Broadcasters, Washington, D.C., 1989.
p y
g g
Section
Transmitting Antennas and Systems

5
The purpose of an antenna is to radiate efficiently the power supplied to it by the transmitter. A
simple antenna, consisting of a single vertical element over a ground plane can do this job quite
well at low-to-medium frequencies. Antenna systems may also be required to concentrate the
radiated power in a given direction and to minimize radiation in the direction of other stations
sharing the same or adjacent frequencies. To achieve such directionality may require a compli-
cated antenna system that incorporates a number of individual elements and matching networks.
As the operating frequency increases into VHF and above, the short wavelengths permit the
design of specialized antennas that offer high directivity and gain.
Wavelength λ is the distance traveled by one cycle of a radiated electric signal. The frequency
f of the signal is the number of cycles per second (Hz). It follows that the frequency is inversely
proportional to the wavelength. Both wavelength and frequency are related to the speed of light c
per the formula c = f λ. The velocity of electric signals in air is essentially the same as that of
light in free space (2.9983 × 1010 cm/s).
Bandwidth is a general classification of the frequencies over which an antenna is effective.
This parameter requires specification of acceptable tolerances relating to the uniformity of
response over the intended operating band. Strictly speaking, antenna bandwidth is the difference
in frequency between two points at which the power output of the device drops to one-half the
midrange value. These points are called half-power points. A half-power point is equal to a
VSWR of 5.83:1, or the point at which the voltage response drops to 0.7071 of the midrange
value. In most communications systems a VSWR of less than 1.2:1 within the occupied band-
width of the radiated signal is preferable, and modern broadcast systems typically exhibit VSWR
of 1.1:1 or lower.
Antenna bandwidth depends upon the radiating element impedance and the rate at which the
reactance of the antenna changes with frequency. Bandwidth and RF coupling go hand in hand
regardless of the method used to excite the antenna. All elements between the transmitter output
circuit and the antenna must be analyzed, first by themselves, and then as part of the overall sys-
tem bandwidth. In any transmission system, the composite bandwidth, not just the bandwidths of
individual components, is of primary concern.
The dipole antenna is simplest of all antennas, and the building block of most other designs.
The dipole consists of two in-line rods or wires with a total length equal to 1/2-wave at the oper-
ating frequency. The radiation resistance of a dipole is on the order of 73 Ω. The bandwidth of
5-1
g y
5-2 Section Five
the antenna may be increased by increasing the diameter of the elements, or by using cones or
cylinders rather than wires or rods. Such modifications also increase the impedance of the
antenna.
The dipole can be straight (in-line) or bent into a V-shape. The impedance of the V-dipole is a
function of the V angle. Changing the angle effectively tunes the antenna. The vertical radiation
pattern of the V-dipole antenna is similar to the straight dipole for angles of 120º or less.
These and other basic principles have led to the development of many different types of
antennas for broadcast applications. With the implementation of digital radio and television
broadcasting, antenna designs are being pushed to new limits as the unique demands of digital-
services are identified and solutions implemented.
In This Section:
Chapter 5.1: Radio Antenna Principles 5-11

Introduction 5-11
Fundamental Parameters 5-11
Polarization 5-12
Antenna Beamwidth 5-12
Antenna Gain 5-12
Space Regions 5-14
Impedance Matching 5-14
L Network 5-14
Tee Network 5-15
Pi Network 5-17
Line Stretcher 5-18
Antenna Types 5-20
Quarter-Wave Monopole 5-20
Log-Periodic Antenna 5-21
Yagi-Uda Antenna 5-21
Spiral Antenna 5-22
Array Antenna 5-22
AM Broadcast Antenna Systems 5-22
Directional AM Antenna Design 5-23
Antenna Pattern Design 5-24
Bearing 5-25
Antenna Monitoring System 5-26
Phase/Current Sample Loop 5-26
Sample lines 5-27
The Common Point 5-28
Monitor Points 5-28
Folded Unipole Antenna 5-28
FM Broadcast Antenna Systems 5-29
Antenna Types 5-29
Bibliography 5-31
Chapter 5.2: Television Antenna Principles 5-33

Introduction 5-33
g y
Transmitting Antennas and Systems 5-3
Power and Voltage Rating 5-33

Effective Radiated Power 5-35
Directional Antennas 5-36
Polarization 5-36
Practical Application 5-37
Gain 5-40
Circular Polarization 5-42
Elevation Pattern Shaping 5-44
Elevation-Pattern Derivation 5-47
Beam Tilting 5-47
Null-Fill 5-47
Azimuthal-Pattern Shaping 5-48
Voltage Standing-Wave Ratio and Impedance 5-52
VSWR Optimization Techniques 5-58
Azimuthal-Pattern Distortion in Multiple-Antenna Installations 5-60
Echoes in Multiple-Tower Systems 5-60
References 5-63
Bibliography 5-64
Chapter 5.3: Television Transmitting Antennas 5-65

Introduction 5-65
Common Antenna Designs 5-66
Multislot Radiator 5-69
Circular Polarization 5-69
UHF Antennas for Tower-Top Installation 5-69
VHF Antennas for Tower-Face Installation 5-70
UHF Antennas for Tower-Face Installation 5-71
Vertically Stacked Multiple Antennas 5-72
Candelabra Systems 5-73
Multiple-Tower Installations 5-73
Multiplexing of Channels 5-73
Broadband Television Antennas 5-74
Bandwidth 5-74
Horizontal Pattern 5-75
DTV Implementation Issues 5-76
Channel-Combining Considerations 5-76
Antenna Systems 5-77
Key Considerations in System Design 5-78
References 5-78
Bibliography 5-79
Chapter 5.4: Tower Construction and Maintenance 5-81

Introduction 5-81
Antenna Tower Lighting and Marking Requirements 5-81
Tower Height 5-82
Ice 5-83
Problems Caused by Icing 5-84
Prevention and Ice Control 5-85
g y
5-4 Section Five
Heating 5-85
References 5-86
Bibliography 5-86
Chapter 5.5: Tower Grounding 5-87

Introduction 5-87
Terms and Codes 5-87
The Need for Grounding 5-87
The Grounding Electrode 5-88
Earth Electrode 5-89
Establishing an Earth Ground 5-89
Grounding Interface 5-89
Ground Electrode Testing 5-93
Chemical Ground Rods 5-94
Ufer Ground System 5-96
Bonding Ground-System Elements 5-99
Exothermic Bonding 5-99
Ground-System Inductance 5-100
Grounding Tower Elements 5-101
Ground-Wire Dressing 5-102
Facility Ground Interconnection 5-103
Personnel Protection 5-105
Grounding on Bare Rock 5-106
Rock-Based Radial Elements 5-106
Transmission-System Grounding 5-108
Cable Considerations 5-108
Satellite Antenna Grounding 5-108
References 5-110
Bibliography 5-112
Chapter 5.6: Lightning Effects 5-115

Introduction 5-115
Sources of Atmospheric Energy 5-115
Characteristics of Lightning 5-118
Cloud-to-Cloud Activity 5-121
Lightning Protection 5-122
Protection Area 5-125
Electrostatic Discharge 5-126
Triboelectric Effect 5-127
EMP Radiation 5-127
Coupling Transient Energy 5-129
Bibliography 5-130
Chapter 5.7: Transmitter Building Grounding Practices 5-133

Introduction 5-133
Designing a Building Ground System 5-133
g y
Bulkhead Panel 5-136

Bulkhead Grounding 5-137
Lightning Protectors 5-141
Typical Installation 5-141
Checklist for Proper Grounding 5-144
References 5-146
Bibliography 5-146

Allnatt, J. W., and R. D. Prosser: “Subjective Quality of Television Pictures Impaired by Long
Delayed Echoes,” Proc. IEEE, vol. 112, no. 3, March 1965.
Bendov, Oded: “Coverage Contour Optimization of HDTV and NTSC Antennas,” Proceedings
of the 1996 NAB Broadcast Engineering Conference, National Association of Broadcasters,
Bendov, Oded.: “Measurement of Circularly Polarized Broadcast Antennas,” IEEE Trans. Broad-
casting, vol. BC-19, no. 1, pp. 28–32, March 1972.
Benson, B., and Whitaker, J.: Television and Audio Handbook for Technicians and Engineers,
Bingeman, Grant: “AM Tower Impedance Matching,” Broadcast Engineering, Intertec Publish-
ing, Overland Park, Kan., July 1985.
Bishop, Don: “How the FCC Decides Whether Your Tower is OK,” Mobile Radio Technology,
Intertec Publishing, Overland Park, Kan., April 1988.
Bishop, Don, “Lightning Devices Undergo Tests at Florida Airports,” Mobile Radio Technology,
PRIMEDIA Intertec, Overland Park, Kan., May 1990.
Bishop, Don, “Lightning Sparks Debate: Prevention or Protection?,” Mobile Radio Technology,
PRIMEDIA Intertec, Overland Park, Kan., January 1989.
Bishop, Don: “Tower Maintenance,” Mobile Radio Technology, Intertec Publishing, Overland
Park, Kan., April 1988.
Bishop, Don: “When to Use Strobes on Communications Towers,” Mobile Radio Technology,
Intertec Publishing, Overland Park, Kan., March 1989.
Bixby, Jeffrey: “AM DAs—Doing it Right,” Broadcast Engineering, Intertec Publishing, Over-
land Park, Kan., February 1984.
Block, Roger, “Dissipation Arrays: Do They Work?,” Mobile Radio Technology, PRIMEDIA
Intertec, Overland Park, Kan., April 1988.
Block, Roger: “How to Ground Guy Anchors and Install Bulkhead Panels,” Mobile Radio Tech-
nology, PRIMEDIA Intertec, Overland Park, Kan., February 1986.
Block, Roger: “The Grounds for Lightning and EMP Protection,” PolyPhaser Corporation, Gard-
nerville, Nev., 1987.
Brawn, D. A., and B. F. Kellom: “Butterfly VHF Panel Antenna,” RCA Broadcast News, vol.
138, pp. 8–12, March 1968.
g y
5-6 Section Five
Carpenter, Roy, B.: “Improved Grounding Methods for Broadcasters,” Proceedings, SBE
National Convention, Society of Broadcast Engineers, Indianapolis, IN, 1987.
Chick, Elton B.: “Monitoring Directional Antennas,” Broadcast Engineering, Intertec Publish-
Clark, R. N., and N. A. L. Davidson: “The V-Z Panel as a Side Mounted Antenna,” IEEE Trans.
Broadcasting, vol. BC-13, no. 1, pp. 3–136, January 1967.
Davis, Gary, and Ralph Jones: Sound Reinforcement Handbook, Yamaha Music Corporation, Hal
Leonard Publishing, Milwaukee, WI, 1987.
DeDad, John A., (ed.): “Basic Facility Requirements,” in Practical Guide to Power Distribution
for Information Technology Equipment, PRIMEDIA Intertec, Overland Park, Kan., pp. 24,
1997.
Defense Civil Preparedness Agency, EMP and Electric Power Systems, Publication TR-6l-D,
U.S. Department of Commerce, National Bureau of Standards, Washington, D.C., July
1973.
Defense Civil Preparedness Agency, “EMP Protection for AM Radio Stations,” Washington,
D.C., TR-61-C, May 1972.
Defense Civil Preparedness Agency, EMP Protection for Emergency Operating Centers, Federal
Information Processing Standards Publication no. 94, Guideline on Electrical Power for
ADP Installations, U.S. Department of Commerce, National Bureau of Standards, Wash-
ington, D.C., 1983.
DeVito, G: “Considerations on Antennas with no Null Radiation Pattern and Pre-established
Maximum-Minimum Shifts in the Vertical Plane,” Alta Frequenza, vol. XXXVIII, no. 6,
1969.
DeVito, G., and L. Mania: “Improved Dipole Panel for Circular Polarization,” IEEE Trans.
Broadcasting, vol. BC-28, no. 2, pp. 65–72, June 1982.
DeWitt, William E.: “Facility Grounding Practices,” in The Electronics Handbook, Jerry C. Whi-
taker (ed.), CRC Press, Boca Raton, Fla., pp. 2218–2228, 1996.
Drabkin, Mark, and Roy Carpenter, Jr., “Lightning Protection Devices: How Do They Com-
pare?,” Mobile Radio Technology, PRIMEDIA Intertec, Overland Park, Kan., October
1988.
Dudzinsky, S. J., Jr.: “Polarization Discrimination for Satellite Communications,” Proc. IEEE,
vol. 57, no. 12, pp. 2179–2180, December 1969.
Fardo, S., and D. Patrick: Electrical Power Systems Technology, Prentice-Hall, Englewood Cliffs,
N.J., 1985.
Fink, D., and D. Christiansen (eds.): Electronics Engineer's Handbook, 3rd ed., McGraw-Hill,
Fisk, R. E., and J. A. Donovan: “A New CP Antenna for Television Broadcast Service,” IEEE
Trans. Broadcasting, vol. BC-22, no. 3, pp. 91–96, September 1976.
Fowler, A. D., and H. N. Christopher: “Effective Sum of Multiple Echoes in Television,” J.
SMPTE, SMPTE, White Plains, N. Y., vol. 58, June 1952.
g y
Fumes, N., and K. N. Stokke: “Reflection Problems in Mountainous Areas: Tests with Circular
Polarization for Television and VHF/FM Broadcasting in Norway,” EBU Review, Technical
Part, no. 184, pp. 266–271, December 1980.
Heymans, Dennis: “Channel Combining in an NTSC/ATV Environment,” Proceedings of the
Hill, Mark: “Computer Power Protection,” Broadcast Engineering, PRIMEDIA Intertec, Over-
land Park, Kan., April 1987.
Hill, P. C. J.: “Measurements of Reradiation from Lattice Masts at VHF,” Proc. IEEE, vol. III,
no. 12, pp. 1957–1968, December 1964.
Hill, P. C. J.: “Methods for Shaping Vertical Pattern of VHF and UHF Transmitting Aerials,”
Proc. IEEE, vol. 116, no. 8, pp. 1325–1337, August 1969.
IEEE Standard 100: Definitions of Electrical and Electronic Terms, IEEE, New York, N.Y.
IEEE Standard 142: “Recommended Practice for Grounding Industrial and Commercial Power
Systems,” IEEE, New York, N.Y., 1982.
IEEE Standard 1100: “Recommended Practice for Powering and Grounding Sensitive Electron-
ics Equipment,” IEEE, New York, N.Y., 1992.
Jay, Frank, ed., IEEE Standard Directory of Electrical and Electronics Terms, 3rd ed., IEEE, New
York, N.Y., 1984.
Johns, M. R., and M. A. Ralston: “The First Candelabra for Circularly Polarized Broadcast
Antennas,” IEEE Trans. Broadcasting, vol. BC-27, no. 4, pp. 77–82, December 1981.
Johnson, R. C, and H. Jasik: Antenna Engineering Handbook, 2d ed., McGraw-Hill, New York,
N.Y., 1984.
Jordan, Edward C. (ed.): “Reference Data for Engineers—Radio, Electronics, Computer and
Communications, 7th ed., Howard W. Sams, Indianapolis, Ind., 1985.
Kaiser, Bruce A., “Can You Really Fool Mother Nature?,” Cellular Business, PRIMEDIA Inter-
tec, Overland Park, Kan., March 1989.
Kaiser, Bruce A., “Straight Talk on Static Dissipation,” Proceedings of ENTELEC 1988, Energy
Telecommunications and Electrical Association, Dallas, Tex., 1988.
Key, Lt. Thomas, “The Effects of Power Disturbances on Computer Operation,” IEEE Industrial
and Commercial Power Systems Conference paper, Cincinnati, Ohio, June 7, 1978.
Knight, P.: “Reradiation from Masts and Similar Objects at Radio Frequencies,” Proc. IEEE, vol.
114, pp. 30–42, January 1967.
Kraus, J. D.: Antennas, McGraw-Hill, New York, N.Y., 1950.
Lanphere, John: “Establishing a Clean Ground,” Sound & Video Contractor, PRIMEDIA Inter-
tec, Overland Park, Kan., August 1987.
Lawrie, Robert: Electrical Systems for Computer Installations, McGraw-Hill, New York, N.Y.,
1988.
g y
5-8 Section Five
Lehtinen, Rick: “Hardening Towers,” Broadcast Engineering, Overland Park, Kan., pp. 94–106,
March 1990.
Lessman, A. M.: “The Subjective Effect of Echoes in 525-Line Monochrome and NTSC Color
Television and the Resulting Echo Time Weighting,” J. SMPTE, SMPTE, White Plains,
N.Y., vol. 1, December 1972.
Little, Richard: “Surge Tolerance: How Does Your Site Rate?” Mobile Radio Technology, Inter-
tec Publishing, Overland Park, Kan., June 1988.
Lobnitz, Edward A.: “Lightning Protection for Tower Structures,” in NAB Engineering Hand-
book, 9th ed., Jerry C. Whitaker (ed.), National Association of Broadcasters, Washington,
D.C., 1998.
Martzloff, F. D., “The Development of a Guide on Surge Voltages in Low-Voltage AC Power Cir-
cuits,” 14th Electrical/Electronics Insulation Conference, IEEE, Boston, Mass., October
1979.
Mertz, P.: “Influence of Echoes on Television Transmission,” J. SMPTE, SMPTE, White Plains,
N.Y., vol. 60, May 1953.
Midkiff, John: “Choosing the Right Coaxial Cable Hanger,” Mobile Radio Technology, Intertec
Publishing, Overland Park, Kan., April 1988.
Military Handbook 419A: “Grounding, Bonding, and Shielding for Electronic Systems,” U.S.
Government Printing Office, Philadelphia, PA, December 1987.
Moreno, T.: Microwave Transmission Design Data, Dover, New York, N.Y.
Mulherin, Nathan D.: “Atmospheric Icing on Communication Masts in New England.” CRREL
Report 86-17, U.S. Army Cold Regions Research and Engineering Laboratory, December
1986.
Mullaney, John H.: “The Folded Unipole Antenna,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., July 1986.
Mullaney, John H.: “The Folded Unipole Antenna for AM Broadcast,” Broadcast Engineering,
Intertec Publishing, Overland Park, Kan., January 1960.
Mullinack, Howard G.: “Grounding for Safety and Performance,” Broadcast Engineering, PRI-
MEDIA Intertec, Overland Park, Kan., October 1986.
NFPA Standard 70: “The National Electrical Code,” National Fire Protection Association,
Quincy, Mass., 1993.
Nott, Ron, “The Sources of Atmospheric Energy,” Proceedings of the SBE National Convention
and Broadcast Engineering Conference, Society of Broadcast Engineers, Indianapolis, IN,
1987.
Perini, J.: “Echo Performance of TV Transmitting Systems,” IEEE Trans. Broadcasting, vol. BC-
16, no. 3, September 1970.
Perini, J.: “Improvement of Pattern Circularity of Panel Antenna Mounted on Large Towers,”
IEEE Trans. Broadcasting, vol. BC-14, no. 1, pp. 33–40, March 1968.
g y
Perini, J., and M. H. Ideslis: “Radiation Pattern Synthesis for Broadcast Antennas,” IEEE Trans.
Broadcasting, vol. BC-18, no. 3, pg. 53, September 1972.
Plonka, Robert J.: “Can ATV Coverage Be Improved With Circular, Elliptical, or Vertical Polar-
ized Antennas?” Proceedings of the 1996 NAB Broadcast Engineering Conference,
National Association of Broadcasters, Washington, D.C., pg. 155, 1996.
Praba, K.: “Computer-aided Design of Vertical Patterns for TV Antenna Arrays,” RCA Engineer,
vol. 18-4, January–February 1973.
Praba, K.: “R. F. Pulse Measurement Techniques and Picture Quality,” IEEE Trans. Broadcast-
ing, vol. BC-23, no. 1, pp. 12–17, March 1976.
“Predicting Characteristics of Multiple Antenna Arrays,” RCA Broadcast News, vol. 97, pp. 63–
68, October 1957.
Sargent, D. W.: “A New Technique for Measuring FM and TV Antenna Systems,” IEEE Trans.
Broadcasting, vol. BC-27, no. 4, December 1981.
Schneider, John: “Surge Protection and Grounding Methods for AM Broadcast Transmitter
Sites,” Proceedings, SBE National Convention, Society of Broadcast Engineers, Indianapo-
lis, IN, 1987.
Schwarz, S. J.: “Analytical Expression for Resistance of Grounding Systems,” AIEE Transac-
tions, vol. 73, Part III-B, pp. 1011–1016, 1954.
Siukola, M. S.: “Size and Performance Trade Off Characteristics of Horizontally and Circularly
Polarized TV Antennas,” IEEE Trans. Broadcasting, vol. BC-23, no. 1, March 1976.
Siukola, M. S.: “The Traveling Wave VHF Television Transmitting Antenna,” IRE Trans. Broad-
casting, vol. BTR-3, no. 2, pp. 49-58, October 1957.
Siukola, M. S.: “TV Antenna Performance Evaluation with RF Pulse Techniques,” IEEE Trans.
Broadcasting, vol. BC-16, no. 3, September 1970.
Smith, Paul D.: “New Channel Combining Devices for DTV, Proceedings of the 1997 NAB
D.C., pg. 218, 1996.
Sullivan, Thomas: “How to Ground Coaxial Cable Feedlines,” Mobile Radio Technology, PRI-
MEDIA Intertec, Overland Park, Kan., April 1988.
Sunde, E. D.: Earth Conduction Effects in Transmission Systems, Van Nostrand Co., New York,
N.Y., 1949.
Technical Reports LEA-9-1, LEA-0-10, and LEA-1-8, Lightning Elimination Associates, Santa
Fe Springs, Calif.
Ulman, Martin A., The Lightning Discharge, Academic Press, Orlando, Fla., 1987.
“WBAL, WJZ and WMAR Build World’s First Three-Antenna Candelabra,” RCA Broadcast
News, vol. 106, pp. 30–35, December 1959.
Webster's New Collegiate Dictionary.
Wescott, H. H.: “A Closer Look at the Sutro Tower Antenna Systems,” RCA Broadcast News,
vol. 152, pp. 35–41, February 1944.
g y
5-10 Section Five
Westberg, J. M.: “Effect of 90° Stub on Medium Wave Antennas,” NAB Engineering Handbook,
7the ed., National Association of Broadcasters, Washington, D.C., 1985.
Whitaker, Jerry C., AC Power Systems Handbook, 2nd ed., CRC Press, Boca Raton, Fla., 1999.
Whitaker, Jerry C., Maintaining Electronic Systems, CRC Press, Boca Raton, Fla., 1991.
Whitaker, Jerry C., Radio Frequency Transmission Systems: Design and Operation,
Whythe, D. J.: “Specification of the Impedance of Transmitting Aerials for Monochrome and
Color Television Signals,” Tech. Rep. E-115, BBC, London, 1968.
g g
Chapter
5.1
Radio Antenna Principles
5.1.1 Introduction
Transmission is accomplished by the emission of coherent electromagnetic waves in free space
from one or more radiating elements that are excited by RF currents. Although, by definition, the
radiated energy is composed of mutually dependent magnetic and electric vector fields, it is con-
ventional practice to measure and specify radiation characteristics in terms of the electric field
only.
5.1.2 Fundamental Parameters

The electrical length of a radiating element is the most basic parameter of an antenna, specifi-
cally
Ht F o
H = ------------ (5.1.1)
2733
Where:
H = length of the radiating element in electrical degrees
Ht = length of the radiating element in feet
Fo = frequency of operation in kHz
When the radiating element is measured in meters
Ht F o
H = ---------------- (5.1.2)
833.23
The radiation resistance of an antenna is defined by
5-11
p
5-12 Transmitting Antennas and Systems
P
R = ---- (5.1.3)
2
I
Where:
R = radiation resistance
P = power delivered to the antenna
I = driving current at the antenna base
5.1.2a Polarization
Polarization is the angle of the radiated electric field vector in the direction of maximum radia-
tion. Antennas may be designed to provide horizontal, vertical, or circular polarization. Horizon-
tal or vertical polarization is determined by the orientation of the radiating element with respect
to earth. If the plane of the radiated field is parallel to the ground, the signal is said to be horizon-
tally polarized. If it is at right angles to the ground, it is said to be vertically polarized. When the
receiving antenna is located in the same plane as the transmitting antenna, the received signal
strength will be maximum.
Circular polarization (CP) of the transmitted signal results when equal electrical fields in the
vertical and horizontal planes of radiation are out-of-phase by 90° and are rotating a full 360° in
one wavelength of the operating frequency. The rotation can be clockwise or counterclockwise,
depending on the antenna design. This continuously rotating field gives CP good signal penetra-
tion capabilities because it can be received efficiently by an antenna of any random orientation.
Figure 5.1.1 illustrates the principles of circular polarization.
5.1.2b Antenna Beamwidth

Beamwidth in the plane of the antenna is the angular width of the directivity pattern where the
power level of the received signal is down by 50 percent (3 dB) from the maximum signal in the
desired direction of reception.
5.1.2c Antenna Gain

Directivity and gain are measures of how well energy is concentrated in a given direction. Direc-
tivity is the ratio of power density in a specified direction to the power density that would be pro-
duced if the energy were radiated isotropically. The reference can be linearly or circularly
polarized. Directivity is usually given in dBi (decibels above isotropic).
Gain is the field intensity produced in a given direction by a fixed input power to the antenna,
referenced to a dipole. Gain is frequently used as a figure of merit. Gain is closely related to
directivity, which in turn is dependent upon the radiation pattern. High values of gain are usually
obtained with a reduction in beamwidth.
An antenna is typically configured to exhibit “gain” by narrowing the beamwidth of the radi-
ated signal to concentrate energy toward the intended coverage area. The actual amount of energy
being radiated is the same with a unity gain antenna or a high gain antenna, but the useful energy
(commonly referred to as the effective radiated power, or ERP) can be increased significantly.
p
Radio Transmitting Antennas 5-13
Figure 5.1.1 Polarization of the electric field of a transmitted wave.
Electrical beam tilt can also be designed into a high gain antenna. A conventional antenna
typically radiates more than half of its energy above the horizon. This energy is lost for practical
purposes in most applications. Electrical beam tilt, caused by delaying the RF current to the
lower elements of a multi-element antenna, can be used to provide more useful power to the ser-
vice area.
Pattern optimization is another method that may be used to maximize radiation to the
intended service area. The characteristics of the transmitting antenna are affected, sometimes
p
greatly, by the presence of the sup-

porting tower, if side-mounted, or by
nearby tall obstructions (such as
another transmitting tower) if top-
mounted. Antenna manufacturers use
various methods to reduce pattern
distortions. These generally involve
changing the orientation of the radia-
tors with respect to the tower, and
adding parasitic elements.
5.1.2d Space Regions

Insofar as the transmitting antenna is
concerned, space is divided into three Figure 5.1.2 L network parameters.
regions:
• Reactive near-field region. This
region is the area of space immediately surrounding the antenna in which the reactive compo-
nents predominate. The size of the region varies, depending on the antenna design. For most
antennas, the reactive near-field region extends 2 wavelengths or less from the radiating ele-
ments.
• Radiating near-field region. This region is characterized by the predictable distribution of the
radiating field. In the near-field region, the relative angular distribution of the field is depen-
dent on the distance from the antenna.
• Far-field region. This region is characterized by the independence of the relative angular dis-
tribution of the field with varying distance. The pattern is essentially independent of distance.
5.1.3 Impedance Matching

Most practical antennas require some form of impedance matching between the transmission
line and the radiating elements. The implementation of a matching network can take on many
forms, depending upon the operating frequency and output power.
The negative sign convention is generally used in impedance matching analysis. That is, if a
network delays or retards a signal by θ degrees, the phase shift across the network is said to be
minus θ degrees. For example, a 1/4-wavelength of transmission line, if properly terminated, has
a phase shift of –90°. Thus a lagging or low-pass network has a negative phase shift, and a lead-
ing or high-pass network has a positive phase shift. There are three basic network types that can
be used for impedance matching: L, pi, and tee.
5.1.3a L Network
The L network is shown in Figure 5.1.2. Common equations are as follows:
p
R2 X1 R2
Q = ----- –1 = = (5.1.4)
R1 X2
----- -----
R1
±R2
X2 = -------- (5.1.5)
Q
–R1R2
X1 = (5.1.6)
X2
---------------
⎛R ⎞
θ = tan−1⎜ 2 ⎟
⎝ X2 ⎠
(5.1.7)
Where:
R 1 = L network input resistance, Ω
R 2 = L network output resistance, Ω
X 1 = series leg reactance, Ω
X 2 = shunt leg reactance, Ω
Q = loaded Q of the L network
The loaded Q of the network is determined from Equation (5.1.4). Equation (5.1.5) defines
the shunt leg reactance, which is negative (capacitive) when θ is negative, and positive (induc-
tive) when θ is positive. The series leg reactance is found using Equation (5.1.6), the phase shift
via Equation (5.1.7), and the currents and voltages via Ohm's Law. Note that R2 (the resistance
on the shunt leg side of the L network) must always be greater than R1. An L network cannot be
used to match equal resistances, or to adjust phase independently of resistance.
5.1.3b Tee Network

The tee network is shown in Figure
5.1.3. This configuration can be used to
match unequal resistances. The tee net-
work has the added feature that phase
shift is independent of the resistance
transformation ratio. A tee network can
be considered simply as two L networks
back-to-back. Note that there are two
loaded Qs associated with a tee net-
work: an input Q and output Q. In order
to gauge the bandwidth of the tee net-
work, the lower value of Q must be
Figure 5.1.3 T network parameters.
p
ignored. Note also that the Q of a tee network increases with increasing phase shift. Common
equations are as follows:
R1 R2
X3 = ---------------- (5.1.8)
sin θ
R
1
X
1 = ----------- – X
3 (5.1.9)
tan θ
R
2
X
2 = ----------- – X
3 (5.1.10)
tan θ
X
Q = 1 (5.1.11)
1 -----
R
1
X
Q = 2 (5.1.12)
2 -----
R
2
P
I
1 = ----- (5.1.13)
R
1
P
I
2 = ----- (5.1.14)
R
2
2 2
I
3 = I
1 + I 2 – 2I 1 I 2 cos θ (5.1.15)
2
R
3 = ( Q2 + 1) R2 (5.1.16)
⎛X ⎞ ⎛X ⎞
θ = tan−1⎜ 1 ⎟± tan−1⎜ 2 ⎟
⎝ R1 ⎠ ⎝ R2 ⎠
(5.1.17)
Where:
R 1 = tee network input resistance, Ω
R 2 = tee network output resistance, Ω
R 3 = midpoint resistance of the network, Ω
I1 = tee network input current, A
I2 = tee network output current, A
p
Figure 5.1.4 Pi network parameters.
I3 = shunt element current, A

X 1 = network input element reactance, Ω
X 2 = network output element reactance, Ω
X 3 = network shunt element reactance, Ω
P = input power, W
Q1 = input loaded Q
Q2 = output loaded Q
Equations (5.1.8) through (5.1.17) describe the tee network. It is a simple matter to find the
input and output currents via Ohm's Law, and the shunt leg current can be found via the Cosine
Law, Equation (5.1.15). Note that this current increases with increasing phase shift. Equation
(5.1.16) describes the mid-point resistance of a tee network, which is always higher than R1 or
R 2. Equation (5.1.17) is useful when designing a phantom tee network; that is, where X2 is made
up only of the antenna reactance, and there is no physical component in place of X2. Keep in
mind that a tee network is considered as having a lagging or negative phase shift when the shunt
leg is capacitive (X3 negative), and vice versa. The input and output arms can go either negative
or positive, depending on the resistance transformation ratio and desired phase shift.
5.1.3c Pi Network
The pi network is shown in Figure 5.1.4. It can also be considered as two L networks back-to-
back and, therefore, the same comments regarding overall loaded Q apply. Common equations
are as follows:
1
Y3 = ------------------------------- (5.1.18)
– sin θ R1R2
tan θ
Y1 = ----------------- (5.1.19)
R1 – Y 3
p
tan θ
Y2 = ----------------- (5.1.20)
R2 – Y 3
Q1 = R 1 Y 1 (5.1.21)
Q2 = R 2 Y 2 (5.1.22)
V1 = R1P (5.1.23)
V2 = R2P (5.1.24)
2 2
V3 = V 1 + V 2 – 2 V 1 V 2 cos θ (5.1.25)
2
Q +1
R 3 = -----2---------- (5.1.26)
R2
Where:
R 1 = pi network input resistance, Ω
R 2 = output resistance, Ω
V 1 = input voltage, V
V 2 = output voltage, V
V 3 = voltage across series element, V
P = power input to pi network, W
Y1 = input shunt element susceptance, mhos
Y2 = output shunt element susceptance, mhos
Y3 = series element susceptance, mhos
Q1 = input loaded Q
Q2 = output loaded Q
Note that susceptances have been used in Equations (5.1.18) through (5.1.22) instead of reac-
tances in order to simplify calculations. The same conventions regarding tee network currents
apply to pi network voltages—Equations (5.1.23, 5.1.24, and 5.1.25). The mid-point resistance
of a pi network is always less than R1 or R2. A pi network is considered as having a negative or
lagging phase shift when Y3 is positive, and vice versa.
Line Stretcher
A line stretcher makes a transmission line appear longer or shorter in order to produce sideband
impedance symmetry at the transmitter PA (see Figure 5.1.5). This is done to reduce audio dis-
tortion in an envelope detector—the type of detector that most monophonic AM receivers
employ. Symmetry is defined as equal sideband resistances, and equal—but opposite sign—side-
band reactances.
p
Figure 5.1.5 Line stretcher configuration.
There are two possible points of symmetry, each 90° from the other. One produces sideband
resistances greater than the carrier resistance, and the other produces the opposite effect. One
side will create a pre-emphasis effect, and the other a de-emphasis effect.
Depending on the Q of the transmitter output network, one point of symmetry may yield
lower sideband VSWR at the PA than the other. This results from the Q of the output network
opposing the Q of the antenna in one direction, but aiding the antenna Q in the other direction.
Common line stretcher equations include:
R 2s + X 2s
Rp = ------------------ (5.1.27)
Rs
R 2s + X 2s
Xp = ------------------ (5.1.28)
Xs
Where:
Rs = series configuration resistance, Ω
R p = parallel configuration resistance, Ω
Xs = series reactance, Ω
X p = parallel reactance, Ω
p
5.1.4 Antenna Types (a)
The dipole antenna is simplest of all anten-

nas, and the building block of most other
designs. The dipole consists of two in-line (b)
rods or wires with a total length equal to 1/2-
wave at the operating frequency. Figure 5.1.6
shows the typical configuration, with two 1/
4-wave elements connected to a transmission Figure 5.1.6 Half-wave dipole antenna: (a) coni-
line. A folded dipole may be fashioned as cal dipole, (b) conventional dipole.
shown in Figure 5.1.7. Such a configuration
results in increased bandwidth and imped-
ance. Impedance can be further increased by
using rods of different diameter and by vary-
ing the spacing of the elements. The 1/4-
wave dipole elements connected to the
closely-coupled 1/2-wave element act as a
matching stub between the transmission line Figure 5.1.7 Folded dipole antenna.
and the single-piece 1/2-wave element. This
broadbands the folded dipole antenna by a
factor of 2.
A corner-reflector antenna may be formed as shown in Figure 5.1.8. A ground plane or flat
reflecting sheet is placed at a distance of 1/16- to 1/4-wavelengths behind the dipole. Gain in the
forward direction can be increased by a factor of 2 with this type of design.
5.1.4a Quarter-Wave Monopole

A conductor placed above a ground plane forms an image in the ground plane such that the
resulting pattern is a composite of the real antenna and the image antenna (see Figure 5.1.9).
The operating impedance is 1/2 of the impedance of the antenna and its image when fed as a
physical antenna in free space. An example will help illustrate this concept. A 1/4-wave mono-
pole mounted on an infinite ground plane has an impedance equal to 1/2 the free-space imped-
ance of a 1/2-wave dipole. It follows, then, that the theoretical characteristic resistance of a 1/4-
wave monopole with an infinite ground plane is 37 Ω.
Figure 5.1.8 Corner-reflector antenna.

p
For a real-world antenna, an infinite

ground plane is neither possible nor
required. An antenna mounted on a ground
plane that is 2 to 3 times the operating
wavelength has about the same impedance
as a similar antenna mounted on an infi-
nite ground plane.
5.1.4b Log-Periodic Antenna

The log-periodic antenna can take-on a Figure 5.1.9 Vertical monopole mounted above a
number of forms. Typical designs include ground plane.
the conical log spiral, log-periodic V, and
log-periodic dipole. The most common of
these antennas is the log-periodic dipole. The antenna can be fed either by using alternating con-
nections to a balanced line, or by a coaxial line running through one of the feeders from front-to-
back. In theory, the log-periodic antenna may be designed to operate over many octaves. In prac-
tice, however, the upper frequency is limited by the precision required in constructing the small
elements, feed lines, and support structure of the antenna.
5.1.4c Yagi-Uda Antenna

The Yagi-Uda is an end-fire array consisting typically of a single driven dipole with a reflector
dipole behind the driven element, and one or more parasitic director elements in front (Figure
5.1.10). Common designs use from one to 7 director elements. As the number of elements is
increased, directivity increases. Bandwidth, however, decreases as the number of elements is
increased. Arrays of more than 4 director elements are typically narrowband.
The radiating element is 1/2-wavelength at the center of the band covered. The single reflec-
tor element is slightly longer, and the director elements are slightly shorter, all spaced approxi-
mately 1/4-wavelength from each other.
Table 5.1.1 demonstrates how the number of elements determines the gain and beamwidth of
a Yagi-Uda antenna.
Figure 5.1.10 The Yagi-Uda array.

p
Table 5.1.1 Typical Characteristics of Single-Channel Yagi-Uda Arrays
Number of Elements Gain, dB Beamwidth, Degrees

2 3 to 4 65
3 6 to 8 55
4 7 to 10 50
5 9 to 11 45
9 12 to 14 37
15 14 to 16 30
5.1.4d Spiral Antenna

The bandwidth limitations of an antenna are based on the
natural change in the critical dimensions of the radiating
elements caused by differing wavelengths. The spiral
antenna overcomes this limitation because the radiating
elements are specified only in angles. A two-arm equiangu-
lar spiral is shown in Figure 5.1.11. This common design
gives wideband performance. Circular polarization is
inherent in the antenna. Rotation of the pattern corresponds
to the direction of the spiral arms. The gain of a spiral
antenna is typically slightly higher than a dipole. Figure 5.1.11 Two-arm equian-
gular spiral antenna.
The basic spiral antenna radiates on both sides of the
arms. Unidirectional radiation is achieved through the addi-
tion of a reflector or cavity.
5.1.4e Array Antenna

The term “array antenna” covers a wide variety of physical structures. The most common config-
uration is the planar array antenna, which consists of a number of radiating elements regularly
spaced on a rectangular or triangular lattice. The linear array antenna, where the radiating ele-
ments are placed in a single line, is also common. The pattern of the array is the product of the
element pattern and the array configuration. Large array antennas may consist of 20 or more
radiating elements.
Correct phasing of the radiating elements is the key to operation of the system. The radiating
pattern of the structure, including direction, can be controlled through proper adjustment of the
relative phase of the elements.
5.1.5 AM Broadcast Antenna Systems

Vertical polarization of the transmitted signal is used for AM broadcast stations because of its
superior groundwave propagation, and because of the relatively simple antenna designs that it
affords. The FCC and licensing authorities in other countries have established classifications of
AM stations with specified power levels and hours of operation. Protection requirements set
forth by the FCC specify that some AM stations (in the U.S.) reduce their transmitter power at
sunset, and return to full power at sunrise. This method of operation is based on the propagation
p
Figure 5.1.12 Block diagram of an AM directional antenna feeder system for a two tower array.
characteristics of AM band frequencies. AM signals propagate further at nighttime than during

the day.
The different day/night operating powers are designed to provide each AM station with a
specified coverage area that is free from interference. Theory rarely translates directly into prac-
tice insofar as coverage is concerned, however, because of the increased interference that all AM
stations suffer at nighttime.
The tower structure visible at any AM radio station transmitter site is only half of the antenna
system. The second element is a buried ground system. Current on a tower does not simply dis-
appear, rather it returns to earth through the capacitance between the earth and the tower. Ground
losses are greatly reduced if the tower has a radial copper ground system. A typical single tower
ground system is made up of 120 radial ground wires each 140 electrical degrees long (at the
operating frequency) equally spaced out from the tower base. This is often augmented with an
additional 120 interspersed radials 50 ft long.
5.1.5a Directional AM Antenna Design

When a non-directional antenna with a given power does not radiate enough energy to serve the
station's primary service area, or radiates too much energy toward other radio stations on the
same or adjacent frequencies, it is necessary to employ a directional antenna system. Rules set
out by the FCC and regulatory agencies in other countries specify the protection requirements to
be provided by various classes of stations, for both daytime and nighttime hours. These limits
tend to define the shape and size of the most desirable antenna pattern.
A directional antenna functions by carefully controlling the amplitude and phase of the RF
currents fed to each tower in the system. The directional pattern is a function of the number and
spacing of the towers (vertical radiators), and the relative phase and magnitude of their currents.
The number of towers in a directional AM array can range from two to six or even more in a
complex system. One tower is defined as the reference tower. The amplitude and phase of the
other towers are measured relative to this reference.
A complex network of power splitting, phasing, and antenna coupling elements is required to
make a directional system work. Figure 5.1.12 shows a block diagram of a basic two tower array.
A power divider network controls the relative current amplitude in each tower. A phasing net-
p
work provides control of the phase of

each tower current, relative to the refer-
ence tower. Matching networks at the
base of each tower couple the transmis-
sion line impedance to the base operat-
ing impedance of the radiating towers.
In practice, the system shown in the
figure would not consist of individual
elements. Instead, the matching net-
work, power dividing network, and
phasing network are usually combined
into a single unit, referred to as the pha-
sor.
Antenna Pattern Design

The pattern of any AM directional
antenna system (array) is determined by
a number of factors, including:
• Electrical parameters (phase rela-
Figure 5.1.13 Radiation pattern generated with a 3-
tionship and current ratio for each tower in-line directional array using the electrical
tower) parameters and orientation shown.
• Height of each tower
• Position of each tower with respect to
the other towers (particularly with
respect to the reference tower)
A directional array consists of two
or more towers arranged in a specific
manner on a plot of land. Figure 5.1.13
shows a typical 3-tower array, and the
pattern such an array would produce.
This is an in-line array, meaning that all
the elements (towers) are in line with
one another. Notice that the major lobe
is centered on the same line as the line
of towers, and that the pattern nulls
(minima) are positioned symmetrically
about the line of towers, protecting co-
channel stations A and B at true bear-
ings of 315° and 45°, respectively.
Figure 5.1.14 shows the same array,
except that it has been rotated by 10°.
Notice that the pattern shape is not
Figure 5.1.14 Radiation pattern produced when the
changed, but the position of the major array of Figure 5.1.13 is rotated to a new orientation.
lobe and the nulls follow the line of tow-
p
ers. Also notice that the nulls no longer are

pointed at the stations to be protected. Fig-
ures 5.1.15 and 5.1.16 show that the posi-
tion of the nulls can be changed by varying
the electrical parameters so that one or the
other can be pointed in the required direc-
tion, but not both. Also, when this is
attempted, the size and shape of the small
back lobe is changed, as is the shape of the
major lobe, especially the radiation on the
line of towers.
If this directional antenna system were
constructed on a gigantic turntable, the
pattern could be rotated without affecting
the shape. But, to accomplish the required
protections and to have the major lobe(s)
oriented in the right direction, there is only
one correct position. In most cases, the
position of the towers will be specified
with respect to a single reference tower.
Figure 5.1.15 Directional pattern generated with
The location of the other towers will be
the orientation shown in Figure 5.1.13, but with dif-
given in the form of a distance and bearing ferent electrical parameters.
from that reference. Occasionally, a refer-
ence point, usually the center of the array,
will be used for a geographic coordinate
point.
5.1.5b Bearing
The bearing or azimuth of the towers from
the reference tower or point is specified
clockwise in degrees from true north. The
distinction between true and magnetic
north is vital. The magnetic North Pole is
not at the true or geographic North Pole.
(In fact, it is in the vicinity of 74° north,
101° west, in the islands of northern Can-
ada.) The difference between magnetic
and true bearings is called variation or
magnetic declination. Declination, a term
generally used by surveyors, varies for dif-
ferent locations; it is not a constant. The
earth's magnetic field is subject to a num-
ber of changes in intensity and direction.
Figure 5.1.16 Directional pattern generated by the
These changes take place over daily,
tower orientation shown in Figure 5.1.13, but with
yearly, and long-term (or secular) periods. a new set of electrical parameters.
p
Figure 5.1.17 A typical 3-tower directional antenna monitoring system.
The secular changes result in a relatively constant increase or decrease in declination over a
period of many years.
Antenna Monitoring System

Monitoring the operation of an AM directional antenna basically involves measuring the power
into the system, the relative value of currents into the towers, their phase relationships, and the
levels of radiated signal at certain monitoring points some distance from the antenna. Figure
5.1.17 shows a block diagram of a typical monitoring system for a three tower array. For systems
with additional towers, the basic layout is extended by adding more pickup elements, sample
lines, and a monitor with additional inputs.
Phase/Current Sample Loop

Two types of phase/current sample pickup elements are commonly used: the sample loop and
torroidal current transformer (TCT). The sample loop consists of a single turn unshielded loop
of rigid construction, with a fixed gap at the open end for connection of the sample line. The
device must be mounted on the tower near the point of maximum current. The loop can be used
on towers of both uniform and non-uniform cross section. It must operate at tower potential,
except for towers of less than 130° electrical height, where the loop can be operated at ground
potential.
When the sample loop is operated at tower potential, the coax from the loop to the base of the
tower is also at tower potential. In order to bring the sample line across the base of the tower, a
sample line isolation coil is used.
p
Figure 5.1.18 Three possible circuit configurations for phase sample pickup.
A shielded torroidal current transformer can also be used as the phase/current pickup ele-
ment. Such devices offer several advantages over the sample loop including greater stability and
reliability. Because they are located inside the tuning unit cabinet or house, TCTs are protected
from wind, rain, ice, and vandalism.
Unlike the rigid, fixed sample loop, torroidal current transformers are available in several sen-
sitivities, ranging from 0.25–1.0 V per ampere of tower current. Tower currents of up to 40 A or
more can be handled, providing a more usable range of voltages for the antenna monitor.
Figure 5.1.18 shows the various arrangements that can be used for phase/current sample
pickup elements.
Sample lines
The selection and installation of the sampling lines for a directional monitoring system are
important factors in the ultimate accuracy of the overall array.
With critical arrays (antennas requiring operation within tight limits specified in the station
license) all sample lines must be of equal electrical length and installed in such a manner that
corresponding lengths of all lines are exposed to equal environmental conditions.
While sample lines can be run above ground on supports (if protected and properly grounded)
the most desirable arrangement is direct burial using jacketed cable. Burial of sample line cable
p
is standard practice because proper burial offers good protection against physical damage and a
more stable temperature environment.
The Common Point

The power input to a directional antenna is measured at the phasor common point. Power is deter-
mined by the direct method
2
P = I R (5.1.29)
Where:
P = power in W
I = the common point current in A
R = the common point resistance in Ω
Monitor Points
Routine monitoring of a directional antenna involves the measurement of field intensity at cer-
tain locations away from the antenna, called monitor points. These points are selected and estab-
lished during the initial tune-up of the antenna system. Measurements at the monitor points
should confirm that radiation in prescribed directions does not exceed a value that would cause
interference to other stations operating on the same or adjacent frequencies. The field intensity
limits at these points are normally specified in the station license. Measurements at the monitor
points may be required on a weekly or a monthly basis, depending on several factors and condi-
tions relating to the particular station. If the system is not a critical array, quarterly measurements
may be sufficient.
5.1.5c Folded Unipole Antenna

The folded unipole antenna consists of a grounded vertical structure with one or more conductors
folded back parallel to the side of the structure.
It can be visualized as a half-wave folded
dipole perpendicular to the ground and cut in
half (see Figure 5.1.19). This design makes it
possible to provide a wide range of resonant
radiation resistances by varying the ratio of the
diameter of the folded-back conductors in rela-
tion to the tower. Top loading can also be used
to broaden the antenna bandwidth. A side view
of the folded unipole is shown in Figure 5.1.20.
The folded unipole antenna could be consid-
ered a modification of the standard shunt-fed
system. Instead of a slant wire that leaves the
tower at an approximate 45° angle (as used for Figure 5.1.19 The folded unipole antenna
shunt-fed systems), the folded unipole antenna can be thought of as a 1/2-wave folded dipole
has one or more wires attached to the tower at a antenna perpendicular to the ground and cut
in half.
predetermined height. The wires are supported
p
by standoff insulators and run parallel to the sides of the tower

down to the base.
The tower is grounded at the base. The folds, or wires, are
joined together at the base and driven through an impedance
matching network. Depending upon the tuning requirements
of the folded unipole, the wires may be connected to the tower
at the top and/or at predetermined levels along the tower with
shorting stubs.
The folded unipole can be used on tall (130° or greater)
towers. However, if the unipole is not divided into two parts,
the overall efficiency (unattenuated field intensity) will be
considerably lower than the normally expected field for the
electrical height of the tower.
5.1.6 FM Broadcast Antenna Systems

The propagation characteristics of VHF FM radio are much
different than for MF AM. There is essentially no difference
between day and night FM propagation. FM stations have rel-
atively uniform day and night service areas with the same
operating power.
A wide variety of antennas is available for use in the FM
broadcast band. Nearly all employ circular polarization.
Although antenna designs differ from one manufacturer to
another, generalizations can be made that apply to most units. Figure 5.1.20 The folds of the
unipole antenna are arranged
either near the legs of the
5.1.6a Antenna Types tower or near the faces of the
tower.
There are three basic classes of FM broadcast transmitting
antennas in use today: ring stub and twisted ring, shunt- and
series-fed slanted dipole, and multi-arm short helix. While each design is unique, all have the
following items in common:
• The antennas are designed for side mounting to a steel tower or pole.
• Radiating elements are shunted across a common rigid coaxial transmission line.
• Elements are placed along the rigid line every one wavelength.
• Antennas with one to seven bays are fed from the bottom of the coaxial transmission line.
• Antennas with more than seven bays are fed from the center of the array to provide more pre-
dictable performance in the field.
Antennas generally include a means of tuning out reactances after the antenna has been installed
through the adjustment of variable capacitive or inductive elements at the feed point.
Figure 5.1.21 shows a shunt-fed slanted dipole antenna that consists of two half-wave dipoles
offset 90°. The two sets of dipoles are rotated 22.5° (from their normal plane) and are delta-
matched to provide a 50 Ω impedance at the radiator input flange. The lengths of all four dipole
p
(a )
(b )
Figure 5.1.21 Mechanical configuration of one bay of a circularly-polarized FM transmitting

antenna: (a) top view of the antenna, (b) front view.
arms can be matched to resonance by mechanical adjustment of the end fittings. Shunt-feeding
(when properly adjusted) provides equal currents in all four arms.
Wideband panel antennas are a fourth common type of antenna used for FM broadcasting.
Panel designs share some of the characteristics listed previously, but are intended primarily for
specialized installations in which two or more stations will use the antenna simultaneously. Panel
antennas are larger and more complex than other FM antennas, but offer the possibility for
shared tower space among several stations and custom coverage patterns that would be difficult
or even impossible with more common designs.
The ideal combination of antenna gain and transmitter power for a particular installation
involves the analysis of a number of parameters. As shown in Table 5.1.2, a variety of pairings
can be made to achieve the same ERP.
p
Table 5.1.2 Various Combinations of Transmitter Power and Antenna Gain that will Produce 100
KW Effective Radiated Power (ERP) for an FM Station
Number of Bays Antenna Gain Required Transmitter Power (kW)*

3 1.5888 66.3
4 2.1332 49.3
5 2.7154 38.8
6 3.3028 31.8
7 3.8935 27.0
8 4.4872 23.5
10 5.6800 18.5
12 6.8781 15.3
* Effects of transmission line and related losses not included.
5.1.7 Bibliography
Benson, B., and Whitaker, J.: Television and Audio Handbook for Technicians and Engineers,
Bingeman, Grant: “AM Tower Impedance Matching,” Broadcast Engineering, Intertec Publish-
Bixby, Jeffrey: “AM DAs—Doing it Right,” Broadcast Engineering, Intertec Publishing, Over-
land Park, Kan., February 1984.
Chick, Elton B.: “Monitoring Directional Antennas,” Broadcast Engineering, Intertec Publish-
Fink, D., and D. Christiansen (eds.): Electronics Engineer's Handbook, 3rd ed., McGraw-Hill,
Jordan, Edward C. (ed.): “Reference Data for Engineers—Radio, Electronics, Computer and
Communications, 7th ed., Howard W. Sams, Indianapolis, Ind., 1985.
Mullaney, John H.: “The Folded Unipole Antenna for AM Broadcast,” Broadcast Engineering,
Intertec Publishing, Overland Park, Kan., January 1960.
Mullaney, John H.: “The Folded Unipole Antenna,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., July 1986.
Westberg, J. M.: “Effect of 90° Stub on Medium Wave Antennas,” NAB Engineering Handbook,
7the ed., National Association of Broadcasters, Washington, D.C., 1985.
p
g g
Chapter
5.2
Television Antenna Principles
Oded Ben-Dov, Krishna Praba
5.2.1 Introduction
A wide variety of options are available for television station antenna system implementations.
Classic designs, such as the turnstile, continue to be used while advanced systems, based on
panel or waveguide techniques, are implemented to solve specific design and/or coverage issues.
Regardless of its basic structure, any antenna design must focus on three key electrical consider-
ations:
• Adequate power handling capability
• Adequate signal strength over the coverage area
• Distortion-free emissions characteristics
5.2.1a Power and Voltage Rating

Television antennas for conventional (NTSC, PAL, and SECAM) video transmissions are conser-
vatively rated assuming a continuous black level. The nomenclature typically used in rating the
antenna is “peak of sync TV power + 20 percent (or 10 percent) aural.” The equivalent heating
(average) power is 0.8 of the power rating if 20 percent aural power is used and 0.7 if 10 percent
aural power is used. The equivalent heating power value is arrived at as given in Table 5.2.1.
As shown in the table, an antenna power rating increases by 14 percent when the aural output
power is reduced from 20 to 10 percent.
In the design of feed systems, the transmission lines must be derated from the manufacturer’s
catalog values (based on VSWR = 1.0) to allow for the expected VSWR under extraordinary cir-
cumstances, such as ice and mechanical damage. The derating factor is
⎛ 1 ⎞ ⎛ VSWR+1⎞2
2
⎜ ⎟ =⎜ ⎟
⎝1+ Γ ⎠ ⎝ 2 VSWR ⎠
(5.2.1)
5-33
p
Table 5.2.1 Television Antenna Equivalent Heating Power
Parameter Carrier Level (%) Fraction of Time (%) Average Power (%)
Voltage Power
Sync 100 100 8 8
Blanking 75 56 92 52
Visual black-signal power 60
Aural power (percent of sync power) 20 (or 10)
Total transmitted power (percent peak-of-sync) 80 (or 70)
where Γ = the expected reflection coefficient resulting from ice or a similar error condition. This
derating factor is in addition to the derating required because of the normally existing VSWR in
the antenna system feed line.
The manufacturer’s power rating for feed system components is based on a fixed ambient
temperature. This temperature is typically 40°C (104ºF). If the expected ambient temperature is
higher than the quoted value, a good rule of thumb is to lower the rating by the same percentage.
Hence, the television power rating (including 20 percent aural) of the feed system is given by
1 ⎛ TT L ⎞⎛ VSWR +1⎞2
PTV ≈ PT L⎜ ⎟⎜ ⎟
0.8 ⎝ T ⎠⎝ 2 VSWR ⎠
(5.2.2)
Where:
P T/V = quoted average power for transmission line components with VSWR = 1.0
T/TT/L = ratio of expected to quoted ambient temperature
VSWR = worst-possible expected VSWR
Television antennas must also be designed to withstand voltage breakdown resulting from
high instantaneous peak power both inside the feed system and on the external surface of the
antenna. Improper air gaps or sharp edges on the antenna structure and insufficient safety factors
can lead to arcing and blooming. The potential problem resulting from instantaneous peak power
is aggravated when multiplexing two or more stations on the same antenna. In this latter case, if
all stations have the same input power, the maximum possible instantaneous peak power is pro-
portional to the number of the stations squared as derived below.
For a single channel, the maximum instantaneous voltage can occur when the visual and aural
peak voltages are in phase. Thus, with 20 percent aural, the worst-case peak voltage is
Vpeak = 2 Z0 PPS + 0.4 Z0 PPS = 2.047 Z 0 PPS

(5.2.3)
Where:
P PS = peak-of-sync input power
Z0 = characteristic impedance
and the equivalent peak power is
p
2
Vpeak
Ppeak = = 4.189 PPS
Z0
(5.2.4)
For N stations multiplexed on the same antenna, the equivalent peak voltage is
1
Vpeak = 2.047 PPS + 2.047 PPS + ... 2.047 N PPS
Z0
(5.2.5)
and the equivalent peak power is
Ppeak = 4.189 N 2 PPS

(5.2.6)
Experience has shown that the design peak power and the calculated peak power should be
related by a certain safety factor. This safety factor is made of two multipliers. The first value,
typically 3, is for the surfaces of pressurized components. This factor accounts for errors result-
ing from calculation and fabrication and/or design tolerances. The second value, typically 3,
accounts for humidity, salt, and pollutants on the external surfaces. The required peak power
capability, thus, is
Ps = 4.189× Fs × N 2 × PPS
(5.2.7)
Where:
P s = safe peak power
F s = 3 for internal pressurized surfaces
F s = 9 for internal external surfaces
5.2.1b Effective Radiated Power

The effective radiated power (ERP) of an antenna is defined as
PERP = Pinput ×Gantenna
The input power to the antenna (Pinput) is the transmitter output power minus the losses in the
interconnection hardware between the transmitter output and the antenna input. This hardware
consists of the transmission lines and the filtering, switching, and combining complex.
p
The gain of an antenna (Gantenna) denotes its ability to concentrate the radiated energy toward
the desired direction rather than dispersing it uniformly in all directions. It is important to note
that higher gain does not necessarily imply a more effective antenna. It does mean a lower trans-
mitter power output to achieve the licensed ERP.
The visual ERP, which must not exceed the FCC specifications, depends on the television
channel frequency, the geographical zone, and the height of the antenna-radiation center above
average terrain. The FCC emission rules relate to either horizontal or vertical polarization of the
transmitted field. Thus, the total permissible ERP for circularly polarized transmission is dou-
bled.
The FCC-licensed ERP is based on average-gain calculations for omnidirectional antennas
and peak-gain calculations for antennas designed to emit a directional pattern.
5.2.1c Directional Antennas

As defined by the FCC rules, an antenna that is intentionally designed or altered to produce a
noncircular azimuthal radiation pattern is a directional antenna. There are a variety of reasons
for designing directional antennas. In some instances, such designs are necessary to meet inter-
ference protection requirements. In other instances, the broadcaster may desire to improve ser-
vice by diverting useful energy from an unpopulated area toward population centers. Generally
speaking, directional antennas are more expensive than omnidirectional antennas.
5.2.2 Polarization
The polarization of an antenna is the orientation of the electric field as radiated from the device.
When the orientation is parallel to the ground in the radiation direction-of-interest, it is defined
as horizontal polarization. When the direction of the radiated electric field is perpendicular to
the ground, it is defined as vertical polarization. These two states are shown in Figure 5.2.1.
Therefore, a simple dipole arbitrarily can be oriented for either horizontal or vertical polariza-
tion, or any tilted polarization between these two states.
If a simple dipole is rotated at the picture carrier frequency, it will produce circular polariza-
tion (CP), since the orientation of the radiated electric field will be rotating either clockwise or
counterclockwise during propagation. This is shown in Figure 5.2.1. Instead of rotating the
antenna, the excitation of equal longitudinal and circumferential currents in phase quadrature
will also produce circular polarization. Because any state of polarization can be achieved by judi-
cious choice of vertical currents, horizontal currents, and their phase difference, it follows that
the reverse is also true. That is, any state of polarization can be described in terms of its vertical
and horizontal phase and amplitude components.
Perfectly uniform and symmetrical circular polarization in every direction is not possible in
practice. Circular polarization is a special case of the general elliptical polarization, which char-
acterizes practical antennas. Other special cases occur when the polarization ellipse degenerates
into linear polarization of arbitrary orientation.
p
Figure 5.2.1 Polarizations of the electric field of the transmitted wave.
5.2.2a Practical Application

The polarization of the electric field of television antennas was limited in the U.S. to horizontal
polarization during the first 30 years of broadcasting. But in other parts of the world, both verti-
cal and horizontal polarizations were allowed, primarily to reduce co-channel and adjacent chan-
nel interference [1]. The FCC modified the rules in the early 1970s to include circularly
polarized transmission for television broadcasting.
p
Figure 5.2.2 Elliptical polarization as a combination of two circularly polarized signals. As shown,
the resultant ellipse equals the right-hand component plus the left-hand component.
It should be pointed out that the investment required in a circularly polarized transmission
facility is considerably greater than that required for horizontal polarization, mostly because of
the doubling of transmitter power. While doubling of antenna gain instead of transmitter power is
possible, the coverage of high-gain antennas with narrow beam-width may not be adequate
within a 2- to 3-mi (3- to 5-km) radius of the support tower unless a proper shaping of the eleva-
tion pattern is specified and achievable.
Practical antennas do not transmit ideally perfect circular polarization in all directions. A
common figure of merit of CP antennas is the axial ratio. The axial ratio is not the ratio of the
vertical to horizontal polarization field strength. The latter is called the polarization ratio. Practi-
cal antennas produce elliptical polarization; that is, the magnitude of the propagating electric
field prescribes an ellipse when viewed from either behind or in front of the wave. Every ellipti-
cally polarized wave can be broken into two circularly polarized waves of different magnitudes
and sense of rotation, as shown in Figure 5.2.2. For television broadcasting, usually a right-hand
(clockwise) rotation is specified when viewed from behind the outgoing wave, in the direction of
propagation.
Referring to Figure 5.2.2, the axial ratio of the elliptically polarized wave can be defined in
terms of the axes of the polarization ellipse or in terms of the right-hand and left-hand compo-
nents. Denoting the axial ratio as R, then
p
E1 E R + E L
R= =
E 2 ER − E L
(5.2.8)
When evaluating the transfer of energy between two CP antennas, the important performance
factors are the power-transfer coefficient and the rejection ratio of the unwanted signals (echoes)
to the desired signal. Both factors can be analyzed using the coupling-coefficient factor between
two antennas arbitrarily polarized. For two antennas whose axial ratios are R1 and R2, the cou-
pling coefficient is
1⎡ 4 R1 R2 (1− R12 )(1− R22 ) ⎤

f = ⎢1± + cos ( 2α)⎥
⎣ (1+ R1 )(1+ R2 ) (1+ R1 )(1+ R2 )
2⎢ ⎥
2 2 2 2
⎦
(5.2.9)
where α = the angle between the major axes of the individual ellipses of the antennas.
The plus sign in Equation (5.2.9) is used if the two antennas have the same sense of rotation
(either both right hand or left hand). The minus sign is used if the antennas have opposite senses
of polarization.
Two special cases are of importance when coupling between two elliptically polarized anten-
nas is considered. The first is when the major axes of the two ellipses are aligned (α = 0). The
second case is when the major axes are perpendicular to each other (α = ± π / 2).
For case 1, where the major axes of the polarization ellipses are aligned, the maximum power
transfer is
(1± R1 R2 )
2
f=
(1+ R12 )(1+ R22 )
(5.2.10)
and the minimum power rejection ratio is
f− (1− R1 R2 )
2
=
f+ (1+ R R )2
1 2
(5.2.11)
For case 2, where the major axes of the two polarization ellipses are perpendicular, the maxi-
mum power transfer is
p
( R1 ± R2 )
2
f=
(1+ R12 )(1+ R22 )
(5.2.12)
and the minimum power rejection ratio is
f− ( R1 − R2 )
2
=
f+ ( R + R )2
1 2
(5.2.13)
The ability to reject unwanted reflections is of particular importance in television transmis-

sion. Because in many cases the first-order reflections have undergone a reversal of the sense of
rotation of the polarization ellipse, the minimum rejection ratio is given by f- /f+.
5.2.3 Gain
Antenna gain is a figure of merit that describes the antenna’s ability to radiate the input power
within specified sectors in the directions of interest. Broadcast antenna gains are defined relative
to the peak gain of a half-wavelength-long dipole. Gain is one of the most critical figures of
merit of the television broadcast antenna, as it determines the transmitter power required to
achieve a given ERP. Gain is related to the beamwidth of the elevation pattern, which in turn
affects the coverage and sets limits on the allowable windsway. It is related to height (windload)
of the antenna and to the noncircularity of the azimuthal pattern.
The gain of any antenna can be estimated quickly from its height and knowledge of its azi-
muthal pattern. It can be calculated precisely from measurements of the radiation patterns. It can
also be measured directly, although this is rarely done for a variety of practical reasons. The gain
of television antennas is always specified relative to a half-wavelength dipole. This practice dif-
fers from that used in nonbroadcast antennas.
Broadcast antenna gains are specified by elevation (vertical) gain, azimuthal (horizontal)
gain, and total gain. The total gain is specified at either its peak or average (rms) value. In the
U.S., the FCC allows average values for omnidirectional antennas but requires the peak-gain
specification for directional antennas. For circularly polarized antennas, the aforementioned
terms also are specified separately for the vertically and horizontally polarized fields.
For an explanation of elevation gain, consider the superimposed elevation patterns in Figure
5.2.3. The elevation pattern with the narrower beamwidth is obtained by distributing the total
input power equally among 10 vertical dipoles stacked vertically 1 wavelength apart. The wider-
beamwidth, lower peak-amplitude elevation pattern is obtained when the same input power is fed
into a single vertical dipole. Note that at depression angles below 5°, the lower-gain antenna
transmits a stronger signal. In the direction of the peak of the elevation patterns, the gain of the
10 dipoles relative to the single dipole is
p
Figure 5.2.3 Elevation pattern gain of two antennas of different aperture.
⎛ 1.0 ⎞
2
ge =⎜ ⎟ = 10
⎝ 0.316 ⎠
(5.2.14)
It can be seen from Figure 5.2.3 that the elevation gain is proportional to the vertical aperture
of the antenna. The theoretical upper limit of the elevation gain for practical antennas is given by
A
GE = 1.22 η
λ
(5.2.15)
Where:
η = feed system efficiency
A/λ = the vertical aperture of the antenna in wavelengths of the operating television channel
In practice, the elevation gain varies from η × 0.85 A/λ to η × 1.1 A/λ.
To completely describe the gain performance of an antenna, it is necessary to also consider
the azimuth radiation pattern. An example pattern is shown in Figure 5.2.4.
p
Figure 5.2.4 Measured azimuthal pattern of a three panel array.
5.2.3a Circular Polarization

To clarify the factors contributing to the gain figure for circular polarized antennas, the deriva-
tion of the applicable expressions for the gain of each polarization and the total gain follow.
The total gain of a mathematical model of an antenna can be described as the sum of the indi-
vidual gains of the field polarized in the vertical plane and the field polarized in the horizontal
plane regardless of their ratio.
Starting with the standard definition of antenna gain G
2 2
E v + Eh
G = 4 πη
1.64 ∫∫ Ev + Eh sin θ d θ d φ
2 2
(5.2.16)
Where E v and Eh equal the magnitudes of the electric fields of the vertical polarization and the
horizontal polarization, respectively, in the direction of interest. This direction usually is the peak
of the main beam.
p
Next, it is necessary to define
2
Eh
Gh = 4 πη
1.64 ∫∫ Eh sin θ d θ d φ
2
(5.2.17)
as the total (azimuthal and elevation) gain for the horizontal polarization component in the
absence of vertical polarization, and then let
2
Eh
Gv = 4 πη
1.64 ∫∫ Eh sin θ d θ d φ
2
(5.2.18)
be the total gain for the vertical polarization component in the absence of horizontal polarization.
Then
2
1+ Ev Eh
2
G 1+ P
= =
Gh 1+ (G G ) E E 1+ (Gh Gv ) P
2 2
h v v h
(5.2.19)
and
2
1+ Eh Ev
2
G 1+ P
= = 2
Gv 1+ (G G ) E E P + ( G v Gh )
2
v h h v
(5.2.20)
where ⎜P ⎜ is the magnitude of the polarization ratio. When the last two expressions are added
and rearranged, the total gain G of the antenna is obtained as
⎡ 1 ⎤
G =⎡
2⎤
⎣1+ P ⎦⎢ ⎥
⎢ (1 Gh )+ P Gv ⎦
⎥
2
⎣
(5.2.21)
The total gain can be broken into two components whose ratio is ⎜P ⎜2. For horizontal polariza-
tion
p
Gh
G=
1+(Gh Gv ) / P
2
(5.2.22)
and for vertical polarization
Gv
G=
1+(Gv Gh ) / P
2
(5.2.23)
The last three expressions specify completely any antenna provided Gh, Gv, and ⎜P ⎜ are
known. From the definitions it can be seen that the first two can be obtained from measured-pat-
tern integration, and the magnitude of the polarization ratio is either known or can be measured.
When the antenna is designed for horizontal polarization, ⎜P ⎜ = 0 and G = Gh. For circular
polarization, ⎜P ⎜ = 1 in all directions, Gh = Gv and the gain of each polarization is half of the
total antenna gain.
5.2.3b Elevation Pattern Shaping

The elevation pattern of a vertically stacked array of radiators can be computed from the illumi-
nation or input currents to each radiator of the array and the elevation pattern of the single radia-
tor [2–5]. Mutual coupling effects generally can be ignored when the spacing between the
adjacent radiators is approximately a wavelength, a standard practice in most broadcast antenna
designs. The elevation pattern E(θ) of an antenna consisting of N vertically stacked radiators, as
a function of the depression angle θ, is given by
N ⎛ 2π ⎞
E ( θ ) = ∑ Ai Pi ( θ ) exp ( j φ) exp⎜ j di sin θ ⎟
i=1 ⎝ λ ⎠
(5.2.24)
Where:
P i (θ) = vertical pattern of ith panel
Ai = amplitude of current in ith panel
θi = phase of current in ith panel
di = location of ith panel
In television applications, the normalized magnitude of the pattern is of primary interest. For
an array consisting of N identical radiators, spaced uniformly apart (d), and carrying identical
currents, the magnitude of the elevation, or vertical, radiation pattern, is given by
p
⎣( N π λ) d sin θ ⎦
sin⎡ ⎤
E (θ) = p (θ)
⎡( πd λ) sin θ ⎦
sin⎣ ⎤
(5.2.25)
where the first part of the expression on the right is commonly termed the array factor. The ele-
vation pattern of a panel antenna comprising six radiators, each 6 wavelengths long, is given in
Figure 5.2.5. The elevation-pattern power gain ge of such an antenna can be determined by inte-
grating the power pattern, and is given by the expression
ge ≅
Nd
------ (5.2.26)
λ
Thus, for an antenna with N half-wave dipoles spaced 1 wavelength apart, the gain is essen-
tially equal to the number of radiators N. In practice, slightly higher gain can be achieved by the
use of radiators which have a gain greater than that of a half-wave dipole.
The array factor becomes zero whenever the numerator becomes zero, except at θ = 0 when
its value equals 1. The nulls of the pattern can be easily determined from the equation
N πd
sin θ m = m
λ
(5.2.27)
or
m
θ m = sin−1
ge
(5.2.28)
where m = 1, 2, 3 ... refers to the null number and θ m = depression angle at which a null is
expected in radians.
The approximate beamwidth corresponding to 70.7 percent of the field (50 percent of power),
or 3 dB below the maximum, can be determined from the array factor. The minimum beamwidth
is given by the expression
50.8 50.8
BWmin = = deg
N ( d λ) ge
(5.2.29)
It is interesting to note that the gain-beamwidth product is essentially a constant. The gain-
beamwidth product of practical antennas varies from 50.8 to 68, depending on the final shaping
of the elevation pattern.
p
Figure 5.2.5 Elevation pattern of an antenna array of six radiators, each six wavelengths long. All
radiators feed equally and in-phase.
In television broadcasting, the desired gain of an omnidirectional antenna generally is deter-

mined by the maximum-allowable effective radiated power and the transmitter power output to
the antenna. Provision for adequate signal strength in the coverage area of interest requires syn-
thesis of the antenna pattern to ensure that the nulls of the pattern are filled to an acceptable
level. In addition, in order to improve the transmission efficiency, the main beam is tilted down
to reduce the amount of unusable radiated power directed above the horizontal.
p
Although the previous discussion has been concerned with arrays of discrete elements, the
concept can be generalized to include antennas that are many wavelengths long in aperture and
have a continuous illumination. In such cases, the classic techniques of null filling by judicious
antenna aperture current distribution, such as cosine-shaped or exponential illuminations, are
employed.
Elevation-Pattern Derivation
The signal strength at any distance from an antenna is directly related to the maximum ERP of
the antenna and the value of the antenna elevation pattern toward that depression angle. For a sig-
nal strength of 100 mV/m, which is considered adequate for analog television applications, and
assuming the FCC (50,50) propagation curves, the desired elevation pattern of an antenna can be
derived. Such curves provide an important basis for antenna selection.
Beam Tilting
Any specified tilting of the beam below the horizontal is easily achieved by progressive phase
shifting of the currents (It) in each panel in an amount equal to the following
I t =−2 ( πdi λ) sin θT

(5.2.30)
where θ T is the required amount of beam tilt in degrees. Minor corrections, necessary to account
for the panel pattern, can be determined easily, by iteration.
In some cases, a progressive phase shifting of individual radiators may not be cost effective,
and several sections of the array can have the illuminating current of the same phase, thus reduc-
ing the number of different phasing lengths. In this case, the correct value of the phase angle for
each section can be computed.
When the specified beam tilt is comparable with the beam width of the elevation pattern, the
steering of the array reduces the gain from the equivalent array without any beam tilt. To improve
the antenna gain, mechanical tilt of the individual panels, as well as the entire antenna, is
resorted to in some cases. However, mechanically tilting the entire antenna results in variable
beam tilt around the azimuth.
Null-Fill
Another important antenna criterion is the null-fill requirement. If the antenna is near the center
of the coverage area, depending on the minimum gain, the nulls in the coverage area must be
filled in to provide a pattern that lies above the 100-mV/m curve for that particular height. For
low-gain antennas, this problem is not severe, especially when the antenna height is lower than
2000 ft (610 m) and only the first null has to be filled. But in the case of UHF antennas, with
gains greater than 20, the nulls usually occur close to the main beam and at least two nulls must
be filled.
When the nulls of a pattern are filled, the gain of the antenna is reduced. A typical gain loss of
1 to 2 dB generally is encountered.
The variables for pattern null-filling are the spacing of the radiators and the amplitudes and
phases of the feed currents. The spacings generally are chosen to provide the minimum number
p
of radiators necessary to achieve the required gain. Hence, the only variables are the 2(N – 1) rel-
ative amplitudes and phases.
The distance from the antenna to the null can be approximated if the gain and the height of the
antenna above average terrain are known. Because the distance at any depression angle θ can be
approximated as
H
D = 0.0109
θ
(5.2.31)
with H equal to the antenna height in feet, and the depression angle of the mth null equal to
⎛m⎞
θ m = 57.3sin−1⎜ ⎟ ge =elevation power gain of antenna
⎝ ge ⎠
(5.2.32)
then
H ge m m
Dm = 0.00019 miles for sin ≈
m ge ge
(5.2.33)
is the distance from the antenna to the mth null.

A simple method of null-filling is by power division among the vertically stacked radiators.
For example, a 70:30 power division between each half of an array produces a 13 percent fill of
the first null. Power division usually is employed where only the first null is to be filled.
Another approach to null-filling by changing only the phases of the currents is useful because
the input power rating of the antenna is maximized when the magnitude of the current to each
radiator in the array is adjusted to its maximum value. For example, if the bottom and the top lay-
ers of an N-layer array differ in phase by θ from the rest, the first (N/2) – 1 nulls are filled.
In practice, the elevation pattern is synthesized, taking into consideration all the design con-
straints, such as power-rating of the individual radiators and the restrictions imposed on the feed
system because of space, access, and other parameters. Beam tilting is achieved by progressive
or discrete phasing of sections of the antenna. The pattern shown in Figure 5.2.6 illustrates the
final design of the array pattern of a high-gain antenna, determined by a computer-aided iteration
technique.
5.2.3c Azimuthal-Pattern Shaping

For omnidirectional antennas, a circular azimuthal pattern is desired. However, in practice, the
typical circularity may differ from the ideal by ±2 dB. The omnidirectional pattern is formed by
the use of several radiators arranged within a circle having the smallest-possible diameter. If a
single radiator pattern can be idealized for a sector, several such radiators can produce truly cir-
p
Figure 5.2.6 High-gain UHF antenna (ERP = 5 MW, gain = 60, beam tilt = 0.75°).
cular patterns (see Figure 5.2.7). Practical single-radiator-element patterns do not have sharp
transitions, and the resultant azimuthal pattern is not a perfect circle. Furthermore, the interradi-
ator spacing becomes important, because for a given azimuth, the signals from all the radiators
add vectorially. The resultant vector, which depends on the space-phase difference of the individ-
ual vectors, varies with azimuth, and the circularity deteriorates with increased spacing.
In the foregoing discussion, it is assumed that the radiators around the circular periphery are
fed in-phase. Similar omnidirectional patterns can be obtained with radial-firing panel antennas
when the panels differ in phase uniformly around the azimuth wherein the total phase change is a
multiple of 360°. The panels then are offset mechanically from the center lines as shown in Fig-
ure 5.2.8. The offset is approximately 0.19 wavelength for triangular-support towers and 0.18
wavelength for square-support towers. The essential advantage of such a phase rotation is that,
when the feedlines from all the radiators are combined into a common junction box, the first-
order reflections from the panels tend to cancel at the input to the junction box, resulting in a
considerable improvement in the input impedance.
The azimuthal pattern of a panel antenna is a cosine function of the azimuthal angle in the
front of the radiator, and its back lobe is small. The half-voltage width of the frontal main lobe
ranges from 90° for square-tower applications to 120° for triangular towers. The panels are
affixed on the tower faces. Generally, a 4-ft-wide tower face is small enough for U. S. channels 7
p
Figure 5.2.7 Horizontal pattern formation, ideal case.
to 13. For channels 2 to 6, the tower-face size could be as large as 10 ft (3 m). Table 5.2.2 lists
typical circularity values.
In all the previous cases, the circular omnidirectional pattern of the panel antenna is achieved
by aligning the main beam of the panels along the radials. This is the radial-fire mode. Another
technique utilized in the case of towers with a large cross section (in wavelengths) is the tangen-
tial-fire mode. The panels are mounted in a skewed fashion around a triangular or square tower,
as shown in Figure 5.2.9. The main beam of the panel is directed along the tangent of the circum-
scribed circle as indicated in the figure. The optimum interpanel spacing is an integer number of
wavelengths when the panels are fed in-phase. When a rotating-phase feed is employed, correc-
tion is introduced by modifying the offset—for example, by adding 1/3 wavelength when the
rotating phase is at 120°. The table of Figure 5.2.9 provides the theoretical circularities for ideal
elements. Optimization is achieved in practice by model measurements to account for the back
lobes and the effect of the tower members.
A measured pattern of such a tangential-fire array is shown in Figure 5.2.10.
In the case of directional antennas, the desired pattern is obtained by choosing the proper
input-power division to the panels of each layer, adjusting the phase of the input signal, and/or
mechanically tilting the panels. In practice, the azimuthal-pattern synthesis of panel antennas is
done by superposition of the azimuthal phase and amplitude patterns of each radiator, while
adjusting the amplitudes and phases of the feed currents until a suitable configuration of the pan-
els on the tower yields the desired azimuthal pattern.
p
Figure 5.2.8 Offset radial firing panels.
The turnstile antenna is a top-mounted omnidirectional pole antenna. Utilizing a phase rota-
tion of four 90° steps, the crossed pairs of radiators act as two dipoles in quadrature, resulting in
a fairly smooth pattern. The circularity improves with decreasing diameter of the turnstile sup-
port pole.
The turnstile antenna can be directionalized. As an example, a peanut-shaped azimuthal pat-
tern can be synthesized, either by power division between the pairs of radiators or by introducing
proper phasing between the pairs. The pattern obtained by phasing the pairs of radiators by 70°
instead of the 90° used for an omnidirectional pattern is shown in Figure 5.2.11.
A directional pattern of a panel antenna with unequal power division is shown in Figure
5.2.12. The panels are offset to compensate for the rotating phase employed to improve the input
impedance of the antenna.
p
Table 5.2.2 Typical Circularities of Panel Antennas
Shape Tower-face Size, ft (m) Circularity, ± dB

Channels 2–6 Channels 7–13
Triangular 5 (1.5) 0.9 1.8
6 (1.8) 1.0 2.0
7 (2.1) 1.1 2.3
10 (3.0) 1.3 3.0
4 (1.2) 0.5 1.6
Square 5 (1.5) 0.6 1.9
6 (1.8) 0.7 2.4
7 (2.1) 0.8 2.7
10 (3.0) 1.2 3.2
Figure 5.2.9 Tangential-fire mode. Note that with back lobes and tower reflections, the circularities
tend to be worse by about 2 dB than those given here, especially for large tower sizes.
5.2.4 Voltage Standing-Wave Ratio and Impedance

The transmission line connecting the transmitter to the antenna is never fully transparent to the
incoming or incident wave. Because of imperfections in manufacture and installation, some of
p
Figure 5.2.10 Measured azimuthal pattern of a “tangential-fire” array of three panels fed in-phase
around a triangular tower with D = 7.13 wavelengths.
the power in the incident wave can be reflected at a number of points in the line. Additional
reflection occurs at the line-to-antenna interface, because the antenna per se presents an imper-
fect match to the incident wave. These reflections set up a reflected wave that combines with the
incident wave to form a standing wave inside the line. The characteristic of the standing-wave
pattern is periodic maximums and minimums of the voltage and current along the line. The ratio
of the maximum to minimum at any plane is called the voltage standing-wave ratio (VSWR).
Because the VSWR is varying along the transmission line, the reference plane for the VSWR
measurement must be defined. When the reference plane is at the gas stop input in the transmit-
ter room, the measured value is system VSWR. When the reference plane is at the input to the
antenna on the tower, the measured value is the antenna VSWR. The system VSWR differs from
the antenna VSWR owing to the introduction of standing waves by the transmission line. When
two sources of standing waves S1 and S2 exist, then the maximum VSWR is
VSWRmax = S1S2 (5.2.34)

p
Figure 5.2.11 Directionalizing the superturnstile antenna by phasing.
and the minimum VSWR is
S2
VSWR min = for S1 < S2
S1
(5.2.35)
More generally, the expected system VSWR for n such reflections is, for the maximum case
VSWR max = S1S2 S3 ... S n

(5.2.36)
and for the minimum case

p
Figure 5.2.12 Directional panel antenna implemented using power division and offsetting.
Sn
VSWR min = for S1 < S2 ... Sn
S1S2 S3 ... Sn−1
(5.2.37)
If the calculated minimum VSWR is less than 1.00, then the minimum VSWR = 1.00.
As an example, consider an antenna with an input VSWR of 1.05 at visual and 1.10 at aural
carrier frequencies. If the transmission line per se has a VSWR of 1.05 at the visual and 1.02 at
aural carriers, the system VSWR will be between
1.05
1.00 = ≤ S vis ≤ 1.05×1.05 = 1.103
1.05
(5.2.38)
p
1.10
1.078 = ≤ Saural ≤1.10 ×1.02 = 1.122
1.02
(5.2.39)
The VSWR resulting from any reflection is defined as
1+ Γ
S=
1− Γ
(5.2.40)
where ⎮Γ ⎪ is the magnitude of the reflection coefficient at that frequency. For example, if 2.5
percent of the incident voltage is reflected at the visual carrier frequency, the VSWR at that fre-
quency is
1+ 0.025
S= = 1.05
1− 0.025
(5.2.41)
A high value of VSWR is undesirable because it contributes to:

• Visible ghosts if the source of the VSWR is more than 250 ft (76 m) away from the transmit-
ter
• Short-term echoes (<0.1-μs delay)
• Aural distortion
• Reduction of the transmission line efficiency
Of all the undesirable effects of the system VSWR, the first, a visible ghost resulting from
input VSWR, is the most critical for conventional (analog) television applications. The further
the antenna is from the transmitter, the higher the subjective impairment of the picture because of
the reflection at the input of the antenna, will be.
Antenna input specification in terms of VSWR is not an effective figure of merit to obtain the
best picture quality. For example, Figure 5.2.13 shows a comparative performance of two anten-
nas. Antenna A has a maximum VSWR of 1.08 across the channel and a VSWR of 1.06 at pic-
ture (pix) carrier. Antenna B has a maximum VSWR of 1.13 and a VSWR of 1.01 at pix. It is
hard to tell anything about the relative picture impairment of these two antennas by analyzing the
VSWR alone. However, the reflection of a 2-T sine-squared pulse by each antenna is also shown
in the figure. It can be seen that antenna A, with maximum VSWR of 1.08, produces a reflection
of more than 3 percent. This reflection results in a ghost that could be perceptible if the transmis-
sion line to the antenna is at least 600 ft long. Antenna B with a maximum VSWR of 1.13 pro-
duces only 1 percent reflection for the same pulse.
The pulse response of an antenna mounted on top of a tower can be measured if the transmis-
sion line is sufficiently long and “clean” to resolve the incident from the reflected pulse. If the
line is not sufficiently long or if knowledge of the antenna’s pulse response is required prior to
installation, a calculation is possible. Pulse response cannot be calculated from the VSWR data
p
(a )
(b )
(c )
Figure 5.2.13 Example VSWR considerations for two antennas: (a) impedance, (b) reflected
pulse, (c) measured VSWR.
alone because it contains no information with respect to the phase of the reflection at each fre-
quency across the channel. Because the impedance representation contains the amplitude and
phase of the reflection coefficient, the pulse response can be calculated if the impedance is
known. The impedance representation is typically done on a Smith chart, which is shown in Fig-
p
ure 5.2.13 for antennas A and B. Some attempts at relating various shapes of VSWR curves to
subjective picture quality have been made. A good rule of thumb is to minimize the VSWR in the
region from –0.25 to + 0.75 MHz of the visual carrier frequencies to a level below 1.05 and not
to exceed the level of 1.08 at color subcarrier.
5.2.4a VSWR Optimization Techniques

As noted in the previous section, the shape of the antenna VSWR across the channel spectrum
must be optimized to minimize the subjective picture impairment. Frequently, it may be desirable
to perform the same VSWR optimization on the transmission line, so that the entire system
appears transparent to the incoming wave. The optimization of the transmission line VSWR is a
relatively time-consuming and expensive task because it requires laying out the entire length of
line on the ground and slugging it at various points. Slugging describes a technique of placing a
metallic sleeve of a certain diameter and length that is soldered over the inner conductor of a
coaxial transmission line. In some instances, it is more convenient to use a section of line with
movable capacitive probes instead of a slug. This section is usually called variously a variable
transformer, fine tuner, or impedance matcher.
The single-slug technique is the simplest approach to VSWR minimization. At any frequency,
if the VSWR in the line is known, the relationship between the slug length in wavelengths and its
characteristic impedance is given by
⎡ ⎤
L 1 −1⎢ S −1 ⎥
= tan ⎢ ⎥
λ 2π ⎢
⎣ ( )(
S − R2 ⎡
⎣ 1 R )
2
− S ⎤⎥
⎦⎦
(5.2.42)
Where:
S = existing VSWR
R = slug characteristic impedance ÷ line characteristic impedance
L/λ = length of slug
A graphic representation of this expression is given in Figure 5.2.14. The effect of the fringe
capacitance, resulting from the ends of the slug, is not included in the design chart because it is
negligible for all television channels. After the characteristic impedance of the slug is known, its
diameter is determined from
D
Z c = 138log10
d
(5.2.43)
Where:
D = the outside diameter of the slug conductor
d = the outside diameter of the inner conductor
p
Figure 5.2.14 Single-slug matching of a coaxial line.
The slug thus constructed is slowly slid over the inner conductor until the VSWR disappears.
This occurs within a line length of 1/2 wavelength.
There are two shortcomings in the single-slug technique:
• Access to the inner conductor is required
• The technique is not applicable if the VSWR at two frequencies must be minimized
The first shortcoming can be eliminated by using the variable transformer mentioned previ-
ously. While it is more expensive, slug machining and installation sliding adjustments are
avoided. The second shortcoming can be overcome with the double-slug technique, or more con-
veniently, with two variable transformers.
The single- and double-slug techniques can have an undesirable effect if not properly applied.
The slugs should be placed as near as possible to the source of the undesirable VSWR. Failure to
do so can lead to higher VSWRs at other frequencies across the bandwidth of interest. Thus, if
both the system and the antenna VSWRs are high at the visual carrier, slugging at the transmitter
room will lower the system VSWR but will not eliminate the undesirable echo. Another effect of
slugging is the potential alteration of the power and voltage rating. This is of particular impor-
tance if the undesirable VSWR is high. Generally, the slugging should be limited to situations
where the VSWR does not exceed 1.25.
p
5.2.5 Azimuthal-Pattern Distortion in Multiple-Antenna Installations

In a stacked arrangement, the cross sections of the support structure of the antenna on the lower
levels are large in order to support the antenna above. Hence, the circularity of the azimuthal pat-
tern of the lower antennas will not be as uniform as for the upper antennas where the support
structure is smaller.
In the case of a candelabra arrangement, the centers of radiation of most antennas are approx-
imately equal. However, the radiated signal from each antenna is scattered by the other opposing
antennas and, owing to the reflected signal, there results a deterioration of azimuthal-pattern cir-
cularity and video-signal performance criteria. When the proximity of one antenna to others is
equal to its height or less, the shape of its elevation pattern at the interfering antennas is essen-
tially the same as its aperture illumination. Consequently the reflections of significance are from
the sections of the interfering structures parallel to the aperture of the radiating antenna. In this
case, a two-dimensional analysis of the scatter pattern can be utilized to estimate the reflected
signal and its effect on the free-space azimuthal pattern.
Owing to the physical separation between the transmitting antenna and the reflecting struc-
ture, the primary and reflected signals add in-phase in some azimuthal directions and out-of-
phase in others. Thus, the overall azimuthal pattern is serrated. The minimum-to-maximum level
of serrations in any azimuthal direction is a function of the cross section of the opposing struc-
ture. It also is directly proportional to the incident signal on the reflecting structure and inversely
proportional to the square root of the spacing, expressed in wavelengths.
Furthermore, the reflections resulting from the vertical polarization component are higher
than those resulting from the horizontal polarization component. Consequently, candelabras for
circularly polarized antennas require larger spacing or fewer antennas. Figures 5.2.15 and 5.2.16
show the azimuthal pattern distortion of a channel-4 circularly polarized antenna 35 ft (11 m)
away from a channel-10 circularly polarized antenna, for horizontal and vertical polarizations,
respectively.
The calculated in-place pattern is based on ideal assumptions, and the exact locations of the
nulls and peaks of the in-place pattern cannot always be determined accurately prior to installa-
tion. However, the outer and inner envelopes of the pattern provide a reasonable means for esti-
mating the amount of deterioration that can result in the pattern circularity. For larger cross
sections of the opposing cylinder, the scatter pattern can he approximated by a uniformly con-
stant value around a major portion of the azimuth and a larger value over the shadow region. The
former is very close to the rms value of the scatter serration. The maximum value occurs toward
the shadow region except for the very small diameter of the opposing structure. The variation of
the rms and peak values of g(φ), the reflection coefficient, is shown in Figure 5.2.17. The rms
value of the reflection coefficient can be utilized to estimate the in-place circularity from an
obstructing cylinder, over most of the periphery, and the peak value can be used to judge the sig-
nal toward the shadow region.
5.2.5a Echoes in Multiple-Tower Systems

In multiple-tower systems, at least two towers are utilized to mount the antennas for the same
coverage area. The discussion relating to multiple-antenna systems in the previous section is
applicable here when the towers are located within 100 ft of each other. The in-place azimuthal
patterns can be determined and the echoes are not perceptible because the delay is less than 0.2
μs if the reradiation due to coupling into the feed system of the interfering antennas is negligible.
p
Figure 5.2.15 In-place azimuthal pattern of one antenna in the presence of another 35 ft away
from it (horizontal polarization). The self pattern is the dotted curve.
When towers are located at a spacing greater than 100 ft (30 m), both the problem of azi-
muthal pattern deterioration and the magnitude of the ghost have to be considered. The magni-
tude of the reflection from an opposing structure decreases as the spacing increases, but not
linearly. For example, if the spacing is quadrupled, the magnitude of the reflection is reduced by
only one-half. When the antennas are located more than several hundred feet from each other, the
magnitude of the reflection is usually small enough to be ignored, as far as pattern deterioration
is concerned. However, a reflection of even 3 percent is noticeable to a critical viewer. Thus,
large separations, usually more than a thousand feet, are necessary to reduce the echo visibility
to an acceptably low level.
In the previous analysis for the illumination of the interfering structure by the antenna, it was
assumed that only the portion of the interfering structure in the aperture of the antenna was of
importance. This is true for separation distances comparable with the antenna aperture. However,
as the separation distance from the antenna increases, the elevation pattern in any vertical plane
changes its shape from a near-field to a far-field pattern. As the elevation pattern changes, more
of the opposing structure is illuminated by the primary signal. This effect of distance from the
antenna on the elevation pattern is illustrated in Figure 5.2.18. Note that the elevation pattern
shown is plotted against the height of the opposing structure, rather than the elevation angle.
p
Figure 5.2.16 In-place azimuthal pattern of one antenna in the presence of another 35 ft away
from it (vertical polarization). The self pattern is the dotted curve.
Figure 5.2.17 Reflection coefficient from a cylinder.

p
Figure 5.2.18 Illustration of opposing tower illumination for various spacings, given in wavelengths
between the transmitting antenna and the opposing structure.
In practice, optimization of a multiple-tower system may require both a model study and com-
puter simulation. The characteristics of the reflections of the particular tower in question can best
be determined by model measurements and the effect of spacing determined by integration of the
induced illumination on the interfering structure.
5.2.6 References
1. Fumes, N., and K. N. Stokke: “Reflection Problems in Mountainous Areas: Tests with Cir-
cular Polarization for Television and VHF/FM Broadcasting in Norway,” EBU Review,
Technical Part, no. 184, pp. 266–271, December 1980.
2. Hill, P. C. J.: “Methods for Shaping Vertical Pattern of VHF and UHF Transmitting Aeri-
als,” Proc. IEEE, vol. 116, no. 8, pp. 1325–1337, August 1969.
3. DeVito, G: “Considerations on Antennas with no Null Radiation Pattern and Pre-estab-
lished Maximum-Minimum Shifts in the Vertical Plane,” Alta Frequenza, vol. XXXVIII,
no. 6, 1969.
4. Perini, J., and M. H. Ideslis: “Radiation Pattern Synthesis for Broadcast Antennas,” IEEE
Trans. Broadcasting, vol. BC-18, no. 3, pg. 53, September 1972.
5. Praba, K.: “Computer-aided Design of Vertical Patterns for TV Antenna Arrays,” RCA
Engineer, vol. 18-4, January–February 1973.
p
5.2.7 Bibliography
Allnatt, J. W., and R. D. Prosser: “Subjective Quality of Television Pictures Impaired by Long
Delayed Echoes,” Proc. IEEE, vol. 112, no. 3, March 1965.
Bendov, O.: “Measurement of Circularly Polarized Broadcast Antennas,” IEEE Trans. Broad-
casting, vol. BC-19, no. 1, pp. 28–32, March 1972.
Dudzinsky, S. J., Jr.: “Polarization Discrimination for Satellite Communications,” Proc. IEEE,
vol. 57, no. 12, pp. 2179–2180, December 1969.
Fowler, A. D., and H. N. Christopher: “Effective Sum of Multiple Echoes in Television,” J.
SMPTE, SMPTE, White Plains, N. Y., vol. 58, June 1952.
Johnson, R. C, and H. Jasik: Antenna Engineering Handbook, 2d ed., McGraw-Hill, New York,
N.Y., 1984.
Kraus, J. D.: Antennas, McGraw-Hill, New York, N.Y., 1950.
Lessman, A. M.: “The Subjective Effect of Echoes in 525-Line Monochrome and NTSC Color
Television and the Resulting Echo Time Weighting,” J. SMPTE, SMPTE, White Plains,
N.Y., vol. 1, December 1972.
Mertz, P.: “Influence of Echoes on Television Transmission,” J. SMPTE, SMPTE, White Plains,
N.Y., vol. 60, May 1953.
Moreno, T.: Microwave Transmission Design Data, Dover, New York, N.Y.
Perini, J.: “Echo Performance of TV Transmitting Systems,” IEEE Trans. Broadcasting, vol. BC-
16, no. 3, September 1970.
Praba, K.: “R. F. Pulse Measurement Techniques and Picture Quality,” IEEE Trans. Broadcast-
ing, vol. BC-23, no. 1, pp. 12–17, March 1976.
Sargent, D. W.: “A New Technique for Measuring FM and TV Antenna Systems,” IEEE Trans.
Broadcasting, vol. BC-27, no. 4, December 1981.
Siukola, M. S.: “TV Antenna Performance Evaluation with RF Pulse Techniques,” IEEE Trans.
Broadcasting, vol. BC-16, no. 3, September 1970.
Whythe, D. J.: “Specification of the Impedance of Transmitting Aerials for Monochrome and
Color Television Signals,” Tech. Rep. E-115, BBC, London, 1968.
g g
Chapter
5.3
Television Transmitting Antennas
Oded Ben-Dov, Krishna Praba
5.3.1 Introduction
Broadcasting is accomplished by the emission of coherent electromagnetic waves in free space
from a single or group of radiating-antenna elements, which are excited by modulated electric
currents. Although, by definition, the radiated energy is composed of mutually dependent mag-
netic and electric vector fields, conventional practice in television engineering is to measure and
specify radiation characteristics in terms of the electric field [1–3].
The field vectors may be polarized, or oriented, linearly, horizontally, vertically, or circularly.
Linear polarization is used for some types of radio transmission. Television broadcasting has
used horizontal polarization for the majority of the system standards in the world since its incep-
tion. More recently, the advantages of circular polarization have resulted in an increase in the use
of this form of transmission, particularly for VHF channels.
Both horizontal and circular polarization designs are suitable for tower-top or tower-face
installations. The latter option is dictated generally by the existence of previously installed tower-
top antennas. On the other hand, in metropolitan areas where several antennas must be located on
the same structure, either a stacking or a candelabra-type arrangement is feasible. Figure 5.3.1
shows an example of antenna stacking.
The use of stacked antennas provides a number of operational benefits, not the least of which
is reduced cost. There are numerous examples of such installations, including the Sears Tower in
Chicago (Figure 5.3.2) and Mt. Sutro in San Francisco, where antennas are mounted on a cande-
labra assembly (Figure 5.3.3). The implementation of DTV operation has generated considerable
interest in such designs primarily because of their ability to accommodate a great number of
antennas on a given structure.
Another solution to the multichannel location problem, where space or structural limitations
prevail, is to diplex two stations on the same antenna. This approach, while economical from the
installation aspect, can result in transmission degradation because of the diplexing circuitry, and
antenna-pattern and impedance broadbanding usually is required.
5-65
g
Figure 5.3.1 Stacked antenna array.
5.3.2 Common Antenna Designs

The typical television broadcast antenna is a broadband radiator operating over a bandwidth of
several megahertz with an efficiency of over 95 percent. Reflections in the system between the
transmitter and antenna must be small enough to introduce negligible picture degradation. Fur-
g
Figure 5.3.2 Twin tower antenna array atop the Sears Tower in Chicago.
thermore, the gain and pattern characteristics must be designed to achieve the desired coverage
within acceptable tolerances, and operationally with a minimum of maintenance. Tower-top,
pole-type antennas designed to meet these requirements can be classified into two broad catego-
ries:
g
Figure 5.3.3 The triangular candelabra of antennas at Mt. Sutro in San Francisco.
• Resonant dipoles and slots

• Multiwavelength traveling-wave elements
• Turnstile configuration
The main consideration in top-mounted antennas is the achievement of excellent omnidirec-
tional azimuthal fields with a minimum of windload. The earliest and most popular resonant
antenna for VHF applications is the turnstile, which is made up of four batwing-shaped elements
mounted on a vertical pole in a manner resembling a turnstile. The four “batwings” are, in effect,
two dipoles that are fed in quadrature phase. The azimuthal-field pattern is a function of the
diameter of the support mast and is within a range of 10 to 15 percent from the average value.
The antenna is made up of several layers, usually six for channels 2 to 6 and twice that number
for channels 7 to 13. This antenna is suitable for horizontal polarization only. It is unsuitable for
g
side-mounting, except for standby applications

where coverage degradation can be tolerated.
5.3.2a Multislot Radiator

A multislot antenna consists of an array of axial
slots on the outer surface of a coaxial transmis-
sion line. The slots are excited by an exponen-
tially decaying traveling wave inside the slotted
pole. The omnidirectional azimuthal pattern
deviation is less than 5 percent from the aver-
age circle. The antenna is generally about fif-
teen wavelengths long. A schematic of such an
antenna is shown in Figure 5.3.4; the principle
of slot excitation is illustrated in Figure 5.3.5.
5.3.2b Circular Polarization

For circular polarization, both resonant and
traveling-wave antennas are available. The trav-
eling-wave antenna is essentially a side-fire
helical antenna supported on a pole. A suitable
number of such helices around the pole provide
excellent azimuthal pattern circularity. This
type of antenna is especially suited for applica-
tion with channels 7 to 13 because only 3 per-
cent pattern and impedance bandwidth are
required. For channels 2 to 6 circular polariza-
tion applications where the bandwidth is
approximately 10 percent, resonant dipoles
around a support pole are a preferred configu-
ration.
5.3.2c UHF Antennas for Tower-Top Installation Figure 5.3.4 Schematic of a multislot travel-
The slotted-cylinder antenna, commonly ing-wave antenna.
referred to as the pylon antenna, is a popular
top-mounted antenna for UHF applications.
Horizontally polarized radiation is achieved using axial resonant slots on a cylinder to generate
circumferential current around the outer surface of the cylinder. An excellent omnidirectional
azimuthal pattern is achieved by exciting four columns of slots around the circumference of the
cylinder, which is a structurally rigid coaxial transmission line.
The slots along the pole are spaced approximately one wavelength per layer and a suitable
number of layers are used to achieve the desired gain. Typical gains range from 20 to 40.
g
Figure 5.3.5 Principle of slot excitation to produce an omnidirectional pattern.
By varying the number of slots around the periphery of the cylinder, directional azimuthal
patterns are achieved. It has been found that the use of fins along the edges of the slot provide
some control over the horizontal pattern.
The ability to shape the azimuthal field of the slotted cylinder is somewhat restricted. Instead,
arrays of panels have been utilized, such as the zigzag antenna [4]. In this antenna, the vertical
component of the current along the zigzag wire is mostly cancelled out, and the antenna can be
considered as an array of dipoles. Several such panels can be mounted around a polygonal
periphery, and the azimuthal pattern can be shaped by the proper selection of the feed currents to
the various panels.
5.3.2d VHF Antennas for Tower-Face Installation

There are a number of possible solutions to the tower-face mounting requirement. A common
panel antenna used for tower face applications is the so-called butterfly [5] or batwing panel
antenna developed from the turnstile radiator. This type of radiator is suitable for the entire range
of VHF applications. Enhancements to the basic structure include modification of the shape of
the turnstile-type wings to rhombus or diamond shape. Another version is the multiple dipole
panel antenna used in many installations outside the U.S. For circularly polarized applications,
g
Table 5.3.1 Circularities of Panel Antennas for VHF Operation
Shape Tower-face Size, ft (m) Circularity, ± dB1

Channels 2–6 Channels 7–13
Triangular 5 (1.5) 0.9 1.8
6 (1.8) 1.0 2.0
7 (2.1) 1.1 2.3
10 (3.0) 1.3 3.0
Square 4 (1.2) 0.5 1.6
5 (1.5) 0.6 1.9
6 (1.8) 0.7 2.4
7 (2.1) 0.8 2.7
10 (3.0) 1.2 3.2
1
Add up to ±0.3 dB for horizontally polarized panels and ± 0.6 dB for circularly polarized panels.
These values are required to account for tolerances and realizable phase patterns of practical hard-
ware assemblies.
two crossed dipoles or a pair of horizontal and vertical dipoles are used [6]. A variety of cavity-
backed crossed-dipole radiators are also utilized for circular polarization transmission.
The azimuthal pattern of each panel antenna is unidirectional, and three or four such panels
are mounted on the sides of a triangular or square tower to achieve omnidirectional azimuthal
patterns. The circularity of the azimuthal pattern is a function of the support tower size [7].
Calculated circularities of these antennas are shown in Table 5.3.1 for idealized panels. The
panels can be fed in-phase, with each one centered on the face of the tower, or fed in rotating
phase with proper mechanical offset on the tower face. In the latter case, the input impedance
match is far better.
Directionalization of the azimuthal pattern is realized by proper distribution of the feed cur-
rents to the individual panels in the same layer. Stacking of the layers provides gains comparable
with those of top-mounted antennas.
The main drawbacks of panel antennas are high wind-load, complex feed system inside the
antenna, and the restriction on the size of the tower face in order to achieve smooth omnidirec-
tional patterns. However, such devices provide an acceptable solution for vertical stacking of sev-
eral antennas or where installation considerations are paramount.
5.3.2e UHF Antennas for Tower-Face Installation

Utilization of panel antennas in a manner similar to those for VHF applications is not always
possible for the UHF channels. The high gains, which are in the range of 20 to 40 compared with
those of 6 to 12 for VHF, require far more panels with the associated branch feed system. It is
also difficult to mount a large number of panels on all the sides of a tower, the cross section of
which must be restricted to achieve a good omnidirectional azimuthal pattern.
The zigzag panel described previously has been found to be applicable for special omnidirec-
tional and directional situations. For special directional azimuthal patterns, such as a cardioid
shape, the pylon antenna can be side-mounted on one of the tower legs.
The use of tangential-firing panels around the periphery of a large tower has resulted in prac-
tical antenna systems for UHF applications. Zigzag panels or dipole panels are stacked vertically
g
Figure 5.3.6 Various antenna designs for VHF and UHF broadcasting. Note that not all these
designs have found commercial use.
at each of the corners of the tower and oriented such that the main beam is along the normal to
the radius through the center of the tower. The resultant azimuthal pattern is usually acceptable
for horizontal polarization transmission.
Figure 5.3.6 summarizes the most common TV antenna technologies for VHF and UHF
applications.
5.3.2f Vertically Stacked Multiple Antennas

In metropolitan areas where there are several stations competing for the same audience, usually
there may be only one preferred location for all the transmitting antennas [8, 9]. A straightfor-
ward approach to the problem is stacking the antennas one on top of the other. One of the earliest
installations of stacked antennas was on the Empire State Building in New York City. Because
the heights of the centers of radiation decrease from the antenna at the top to that at the lowest
level, there is a preference to be at a higher level. Thus, the final arrangement depends on both
technical and commercial constraints. For relatively uniform coverage, technical considerations
usually dictate that the top of the mast be reserved for the higher channels and the bottom of the
mast for the low channels. However, because of contractual stipulations, this is not always possi-
ble.
Vertical stacking provides the least amount of interference among the antennas. However,
because the antennas in the lower levels are panels on the support tower faces, which tend to be
larger because the level is lower, the azimuthal pattern characteristics can be less than optimum.
Furthermore, the overall height constraint can compromise the desirable gain for each channel.
g
5.3.2g Candelabra Systems

In order to provide the same, as well as the highest, center of radiation to more than one station,
antennas can be arranged on platform in a candelabra style. There are many tee bar arrange-
ments in which there are two stacks of antennas. A triangular or even a square candelabra is uti-
lized in some cases. When several antennas are located within the same aperture, the radiated
signal from one antenna is partly reflected and partly reradiated by the opposing antenna or
antennas. Because the interference signal is received with some time delay, with respect to the
primary signal, it can introduce picture distortion and radiation-pattern deterioration. The choice
of opposing antennas and the interantenna spacing and orientation of the antennas determine the
trade-offs among the in-place performance characteristics, which include:
• Azimuthal pattern
• Video bandwidth response
• Echoes
• Differential windsway
• Isolation among channels
5.3.2h Multiple-Tower Installations

The candelabra multiple-antenna system requires cooperation of all the broadcasters in an area
and considerable planning prior to construction [10]. In many cases, new channels are licensed
after the first installation and addition of the antenna on an existing tower is not possible. Conse-
quently, location on a nearby new tower is the only solution. In some cases, the sheer size of the
candelabra may make it more expensive than multiple towers. The antenna farms around Phila-
delphia and Miami are good examples of several towers located in the same area.
In the case of multiple towers, the creation of long-delayed echoes is of major concern, if eco-
nomically or practically towers cannot be located close enough to result in echoes that either (a)
cannot be resolved by the television system, or (b) will not affect color subcarrier phase and
amplitude (for analog systems). This would dictate an impractical spacing of under 100 ft (30
m). Circularly polarized antennas provide an advantage because the reflected signal sense of
polarization rotation is usually reversed, and the effect of echoes can he reduced with a proper
circularly polarized receiving antenna [11, 12]. The twin stack of antennas atop the Sears Tower
in Chicago is illustrative of a multiple-tower system.
5.3.2i Multiplexing of Channels

Another technique of accommodating more than one channel in the same antenna location is by
combining the signals from these stations and feeding them to the same antenna for radiation.
Broadband antennas are designed for such applications. The antenna characteristics must be
broadband in more terms than input impedance. Pattern bandwidth and, in the case of circularly
polarized antennas, axial-ratio bandwidth are equally important. Generally, it is more difficult to
broadband antennas if the required bandwidth will be in excess of 20 percent of the design-cen-
ter frequency. Another of the problems of multiplexing is that the antenna must be designed for
the peak voltage breakdown, which is proportional to the square of the number of channels. A
g
third problem is the resolution of all technical, com- (a) (b )

mercial, and legal responsibilities that arise from
joint usage of the same antenna.
Multiplexed antennas typically are designed with
integral redundancy. The antenna is split into two
halves with two input feedlines. This provides pro-
tection from failure as a result of the potential
reduction in the high-power breakdown safety mar-
gin.
5.3.2j Broadband Television Antennas

As touched on in the previous section, the radiation
of multiple channels from a single antenna requires
the antenna to be broadband in both pattern and
impedance (VSWR) characteristics. As a result, a
broadband TV antenna represents a significant
departure from the narrowband, single channel pole
antennas commonly used for VHF and UHF. The
Figure 5.3.7 Antenna feed configura-
typical single channel UHF antenna uses a series
tions: (a) series feed, (b) branch feed.
feed to the individual radiating elements, while a
broadband antenna has a branch feed arrangement.
The two feed configurations are shown in Figure
5.3.7.
At the designed operating center frequency, the series feed provides co-phased currents to its
radiating elements. As the frequency varies, however, the electrical length of the series line feed
changes such that the radiating elements are no longer in-phase outside of the designed channel.
This electrical length change causes significant beam tilt out of band, and an input VSWR that
varies widely with frequency.
In contrast, the branch feed configuration employs feed lines that are nominally of equal
length. Therefore the phase relationships of the radiating elements are maintained over a wide
span of frequencies. This provides vertical patterns with stable beam tilt, a requirement for multi-
channel applications.
The basic building block of the multi-channel antenna is the broadband panel radiator. The
individual radiating elements within a panel are fed by a branch feeder system that provides the
panel with a single input cable connection. These panels are then stacked vertically and arranged
around a supporting spine or existing tower to produce the desired vertical and horizontal radia-
tion patterns.
Bandwidth
The ability to combine multiple channels in a single transmission system depends upon the band-
width capabilities of the antenna and waveguide or coax. The antenna must have the necessary
bandwidth in both pattern and impedance (VSWR). It is possible to design an antenna system for
low power applications using coaxial transmission line that provides whole-band capability. For
high power systems, waveguide bandwidth sets the limits of channel separation.
g
(a ) (b) (c )
Figure 5.3.8 Measured antenna patterns for three types of panel configurations at various operat-
ing frequencies: (a) 5 panels per bay, (b) 6 panels per bay, (c) 8 panels per bay.
Antenna pattern performance is not a significant limiting factor. As frequency increases, the
horizontal pattern circularity deteriorates, but this effect is generally acceptable, given the pri-
mary project objectives. Also, the electrical aperture increases with frequency, which narrows
the vertical pattern beamwidth. If a high gain antenna were used over a wide bandwidth, the
increase in electrical aperture might make the vertical pattern beamwidth unacceptably narrow.
This is, however, usually not a problem because of the channel limits set by the waveguide.
Horizontal Pattern
Because of the physical design of a broadband panel antenna, the cross-section is larger than the
typical narrowband pole antenna. Therefore, as the operating frequencies approach the high end
of the UHF band, the circularity (average circle to minimum or maximum ratio) of an omnidirec-
tional broadband antenna generally deteriorates.
Improved circularity is possible by arranging additional panels around the supporting struc-
ture. Previous installations have used 5, 6, and 8 panels per bay. These are illustrated in Figure
5.3.8 along with measured patterns at different operating channels. These approaches are often
required for power handling considerations, especially when three or four transmitting channels
are involved.
The flexibility of the panel antenna allows directional patterns of unlimited variety. Two of
the more common applications are shown in Figure 5.3.9. The peanut and cardioid types are
often constructed on square support spines (as indicated). A cardioid pattern can also be pro-
duced by side-mounting on a triangular tower. Different horizontal radiation patterns for each
channel can also be provided, as indicated in Figure 5.3.10. This is accomplished by changing the
power and/or phase to some of the panels in the antenna with frequency.
Most of these antenna configurations are also possible using a circularly-polarized panel. If
desired, the panel can be adjusted for elliptical polarization, with the vertical elements receiving
g
(a ) (b)
Figure 5.3.9 Common directional antenna patterns: (a) peanut, (b) cardioid.
less than 50 percent of the power. Using a circu-

larly-polarized panel will reduce the horizontally-
polarized ERP by half (assuming the same trans-
mitter power).
5.3.3 DTV Implementation Issues

The design and installation of a transmitting
antenna for DTV operation is a complicated pro-
cess that must take into consideration a number of
variables. Fortunately, new technologies and
refinements to classic antenna designs have pro-
vided additional flexibility in this process. Figure 5.3.10 Use of a single antenna to
produce two different radiation patterns,
omnidirectional (trace A) and peanut (trace
5.3.3a Channel-Combining Considerations B).
A number of techniques are practical to utilize an
existing tower for both NTSC and DTV transmis-
sions. Combining RF signals allows broadcasters to use an existing structure to transmit NTSC
and DTV from a common line and antenna or, in the case of a VHF and UHF combination, to
utilize the same line to feed two separate antennas.
In the transition period from NTSC to DTV, many broadcasters have chosen to use their exist-
ing tower to transmit both NTSC and DTV channels. Some have chosen to add a new line and
DTV antenna; others to combine their DTV with NTSC and transmit from a common antenna
and line; and still others to consolidate to a new structure common to many local channels. For
most stations, it is a matter of cost and feasibility.
Channel combiners, also known as multiplexers or diplexers, are designed for various appli-
cations. These systems can be generally classified as follows [13]:
g
• Constant impedance—designs that consist of two identical filters placed between two
hybrids.
• Starpoint—designs that consist of single bandpass filters phased into a common output tee.
• Resonant loop—types that utilize two coaxial lines placed between two hybrids; the coaxial
lines are of a calculated length.
• Common line—types that use a combination of band-stop filters matched into a common out-
put tee.
The dual-mode channel combiner is a device that permits a single transmission line on a
tower to feed two separate antennas. Dual-mode channel combining is the process by which two
channels are combined within the same waveguide, but in separate orthogonal modes of propa-
gation [14].
The device combines two different television channels from separate coaxial feedlines into a
common circular waveguide. Within the circular waveguide, one channel propagates in the TE11
mode while the other channel propagates in the TM01 mode. The dual-mode channel combiner is
reciprocal and, therefore, also may be used to efficiently separate two TE 11/TM 01 mode-isolated
channels that are propagating within the same circular waveguide into two separate coaxial lines.
This provides a convenient method for combining at the transmitters and splitting at the anten-
nas. The operating principles of the dual-mode channel combiner are described in [14].
5.3.3b Antenna Systems

The availability of suitable locations for new television transmission towers is diminishing, even
in the secondary markets, and sites are practically nonexistent in major markets [15]. After the
hurdles of zoning variance and suitable tower location are overcome, FAA restrictions and envi-
ronmental concerns can delay the construction of a new tower for years. Not surprisingly, many
broadcasters have looked at the pros and cons of using existing towers to support their new DTV
antennas even though the prime tower-top spots are occupied.
For any given antenna, directional or omnidirectional, the tower will modify the as-designed
antenna pattern. For optimum coverage, the as-installed pattern must be known—not just at the
carrier frequency, but throughout the entire channel—before the relative position of the antenna
and its azimuthal pattern orientation can be fixed. There is usually one position that will provide
the optimum coverage without exceeding the structural limitations of the tower. This optimum
position can be calculated (see [15]).
Coverage considerations are particularly important to DTV because undesired energies, such
as reflections, translate into a loss of coverage, whereas the undesired energies in NTSC translate
primarily into a loss of picture quality.
Another transmission-optimization technique that holds promise for DTV is circular polariza-
tion. The transmission of CP has obvious drawbacks in the form of a 2× increase in required
transmitter power and transmission line, as well as a more complex antenna. Still, for the DTV
signal, polarization diversity can be achieved if the vertically polarized signal is transmitted
through CP. A polarization-diversity system at the receive antenna can provide missing signal
level when one of the horizontal or vertical signal components experiences a deep fade. It fol-
lows that the inherent diversity attributes of CP operation could be put to good use in reducing
the cliff-edge effect of the terrestrial DTV signal [16].
g
5.3.4 Key Considerations in System Design

Antenna system design is always an iterative process of configuration analysis, subject to some
well-defined and many ill-defined constraints. Each iteration basically is a study with a sequence
of basic technical and commercial decisions. The number of antennas at the same location
increases the complexity of such a study, The main reason for this is that most constraints must
be viewed from three angles: engineering, legal, and commercial. Following is a checklist of
items that are key considerations in system design:
• Coverage and picture quality
• Transmitter power
• Antenna mechanical aperture
• Channel multiplexing desirability or necessity
• System maintenance requirements for the broadcaster and service organization
• Performance deterioration resulting from ice, winds, and earthquakes
• Initial investment
• Project implementation schedule
• Building code requirements
• Aesthetic desires of the broadcaster and the community
• Radiation hazards protection
• Environmental protection
• Beacon height
• Accessibility of beacon for maintenance
It is worth reemphasizing that these considerations are interrelated and cover more than iso-
lated technical issues.
5.3.5 References
1. Kraus, J. D.: Antennas, McGraw-Hill, New York, N.Y., 1950.
2. Moreno, T.: Microwave Transmission Design Data, Dover, New York, N.Y.
3. Johnson, R. C, and H. Jasik: Antenna Engineering Handbook, 2d ed., McGraw-Hill, New
York, N.Y., 1984.
4. Clark, R. N., and N. A. L. Davidson: “The V-Z Panel as a Side Mounted Antenna,” IEEE
Trans. Broadcasting, vol. BC-13, no. 1, pp. 3–136, January 1967.
5. Brawn, D. A., and B. F. Kellom: “Butterfly VHF Panel Antenna,” RCA Broadcast News,
vol. 138, pp. 8–12, March 1968.
6. DeVito, G. G., and L. Mania: “Improved Dipole Panel for Circular Polarization,” IEEE
Trans. Broadcasting, vol. BC-28, no. 2, pp. 65–72, June 1982.
g
7. Perini, J.: “Improvement of Pattern Circularity of Panel Antenna Mounted on Large Tow-
ers,” IEEE Trans. Broadcasting, vol. BC-14, no. 1, pp. 33–40, March 1968.
8. “Predicting Characteristics of Multiple Antenna Arrays,” RCA Broadcast News, vol. 97, pp.
63–68, October 1957.
9. Hill, P. C. J.: “Measurements of Reradiation from Lattice Masts at VHF,” Proc. IEEE, vol.
III, no. 12, pp. 1957–1968, December 1964.
10. “WBAL, WJZ and WMAR Build World’s First Three-Antenna Candelabra,” RCA Broad-
cast News, vol. 106, pp. 30–35, December 1959.
11. Knight, P.: “Reradiation from Masts and Similar Objects at Radio Frequencies,” Proc.
IEEE, vol. 114, pp. 30–42, January 1967.
12. Siukola, M. S.: “Size and Performance Trade Off Characteristics of Horizontally and Cir-
cularly Polarized TV Antennas,” IEEE Trans. Broadcasting, vol. BC-23, no. 1, March
1976.
13. Heymans, Dennis: “Channel Combining in an NTSC/ATV Environment,” Proceedings of
14. Smith, Paul D.: “New Channel Combining Devices for DTV, Proceedings of the 1997 NAB
D.C., pg. 218, 1996.
15. Bendov, Oded: “Coverage Contour Optimization of HDTV and NTSC Antennas,” Proceed-
ings of the 1996 NAB Broadcast Engineering Conference, National Association of Broad-
casters, Washington, D.C., pg. 69, 1996.
16. Plonka, Robert J.: “Can ATV Coverage Be Improved With Circular, Elliptical, or Vertical
Polarized Antennas?” Proceedings of the 1996 NAB Broadcast Engineering Conference,
National Association of Broadcasters, Washington, D.C., pg. 155, 1996.
5.3.6 Bibliography
Fisk, R. E., and J. A. Donovan: “A New CP Antenna for Television Broadcast Service,” IEEE
Trans. Broadcasting, vol. BC-22, no. 3, pp. 91–96, September 1976.
Johns, M. R., and M. A. Ralston: “The First Candelabra for Circularly Polarized Broadcast
Antennas,” IEEE Trans. Broadcasting, vol. BC-27, no. 4, pp. 77–82, December 1981.
Siukola, M. S.: “The Traveling Wave VHF Television Transmitting Antenna,” IRE Trans. Broad-
casting, vol. BTR-3, no. 2, pp. 49-58, October 1957.
Wescott, H. H.: “A Closer Look at the Sutro Tower Antenna Systems,” RCA Broadcast News,
vol. 152, pp. 35–41, February 1944.
g
g g
Chapter
5.4
Tower Construction and Maintenance
5.4.1 Introduction
The transmitting tower is the most visible component of any broadcast transmission facility. It is
also the component most vulnerable to hostile elements. The basic task of the station engineer is
to keep the tower erect in the face of natural and man-made forces until the structure is no longer
needed. The most common reasons for tower failure are poor construction, poor maintenance,
overloading, icing, and accidents.
5.4.2 Antenna Tower Lighting and Marking Requirements

In the U.S., the FCC has been given the authority by Congress to require the painting and/or illu-
mination of antenna towers when it determines that such towers may constitute a menace to air
navigation [47 U.S.C. § 303(q)]. The FCC's rules governing antenna tower lighting and painting
requirements are based upon the advisory recommendations of the FAA, which are set forth in
two FAA Advisory Circulars [47 CFR §§ 17.21–17.58]. Although the FAA lighting and painting
standards are advisory in nature, FCC rules make the standards mandatory.
The FCC always requires an FAA determination that an antenna tower will not pose an avia-
tion hazard before it will grant permission to build the structure. The FAA's determination takes
into consideration the location and height of the proposed tower, and its safety lighting and mark-
ing.
Each new or altered antenna tower structure registered must conform to the FAA's painting
and lighting recommendations set forth on the structure's FAA determination of “no hazard,” and
must be cleared with the FAA and filed with the FCC. If the FAA determines that the tower
would be a physical hazard, the FCC will not approve the construction permit application. When,
however, the FAA determines that there is an aviation hazard due to possible radio frequency
interference (RFI) with aviation communication signals, the FCC makes an independent analysis
of who will be responsible for resolving possible conflicts, and may not automatically defer to
the FAA determination as to what party should bear the cost of any needed equipment changes.
The FAA advisory circulars set forth detailed specifications for lighting and painting. If, how-
ever, the FAA's standards allow more than one lighting or painting option for a particular struc-
ture, the organization seeking FAA clearance for a tower may indicate which of the specified
5-81
types of marking and lighting systems is desired. With respect to telecommunications towers, the
most common option approved by the FAA is the substitution of white flashing lights for a com-
bination of red lights and painting. Any preferences or requests for deviation from standards
must be submitted to the FAA regional office that services the area where the structure would be
located. The FAA regional office then conducts an aeronautical study of the safety of the struc-
ture and considers the proposed deviations or preferences in conducting its analysis.
Where the FAA approves the substitution of high intensity white lights for a combination of
red lights and painting, and the antenna tower is located in a residential neighborhood, the Com-
mission requires the applicant to prepare an environmental assessment [47 CFR § 1.1307(a)(8)].
The Commission, upon review of the environmental assessment, may determine that the pro-
posed substitution of high intensity white lights would not have a significant impact, and may
process the application without further review [47 CFR § 1.1308(d)]. If, however, based upon a
review of the environmental assessment, the Commission determines that the proposed high
intensity lights would have a significant environmental impact upon the human environment, the
Commission will inform the applicant. The applicant will have the opportunity to amend its
application to eliminate the environmental problem. If the problem is not eliminated, the Com-
mission will publish in the Federal Register a Notice of Intent that an Environmental Impact
Statement be prepared [47 CFR § 1.1308(c)]. The Commission may, to assist in the preparation
of an Environmental Impact Statement, request further information from the applicant, interested
persons, and other agencies or authorities. The Commission may also direct that objections to the
proposed lighting be raised with the appropriate state or local authorities [47 CFR § 1.1314(d)].
As part of the its aeronautical study, the FAA may—if it considers it necessary—solicit com-
ments from or convene a meeting of all interested persons for the purpose of gathering all facts
relevant to the effect of the proposed construction on the safe and efficient utilization of the nav-
igable airspace. (See 14 CFR §§ 77.35, 77.41–77.69.) The FAA regional office forwards its rec-
ommendation to FAA headquarters in Washington for final approval. The final FAA
determination also must be submitted to the FCC with any antenna construction permit applica-
tion that requires FAA notification. These structures are subject to inspection and enforcement of
marking and lighting requirements by the FCC.
Examples of various tower lighting options are given in Figure 5.4.1.
5.4.2a Tower Height

Although there is no absolute height limit for antenna towers, both the FCC and FAA have estab-
lished a rebuttable presumption against structures over 2,000 ft above ground level. The FCC has
a policy that applications filed with the FCC for antenna towers higher than 2,000 ft above
ground will be presumed to be inconsistent with the public interest and the applicant will have a
burden of overcoming that strong presumption. The applicant must accompany its application
with a detailed showing directed to meeting this burden. Only in the exceptional case, where the
Commission concludes that a clear and compelling showing has been made that there are public
interest reasons requiring a tower higher than 2,000 ft above ground, and after the parties have
complied with applicable FAA procedures, and full Commission coordination with FAA on the
question of menace to air navigation, will a grant be made. (See 47 CFR § 1.61 Note.)
Figure 5.4.1 Common approaches to tower obstruction marking and lighting. (After [1].)
5.4.3 Ice
Tower owners have long known that atmospheric icing of transmission towers can cause prob-
lems ranging in severity from transmission pattern distortion to complete tower collapse. Ice
forming between antenna radiating elements can cause electrical shorting and equipment burn-
out. Ice can stretch guy lines. Also, towers near populated areas are subject to the added liability
of falling ice, which threatens lives and surrounding property.
There are two recognized sources of ice accretion. The first is in-cloud icing, in which super-
cooled water droplets float in the air and contact a surface because of air movement. The second
is precipitational icing, where the droplets are massive enough to fall from the atmosphere onto
the tower structure.
These two sources form three types of ice, as illustrated in Figure 5.4.2. Glaze ice is usually the
product of freezing rain or of airborne spray from nearby bodies of water. It forms at relatively
high temperatures (0°C to −3°C) and forms on surfaces as a tightly bonded, clear, dense, glass-
like coating. This type of icing is the most serious threat to structures because of its density and
the large additional loads it may impart.
Rime, or fluffy, white ice, forms more frequently than glaze in mountainous areas. Rime ice
varies from “soft” to “hard” depending on its density, clarity, and crystal structure. Soft rime
forms at low temperatures (−5°C to −25°C) and low wind speeds. The impinging droplets freeze
quickly, trapping air as the accretion grows. The large

amount of entrapped air is responsible for the opaque-
white and fluffy appearance of rime.
Because of its lower density, soft rime is not usually
problematic. Hard rime, on the other hand, is halfway
between glaze and soft rime in terms of density, clarity,
and hardness; it can be as dangerous as glaze ice. Frost, a
fairly harmless form of icing, forms when in-cloud mois-
ture freezes on a surface, in still air.
Many forms of ice can form simultaneously on the
same structure, depending on surface features such as
shape, exposure, and heat-dissipation characteristics of
the structure.
5.4.3a Problems Caused by Icing

Icing imparts additional dead weight to the structure and
also presents a larger surface area to the wind. Towers
must be periodically surveyed for signs of structural
Figure 5.4.2 Common types of ice
fatigue from repeated ice accretion and windloading, and
formation on tower structures. (After
guy lines should be inspected for wear and retensioned to [1].)
counteract stretching.
When ice falls, it is called “shedding.” Guyed towers
are especially prone to failure from uneven shedding. Heavy ice-loading on the guys exerts tre-
mendous tension, and when individual guy lines suddenly release a load, torsional forces may
overcome the strength of the tower.
Harmonic oscillation of guy lines, or “guy galloping,” is a rare but extreme type of ice-
induced stress. It has been proposed that a small amount of ice building up on the windward side
causes a cable to assume the shape on an airfoil. A moderate wind then can induce the cable to
move because of an aerodynamic lift and drop phenomenon. Galloping occurs when the move-
ment matches the resonant frequency of the cable, resulting in increasing oscillation amplitudes.
The danger lies in the fact that galloping usually affects only one or two lines of an entire guy
system, which can produce violent twisting of the tower. An added danger is the fact that metal
becomes more brittle and subject to failure when cold. Damping guys can be installed, which
tend to limit such oscillation.
Tower-mounted items are subject to damage from falling ice shed from the upper levels.
Threatened items include transmission lines, reflector dishes, and antenna elements. Falling ice
chunks of considerable size, weighing tens of kilograms, are common during shedding events.
Icing and shedding are usually the result of specific storm patterns, and facility personnel
often can predict from past experience the onset of a dangerous situation. In light of the likely
storm track and the associated wind directions, buildings beneath the tower normally are situated
to the windward side, for protection from falling ice. Transmitter roof buildings are likely to be
constructed to absorb impacts and resist punctures. Vulnerable items on the tower can be
shielded from above with wood, sheet metal, or wire-frame construction.
5.4.3b Prevention and Ice Control

Falling ice is a difficult problem because there are no feasible proven means of prevention avail-
able for tall masts. The best way to guard against damage to adjacent property is to restrict land
usage in the icefall shadow of the tower. Initially, the tower should be constructed on a vacant
parcel of land large enough to encompass the highly probable fall zone. Thereafter, land-use
planners should be cognizant of the danger and restrict development in this zone.
Many different approaches have been taken to prevent ice accretion, to minimize its severity,
or to aid in its removal. “Anti-icing” methods minimize or prevent accretion, whereas “de-icing”
methods remove the ice once it has formed. Because of the large size of transmitting towers,
many of the traditional anti-icing and de-icing methods are not cost-effective when applied to the
whole tower. They usually are applied only to sections immediately surrounding the antennas.
Popular techniques for ice control include the following:
• Shrouding. Atmospheric icing has been shown, in theory, to be diminished by increasing the
diameter of superstructure elements. This reduces the ability of the structure to collect water
droplets. This idea has been used with success on arctic oil drilling platforms by enclosing the
superstructure in a solid panelwork. Radomes are an example of the use of this principle.
Application of this concept to broadcasting is limited to short, sturdy towers that are not sub-
ject to excessive windloading.
• Flexure. Outfitting radomes or exposed elements with a flexible sheathing has been success-
ful at some installations. Flexure is caused either passively by wind and vibratory action, or
by an active pneumatic system.
• Low-adhesion coatings. Another approach to icing protection has been in the areas of icepho-
bic or low-ice-adhesion coatings. The types of coatings studied have been freezing-point
depressants and low-wettability substances. Freezing-point depressants, such as glycol solu-
tions, soluble salt solutions, and gas-evolving coatings, function by contaminating the accret-
ing droplets and reducing the freezing point to below that of pure water. Sloping or vertical
surfaces then will shed the liquid so ice does not form. As such, freezing-point depressants
are classified as “sacrificial coatings” because they are continually being washed away and
must somehow be replenished. Highway salting and aircraft wing de-icing are common appli-
cations of these materials.
Low-wettability oils, greases, and permanent coatings have been pursued because of their
hydrophobicity. However, it is incorrect to assume that because a coating sheds water it will nec-
essarily shed ice. During the early stages of icing, these substances allow the droplets to run off a
sloping surface more rapidly before freezing can occur. Eventually, some droplets accrete before
they can be shed. In turn, these create sites for further accretion and the hydrophobic coating
thereafter rapidly becomes coated with ice.
Studies have shown that certain polymer coatings exhibit a lower adhesive strength for ice
than bare metal surfaces. Some broadcasters insist that the glossy surfaces of their antennas’
radiating elements help prevent the formation of ice.
Heating
The only totally effective anti-icing method commonly available is heating, and it is the method
of choice for most station owners. Given the large power demands, heating is—in general—used
only to prevent icing of the radiating elements of FM and TV antennas. Heating units are fac-
tory-built into antennas and must be activated before icing can begin. These low-wattage heaters
usually cannot keep up with the accretion rate if ice is allowed to accumulate appreciably before
the heaters are activated. Some station operators manually activate heaters based on the local
weather forecast or individual judgment. Others prefer the more cautious alternative of operating
de-icers for the entire season. A third alternative is to provide for automatic activation via ther-
mal, precipitation, and/or icing sensors.
5.4.4 References
1. Lehtinen, Rick: “Hardening Towers,” Broadcast Engineering, Overland Park, Kan., pp.
94–106, March 1990.
5.4.5 Bibliography
Bishop, Don: “How the FCC Decides Whether Your Tower is OK,” Mobile Radio Technology,
Bishop, Don: “Tower Maintenance,” Mobile Radio Technology, Intertec Publishing, Overland
Park, Kan., April 1988.
Bishop, Don: “When to Use Strobes on Communications Towers,” Mobile Radio Technology,
Mulherin, Nathan D.: “Atmospheric Icing on Communication Masts in New England.” CRREL
Report 86-17, U.S. Army Cold Regions Research and Engineering Laboratory, December
1986.
g g
Chapter
5.5
Tower Grounding
5.5.1 Introduction
The attention given to the design and installation of a tower ground system is a key element in
the day-to-day reliability of the transmission plant. A well-designed and -installed ground net-
work is invisible to the engineering staff. A marginal ground system, however, will cause prob-
lems on a regular basis. Grounding schemes can range from simple to complex, but any system
serves three primary purposes:
• Provides for operator safety.
• Protects electronic equipment from damage caused by transient disturbances.
• Diverts stray radio frequency energy from sensitive audio, video, control, and computer
equipment.
Any ground system consists of two key elements: 1) the earth-to-grounding electrode inter-
face, and 2) the RF, ac power, and signal-wiring systems.
5.5.2 Terms and Codes

A facility can be defined as something that is built, installed, or established to serve a particular
purpose [1]. A facility is usually thought of as a single building or group of buildings. The
National Electrical Code (NEC) uses the term premises to refer to a facility when it defines pre-
mises wiring as the interior and exterior (facility) wiring, such as power, lighting, control, and
signal systems. Premises wiring includes the service and all permanent and temporary wiring
between the service and the load equipment. Premises wiring does not include wiring internal to
any load equipment.
5.5.2a The Need for Grounding

The Institute of Electrical and Electronic Engineers (IEEE) defines grounding as “a conducting
connection, whether intentional or accidental, by which an electric circuit or equipment is con-
nected to the earth, or to some conducting body of relatively large extent that serves in place of
5-87
g
the earth. It is used for establishing and maintaining the potential of the earth (or of the conduct-
ing body), or approximately that potential, on conductors connected to it, and for conducting
ground current to and from the earth (or the conducting body).” [2] Based on this definition, the
reasons for grounding can be identified as:
• Personnel safety by limiting potentials between all non-current-carrying metal parts of an
electrical distribution system.
• Personnel safety and control of electrostatic discharge (ESD) by limiting potentials between
all non-current-carrying metal parts of an electrical distribution system and the earth.
• Fault isolation and equipment safety by providing a low-impedance fault return path to the
power source to facilitate the operation of overcurrent devices during a ground fault.
The IEEE definition makes an important distinction between ground and earth. Earth refers
to mother earth, and ground refers to the equipment grounding system, which includes equip-
ment grounding conductors, metallic raceways, cable armor, enclosures, cabinets, frames, build-
ing steel, and all other non-current-carrying metal parts of the electrical distribution system.
There are other reasons for grounding not implicit in the IEEE definition. Overvoltage con-
trol has long been a benefit of proper power system grounding, and is described in IEEE Stan-
dard 142, also known as the Green Book [3]. With the increasing use of electronic computer
systems, noise control has become associated with the subject of grounding, and is described in
IEEE Standard 1100, the Emerald Book. [4].
5.5.3 The Grounding Electrode

The process of connecting the grounding system to earth is called earthing, and consists of
immersing a metal electrode or system of electrodes into the earth [5]. The conductor that con-
nects the grounding system to earth is called the grounding electrode conductor. The function of
the grounding electrode conductor is to keep the entire grounding system at earth potential (i.e.,
voltage equalization during lightning and other transients) rather than for conducting ground-
fault current. Therefore, the NEC allows reduced sizing requirements for the grounding elec-
trode conductor when connected to made electrodes [6].
The basic measure of effectiveness of an earth electrode system is the earth electrode resis-
tance. Earth electrode resistance is the resistance, in ohms, between the point of connection and
a distant point on the earth called remote earth. Remote earth, about 25 feet from the driven elec-
trode, is the point where earth electrode resistance does not increase appreciably when this dis-
tance is increased. Earth electrode resistance consists of the sum of the resistance of the metal
electrode (negligible) plus the contact resistance between the electrode and the soil (negligible)
plus the soil resistance itself. Thus, for all practical purposes, earth electrode resistance equals
the soil resistance. The soil resistance is nonlinear, with most of the earth resistance contained
within several feet of the electrode. Furthermore, current flows only through the electrolyte por-
tion of the soil, not the soil itself. Thus, soil resistance varies as the electrolyte content (moisture
and salts) of the soil varies. Without electrolyte, soil resistance would be infinite.
Soil resistance is a function of soil resistivity. A one cubic meter sample of soil with a resis-
tivity ρ of one ohm-meter will present a resistance R of one ohm between opposite faces. A
broad variation of soil resistivity occurs as a function of soil types, and soil resistivity can be esti-
mated or measured directly. Soil resistivity is usually measured by injecting a known current into
g
a given volume of soil and measuring the resulting voltage drop. When soil resistivity is known,
the earth electrode resistance of any given configuration (single rod, multiple rods, or ground
ring) can be determined by using standard equations developed by Sunde [7], Schwarz [8], and
others.
Earth resistance values should be as low as practicable, but are a function of the application.
The NEC approves the use of a single made electrode if the earth resistance does not exceed 25
Ω. Methods of reducing earth resistance values include the use of multiple electrodes in parallel,
the use of ground rings, increased ground rod lengths, installation of ground rods to the perma-
nent water level, increased area of coverage of ground rings, and the use of concrete-encased
electrodes, ground wells, and electrolytic electrodes.
5.5.3a Earth Electrode

Earth electrodes may be made electrodes, natural electrodes, or special-purpose electrodes [5].
Made electrodes include driven rods, buried conductors, ground mats, buried plates, and ground
rings. The electrode selected is a function of the type of soil and the available depth. Driven elec-
trodes are used where bedrock is 10 ft or more below the surface. Mats or buried conductors are
used for lesser depths. Buried plates are not widely used because of the higher cost when com-
pared to rods. Ground rings employ equally spaced driven electrodes interconnected with buried
conductors. Ground rings are used around large buildings, around small unit substations, and in
areas having high soil resistivity.
Natural electrodes include buried water pipe electrodes and concrete-encased electrodes. The
NEC lists underground metal water piping, available on the premises and not less than 10 feet in
length, as part of a preferred grounding electrode system. Because the use of plastic pipe in new
water systems will impair the effectiveness of water pipe electrodes, the NEC requires that metal
underground water piping be supplemented by an additional approved electrode. Concrete below
ground level is a good electrical conductor. Thus, metal electrodes encased in such concrete will
function as excellent grounding electrodes. The application of concrete-encased electrodes is
covered in IEEE Standard 142.
5.5.4 Establishing an Earth Ground

The grounding electrode is the primary element of any ground system. The electrode can take
many forms. In all cases, its purpose is to interface the electrode (a conductor) with the earth (a
semiconductor). Grounding principles have been refined to a science. Still, however, many mis-
conceptions exist regarding grounding. An understanding of proper grounding procedures begins
with the basic earth-interface mechanism.
5.5.4a Grounding Interface

The grounding electrode (or ground rod) interacts with the earth to create a hemisphere-shaped
volume, as illustrated in Figure 5.5.1. The size of this volume is related to the size of the ground-
ing electrode. The length of the electrode has a much greater effect than the diameter. Studies
have demonstrated that the earth-to-electrode resistance from a driven ground rod increases
exponentially with the distance from that rod. At a given point, the change becomes insignifi-
g
cant. It has been found that for maximum effectiveness

of the earth-to-electrode interface, each ground rod
requires a hemisphere-shaped volume with a diameter
that is approximately 2.2 times the rod length [9].
The constraints of economics and available real
estate place practical limitations on the installation of
a ground system. It is important, however, to keep the
2.2 rule in mind because it allows the facility design
engineer to use the available resources to the best
advantage. Figure 5.5.2 illustrates the effects of locat-
ing ground rods too close (less than 2.2 times the rod
length). An overlap area is created that effectively
wastes some of the earth-to-electrode capabilities of
the two ground rods. For example, two 10-ft ground
rods driven only 1 ft apart provide about the same
resistivity as a single 10-ft rod. Figure 5.5.1 The effective earth-inter-
There are three schools of thought with regard to face hemisphere resulting from a sin-
ground-rod length. The first approach suggests that gle driven ground rod. The 90 percent
effective area of the rod extends to a
extending ground-rod length beyond about 10 ft is of
radius of approximately 1.1 times the
little value for most types of soil. The reasoning length of the rod. (After [10].)
behind this conclusion is presented in Figure 5.5.3a,
where ground resistance is plotted as a function of
ground-rod length. Beyond 10 ft in length, a point of
diminishing returns is reached.
The second school of thought suggests that opti-
mum earth-to-electrode interface is achieved with long
(40 ft or greater) rods, driven to penetrate the local
water table. When planning this type of installation,
consider the difficulty that may be encountered when
attempting to drive long ground rods. The foregoing
discussion assumes that the soil around the grounding
electrode is reasonably uniform in composition.
Depending upon the location, however, this assump-
tion may not hold true. Figure 5.5.2 The effect of overlapping
The third school of thought suggests that the opti- earth interface hemispheres by plac-
mum ground rod installation is achieved by using the ing two ground rods at a spacing less
than 2.2 times the length of either rod.
longest possible rod depth (length). Data to support
(After [10].)
this conclusion is given in Figure 5.5.3b, which plots
ground rod performance as a function of depth in soil
of uniform resistivity. This curve does not take into account seasonal moisture content, changing
chemical composition at different soil layers, and frozen soil conditions.
Given these three conflicting approaches, the benefits of hiring an experienced, licensed pro-
fessional engineer to design the ground system can be readily appreciated.
Horizontal grounding electrodes provide essentially the same resistivity as an equivalent-
length vertical electrode, given uniform soil conditions. As Figure 5.5.4 demonstrates, the differ-
ence between a 10-ft vertical and a 10-ft horizontal ground rod is negligible (275 Ω vs. 250 Ω).
This comparison includes the effects of the vertical connection element from the surface of the
g
(a )
(b )
Figure 5.5.3 Charted grounding resistance as a function of ground-rod length: (a) data demon-
strating that ground-rod length in excess of 10 ft produces diminishing returns (1-in.-diameter rod)
[10], (b) data demonstrating that ground system performance continues to improve as depth
increases. (Chart b from [1]. Used with permission.)
ground to the horizontal rod. Taken by itself, the horizontal ground rod provides an earth-inter-
face resistivity of approximately 308 Ω when buried at a depth of 36 in.
Ground rods come in many sizes and lengths. The more popular sizes are 1/2, 5/8, 3/4, and 1
in. The 1/2-in. size is available in steel with stainless-clad, galvanized, or copper-clad rods. All-
stainless-steel rods also are available. Ground rods can be purchased in unthreaded or threaded
(sectional) lengths. The sectional sizes are typically 9/16-in. or 1/2-in. rolled threads. Couplers
are made from the same materials as the rods. Couplers can be used to join 8- or 10-ft-length
rods together. A 40-ft ground rod, for example, is driven one 10-ft section at a time.
The type and size of ground rod used is determined by how many sections are to be connected
and how hard or rocky the soil is. Copper-clad 5/8-in. × 10-ft rods are probably the most popular.
Copper cladding is designed to prevent rust. The copper is not primarily to provide better con-
ductivity. Although the copper certainly provides a better conductor interface to earth, the steel
g
Figure 5.5.4 The effectiveness of vertical ground rods compared with horizontal ground rods.
(After [10].)
Table 5.5.1 Typical Resistivity of Common Soil Types
Type of Soil Resistivity in Ω / cm Average Minimum Maximum

Filled land, ashes, salt marsh 2,400 600 7,000
Top soils, loam 4,100 340 16,000
Hybrid soils 6,000 1,000 135,000
Sand and gravel 90,000 60,000 460,000
that it covers is also an excellent conductor when compared with ground conductivity. The thick-
ness of the cladding is important only insofar as rust protection is concerned.
Wide variations in soil resistivity can be found within a given geographic area, as documented
in Table 5.5.1. The wide range of values shown results from differences in moisture content,
mineral content, and temperature.
Temperature is a major concern in shallow grounding systems because it has a significant
effect on soil resistivity [11]. During winter months, the ground system resistance can rise to
unacceptable levels because of the freezing of liquid water in the soil. The same shallow ground-
ing system can also suffer from high resistance in the summer as moisture is evaporated from
soil. It is advisable to determine the natural frost line and moisture profile for an area before
attempting design of a ground system.
Figure 5.5.5 describes a four-point method for in-place measurement of soil resistivity. Four
uniformly spaced probes are placed in a linear arrangement and connected to a ground resistance
test meter. An alternating current (at a frequency other than 60 Hz) is passed between the two
most distant probes, resulting in a potential difference between the center potential probes. The
meter display in ohms of resistance can then be applied to determine the average soil resistivity
in ohm-centimeters for the hemispherical area between the C1 and P2 probes.
Soil resistivity measurements should be repeated at a number of locations to establish a resis-
tivity profile for the site. The depth of measurement can be controlled by varying the spacing
between the probes. In no case should the probe length exceed 20 percent of the spacing between
probes.
g
Figure 5.5.5 The four-point method for soil resistivity measurement. (From [11]. Used with permis-
sion.)
After the soil resistivity for a site is known, calculations can be made to determine the effec-
tiveness of a variety of ground system configurations. Equations for several driven rod and radial
cable configurations are given in [11], which—after the solid resistivity is known—can be used
for the purpose of estimating total system resistance. Generally, driven rod systems are appropri-
ate where soil resistivity continues to improve with depth or where temperature extremes indicate
seasonal frozen or dry soil conditions. Figure 5.5.6 shows a typical soil resistivity map for the
U.S.
5.5.4b Ground Electrode Testing

Testing of all ground electrodes before they are connected to form a complex network is a fairly
simple process that is well described in the documentation included with ground electrode
meters. This instructional process, therefore, will not be described here. Also, the system as a
whole, should be tested after all interconnections are made, providing a benchmark for future
tests.
At a new site, it is often advisable to perform ground system tests before the power company
ground/neutral conductor is attached to the system. Conduct a before-and-after test with probes
in the same position to determine the influence of the power company attachment [11]. It is also
worthwhile to install permanent electrodes and marker monuments at the original P2 and C2
probe positions to ensure the repeatability of future tests.
g
Figure 5.5.6 Typical soil resistivity map for the U.S. (From [11]. Used with permission.)
5.5.4c Chemical Ground Rods

A chemically activated ground system is an alternative to the conventional ground rod. The con-
cept behind the chemical ground rod is to increase the earth-to-electrode interface by condition-
ing the soil surrounding the rod. Experts have known for many years that the addition of ordinary
table salt (NaCl) to soil will reduce the resistivity of the earth-to-ground electrode interface.
With the proper soil moisture level (4 to 12 percent), salting can reduce soil resistivity from per-
haps 10,000 Ω/m to less than 100 Ω/m. Salting the area surrounding a ground rod (or group of
rods) follows a predictable life-cycle pattern, as illustrated in Figure 5.5.7. Subsequent salt appli-
cations are rarely as effective as the initial salting.
Various approaches have been tried over the years to solve this problem. One such product is
shown in Figure 5.5.8. This chemically activated grounding electrode consists of a 2-1/2-in.-
diameter copper pipe filled with rock salt. Breathing holes are provided on the top of the assem-
bly, and seepage holes are located at the bottom. The theory of operation is simple. Moisture is
absorbed from the air (when available) and is then absorbed by the salt. This creates a solution
that seeps out of the base of the device and conditions the soil in the immediate vicinity of the
rod.
Another approach is shown in Figure 5.5.9. This device incorporates a number of ports
(holes) in the assembly. Moisture from the soil (and rain) is absorbed through the ports. The
metallic salts subsequently absorb the moisture, forming a saturated solution that seeps out of the
ports and into the earth-to-electrode hemisphere. Figure 5.5.10 shows the measured performance
of a typical chemical ground rod in three types of soil.
g
Figure 5.5.7 The effect of soil salting on ground-rod resistance with time. The expected resalting
period, shown here as two years, varies depending on the local soil conditions and the amount of
moisture present. (After [10].)
(a) (b)
Figure 5.5.8 An air-breathing chemically activated ground rod: (a) breather holes at the top of the
device permit moisture penetration into the chemical charge section of the rod, (b) a salt solution
seeps out of the bottom of the unit to form a conductive shell. (After [10].)
Implementations of chemical ground-rod systems vary depending on the application. Figure

5.5.11 illustrates a counterpoise ground consisting of multiple leaching apertures connected in a
spoke fashion to a central hub. The system is serviceable in that additional salt compound can be
added to the hub at required intervals to maintain the effectiveness of the ground. Figure 5.5.12
g
Figure 5.5.9 An alternative approach to the chemically activated ground rod. Multiple holes are
provided on the ground-rod assembly to increase the effective earth-to-electrode interface. Note
that chemical rods can be produced in a variety of configurations. (After [10].)
Figure 5.5.10 Measured performance of a chemical ground rod. (After [10].)
shows a counterpoise system made up of individual chemical ground rods interconnected with
radial wires buried below the surface.
5.5.4d Ufer Ground System

Driving ground rods is not the only method of achieving a good earth-to-electrode interface [9].
The concept of the Ufer ground has gained interest because of its simplicity and effectiveness.
g
Figure 5.5.11 Hub and spoke counterpoise ground system. (After [10].)
Figure 5.5.12 Tower grounding scheme using buried copper radials and chemical ground rods.
(After [10].)
The Ufer approach (named for its developer), however, must be designed into a new structure. It
cannot be added on later. The Ufer ground takes advantage of the natural chemical- and water-
retention properties of concrete to provide an earth ground. Concrete typically retains moisture
for 15 to 30 days after a rain. The material has a ready supply of ions to conduct current because
of its moisture-retention properties, mineral content, and inherent pH. The large mass of any
concrete foundation provides a good interface to ground.
A Ufer system, in its simplest form, is made by routing a solid-copper wire (no. 4 gauge or
larger) within the foundation footing forms before concrete is poured. Figure 5.5.13 shows one
such installation. The length of the conductor run within the concrete is important. Typically a
20-ft run (10 ft in each direction) provides a 5 Ω ground in 1000 Ω/m soil.
g
As an alternative, steel reinforcement

bars (rebar) can be welded together to
provide a rigid, conductive structure. A
ground lug is provided to tie equipment
to the ground system in the foundation.
The rebar must be welded, not tied,
together. If it is only tied, the resulting
poor connections between rods can
result in arcing during a current surge.
This can lead to deterioration of the
concrete in the affected areas.
The design of a Ufer ground is not to
be taken lightly. Improper installation
can result in a ground system that is
subject to problems. The grounding
electrodes must be kept a minimum of 3 Figure 5.5.13 The basic concept of a Ufer ground
in. from the bottom and sides of the con- system. (After [9].)
crete to avoid the possibility of founda-
tion damage during a large lightning
surge. If an electrode is placed too near
the edge of the concrete, a surge could
turn the water inside the concrete to
steam and break the foundation apart.
The Ufer approach also can be
applied to guy-anchor points or a tower
base, as illustrated in Figure 5.5.14.
Welded rebar or ground rods sledged in
place after the rebar cage is in position
can be used. By protruding below the
bottom concrete surface, the ground
rods add to the overall electrode length
to help avoid thermal effects that can
crack the concrete. The maximum
length necessary to avoid breaking the Figure 5.5.14 The Ufer ground system as applied to
concrete under a lightning discharge is a transmission-tower base or guy-anchor point.
determined by the following: (After [9].)
• Type of concrete (density, resistivity,

and other factors)
• Water content of the concrete
• How much of the buried concrete surface area is in contact with the ground
• Ground resistivity
• Ground water content
• Size and length of the ground rod
• Size of lightning flash
g
The last variable is a bit of a gamble. The 50 percent mean occurrence of lightning strikes is
perhaps 18 A, but superstrikes can occur that approach 100 to 200 kA.
Before implementing a Ufer ground system, consult a qualified contractor. Because the Ufer
ground system will be the primary grounding element for the facility, it must be done correctly.
5.5.5 Bonding Ground-System Elements

A ground system is only as good as the methods used to interconnect the component parts [9].
Do not use soldered-only connections outside the equipment building. Crimped/brazed and exo-
thermic (Cadwelded) connections are preferred. (Cadweld is a registered trademark of Erico
Corp.) To make a proper bond, all metal surfaces must be cleaned, any finish removed to bare
metal, and surface preparation compound applied. Protect all connections from moisture by
appropriate means, usually sealing compound and heat-shrink tubing.
It is not uncommon for an untrained installer to use soft solder to connect the elements of a
ground system. Such a system is doomed from the start. Soft-soldered connections cannot stand
up to the acid and mechanical stress imposed by the soil. The most common method of connect-
ing the components of a ground system is silver soldering. This process requires the use of braz-
ing equipment, which may be unfamiliar to many facility engineers. The process uses a high-
temperature/high-conductivity solder to complete the bonding process. For most grounding sys-
tems, however, the best approach to bonding is the exothermic process.
5.5.5a Exothermic Bonding

Exothermic bonding is the preferred method of connecting the elements of a ground system [9].
Molten copper is used to melt connections together, forming a permanent bond. This process is
particularly useful in joining dissimilar metals. In fact, if copper and galvanized cable must be
joined, exothermic bonding is the only acceptable means. The completed connection will not
loosen or corrode and will carry as
much current as the cable connected to
it. Figure 5.5.15 illustrates the bonding
that results from the exothermic pro-
cess.
The bond accomplished by dumping
powdered metals (copper oxide and alu-
minum) from a container into a graphite
crucible and igniting the material by
means of a flint lighter. Reduction of the
copper oxide by the aluminum produces
molten copper and aluminum oxide
slag. The molten copper flows over the
conductors, bonding them. Figure
5.5.16 shows a typical mold. A variety
of special-purpose molds are available
to join different-size cables and copper Figure 5.5.15 The exothermic bonding process.
strap. Figure 5.5.17 shows the bonding (After [9].)
g
process for a copper-strap-to-ground-

rod interface.
Variations on the basic Cadweld sys-
tem include a “smokeless” exothermic
process, recommended for indoor use,
and disposable “one-shot” devices that
simplify installation. The one-shot
approach is particularly well suited to
small jobs or when work is performed
by less experienced installers.
5.5.5b Ground-System Inductance

Conductors interconnecting sections or
components of an earth ground system
must be kept as short as possible to be
fully effective [9]. The inductance of a
conductor is a major factor in its charac-
teristic impedance to surge energy. For
Figure 5.5.16 Exothermic bonding mold for connect-
example, consider a no. 6 AWG copper
ing a cable to a ground rod. (After [9].)
wire 10 m in length. The wire has a dc
resistance of 0.013 Ω and an inductance
of approximately 10 μH. For a 1000 A
lightning surge with a 1 μs rise time, the
resistive voltage drop will be 13 V, but
the reactive voltage drop will be 10 kV.
Furthermore, any bends in the conduc-
tor will increase its inductance and fur-
ther decrease the effectiveness of the
wire. Bends in ground conductors
should be gradual. A 90° bend is elec-
trically equivalent to a 1/4-turn coil. The
sharper the bend, the greater the induc-
tance.
Because of the fast rise time of most
lightning discharges and power-line
transients, the skin effect plays an
important role in ground-conductor Figure 5.5.17 Mold for connecting a copper strap to
selection. When planning a facility a ground rod. (After [9].)
ground system, view the project from an
RF standpoint.
The effective resistance offered by a conductor to radio frequencies is considerably higher
than the ohmic resistance measured with direct currents. This is because the skin effect causes
the currents to be concentrated in certain parts of the conductor and leaves the remainder of the
cross section to contribute little or nothing toward carrying the applied current.
g
When a conductor carries an alternating

current, a magnetic field is produced that
surrounds the wire. This field continually is Magnetic flux
expanding and contracting as the ac current

wave increases from zero to its maximum
positive value and back to zero, then through
its negative half-cycle. The changing mag-
netic lines of force cutting the conductor
induce a voltage in the conductor in a direc-
tion that tends to retard the normal flow of
current in the wire. This effect is more pro-
nounced at the center of the conductor. Thus,
current within the conductor tends to flow Distribution of
current density
more easily toward the surface of the wire.
The higher the frequency, or the faster the Figure 5.5.18 Skin effect on an isolated round
rise time of the applied waveform, the conductor carrying a high-frequency signal.
greater the tendency for current to flow at the
surface. Figure 5.5.18 shows the distribution of current in a radial conductor.
When a circuit is operating at high frequencies, the skin effect causes the current to be redis-
tributed over the conductor cross section in such a way as to make most of the current flow where
it is encircled by the smallest number of flux lines. This general principle controls the distribu-
tion of current regardless of the shape of the conductor involved. With a flat-strip conductor, the
current flows primarily along the edges, where it is surrounded by the smallest amount of flux.
5.5.6 Grounding Tower Elements

Guyed towers usually are better than self-supporting towers at dissipating lightning surge cur-
rents [9]. This is true, however, only if the guy anchors are grounded properly. Use of the Ufer
technique is one way of effectively grounding the anchors. For anchors not provided with a Ufer
ground during construction, other, more conventional, techniques can be used. For guyed towers
whose guy lines are not electrically continuous between the tower and the base, such as section-
alized towers (certain FM and TV broadcast towers and AM broadcast towers), surge current dis-
sipation is essentially equivalent to a self-supporting tower.
Never rely on the turnbuckles of a guy anchor as a path for lightning energy. The current
resulting from a large flash can weld the turnbuckles in position. If the turnbuckles are provided
with a safety loop of guy cable (as they should be), the loop can be damaged where it contacts
the guys and turnbuckle. Figure 5.5.19 shows the preferred method of grounding guy wires: tie
them together above the loop and turnbuckles. Do not make these connections with copper wire,
even if they are Cadwelded. During periods of precipitation, water shed from the top copper wire
will carry ions that may react with the lower galvanized (zinc) guy wires. This reaction washes
off the zinc coating, allowing rust to develop.
The best way to make the connection is with all-galvanized materials. This includes the
grounding wire, clamps, and ground rods. It may not be possible to use all galvanized materials
because, at some point, a connection to copper conductors will be required. Battery action
caused by the dissimilar metal junction may allow the zinc to act as a sacrificial anode. The zinc
g
eventually will disappear into the soil,

leaving a bare steel conductor that can
fall victim to rust.
Ruling out an all-galvanized system,
the next best scheme uses galvanized
wire (guy-wire material) to tie the guy
wires together. Just above the soil, Cad-
weld the galvanized wire to a copper
conductor that penetrates below grade to
the perimeter ground system. The height
above grade for the connection is deter-
mined by the local snowfall or flood
level. The electric conductivity of snow,
although low, can cause battery action
from the copper through the snow to the
zinc. The Cadwelded joint should be
positioned above the usual snow or Figure 5.5.19 Recommended guy-anchor ground-
flood level. ing procedure. (After [9].)
5.5.6a Ground-Wire Dressing

Figure 5.5.20 illustrates the proper way
to bond the tower base ground leads to
the buried ground system [9]. Dress the
leads close to the tower from the lowest
practical structural element at the base.
Keep the conductors as straight and
short as possible. Avoid any sharp
bends. Attach the ground wires to the
tower only at one or more existing bolts
(or holes). Do not drill any holes into
the tower. Do not loosen any bolts to
make the ground-wire attachment. Use
at least two 3- to 4-in copper straps
between the base of the tower and the
buried ground system. Position the
straps next to the concrete pier of the
tower base. For towers more than 200 ft
in height, use four copper straps, one on
each side of the pier.
Figure 5.5.21 illustrates the proper
way to bond guy wires to the buried
ground system. The lead is dressed
straight down from the topmost to the
lowest guy. It should conform as close Figure 5.5.20 Ground-conductor dressing for the ba
to vertical as possible, and be dressed of a guyed tower. (After [9].)
g
Figure 5.5.21 Top view of recommended guy-anchor grounding technique. (After [9].)
downward from the lower side of each guy wire after connecting to each wire. To ensure that no
arcing will occur through the turnbuckle, a connection from the anchor plate to the perimeter
ground circle is recommended (use no. 2 gauge copper wire). This helps minimize the unavoid-
able inductance created by the conductor being in the air. Interconnect leads that are suspended
in air must be dressed so that no bending radius is less than 8 in.
A perimeter ground—a circle of wire connected at several points to ground rods driven into
the earth—should be installed around each guy-anchor point. The perimeter system provides a
good ground for the anchor, and when tied together with the tower base radials, acts to rapidly
dissipate lightning energy in the event of a flash. Tower base radials are buried wires intercon-
nected with the tower base ground that extend away from the center point of the structure.
The required depth of the perimeter ground and the radials depends upon soil conductivity.
Generally speaking, however, about 8 in. below grade is sufficient. In soil with good conductiv-
ity, the perimeter wire may be as small as no. 10 gauge. Because no. 2 gauge is required for the
segment of conductor suspended in air, it may be easier to use no. 2 throughout. An added advan-
tage is that the same size Cadweld molds may be used for all bonds.
5.5.6b Facility Ground Interconnection

Any radial that comes within 2 ft of a conductive structure must be tied into the ground system
[9]. Bury the interconnecting wire, if possible, and approach the radial at a 45° angle, pointing
toward the expected surge origin (usually the tower). Cadweld the conductor to the radial and to
the structure.
For large-base, self-supporting towers, the radials should be split among each leg pad, as
shown in Figure 5.5.22. The radials may be brought up out of the soil (in air) and each attached at
spaced locations around the foot pad. Some radials may have to be tied together first and then
g
joined to the foot pad. Remember, if space

between the radial lines can be maintained, less
mutual inductance (coupling) will exist, and the
system surge impedance will be lower.
It is desirable to have a continuous one-piece
ring with rods around the equipment building.
Connect this ring to one, and only one, of the
tower radials, thus forming no radial loops. See
Figure 5.5.23. Bury the ring to the same depth as
the radials to which it interconnects. Connect
power-line neutral to the ring. Warning: Substan-
tial current may flow when the power-line neutral
is connected to the ring. Follow safety procedures
when making this connection.
Install a ground rod (if the utility company has
not installed one) immediately outside the genera-
tor or utility company vault, and connect this rod
to the equipment building perimeter ground ring.
Route a no. 1/0 insulated copper cable from the
main power panel inside the generator vault to the
ground rod outside the vault. Cut the cable to
length, strip both ends, and tie one end to the
power-line neutral at the main power panel in the
generator vault. Connect the other end to the Figure 5.5.22 Interconnecting a self-sup-
porting tower to the buried ground system.
ground rod. Warning: Use care when making this
(After [9].)
Figure 5.5.23 Interconnecting the metal structures of a facility to the ground system. (After [9].)
g
Figure 5.5.24 Protection methods for personnel at an exposed site. (From [11]. Used with permis-
sion.)
connection. Hazardous voltage may exist between the power-line neutral and any point at earth
potential.
Do not remove any existing earth ground connections to power-line neutral, particularly if
they are installed by the power company. To do so may violate the local electrical code. The goal
of this interconnection is to minimize noise that may be present on the neutral, and to conduct
this noise as directly as possible outside to earth ground.
5.5.6c Personnel Protection

The threat to personnel during a lightning strike ranges from the obvious danger of direct contact
with lightning to the more obscure effects of step and touch voltages [11]. Protection from a
direct strike when near or within structures is accomplished with traditional rolling sphere con-
cept methods. Step and touch potentials, however, are created as a lightning current passes
through resistive soil and other available paths as it dissipates into the earth. A person in contact
with only one point of the gradient will simply rise and fall in potential with the gradient without
injury. A person in contact with multiple points on the earth or objects at different potentials
along the gradient, however, will become part of the current path and may sustain injury or death
[12].
Figure 5.5.24 [13, 14] illustrates a number of methods for protecting personnel from the
direct and secondary effects of lightning. A typical tower/transmitter site is used as an example.
A technician responding to a service problem during a thunderstorm would likely exit his or her
g
vehicle outside the gate, unlock and open the gate, and move the vehicle into the inside yard. The
technician would then leave the vehicle, and enter the building.
The threat of a direct lightning strike to the technician has been minimized by establishing a
protective zone over the areas to be traversed. This zone is created by the tower and air terminals
mounted atop light poles.
Step potentials are minimized through the use of a ground mat buried just below the surface
of the area where the technician is expected to be outside the vehicle. Ground mats are commer-
cially available, fabricated in a 6-in. × 6-in. square pattern using no. 8 AWG bare copper wire.
Each intersection is welded, creating—for all practical purposes—an equipotential area that
short-circuits the step potential gradient in the area above the mat. The mat, as a whole, will rise
and fall in potential as a result of the lightning current discharges; however, there will be very lit-
tle difference in potential between the technician's feet. Mats should be covered with six inches
of crushed stone or pavement.
The threat of dangerous touch potentials is minimized by bonding the following elements to
the ground system:
• Personnel ground mat
• Fence at each side of the gate opening
• Door frame of the transmitter building
• Flexible bonding connection between the swing gate and its terminal post
Such bonding will ensure that the object being touched by the technician is at or near the
same potential as his or her feet.
Bonding both sides of the gate opening to the mat helps to ensure that the technician and both
sides of the gate are at approximately the same potential while the gate is being handled. The
flexible bond between the gate and its support post can be accomplished using a commercially
available kit or by Cadwelding a short length of flexible 2/0 AWG welding cable between the two
elements.
5.5.7 Grounding on Bare Rock

A bare rock mountaintop location provides special challenges to the facility design engineer [9].
There is no soil, thus there are no ground rods. Radials are the only means to develop a ground
system. Install a large number of radials, laid straight, but not too taut. The portions not in con-
tact with the rock are in air and form an inductance that will choke the surge current. Because
rock is not conductive when it is dry, keep the radials short. Only a test measurement will deter-
mine how short the radials should be. A conventional earth-resistance tester will tell only half the
story (besides, ground rods cannot be placed in rock for such a measurement). A dynamic
ground tester offers the only way to obtain the true surge impedance of the system.
5.5.7a Rock-Based Radial Elements

On bare rock, a radial counterpoise will conduct and spread the surge charge over a large area. In
essence, it forms a leaky capacitor with the more conductive earth on or under the mountain [9].
The conductivity of the rock will be poor when dry, but quite good when wet. If the site experi-
g
ences significant rainfall before a lightning flash, protection will be enhanced. The worst case,
however, must be assumed: an early strike under dry conditions.
The surge impedance, measured by a dynamic ground tester, should be 25 Ω or less. This
upper-limit number is chosen so that less stress will be placed on the equipment and its surge
protectors. With an 18 kA strike to a 25 Ω ground system, the entire system will rise 450 kV
above the rest of the world at peak current. This voltage has the potential to jump almost 15.750
in. (0.35 in./10 kV at standard atmospheric conditions of 25° C, 30 in. of mercury and 50 percent
relative humidity).
For nonsoil conditions, tower anchor points should have their own radial systems or be encap-
sulated in concrete. Configure the encapsulation to provide at least 3 in. of concrete on all sides
around the embedded conductor. The length will depend on the size of the embedded conductor.
Rebar should extend as far as possible into the concrete. The dynamic ground impedance mea-
surements of the anchor grounds should each be less than 25 Ω.
The size of the bare conductor for each tower radial (or for an interconnecting wire) will vary,
depending on soil conditions. On rock, a bare no. 1/0 or larger wire is recommended. Flat, solid-
copper strap would be better, but may be blown or ripped if not covered with soil. If some
amount of soil is available, no. 6 cable should be sufficient. Make the interconnecting radial
wires continuous, and bury them as deep as possible; however, the first 6 to 10 in will have the
most benefit. Going below 18 in will not be cost-effective, unless in a dry, sandy soil where the
water table can be reached and ground-rod penetration is shallow. If only a small amount of soil
exists, use it to cover the radials to the extent possible. It is more important to cover radials in the
area near the tower than at greater distances. If, however, soil exists only at the outer distances
and cannot be transported to the inner locations, use the soil to cover the outer portions of the
radials.
If soil is present, install ground rods along the radial lengths. Spacing between ground rods is
affected by the depth that each rod is driven; the shallower the rod, the closer the allowed spac-
ing. Because the ultimate depth a rod can be driven cannot always be predicted by the first rod
driven, use a maximum spacing of 15 ft when selecting a location for each additional rod. Drive
rods at building corners first (within 24 in. but not closer than 6 in. to a concrete footer unless
that footer has an encapsulated Ufer ground), then fill in the space between the corners with
additional rods.
Drive the ground rods in place. Do not auger; set in place, then back-fill. The soil compact-
ness is never as great on augured-hole rods when compared with driven rods. The only exception
is when a hole is augured or blasted for a ground rod or rebar and then back-filled in concrete.
Because concrete contains lime (alkali base) and is porous, it absorbs moisture readily, giving it
up slowly. Electron carriers are almost always present, making the substance a relatively good
conductor.
If a Ufer ground is not being implemented, the radials may be Cadwelded to a subterranean
ring, with the ring interconnected to the tower foot pad via a minimum of three no. 1/0 wires
spaced at 120° angles and Cadwelded to the radial ring.
g
5.5.8 Transmission-System Grounding

Nature can devastate a communications site. Lightning can seriously damage an unprotected
facility with little or no warning, leaving an expensive and time-consuming repair job. The first
line of defense is proper grounding of the communications system.
5.5.8a Transmission Line

All coax and waveguide lines must include grounding kits for bonding the transmission line at
the antenna [9]. On conductive structures, this can be accomplished by bonding the tail of the
grounding kit to the structure itself. Remove all nonconductive paint and corrosion before attach-
ment. Do not drill holes, and do not loosen any existing tower member bolts. Antenna clamping
hardware can be used, or an all-stainless-steel hose clamp of the appropriate size can be substi-
tuted. The location of the tower-top ground is not as critical as the bottom grounding kit.
On nonconductive structures, a no. 1/0 or larger wire must be run down the tower [9]. Bond
the transmission-line grounding kit to this down-run. Keep the wire as far away from all other
conductive runs (aviation lights, coax, and waveguide) as possible. Separation of 2 ft is pre-
ferred; 18 in. is the minimum. If any other ground lines, conduit, or grounded metallic-structure
members must be traversed that are closer than 18 in., they too must be grounded to the down-
lead ground line to prevent flashover.
At the point where the coax or waveguide separates from the conductive structure (metal
tower), a coax or waveguide grounding kit must be installed. Secure the connection to a large
vertical structure member with a small number of joints. Attach to a structural member as low as
possible on the tower.
Dress the grounding kit tails in a nearly straight 45° downward angle to the tower member.
On nonconductive structures, a metal busbar must be used. Ground the bar to one or more of the
vertical downconductors, and as close to ground level as possible.
Coaxial cables, lighting conduit, and other lines on the tower must be secured properly to the
structure. Figure 5.5.25 illustrates several common attachment methods.
5.5.8b Cable Considerations

Ground-strap connections must withstand weathering and maintain low electrical resistance
between the grounded component and earth [9]. Corrosion impairs ground-strap performance.
Braided-wire ground straps should not be used in outside installations. Through capillary action
resembling that of a wick, the braid conveys water, which accelerates corrosion. Eventually,
advancing corrosion erodes the ground-strap cable. Braid also can act as a duct that concentrates
water at the bond point. This water speeds corrosion, which increases the electrical resistance of
the bond. A jacketed, seven-strand copper wire strap (no. 6 to no. 2) is recommended for trans-
mission-line grounding at the tower.
5.5.9 Satellite Antenna Grounding

Most satellite receiving/transmitting antenna piers are encapsulated in concrete [9]. Consider-
ation should be given, therefore, to implementing a Ufer ground for the satellite dish. A 4-in.-
g
Figure 5.5.25 Transmission-line mounting and grounding procedures for a communications site.
g
Figure 5.5.26 Grounding a satellite receiving antenna. (After [9].)
diameter pipe, submerged 4 to 5 ft in an 18-in.-diameter (augered) hole will provide a good start
for a Ufer-based ground system. It should be noted that an augered hole is preferred because dig-
ging and repacking the soil around the pier will create higher ground resistance. In areas of good
soil conductivity (100 Ω / m or less), this basic Ufer system may be adequate for the antenna
ground.
Figure 5.5.26 shows the preferred method: a hybrid Ufer/ground-rod and radial system. A
cable connects the mounting pipe (Ufer ground) to a separate driven ground rod. The cable then
is connected to the facility ground system. In areas of poor soil conductivity, additional ground
rods are driven at increments (2.2 times the rod length) between the satellite dish and the facility
ground system. Run all cables underground for best performance. Make the interconnecting cop-
per wire no. 10 size or larger; bury the wire at least 8 in. below finished grade. Figure 5.5.27
shows the addition of a lightning rod to the satellite dish.
g
Figure 5.5.27 Addition of a lightning rod to a satellite antenna ground system. (After [9].)
5.5.10 References
1. Webster's New Collegiate Dictionary.
2. IEEE Standard 100: Definitions of Electrical and Electronic Terms, IEEE, New York, N.Y.
3. IEEE Standard 142: “Recommended Practice for Grounding Industrial and Commercial
Power Systems,” IEEE, New York, N.Y., 1982.
4. IEEE Standard 1100: “Recommended Practice for Powering and Grounding Sensitive Elec-
tronics Equipment,” IEEE, New York, N.Y., 1992.
5. DeWitt, William E.: “Facility Grounding Practices,” in The Electronics Handbook, Jerry C.
Whitaker (ed.), CRC Press, Boca Raton, Fla., pp. 2218–2228, 1996.
6. NFPA Standard 70: “The National Electrical Code,” National Fire Protection Association,
Quincy, Mass., 1993.
7. Sunde, E. D.: Earth Conduction Effects in Transmission Systems, Van Nostrand Co., New
York, N.Y., 1949.
g
8. Schwarz, S. J.: “Analytical Expression for Resistance of Grounding Systems,” AIEE Trans-
actions, vol. 73, Part III-B, pp. 1011–1016, 1954.
9. Block, Roger: “The Grounds for Lightning and EMP Protection,” PolyPhaser Corporation,
Gardnerville, Nev., 1987.
10. Carpenter, Roy, B.: “Improved Grounding Methods for Broadcasters,” Proceedings, SBE
National Convention, Society of Broadcast Engineers, Indianapolis, IN, 1987.
11. Lobnitz, Edward A.: “Lightning Protection for Tower Structures,” in NAB Engineering
Handbook, 9th ed., Jerry C. Whitaker (ed.), National Association of Broadcasters, Wash-
ington, D.C., 1998.
12. DeDad, John A., (ed.): “Basic Facility Requirements,” in Practical Guide to Power Distri-
bution for Information Technology Equipment, PRIMEDIA Intertec, Overland Park, Kan.,
pp. 24, 1997.
13. Military Handbook 419A: “Grounding, Bonding, and Shielding for Electronic Systems,”
U.S. Government Printing Office, Philadelphia, PA, December 1987.
14. IEEE 142 (Green Book): “Grounding Practices for Electrical Systems,” IEEE, New York,
N.Y.
5.5.11 Bibliography
Benson, K. B., and Jerry C. Whitaker: Television and Audio Handbook for Engineers and Tech-
nicians, McGraw-Hill, New York, N.Y., 1989.
Block, Roger: “How to Ground Guy Anchors and Install Bulkhead Panels,” Mobile Radio Tech-
Davis, Gary, and Ralph Jones: Sound Reinforcement Handbook, Yamaha Music Corporation, Hal
Leonard Publishing, Milwaukee, WI, 1987.
Defense Civil Preparedness Agency: “EMP Protection for AM Radio Stations,” Washington,
D.C., TR-61-C, May 1972.
Fardo, S., and D. Patrick: Electrical Power Systems Technology, Prentice-Hall, Englewood Cliffs,
N.J., 1985.
Hill, Mark: “Computer Power Protection,” Broadcast Engineering, PRIMEDIA Intertec, Over-
land Park, Kan., April 1987.
Lanphere, John: “Establishing a Clean Ground,” Sound & Video Contractor, PRIMEDIA Inter-
tec, Overland Park, Kan., August 1987.
1988.
Little, Richard: “Surge Tolerance: How Does Your Site Rate?” Mobile Radio Technology, Inter-
Midkiff, John: “Choosing the Right Coaxial Cable Hanger,” Mobile Radio Technology, Intertec
g
Mullinack, Howard G.: “Grounding for Safety and Performance,” Broadcast Engineering, PRI-
MEDIA Intertec, Overland Park, Kan., October 1986.
Schneider, John: “Surge Protection and Grounding Methods for AM Broadcast Transmitter
lis, IN, 1987.
Sullivan, Thomas: “How to Ground Coaxial Cable Feedlines,” Mobile Radio Technology, PRI-
MEDIA Intertec, Overland Park, Kan., April 1988.
Fe Springs, Calif.
g
g g
Chapter
5.6
Lightning Effects
5.6.1 Introduction
Natural phenomena of interest to facility managers consist mainly of lightning and related distur-
bances. The lightning effect can be compared to that of a capacitor, as shown in Figure 5.6.1. A
charged cloud above the earth will create an oppositely charged area below it of about the same
size and shape. When the voltage difference is sufficient to break down the dielectric (air), the
two “plates” of the “capacitor” will arc over and neutralize their respective charges. If the dielec-
tric spacing is reduced, as in the case of a conductive steel structure (such as a transmitting
tower), the arc-over will occur at a lower-than-normal potential, and will travel through the con-
ductive structure.
The typical duration of a lightning flash is approximately 0.5 s. A single flash is made up of
various discharge components, among which are typically three or four high-current pulses
called strokes. Each stroke lasts about one 1 ms; the separation between strokes is typically sev-
eral tens of milliseconds. Lightning often appears to flicker because the human eye can just
resolve the individual light pulses that are produced by each stroke.
5.6.2 Sources of Atmospheric Energy

Lightning is one of the more visible effects of atmospheric electricity. Stellar events that
occurred light-years ago spray the earth and its atmosphere with atoms that have been stripped of
most or all of their electrons. In the process of entering the atmosphere, these particles collide
with air molecules, which are knocked apart, creating billions more ion pairs each second. Even
though these ions may exist for only about 100 s, they constantly are being replenished from
deep space. The existence of ions in the atmosphere is the fundamental reason for atmospheric
electricity. The primary sources of this energy are:
• Cosmic rays: charged particles emitted by all radiating bodies in space. Most of these parti-
cles (ions) expend their energy in penetrating the envelope of air surrounding the earth.
Through this process, they create more ions by colliding with air atoms and molecules. One
high-energy particle may create up to a billion pairs of ions, many of which will become
atmospheric electricity.
5-115
g g
Figure 5.6.1 The lightning effect and how it can be compared to a more familiar mechanism, the
capacitor principle. Also shown are the parameters of a typical lightning strike.
• Solar wind: charged particles from the sun that continuously bombard the surface of the earth.
Because about half of the earth’s surface is always exposed to the sun, variations are experi-
enced from day to night. Solar wind particles travel at only 200 to 500 miles per second, com-
pared with cosmic particles that travel at near the speed of light. Because of their slower
speed, solar wind particles have less of an effect on air atoms and molecules.
• Natural radioactive decay: the natural disintegration of radioactive elements. In the process
of radioactive decay, air molecules are ionized near the surface of the earth. One of the results
is radon gas.
• Static electricity: energy generated by the interaction of moving air and the earth.
• Electromagnetic generation: energy generated by the movement of air molecules through the
magnetic field of the earth.
The combined effects of cosmic rays and solar wind account for most atmospheric electrical
energy.
Atmospheric energy is present at all times, even during clear weather conditions. This energy
takes the form of a voltage differential of 300 to 400 kV between the surface of the earth and the
ionosphere. The voltage gradient is nonlinear; near the surface it may be 150 V/m of elevation,
but it diminishes significantly at higher altitudes. Under normal conditions, the earth is negative
with respect to the ionosphere, and ions flow between the two entities. Because there are fewer
free ions near the earth than the ionosphere, the volts/meter value is thought to be greater
because of the lower effective conductivity of the air. This concept is illustrated in Figure 5.6.2.
Thermodynamic activity in a developing storm cloud causes it to become a powerfully
charged cell, usually negatively charged on the bottom and positively charged on the top. (See
Figure 5.6.3.) This voltage difference causes a distortion in the voltage gradient and, in fact, the
g g
Figure 5.6.2 The effects of atmospheric conductivity.
Figure 5.6.3 A developing thunderstorm cell.
polarity inverts, with the earth becoming positive with reference to the bottom of the cloud. This
voltage gradient increases to a high value, sometimes exceeding 10 kV/m of elevation. The over-
all charge between the earth and the cloud may be on the order of 10 to 100 MV, or more. When
sufficient potential difference exists, a lightning flash may occur.
Figure 5.6.4 shows the flash waveform for a typical lightning discharge. The rise time is very
fast, in the microsecond range, as the lightning channel is established. The trailing edge exhibits
a slow decay; the decay curve is known as a reciprocal double exponential waveform. The trail-
g g
Figure 5.6.4 Discharge waveform for a typical lightning strike.
ing edge is the result of the resistance of the ionized channel depleting energy from the cloud.
The path length for a lightning discharge is measured in kilometers. The most common source of
lightning is cumulonimbus cloud forms, although other types of clouds (such as nimbostratus)
occasionally can produce activity.
Although most lightning strikes are negative (the bottom of the cloud is negative with respect
to the earth), positive strikes also can occur. Such strikes have been estimated to carry as much as
10 times the current of a negative strike. A positive flash can carry 200 kiloamps (kA) or more of
discharge current. Such “hot strikes,” as they are called, can cause considerable damage. Hot
strikes can occur in the winter, and are often the after-effect of a particularly active storm. After a
number of discharges, the lower negative portion of the cloud will become depleted. When
charged, the lower portion may have functioned as a screen or shield between the earth and the
upper, positively charged portion of the cloud. When depleted, the shield is removed, exposing
the earth to the massive charge in the upper cloud containing potentials of perhaps 500 MV or
more.
5.6.3 Characteristics of Lightning

A typical lightning flash consists of a stepped leader that progresses toward the ground at a
velocity that can exceed 50 m/μs. When sufficient potential difference between the cloud and the
ground exists, arcs move from the ground to the leader column, completing the ionized column
from cloud to ground. A fast and bright return stroke then moves upward along the leader col-
umn at about one-third the speed of light. Peak currents from such a lightning flash may exceed
g g
Figure 5.6.5 Four general types of lightning activity. (After [1].)
100 kA, with a total charge as high as 100 coulombs (C). Although averages are difficult to
assess where lightning is concerned, a characteristic flash exhibits a 2 μs rise time, and a 10 to 40
μs decay to a 50 percent level. The peak current will average 18 kA for the first impulse, and
about half that for the second and third impulses. Three to four strokes per flash are common.
A lightning flash is a constant-current source. Once ionization occurs, the air becomes a con-
ductive plasma reaching 60,000°F, and becomes luminous. The resistance of an object struck by
lightning is of small consequence except for the power dissipation on that object, which is equiv-
alent to I2R. Fifty percent of all strikes will have a first discharge of at least 18 kA, 10 percent
will exceed 60 kA, and only one percent will exceed 120 kA.
Four specific types of cloud-to-ground lightning have been identified. They are categorized in
terms of the direction of motion (upward or downward) and the sign of the electric charge (posi-
tive or negative) of the initiating leader. The categories, illustrated in Figure 5.6.5, are defined as
follows:
• Category 1: negative leader cloud-to-ground discharge. By far the most common form of
lightning, such discharges account for 90 percent or more of the cloud-to-ground flashes
worldwide. Such events are initiated by a downward-moving negatively charged leader.
g g
Figure 5.6.6 The mechanics of a lightning flash. (After [2].)
• Category 2: positive leader ground-to-cloud discharge. This event begins with an upward-ini-
tiated flash from earth, generally from a mountaintop or tall steel structure. Category 2 dis-
charges are relatively rare.
• Category 3: positive leader cloud-to-ground discharge. Less than 10 percent of
cloud-to-ground lightning worldwide is of this type. Positive discharges are initiated by lead-
ers that do not exhibit the distinct steps of their negative counterparts. The largest recorded
peak currents are in the 200–300 kA range.
• Category 4: negative leader ground-to-cloud discharge. Relatively rare, this form of lightning
begins with an upward leader that exhibits a negative charge. Similar to Category 2 dis-
charges, Category 4 discharges occur primarily from a mountaintop or tall steel structure.
An idealized lightning flash is shown in Figure 5.6.6. The stepped leader initiates the first
return stroke in a negative cloud-to-ground flash by propagating downward in a series of discrete
steps, as shown. The breakdown process sets the stage for a negative charge to be lowered to the
ground. A fully developed leader lowers 10 C or more of negative cloud charge to near the
g g
Figure 5.6.7 The mechanisms of lighting surge on an overhead power-distribution line.
ground within a few tens of milliseconds. The average return leader current measures from 100
A to 1 kA. During its trip toward earth, the stepped leader branches in a downward direction,
producing the characteristic lightning discharge.
The electrical potential difference between the bottom of the negatively charged leader chan-
nel and the earth can exhibit a magnitude in excess of 100 MV. As the leader tip nears ground
level, the electric field at sharp objects on the ground increases until the breakdown strength of
the atmosphere is exceeded. At that point, one or more upward-moving discharges are initiated,
and the attachment process begins. The leader channel is discharged when the first return stroke
propagates up the previously ionized and charged leader path. This process will repeat if suffi-
cient potential exists after the initial stroke. The time between successive strokes in a flash is
usually several tens of milliseconds.
5.6.3a Cloud-to-Cloud Activity

A cloud discharge can be defined as any lightning event that does not connect with the earth.
Most lightning discharges occur within the confines of the cloud. Cloud discharges can be subdi-
vided into intra-cloud, inter-cloud, and cloud-to-air flashes. Although relatively insignificant
g g
insofar as earthbound equipment is concerned,

current movement between clouds can create a
corresponding earth current.
It is estimated that only about 10 to 25 percent
of lightning occurs from cloud to ground; most
discharges consist of intra-cloud activity. The rea-
son is the enormous voltage difference that builds
up between the top and bottom of a storm cloud.
Furthermore, the region between the top and bot-
tom of the cloud can be more highly ionized than
the region between the bottom of the cloud and
the earth. Currents developed by cloud-to-cloud
discharges can induce significant voltages in con-
ductors buried in-line with the charge movement.
The windstorm effect also can induce voltages in
above- or below-ground conductors as a result of
rapid changes in the electrical potential of the
atmosphere.
It is unnecessary, therefore, for atmospheric
charge energy to actually strike a conductor of Figure 5.6.8 Various types static dissipa-
concern, such as a transmitting tower or utility tion arrays.
company pole. In many cases, significant voltage
transients can be generated solely by induction. Cloud-to-cloud charge movements generate hor-
izontally polarized radiation, and cloud-to-ground discharges generate vertically polarized radia-
tion. Field strengths exceeding 70 V/m can be induced in conductors a mile or so from a large
strike.
Figure 5.6.7 illustrates the mechanisms of lightning damage. Traveling waves of voltage and
current follow all conductive paths until the flash energy has been dissipated. Reflections occur
at discontinuities, such as lightning arresters (points 1, 2, 3, and 5) and transformers (points 4
and 6).
5.6.3b Lightning Protection

Research into the physical properties of lightning and related phenomena has two basic goals: 1)
to identify the character and severity of the threat, and 2) to devise methods to prevent damage
resulting from atmospheric activity. Many different approaches have been taken over the years
for controlling the damaging potential of lightning; some have become widely accepted, others
remain controversial. The issue of lightning prevention clearly falls into the second category.
Application of the point discharge theory, the basis of lightning prevention schemes, is con-
troversial to begin with. Still, it offers the promise of a solution to a serious problem faced by
nearly all telecommunications operators. The goal is to dissipate static charges around a given
structure at a rate sufficient to maintain the charge below the value at which a lightning flash will
occur. The theory holds that discharge from the point of an electrode to a surrounding medium
will follow predictable rules of behavior. The sharper the point, the greater the discharge. The
greater the number of discharge points, the more efficient the dissipation system. Several static
g g
Figure 5.6.9 Umbrella-type dissipation array.
dissipators based on this theory are shown in Figure 5.6.8. Key design elements for such dissipa-
tors include:
• Radius of the dissipator electrode. The purpose of the dissipator is to create a high field inten-
sity surrounding the device. Theory states that the electric field intensity will increase as the
electrode radius is reduced. Dissipators, therefore, generally use the smallest-radius elec-
trodes possible, consistent with structural integrity. There is, however, disagreement among
certain dissipation-array manufacturers on this point. The “optimum wire size,” according to
available literature, varies from 0.005-in.- to 1/8-in.-thick tapered spikes.
• Dissipator construction material. Important qualities include conductivity and durability. The
dissipator should be a good conductor to provide: 1) the maximum discharge of current dur-
ing normal operation, and 2) an efficient path for current flow in the event of a direct strike.
• Number of dissipator electrodes. Calculating the number of dissipator points is, again, the
subject of some debate. However, because the goal of the system is to provide a low-resis-
tance path to the atmosphere, it generally is assumed that the more discharge points, the more
g g
Figure 5.6.10 Measured discharge current of a lightning dissipation array.
effective the system. Dissipator electrode requirements are determined by the type of struc-
ture being protected as well as the environmental features surrounding it.
• Density of dissipator electrodes. Experimentation by some manufacturers has shown that the
smaller the radius of the dissipator electrodes, the more closely they can be arranged without
reducing the overall efficiency of the dissipator. Although this convention seems reasonable,
disagreement exists among dissipation-array manufacturers. Some say the points should be
close together; others say they should be far apart.
• Configuration of the dissipator on the tower. Disagreement abounds on this point. One school
of thought supports the concept of a dedicated “umbrella-type” structure at the top of the
tower as the most efficient method of protecting against a lightning flash (Figure 5.6.9).
Another view holds that the dissipator need not be at the highest point, and that it may be
more effective if one or more dissipators are placed at natural dissipation points on the struc-
ture. Such points include side-mounted antennas and other sharp elements on the tower.
• Size and deployment of grounding electrodes. Some systems utilize an extensive ground sys-
tem, others do not. One manufacturer specifies a “collector” composed of wire radials
extending from the base of the tower and terminated by ground rods. Another manufacturer
does not require a ground connection to the dissipator.
Available literature indicates that from 10 μA to 10 mA flow through a properly designed dis-
sipative system into the surrounding air during a lightning storm. Figure 5.6.10 charts the dis-
charge current recorded during a period of storm activity at a protected site. Although a lightning
stroke can reach several hundreds of thousands of amperes, this energy flows for a very short
g g
Figure 5.6.11 Intended protection area for a dissipation array.
period of time. The concept of the dissipative array is to continuously bleed off space charge cur-
rent to prevent a direct hit.
Proof that static dissipators work as intended is elusive and depends upon the definition of
“proof.” Empirical proof is difficult to obtain because successful performance of a static dissipa-
tor is evidenced by the absence of any results. Supporting evidence, both pro and con, is avail-
able from end-users of static dissipators, and from those who have studied this method of
reducing the incidence of lightning strikes to a structure.
Protection Area
The placement of a tall structure over low-profile structures tends to protect the facilities near the
ground from lightning flashes. The tall structure, typically a communications tower, is assumed
to shield the facility below it from hits. This cone of protection is determined by the following:
• Height of the tall structure
• Height of the storm cloud above the earth
g g
Figure 5.6.12 Protection zone for a tall tower under the “rolling sphere” theory.
The higher the cloud, the larger the radius of the base of the protecting cone. The ratio of radius
to base to height varies approximately from one to two, as illustrated in Figure 5.6.11.
Conventional wisdom has held that a tower, whether protected with a static dissipation array
or simply a tower-top lightning rod, provided a cone of protection stretching out on all sides of
the structure at an angle of about 45°. Although this theory held favor for many years, modifica-
tions have been proposed. One school of thought suggests that a smaller cone of perhaps 30°
from the tower is more realistic. Another suggests that the cone theory is basically flawed and,
instead, proposes a “rolling sphere” approach. This theory states that areas enclosed below a
150-ft rolling sphere will enjoy protection against lightning strikes. The concept is illustrated in
Figure 5.6.12. Note that the top of the tower shown in the figure is assumed to experience limited
protection. The concept, as it applies to side-mounted antennas, is shown in Figure 5.6.13. The
antenna is protected through the addition of two horizontally mounted lightning rods, one above
the antenna and one below.
5.6.4 Electrostatic Discharge

A static charge is the result of an excess or deficiency of electrons on a given surface. The rela-
tive level of electron imbalance determines the static charge. Simply stated, a charge is generated
by physical contact between, and then separation of, two materials. One surface loses electrons to
the other. The types of materials involved, and the speed and duration of motion between the
materials, determine the charge level. Electrostatic energy is a stationary charge phenomenon
that can build up in either a nonconductive material or in an ungrounded conductive material.
The charge can occur in one of two ways:
g g
• Polarization: charge buildup when a conductive

material is exposed to a magnetic field.
• Triboelectric effects: charge buildup that occurs
when two surfaces contact and then separate,
leaving one positively charged and the other neg-
atively charged.
Friction between two materials increases the tri-
boelectric charge by increasing the surface area that
experiences contact. For example, a person accumu-
lates charge by walking across a nylon carpet; dis-
charge occurs when the person touches a conductive
surface.
5.6.4a Triboelectric Effect

Different materials have differing potentials for
charge. Nylon, human and animal hair, wool, and
asbestos have high positive triboelectric potential.
Silicon, polyurethane, rayon, and polyester have
negative triboelectric potentials. Cotton, steel,
paper, and wood all tend to be relatively neutral
materials. The intensity of the triboelectric charge is
inversely proportional to the relative humidity (RH).
As humidity increases, electrostatic discharge
(ESD) problems decrease. For example, a person
walking across a carpet can generate a 1.5 kV
charge at 90 percent RH, but will generate as much
as 35 kV at 10 percent RH.
Figure 5.6.13 Protection zone for a side-
When a charged object comes in contact with
mounted antenna under the “rolling
another object, the electrostatic charge will attempt sphere” theory.
to find a path to ground, discharging into the con-
tacted object. The current level is very low (typi-
cally less than 0.1 nA), but the voltage can be high (25 to 50 kV).
ESD also can collect on metallic furnishings, such as chairs and equipment racks. Sharp cor-
ners and edges, however, encourage a corona that tends to bleed the charge off such objects. The
maximum voltage normally expected for furniture-related ESD is about 6 to 8 kV. Because
metallic furniture is much more conductive than humans, however, furniture-related ESD gener-
ally will result in higher peak discharge currents. Figure 5.6.14 shows a discharge waveform of a
typical ESD event.
5.6.5 EMP Radiation

Electromagnetic pulse (EMP) radiation is the result of an intense release of electromagnetic
waves that follows a nuclear explosion. (See Figure 5.6.15.) The amount of damaging energy is a
g g
function of the altitude of detonation and the

size of the device. A low-altitude or surface
burst would generate a strong EMP covering
a few thousand square kilometers. However,
the effects of the radiation would be mean-
ingless, because the blast would destroy most
structures in the area. A high-altitude burst,
on the other hand, presents a real threat to all
types of communications and electronic sys-
tems. Such an explosion would generate an
EMP with a radius of more than 1000 km, a
large portion of the United States.
The sudden release of gamma rays in a
nuclear explosion would cause almost instant Figure 5.6.14 Discharge waveform for an ESD
ionization (the removal of electrons from event. (After [3].)
atoms) of the atmospheric gases that sur-
round the detonation area. Free electrons are
driven outward. In a high-altitude event, the gamma rays travel great distances before ionizing
the upper atmosphere. The forced movement of these electrons, which will again recombine with
atoms in the atmosphere, creates a pulsed electromagnetic field.
Figure 5.6.15 The EMP effect and how it can induce damaging voltages onto utility company lines
and antenna structures. The expected parameters of an EMP event also are shown.
g g
Figure 2.17 Discharge waveform for a nuclear EMP event.
Table 5.6.1 Effects of an EMP Event on Various Systems
Type of Conductor Rise Time (s) Peak Voltage (V) Peak Current (A)
Long unshielded cable (power line, long 10–8 to 10–7 105 to 5 × 106 103 to 104
antenna)
HF antenna system 10–8 to 10–7 104 to 106 500 to 104
VHF antenna system 10–9
to 10–8 103 to 105 100 to 103
UHF antenna system 10–9
to 10 –8
100 to 10 4 10 to 100
Shielded cable 10–6 to 10–4 1 to 100 0.1 to 50
The amplitude and polarization of the field produced by a high-altitude detonation depends
on the altitude of the burst, the yield of the device, and the orientation of the burst with respect to
the receiving point. The EMP field creates a short but intense broadband radio frequency pulse
with significant energy up to 100 MHz. Most of the radiated energy, however, is concentrated
below 10 MHz. Figure 5.6.16 shows the distribution of energy as a function of frequency. The
electric field can be greater than 50 kV/m, with a rise time measured in the tens of nanoseconds.
Figure 5.6.17 illustrates the field of a simulated EMP discharge.
Many times, lightning and other natural occurrences cause problems not because they strike a
given site, but because they strike part of the utility power system and are brought into the facil-
ity via the ac lines. Likewise, damage that could result from EMP radiation would be most severe
to equipment connected to the primary power source, because it is generally the most exposed
part of any facility. Table 5.6.1 lists the response of various systems to an EMP event.
5.6.6 Coupling Transient Energy

The utility power-distribution system can couple transient overvoltages into a customer’s load
through induction or direct-charge injection. As stated previously, a lightning flash a mile away
from a 12 kV line can create an electromagnetic field with a strength of 70 V/m or more. Given a
g g
sufficiently long line, substantial voltages can be coupled to the primary power system without a
direct hit. Likewise, the field created by EMP radiation can be coupled to the primary power
lines, but in this case at a much higher voltage (50 kV/m). Considering the layout of many parts
of the utility company power system—long exposed lines over mountaintops and the like—the
possibility of a direct lightning flash to one or more legs of the system is a distinct one.
Lightning is a point charge-injection process, with pulses moving away from the point of
injection. The amount of total energy (voltage and current) and the rise and decay times of the
energy seen at the load as a result of a lightning flash are functions of the distance between the
flash and the load and the physical characteristics of the power distribution system. Determining
factors include:
• Wire size
• Number and sharpness of bends
• Types of transformers
• Types of insulators
• Placement of lightning suppressors
The character of a lightning flash covers a wide range of voltage, current, and rise-time
parameters. Making an accurate estimate of the damage potential of a flash is difficult. A direct
hit to a utility power line causes a high-voltage, high-current wave to travel away from the point
of the hit in both directions along the power line. The waveshape is sawtooth in form, with a rise
time measured in microseconds or nanoseconds. The pulse travels at nearly the speed of light
until it encounters a significant change in line impedance. At this point, a portion of the wave is
reflected back down the line in the direction from which it came. This action creates a standing
wave containing the combined voltages of the two pulses. A high-energy wave of this type can
reach a potential sufficient to arc over to another parallel line, a distance of about 8 ft on a local
feeder (typically 12 kV) power pole.
5.6.7 Bibliography
Bishop, Don, “Lightning Devices Undergo Tests at Florida Airports,” Mobile Radio Technology,
PRIMEDIA Intertec, Overland Park, Kan., May 1990.
Bishop, Don, “Lightning Sparks Debate: Prevention or Protection?,” Mobile Radio Technology,
PRIMEDIA Intertec, Overland Park, Kan., January 1989.
Block, Roger, “Dissipation Arrays: Do They Work?,” Mobile Radio Technology, PRIMEDIA
Intertec, Overland Park, Kan., April 1988.
Block, Roger, “The Grounds for Lightning and EMP Protection,” PolyPhaser Corp., Gardnerv-
ille, Nev., 1987.
Defense Civil Preparedness Agency, EMP and Electric Power Systems, Publication TR-6l-D,
U.S. Department of Commerce, National Bureau of Standards, Washington, D.C., July
1973.
Defense Civil Preparedness Agency, EMP Protection for Emergency Operating Centers, Federal
Information Processing Standards Publication no. 94, Guideline on Electrical Power for
g g

ington, D.C., 1983.
Drabkin, Mark, and Roy Carpenter, Jr., “Lightning Protection Devices: How Do They Com-
pare?,” Mobile Radio Technology, PRIMEDIA Intertec, Overland Park, Kan., October
1988.
Fink, D., and D. Christiansen, Electronics Engineers’ Handbook, 3rd ed., McGraw-Hill, New
York, N.Y., 1989.
Jay, Frank, ed., IEEE Standard Directory of Electrical and Electronics Terms, 3rd ed., IEEE, New
York, N.Y., 1984.
Jordan, Edward C. (ed.), Reference Data for Engineers: Radio, Electronics, Computer, and Com-
munications, 7th ed., Howard W. Sams, Indianapolis, Ind., 1985.
Kaiser, Bruce A., “Can You Really Fool Mother Nature?,” Cellular Business, PRIMEDIA Inter-
Kaiser, Bruce A., “Straight Talk on Static Dissipation,” Proceedings of ENTELEC 1988, Energy
Telecommunications and Electrical Association, Dallas, Tex., 1988.
Key, Lt. Thomas, “The Effects of Power Disturbances on Computer Operation,” IEEE Industrial
and Commercial Power Systems Conference paper, Cincinnati, Ohio, June 7, 1978.
Martzloff, F. D., “The Development of a Guide on Surge Voltages in Low-Voltage AC Power Cir-
cuits,” 14th Electrical/Electronics Insulation Conference, IEEE, Boston, Mass., October
1979.
Nott, Ron, “The Sources of Atmospheric Energy,” Proceedings of the SBE National Convention
and Broadcast Engineering Conference, Society of Broadcast Engineers, Indianapolis, IN,
1987.
Ulman, Martin A., The Lightning Discharge, Academic Press, Orlando, Fla., 1987.
Whitaker, Jerry C., Maintaining Electronic Systems, CRC Press, Boca Raton, Fla., 1991.
Whitaker, Jerry C., Radio Frequency Transmission Systems: Design and Operation,
g g
g g
Chapter
5.7
Transmitter Building Grounding
Practices
5.7.1 Introduction
Proper grounding is basic to protection against ac line disturbances. This applies whether the
source of the disturbance is lightning, power-system switching activities, or faults in the distribu-
tion network. Proper grounding is also a key element in preventing radio frequency interference
in transmission or computer equipment. A facility with a poor ground system can experience
RFI problems on a regular basis. Implementing an effective ground network is not an easy task.
It requires planning, quality components, and skilled installers. It is not inexpensive. However,
proper grounding is an investment that will pay dividends for the life of the facility.
5.7.2 Designing a Building Ground System

After the required grounding elements have been determined, they must be connected into a uni-
fied system [1]. Many different approaches can be taken, but the goal is the same: establish a
low-resistance, low-inductance path to surge energy. Figure 5.7.1 shows a building ground sys-
tem using a combination of ground rods and buried bare-copper radial wires. This design is
appropriate when the building is large or located in an urban area.
Figure 5.7.2 shows another approach in which a perimeter ground strap is buried around the
building and ground rods are driven into the earth at regular intervals (2.2 times the rod length).
The ground ring consists of a one-piece copper conductor that is bonded to each ground rod.
If a transmission or microwave tower is located at the site, connect the tower ground system to
the main ground point via a copper strap. The width of the strap must be at least 1 percent of the
length and, in any event, not less than 3 in. wide. The building ground system is not a substitute
for a tower ground system, no matter what the size of the tower. The two systems are treated as
independent elements, except for the point at which they interconnect.
Connect the utility company power-system ground rod to the main facility ground point as
required by the local electrical code. Do not consider the building ground system to be a substi-
tute for the utility company ground rod. The utility rod is important for safety reasons and must
5-133
g g
Figure 5.7.1 A facility ground system using the hub-and-spoke approach. The available real
estate at the site will dictate the exact configuration of the ground system. If a tower is located at
the site, the tower ground system is connected to the building ground as shown.
Figure 5.7.2 Facility ground using a perimeter ground-rod system. This approach works well for
buildings with limited available real estate.
g g
Figure 5.7.3 A typical guy-anchor and tower-radial grounding scheme. The radial ground is no. 6
copper wire. The ground rods are 5/8 in. × 10 ft. (After [1].)
not be disconnected or moved. Do not remove any existing earth ground connections to the
power-line neutral connection. To do so may violate local electrical code.
Bury all elements of the ground system to reduce the inductance of the overall network. Do
not make sharp turns or bends in the interconnecting wires. Straight, direct wiring practices will
reduce the overall inductance of the system and increase its effectiveness in shunting fast-rise-
time surges to earth. Figure 5.7.3 illustrates the interconnection of a tower and building ground
system. In most areas, soil conductivity is high enough to permit rods to be connected with no. 6
bare-copper wire or larger. In areas of sandy soil, use copper strap. A wire buried in low-conduc-
tivity, sandy soil tends to be inductive and less effective in dealing with fast-rise-time current
surges. Make the width of the ground strap at least 1 percent of its overall length. Connect buried
elements of the system as shown in Figure 5.7.4.
For small installations with a low physical profile, a simplified grounding system can be
implemented, as shown in Figure 5.7.5. A grounding plate is buried below grade level, and a
ground wire ties the plate to the microwave tower mounted on the building.
g g
Figure 5.7.4 Preferred bonding method for below-grade elements of the ground system. (After
[1].)
5.7.2a Bulkhead Panel

The bulkhead panel is the cornerstone of an effective facility grounding system [1]. The concept
of the bulkhead is simple: establish one reference point to which all cables entering and leaving
the equipment building are grounded and to which all transient-suppression devices are
mounted. Figure 5.7.6 shows a typical bulkhead installation for a broadcast or communications
facility. The panel size depends on the spacing, number, and dimensions of the coaxial lines,
power cables, and other conduit entering or leaving the building.
To provide a weatherproof point for mounting transient-suppression devices, the bulkhead
can be modified to accept a subpanel, as shown in Figure 5.7.7. The subpanel is attached so that
it protrudes through an opening in the wall and creates a secondary plate on which transient sup-
pressors are mounted and grounded. A typical cable/suppressor-mounting arrangement for a
communications site is shown in Figure 5.7.8. To handle the currents that may be experienced
during a lightning strike or large transient on the utility company ac line, the bottom-most sub-
panel flange (which joins the subpanel to the main bulkhead) must have a total surface-contact
area of at least 0.75 in.2 per transient suppressor.
Because the bulkhead panel will carry significant current during a lightning strike or ac line
disturbance, it must be constructed of heavy material. The recommended material is 1/8-in C110
(solid copper) 1/2-hard. This type of copper stock weighs nearly 5-1/2 lb/ft 2 and is rather expen-
sive. Installing a bulkhead, while sometimes difficult and costly, will pay dividends for the life of
the facility. Use 18-8 stainless-steel mounting hardware to secure the subpanel to the bulkhead.
Because the bulkhead panel establishes the central grounding point for all equipment within
the building, it must be tied to a low-resistance (and low-inductance) perimeter ground system.
The bulkhead establishes the main facility ground point, from which all grounds inside the build-
g g
Figure 5.7.5 Grounding a small microwave transmission tower.
ing are referenced. A typical bulkhead installation for a small communications site is shown in
Figure 5.7.9.
Bulkhead Grounding
A properly installed bulkhead panel will exhibit lower impedance and resistance to ground than
any other equipment or cable grounding point at the facility [1]. Waveguide and coax line
grounding kits should be installed at the bulkhead panel as well as at the tower. Dress the kit tails
downward at a straight 45° angle using 3/8-in. stainless-steel hardware to the panel. Position the
g g
Figure 5.7.6 The basic design of a bulkhead panel for a facility. The bulkhead establishes the
grounding reference point for the plant.
Figure 5.7.7 The addition of a subpanel to a bulkhead as a means of providing a mounting surface
for transient-suppression components. To ensure that the bulkhead is capable of handling high
surge currents, use the hardware shown. (After [1].)
g g
Figure 5.7.8 Mounting-hole layout for a communications site bulkhead subpanel.
Figure 5.7.9 Bulkhead installation at a small communications site. (After [1].)

g g
Figure 5.7.10 The proper way to ground a bulkhead panel and provide a low-inductance path for
surge currents stripped from cables entering and leaving the facility. The panel extends along the
building exterior to below grade. It is silver-soldered to a no. 2/0 copper wire that interconnects
with the outside ground system. (After [1].)
stainless-steel lug at the tail end, flat against a cleaned spot on the panel. Joint compound will be
needed for aluminum and is recommended for copper panels.
Because the bulkhead panel will be used as the central grounding point for all the equipment
inside the building, the lower the inductance to the perimeter ground system, the better. The best
arrangement is to simply extend the bulkhead panel down the outside of the building, below
grade, to the perimeter ground system. This will give the lowest resistance and the smallest
inductive voltage drop. This approach is illustrated in Figure 5.7.10.
If cables are used to ground the bulkhead panel, secure the interconnection to the outside
ground system along the bottom section of the panel. Use multiple no. 1/0 or larger copper wire
or several solid-copper straps. If using strap, attach with stainless-steel hardware, and apply joint
compound for aluminum bulkhead panels. Clamp and Cadweld, or silver-solder for copper/brass
panels. If no. 1/0 or larger wire is used, employ crimp lug and stainless-steel hardware. Measure
the dc resistance. It should be less than 0.01 Ω between the ground system and the panel. Repeat
this measurement on an annual basis.
If the antenna feed lines do not enter the equipment building via a bulkhead panel, treat them
in the following manner:
g g
• Mount a feed-line ground bar on the wall of the building approximately 4 in. below the feed-
line entry point.
• Connect the outer conductor of each feed line to the feed-line ground bar using an appropriate
grounding kit.
• Connect a no. 1/0 cable or 3- to 6-in.-wide copper strap between the feed-line ground bar and
the external ground system. Make the joint a Cadweld or silver-solder connection.
• Mount coaxial arrestors on the edge of the bar.
• Weatherproof all connections.
5.7.2b Lightning Protectors

A variety of lightning arrestors are available for use on coaxial transmission lines, utility ac
feeds, and telephone cables. The protector chosen must be carefully matched to the requirements
of the application. Do not use air gap protectors because these types are susceptible to air pollu-
tion, corrosion, temperature, humidity, and manufacturing tolerances. The turn-on speed of an
air gap device is a function of all of the foregoing elements. A simple gas-tube-type arrestor is an
improvement, but neither of these devices will operate reliably to protect shunt-fed cavities, iso-
lators, or receivers that include static drain inductors to ground (which most have). Such voltage-
sensitive crowbar devices are short-circuited by the dc path to ground found in these circuits. The
inductive change in current per unit time (Δdi/dt) voltage drop may not be enough to fire the pro-
tector, but it can often be sufficient to destroy the inductor and then the receiver front end.
Instead, select a protector that does not have dc continuity on the coaxial center pin from input to
output. Such units have a series capacitor between the gas tube and the equipment center pin that
will allow the voltage to build up so that the arrestor can fire properly.
Typical Installation
Figure 5.7.11 illustrates a common grounding arrangement for a remotely located grounded-
tower (FM, TV, or microwave radio) transmitter plant. The tower and guy wires are grounded
using 10-ft-long copper-clad ground rods. The antenna is bonded to the tower, and the transmis-
sion line is bonded to the tower at the point where it leaves the structure and begins the horizontal
run into the transmitter building. Before entering the structure, the line is bonded to a ground rod
through a connecting cable. The transmitter itself is grounded to the transmission line and to the
ac power-distribution system ground. This, in turn, is bonded to a ground rod where the utility
feed enters the building. The goal of this arrangement is to strip all incoming lines of damaging
overvoltages before they enter the facility. One or more lightning rods are mounted at the top of
the tower structure. The rods extend at least 10 ft above the highest part of the antenna assembly.
Such a grounding configuration, however, has built-in problems that can make it impossible
to provide adequate transient protection to equipment at the site. Look again at the example. To
equipment inside the transmitter building, two grounds actually exist: the utility company ground
and the antenna ground. One ground will have a lower resistance to earth, and one will have a
lower inductance in the connecting cable or copper strap from the equipment to the ground sys-
tem.
Using the Figure 5.7.11 example, assume that a transient overvoltage enters the utility com-
pany meter panel from the ac service line. The overvoltage is clamped by a protection device at
g g
Figure 5.7.11 A common, but not ideal, grounding arrangement for a transmission facility using a
grounded tower. A better configuration involves the use of a bulkhead panel through which all
cables pass into and out of the equipment building.
the meter panel, and the current surge is directed to ground. But which ground, the utility ground
or the antenna ground?
The utility ground surely will have a lower inductance to the current surge than the antenna
ground, but the antenna probably will exhibit a lower resistance to ground than the utility side of
the circuit. Therefore, the surge current will be divided between the two grounds, placing the
transmission equipment in series with the surge suppressor and the antenna ground system. A
transient of sufficient potential will damage the transmission equipment.
Transients generated on the antenna side because of a lightning discharge are no less trouble-
some. The tower is a conductor, and any conductor is also an inductor. A typical 150-ft self-sup-
porting tower may exhibit as much as 40 μH inductance. During a fast-rise-time lightning strike,
an instantaneous voltage drop of 360 kV between the top of the tower and the base is not
unlikely. If the coax shield is bonded to the tower 15 ft above the earth (as shown in the previous
figure), 10 percent of the tower voltage drop (36 kV) will exist at that point during a flash. Figure
5.7.12 illustrates the mechanisms involved.
The only way to ensure that damaging voltages are stripped off all incoming cables (coax, ac
power, and telephone lines) is to install a bulkhead entrance panel and tie all transient-suppres-
sion hardware to it. Configuring the system as shown in Figure 5.7.13 strips away all transient
voltages through the use of a single-point ground. The bulkhead panel is the ground reference for
g g
Figure 5.7.12 The equivalent circuit of the facility shown in Figure 5.7.11. Note the discharge cur-
rent path through the electronic equipment.
Figure 5.7.13 The preferred grounding arrangement for a transmission facility using a bulkhead
panel. With this configuration, all damaging transient overvoltages are stripped off the coax,
power, and telephone lines before they can enter the equipment building.
g g
Figure 5.7.14 The equivalent circuit of the facility shown in Figure 5.7.13. Discharge currents are
prevented from entering the equipment building.
the facility. With such a design, secondary surge current paths do not exist, as illustrated in Fig-
ure 5.7.14.
Protecting the building itself is another important element of lightning surge protection. Fig-
ure 5.7.15 shows a common system using multiple lightning rods. As specified by NFPA 78, the
dimensions given in the figure are as follows:
• A = 50-ft maximum spacing between air terminals
• B = 150-ft maximum length of a coursing conductor permitted without connection to a main
perimeter or downlead conductor
• C = 20- or 25-ft maximum spacing between air terminals along an edge
5.7.2c Checklist for Proper Grounding

A methodical approach is necessary in the design of a facility ground system. Consider the fol-
lowing points:
1. Install a bulkhead panel to provide mechanical support, electric grounding, and lightning
protection for coaxial cables, power feeds, and telephone lines entering the equipment
building.
2. Install an internal ground bus using no. 2 or larger solid-copper wire. (At transmission
facilities, use copper strap that is at least 3 in wide.) Form a star grounding system. At
larger installations, form a star-of-stars configuration. Do not allow ground loops to exist
in the internal ground bus. Connect the following items to the building internal ground sys-
tem:
g g
Figure 5.7.15 Conduction lightning-protection system for a large building. (From [12]. Used with
permission.)
• Chassis racks and cabinets of all hardware

• All auxiliary equipment
• Battery charger
• Switchboard
• Conduit
• Metal raceway and cable tray
3. Install a tower earth ground array by driving ground rods and laying radials as required to
achieve a low earth ground impedance at the site.
4. Connect outside metal structures to the earth ground array (towers, metal fences, metal
buildings, and guy-anchor points).
5. Connect the power-line ground to the array. Follow local electrical code to the letter.
6. Connect the bulkhead to the ground array through a low-inductance, low-resistance bond.
7. Do not use soldered-only connections outside the equipment building. Crimped, brazed,
and exothermic (Cadwelded) connections are preferable. For a proper bond, all metal sur-
g g
faces must be cleaned, any finish removed to bare metal, and surface preparation com-
pound applied (where necessary). Protect all connections from moisture by appropriate
means (sealing compound and heat sink tubing).
5.7.3 References
1. Block, Roger, “The Grounds for Lightning and EMP Protection,” PolyPhaser Corporation,
Gardnerville, Nev., 1987.
5.7.4 Bibliography
Benson, K. B., and Jerry C. Whitaker, Television and Audio Handbook for Engineers and Techni-
cians, McGraw-Hill, New York, N.Y., 989.
Block, Roger, “How to Ground Guy Anchors and Install Bulkhead Panels,” Mobile Radio Tech-
Fardo, S., and D. Patrick, Electrical Power Systems Technology, Prentice-Hall, Englewood Cliffs,
N.J., 1985.
Little, Richard, “Surge Tolerance: How Does Your Site Rate?” Mobile Radio Technology, Inter-
Schneider, John, “Surge Protection and Grounding Methods for AM Broadcast Transmitter
lis, Ind., 1987.
Fe Springs, Calif.
Whitaker, Jerry C., AC Power Systems Handbook, 2nd ed., CRC Press, Boca Raton, Fla., 1999.
g g
Section
Radio Receivers
6
The development of radio transmission and reception is one of the major technical achievements
of the twentieth century. The impact of voice broadcasts to the public, whether by commercial
stations or government-run organizations, have expanded the horizons of everyday citizens in
virtually every country on earth. It is hard to overestimate the power of radio broadcasting.
Technology has dramatically reshaped the transmission side of AM and FM broadcasting.
Profound changes have also occurred in receiver technology. Up until 1960, radio broadcasting
was basically a stationary medium. The receivers of that time were physically large and heavy,
and required 120 V ac power to drive them. The so-called portable radios of the day relied on
bulky batteries that offered only a limited amount of listening time. Automobile radios incorpo-
rated vibrator-choppers to simulate ac current. All the receivers available for commercial use
during the 1940s and 1950s used vacuum tubes exclusively.
The first technical breakthrough for radio broadcasting was the transistor, available commer-
cially at reasonable prices during the early 1960s. The transistor brought with it greatly reduced
physical size and weight, and even more importantly, it eliminated the necessity of ac line current
to power the radio. The first truly portable AM radios began to appear during the early 1960s,
with AM-FM versions following by the middle of the decade.
Many of the early receiver designs were marginal from a performance stand-point. The really
good receivers were still built around vacuum tubes. As designers learned more about transistors,
and as better transistors became available, tube-based receivers began to disappear. By 1970,
transistorized radios, as they were called, commanded the consumer market.
The integrated circuit (IC) was the second technical breakthrough in consumer receiver
design. This advance, more than anything else, made high-quality portable radios possible. It
also accelerated the change in listening habits from AM to FM. IC-based receivers allowed
designers to put more sophisticated circuitry into a smaller package, permitting consumers to
enjoy the benefits of FM broadcasting without the penalties of the more complicated receiver
required for FM stereo.
The move to smaller, lighter, more power-efficient radios have led to fundamental changes in
the way radios are built and serviced. In the days of vacuum-tube and transistor-based receivers,
the designer would build a radio out of individual stages that were interconnected to provide a
working unit. The stages for a typical radio included:
• RF amplifier
6-1
6-2 Section Six
• Local oscillator
• Intermediate frequency (IF) amplifier
• Detector and audio preamplifier
Today, however, large-scale integration (LSI) or even very large scale integration (VLSI)
techniques have permitted virtually all the active circuits of an AM-FM radio to be placed on a
single IC. Advanced circuitry has also permitted radio designers to incorporate all-electronic
tuning, eliminating troublesome and sometimes expensive mechanical components. Electroni-
cally tuned radios (ETRs) have made features such as “station scan” and “station seek” possible.
Some attempts were made to incorporate scan and seek features in mechanically tuned radios.
but the results were never very satisfactory.
The results of LSI-based receiver design have been twofold. First, radios based on advanced
chip technologies are much easier to build and are, therefore, usually less expensive to consum-
ers. Second, such radios are not serviceable. Most consumers today would not bother to have a
broken radio repaired. They would simply buy a new one and throw the old one away.
Still, however, it is important to know what makes a radio work. Although radios being built
with LSI and VLSI technology do not lend themselves to stage-by-stage troubleshooting as ear-
lier radios did, it is important to understand how each part of the system functions to make a
working unit. Regardless of the sophistication of a VLSI-based receiver, the basic principles of
operation are the same as a radio built of discrete stages.
In This Section:
Chapter 6.1: Receiver Characteristics 6-5

Introduction 6-5
Practical Receivers 6-6
The Receiving System 6-7
Gain, Sensitivity, and Noise Figure 6-8
NF 6-8
Minimum Detectable Signal 6-10
Selectivity 6-11
Dynamic Range 6-12
Desensitization 6-14
AM Cross Modulation 6-14
IM 6-15
Error Vector Magnitude 6-18
Gain Control 6-18
Digital Receiver Characteristics 6-22
Bit Error Rate 6-23
Bibliography 6-24
Chapter 6.2: The Radio Channel 6-27

Introduction 6-27
Radio Channel Characteristics 6-28
Channel Impulse Response 6-28
Doppler Effect 6-34
Radio Receivers 6-3
Transfer Function 6-35

Time Response of Channel Impulse Response and Transfer Function 6-36
References 6-38
Chapter 6.3: AM and FM Receivers 6-39

Introduction 6-39
Superheterodyne Receiver 6-39
Radio Wave Propagation 6-41
AM Band Propagation 6-41
FM Band Propagation 6-41
Radio Receivers 6-42
Antenna Systems 6-43
Antenna Coupling Network 6-43
Whip Antenna 6-43
Loop Antenna 6-44
Filter Types 6-45
LC Filter 6-46
Quartz Crystal Filter 6-46
Monolithic Quartz Filter 6-46
Ceramic Filter 6-47
RF Amplifier and AGC 6-47
The AGC Loop 6-48
Mixer 6-49
Passive Mixer 6-50
Active Mixer 6-51
Local Oscillator 6-52
PLL Synthesizer 6-53
Frequency Divider 6-54
Variable-Frequency Oscillator 6-54
Diode Switching 6-55
Crystal-Controlled Oscillator 6-55
AM-FM Demodulation 6-57
AM Demodulation 6-60
FM Demodulation 6-62
Amplitude Limiter 6-64
Stereo Systems 6-65
FM Stereo 6-66
Generating the Stereo Signal 6-66
Decoding the Stereo Signal 6-67
AM Stereo 6-68
Decoding the C-QUAM Signal 6-70
References 6-71
Bibliography 6-71

Amos, S. W.: “FM Detectors,” Wireless World, vol. 87, no. 1540, pg. 77, January 1981.
6-4 Section Six
Benson, K. Blair, and Jerry C. Whitaker: Television and Audio Handbook for Engineers and
Technicians, McGraw-Hill, New York, N.Y., 1990.
Engelson, M., and J. Herbert: “Effective Characterization of CDMA Signals,” Microwave Jour-
nal, pg. 90, January 1995.
Howald, R.: “Understand the Mathematics of Phase Noise,” Microwaves & RF, pg. 97, December
1993.
Johnson, J. B:, “Thermal Agitation of Electricity in Conduction,” Phys. Rev., vol. 32, pg. 97, July
1928.
Nyquist, H.: “Thermal Agitation of Electrical Charge in Conductors,” Phys. Rev., vol. 32, pg.
110, July 1928.
Pleasant, D.: “Practical Simulation of Bit Error Rates,” Applied Microwave and Wireless, pg. 65,
Spring 1995.
Rohde, Ulrich L.: Digital PLL Frequency Synthesizers, Prentice-Hall, Englewood Cliffs, N.J.,
1983.
Rohde, Ulrich L.: “Key Components of Modern Receiver Design—Part 1,” QST, pg. 29, May
1994.
Rohde, Ulrich L. Rohde and David P. Newkirk: RF/Microwave Circuit Design for Wireless
Applications, John Wiley & Sons, New York, N.Y., 2000.
Rohde, Ulrich L, and Jerry C. Whitaker: Communications Receivers, 3rd ed., McGraw-Hill,
“Standards Testing: Bit Error Rate,” application note 3SW-8136-2, Tektronix, Beaverton, Ore.,
July 1993.
Using Vector Modulation Analysis in the Integration, Troubleshooting and Design of Digital RF
Communications Systems, Product Note HP89400-8, Hewlett-Packard, Palo Alto, Calif.,
1994.
Watson, R.: “Receiver Dynamic Range; Pt. 1, Guidelines for Receiver Analysis,” Microwaves &
RF, vol. 25, pg. 113, December 1986.
“Waveform Analysis: Noise and Jitter,” application note 3SW8142-2, Tektronix, Beaverton,
Ore., March 1993.
Wilson, E.: “Evaluate the Distortion of Modular Cascades,” Microwaves, vol. 20, March 1981.
Whitaker, Jerry C. (ed.): NAB Engineering Handbook, 9th ed., National Association of Broad-
g g
Chapter
6.1
Receiver Characteristics
Ulrich L. Rohde
6.1.1 Introduction1
The superheterodyne receiver makes use of the heterodyne principle of mixing an incoming sig-
nal with a signal generated by a local oscillator (LO) in a nonlinear element (Figure 6.1.1). How-
ever, rather than synchronizing the frequencies, the superheterodyne receiver uses a LO
frequency offset by a fixed intermediate frequency (IF) from the desired signal. Because a non-
linear device generates identical difference frequencies if the signal frequency is either above or
below the LO frequency (and also a number of other spurious responses), it is necessary to pro-
vide sufficient filtering prior to the mixing circuit so that this undesired signal response is sub-
stantially suppressed. The frequency of the undesired signal is referred to as an image frequency,
and a signal at this frequency is referred to as an image. The image frequency is separated from
the desired signal frequency by a difference equal to twice the IF. The preselection filtering
required at the signal frequency is much broader than if the filtering of adjacent channel signals
were required. The channel filtering is accomplished at IF. This is a decided advantage when the
receiver must cover a wide frequency band, because it is much more difficult to maintain con-
stant bandwidth in a tunable filter than in a fixed one. Also, for receiving different signal types,
the bandwidth can be changed with relative ease at a fixed frequency by switching filters of dif-
ferent bandwidths. Because the IF at which channel selectivity is provided is often lower than the
signal band frequencies, it may be easier to provide selectivity at IF, even if wide-band RF tuning
is not required.
1. This chapter is based on: Rohde, Ulrich L., and Jerry C. Whitaker: Communications Receiv-
ers: Principles and Design, 3rd ed., McGraw-Hill, New York, N.Y., 2000. Used with permis-
sion.
6-5
6-6 Radio Receivers
Figure 6.1.1 Block diagram of a superheterodyne receiver.
6.1.1a Practical Receivers

Because of the nature of active electronic devices, it is generally easier to provide high stable
gain in a fixed-frequency amplifier than in a tunable one, and gain is generally more economical
at lower frequencies. Thus, although the superheterodyne receiver does introduce a problem of
spurious responses not present in the other receiver types, its advantages are such that it has
replaced other types except for special applications.
Referring again to Figure 6.1.1, the signal is fed from the antenna to a preselector filter and
amplifier. The input circuit is aimed at matching the antenna to the first amplifying device so as
to achieve the best sensitivity while providing sufficient selectivity to reduce the probability of
overload from strong undesired signals in the first amplifier. Losses from the antenna coupling
circuit and preselection filters decrease the sensitivity. Because sufficient selectivity must be
provided against the image and other principal spurious responses prior to the mixing circuit, the
preselection filtering is often broken into two or more parts with intervening amplifiers to mini-
mize the effects of the filter loss on the noise factor (NF). The LO provides a strong stable signal
at the proper frequency to convert the signal frequency to IF. This conversion occurs in the mixer.
(This element has also been referred to as the first detector, converter, or frequency changer.)
The output from the mixer is applied to the IF amplifier, which amplifies the signal to a suit-
able power level for the demodulator. This circuit derives from the IF signal the modulation sig-
nal, which may be amplified by the baseband amplifier to the level required for output.
Normally, the output of an audio amplifier is fed to a headphone or loudspeaker at the radio, or
coupled externally for remote use. In other cases, the output may be supplied to a data demodula-
tor to produce digital data signals from the baseband signal. The data modem may also be fed
directly from the receiver at IF. Data demodulation is typically accomplished using digital pro-
cessing circuits rather than analog demodulators and amplifiers. In this case, the IF amplifier
must be designed to provide the appropriate level to an A/D converter so that digital processing
can be carried out. Additional IF filtering, data demodulation, and error control coding can all be
performed by digital circuits or a microprocessor.
An alternative to IF sampling and A/D conversion is the conversion of the signal to baseband
in two separate coherent demodulators driven by quadrature LO signals at the IF. The two out-
puts are then sampled at the appropriate rate for the baseband by two A/D converters or a single
multiplexed A/D converter, providing the in-phase and quadrature samples of the baseband sig-
nal. Once digitized, these components can be processed digitally to provide filtering, frequency
changing, phase and timing recovery, data demodulation, and error control.
6.1.2 The Receiving System

The first essential function of any radio receiver is to effect the transfer of energy picked up by
the antenna to the receiver itself through the input circuits. Maximum energy is transferred if the
impedance of the input circuit matches that of the antenna (inverse reactance and same resis-
tance) throughout the frequency band of the desired signal. This is not always feasible, and the
best energy transfer is not essential in all cases.
There are at least three antenna matching problems in a receiver. The first and, in many cases,
most crucial problem is that the receiver may be used from time to time with different antennas
whose impedances the potential users cannot specify fully. Second, antennas may be used in
mobile applications or in locations subject to changing foliage, buildings, or waves at sea, so that
the impedance—even if measured accurately at one time—is subject to change from time to
time. Third, at some frequencies, the problems of matching antennas are severely limited by
available components, and the losses in a matching network may prove greater than for a simpler
lower-loss network with poorer match.
When antenna matching is important over a wide band, it may be necessary to design a net-
work that can be tuned mechanically or electrically under either manual or automatic control in
response to a performance measure in the system. In older receivers with a wide tuning range, it
was common to have a mechanically tuned preselector that could be adjusted by hand and was
generally connected directly to the variable-frequency oscillator (VFO) used for the first conver-
sion. At times a trimmer was used in the first circuit to compensate for small antenna mis-
matches. Thus, tuning of the circuit could be modified to match the effects of the expected
antenna impedance range.
Often the problem of antenna matching design is solved by the user specification that defines
one or more “dummy antenna” impedances to be used with a signal generator to test the perfor-
mance of the receiver for different receiver input circuits. In this case, the user’s system is
designed to allow for the mismatch losses in performance that result from the use of actual
antennas.
A number of other receiver input considerations may occur in certain cases. The input circuits
may be balanced or unbalanced, or may need to be connectable either way. The input circuits may
require grounding, isolation from ground, or either connection. The circuits may need protection
from high-voltage discharges or from impulses. They may need to handle, without destruction,
high-power nearby signals, both tuned to the receiver frequency and off-tune. Thus, the input cir-
cuit can—at times—present significant design challenges.
6-8 Radio Receivers
6.1.2a Gain, Sensitivity, and Noise Figure

Any given receiver is usually required to receive and process a wide range of signal powers, but
in most cases it is important that they be capable of receiving distant signals whose power has
been attenuated billions of times during transmission. The extent to which such signals can be
received usefully is determined by the noise levels received by the antenna (natural and man-
made) and those generated within the receiver itself. It is also necessary that the receiver produce
a level of output power suitable for the application. Generally the ratio of the output power of a
device to the input power is known as the gain. The design of a receiver includes gain distribu-
tion among the various stages so as to provide adequate receiver gain and an optimum compro-
mise among the other operating characteristics.
While there are limits to the amount of gain that can be achieved practically at one frequency
because of feedback, modern receivers need not be gain-limited. When the gain is sufficiently
high, the weakest signal power that may be processed satisfactorily is noise-limited. This signal
level is referred to as the sensitivity of the system at a particular time and varies depending on the
external noise level. It is possible in some systems for the external noise to fall sufficiently so
that the system sensitivity is established by the internal noise of the receiver. A receiver’s sensi-
tivity is one of its most important characteristics. There are no universal standards for measure-
ment of sensitivity, although standards have been adopted for specific applications and by
specific user groups. Figure 6.1.2 shows a block diagram of the test setup and the typical steps
involved in determining receiver sensitivity.
NF
Sensitivity measures depend upon specific signal characteristics. The NF measures the effects of
inherent receiver noise in a different manner. Essentially it compares the total receiver noise with
the noise that would be present if the receiver generated no noise. This ratio is sometimes called
the noise factor F, and when expressed in dB, the NF. F is also defined equivalently as the ratio of
the S/N of the receiver output to the S/N of the source. The source generally used to test receivers
is a signal generator at local room temperature. An antenna, which receives not only signals but
noises from the atmosphere, the galaxy, and man-made sources, is unsuitable to provide a mea-
sure of receiver NF. However, the NF required of the receiver from a system viewpoint depends
on the expected S/N from the antenna. The effects of external noise are sometimes expressed as
an equivalent antenna NF.
For the receiver, we are concerned with internal noise sources. Passive devices such as con-
ductors generate noise as a result of the continuous thermal motion of the free electrons. This
type of noise is referred to generally as thermal noise, and is sometimes called Johnson noise
after the person who first demonstrated it. Using the statistical theory of thermodynamics,
Nyquist showed that the mean-square thermal noise voltage generated by any impedance
between two frequencies f1 and f2 can be expressed as
2 f2
Vn =4kt ∫ f1
R( f ) d f
(6.1.1)
where R(f) is the resistive component of the impedance.

(a)
(b )
Figure 6.1.2 Receiver sensitivity measurement: (a) test setup, (b) procedure.
Magnetic substances also produce noise, depending upon the residual magnetization and the
applied dc and RF voltages. This is referred to as the Barkhausen effect, or Barkhausen noise.
The greatest source of receiver noise, however, is generally that generated in semiconductors.
Like the older thermionic tubes, transistors and diodes also produce characteristic noise. Shot
noise resulting from the fluctuations in the carrier flow in semiconductor devices produces wide-
band noise, similar to thermal noise. Low-frequency noise or 1/f noise, also called flicker effect,
is roughly inversely proportional to frequency and is similar to the contact noise in contact resis-
tors. All of these noise sources contribute to the “excess noise” of the receiver, which causes the
NF to exceed 0 dB.
The NF is often measured in a setup similar to that of Figure 6.1.2, using a specially designed
and calibrated white-noise generator as the input. The receiver is tuned to the correct frequency
and bandwidth, and the output power meter is driven from a linear demodulator or the final IF
amplifier. The signal generator is set to produce no output, and the output power is observed. The
generator output is then increased until the output has risen 3 dB. The setting on the generator is
the NF in decibels.
The NF of an amplifier can also be calculated as the ratio of input to output S/N, per the equa-
tion
6-10 Radio Receivers
⎡ (S N )1 ⎤
NF = 10 log ⎢ ⎥
⎢⎣( S N ) 0 ⎥⎦
(6.1.2)
where NF is the noise figure in dB and (S/N)1 and (S/N)2 are the amplifier input and output SNR,
respectively.
The NF for a noiseless amplifier or lossless passive device is 0 dB; it is always positive for
nonideal devices. The NF of a lossy passive device is numerically equal to the device insertion
loss. If the input of a nonideal amplifier of gain G (dB) and noise figure NF (dB) were connected
to a matched resistor, the amplifier output noise power PNo (dB) would be
PNo = 10 log ( kT ) + 10 log ( B) + G + NF

(6.1.3)
where k is Boltzmann’s constant (mW/°K), T is the resistor temperature in °K, and B is the noise
bandwidth in Hz.
When amplifiers are cascaded, the noise power rises toward the output as noise from succeed-
ing stages is added to the system. Under the assumption that noise powers add noncoherently, the
noise figure NFT of a cascade consisting of two stages of numerical gain A 1 and A2 and noise
factor N1 and N2, is given by Friis’ equation
⎡ N + ( N 2 – 1) ⎤
NFT = 10 log ⎢ 1 ⎥
⎣ A1 ⎦
(6.1.4)
where the noise factor is N = 10(NF/10) and the numerical gain is A = 10(G/10). The system NF,
therefore, is largely determined by the first stage NF when A1 is large enough to make
( N 2 – 1 ) ⁄ A 1 much smaller than N1.
Minimum Detectable Signal

Another measure of sensitivity is the minimum detectable signal (MDS). The measurement pro-
cedure is similar to the NF measurement except that a sinusoidal signal generator replaces the
noise generator to produce the doubling of output power over noise power alone. This signal
power, just equal to the noise power, is defined as the MDS. Because receiver noise varies with
bandwidth, so does the MDS, which can be expressed as
MDS = k T B n F
(6.1.5)
In dBm, MDS = – 174 + 10 log ( B n + NF ) , where Bn is the noise bandwidth of the receiver. (0
dbm = decibels referenced to 1mW.)
The available thermal noise power per hertz is –174 dBm at 290 K (63ºF), an arbitrary refer-
ence temperature near standard room temperatures. When any two of the quantities in the expres-
sion are known, the third may be calculated. As in the case of NF measurements, care is required
in measuring MDS, because a large portion of the power being measured is noise, which pro-
duces MDS’ typical fluctuations.
6.1.3 Selectivity
Selectivity is the property of a receiver that allows it to separate a signal or signals at one fre-
quency from those at all other frequencies. At least two characteristics must be considered simul-
taneously in establishing the required selectivity of a receiver. The selective circuits must be
sufficiently sharp to suppress the interference from adjacent channels and spurious responses.
On the other hand, they must be broad enough to pass the highest sideband frequencies with
acceptable distortion in amplitude and phase. Each class of signals to be received may require
different selectivity to handle the signal sidebands adequately while rejecting interfering trans-
missions having different channel assignment spacings. However, each class of signal requires
about the same selectivity throughout all the frequency bands allocated to that class of service.
Older receivers sometimes required broader selectivity at higher frequencies to compensate for
greater oscillator drift. This requirement has been greatly reduced by the introduction of synthe-
sizers for control of LOs and economical high-accuracy and high-stability crystal standards for
the reference frequency oscillator.
Quantitatively the definition of selectivity is the bandwidth for which a test signal x decibels
stronger than the minimum acceptable signal at a nominal frequency is reduced to the level of
that signal. This measurement is relatively simple for a single selective filter or single-frequency
amplifier, and a selectivity curve can be drawn showing the band offset both above and below the
nominal frequency as the selected attenuation level is varied. Ranges of 80 to 100 dB of attenua-
tion can be measured readily, and higher ranges—if required—can be achieved with special care.
A test setup similar to Figure 6.1.2 may be employed with the receiver replaced by the selective
element under test. Proper care must be taken to achieve proper input and output impedance ter-
mination for the particular unit under test. The power output meter need only be sufficiently sen-
sitive, have uniform response over the test bandwidth, and have monotonic response so that the
same output level is achieved at each point on the curve. A typical IF selectivity curve is shown
in Figure 6.1.3.
The measurement of overall receiver selectivity, using the test setup of Figure 6.1.2, presents
some difficulties. The total selectivity of the receiving system is divided among RF, IF, and base-
band selective elements. There are numerous amplifiers and frequency converters, and at least
one demodulator intervening between input and output. Hence, there is a high probability of non-
linearities in the nonpassive components affecting the resulting selectivity curves. Some of the
effects that occur include overload, modulation distortion, spurious signals, and spurious
responses. If there is an AGC, it must be disabled so that it cannot change the amplifier gain in
response to the changing signal levels in various stages of the receiver. If there is only an AM or
FM demodulator for use in the measurement, distortions occur because of the varying attenua-
tion and phase shift of the circuits across the sidebands.
When measuring complete receiver

selectivity, with either a modulated or non-
modulated signal, it is wise to use an out-
put power meter calibrated in decibels.
The measurement proceeds as described
previously. However, if any unusual
changes in attenuation or slope are
observed, the generator level may be
increased in calibrated steps; it should be
noted whether the output changes decibel
for decibel. If not, what is being observed
at this point is not the selectivity curve, but
one of the many nonlinearities or
responses of the system.
Figure 6.1.3 Example of an IF selectivity curve.
6.1.4 Dynamic Range

The term dynamic range, especially in product literature, has been used to mean a variety of
things. We must be especially careful in using a common definition when comparing this charac-
teristic of receivers. In some cases, the term has been used to indicate the ratio in decibels
between the strongest and weakest signals that a receiver could handle with acceptable noise or
distortion. This is the ratio between the signal that is so strong that it causes maximum tolerable
distortion and the one that is so weak that it has the minimum acceptable S/N. This measure is of
limited value in assessing performance in the normal signal environment where the desired sig-
nal may have a range of values, but is surrounded by a dense group of other signals ranging from
very weak to very strong. The selective circuits of a receiver can provide protection from many
of these signals. However, the stronger ones, because of the nonlinearity of the active devices
necessary to provide amplification and frequency conversion, can degrade performance substan-
tially. In modern parlance, dynamic range refers to the ratio of the level of a strong out-of-band
signal that in some manner degrades signal performance of the receiver to a very weak signal.
The most common weak signal considered is the minimum discernible signal (MDS), and differ-
ing strong degrading signal levels may be used. It is, therefore, important to know which defini-
tion is meant when evaluating the meaning of the term dynamic range.
If the foregoing discussion of dynamic range seems vague, it is because there is not one char-
acteristic but several that are encompassed by the term. Each may have a different numeric value.
A receiver is a complex device with many active stages separated by different degrees of selec-
tivity. The response of a receiver to multiple signals of different levels is extremely complex, and
the results do not always agree with simple theory. However, such theory provides useful com-
parative measures. If we think of an amplifier or mixer as a device whose output voltage is a
function of the input voltage, we may expand the output voltage in a power series of the input
voltage
V o = Σ an V i
n
(6.1.6)
where a1 is the voltage amplification of the device and the higher-order an cause distortion.
Because the desired signal and the undesired interference are generally narrow-band signals,
we may represent Vi as a sum of sinusoids of different amplitudes and frequencies. Generally,
n
( A1 sin 2 π f1 t + A 2 sin 2 πf 2 t )
as a result of trigonometric identities, produces a number of components with different frequen-

cies,
mf1 ± ( n – m ) f2
with m taking on all values from 0 to n. These intermodulation (IM) products may have the same
frequency as the desired signal for appropriate choices of f1 and f2. When n is even, the minimum
difference between the two frequencies for this to happen is the desired frequency itself. This
type of even IM interference can be reduced substantially by using selective filters.
When n is odd, however, the minimum difference can be very small. Because m and n – m can
differ by unity, and each can be close to the signal frequency, if the adjacent interferer is ξ f from
the desired signal, the second need be only
( 2ξ f ) ⁄ ( n – 1 )
further away for the product to fall at the desired frequency. Thus, odd-order IM products can be
caused by strong signals only a few channels removed from the desired signal. Selective filtering
capable of reducing such signals substantially is not available in most superheterodyne receivers
prior to the final IF. Consequently, odd-order IM products generally limit the dynamic range sig-
nificantly.
Other effects of odd-order distortion are desensitization and cross modulation. For the case
where n is odd, the presence of the desired signal and a strong interfering signal results in a prod-
uct of the desired signal with an even order of the interfering signal. One of the resulting compo-
nents of an even power of a sinusoid is a constant, so the desired signal is multiplied by that
constant and an even power of the interferer’s signal strength. If the interferer is sufficiently
strong, the resulting product will subtract from the desired signal product from the first power
term, reducing the effective gain of the device. This is referred to as desensitization. If the inter-
ferer is amplitude-modulated, the desired signal component will also be amplitude-modulated by
the distorted modulation of the interferer. This is known as cross modulation of the desired signal
by the interfering signal.
This discussion provides a simple theory that can be applied in considering strong signal
effects. However, the receiver is far more complicated than the single device, and strong signal
performance of single devices by these techniques can become rapidly intractable as higher-
order terms must be considered. Another mechanism also limits the dynamic range. LO noise
sidebands at low levels can extend substantially from the oscillator frequency. A sufficiently
strong off-tune signal can beat with these noise sidebands in a mixer, producing additional noise
in the desired signal band. Other characteristics that affect the dynamic range are spurious sig-
nals and responses, and blocking.
The effects described here occur in receivers, and tests to measure them are essential to deter-
mining the dynamic range. Most of these measurements involving the dynamic range require
more than one signal input. They are conducted using two or three signal generators in a test
setup such as that indicated in Figure 6.1.4.
6.1.4a Desensitization
Desensitization measurements are related to the 1-dB compression point and general linearity of
the receiver. Two signal generators are used in the setup of Figure 6.1.4. The controls of the
receiver under test are set as specified, usually to one of the narrower bandwidths and with AGC
defeated or manual gain control (MGC) set so as to avoid effects of the AGC system. The signal
in the operating channel is modulated and set to a specified level, usually to produce an output S/
N measurement of a particular level, for example, 13 dB. The interfering signal is moved off the
operating frequency by a predetermined amount so that it does not affect the S/N measurement
because of beat notes and is then increased in level until the S/N measurement is reduced by a
specified amount, such as 3 dB. More complete information can be obtained by varying the fre-
quency offset and plotting a desensitization selectivity curve. In some cases, limits for this curve
are specified. The curve may be carried to a level of input where spurious responses, reciprocal
mixing, or other effects prevent an unambiguous measurement. Measurements to 120 dB above
sensitivity level can often be achieved.
If the degradation level at which desensitization is measured is set to 1-dB, and the desensitiz-
ing signal is well within the passband of the preselector filters, the desensitization level corre-
sponds to 1 dB gain compression (GC), which is experienced by the system up to the first mixer.
(See the subsequent discussion of intermodulation and intercept points.) A gain compression (or
blocking) dynamic range can be defined by comparing the input signal level at 1-dB GC to the
MDS; i. e., dynamic range (dB) equals the GC (input dBm) minus the MDS (input dBm). This is
sometimes referred to as the single-tone dynamic range, because only a single interfering signal
is needed to produce GC.
6.1.4b AM Cross Modulation

Although many saturation effects in receivers have been called cross modulation, SSB and FM
are not cross-modulated in the same sense as described previously. Cross modulation occurs in
AM and VSB signals by a strong modulated signal amplitude-modulating a weak signal through
the inherent nonlinearities of the receiver. Cross modulation typically occurs in a band allocated
for AM use and requires a much higher interfering signal level than for the generation of IM
products. The typical measurement setup is similar to that for overload measurements, except
that the interfering signal is amplitude-modulated, usually at a high level, such as 90 percent. The
modulation is at a different frequency than that for the operating channel (if it is modulated), and
a band-pass filter is used in the output to assure that the transferred modulation is being mea-
sured. The out-of-channel interfering signal is increased in level until the desired signal has a
specified level of output at the cross modulation frequency, for example, the equivalent of 10
percent modulation of the desired carrier. One or more specific offsets may be specified for the
measurement, or a cross-modulation selectivity curve may be taken by measuring carrier level
versus frequency offset to cause the specified degree of cross modulation.
Figure 6.1.4 Test setup for measuring the dynamic range properties of a receiver.
6.1.4c IM
As described in previous sections, IM produces sum and difference frequency products of many
orders that manifest themselves as interference. The measurement of the IM distortion perfor-
mance is one of the most important tests for a communications receiver. No matter how sensitive
a receiver may be, if it has poor immunity to strong signals, it will be of little use. Tests for even-
order products determine the effectiveness of filtering prior to the channel filter, while odd-order
products are negligibly affected by those filters. For this reason, odd-order products are generally
much more troublesome than even-order products and are tested for more frequently. The sec-
ond- and third-order products are generally the strongest and are the ones most frequently tested.
A two-signal generator test set is required for testing, depending on the details of the specified
test.
For IM measurements, the controls of the receiver under test are set to the specified band-
widths, operating frequency, and other settings as appropriate, and the gain control is set on man-
ual (or AGC disabled). One signal generator is set on the operating frequency, modulated and
adjusted to provide a specified S/N (that for sensitivity, for example). The modulation is dis-
abled, and the output level of this signal is measured. This must be done using the IF output, the
SSB output with the signal generator offset by a convenient audio frequency, or with the BFO on
and offset. Alternatively, the dc level at the AM demodulator can be measured, if accessible. The
signal generator is then turned off. It may be left off during the remainder of the test or retuned
and used to produce one of the interfering signals.
For second-order IM testing, two signal generators are now set to two frequencies differing
from each other by the operating frequency. These frequencies can be equally above and below
the carrier frequency at the start, and shifted on successive tests to assure that the preselection
filters do not have any weak regions. The signal with the frequency nearest to the operating fre-
quency must be separated far enough to assure adequate channel filter attenuation of the signal
(several channels). For third-order IM testing, the frequencies are selected in accordance with the
formula given previously so that the one further from the operating frequency is twice as far as
the one nearer to the operating frequency. For example, the nearer interferer might be three chan-
nels from the desired frequency and the further one, six channels in the same direction.
In either case, the voltage levels of the two interfering signal generators are set equal and are
gradually increased until an output equal to the original channel output is measured in the chan-
nel. One of several performance requirements may be specified. If the original level is the sensi-
tivity level, the ratio of the interfering generator level to the sensitivity level may have a specified
minimum. Alternatively, for any original level, an interfering generator level may be specified
that must not produce an output greater than the original level. Finally, an intercept point (IP)
may be specified.
The IP for the nth order of intermodulation occurs because the product is a result of the inter-
fering signal voltages being raised to the nth power. With equal voltages, as in the test, the result-
ant output level of the product increases as
V dn = c n V n
(6.1.7)
where cn is a proportionality constant and V is the common input level of the two signals.
Because the output from a single signal input V at the operating frequency is GvV, there is a the-
oretical level at which the two outputs would be equal. This value VIPn is the nth IP, measured at
the input. It is usually specified in dBm. In practice the IPs are not reached because as the ampli-
fiers approach saturation, the voltage at each measured frequency becomes a combination of
components from various orders of n. Figure 6.1.5 indicates the input-output power relationships
in second- and third-order IPs.
In Equation 6.1.7 we note that at the IP
V dn = c n (V IPn ) n and (V IPn ) out = G v (V IPn ) in

(6.1.8)
This leads to
1 – n
⎡V ⎤
c n = G v (V IPn ) 1 – n
and V dn = G v V ⎢ ⎥
⎣V IPn ⎦
(6.1.9)
The ratio of signal to distortion becomes ( VIP n ) ⁄ V n – 1 . In decibels it becomes

Figure 6.1.5 Input/output power relationships for second- and third-order intercept points.
⎡V ⎤
Rdn = 20 log ⎢ ⎥ = ( n – 1) [ 20 log V IPn – 20 log V ]
⎣V dn ⎦
(6.1.10)
If the intercept level is expressed in dBm rather than voltage, then the output power represented
by V must be similarly expressed.
The IM products we have been discussing originate in the active devices of the receiver, so
that the various voltages or power levels are naturally measured at the device output. The IP is
thus naturally referred to the device output and is so specified in most data sheets. In the forego-
ing discussion, we have referred the IP to the voltage level at the device input. If the input power
is required, we subtract from the output intercept level in decibels, the amplifier power gain or
loss. The relationship between input and output voltage at the IP is given in Equation 6.1.8. Ref-
erence of the IP to the device input is somewhat unnatural but is technically useful because the
receiver system designer must deal with the IP generation in all stages and needs to know at what
antenna signal level the receiver will produce the maximum tolerable IM products.
Consider the input power (in each signal) that produces an output IM product equal to the
MDS. The ratio of this power to the MDS may be called the third-order IM dynamic range. It
also is sometimes referred to as the two-tone dynamic range. Expressing Equation 6.1.10 in
terms of input power and input IP measured in dBm, we have
Rdn = ( n – 1) [ IPn ( in ) – P( in ) ]
(6.1.11)
When we substitute MDS for the distortion and MDS + DR for P(in) we obtain
DR = ( n – 1) [ IPn ( in ) – MDS – DR] ,
nDR = ( n – 1) [ IPn ( in ) – MDS ]

(6.1.12)
When n is 3, we find the relationship
2 [ IP3 ( in ) – MDS ]
DR =
3
(6.1.13)
A dynamic range could presumably be defined for other orders of IM, but it is not common to do
so. From the three different definitions of dynamic range described in this section, it should be
clear why it is important to be careful when comparing receiver specifications for this character-
istic.
6.1.4d Error Vector Magnitude

As shown in Figure 6.1.6, particular sources of amplitude and/or phase error can shift the values
of digital emission data states toward decision boundaries, resulting in increased BER because of
intersymbol interference. Figures 6.1.7, 6.1.8, and 6.1.9 show three additional sources of such
errors.
A figure of merit known as error vector magnitude (EVM) has been developed as a sensitive
indicator of the presence and severity of such errors. The error vector magnitude of an emission
is the magnitude of the phasor difference as a function of time between an ideal reference signal
and the measured transmitted signal after its timing, amplitude, frequency, phase, and dc offset
have been modified by circuitry and/or propagation. Figure 6.1.10 illustrates the EVM concept.
6.1.5 Gain Control

Receivers must often be capable of handling a signal range of 100 dB or more. Most amplifiers
remain linear over only a much smaller range. The later amplifiers in a receiver, which must pro-
Figure 6.1.6 Phase noise is critical to digitally modulated communication systems because of
the modulation errors it can introduce. Inter-symbol interference (ISI), accompanied by a rise
in BER, results when state values become so badly error-blurred that they fall into the
regions of adjacent states. This drawing depicts only the results of phase errors introduced
by phase noise; in actual systems, thermal noise, AM-to-PM conversion, differential group delay,
propagation, and other factors may also contribute to the spreading of state amplitude and
phase values. (Courtesy of Rohde & Schwarz.)
Figure 6.1.7 Effect of gain imbalance between I and Q channels on data signal phase con-stella-
tion. (Courtesy of Rohde & Schwarz.)
vide the demodulator with about 1 V on weak signals, would need the capability to handle thou-
sands of volts for strong signals without some form of gain control. Consequently, receivers
customarily provide means for changing the gain of the RF or IF amplifiers, or both.
For applications where the received signal is expected to remain always within narrow limits,
some form of manually selectable control can be used, which may be set on installation and sel-
dom adjusted. There are few such applications. Most receivers, however, even when an operator
is available, must receive signals that vary by tens of decibels over periods of fractions of sec-
Figure 6.1.8 Effect of quadrature offset on data signal phase constellation. (Courtesy of Rohde &
Schwarz.)
Figure 6.1.9 Effect of LO-feedthrough-based IQ offset on data signal phase constellation. (Cour-
tesy of Rohde & Schwarz.)
onds to minutes. The level also changes when the frequency is reset to receive other signals that
may vary over similar ranges but with substantially different average levels. Consequently, an
AGC is very desirable.
Some angle modulation receivers provide gain control by using amplifiers that limit on strong
signals. Because the information is in the angle of the carrier, the resulting amplitude distortion
is of little consequence. Receivers that must preserve AM or maintain very low angle modulation
distortion use amplifiers that can be varied in gain by an external control voltage. In some cases,
this has been accomplished by varying the operating points of the amplifying devices, but most
modern systems separate solid-state circuits or switched passive elements to obtain variable
attenuation between amplifier stages with minimum distortion. For manual control, provision
can be made to let an operator set the control voltage for these variable attenuators. For automatic
control, the output level from the IF amplifiers or the demodulator is monitored by the AGC cir-
cuit and a low-pass negative-feedback voltage is

derived to maintain a relatively constant signal level.
A block diagram of a dual AGC loop system is
illustrated in Figure 6.1.11. One loop is driven by
first IF energy that is band-limited, and the other
loop is driven by second IF energy that is band-lim-
ited by optional second IF filters. The first loop
controls a PIN diode pi attenuator ahead of the first
mixer. The second loop controls the second IF
amplifier stages. In this design, a microprocessor
adjusts the time constants of both loops so that time
delays introduced by the filters do not cause AGC Figure 6.1.10 The concept of error vec-
oscillation. tor magnitude. (Courtesy of Hewlett-
The principal AGC characteristics of importance Packard.)
are the steady-state control range and output-input
curve, and the attack and decay times. In a some systems, a variety of time constants are pro-
vided for the AGC to allow for different modulation types. For AM voice modulation, the radi-
ated carrier is continually present and the lowest sidebands are usually several thousand hertz
removed from the carrier. At the receiver, the carrier component can be separated from the
demodulated wave by a low-pass filter and can serve as the AGC control voltage. The response
time of the filter, which is often just an RC network, need only be fast enough to respond to the
fading rate of the medium, which is a maximum of 5 or 10 dB/s in most AM applications. A
response time of 0.1 to 0.2 s is required for such a fading rate. For the more common slower
rates, responses up to a second or more can be used.
Figure 6.1.11 Block diagram of a dual loop AGC system for a communications receiver.
Figure 6.1.12 Representative input-output AGC curves.
To test for the AGC range and input-output curve, a single signal generator is used (as in Fig-
ure 6.1.2) in the AM mode with the receiver’s AGC actuated. The signal generator is set to sev-
eral hundred microvolts, and the baseband output level is adjusted to a convenient level for
output power measurement. The signal generator is then tuned to its minimum level and the out-
put level is noted. The signal is gradually increased in amplitude, and the output level is mea-
sured for each input level, up to a maximum specified level, such as 2 V. Figure 6.1.12 shows
some typical AGC curves. In most cases, there will be a low-input region where the signal out-
put, rising out of the noise, varies linearly with the input. At some point, the output curve bends
over and begins to rise very slowly. At some high level, the output may drop off because of satu-
ration effects in some of the amplifiers. The point at which the linear relationship ends is the
threshold of the AGC action. The point at which the output starts to decrease, if within a speci-
fied range, is considered the upper end of the AGC control range. The difference between these
two input levels is the AGC control range. If the curve remains monotonic to the maximum input
test level, that level is considered the upper limit of the range. A measure of AGC effectiveness is
the increase in output from a specified lower input voltage level to an upper input voltage level.
For example, a good design might have an AGC with a threshold below 1 μV that is monotonic
to a level of 1 V and has the 3 dB increase in output between 1 μV and 0.1 V.
The foregoing applies to a purely analog system. The advantage of a digital signal processor
(DSP)-based system include that AGC is handled internally. The receiver still requires a dual-
loop AGC of which the input stage AGC will remain analog.
6.1.6 Digital Receiver Characteristics

The foregoing has not exhausted the analog receiver characteristics that may be of interest but
has reviewed some of the more significant ones. For example, in FM sets, the capture ratio is
important. Clearly, an area of increasing interest is the characterization of systems utilizing digi-
Figure 6.1.13 Test setup for eye pattern measurement.
tal modulation techniques. Because a digital radio system is a hybrid analog/digital device, many
of the test procedures outlined previously for analog receivers are useful and important in charac-
terizing a digital radio system. Additional tests, primarily involving the analysis of bit error rates
(BER), must also be run to properly identify any weak points in a particular receiver design or
basic architecture.
6.1.6a Bit Error Rate

The primary method for testing the quality of transmission over a high speed digital communica-
tions link is the BER, defined as the number of bit errors divided by the number of bits transmit-
ted. The BER is also used to qualify the sensitivity and noise characteristics of a receiver. The
major contributor to BER is jitter, which results from noise in the system. This noise causes the
output comparator to vary the time of its transition relative to the data clock. If the transition time
changes too much, an error will result.
Using a signal analyzer specifically designed for BER testing, jitter can be displayed directly,
or the BER is simply tabulated by the analyzer. The format of the testing signal is determined by
the application to which the digital radio system is applied. A variety of formats and protocols
are used. For end-to-end testing, the analyzer feeds a reference RF signal generator whose char-
acteristics are known, and the output signal is applied to the receiver antenna input.
The noise and jitter on a data waveform provides vital information about the quality of the
signal. A typical setup for capturing an eye pattern is shown in Figure 6.1.13. Eye patterns are
the traditional method of displaying high-speed digital data (Figure 6.1.14). Some communica-
tions signal analyzers augment this information with a built-in statistical database, which allows
additional analysis, including automated noise and jitter measurements on random data. Sophis-
ticated software can also analyze the form of the distribution, providing mean, rms, and standard
deviation results.
Proper receiver design involves identifying the areas of the system that are likely to cause
problems. LO phase noise is one such area. Phase noise can seriously impair the performance of
a digital receiver. Excessive phase noise can increase the BER, especially in systems using
Figure 6.1.14 Eye pattern display of BER measurement. (Courtesy of Tektronix.)
phase-based modulation schemes, such as binary PSK and quadrature PSK. For a given statisti-
cal phase-error characteristic, BER is degraded according to the percentage of time that the
phase error causes the signal position in signal space to cross a decision boundary.
6.1.7 Bibliography
Engelson, M., and J. Herbert: “Effective Characterization of CDMA Signals,” Microwave Jour-
nal, pg. 90, January 1995.
Howald, R.: “Understand the Mathematics of Phase Noise,” Microwaves & RF, pg. 97, December
1993.
Johnson, J. B:, “Thermal Agitation of Electricity in Conduction,” Phys. Rev., vol. 32, pg. 97, July
1928.
Nyquist, H.: “Thermal Agitation of Electrical Charge in Conductors,” Phys. Rev., vol. 32, pg.
110, July 1928.
Pleasant, D.: “Practical Simulation of Bit Error Rates,” Applied Microwave and Wireless, pg. 65,
Spring 1995.
Rohde, Ulrich L.: “Key Components of Modern Receiver Design—Part 1,” QST, pg. 29, May
1994.
Rohde, Ulrich L. Rohde and David P. Newkirk: RF/Microwave Circuit Design for Wireless
Applications, John Wiley & Sons, New York, N.Y., 2000.
“Standards Testing: Bit Error Rate,” application note 3SW-8136-2, Tektronix, Beaverton, Ore.,
July 1993.
Using Vector Modulation Analysis in the Integration, Troubleshooting and Design of Digital RF
Communications Systems, Product Note HP89400-8, Hewlett-Packard, Palo Alto, Calif.,
1994.
Watson, R.: “Receiver Dynamic Range; Pt. 1, Guidelines for Receiver Analysis,” Microwaves &
RF, vol. 25, pg. 113, December 1986.
“Waveform Analysis: Noise and Jitter,” application note 3SW8142-2, Tektronix, Beaverton,
Ore., March 1993.
g g
Chapter
6.2
The Radio Channel
Ulrich L. Rohde
6.2.1 Introduction1
The transmission of information from a fixed station to one or more mobile stations is governed
by the characteristics of the radio channel [1]. The RF signal arrives at the receiving antenna not
only on the direct path but is normally reflected by natural and artificial obstacles in its way.
Consequently, the signal arrives at the receiver several times in the form of echoes that are super-
imposed on the direct signal, as illustrated in Figure 6.2.1. This superposition may be an advan-
tage as the energy received in this case is greater than in single-path reception. However, this
characteristic may be a disadvantage when the different waves cancel each other under unfavor-
able phase conditions. In conventional car radio reception this effect is known as fading. It is par-
ticularly annoying when the vehicle stops in an area where the field strength is reduced because
of fading (for example, at traffic lights). Additional difficulties arise when digital signals are
transmitted. If strong echo signals (compared to the directly received signal) arrive at the receiver
with a delay in the order of a symbol period or more, time-adjacent symbols interfere with each
other. In addition, the receive frequency may be falsified at high vehicle speeds because of the
Doppler effect so that the receiver may have problems estimating the instantaneous phase in the
case of angle-modulated carriers. Both effects lead to a high symbol error rate even if the field
strength is sufficiently high.
Radio broadcasting systems using conventional frequency modulation are not seriously
affected by these interfering effects in most cases. If an analog system is replaced by a digital one
that is expected to offer advantages over the previous system, the designer must ensure that the
expected advantages—for example, improved audio S/N and the possibility to offer supplemen-
tary services to the subscriber—are not achieved at the expense of reception in hilly terrain or at
high vehicle speeds because of extreme fading. For this reason, a modulation method combined
1. This chapter is based on: Rohde, Ulrich L., and Jerry C. Whitaker: Communications Receiv-
ers: Principles and Design, 3rd ed., McGraw-Hill, New York, N.Y., 2000. Used with permis-
sion.
6-27
Figure 6.2.1 Mobile receiver affected by fading. (Courtesy of Rohde & Schwarz.)
with suitable error protection must be found for mobile reception in a typical radio channel that
is immune to fading, echo, and Doppler effects.
6.2.2 Radio Channel Characteristics

With a view to this design objective, more detailed information on the radio channel is required.
The channel can be described by means of a model. In the worst case, which may be the case for
reception in urban areas, it can be assumed that the mobile receives the signal on several indirect
paths but not on a direct one. The signals are reflected, for example, by large buildings; the
resulting signal delays are relatively long. In the vicinity of the receiver, these paths are split up
into a great number of subpaths; the delays of these signals are relatively short. These signals
may again be reflected by buildings and by other vehicles or natural obstacles such as trees.
Assuming the subpaths being statistically independent of each other, the superimposed signals at
the antenna input cause considerable time- and position-dependent field-strength variations with
an amplitude obeying the Rayleigh distribution (Figures 6.2.2 and 6.2.3). If a direct path is
received in addition to the reflected ones, the distribution changes to the Rice distribution and
finally, when the direct path becomes dominant, the distribution follows the Gaussian distribu-
tion with the field strength of the direct path being used as the center value.
In a Rayleigh channel, the bit error rate increases dramatically compared to the BER in an
AWGN (additive white Gaussian noise) channel (Figure 6.2.4).
6.2.2a Channel Impulse Response

The scenario outlined in the previous section can be demonstrated by means of the channel
impulse response [1]. Assume that a very short pulse of extremely high amplitude—in the ideal
case a Dirac pulse δ(t)—is sent by the transmitting antenna at a time t0 = 0. This pulse arrives at
the receiving antenna direct and in the form of reflections with different delays τ i and different
amplitudes because of path losses. The impulse response of the radio channel is the sum of all
received pulses (Figure 6.2.5). Because the mobile receiver and some of the reflecting objects are
Figure 6.2.2 Receive signal as a function of time or position. (Courtesy of Rohde & Schwarz.)
Figure 6.2.3 Rayleigh and Rice distribution. (Courtesy of Rohde & Schwarz.)
moving, the channel impulse response is a function of time and of delays τ i; that is, it corre-
sponds to
h(t, τ) = ∑ a i δ(t − τ i )
N
(6.2.1)
Figure 6.2.4 Bit error rate in a Rayleigh channel. (Courtesy of Rohde & Schwarz.)
Figure 6.2.5 Channel impulse response. (Courtesy of Rohde & Schwarz)
This shows that delta functions sent at different times t cause different reactions in the radio
channel.
In many experimental investigations, different landscape models with typical echo profiles
were created. The most important are:
Figure 6.2.6 Calculation of delay spread. (Courtesy of Rohde & Schwarz.)
• Rural area (RA)

• Typical urban area (TU)
• Bad urban area (BA)
• Hilly terrain (HT)
The channel impulse response tells us how the received power is distributed to the individual
echoes. A useful parameter, the delay spread can be calculated from the channel impulse
response, permitting an approximate description of typical landscape models, as illustrated in
Figure 6.2.6.
The delay spread also roughly characterizes the modulation parameters, carrier frequency,
symbol period, and duration of guard interval, which have to be selected in relation to each other.
If the receiver is located in an area with a high delay spread (for example, in hilly terrain), echoes
of the symbols sent at different times are superimposed when broadband modulation methods
with a short symbol period are used. An adjacent transmitter emitting the same information on
the same frequency has the effect of an artificial echo (Figure 6.2.7).
A constructive superposition of echoes is only possible if the symbol period is much greater
than the delay spread. The following holds
T s >10T d
(6.2.2)
Figure 6.2.7 Artificial and natural echoes in the single-frequency network. (Courtesy of Rohde &
Schwarz.)
Figure 6.2.8 Channel estimation. (Courtesy of Rohde & Schwarz.)
This has the consequence that relatively narrowband modulation methods have to be used. If this
is not possible, channel equalizing is required.
For channel equalizing, a continuous estimation of the radio channel is necessary. The estima-
tion is performed with the aid of a periodic transmission of data known to the receiver. In net-
works, a midamble consisting of, for example, 26 bits—the training sequence—can be
transmitted with every burst. The training sequence corresponds to a characteristic pattern of I/Q
signals that is held in a memory at the receiver. The baseband signals of every received training
sequence are correlated with the stored ones. From this correlation, the channel can be esti-
mated; the properties of the estimated channel will then be fed to the equalizer, as shown in Fig-
ure 6.2.8.
The equalizer uses the Viterbi algorithm (maximum sequence likelihood estimation) for the
estimation of the phases that most likely have been sent at the sampling times. From these phases
the information bits are calculated (Figure 6.2.9). A well designed equalizer then will superim-
pose the energies of the single echoes constructively, so that the results in an area where the ech-
oes are moderately delayed (delay times up to 16 μs at the receiver in this example) are better
Figure 6.2.9 Channel equalization. (Courtesy of Rohde & Schwarz.)
Figure 6.2.10 BERs after the channel equalizer in different areas. (Courtesy of Rohde &
Schwarz.)
than in an area with no significant echoes (Figure 6.2.10). The remaining bit errors are elimi-
nated using another Viterbi decoder for the transmitter convolutionally encoded data sequences.
The ability of a mobile receiver to work in an hostile environment such as the radio channel
with echoes must be proven. The test is performed with the aid of a fading simulator, which sim-
ulates various scenarios with different delay times and different Doppler profiles. A signal gener-
ator produces undistorted I/Q modulated RF signals that are downconverted into the baseband.
Next, the I/Q signals are digitized and split into different channels where they are delayed and
attenuated, and where Doppler effects are superimposed. After combination of these distorted
Figure 6.2.11 Fading simulator. (Courtesy of Rohde & Schwarz.)
signals at the output of the baseband section of the simulator, the signals modulate the RF carrier,
which is the test signal for the receiver under test (Figure 6.2.11).
6.2.2b Doppler Effect

Because the mobile receiver and (usually) some of the reflecting objects are in motion, the
receive frequency is shifted as a result of the Doppler effect [1]. In the case of single-path recep-
tion, this shift is calculated as follows
v
fd = f c cos α
c
(6.2.3)
Where:
v = speed of vehicle
c = speed of light
f = carrier frequency
α = angle between v and the line connecting transmitter and receiver
In the case of multipath reception, the signals on the individual paths arrive at the receiving
antenna with different Doppler shifts because of the different angles αi, and the receive spectrum
is spread. Assuming an equal distribution of the angles of incidence, the power density spectrum
can be calculated as follows
Figure 6.2.12 Doppler spread. (Courtesy of Rohde & Schwarz.)
1 1
P( f ) = for f < f d
π f −f
2 2
d
(6.2.4)
where fd = maximum Doppler frequency.

Of course, other Doppler spectra are possible in addition to the pure Doppler shift; for exam-
ple, spectra with a Gaussian distribution using one or several maxima. A Doppler spread can be
calculated from the Doppler spectrum analogously to the delay spread shown in Figure 6.2.12.
6.2.2c Transfer Function

The FFT value of the channel impulse response is the transfer function H(f,t) of the radio chan-
nel, which is also time-dependent. The transfer function describes the attenuation of frequencies
in the transmission channel. When examining the frequency dependence, it will be evident that
the influence of the transmission channel on two sine-wave signals of different frequencies
becomes greater with increasing frequency difference. This behavior can be adequately
described by the coherence bandwidth, which is approximately equal to the reciprocal delay
spread; that is
1
(Δf ) c =
Td
(6.2.5)
If the coherence bandwidth is sufficiently wide and—consequently—the associated delay

spread is small, the channel is not frequency-selective. This means that all frequencies are sub-
Figure 6.2.13 Effect of transfer function on modulated RF signals. (Courtesy of Rohde &
Schwarz.)
ject to the same fading. If the coherence bandwidth is narrow and the associated delay spread
wide, even very close adjacent frequencies are attenuated differently by the channel. The effect
on a broadband-modulated carrier with respect to the coherence bandwidth is obvious. The side-
bands important for the transmitted information are attenuated to a different degree. The result is
a considerable distortion of the receive signal combined with a high bit error rate even if the
received field strength is high. (See Figure 6.2.13).
6.2.2d Time Response of Channel Impulse Response and Transfer Function

The time response of the radio channel can be derived from the Doppler spread. It is assumed
that the channel rapidly varies at high vehicle speeds. The time variation of the radio channel can
be described by a figure, the coherence time, which is analogous to the coherence bandwidth.
This calculated value is the reciprocal bandwidth of the Doppler spectrum. A wide Doppler spec-
trum therefore indicates that the channel impulse response and the transfer function vary rapidly
with time, as shown in Figure 6.2.14. If the Doppler spread is reduced to a single line, the chan-
nel is time-invariant. In other words, if the vehicle has stopped or moves at a constant speed in a
Figure 6.2.14 Channel impulse response and transfer function as a function of time. (Courtesy of
Rohde & Schwarz.)
Figure 6.2.15 Phase uncertainty caused by Doppler effect. (Courtesy of Rohde & Schwarz.)
terrain without reflecting objects, the channel impulse response and the transfer function mea-
sured at different times are the same.
The effect on information transmission can be illustrated with a simple example. In the case
of MPSK modulation using hard keying, the transmitter holds the carrier phase for a certain
period of time; that is, for the symbol period T. In the case of soft keying with low-pass-filtered
baseband signals for limiting the modulated RF carrier, the nominal phase is reached at a spe-
cific time—the sampling time. In both cases the phase error ϕf = fdTS is superimposed onto the
nominal phase angle, which yields a phase uncertainty of Δϕ = 2ϕf at the receiver. The longer the
symbol period, the greater the angle deviation (Figure 6.2.15). Considering this characteristic of
the transmission channel, a short symbol period of Ts << (Δt)c should be used. However, this
requires broadband modulation methods.
Figure 6.2.16 shows the field strength or power arriving at the mobile receiver if the vehicle
moves in a Rayleigh distribution channel. Because the phase depends on the vehicle position, the
receiver moves through positions of considerably differing field strength at different times (time-
dependence of radio channel). In the case of frequency-selective channels, this applies to one fre-
quency only; that is, to a receiver using a narrowband IF filter for narrowband emissions. As Fig-
Figure 6.2.16 Effect of bandwidth on fading. (Courtesy of Rohde & Schwarz.)
ure 6.2.16 shows, this effect can be reduced by increasing the bandwidth of the emitted signal
and consequently the receiver bandwidth.
6.2.3 References
1. Rohde, Ulrich L., and David P. Newkirk: RF/Microwave Circuit Design for Wireless Appli-
cations, John Wiley & Sons, New York, N.Y., 2000.
g g
Chapter
6.3
AM and FM Receivers
6.3.1 Introduction1
A simplex transmitter-receiver system is the simplest type of communication system possible. A
single source transmits to a single receiver. When the receiver can respond to a transmission
from the source and relay voice or other information back to the transmission site, a duplex
arrangement is realized. A two-way business radio system operating on different transmit and
receive frequencies employs a duplex arrangement. AM and FM broadcast systems utilize a sim-
plex star configuration in which one transmitter feeds many receivers. For greatest efficiency,
the receivers in a simplex star should be made as simple as possible, concentrating technical
complexity at the transmit site. Radio broadcasting today follows this basic rule.
AM and FM stations transmit at high power levels to facilitate simpler radio designs. AM sta-
tions in the U.S. can operate at up to 50 kW, while FM stations can operate at up to 100 kW. This
essentially eliminates the need for a sensitive antenna at the receiver. Stereo AM and stereo FM
systems are, likewise, designed to permit the more complex operations to be performed in the
encoding stage at the transmitter, rather than in the decoding circuits of the receiver.
6.3.1a Superheterodyne Receiver

Virtually every AM-FM radio manufactured today incorporates the superheterodyne method of
reception. This system makes use of the heterodyne principle of mixing an incoming signal with
a signal generated by a local oscillator (LO), as illustrated in Figure 6.3.1. The LO is offset by a
fixed intermediate frequency (IF) from the desired signal. Because the mixer (a nonlinear
device) generates a difference frequency that is identical if the desired signal is either above or
below the LO frequency (and also a number of other spurious responses), filtering is required
prior to the mixer to suppress the undesired signal. The frequency of the undesired signal is
referred to as an image frequency and is separated from the desired signal frequency by a differ-
ence equal to twice the IF.
1. Portions of this chapter were adapted form Rohde, Ulrich L., and Jerry C. Whitaker: Commu-
nications Receivers, 3rd ed., McGraw-Hill, New York, N.Y., 2000. Used with permission.
6-39
Figure 6.3.1 Simplified block diagram of a superheterodyne receiver. Depending on the sophisti-
cation of the radio, some of the stages shown may be combined into a single circuit. (From [1].
Channel filtering is accomplished by one or more fixed-frequency filters in the IF. This is a
decided advantage when the receiver must cover a wide frequency band, because it is more diffi-
cult to maintain constant bandwidth in a tunable filter than in a fixed one. An IF frequency of
455 kHz is used for AM receivers and 10.7 MHz for FM receivers.
As shown in Figure 6.3.1, the input signal is fed from the antenna to a preselector filter and
RF amplifier. The input circuit matches the antenna to the first amplifying device to achieve the
best sensitivity. It also provides the necessary selectivity to reduce the possibility of overload in
the first amplifier stage caused by strong, undesired signals. Because sufficient selectivity must
be provided to eliminate the image and other spurious signals prior to the mixer, preselection fil-
tering may he broken into two or more parts with intervening amplifiers (in premium receiver
designs) to minimize the effects of filter losses on the receiver noise figure (NF).
The LO provides a strong stable signal at the proper frequency to the mixer for conversion of
the desired signal to the IF. (The mixer stage may also be called the first detector, converter, or
frequency changer.) The operating frequency of the LO may be controlled by a variable capacitor
or varactor diode operating in a phase-locked-loop (PLL) system. Frequency synthesizer PLL
designs are the basis for electronically-tuned radios (ETRs).
The output of the mixer is applied to the IF amplifier, which raises the desired signal to a suit-
able power level for the demodulator. The demodulator derives from the IF signal the modulated
baseband waveform, which may be amplified by a baseband amplifier before being applied to a
decoder (for stereo broadcasts) or audio preamplifier (for monophonic broadcasts).
6.3.2 Radio Wave Propagation

Radio waves are electromagnetic energy that can be attenuated, reflected, refracted, and scattered
by changes in the media through which they propagate. Radio waves in free space have electric
and magnetic field components that are mutually perpendicular and lie in a plane transverse to
the direction of propagation. Radio waves travel at a velocity of about 300,000 km/s.
The received signal field is accompanied by noise generated as the result of natural sources
(electrical activities in the atmosphere and space) and/or artificial sources (other radio transmis-
sions and machinery of various types). In addition, the receiver itself is a source of noise. Electri-
cal noise limits the range and performance of radio communications by requiring a sufficient
signal strength at the receiver to overcome the undesired noise.
6.3.2a AM Band Propagation

AM broadcasting falls in the middle of the medium-frequency (MF) band, which extends from
300 kH to 3 MHz. During daylight hours, ground wave propagation is predominate, and the
effects of atmospheric noise are minimal. The receiver NF has little effect on overall signal qual-
ity unless the antenna system is very inefficient. At night, sky wave propagation is significant,
permitting reception of signals hundreds or even thousands of miles away. Unfortunately, atmo-
spheric noise is also greatest at night.
Fading at the receiver occurs in portions of a station's coverage area where both the ground
wave and sky wave are comparable in strength. Fading can become quite deep during periods
when the two waves are nearly equal. The ground wave will always reach the receiver through the
most direct route. The sky wave will arrive later in time, out of phase with the ground wave,
resulting in attenuation of the composite signal. Fading can also occur as a result of two or more
sky waves with different numbers of reflections combining at the receiver.
When fading is caused by two or more waves that interfere as a result of having traveled over
paths of different lengths, different frequencies within the transmitted spectrum can be attenu-
ated to various extents. This phenomenon is known as selective fading. It can result in severe dis-
tortion of the signal.
Noise-free AM reception is particularly difficult at night because of the tremendous amount
of noise present on the AM dial caused by sky wave interference from distant stations on the
same or adjacent frequencies. The only solution to this problem is to increase the signal strength
of the desired station at the antenna of the receiver. This essentially means that the effective ser-
vice area of an AM station will be restricted during nighttime hours to locations in which a
strong local ground wave signal is available.
6.3.2b FM Band Propagation

FM broadcasting occupies a portion of the very-high-frequency (VHF) band, which extends
from 30 to 300 MHz. For greatest reliability, most VHF communication relies on a line-of-sight
path from the transmitter to the receivers. For this reason, FM transmitting antennas are located
on high towers, tall buildings, or mountaintop sites. The wavelength at FM frequencies is suffi-
ciently short that FM can penetrate automobiles and steel-frame buildings with a minimum of
loss. FM is not affected by nighttime sky wave interference. The propagation properties of FM
radio are essentially the same from day to night.
Figure 6.3.2 The causes of multipath in FM broadcasting. Distortion of the demodulated FM sig-
nal is related to both the direct-reflected signal ratio and the delay time of the reflected signal. Dis-
tortion increases as the signal ratio approaches 1:1, and as the secondary path delay increases.
FM is, however, affected by reflections of the transmitted signal that can mix at the receiver
and cause fading or distortion. This phenomenon is known as multipath (see Figure 6.3.2).
Reflections can be caused by mountains, steel-frame buildings, vehicles, and other objects. The
interference patterns set up as a result of multipath cause the signal strength to vary from one
location to another in an apparently random manner. It is often possible, for example, to improve
reception simply moving the antenna 2 ft or so (about 1/4 wavelength at FM frequencies).
The noise level at VHF is low compared to the MF band. Artificial noise can produce impul-
sive interference, but the use of hard limiting in receivers will eliminate most amplitude-related
noise. Interference from other FM stations can, however, produce noise at the receiver that usu-
ally cannot be stripped off by limiting.
6.3.3 Radio Receivers

The design of an AM or FM receiver is a complicated process involving a huge number of tech-
nical, economic, and perceptual characteristics. It is not enough for a receiver to perform well—
it must also meet a particular price point and respond to the needs (or wants) of the consumer.
Take, for example, the small, portable (“Walkman”-type) AM-FM radios available today. The
consumer wants a radio that is small and light enough to clip on a belt or slip into a shirt pocket.
The radio does not need to be feature-rich, but it must perform well without any external antenna
in a wide variety of working environments. The price point of the radio to consumers must be
below $20.
These specifications place enormous demands on the manufacturer. It is far easier to build a
large, heavy, feature-rich radio that uses an outdoor antenna, is powered from the ac line, and can
be priced upwards of $400 than it is to build the radio just described. The design of any AM-FM
radio for use by consumers is an exercise in compromise.
Most radios today have been reduced to just a handful of LSI chips, or even a single VLSI
device. This dramatic move toward miniaturization has eliminated the traditional stages that
technicians are familiar with. Still, each stage exists in one form or another in virtually every
radio produced today. The circuits may be hidden on a slab of silicon, but they are there just the
same.
6.3.3a Antenna Systems

The antenna must be designed to efficiently capture the transmitted signal and meet the physical
limitations of the receiver. This is easier said than done. The ideal antenna for an AM radio
would be a tall vertical wire or tower placed over a buried ground system. The FM equivalent
would be a high-gain Yagi antenna mounted 10 to 30 ft above the ground. Such antennas are
completely impractical for all but a few specialized, fixed installations. For most applications,
especially mobile or portable radios, compromises in antenna design must be made.
Antenna efficiency, impedance, bandwidth, and reception pattern are all a function of the
antenna dimensions relative to wavelength. A ferrite loop antenna is typically used for fixed and
portable AM radios, while a short (less than 3 ft) telescoping vertical antenna is normally used
for automobile AM radios. FM antennas usually consist of simply a length of wire (about 18 in)
that is tacked up on a nearby wall or run along a baseboard for fixed-location receivers, or a short
telescoping antenna for portable radios. Automobile FM radios use the AM auto antenna dis-
cussed previously. Shirt-pocket-sized radios simply rely on a length of wire run within the case
of the unit.
None of the antennas just discussed are particularly effective in converting radio waves into
electric signals. The constraints of space and cost have led manufacturers to sacrifice antenna
efficiency for a compact, convenient product. Supplemental external antennas are used only in
fixed applications requiring exceptional performance.
Antenna Coupling Network

Whatever form the antenna takes, it must be coupled efficiently to the first RF stage of the
receiver. The coupling network must exhibit low loss and adequate bandwidth to cover the band
selected (AM or FM). Until recently, it was customary practice to couple the antenna to a tuned
circuit connected to the input RF amplifier. Figure 6.3.3 illustrates several common coupling
arrangements. The examples differ mainly in the details of coupling. With ETR systems,
mechanical tuning has become essentially obsolete. Voltage-tuned capacitors (varactors) are typ-
ically used in radios today.
Whip Antenna
Nearly all automobile and some portable radios use whip antennas of about 2 to 3 ft in length for
both AM and FM reception. Seldom does the mounting surface resemble a plane. The problem
of coupling a whip optimally to the first active circuit is a difficult one. The antenna represents a
complex set of impedance, capacitance, reactance, and resistive components that vary with the
frequency and the surrounding physical structures. Still, given the ability to switch coupling cir-
Figure 6.3.3 Typical circuits used for coupling an antenna to a tuned resonant circuit. (From [1].
cuitry between AM and FM bands, coupling to the RF amplifier can be satisfactory for reason-
able efficiency and bandwidth.
Loop Antenna
The loop antenna has been used in portable AM receivers for many years. Its response differs
from the monopole in that when the face of the loop is vertical, it responds to the magnetic field
rather than the electric field. Instead of being omnidirectional in azimuth (like a whip), the loop
responds to the cosine of the angle between its face and the direction of the desired transmission.
This yields the familiar figure-eight pattern, which makes the loop useful for direction finding
by providing a sharp null for waves arriving perpendicular to the face.
Loops used for AM broadcast reception incorporate a high-permeability (ferrite) core to
facilitate reduced size. Such a loop may be tuned by a capacitance and connected directly to the
input device of the receiver. Coupling is usually simple, as illustrated in Figure 6.3.4. If the loop
Figure 6.3.4 Examples of coupling circuits used for AM broadcast reception using a loop antenna.
has an inductance lower than that required for proper input impedance, it may be connected in
series with an additional inductance for tuning (as shown). If the loop impedance is too high, the
receiver input may be tapped down on the circuit.
6.3.3b Filter Types

Filters are perhaps the most basic element of any receiver. If you are old enough to remember
“crystal radio” hobbyist sets popular many years ago, you will recall that these childhood toys
consisted of little more than a tuned circuit (filter) and detector. Although these sets have pretty
much been retired to museums, the principle still holds true.
Filter design has come a long way within the past couple of decades. Filters that once were
used only in military applications because of their high cost are now produced by the tens of
thousands. With reduced costs have come better designs. Modern radios include filters that
designers of consumer products could have only dreamed about 20 years ago.
Conventional filter design may be implemented using a number of different types of resona-
tors. The principal available technologies include:
• LC filter
• Electrical resonator
• Quartz crystal filter
• Monolithic quartz filter
• Ceramic filter
The classical approach to radio filtering involved cascading single- or dual-resonator filters
separated by amplifier stages. Overall selectivity was provided by this combination of one- or
two-pole filters. This approach, however, had two distinct disadvantages: 1) the circuits were dif-
ficult to align properly, and 2) the system was susceptible to IM and overload (even in the early
IF stages) from out-of-band signals. The classic approach did have some advantages, though.
First, limiting from strong impulse noise would occur in early stages where the broad bandwidth
would reduce the noise energy more than after complete selectivity had been achieved. Second,
such designs were relatively inexpensive to manufacture.
Modern radios use multiresonator filters inserted as early as possible in the amplification
chain to reduce nonlinear distortion, simplify alignment, and permit easy attainment of a variety
of selectivity patterns. The simple single- or dual-resonator pairs are now used primarily for
impedance matching between stages or to reduce noise between broadband cascaded amplifiers.
LC Filter
lnductor-capacitor resonators are limited to Q values on the order of a few hundred for reason-
able sizes. In most cases, designers must be satisfied with rather low Q. The size of the filter
depends on the center frequency. Two separate LC filters can easily cover the AM and FM broad-
cast bands. Skirt selectivity depends on the number of resonators used. Ultimate filter rejection
can be made higher than 100 dB with careful design. Filter loss depends on the percentage band-
width required and the resonator Q. It can be as high as 1 dB per resonator at narrow bandwidths.
This type of filter does not generally suffer from nonlinearities. Frequency stability is limited
by the individual components and cannot be expected to achieve much better than 0.1 percent of
center frequency under extremes of temperature and aging. Except for front ends that require
broad bandwidth filters, LC filters have been largely superseded in modern radios by new filter
technologies.
Quartz Crystal Filter

While several piezoelectric materials have been used for filter resonators, quartz crystal units
have proved to be the most satisfactory. Filters are available in frequencies that can easily cover
the AM broadcast band, and the FM band IF frequency (10.7 MHz). Standard filter response
shapes are available to simplify the receiver design process and keep costs down. Ultimate filter
rejection can be in excess of 100 dB. Input and output impedances are determined by input and
output matching networks in the filters. Insertion loss varies from about 1 to 10 dB, depending
on filter bandwidth and complexity. Frequency accuracy and temperature stability can be main-
tained to tight tolerances.
Monolithic Quartz Filter

A monolithic quartz filter is made up of a number of resonators constructed on a single quartz
substrate. The principal energy of each resonator is confined primarily to the region between the
plated electrodes, with a small amount of energy escaping to provide coupling. Usually these fil-
ters are limited to about four resonators. Monolithic quartz filters are available for operation
through the FM band with characteristics generally similar to discrete quartz resonator filters,
except that the bandwidth is limited to several tenths of a percent. Monolithic quartz filters are
also smaller and lighter than discrete resonator filters.
Ceramic Filter
Piezoelectrie ceramics can be used to achieve some of the characteristics of a quartz filter, but at
a lower cost. Such filters are comparable in size to monolithic quartz filters but are available over
a limited center frequency range (100 to 700 kHz). This limits ceramic filters to IF applications
for AM reception (455 kHz). The cutoff rate, stability, and accuracy of a ceramic filter are not as
good as quartz but are adequate for many applications. Single- and double-resonator structures
are available. Multiple-resonator filters use electrical coupling between sections.
6.3.3c RF Amplifier and AGC

Depending on the receiver design, an amplifier stage may be used to increase the voltage of the
signal received by the antenna (which may be 1 mV or less) to a level sufficient to drive the
mixer. Because of the wide range of signals to which a receiver must respond, the input device
must be capable of wide dynamic range. It must also be as linear as possible to minimize the gen-
eration of IM products from strong signals at the input. It follows that the number of strong sig-
nals should be minimized by restricting receiver bandwidth at as low a gain level as possible.
Thus, gain should be low prior to the most narrow bandwidth stage in the receiver (the IF).
At AM frequencies, it is common practice to avoid RF amplification and use the mixer as the
input device of the receiver. Bandwidth restriction is handled by filters in the first IF amplifier
stage.
When the desired signal is relatively strong, the RF amplifier may raise it to a level that will
cause distortion in later stages. Automatic gain control (AGC) is, therefore, provided for most
RF amplifiers. AGC circuits are basically low-frequency feedback systems. They are necessary
in a receiver to maintain a relatively constant output level when the input signal changes fre-
quently. The successful design of an AGC circuit that will perform satisfactorily under all
expected signal conditions is a major engineering challenge.
Gain control is generally distributed over a number of stages. The gain in later IF stages is
reduced first, and gain in earlier (RF and first IF) stages is reduced only for signal levels suffi-
ciently high to assure the desired S/N. Variable gain amplifiers are controlled electronically.
When attenuators are used in a receiver, they are usually operated electrically by either variable
voltages for continuous attenuators or by electric switches (relays or diodes) for fixed or stepped
attenuators.
The simplest method of gain control is to design one or more amplifier stages to change gain
in response to a control voltage. One common configuration uses a differential pair of common-
emitter amplifiers whose emitters are supplied through a separate common-emitter stage. The
gain-control voltage is applied to the base of the latter stage, while the signal is applied to one (or
both, if balanced) of the bases of the differential pair. This approach has been implemented in a
variety of linear integrated circuits. Figure 6.3.5 shows a schematic diagram of one such IC. The
gain-control voltage curves are given in Figure 6.3.6.
A PIN diode attenuator can also provide the low-distortion gain control that is especially
important prior to the mixer. Figure 6.3.7 shows such a circuit and its control elements. The con-
trol curve has approximately linear decibel variation over most of its 60-dB range. The pi-type
attenuator circuit provides a good match between terminations over the control range. The mini-
Figure 6.3.5 A gain-controlled RF amplifier integrated circuit suitable for use in AM-FM receivers.
mum useful frequency for a PIN diode attenuator varies inversely with the minority carrier life-
time. For available diodes, the low end of the HF band is near this limit.
The AGC Loop

The design objective of the AGC loop is to provide a substantially constant signal level to the
demodulator despite changes in the input signal level. Figure 6.3.8 shows a simplified block dia-
gram of a receiver with AGC. The input voltage to the RF amplifier stage may range between 1
μV and 1 V. The envelope of this voltage is detected at the input to the detector stage. This signal
is processed to produce the control voltages necessary for the variable-gain device (or devices) in
the IF amplifier chain. This control voltage may be increased in level by an amplifier stage, as
shown in Figure 6.3.9.
As the AGC tries to maintain constant output voltage with a varying input voltage, the issue
of attack and decay times comes into play. The AGC system has a finite delay in its response to a
change at the input of the receiver. In practice, it is not desirable for the AGC to have too fast a
reaction time. In such a case, any static pulse, ignition noise, or other impulsive interference with
a fast rise time would be detected by the AGC and desensitize the receiver for a given “hold
time.” The loop must be fast enough, though, to respond to signal level variations caused by ter-
rain, artificial structures, and multipath.
Figure 6.3.6 Gain-control curves of the CA3002 RF amplifier shown in Figure 6.3.5. The “Mode A”
and “Mode B” traces refer to different circuit configurations for the device.
6.3.3d Mixer
In the mixer circuit, the RF and LO signals are acted upon by the nonlinear properties of a device
(or devices) to produce a third frequency, the IF. At AM frequencies, receivers are usually built
without RF preamplifiers; the antenna is fed directly to the mixer stage. In this frequency range,
artificial and atmospheric noise is usually greater than the receiver NF. At FM frequencies, an
RF amplifier is typically used. The mixer is located in the signal chain prior to the narrow filter-
ing of the first IF.
Ideally, the mixer should accept the RF and LO inputs and produce an output having only one
frequency (sum or difference), with signal modulation precisely transferred to this IF. In actual
practice, however, mixers produce the desired IF but also many undesired components that must
be filtered. Any device with nonlinear transfer characteristics can act as a mixer. For the pur-
poses of this discussion, two classes will be discussed:
• Passive mixers, which use diodes as the mixing elements
• Active mixers, which employ gain devices (e.g., bipolar transistors or FETs)
Figure 6.3.7 Schematic of a PIN diode attenuator. (From [1]. Used with permission.)
Figure 6.3.8 Simplified block diagram of a receiver with AGC. (From [1]. Used with permission.)
Passive Mixer
Passive mixers have been built using germanium and silicon diodes. The development of hot car-
rier diodes, however, has resulted in a significant improvement in passive mixers.
A single diode can be used to build a mixer. The performance of such a circuit is poor,
though, because the RF and LO frequencies (as well as their harmonics and other odd and even
mixing products) all appear at the output. As a result, a large number of spurious components are
produced that are difficult to remove. Moreover, there is no isolation of the LO and its harmonics
from the input circuit, necessitating the use of an RF amplifier to prevent oscillator radiation
from the antenna.
A better approach can be found in the double-

balanced mixer, shown in Figure 6.3.10. This cir-
cuit, with its balanced diodes and transformers,
cancels even harmonics of both RF and LO fre-
quencies and provides isolation among the vari-
ous ports. For optimum performance, careful
matching of the diodes and transformers is neces-
sary. The manufacturing process for hot carrier
diodes has provided the tight tolerances that make Figure 6.3.9 Block diagram of a receiver
them substantially better than other diode types with amplified AGC. (From [1]. Used with
permission.)
available for mixer applications. Passive mixers
can be described as low-, medium- or high-level
mixers, depending on the diodes used and the number of diodes in the ring.
Active Mixer
The simplest type of active mixer uses a FET or bipolar transistor with the LO and RF signals
applied to the gate-source or base-emitter junction. This unbalanced mixer has the same draw-
backs as the simple diode mixer and is not used for high-performance receivers.
An improved configuration uses a dual-gate FET or cascode bipolar circuit with the LO and
RF signals applied to different gates (bases). The balanced transistor arrangement of Figure 6.3.5
can also be used as a mixer with the LO applied to the base of Q3 and the signal applied to the
bases of Q1 and/or Q5.
Active mixers can be implemented using a wide variety of devices and configurations,
depending on the specifications and cost structure of the receiver. Figure 6.3.11 shows a push-
pull balanced FET mixer. The circuit uses two dual-gate FETs in a push-pull arrangement
Figure 6.3.10 Schematic of a double-balanced passive mixer. (From [1]. Used with permission.)
Figure 6.3.11 Schematic diagram of a push-pull dual gate FET balanced mixer.
between the RF input (applied to the first gates) and the IF output. The oscillator is injected in
parallel on the second gates.
Active mixers have gain and are sensitive to mismatch conditions. If operated at high levels,
the collector or drain voltage can become so high that the base-collector or gate-drain junction
can open during a cycle and cause severe distortion. Control of the RF input by filtering out-of-
band signals and AGC are important considerations for active mixer designs. Advantages of the
active mixer include lower LO drive requirements and the possible elimination of an RF pream-
plifier stage.
6.3.3e Local Oscillator

Most modern high-performance receivers utilize a frequency-synthesized local oscillator to gen-
erate all individual frequencies needed over the required band(s). Synthesized oscillators gener-
ally use varactor diodes for the required variable-tuning capability. This arrangement offers a
stable frequency source that is easily tunable in precise steps and is easily adapted to prepro-
grammed stations. The days of station presets based on moving electromechanical elements are,
thankfully, gone for good. The LO must meet three basic requirements:
• Spectral purity to simplify IF filtering requirements
• Frequency agility to facilitate rapid station selection
Figure 6.3.12 Block diagram of a single-loop frequency synthesizer.
• Frequency adjustment accuracy to match the center carrier frequencies of AM and FM broad-
cast stations
Synthesizers may categorized into two basic classes:
• Direct, in which the LO output is derived from the product of multiple mixing and filtering
• Indirect, in which the LO output is derived from a phase-locked loop that samples the direct
output to reduce spurious signals.
PLL Synthesizer
Most AM-FM receivers incorporating a frequency synthesized LO use a single-loop digital PLL
of the type shown in Figure 6.3.12. When describing frequency synthesizers mathematically, a
linearized model is generally used. Because most effects occurring in the phase detector are
highly nonlinear, however, only the so-called piecewise linear treatment allows adequate approx-
imation. The PLL is nonlinear because the phase detector is nonlinear. However, it can be accu-
rately approximated by a linear model when the loop is in lock.
Assume that the voltage-controlled oscillator (VCO) of Figure 6.3.12 is tunable over a range
of 88 to 108 MHz. The output is divided to the reference frequency in a programmable divider
stage whose output is fed to one of the inputs of the phase-frequency detector and compared with
the reference frequency (fed to the other input). The loop filter at the output of the phase detector
suppresses the reference frequency components, while also serving as an integrator. The dc con-
trol voltage output of the loop filter pulls the VCO until the divided frequency and phase equal
those of the reference. A fixed division of the frequency standard oscillator (not shown in Figure
6.3.12) produces the reference frequency of appropriate step size. The operating range of the
PLL is determined by the maximum operating frequency of the programmable divider, its divi-
sion range ratio, and the tuning range of the VCO.
There are various choices of loop filter types and response. Because the VCO by itself is an
integrator, a simple RC filter following the phase detector can be used. If the gain of the passive
loop is too low to provide adequate drift stability of the output phase (especially if a high division
ratio is used), an active amplifier may be used as an integrator. In most frequency synthesizers,
an active filter-integrator is preferred to a passive one. Figure 6.3.13 shows a passive RC filter
for the second-order loop typically used in PLL synthesiz-

ers. Figure 6.3.14 shows an active filter for the second-
order loop.
Frequency Divider
Frequency dividers are commonly built using transistor-
transistor logic (TTL), complementary MOS (CMOS), and Figure 6.3.13 Schematic dia-
low-power emitter-coupled logic (ECL) IC technologies. gram of a PLL passive RC filter.
Dividers come in two common categories: synchronous
counters and asynchronous counters. The frequency range
of the CMOS, depending on the process, is limited to 10 to
30 MHz. TTL operates successfully up to 100 MHz in a
ripple counter configuration. In a synchronous counter
configuration, TTL is limited to perhaps 30 MHz.
Frequency extension is possible through the use of an
ECL prescaler, available in variable-ratio and fixed-ratio
configurations. The term prescaling is generally used in the
Figure 6.3.14 Schematic dia-
sense of a predivider that is nonsynchronous with the rest
gram of an active filter for a sec-
of the chain. Fixed-ratio prescalers are used as ripple ond-order PLL.
counters preceding a synchronous counter. A single-loop
synthesizer loses resolution by the amount of prescaling.
Figure 6.3.15 shows a block diagram of the MC12012 (Motorola) variable-ratio dual-modu-
lus presealer. Through external programming, this ECL divider can be made to divide in various
ratios. With proper clocking, the device can be considered a synchronous counter. With such a
system, it is possible to increase the maximum operating frequency to about 400 MHz without
losing resolution.
Variable-Frequency Oscillator
The LO in the receiver must be capable of being turned over a specified frequency range, offset
from the desired operating band(s) by the IF. Prior to the advent of the varactor diode and good
switching diodes, it was customary to tune an oscillator mechanically using a variable capacitor
with an air dielectric, or in some cases by moving a powdered iron core inside a coil to make a
variable inductor. Automobile radios typically used the variable-inductor method of tuning the
AM broadcast band. Figure 6.3.16 shows the classic VFO circuits commonly used in receivers.
Different configurations are used in different applications, depending on the range of tuning and
whether the tuning elements are completely independent or have a common element (such as the
rotor of a tuning capacitor).
Newer receivers control the oscillator frequency by electronic rather than mechanical means.
Tuning is accomplished by a voltage-sensitive capacitor (varactor diode). Oscillators that are
tuned by varying the input voltage are referred to as voltage-controlled oscillators (VCOs).
The capacitance versus voltage curves of a varactor diode depend on the physical composi-
tion of the diode junction. Maximum values range from a few hundred picofarads, and useful
capacitance ratios range from about 5 to 15. Figure 6.3.17 shows three typical tuning circuits
incorporating varactor diodes. In all cases the voltage is applied through a large value resistor.
Figure 6.3.15 Block diagram of a divide-by-10/11 dual-modulus prescaler IC, the Motorola
MC12012. (Courtesy of Motorola.)
Diode Switching
Because diodes have a low resistance when biased in one direction and a very high resistance
when biased in the other, they may be used to switch RF circuits. A sufficiently large bias voltage
may be applied to keep the diode on when it is carrying RF currents, or off when it is subjected to
RF voltages. It is important that, in the forward-biased condition, the diode add as little resis-
tance as possible to the circuit and that it be capable of handling the maximum RF current plus
the bias current. When the diode is reverse-biased, the breakdown voltage must be higher than
the combined bias and RF peak voltage in the circuit. Almost any type of diode can perform
switching, but at high frequencies, PIN diodes are especially useful. Figure 6.3.18 shows three
examples of diode switching in RF circuits.
The advantage of electronic tuning using varactor diodes is only fully realized when band
selection also takes place electronically. Diode switches are preferable to mechanical switches
because of their high reliability. Diode switches eliminate the need for a mechanical link between
front panel controls and the tuned circuits to be switched.
Crystal-Controlled Oscillator
Piezoelectric quartz crystals are the basis for most PLL reference oscillators. Quartz crystals
have resonances that are much more stable than the LC circuits discussed so far and also have
very high Q. Consequently, quartz crystal resonators are typically used for high-stability fixed-
frequency oscillators. A piezoelectric material is one that develops a voltage when it is under a
mechanical strain or is placed under strain by an applied voltage. A physical piece of such mate-
rial, depending upon its shape, can have a number of mechanical resonances. By appropriate
shaping and location of the electrodes, one or another resonant mode of vibration can be favored,
so that the resonance may be excited by an external voltage.
The crystal exhibits at its frequency of oscillation the equivalent electric circuit shown in Fig-
ure 6.3.19. The series resonant circuit represents the effect of the crystal vibrator, and the shunt
capacitance is the result of the coupling plates and of capacitance to surrounding metallic objects
(such as the metal case). The resonant circuit represents the particular vibrating mode that is
Figure 6.3.16 Schematic diagrams of common oscillator circuits using vacuum-tube, transistor,
and FET active circuits. (From [1]. Used with permission.)
(a) (b )
(c)
Figure 6.3.17 Typical tuning circuits using varactor diodes as the control element: (a) single diode
in the circuit low side, (b) single diode in the circuit high side, (c) two diodes in a series back-to-
back arrangement. (From [2]. Used with permission.)
excited. If more than one mode can be excited, a more complex circuit would be required to rep-
resent the crystal.
The most common type of circuit using a fundamental (AT) crystal is an aperiodic oscillator,
which has no selective circuits other than the crystal. Such oscillators, often referred to as paral-
lel resonant oscillators, use the familiar Pierce and Clapp configurations (see Figure 6.3.20).
6.3.3f AM-FM Demodulation

The function of any receiver is to recover the original information used to modulate the transmit-
ter. This process is referred to as demodulation, and the circuits that perform the recovery are
called demodulators. The term detector is also used, and the demodulator in a single-superhet-
erodyne receiver is sometimes called a second detector. Today, however, the term detector is sel-
dom used in this fashion.
Because of transmission and reception system distortions and noise caused by thermal, atmo-
spheric, and artificial sources, the demodulated signal is—to some extent—a distorted version of
the original modulating signal and is corrupted by the addition of noise. It is the obvious goal of
the demodulator to minimize these corrupting effects and provide an output signal that is as close
to the original modulating waveform as possible.
Analog modulated waves come in a variety of forms, including:
• Conventional AM
• Double-sideband suppressed carrier (DSSC) AM
(a)
(b)
(c)
Figure 6.3.18 Typical circuits using diodes for band switching: (a) series diode arrangement, (b)
shunt-diode arrangement, (c) use of both series and shunt diodes. (From [1]. Used with permis-
sion.)
Figure 6.3.19 The equivalent electric circuit of a crystal at resonance (spurious and overtone
modes not shown).
(a) (b )
(c )
Figure 6.3.20 Common parallel resonant circuits used in fundamental crystal oscillators: (a)
Pierce circuit, (b) Clapp circuit, collector grounded, (c) Clapp circuit, base grounded. (From [2].
(a)
(b)
Figure 6.3.21 AM demodulators with idealized waveforms: (a) average demodulator and resulting
waveform, (b) envelope detector and resulting waveform. (From [1]. Used with permission.)
• Single-sideband (SSB) AM
• Vestigial-sideband (VSB) AM
• Phase modulation (PM)
• Frequency modulation (FM)
For the purposes of this discussion, we will concentrate on the two forms used for AM-FM
broadcast transmission.
AM Demodulation
An AM signal is made up of an RF sinusoid whose envelope varies at a relatively slow rate about
an average (carrier) level. Any sort of rectifier circuit will produce an output component at the
modulation frequency. Figure 6.3.21 illustrates two of the simple diode rectifier circuits that may
be used, along with idealized waveforms. The average output of the rectifier of Figure 6.3.21a is
proportional to the carrier plus the signal. The circuit exhibits, however, significant output
energy at the RF and its harmonics. A low-pass filter is necessary to eliminate these components.
If the selected filter incorporates a sufficiently large capacitor at its input, the effect is to produce
a peak rectifier, with the idealized waveform shown in Figure 6.3.21b. In this case the demodu-
lated output is increased from the average of a half a sine wave (0.637 peak) to the full peak, and
the RF components are substantially reduced. A peak rectifier used in this way is often referred
to as an envelope detector or demodulator. It is the circuit most frequently used for demodulating
AM broadcast signals.
AM signals may also be demodulated by using a coherent or synchronous demodulator. This
type of demodulator uses a mixer circuit, with an LO signal synchronized in frequency and phase
to the carrier of the AM input. Figure 6.3.22 illustrates three approaches to synchronous demod-
(a) (b)
(c )
Figure 6.3.22 Three types of synchronous demodulators: (a) diode-based circuit, (b) dual-gate
MOSFET-based circuit, (c) bipolar IC (CA3005/CA3006) basic circuit.
ulation. The synchronous component can be generated by an oscillator phase-locked to the car-
rier, as illustrated in Figure 6.3.23.
The synchronous demodulator translates the carrier and sidebands to baseband. As long as
the LO is locked to the carrier phase, baseband noise results only from the in-phase component
of the noise input. Consequently the noise increase and S/N reduction that occur at low levels in
the envelope demodulator are absent in the synchronous demodulator. The recovered carrier fil-
tering is narrow band, so that phase lock can be maintained at carrier-to-noise levels below useful
modulation output levels. This type of circuit, while better than an envelope demodulator, is not
(a)
(b)
Figure 6.3.23 Two approaches to AM carrier recovery: (a) system based on a filter, clipper, and
amplifier; (b) system based on a filter, clipper, and PLL. (From [1]. Used with permission.)
generally used for AM broadcast demodulation because of its complexity. Most stereo AM
receivers, however, incorporate synchronous demodulators as part of the decoding circuit.
FM Demodulation
The most common technique for FM demodulation incorporates the use of linear circuits to con-
vert frequency variations to envelope variations, followed by an envelope detector. Another tech-
nique used with linear integrated circuits involves the conversion of frequency variations to
phase variations that are then applied to a phase demodulator. Still other FM demodulators
employ PLLs and frequency-locked loops (FM feedback circuits), or counter circuits whose out-
put is proportional to the rate of zero crossings of the wave. Frequency demodulators are often
referred to as discriminators or frequency detectors.
Resonant circuits are used in discriminators to provide adequate sensitivity to small-percent-
age frequency changes. To eliminate the dc components, two circuits can be used, one tuned
above and one tuned below the carrier frequency. The outputs are demodulated by envelope
demodulators and are then subtracted, eliminating the dc component. Voltage sensitivity is dou-
bled compared to the use of a single circuit. This balanced arrangement also eliminates even-
order distortion so that the first remaining distortion term is third-order. Figure 6.3.24 shows one
implementation of this scheme, known in the U.S. as the Travis discriminator. Because the circuit
depends on the different amplitude responses of two circuits, it has sometimes been called an
amplitude discriminator.
Figure 6.3.24 Schematic diagram of the Travis discriminator. (From [3]. Used with permission.)
The Foster-Seeley discriminator, shown in Figure 6.3.25, is a more common approach to FM

demodulation. In this circuit, the voltage across the primary is added to the voltage across each
of the two halves of the tuned secondary. At resonance the secondary voltage is in quadrature
with the primary voltage, but as the frequency changes, so do the phase shifts. The voltages from
the upper and lower halves of the secondary add to the primary voltage in opposition. As the fre-
quency rises, the phase shift increases, and as the frequency falls, it decreases. The opposite
phase additions cause the resulting amplitudes of the upper and lower voltages to differ, produc-
ing the discriminator effect. When the primary circuit is also tuned to the center frequency
(which produces much higher demodulation sensitivity), the phase of the primary voltage also
varies slightly, as does its amplitude. The proper selection of coupling factor is required to pro-
duce optimum sensitivity and linearity of the discriminator. Because of the method of arriving at
the amplitude difference in this demodulator, it is sometimes referred to as a phase discrimina-
tor.
The ratio detector is a variant of the phase discriminator that has an inherent degree of AM
suppression. The circuit tolerates less effective limiting in prior circuits and thus can reduce the
cost of the receiver. Figure 6.3.26 shows the basic concept of the ratio detector. It resembles the
Foster-Seeley circuit, except that the diodes are reversed. The combination of R1, R2, and C3 has
a time constant that is long compared to the lowest modulating frequency (on the order of 0.1 s
for audio modulation). The result is that during modulation, the voltage to the (grounded) center
tap across load resistor R2 is (E1 + E2)/2, and across R1 it is – (E1 + E2)/2. Following the circuit
from ground through R2 and C2, we see that the voltage at the center tape of the capacitors is
[(E1 + E2)/2] – E2 = (E1 – E2)/2, or half of the value of the Foster-Seeley discriminator.
The long time constant associated with C3 reduces the required current from the diodes when
E 1 + E2 drops and increases it when E1 + E2 rises. This changes the load on the RF circuit and
causes higher drive when the output falls and lower drive when it rises. This tends to further sta-
bilize the voltage E 1 + E2 against incidental AM. The sum voltage can also be used to generate
an AGC signal, so that the prior circuits need not limit. This can be advantageous when a mini-
mum number of circuits is required and the selectivity is distributed.
Figure 6.3.25 The Foster-Seeley FM discriminator circuit with tuned primary. (From [1]. Used with
permission.)
Figure 6.3.26 Schematic diagram of a basic ratio detector circuit. (From [1]. Used with permis-
sion.)
Amplitude Limiter
Amplitude limiting is essential for FM demodulators using analog circuits. Although solid-state
amplifiers tend to limit when the input signal becomes excessive, limiters that make use of this
characteristic often limit the envelope asymmetrically. For angle demodulation, symmetrical lim-
iting is desirable. AGC circuits, which can keep the signal output constant over wide ranges of
input signals, are unsuitable for limiting because they cannot be designed with sufficiently rapid
response to eliminate the envelope variations encountered in angle modulation interference. One
or more cascaded limiter stages are required for good FM demodulation.
Almost any amplifier circuit, when sufficiently driven, provides limiting. However, balanced
limiting circuits produce better results than those that are not balanced. In general, current cutoff
is more effective than current saturation in producing sharp limiting thresholds. Nonetheless,
overdriven amplifiers have been used in many FM systems to provide limiting. If the amplifier is
operated with low supply voltage and near cutoff, it becomes a more effective limiter. The tran-
(a)
(b) (c)
Figure 6.3.27 Typical FM limiter circuits: (a) balanced-transistor amplifier, (b) shunt-diode lim-
iter, (c) series diode limiter. (From [1]. Used with permission.)
sistor differential amplifier shown in Figure 6.3.27a is an excellent limiter when the bias of the
emitter load transistor is adjusted to cause cutoff to occur at small base-emitter input levels.
The classic shunt-diode limiter is shown in Figure 6.3.27b. It is important that the off resis-
tance of the diodes be much higher than the driving and load impedances, and the on resistance
be much lower. Figure 6.3.27c shows the classic series diode limiter. In this example the diodes
are normally biased on, so that they permit current flow between the driver and the load. As the
RF input voltage rises, one diode is cut off—and as it falls, the other is cut off. The effectiveness
of limiting is determined by the difference in off and on resistances of the diode, compared to the
driving and load impedances.
6.3.4 Stereo Systems

Stereophonic broadcasting has reshaped the radio industry. Stereo FM, introduced during the
early 1960s, gave FM broadcasters a powerful new marketing tool. Receiver manufacturers
responded with new radios that offered high-quality performance at an affordable price.
Figure 6.3.28 The composite baseband signal of the FM stereo system.
6.3.4a FM Stereo
The system devised for broadcasting stereo audio over FM has served the industry well. Key
requirements for the scheme were: 1) compatibility with monophonic receivers that existed at the
time the standard was developed, and 2) a robust signal that would not be degraded significantly
by multipath. Figure 6.3.28 shows the composite baseband that modulates the FM carrier for ste-
reophonic broadcasting. The two-channel baseband has a bandwidth of 53 kHz and is made up
of:
• A main channel (L + R) signal, which consists of the sum of left plus right audio signals (the
same signal broadcast by a monaural station). A fully modulated main channel will modulate
the FM transmitter to 45 percent when broadcasting stereo programming.
• A stereophonic subchannel (L – R), which consists of a double-sideband AM modulated car-
rier with a 38-kHz center frequency. The modulating signal is equal to the difference of the
left and right audio inputs. The subcarrier is suppressed to conserve modulation capability. As
a result, the AM sidebands have the same modulation potential as the main channel. A fully
modulated subchannel will modulate the FM transmitter to 45 percent when broadcasting ste-
reo programming.
• A 19-kHz subcarrier pilot, which is one-half the frequency of the stereophonic subcarrier and
in phase with it. The pilot supplies the reference signal needed by stereo receivers to reinsert
the 38-kHz carrier for demodulation of the double-sideband suppressed carrier transmission.
The pilot, in other words, is used to synchronize the decoder circuitry in the receiver to the
stereo generator at the transmitter. The frequency tolerance of the pilot is ±2 Hz. The pilot
modulates the transmitter 8 to 10 percent.
Generating the Stereo Signal

Two basic analog approaches have been used to generate the stereophonic subchannel: time-divi-
sion multiplexing (TDM), the switching method; and frequency-division multiplexing (FDM),
the matrix method.
A simplified block diagram of the FDM approach is shown in Figure 6.3.29. The left and
right audio channels first pass through a preemphasis circuit and low-pass filter. They are then
Figure 6.3.29 Functional block diagram of a frequency-division multiplexing (FDM) FM stereo

generator.
supplied to the matrix, which produces sum and difference components. The audio signals are
added to form the L + R main channel signal.
The difference signal is fed to a balanced modulator that generates the L – R subchannel.
Because a balanced modulator is used, the 38-kHz carrier is suppressed, leaving only the modu-
lated sidebands. The 19-kHz pilot signal is derived by dividing the 38-kHz oscillator by 2. The
main channel, stereophonic sub-channel, and pilot are then combined in the proper (45/45/10
percent) ratio to form the composite baseband.
The TDM method of generating a stereo signal is shown in block diagram form in Figure
6.3.30. The L + R and L – R signals are generated by an electronic switch that is toggled at a 38-
kHz rate. The switch samples one audio channel and then the other. Considerable harmonic
energy is generated in this process, requiring the use of a low-pass filter. When the harmonics are
filtered out, the proper composite waveform results. This approach, while simple and stable, may
produce unwanted artifacts, most notably reduced stereo separation, because of the filtering
requirements.
An improvement to the basic TDM concept is shown in Figure 6.3.31. By using a “soft
switch” to sample the left and right channels, it is possible to eliminate the low-pass filter and its
side-effects. The variable element shown in the figure consists of an electronic attenuator that is
capable of swinging between its minimum and maximum attenuation values at a 38-kHz rate.
Like the fast-switching TDM system, the L + R and L – R channels are generated in one opera-
tion. No filter is required at the output of the generator as long as the 38-kHz sine wave is free
from harmonics and the variable attenuator has good linearity.
Decoding the Stereo Signal

All stereo FM receivers include a circuit to convert the multiplexed signal at the FM detector to
the left and right audio channels originally transmitted by the station. There are a number of
ways to accomplish this task. In practice, however, one type of decoder is most commonly found,
built around a PLL integrated circuit (see Figure 6.3.32). This system offers both high perfor-
mance and low cost.
Figure 3.30 Functional block diagram of a time-division multiplexing (TDM) FM stereo generator.
Figure 3.31 Functional block diagram of a time-division multiplexing stereo generator using a vari-
able electronic attenuator.
The composite signal from the demodulator is fed to a buffer amplifier and sampled by a PLL
within the decoder IC. A voltage controlled oscillator, typically running at 76 kHz (four times
the pilot frequency) is locked in phase with the pilot by the error output voltage of the PLL. The
oscillator signal is divided by 2, resulting in a square wave at 38 kHz with nearly perfect duty
cycle and fast rise and fall times. This signal drives the audio switcher (demultiplexer) to transfer
the composite baseband to the left and right audio outputs in synchronization with the station’s
stereo generator. A deemphasis circuit follows the matrix to complement the signal preemphasis
at the FM transmitter.
6.3.4b AM Stereo
AM stereo operation was approved by the FCC in 1981 using the C-QUAM (Motorola) system.
A modification of simple quadrature modulation, C-QUAM was designed to maintain compati-
bility with monophonic transmissions.
Figure 6.3.32 Block diagram of a stereo decoder using PLL-controlled time-division multiplexing.
Figure 6.3.33 Block diagram of a C-QUAM AM stereo exciter.
The C-QUAM encoder is shown in Figure 6.3.33. As in FM stereo broadcasting, sum and dif-
ference signals of the left and right audio inputs are produced. Pure quadrature is generated by
taking the L + R and L – R signals and modulating two balanced modulators fed with RF signals
that are out of phase by 90° (producing components referred to as I and Q). As shown in the fig-
Figure 6.3.34 Block diagram of a C-QUAM AM stereo decoder.
ure, the 90° phase shift is derived by using a Johnson counter, which divides an input frequency
(4 times the station carrier frequency) by 4 and provides digital signals precisely 90° out of phase
for the balanced modulators. The carrier is inserted directly from the Johnson counter. At the
output of the summing network, the result is a pure quadrature AM stereo signal. From there it is
passed through a limiter that strips the incompatible AM components from the signal. The output
of the limiter is amplified and sent to the transmitter in place of the crystal oscillator.
The left and right audio signals are summed and sent as compatible L + R to the audio input
terminals of the transmitter.
Decoding the C-QUAM Signal

C-QUAM AM stereo is decoded by converting the demodulated AM broadcast waveform (which
is already close to a quadrature signal) to pure quadrature and then using a quadrature detector to
extract the L – R component (see Figure 6.3.34). In order to prepare the received signal for the
quadrature demodulator, it has to be converted from the envelope-detector-compatible signal that
is broadcast to the original quadrature signal that was not envelope detector compatible. This is
accomplished by demodulating the broadcast signal in two ways: with an envelope detector and
with an I detector. The two signals are compared, and the resultant error is used to gain-modulate
the input of the I and Q demodulators.
When the transmitted signal is L + R (monaural, no stereo), it is pure AM (only I sidebands).
In this case the envelope detector and the I demodulator see the same thing. There is no error sig-
nal, the inverse modulator does nothing, and the signal passes without change. However, when a
left- or right-only signal is transmitted, both AM and PM are present, and the input signal is
shifted in phase to the I demodulator, which loses some of its I amplitude. The envelope detector
sees no difference in the AM because of the phase modulation. When the envelope detector and
the I demodulator are compared, there is an error signal. The error signal increases the input level
to the detector. This makes the input signal to the I and Q demodulators look like a pure quadra-
ture signal, and the audio output yields the L – R information. The demodulator output is com-
bined with the envelope-detector output in a matrix to reconstruct the left and right audio
channels.
6.3.5 References
1. Rohde, Ulrich L, and Jerry C. Whitaker: Communications Receivers, 3rd ed., McGraw-
2. Rohde, Ulrich L.: Digital PLL Frequency Synthesizers, Prentice-Hall, Englewood Cliffs,
N.J., 1983.
3. Amos, S. W.: “FM Detectors,” Wireless World, vol. 87, no. 1540, pg. 77, January 1981.
6.3.6 Bibliography
Benson, K. Blair, and Jerry C. Whitaker: Television and Audio Handbook for Engineers and
Technicians, McGraw-Hill, New York, N.Y., 1990.
Whitaker, Jerry C. (ed.): NAB Engineering Handbook, 9th ed., National Association of Broad-
g g
Section
Television Receivers and Cable/Satellite

7
Distribution Systems
The familiar—and ubiquitous—television receiver has gone through significant and far reaching
changes within the past few years. Once a relatively simple, single purpose device intended for
viewing over-the-air broadcasts, the television set has become the focal point of entertainment
and information services in the home. What began as an all-in-one-box receiver has evolved into
a suite of devices intended to serve a viewing public that demands more selections, more flexibil-
ity, simpler control, and better pictures and sound. The TV set is—in fact—going through the
same metamorphosis that audio systems did in the 1970s. The “console stereo” of the 1960s
evolved from a multipurpose, albeit inflexible, aural entertainment system into the component
scenario that is universal today for high-end audio devices.
The move to such a component approach to television is important because it permits the con-
sumer to select the elements and features that suit his or her individual needs and tastes. Upgrade
options also are simplified. By separating the display device and its related circuitry from the
receiver makes it possible to upgrade from, for example, an NTSC receiver to a NTSC/DTV-
compliant receiver without sacrificing the display—where most of the cost is concentrated and
where most of the really significant advancements are likely to be seen in the coming years.
Cable and satellite systems also play into this component scenario, as new services are rolled out
to consumers.
While the arguments in favor of a component approach to video devices are compelling, the
tradeoff is complexity and—of course—cost. Consumers have clearly indicated that they want
entertainment devices to be simpler to operate and simpler to install, and price is always an issue.
Here again, the integration of audio devices points the way for video manufacturers. There is
ample evidence that consumers will pay a premium if a given device or system gives them what
they want.
Apart from consumer television, videoconferencing is a related area of receiver system devel-
opment. Although not strictly a receiver (in many cases, it is really a computer), emerging desk-
top videoconferencing systems promise to bring visual communication to businesses with the
ease of a phone call or e-mail. Long hamstrung by bandwidth bottlenecks, the era of high-speed
real-time networking has—at last—made desktop videoconferencing possible, and more impor-
tantly, practical.
7-1
y
7-2 Section Seven
In This Section:
Chapter 7.1: Television Reception Principles 7-11

Introduction 7-11
Basic Operating Principles 7-11
Tuner Principles 7-12
Selectivity 7-13
VHF Tuner 7-14
UHF Tuner 7-17
Tuner-IF Link Coupling 7-17
Electronically Tuned Systems 7-18
Varactor Tuner 7-19
Tuning Systems 7-20
Intermediate-Frequency (IF) Amplifier Requirements 7-22
Gain Characteristics 7-23
Response Characteristics 7-23
Surface Acoustic Wave (SAW) Filter 7-25
Gain Block Requirements 7-26
Video Signal Demodulation 7-27
Envelope Detector 7-28
Transistor Detector 7-30
Synchronous Detector 7-30
Automatic Gain Control (AGC) 7-30
AGC Delay 7-31
System Analysis 7-31
Automatic Frequency Control (AFC) 7-33
Sound Carrier Separation Systems 7-33
Video Amplifiers 7-36
Picture Controls 7-36
AC and dc Coupling 7-36
Low-Frequency Response Requirements 7-37
Color Receiver Luminance Channel 7-39
Chroma Subcarrier Processing 7-40
Burst Separation 7-40
Burst Gating Signal Generation 7-41
Color Subcarrier Reference Separation 7-41
Reference Systems 7-42
Chroma Demodulation 7-43
Demodulation Axis 7-43
RGB Matrixing and CRT Drive 7-43
Comb Filter Processing 7-44
Automatic Circuits 7-46
Tint 7-46
Scanning Synchronization 7-47
Vertical Countdown Synchronizing Systems 7-49
Impulse Noise Suppression 7-50
Horizontal Automatic Phase Control 7-50
Vertical Scanning 7-51
y
Television Receivers and Cable/Satellite Distribution Systems 7-3
Retrace Power Reduction 7-51

Horizontal Scanning 7-52
High Voltage Generation 7-54
Flyback with Half-Wave Rectifier 7-56
Voltage Multiplier Circuits 7-56
Integrated Flybacks 7-56
Power Supplies 7-56
References 7-58
Chapter 7.2: ATSC DTV Receiver Systems 7-61

Introduction 7-61
Noise Figure 7-61
Receiver System Overview 7-62
Tuner 7-63
Channel Filtering and VSB Carrier Recovery 7-65
Segment Sync and Symbol Clock Recovery 7-68
Noncoherent and Coherent AGC 7-69
Data Field Synchronization 7-70
Interference-Rejection Filter 7-71
Channel Equalizer 7-73
Phase Tracker 7-75
Trellis Decoder 7-77
Data De-Interleaver 7-79
Reed-Solomon Decoder 7-79
Data De-Randomizer 7-80
Receiver Loop Acquisition Sequencing 7-80
High Data Rate Mode 7-81
Receiver Equalization Issues 7-81
References 7-82
Bibliography 7-83
Chapter 7.3: Receiver Antenna Systems 7-85

Introduction 7-85
Basic Characteristics and Definitions of Terms 7-85
VHF/UHF Receiving Antennas 7-88
Dipole Antenna 7-88
Folded Dipole 7-90
V Dipole 7-90
Quarter-Wave Monopole 7-92
Loop Antenna 7-92
Multielement Array 7-92
DBS Reception 7-93
Phased Arrays 7-95
Transmission Lines 7-95
Active Antenna Systems 7-96
CEA Antenna Control Interface Standard 7-101
References 7-103
y
7-4 Section Seven
Chapter 7.4: Cable Television Systems 7-105

Introduction 7-105
Evolution of Cable Television 7-105
Coax-Based Cable Systems 7-106
Program Sources 7-106
Channel Assignments 7-106
Elements of a Cable System 7-107
Interactive Service 7-107
Head End 7-108
Signal Processing 7-110
Heterodyne Processors 7-111
Demodulator/Modulator Pair 7-113
Single-Channel Amplifiers 7-115
Supplementary Service Equipment 7-115
Head-End Signal Distribution 7-116
Amplification 7-117
Bridging Amplifier 7-117
Subscriber-Area Distribution 7-117
Trunk and Distribution Amplifiers 7-119
Gain Control 7-120
Cross-Modulation 7-120
Frequency Response 7-121
Second-Order Distortion 7-121
Noise Figure 7-122
Hum Modulation 7-122
Subscriber-Premises Equipment 7-122
Converters and Descramblers 7-123
Off-Premises Systems 7-124
Bidirectional Cable Television Systems 7-125
Fiber-Based Cable Systems 7-125
Hybrid Fiber/Coax 7-125
HFC Architectures 7-126
DWDM Architectures 7-127
Deep Fiber-Node Segmentation 7-128
Application Considerations 7-128
Return Path 7-129
Open Cable 7-130
Regulatory Issues 7-131
Initial Jurisdiction and Rules 7-131
1984 Congressional Policy and Rules 7-132
Cable Television Defined 7-133
Signal Carriage Requirements 7-134
Noncommercial Educational Television Stations 7-135
Low Power Television Stations 7-135
Radio Programming 7-135
Manner of Carriage 7-135
y
Technical Standards 7-136

References 7-136
Chapter 7.5: Satellite Delivery Systems 7-137

Introduction 7-137
The Satellite Communications Link 7-137
The Satellite 7-138
Antennas 7-139
Transponders 7-139
Power Sources 7-140
Control of Satellites 7-142
Positioning and Attitude Control 7-142
The Up-Link 7-144
The Down-Link 7-144
Performance Considerations 7-145
Home Terminal Variables 7-147
Rain Attenuation 7-147
Digital Transmission 7-149
References 7-150
Chapter 7.6: Content Distribution 7-151

Introduction 7-151
Internet and TV Convergence 7-151
Trends in Internet Content 7-152
Limitations of Internet Infrastructure 7-153
Content Distribution Network 7-154
Broadcast Internet 7-154
DTV Content Distribution Models 7-157
PC-Based DTV Receivers 7-158
Practical Implementation Considerations 7-158
Software DTV Decoding 7-159
Hardware Assisted Software DTV Decoding 7-159
Hardware DTV Decoding 7-159
Video Interconnection and Networking 7-160
DVI 7-161
IEEE 1394 7-162
Digital Home Network 7-163
HAVi Specifications 7-164
PVRs 7-164
DTV Product Classification 7-164
Videoconferencing 7-166
Infrastructure Issues 7-166
Full Motion Video 7-166
Desktop Systems—Hardware and Software 7-167
Small Group-vs.-Large Group Systems 7-167
System Implementations 7-167
References 7-169
Bibliography 7-169
y
7-6 Section Seven

ACATS: “Field Test Results of the Grand Alliance Transmission Subsystem,” Document SS/
WP2-1354, September 16, 1994.
ACATS: “Final Report of the Spectrum Utilization and Alternatives Working Party of the Plan-
ning Subcommittee of ACATS.”
Applebaum. S. P.: “Adaptive Arrays,” IEEE Trans., vol. AP-24, pg. 585, September 1976.
ATSC: “ATSC Transmission System, Equipment and Future Directions: Report of the ATSC
Task Force on RF System Performance,” Advanced Television Sysems Committee, Wash-
ington, D.C., April 12, 2001.
ATSC: “Guide to the Use of the ATSC Digital Television Standard,” Advanced Television Sys-
tems Committee, Washington, D.C., Doc. A/54, October 4, 1995.
ATSC: “Recommended Practice A/74, Receiver Performance Guidelines,” Advanced Television
Systems Committee, Washington, D.C., June 18, 2004
ATSC: “Technology Group Report T3-600, DTV Signal Reception and Processing Consider-
ations,” Advanced Television Systems Committee, Washington, D.C., September 18, 2003.
Baldwin, T. F., and D. S. McVoy: Cable Communications, Prentice-Hall, Englewood Cliffs, N.J.,
1983.
Bendov, O.: “On the Validity of the Longley-Rice (50,90/10) Propagation Model for HDTV Cov-
erage and Interference Analysis,” Proceedings of the Broadcast Engineering Conference,
Bendov, O., J. F. X. Brown, C. W. Rhodes, Y. Wu, and P. Bouchard: “DTV Coverage and Service
Prediction, Measurement and Performance Indices,” IEEE Trans. on Broadcasting, vol. 47,
no. 3, pg. 207, September 2001.
Bonang, C., and C. Auvray-Kander: “Next Generation Broadband Networks for Interactive Ser-
vices,” in Proceedings of IBC 2000, International Broadcasting Convention, Amsterdam,
2000.
“Cable Television Information Bulletin,” Federal Communications Commission, Washington,
D.C., June 2000.
Ciciora, Walter S.: “Cable Television,” in NAB Engineering Handbook, 9th ed., Jerry C. Whi-
taker (ed.), National Association of Broadcasters, Washington, D.C., pp. 1339– 1363, 1999.
Ciciora, Walter, et. al.: “A Tutorial on Ghost Canceling in Television Systems,” IEEE Transac-
tions on Consumer Electronics, IEEE, New York, N.Y., vol. CE-25, no. 1, pp. 9–44, Febru-
ary 1979.
Citta, R.: “Automatic phase and frequency control system,” U. S. patent No. 4 072 909, 7 Febru-
ary 1978.
Citta, R. W., and D. A. Wilming: “Receiver for a Trellis Coded Digital Television Signal,” U.S.
patent No. 5 636 251, 3 June 1997.
Citta, R. W., D. M. Mutzabaugh, and G. J. Sgrignoli: “Digital Television Synchronization System
and Method,” U.S. patent No. 5 416 524, 16 May 1995.
y
Citta, R. W., G. J. Sgrignoli, and R. Turner: “Digital Signal with Multilevel Signals and Sync
Recognition,” U.S. patent No. 5 598 220, 28 January 1997.
Compton, R. T., Jr., R. J. Huff, W. G. Swarner, and A. A. Ksienski: “Adaptive Arrays for Com-
munication Systems: An Overview of Research at the Ohio State University,” IEEE Trans.,
vol. AP-24, pg. 599, September 1976.
Cook, James H., Jr., Gary Springer, Jorge B. Vespoli: “Satellite Earth Stations,” in NAB Engi-
neering Handbook, Jerry C. Whitaker (ed.), National Association of Broadcasters, Wash-
ington, D.C., pp. 1285–1322, 1999.
DeVries, A., et al: “Characteristics of Surface-Wave Integratable Filters (SWIFS),” IEEE Trans.,
vol. BTR-17, no. 1, pg. 16.
Di Toro, M. J.: “Communications in Time-Frequency Spread Media,” Proc. IEEE, vol. 56, Octo-
ber 1968.
EIA/CEA-909, “Smart Antenna Interface,” Electronics Industries Alliance, Arlington, VA, Feb-
ruary 26, 2002.
Elliot, R. S.: Antenna Theory and Design, Prentice-Hall, Englewood Cliffs, N.J., pg. 64, 1981.
FCC Regulations, 47 CFR, 15.65, Washington, D.C.
Feinstein, J.: “Passive Microwave Components,” Electronic Engineers’ Handbook, D. Fink and
D. Christiansen (eds.), Handbook, McGraw-Hill, New York, N.Y., 1982.
Fink, D. G., and D. Christiansen (eds.): Electronic Engineer’s Handbook, 2nd ed., McGraw-Hill,
New York, 1982.
Fockens, P., and C. G. Eilers: “Intercarrier Buzz Phenomena Analysis and Cures,” IEEE Trans.
Consumer Electronics, vol. CE-27, no. 3, pg. 381, August 1981.
Gibson, E. D.: “Automatic Equalization Using Time-Domain Equalizers,” Proc. IEEE, vol. 53,
pg. 1140, August 1965.
Ghosh, M: “Blind Decision Feedback Equalization for Terrestrial Television Receivers,” Proc. of
the IEEE, vol. 86, no. 10, pp. 2070–2081, October, 1998.
Gibson, J. J., and R. M. Wilson: “The Mini-State—A Small Television Antenna,” RCA Engineer,
vol. 20, pp. 9–19, 1975.
Grossner, N.: Transformers for Electronic Circuits, 2nd ed., McGraw-Hill, New York, N.Y., pp.
344–358, 1983.
Hoff, L. E., and A. R. King: “Skywave Communication Techniques,” Tech. Rep. 709, Naval
Ocean Systems Center, San Diego, Calif, March 30,1981.
Hoffman, Gary A.: “IEEE 1394: The A/V Digital Interface of Choice,” 1394 Technology Associ-
ation Technical Brief, 1394 Technology Association, Santa Clara, Calif., 1999.
Horwitz, T. P., R. B. Lee, and G. Krishnamurthy: “Error Tracking Loop,” U.S. patent No. 5 406
587, 11 April 1995.
Hufford, G. A: “A Characterization of the Multipath in the HDTV Channel,” IEEE Trans. on
Broadcasting, vol. 38, no. 4, December 1992.
y
7-8 Section Seven
Hulst, G. D.: “Inverse Ionosphere,” IRE Trans., vol. CS-8, pg. 3, March 1960.
IEEE Guide for Surge Withstand Capability, (SWC) Tests, ANSI C37.90a-1974/IEEE Std. 472-
1974, IEEE, New York, N.Y., 1974.
Jasik, H., Antenna Engineering Handbook, McGraw-Hill, New York, Chapter 24, 1961.
Kahn, L. R.: “Ratio Squarer,” Proc. IRE, vol. 42, pg. 1704, November 1954.
Kase, C. A., and W. L. Pritchard: “Getting Set for Direct-Broadcast Satellites,” IEEE Spectrum,
IEEE, New York, N.Y., vol. 18, no. 8, pp. 22–28, 1981.
Kim, J. G., K. S. Kim, and S. W. Jung: “Phase Error Corrector for HDTV Reception System,”
U.S. patent No. 5 602 601, 11 February 1997.
Kraus, J. D.: Antennas, McGraw-Hill, New York, N.Y., Chapter 12, 1950.
Krishnamurthy, G., and R. B. Lee: “Error Tracking Loop Incorporating Simplified Cosine Look-
up Table,” U.S. patent No. 5 533 071, 2 July 1996.
Krishnamurthy, G., T. G. Laud, and R. B. Lee: “Polarity Selection Circuit for Bi-phase Stable
FPLL,” U. S. patent No. 5 621 483, 15 April 1997.
Limberg, A. L. R: “Plural-Conversion TV Receiver Converting 1st IF to 2nd IF Using Oscilla-
tions of Fixed Frequency Above 1st IF,” U. S. patent No. 6 307 595, 23 October 2001.
Lo, Y. T.: “TV Receiving Antennas,” in Antenna Engineering Handbook, H. Jasik (ed.),
McGraw-Hill, New York, N.Y., pp. 24–25, 1961.
Lucky, R. W.: “Automatic Equalization for Digital Communication,” Bell Sys.Tech. J., vol. 44,
pg. 547, April 1965.
Miyazawa, H.: “Evaluation and Measurement of Airplane Flutter Interference,” IEEE Trans. on
Broadcasting, vol. 35, no. 4, pp 362–367, December 1989.
Monsen, P.: “Fading Channel Communications,” IEEE Commun.., vol. 18, pg. 16, January 1980.
Mycynek, V., and G. J. Sgrignoli: “FPLL With Third Multiplier in an AC Path in the FPLL,” U.S.
patent No. 5 745 004, 28 April 1998.
NAB V TechCheck: “Antenna Control Interface Standard and Implementation,” National Associ-
ation of Broadcasters, Washington, D.C., April 22, 2002.
NAB TV TechCheck: “CEA Establishes Definitions for Digital Television Products,” National
Association of Broadcasters, Washington, D.C., September 1, 2000.
NAB TV TechCheck: “Consumer Electronic Consortium Publishes Updated Specifications for
Home Audio Video Interoperability,” National Association of Broadcasters, Washington,
D.C., May 21, 2001.
NAB TV TechCheck: “FCC Adopts Rules for Labeling of DTV Receivers,” National Association
of Broadcasters, Washington, D.C., September 25, 2000.
Neal, C. B., and S. Goyal: “Frequency and Amplitude Phase Effects in Television Broadcast Sys-
tems,” IEEE Trans., vol. CE-23, no. 3, pg. 241, August 1977.
O’Connor, Robert: “Understanding Television’s Grade A and Grade B Service Contours,” IEEE
Transactions on Broadcasting, vol. BC-14, no. 4, pp. 137–143, December 1968.
y
Plonka, Robert: “Can ATV Coverage be Improved with Circular, Elliptical or Vertical Polarized
Antennas?,” NAB Engineering Conference Proceedings, National Association of Broad-
Price, R., and P. E. Green, Jr.: “A Communication Technique for Multipath Channels,” Proc. IRE,
vol. 46, pg. 555, March 1958.
Proakis, J. G.: “Advances in Equalization for Intersymbol Interference,” in Advances in Commu-
nication Systems: Theory & Application, A. V. Balakrishnan and A. J. Viterbi (eds.), Aca-
demic Press, New York, N.Y., 1975.
Qureshi, S.: “Adaptive Equalization,” IEEE Commun., vol. 20, pg. 9, March 1982.
Qureshi, Shahid U. H.: “Adaptive Equalization,” Proceedings of the IEEE, IEEE, New York,
N.Y., vol. 73, no. 9, pp. 1349–1387, September 1985.
Radio Amateur’s Handbook, American Radio Relay League, Newington, Conn., 1983.
“Receiver Planning Factors Applicable to All ATV Systems,” Final Report of PS/WP3,
Advanced Television Systems Committee, Washington, D.C., December 1, 1994.
Riegler, R. L., and R. T. Compton, Jr.: “An Adaptive Array for Interference Rejection,” Proc.
IEEE, vol. 61, June 1973.
Rossweiler, G. C., F. Wallace, and C. Ottenhoff: “Analog versus Digital Null-Steering Control-
lers,” ICC ‘77 Conf. Rec., 1977.
Sgrignoli, G. J.: “Pilot Recovery and Polarity Detection System,” U.S. patent No. 5 675 283, 7
October1997.
“Television Receivers and Video Products,” UL 1410, Sec. 71, Underwriters Laboratories, Inc.,
Ungerboeck, Gottfried: “Fractional Tap-Spacing Equalizer and Consequences for Clock Recov-
ery in Data Modems,” IEEE Transactions on Communications, IEEE, New York, N.Y., vol.
COM-24, no. 8, pp. 856–864, August 1976.
Whitaker, Jerry C.: Interactive TV Survival Guide, McGraw-Hill, New York, N.Y., 2001.
Widrow, B., and J. M. McCool: “A Comparison of Adaptive Algorithms Based on the Methods
of Steepest Descent and Random Search,” IEEE Trans., vol. AP-24, pg. 615, September
1976.
Widrow, B., and M. E. J. Hoff: “Adaptive Switching Circuits,” IRE 1960 Wescon Conv. Record,
pp. 563–587.
Widrow, B., P. E. Mantey, L. J. Griffiths, and B. B. Goode: “Adaptive Antenna Systems,” Proc.
IEEE, vol. 55, pg. 2143, December 1967.
Wilming, D. A.: “Data Frame Structure and Synchronization System for Digital Television Sig-
nal,” U.S. patent No. 5 629 958, 13 May 1997.
Yamada and Uematsu: “New Color TV with Composite SAW IF Filter Separating Sound and
Picture Signals,” IEEE Trans., vol. CE-28, no. 3, pg. 193.
y
g g
Chapter
7.1
Television Reception Principles
K. Blair Benson
L. H. Hoke, Jr., L. E. Donovan, J. D. Knox, D. E. Manners, W. G. Miller,

R. J. Peffer, J. G. Zahnen
7.1.1 Introduction
Television receivers provide black-and-white or color reproduction of pictures and the accompa-
nying monaural or stereophonic sound from signals broadcast through the air or via cable distri-
bution systems. The broadcast channels in the U.S. are 6 MHz wide for transmission of
conventional 525-line NTSC signals and DTV signals.
7.1.2 Basic Operating Principles

The minimum signal level at which analog television receivers provide usable pictures and
sound, called the sensitivity level, generally is on the order of 10 to 20 μV. The maximum level
encountered in locations near transmitters may be as high as several hundred millivolts. The FCC
has set up two standard signal level classifications, Grades A and B, for the purpose of licensing
television stations and allocating coverage areas. Grade A is to be used in urban areas relatively
near the transmitting tower, and Grade B use ranges from suburban to rural and fringe areas a
number of miles from the transmitting antenna. The FCC values are expressed in microvolts per
meter (μV/m) [1].
The standard transmitter field-strength values for the outer edges of these services for Chan-
nels 2 through 69 are listed in Table 7.1.1. Included for reference in the table are the signal levels
for what may be considered “city grade” in order to give an indication of the wide range in signal
level that a receiver may be required to handle. The actual antenna terminal voltage into a
matched receiver load, listed in the second column, is calculated from
e = E -96.68
------------ (7.1.1)
f 1 f2
7-11
p p
7-12 Television Receivers and Cable/Satellite Distribution Systems
Table 7.1.1 Television Service Operating Parameters
Band and Frequency City grade Grade A Grade B

Channels (MHz)
μV/m μV μV/m μV μV/m μV
VHF 2–6 54–88 MHz 5,010 7030 2510 3520 224 314
VHF 7–13 174–216 MHz 7,080 3550 3550 1770 631 315
UHF 14–69 470–806 MHz 10,000 1570 5010 787 1580 248
UHF 70–831 806–890 MHz 10,000 1570 5010 571 1580 180
1. Receiver coverage of Channels 70 to 83 has been on a voluntary basis since July 1982. This frequency band
was reallocated by the FCC to land mobile use in 1975 with the provision that existing transmitters could continue
indefinitely.
Where
e = terminal voltage, μV, 300 Ω
E = field, μV/m
f1 and f2 = band-edge frequencies, MHz
Many sizes and form factors of receivers are manufactured. Portable personal types include
pocket-sized or hand-held models with picture sizes of 2 to 4 in (5 to 10 cm) diagonal for mono-
chrome and 5 to 6 in (13 to 15 cm) for color powered by either batteries or ac. Conventional cath-
ode ray tubes (CRTS) for picture displays in portable sets have essentially been supplanted by
flat liquid crystal displays and other flat-panel technologies.
Larger screen sizes are available in monochrome where low cost and light weight are prime
requirements. However, except where extreme portability is important, the vast majority of tele-
vision program viewing is in color. The 19-in (48-cm) and 27-in (69-cm) sizes dominate the mar-
ket, although the smaller 13-in (33-cm) size is popular as a second or semiportable set.
Television receiver functions can be broken down into several interconnected blocks. With the
rapidly increasing use of large-scale integrated circuits, the isolation of functions has become
more evident in the design and service of receivers, while at the same time the actual parts count
has dropped dramatically. The typical functional configuration of a receiver using a tri-gun pic-
ture tube, shown in Figure 7.1.1, will serve as a guide for the following description of receiver
design and operation. The discussions of each major block, in turn, are accompanied with more
detailed subblock diagrams.
7.1.2a Tuner Principles

The purpose of the tuner, and the following intermediate amplifier (IF), is to select the desired
radio frequency (RF) signals in a 6 MHz channel, to the exclusion of all other signals, available
from the antenna or cable system and to amplify the signals to a level adequate for demodulation.
Channel selection is accomplished with either mechanically switched and manually tuned cir-
cuits, or in varactor tuners with electrically switched and controlled circuit components. A
mechanical tuner consists of two units, one for the VHF band from 54 to 88 and 174 to 216
MHz, and the other for the UHF band from 470 to 806 MHz. Two separate antenna connections
are provided for the VHF and UHF sections of the tuner.
Varactor tuners, on the other hand, have no moving parts or mechanisms and consequently are
less than a third the volume of their mechanical equivalent. Part of this smaller size is the result
p p
Figure 7.1.1 Fundamental block diagram of a color receiver with a tri-gun picture tube display.
of combining the VHF and UHF circuits on a single printed circuit board in the same shielded
box with a common antenna connection, thus eliminating the need for an outrigger coupling unit.
Selectivity
The tuner bandpass generally is 10 MHz in order to ensure that the picture and sound signals of
the full 6 MHz television channel are amplified with no significant imbalance in levels or phase
distortion by the skirts of the bandpass filters. This bandpass characteristic usually is provided by
three tuned circuits:
• A single-tuned preselector between the antenna input and the RF amplifier stage
• A double-tuned interstage network between the RF and mixer stages
• A single-tuned coupling circuit at the mixer output.
The first two circuits are frequency-selective to the desired channel by varying either or both the
inductance and capacitance. The mixer output is tuned to approximately 44 MHz, the center fre-
quency of the IF channel.
The purpose of the RF selectivity function is to reduce all signals that are outside of the
selected television channel. For example, the input section of VHF tuners usually contains a
p p
Table 7.1.2 Potential VHF Interference Problems
Desired Channel Interfering Signals Mechanism

5 Channel. 11 picture 2 × ch. 5 osc. – ch. 11 pix = IF
6 Channel 13 picture 2 × ch. 6 osc. – ch. 13 pix = IF
7 and 8 Channel 5, FM (98–108 MHz) Ch. 5 pix + FM = ch. 7 and 8
2–6 Channel 5, FM (97–99 MHz) 2 × (FM – ch. 5) = IF
7–13 FM (88–108 MHz) 2 × FM = ch. 7–13
6 FM (89–92 MHz) Ch. 6 pix + FM – ch. 6 osc. = IF
2 6 m amateur (52–54 MHz) 2 × ch. 2 pix – 6 m = ch. 2
2 CB (27 MHz) 2 × CB = ch. 2
5 and 6 CB (27 MHz) 3 × CB = ch. 5 and 6
Table 7.1.3 Potential UHF Interference Problems
Interference Type Interfering Channels Channel 30 Example

IF beat N ± 7, ± 8 22, 23, 37, 38
Intermodulation N ± 2, ± 3, ± 4, ± 5 25–28, 32–35
Adjacent channel N + 1, – 1 29, 31
Local oscillator N±7× 23, 37
Sound image N + 1/6 (2 × 41.25) 44
Picture image N + 1/6 (2 × 45.75) 45
high-pass filter and trap section to reject signals lower than Channel 2 (54 MHz), such as stan-
dard broadcast, amateur, and citizen's band (CB) emissions. In addition, a trap is provided to
reduce FM broadcast signals in the 88 to 108 MHz band. A list of the major interference prob-
lems is tabulated in Table 7.1.2 for VHF channels. In Table 7.1.3 for UHF channels, the formula
for calculation of the interfering channels is given in the second column, and the calculation for a
receiver tuned to Channel 30 is given in the third column.
VHF Tuner
A block diagram of a typical mechanical tuner is shown in Figure 7.1.2. The antenna is coupled
to a tunable RF stage through a bandpass filter to reduce spurious interference signals in the IF
band, or from FM broadcast stations and CB transmitters. Another bandpass filter is provided in
the UHF section for the same purpose. The typical responses of these filters are shown in Figures
7.1.3.a and b.
The RF stage provides a gain of 15 to 20 dB (approximately a 10:1 voltage gain) with a band-
pass selectivity of about 10 MHz between the –3 dB points on the response curve. The response
between these points is relatively flat with a dip of only a decibel or so at the midpoint. There-
fore, the response over the narrower 6 MHz television channel bandwidth is essentially flat.
VHF tuners have a rotary shaft that switches a different set of three or four coils or coil taps
into the circuit at each VHF channel position (2 to 13). The circuits with these switched coils
include:
• RF input preselection
• RF input coupling (single-tuned for monochrome, double-tuned for color)
• RF-to-mixer interstage
p p
Figure 7.1.2 Typical mechanical-tuner configuration.
In the first switch position (Channel 1), the RF stage is disabled and the mixer stage becomes
an IF amplifier stage, centered on 44 MHz for the UHF tuner.
The mixer stage combines the RF signal with the output of a tunable local oscillator to pro-
duce an IF of 43.75 MHz for the picture carrier signal and 42.25 MHz for the sound carrier sig-
nal. The local oscillator signal thus is always 45.75 MHz above that of the selected incoming
picture signal. For example, the frequencies for Channel 2 are listed in Table 7.1.4.
These frequencies were chosen to minimize interference from one television receiver into
another by always having the local-oscillator signal above the VHF channels. Note that the oscil-
lator frequencies for the low VHF channels (2 to 6) are between Channels 6 (low VHF) and 7
(high VHF), and the oscillator frequencies for high VHF fall above these channels.
p p
(a)
(b)
Figure 7.1.3 Filter response characteristics: (a) response of a tuner input FM bandstop filter, (b)
response of a tuner input with CB and IF traps.
Table 7.1.4 Local Oscillator and IF Frequencies
Parameter Channel 2, MHz Channel 6, MHz Channel 7, MHz

Channel width 54–60 82–88 174–180
Local oscillator 101.00 129.00 221.00
Less picture signal 55.25 83.25 175.25
IF picture signal 45.75 45.75 45.75
The picture and sound signals of the full 6 MHz television channel are amplified with no sig-
nificant imbalance in levels or phase distortion by the skirts of the bandpass filters. This band-
pass characteristic usually is provided by three tuned circuits:
• A single-tuned preselector between the antenna input and the RF amplifier stage
• Double-tuned interstage network between the RF and mixer stages
• Single-tuned coupling circuit at the mixer output
The first two are frequency-selective to the desired channel by varying either or both the induc-
tance and capacitance. The mixer output is tuned to approximately 44 MHz, the center frequency
of the IF channel.
p p
Figure 7.1.4 UHF rotary turner. (Courtesy of Philips Consumer Electronics.)
UHF Tuner
The UHF tuner contains a tunable input circuit to select the desired channel, followed by a diode
mixer. As in a VHF tuner, the local oscillator is operated at 45.75 MHz above the selected input
channel signal. The output of the UHF mixer is fed to the mixer of the accompanying VHF tuner,
which functions as an IF amplifier. Selection between UHF and VHF is made by applying power
to the appropriate tuner RF stage.
Mechanical UHF tuners have a shaft that when rotated moves one set of plates of variable air-
dielectric capacitors in three resonant circuits. The first two are a double-tuned preselector in the
amplifier-mixer coupling circuit, and the third is the tank circuit of the local oscillator. In order
to meet the discrete selection requirement of the FCC, a mechanical detent on the rotation of the
shaft and a channel-selector indicator are provided, as illustrated in Figure 7.1.4.
The inductor for each tuned circuit is a rigid metal strip, grounded at one end to the tuner
shield and connected at the other end to the fixed plate of a three-section variable capacitor with
the rotary plates grounded. The three tuned circuits are separated by two internal shields that
divide the tuner box into three compartments.
Tuner-IF Link Coupling

With mechanically switched tuners, it has usually been necessary to place the tuner behind the
viewer control panel and connect it to the IF section, located on the chassis, with a foot or so
length of shielded 50 or 75 Ω coaxial cable. Because the output of the tuner and the input of the
IF amplifier are high-impedance-tuned circuits, for maximum signal transfer, it is necessary to
couple these to the cable with impedance-matching networks.
Two common resonant circuit arrangements are shown in Figure 7.1.5. The low-side capaci-
tive system has a low-pass characteristic that attenuates the local oscillator and mixer harmonic
currents ahead of the IF amplifier. This can be an advantage in controlling local-oscillator radia-
p p
(a )
(b )
Figure 7.1.5 Tuner-to-IF section link coupling: (a) low-side capacitive coupling, (b) low-side induc-
tive coupling.
tion and in reducing the generation of spurious signals in the IF section. On the other hand, the
low-side inductance gives a better termination to the link cable and therefore reduces interstage
cable loss. The necessary bandpass characteristics can be obtained either by undercoupled stag-
ger tuning or by overcoupled synchronous tuning, as illustrated in Figure 7.1.5.
7.1.2b Electronically Tuned Systems

Mechanically tuned television receivers were the mainstay of consumer sets since the beginning
of TV broadcasting. It was not until the 1970s and 1980s that electronically tuned systems
became practical. More recent technological trends include microprocessor-based control of tun-
ing—and nearly all other—functions. Many consumer models have, in fact, completely dis-
pensed with conventional set controls, in favor of using the remote control as the primary user
interface. This being the case, considerable integration of functions can be gained, resulting in
performance improvements and cost savings.
The preceding sections on mechanically tuned receivers are, however, still important because
they form the foundations for the all-electronic tuning systems that are prevalent today.
p p
Figure 7.1.6 Typical varactor-tuned block configuration.
Varactor Tuner
The varactor diode forms the basis for electronic tuning, which is accomplished by a change in
capacitance with the applied dc voltage to the device. One diode is used in each tuned circuit.
Unlike variable air-dielectric capacitors, varicaps have a resistive component in addition to their
capacitance that lowers the Q and results in a degraded noise figure. Therefore, varactor UHF
tuners usually include an RF amplifier stage, making it functionally similar to a VHF tuner. (See
Figure 7.1.6.)
The full UHF band can be covered by a single varicap in a tuned circuit because the ratio of
highest and lowest frequencies in the UHF bands is less than 2:1 (1.7). However, the ratio of the
highest to lowest frequencies in the two VHF bands is over twice (4.07) that of the UHF band.
This is beyond the range that typically can be covered by a tuned circuit using varicaps. This
problem is solved by the use of band switching between the low and high VHF channels. This is
accomplished rather simply by short-circuiting a part of the tuning coil in each resonant tank cir-
cuit to reduce its inductance. The short circuit is provided by a diode that has a low resistance in
the forward-biased condition and a low capacitance in the reverse-biased condition. A typical RF
p p
Figure 7.1.7 VHF tuner band-switching and tuning circuits. +VB = high VHF (active tuning induc-
tors = L3 in parallel with L1 and L2, L10, L11, L21). –VB = low VHF (active tuning inductors = L1 + L2,
L12, LI1 + L13, L14, L21 + L22).
input and oscillator circuit arrangement is shown in Figure 7.1.7. Applying a positive voltage to
V, switches the tuner to high VHF by causing the diodes to conduct and lower the inductance of
the tuning circuits.
Tuning Systems
The purpose of the tuning system is to set the tuner, VHF or UHF, to the desired channel and to
fine-tune the local oscillator for the video carrier from the mixer to be set at the proper IF fre-
quency of 45.75 MHz. In mechanical tuners, this obviously involves an adjustment of the rotary
selector switch and the capacitor knob on the switch shaft. In electronically tuned systems, the dc
tuning voltage can be supplied from the wiper arm of a potentiometer control connected to a
fixed voltage source as shown in Figure 7.1.8a.
Alternatively, multiples of this circuit, as shown in Figure 7.1.8b, can provide preset fine-tun-
ing for each channel. This arrangement most commonly is found in cable-channel selector boxes
supplied with an external cable processor.
In digital systems, such as that shown in Figure 7.1.8c, the tuning voltage can be read as a
digital word from the memory of a keyboard and display station (or remote control circuit). After
conversion from a digital code to an analog voltage, the tuning control voltage is sent to the
tuner.
Figure 7.1.8d shows a microprocessor system using a phase-locked loop to compare a
medium-frequency square-wave signal from the channel selector keyboard, corresponding to a
specific channel, with a signal divided down by 256 from the local oscillator. The error signal
generated by the difference in these two frequencies is filtered and used to correct the tuning
voltage being supplied to the tuner.
p p
(a ) (b)
(c )
(d )
Figure 7.1.8 Varactor-tuned systems: (a) simple potentiometer controlled varactor-tuned system,
(b) multiple potentiometers providing n-channel selection, (c) simplified memory tuning, (d) micro-
processor-based PLL tuning system.
p p
(a )
(b )
Figure 7.1.9 IF amplifier system: (a) typical IF amplifier strip block diagram and gain distribution,
(b) receiver AGC system.
7.1.2c Intermediate-Frequency (IF) Amplifier Requirements

The IF picture and sound carrier frequencies standardized for conventional television receivers
were chosen with prime consideration of possible degradation of the picture from interfering sig-
nals. The picture-carrier frequency is 45.75 MHz and, with the local oscillator above the received
RF signal, the sound carrier frequency is 41.25 MHz. The three factors given greatest emphasis
in the choice of these parameters were:
• Interference from other nontelevision services
• Interference from the fundamental and harmonics of local oscillators in other television
receivers
• Spurious responses from the image signal in the mixer conversion and from harmonics of the
IF signals
Analysis of these relationships indicates the soundness of the choice of 45.75 MHz for the IF
picture carrier. The important advantages include the following:
• No images from the mixer conversion process fall within the VHF band selected by the tuner
except for a negligible interference on the edge of the Channel 7 passband from the image
from another receiver tuned to Channel 6.
• All channels are clear of picture harmonics except that the fourth harmonic of the IF picture
carrier falls near the Channel 8 picture carrier. This can cause a noticeable beat pattern in
another receiver if the offending receiver has not been designed with adequate shielding.
p p
• Local oscillator radiation does not interfere with another receiver on any channel or on any
channel image.
• No UHF signal falls on the image frequency of another station.
On the other hand, it should be pointed out that channels for certain public safety communica-
tions are allocated in the standard IF band. Because these transmitters radiate high power levels,
receivers require thorough shielding of the IF amplifier and IF signal rejection traps in the tuner
ahead of the mixer. In locations where severe cases of interference are encountered, the addition
of a rejection filter at the antenna input may be necessary.
Gain Characteristics
The output level of the picture and sound carriers from the mixer in the tuner is about 200 μV.
The IF section provides amplification to boost this level to about 2 V, which is required for linear
operation of the detector or demodulator stage. This is an overall gain of 80 dB. The gain distri-
bution in a typical IF amplifier using discrete gain stages is shown in Figure 7.1.9a.
Automatic gain control (AGC) in a closed feedback loop is used to prevent overload in the IF,
and the mixer stage as well, from strong signals (see Figure 7.1.9b). Input-signal levels may
range from a few microvolts to several hundred microvolts, thus emphasizing the need for AGC.
The AGC voltage is applied only to the IF for moderate signal levels so that the low-noise RF
amplifier stage in the tuner will operate at maximum gain for relatively weak tuner input signals.
A “delay” bias is applied to the tuner gain control to block application of the AGC voltage except
at very high antenna signal levels. As the antenna signal level increases, the AGC voltage is
applied first to the first and second IF stages. When the input signal reaches about 1 mV, the
AGC voltage is applied to the tuner, as well.
Response Characteristics
The bandpass of the IF amplifier must be wide enough to cover the picture and sound carriers of
the channel selected in the tuner while providing sharp rejection of adjacent channel signals.
Specifically, the upper adjacent-picture carrier and lower adjacent-sound channel must be attenu-
ated 40 and 50 dB, respectively, to eliminate visible interference patterns in the picture. The
sound carrier at 4.5 MHz below the picture carrier must be of adequate level to feed either a sep-
arate sound IF channel or a 4.5 MHz intercarrier sound channel. Furthermore, because in the
vestigial sideband system of transmission the video carrier lower sideband is missing, the
response characteristic is modified from flat response to attenuate the picture carrier by 50 per-
cent (6 dB).
In addition, in color receivers the color subcarrier at 3.58 MHz below the picture carrier must
be amplified without time delay relative to the picture carrier or distortion. Ideally, this requires
the response shown in Figure 7.1.10. Notice that the color IF is wider and has greater attenuation
of the channel sound carrier in order to reduce the visibility of the 920 kHz beat between the
color subcarrier and the sound carrier.
These and other more stringent requirements for color reception are illustrated in Figure
7.1.11. Specifically:
• IF bandwidth must be extended on the high-frequency video side to accommodate the color
subcarrier modulation sidebands that extend to 41.25 MHz (as shown in Figure 7.1.11). The
response must be stable and, except in sets with automatic chroma control (ACC), the
p p
Figure 7.1.10 Ideal IF amplitude response for color and monochrome reception.
Figure 7.1.11 Overall IF bandwidth for color reception.
response must not change with the input signal level (AGC) in order to maintain a constant
level of color saturation.
• More accurate tuning of the received signal must be accomplished in order to avoid shifting
the carriers on the tuner IF passband response. Deviation from their prescribed positions will
alter the ratio of luminance to chrominance (saturation). While this is corrected in receivers
with automatic fine tuning (AFT) and ACC, it can change the time relationship between color
and luminance that is apparent in the color picture as chroma being misplaced horizontally.
• Color subcarrier presence as a second signal dictates greater freedom from overload of ampli-
fier and detector circuits, which can result in spurious intermodulation signals visible as beat
patterns. These cannot be removed by subsequent filtering.
p p
Figure 7.1.12 Surface acoustic wave (SAW) filter response.
• Envelope delay (time delay) of the narrow-band chroma and wide-band luminance signals
must be equalized so that the horizontal position of the two signals match in the color picture.
Surface Acoustic Wave (SAW) Filter

A SAW filter can provide the entire passband shape and adjacent-channel attenuation required
for a television receiver. A typical amplitude response and group delay characteristic are shown
in Figure 7.1.12. The sound carrier (41.25 MHz) attenuation of the SAW filter has been designed
to operate with a synchronous detector, hence the lesser attenuation than in a conventional LC
bandpass filter. The response of an LC discrete stage configuration shows a 60 dB attenuation of
the sound carrier, necessary for suppression of the 920 kHz sound-chroma beat when used with a
diode detector. In addition, with SAW technology it is possible to make wider adjacent-channel
traps, which improve their performance and—in part—makes allowance for the temperature
coefficient of the substrate materials used in SAW filters. This drift may be as great as 59 kHz
per 10°C.
The schematic diagram of a SAW filter IF circuit is shown in Figure 7.1.13. The filter typi-
cally has an insertion loss of 15 to 20 dB and therefore requires a preamplifier to maintain a sat-
isfactory overall receiver SNR.
The SAW filter consists of a piezoelectric substrate measuring 4 to 8 mm by 0.4 mm thick,
upon which has been deposited a pattern of two sets of interleaved aluminum fingers. The width
of the fingers may be on the order of 50 to 500 mm thick and 10 to 20 μm wide. (See Figure
7.1.14.
p p
Figure 7.1.13 Schematic of a block filter/block gain configuration.
Figure 7.1.14 Physical implementation of a SAW filter.
Although quartz and other materials have been used as SAW filter substrates, for television
applications lithium niobate and lithium tantalate are typical. When one set is driven by an elec-
tric signal voltage, an acoustic wave moves across the surface to the other set of fingers that are
connected to the load. The transfer amplitude-frequency response appears as a (sin x)/x (Figure
7.1.15).
A modification to the design in Figure 7.1.14 that gives more optimum television bandpass
and trap response consists of varying the length of the fingers to form a diamond configuration,
illustrated in Figure 7.1.16. This is equivalent to connecting several transducers with slightly dif-
ferent resonant frequencies and bandwidths in parallel. Other modification consists of varying
the aperture spacings, distance between the transducers, and the passive coupler strip patterns in
the space between the transducers.
Gain Block Requirements

Integrated-circuit gain blocks have the same basic requirements as discrete stage IF amplifiers
(see Figure 7.1.17). These are: high gain, low distortion, and a large linear gain-control range
under all operating conditions. One typical differential amplifier configuration used in IC gain
p p
Figure 7.1.15 Response of a uniform interdigitated SAW filter. (After [2].)
blocks that meets these objectives has a gain of nearly 20 dB and a gain-control range of 24 dB.
A direct-coupled cascade of three stages yields an overall gain of 57 dB and a gain-control range
of 64 dB. The gain-control system internal to this IC begins to gain-reduce the third stage at an
IC input level of 100 μV of IF carrier. With increasing input signal level, the third stage gain
reduces to 0 dB and then is followed by the second stage to a similar level, followed by the first
in the same manner. By this means, a noise figure of 7 dB is held constant over an IF input signal
range of 40 dB. The need for a preamplifier ahead of the SAW IF becomes less important when
the IF amplifier noise figure is maintained constant by this cascaded control system.
The high gain and small size of an integrated-circuit IF amplifier places greater importance
on PC layout techniques and ground paths if stability is to be achieved under a wide range of
operating conditions. These considerations also carry over to the external circuits and compo-
nents.
7.1.2d Video Signal Demodulation

The function of the video demodulator is to extract the
picture signal information that has been placed on the
RF carrier as amplitude modulation. The demodulator
receives the modulated carrier signal from the IF
amplifier, which has boosted the peak-to-peak level to
1 or 2 V. The modulation components extend from dc
to 4.5 MHz. The output of the demodulator is fed
directly to the video amplifier.
There are four types of demodulators commonly
used in television receivers:
• Envelope detector
• Transistor detector
Figure 7.1.16 SAW apodized IDT pat-
• Synchronous detector tern.
p p
Figure 7.1.17 Frequency response of SAW filter picture-carrier output. (After [3].)
Figure 7.1.18 Television receiver envelope detector.
• Feedback balanced diode
Envelope Detector
Of the several types of demodulators, the envelope detector is the simplest. It consists of a diode
rectifier feeding a parallel load of a resistor and a capacitor (Figure 7.1.18). In other words, it is a
half-wave rectifier that charges the capacitor to the peak value of the modulation envelope.
Because of the large loss in the diode, a high level of IF voltage is required to recover 1 or 2
volts of demodulated video. In addition, unless the circuit is operated at a high signal level, the
curvature of the diode impedance curve near cutoff results in undesirable compression of peak-
white signals. The requirement for large signal levels and the nonlinearity of detection result in
design problems and certain performance deficiencies, including the following:
• Beat signal products will occur between the color subcarrier (42.17 MHz), the sound carrier
(41.25 MHz), and high-amplitude components in the video signal. The most serious is a 920
p p
(a)
(b)
Figure 7.1.19 Comparison of quadrature distortion in envelope and synchronous detectors: (a)
axis shift, (b) inverted and normal 2T-pulse response. (After [4].)
Hz (color to sound) beat and 60 Hz buzz in sound from vertical sync and peak-white video
modulation.
• Distortion of luminance toward black of as much as 10 percent and asymmetric transient
response. Referred to as quadrature distortion, this characteristic of the vestigial sideband is
aggravated by nonlinearity of the diode. (See Figure 7.1.19.)
• Radiation of the fourth harmonic of the video IF produced by the detection action directly
from the chassis, which can interfere with reception of VHF Channel 8 (180 to 186 MHz).
Even with these deficiencies, the diode envelope detector was used in the majority of the
monochrome and color television receivers dating back to vacuum tube designs up to the era of
discrete transistors.
p p
(a ) (b)
Figure 7.1.20 Transistor detector: (a) schematic diagram, (b) detection characteristic.
Transistor Detector
A transistor biased near collector cutoff and driven with a modulated carrier at an amplitude
greater than the bias level (see Figure 7.1.20) provides a demodulator that can have a gain of 15
or 20 dB over that of a diode. Consequently, less gain is required in the IF amplifier; in fact, in
some receiver designs, the third IF stage has been eliminated. Unfortunately, this is offset by the
same deficiencies in signal detection as the diode envelope detector.
Synchronous Detector
The synchronous detector is basically a balanced rectifier in which the carrier is sampled by an
unmodulated carrier at the same frequency as the modulated carrier. The unmodulated reference
signal is generated in a separate high-Q limiting circuit that removes the modulation. An alterna-
tive system for generating the reference waveform is by means of a local oscillator phase-locked
to the IF signal carrier.
The advantages of synchronous demodulation are:
• Higher gain than a diode detector
• Low level input, which considerably reduces beat-signal generation
• Low level detection operation, which reduces IF harmonics by more than 20 dB
• Little or no quadrature distortion (see Figure 7.1.19), depending upon the lack of residual
phase modulation (purity) of the reference carrier used for detection
• Circuit easily formatted in an IC
7.1.2e Automatic Gain Control (AGC)

Each amplifier, mixer, and detector stage in a television receiver has a number of operating con-
ditions that must be met in order to achieve optimum performance. Specifically, these conditions
include that:
• The input level is greater than the internally generated noise by a factor in excess of the mini-
mum acceptable SNR.
p p
• Input level does not overload the amplifier stages, thus causing a bias shift.
• Bias operates each functional component of the system at its optimum linearity point, that is,
the lowest third-order product for amplifiers, and highest practical second-order for mixers
and detectors.
• Spurious responses in the output are 50 dB below the desired signal.
The function of the automatic gain control system is to maintain signal levels in these stages
at the optimum value over a large range of receiver input levels. The control voltage to operate
the system usually is derived from the video detector or the first video amplifier stage. Common
implementation techniques include the following:
• Average AGC, which operates on the principle of keeping the carrier level constant in the IF.
Changes in modulation of the carrier will affect the gain control, and therefore it is used only
in low-cost receivers.
• Peak or sync-clamp AGC, which compares the video sync-tip level with a fixed dc level. If
the sync-tip amplitude exceeds the reference level, a control voltage is applied to the RF and
IF stages to reduce their gain and thus restore the sync-tip level to the reference level.
• Keyed or gated AGC, which is similar to sync-clamp AGC. The stage where the comparison
of sync-tip and reference signals takes place is activated only during the sync-pulse interval
by a horizontal flyback pulse. Because the AGC circuit is insensitive to spurious noise signals
between sync pulses, the noise immunity is considerably improved over the other two sys-
tems.
AGC Delay
For best receiver SNR, the tuner RF stage is operated at maximum gain for RF signals up to a
threshold level of 1 mV. In discrete amplifier chains, the AGC system begins to reduce the gain
of the second IF stage proportionately as the RF signal level increases from just above the sensi-
tivity level to the second-stage limit of gain reduction (20 to 25 dB). For increasing signals, the
first IF stage gain is reduced. Finally, above the control delay point of 1mV, the tuner gain is
reduced. A plot of the relationships between receiver input RF level and the gain characteristics
of the tuner and IF are shown in Figure 7.1.21a and the noise figure is shown in Figure 7.1.21b.
System Analysis
The interconnection of the amplifier stages (RF, mixer, and IF), the detector, and the lowpass-fil-
tered control voltage is—in effect—a feedback system. The loop gain of the feedback system is
nonlinear, increasing with increasing signal level. There are two principal constraints that
designers must cope with:
• First, the loop gain should be large to maintain good regulation of the detector output over a
wide range of input signal levels. As the loop gain increases, the stability of the system will
decrease and oscillation can occur.
• Second, the ability of a receiver to reject impulse noise is inversely proportional to the loop
gain. Excessive impulse noise can saturate the detector and reduce the RF-IF gain, thereby
causing either a loss in picture contrast or a complete loss of picture. This problem can be
p p
(a ) (b )
Figure 7.1.21 Automatic gain control principles: (a) gain control as a function of input level, (b)
noise figure of RF and IF stages with gain control and the resulting receiver SNR.
alleviated by bandwidth-limiting the video signal fed to the AGC detector, or the use of keyed
or gated AGC to block false input signals (except during the sync pulse time).
A good compromise between regulation of video level and noise immunity is realized with a
loop-gain factor of 20 to 30 dB.
The filter network and filter time constants play an important part in the effectiveness of AGC
operation. The filter removes the 15.750 kHz horizontal sync pulses and equalizing pulses, the
latter in blocks in the 60 Hz vertical sync interval. The filter time constants must be chosen to
eliminate or minimize the following problems:
• Airplane flutter, a fluctuation in signal level caused by alternate cancellation and reinforce-
ment of the received signal by reflections from an airplane flying overhead in the path
between the transmitting and receiving antennas. The amplitude may vary as much as 2 to 1 at
rates from 50 to 150 Hz. If the time constants are too long, especially that of the control volt-
age to the RF stage, the gain will not change rapidly enough to track the fluctuating level of
the signal. The result will be a flutter in contrast and brightness of the picture.
• Vertical sync pulse sag, resulting from the AGC system speed of response being so fast that it
will follow changes in sync pulse energy during the vertical sync interval. Gain increases dur-
ing the initial half-width equalizing pulses, then decreases during the slotted vertical sync
pulse, increases again during the final equalizing pulses, and then returns to normal during
the end of vertical blanking and the next field of picture. Excessive sag can cause loss of
proper interlace and vertical jitter or bounce. Sag can be reduced by limiting the response of
the AGC control loop, or through the use of keyed AGC.
• Lock-out of the received signal during channel switching, caused by excessive speed of the
AGC system. This can result in as much as a 2:1 decrease in pull-in range for the AGC sys-
tem. In keyed AGC systems, if the timing of the horizontal gating pulses is incorrect, exces-
p p
sive or insufficient sync can result at the Table 7.1.5 Typical AFC Closed-Loop Charac-
sync separator, which—in turn—will teristics
upset the operation of the AGC loop. Parameter Value
Pull-in range ± 750 kHz
Hold-in range ± 1.5 MHz
7.1.2f Automatic Frequency Control (AFC) Frequency error for ±500 kHz offset < 50 kHz
Also called automatic fine-tuning (AFT), the
AFC circuit senses the frequency of the pic-
ture carrier in the IF section and sends a correction voltage to the local oscillator in the tuner if
the picture carrier is not on the standard frequency of 45.75 MHz.
Typical AFC systems consist of a frequency discriminator prior to the video detector, a low-
pass filter, and a varactor diode controlling the local oscillator. The frequency discriminator in
discrete transistor IF systems has typically been the familiar balanced-diode type used for FM
radio receivers with the components adjusted for wide-band operation centering on 45.75 MHz.
A small amount of unbalance is designed into the circuit to compensate for the unbalanced side-
band components of vestigial sideband signal characteristics. The characteristics of AFC closed
loops are shown in Table 7.1.5.
In early solid-state designs, the AFC block was a single IC with a few external components.
More current designs include the AFC circuit in the form of a synchronous demodulator on the
same IC die as the other functions of the IF section.
7.1.2g Sound Carrier Separation Systems

Television sound is transmitted as a frequency-modulated signal with a maximum deviation of
± 25 kHz (100 percent modulation) capable of providing an audio bandwidth of 50 to 15,000 Hz.
The frequency of the sound carrier is 4.5 MHz above the RF picture carrier. The basic system is
monaural with dual-channel stereophonic transmission at the option of the broadcaster.
The intercarrier sound system passes the IF picture and sound carriers (45.75 and 41.25
MHz, respectively) through a detector (nonlinear stage) to create the intermodulation beat of 4.5
MHz. The intercarrier sound signal is then amplified, limited, and FM-demodulated to recover
the audio. Block diagrams of intercarrier sound systems are shown in Figures 7.1.22 and 7.1.23.
Figure 7.1.22 A typical monochrome intercarrier sound system.

p p
Figure 7.1.23 An intercarrier sound takeoff ahead of the video detector.
Figure 7.1.24 Split-carrier sound system.
In a discrete component implementation, the intercarrier detector is typically a simple diode

detector feeding a 4.5 MHz resonant network. In an IC IF system, the sound and picture IF sig-
nals are carried all the way to the video detector, where one output port of the balanced synchro-
nous demodulator supplies both the 4.5 MHz sound carrier and the composite baseband video.
The coupling network between the intercarrier detector and the sound IF amplifier usually has
the form of a half-section high-pass filter that is resonant at 4.5 MHz. This form gives greater
attenuation to the video and sync pulses in the frequency range from 4.5 MHz to dc (carrier),
thereby reducing buzz in the recovered audio, especially under the conditions of low picture car-
rier. An alternative implementation uses a piezoelectric ceramic filter that is designed to have a
bandpass characteristic at 4.5 MHz and needs no in-circuit adjustment.
Audio buzz results when video-related phase-modulated components of the visual carrier are
transferred to the sound channel. The generation of incidental carrier phase modulation (ICPM)
can be transmitter-related or receiver-related; however, the transfer to the sound channel takes
place in the receiver. This can occur in the mixer or the detection circuit. Here, a synchronous
p p
Figure 7.1.25 Quasi-parallel sound system.
detector represents little improvement over an envelope diode detector unless a narrow-band fil-
ter is used in the reference channel [5].
Split-carrier sound processes the IF picture and sound carriers as shown in Figure 7.1.24.
Quasi-parallel sound utilizes a special filter such as the SAW filter of Figure 7.1.14 to elimi-
nate the Nyquist slope in the sound detection channel, thereby eliminating a major source of
ICPM generation in the receiver. The block diagram of this system is shown in Figure 7.1.25.
Nearly all sound channels in present-day television receivers are designed as a one- or two-IC
configuration. The single IC contains the functions of sound IF amplifier-limiter, FM detector,
volume control, and audio output. Two-chip systems usually incorporate stereo functionality or
audio power amplification.
Four types of detector circuits typically are used in ICs for demodulation of the FM sound
carrier:
• The quadrature detector, also known as the gated coincidence detector and analog multiplier,
measures the instantaneous phase shift across a reactive circuit as the carrier frequency shifts.
At center frequency (zero deviation) the LC phase network gives a 90° phase shift to V 2 com-
pared with V1. As the carrier deviates, the phase shift changes proportionately to the amount
of carrier deviation and direction.
• The balanced peak detector, which utilizes two peak or envelope detectors, a differential
amplifier, and a frequency-selective circuit or piezoceramic discriminator.
• The differential peak detector, which operates at a low voltage level and does not require
square-wave switching pulses. Therefore, it creates less harmonic radiation than the quadra-
ture detector. In some designs, a low-pass filter is placed between the limiter and peak detec-
tor to further reduce harmonic radiation and increase AM rejection.
• The phase-locked-loop detector, which requires no frequency-selective LC network to
accomplish demodulation. In this system, the voltage-controlled oscillator (VCO) is phase-
locked by the feedback loop into following the deviation of the incoming FM signal. The low-
frequency error voltage that forces the VCO to track is—in fact—the demodulated output.
p p
(a) (b) (c)
Figure 7.1.26 Contrast control circuits: (a) contrast control network in the emitter circuit, (b) equiv-
alent circuit at maximum contrast (maximum gain), (c) minimum contrast.
7.1.2h Video Amplifiers

A range of video signals of 1 to 3 V at the second detector has become standard for many practi-
cal reasons, including optimum gain distribution between the RF, IF, and video sections and dis-
tribution of signal levels so that video detection and sync separation may be effectively
performed. The video amplifier gain and output level are designed to drive the picture tube with
this input level. Sufficient reserve is provided to allow for low percentage modulation and signal
strengths below the AGC threshold.
Picture Controls
A video gain or contrast control and a brightness or background control are provided to allow the
viewer to select the contrast ratio and overall brightness level that produce the most pleasing pic-
ture for a variety of scene material, transmission characteristics, and ambient lighting conditions.
The contrast control usually provides a 4:1 gain change. This is accomplished either by attenua-
tor action between the output of the video stage and the CRT or by changing the ac gain of the
video stage by means of an ac-coupled variable resistor in the emitter circuit. The brightness con-
trol shifts the dc bias level on the CRT to raise or lower the video signal with respect to the CRT
beam cutoff voltage level. (See Figure 7.1.26.)
AC and dc Coupling
For perfect picture transmission and reproduction, it is necessary that all shades of gray are
demodulated and reproduced accurately by the display device. This implies that the dc level
developed by the video demodulator, in response to the various levels of video carrier, must be
carried to the picture tube. Direct coupling or dc restoration is often used, especially in color
receivers where color saturation is directly dependent upon luminance level. (See Figure 7.1.27.)
p p
(a ) (b)
Figure 7.1.27 CRT luminance drive circuit: (a) brightness control in CRT cathode circuit, (b)
brightness control in CRT grid circuit.
Many low cost monochrome designs utilize only ac coupling with no regard for the dc infor-
mation. This eases the high-voltage power supply design as well as simplifying the video cir-
cuitry. These sets will produce a picture in which the average value of luminance remains nearly
constant. For example, a night scene having a level of 15 to 20 IRE units and no peak-white
excursions will tend to brighten toward the luminance level of the typical daytime scene (50 IRE
units). Likewise a full-raster white scene with few black excursions will tend to darken to the
average luminance level condition by use of partial dc coupling in which a high-resistance path
exists between the second detector and the CRT. This path usually has a gain of one-half to one-
fourth that of the ac signal path.
The transient response of the video amplifier is controlled by its amplitude and phase charac-
teristics. The low-frequency transient response, including the effects of dc restoration, if used, is
measured in terms of distortion to the vertical blanking pulse. Faithful reproduction requires that
the change in amplitude over the pulse duration, usually a decrease from initial value called sag
or tilt, be less than 5 percent. In general, there is no direct and simple relationship between the
sag and the lower 3 dB cutoff frequency. However, lowering the 3 dB cutoff frequency will
reduce the tilt, as illustrated in Figure 7.1.28.
Low-Frequency Response Requirements

The effect of inadequate low-frequency response appears in the picture as vertical shading. If the
response is so poor as to cause a substantial droop of the top of the vertical blanking pulse, then
incomplete blanking of retrace lines can occur.
It is not necessary or desirable to extend the low-frequency response to achieve essentially
perfect LF square-wave reproduction. First, the effect of tilt produced by imperfect LF response
is modified if dc restoration is employed. Direct-current restorers, particularly the fast-acting
p p
(a ) (b)
Figure 7.1.28 Video stage low-frequency response: (a) square-wave output showing tilt, (b) RC
time constant circuits in the common-emitter stage that affect low-frequency response.
variety, substantially reduce tilt, and their effect must be considered in specifying the overall
response. Second, extended LF response makes the system more susceptible to instability and
low-frequency interference. Current coupling through a common power supply impedance can
produce the low-frequency oscillation known as “motorboating.” Motorboating is not usually a
problem in television receiver video amplifiers because they seldom employ the number of
stages required to produce regenerative feedback, but in multistage amplifiers the tendency
toward motorboating is reduced as the LF response is reduced.
A more commonly encountered problem is the effect of airplane flutter and line bounce.
Although a fast-acting AGC can substantially reduce the effects of low-frequency amplitude
variations produced by airplane reflections, the effect is so annoying visually as to warrant a sac-
rifice in LF response to bring about further reduction. A transient in-line voltage amplitude,
commonly called a line bounce, also can produce an annoying brightness transient that can simi-
larly be reduced through a sacrifice of LF response. Special circuit precautions against line
bounce include the longest possible power supply time constant, bypassing the picture tube elec-
trodes to the supply instead of ground, and the use of coupling networks to attenuate the response
sharply below the LF cutoff frequency. The overall receiver response is usually an empirically
determined compromise.
The high-frequency transient characteristic is usually expressed as the amplifier response to
an ideal input voltage or current step. This response is shown in Figure 7.1.29 and described in
the following terms:
• Rise time τ R is the time required for the output pulse to rise from 10 to 90 percent of its final
(steady-state) value.
• Overshoot is the amplitude by which the transient rise exceeds its final value, expressed as a
percentage of the final value.
• Preshoot is the amplitude by which the transient oscillatory output waveform exceeds its ini-
tial value.
• Smear is an abnormally slow rise as the output wave approaches its final value.
p p
Figure 7.1.29 Video amplifier response to a step input.
• Ringing is an oscillatory approach to the final value.

In practice, rise times of 0.1 to 0.2 μs are typical. Overshoot, smear, and ringing amplitude are
usually held to 5 percent of the final value, and ringing is restricted to one complete cycle.
7.1.2i Color Receiver Luminance Channel

Suppression of the chroma subcarrier is necessary to reduce objectionable dot crawl in and
around colored parts of the picture, as well as reduce the distortion of luminance levels resulting
from the nonlinear transfer characteristic of CRT electron guns. Traditionally, a simple high-Q
LC trap, centered around the color subcarrier, has been used for rejection, but this necessitates a
trade off between luminance channel bandwidth and the stop band for the chroma sidebands.
Luminance channel comb filtering largely avoids this compromise and is one reason why it is
commonly used.
The luminance channel also provides the time delay required to correct the time delay regis-
tration with the color difference signals, which normally incur delays in the range of from 300 to
1000 ns in their relatively narrow-bandwidth filters.
While delay circuits having substantially flat amplitude and group delay out to the highest
baseband frequency of interest can and have been used, this is not necessarily required nor desir-
able for cost-effective overall design. Because the other links in the chain (i.e., tuner, IF, traps at
4.5 and 3.58 MHz, and CRT driver stage) may all contribute significant linear distortion individ-
ually, it is frequently advantageous to allow these distortions to occur and use the delay block as
an overall group delay and/or amplitude equalizer.
Although it is well known that, for “distortionless” transmission, a linear system must possess
both uniform amplitude and group delay responses over the frequency band of interest, the limi-
tations of a finite bandwidth lead to noticeably slower rise and fall times, rendering edges less
sharp or distinct. By intentionally distorting the receiver amplitude response and boosting the
relative response to the mid and upper baseband frequencies to varying degrees, both faster rise
and fall times can be developed along with enhanced fine detail. If carried too far, however,
p p
objectionable outlining can occur, especially to those transients in the white direction. Further-
more, the visibility of background noise is increased.
For several reasons—including possible variations in transient response of the transmitted
signal, distortion due to multipath, antennas, receiver tolerances, SNR, and viewer preference—
it is difficult to define a fixed response at the receiver that is optimum under all conditions.
Therefore, it is useful to make the amplitude response variable so it can be controlled to best suit
the individual situation. Over the range of adjustment, it is assumed that the overall group delay
shall remain reasonably flat across the video band. The exact shape of the amplitude response is
directly related to the desired time domain response (height and width of preshoot, overshoot,
and ringing) and chroma subcarrier sideband suppression.
Because the peaked signal later operates on the nonlinear CRT gun characteristics, large
white preshoots and overshoots can contribute to excessive beam currents, which can cause CRT
spot defocusing. To alleviate this, circuits have been developed that compress large excursions of
the peaking component in the white direction. For best operation, it is desirable that the signals
being processed have equal preshoot and overshoot.
Low level, high frequency noise in the luminance channel can be removed by a technique
called coring. One coring technique involves nonlinearly operating on the peaking or edge-
enhancement signal, discussed earlier in this section. The peaking signal is passed through an
amplifier having low or essentially no gain in the mid-amplitude range. When this modified
peaking signal is added to the direct video, the large transitions will be enhanced, but the small
ones (noise) will not be, giving the illusion that the picture sharpness has been increased while
the noise has been decreased.
7.1.2j Chroma Subcarrier Processing

In the equiband chroma system, typical of practically all consumer receivers, the chroma ampli-
fier must be preceded by a bandpass filter network that complements the chroma sideband
response produced by the tuner and IF. Frequencies below 3 MHz also must be attenuated to
reduce not only possible video cross-color disturbances but also crosstalk caused by the lower-
frequency I channel chroma information.
Another requirement is that the filter have a gentle transition from passband to stop band in
order to impart a minimum amount of group delay in the chroma signal, which then must be
compensated by additional group delay circuitry in the luminance channel. The fourth-order
high-pass filter is a practical realization of these requirements.
As described previously, this stage also serves as the chroma gain control circuit. The usual
implementation in an IC consists of a differential amplifier having the chroma signal applied to
the current source. The gain-control dc voltage is applied to one side of the differential pair to
divert the signal current away from the output side.
Burst Separation
Complete separation of the color synchronizing burst from video requires time gating. The gate
requirements are largely determined by the horizontal sync and burst specifications, illustrated in
Figure 7.1.30. It is essential that all video information be excluded. It is also desirable that both
the leading and trailing edges of burst be passed so that the complementary phase errors intro-
duced at these points by quadrature distortion average to zero. Widening the gate pulse to mini-
mize the required timing accuracy has a negligible effect on the noise performance of the
p p
Figure 7.1.30 Horizontal blanking interval specification.
reference system and may be beneficial in the presence of echoes. The ≈ 2 μs spacing between
the trailing edges of burst and horizontal blanking determines the total permissible timing varia-
tion. Noise modulation of the gate timing should not permit noise excursions to encroach upon
the burst because the resulting cross modulation will have the effect of increasing the noise
power delivered to the reference system.
Burst Gating Signal Generation

The gate pulse generator must provide both steady-state phase accuracy and reasonable noise
immunity. The horizontal flyback pulse has been widely used for burst gating because it is
derived from the horizontal scan oscillator system, which meets the noise immunity require-
ments and, with appropriate design, can approximate the steady-state requirements. A further
improvement in steady-state phase accuracy can be achieved by deriving the gating pulse directly
from the trailing edge of the horizontal sync pulse. This technique is utilized in several chroma
system ICs.
The burst gate in conventional discrete component circuits has the form of a conventional
amplifier that is biased into linear conduction only during the presence of the gating pulse. In the
IC implementation, the complete chroma signal is usually made available at one input of the
automatic phase control (APC) burst-reference phase detector. The gating pulse then enables the
phase detector to function only during the presence of burst.
Color Subcarrier Reference Separation

The color subcarrier reference system converts the synchronizing bursts to a continuous carrier
of identical frequency and close phase tolerance. Theoretically, the long-term and repetitive
phase inaccuracies should be restricted to the same value, approximately ± 5°. Practically, if
transmission variations considerably in excess of this value are encountered, and if operator con-
trol of phase (“hue control”) is provided, the long-term accuracy need not be so great. Somewhat
p p
Figure 7.1.31 Automatic phase control (APC) system block diagram.
greater instantaneous inaccuracies can be tolerated in the presence of thermal noise so that an
rms phase error specification of 5 to 10° at an S/N of unity may be regarded as typical.
Reference Systems
Three types of reference synchronization systems have been used:
• Automatic phase control of a VCO
• Injection lock of a crystal
• Ringing of a crystal filter
Best performance can be achieved by the APC loop. In typical applications, the figure of
merit can be made much smaller (better) for the APC loop than for the other systems by making
the factor (1/y) + m have a value considerably less than 1, even as small as 0.1. The parts count
for each type system, at one time much higher for the APC system, is no longer a consideration
because of IC implementations where the oscillator and phase detector are integrated and only
the resistors and capacitors of the filter network and oscillator crystal are external.
The APC circuit is a phase-actuated feedback system consisting of three functional compo-
nents:
• A phase detector
• Low-pass filter
• DC voltage-controlled oscillator
The overall system is illustrated in Figure 7.1.31 The characteristics of these three units
define both the dynamic and static loop characteristics and, hence, the overall system perfor-
mance.
The phase detector generates a dc output E whose polarity and amplitude are proportional to
the direction and magnitude of the relative phase difference dφ between the oscillator and syn-
chronizing (burst) signals.
The VCO is an IC implementation that requires only an external crystal and simple phase-
shift network. The oscillator can be shifted ± 45° by varying the phase-control voltage. This
leads to symmetrical pull-in and hold-in ranges.
p p
Chroma Demodulation
The chroma signal can be considered to be made up of two amplitude-modulated carriers having
a quadrature phase relationship. Each of these carriers can be individually recovered through the
use of a synchronous detector. The reference input to the demodulator is that phase which will
demodulate only the I signal. The output contains no Q signal.
Demodulation products of 7.16 MHz in the output signal can contribute to an optical interfer-
ence moiré pattern in the picture. This is related to the line geometry of the shadow mask. The
7.16 MHz output also can result in excessively high line-terminal radiation from the receiver. A
first-order low-pass filter with cutoff of 1 to 2 MHz usually provides sufficient attenuation. In
extreme cases, an LC trap may be required.
Demodulation Axis
Over the double sideband region (± 500 kHz around the subcarrier frequency) the chrominance
signal can be demodulated as pure quadrature AM signals along either the I and Q axis or R–Y
and B–Y axis. The latter signal leads to a simpler matrix for obtaining the color drive signals R,
G, and B.
Modern practice has moved away from the classic demodulation angles for two main reasons:
• Receiver picture tube phosphors have been modified to yield greater light output and can no
longer produce the original NTSC primary colors. The chromaticity coordinates of the pri-
mary colors, as well as the RGB current ratios to produce white balance, vary from one CRT
manufacturer to another.
• The white-point setup has, over the years, moved from the cold 9300 K of monochrome tubes
to the warmer illuminant C and D65, which produce more vivid colors, more representative of
those that can be seen in natural sunlight.
7.1.2k RGB Matrixing and CRT Drive

Because RGB primary color signals driving the display are required as the end output in a televi-
sion receiver, it is necessary to combine or “matrix” the demodulated color-difference signals
with the luminance signal. Several circuit configurations can be used to accomplish this task.
In the color-difference drive matrixing technique, R–Y, G–Y, and B–Y signals are applied to
respective control grids of the CRT while luminance is applied to all three cathodes; the CRT
thereby matrixes the primary colors. This approach has the advantage that gray scale is not a
function of linearity matching among the three channels because at any level of gray, the color-
difference driver stages are at the same dc level. Also, because the luminance driver is common,
any dc drift shows up only as a brightness shift. Luminance channel frequency response unifor-
mity is ensured by the common driver.
RGB drive, wherein RGB signals are applied to respective cathodes and G1 is dc biased,
requires less drive and has none of the potential color fidelity errors of the color-difference sys-
tem. RGB drive, however, places higher demands on the drive amplifiers for linearity, frequency
response, and dc stability, plus requiring a matrixing network in the amplifier chain.
Low-level RGB matrixing and CRT drive are commonly used, especially with CRT devices
that have unitized guns in which the common G1 and G2 elements require differential cathode
bias adjustments and drive adjustments to yield gray-scale tracking. In this technique, RGB sig-
p p
Figure 7.1.32 Simplified R, G, and B output stage.
nals are matrixed at a level of a few volts and then amplified to a higher level (100 to 200 V) suit-
able for CRT cathode drive.
Direct current stability, frequency response, and linearity of the three stages, even if some-
what less than ideal, should be reasonably well matched to ensure overall gray-scale tracking.
Bias and gain adjustments should be independent in their operation, rather than interdependent,
and should minimally affect those characteristics listed previously in this section.
Figure 7.1.32 illustrates a simple example of one of the three CRT drivers. If the amplifier
black-level bias voltage equals the black level from the RGB decoder, drive adjustment will not
change the amplifier black-level output voltage level or affect CRT cutoff. Furthermore, if RB >>
R E, drive level will be independent of bias setting. Note also that frequency response-determin-
ing networks are configured to be unaffected by adjustments.
Frequently, the shunt peaking coil can be made common to all three channels, because differ-
ences between the channels are predominantly color-difference signals of relatively narrow band-
width. Although the frequency responses could be compensated to provide the widest possible
bandwidth, this is usually not necessary when the frequency response of preceding low-level
luminance processing (especially the peaking stage) is factored in. One exception in which out-
put stage bandwidth must be increased to its maximum is in an application, television receiver or
video monitor, where direct RGB inputs are provided for auxiliary services, such as computers
or DVDs.
Comb Filter Processing

The frequency spectrum of a typical NTSC composite video signal is shown in Figure 7.1.33a. A
comb filter, characterized by 100 percent transmission of desired frequencies of a given channel
and substantially zero transmission for the undesired interleaved signal spectrum, can effectively
separate chroma and luminance components from the composite signal. Such a filter can be eas-
p p
(a)
(b)
(c)
Figure 7.1.33 Color system filtering: (a) frequency spectrum of NTSC color system, showing inter-
leaving of signals, (b) simplified 1H delay comb filter block diagram, (c) chrominance VC and lumi-
nance VL outputs of the comb filter.
p p
ily made, in principle, by delaying the composite video signal one horizontal scan period (63.555
μs in NTSC-M) and adding or subtracting to the undelayed composite video signal (Figure
7.1.33b and c).
The output of the sum channel will have frequencies at f (horizontal), and all integral multi-
ples thereof reinforce in phase, while those interleaved frequencies will be out of phase and will
cancel. This can be used as the luminance path. The difference channel will have integral fre-
quency multiples cancel while the interleaved ones will reinforce. This channel can serve as the
chrominance channel. The filter characteristic and interleaving are shown in Figure 7.1.33c.
Automatic Circuits
The relative level of the chroma subcarrier in the incoming signal is highly sensitive to transmis-
sion path disorders, thereby introducing objectionable variations in saturation. These can be
observed between one received signal and another or over a period of time on the same channel
unless some adaptive correction is built into the system. The color burst reference, transmitted at
40 IRE units peak-to-peak, is representative of the same path distortions and is normally used as
a reference for automatic gain controlling the chroma channel. A balanced peak detector or syn-
chronous detector, having good noise rejection characteristics, detects the burst level and pro-
vides the control signal to the chroma gain-controlled stage.
Allowing the receiver chroma channel to operate during reception of a monochrome signal
will result in unnecessary cross color and colored noise, made worse by the ACC increasing the
chroma amplifier gain to the maximum. Most receivers, therefore, cut off the chroma channel
system when the received burst level goes below approximately 5 to 7 percent. Hysteresis has
been used to minimize the flutter or threshold problem with varying signal levels.
Burst-referenced ACC systems perform adequately when receiving correctly modulated sig-
nals with the appropriate burst-to-chroma ratio. Occasionally, however, burst level may not bear a
correct relation to its accompanying chroma signal, leading to incorrectly saturated color levels.
It has been determined that most viewers are more critical of excessive color levels than insuffi-
cient ones. Experience has shown that when peak chroma amplitude exceeds burst by greater
than 2/1, limiting of the chroma signal is helpful. This threshold nearly corresponds to the ampli-
tude of a 75 percent modulated color bar chart (2.2/1). At this level, negligible distortion is intro-
duced into correctly modulated signals. Only those that are nonstandard are affected
significantly. The output of the peak chroma detector also is sent to the chroma gain-controlled
stage.
Tint
One major objective in color television receiver design is to minimize the incidence of flesh-tone
reproduction with incorrect hue. Automatic hue-correcting systems can be categorized into two
classes:
• Static flesh-tone correction, achieved by selecting the chroma demodulating angles and gain
ratios to desensitize the resultant color-difference vector in the flesh-tone region (+I axis).
The demodulation parameters remain fixed, but the effective Q axis gain is reduced. This has
the disadvantage of distorting hues in all four quadrants.
• Dynamic flesh-tone corrective systems, which can adaptively confine correction to the region
within several degrees of the positive I axis, leaving all other hues relatively unaffected. This
is typically accomplished by detecting the phase of the incoming chroma signal and modulat-
p p
Figure 7.1.34 Vertical interval reference (VIR) signal. Note that the chrominance and the program
color burst are in phase.
ing the phase angle of the demodulator reference signal to result in an effective phase shift of
10 to 15° toward the I axis for a chroma vector that lies within 30° of the I axis. This approach
produces no amplitude change in the chroma. In fact, for chroma saturation greater than 70
percent, the system is defeated on the theory that the color is not a flesh tone.
A simplification in circuitry can be achieved if the effective correction area is increased to the
entire positive-I 180° sector. A conventional phase detector can be utilized and the maximum
correction of approximately 20° will occur for chroma signals having phase of ± 45° from the I
axis. Signals with phase greater or less than 45° will have increasingly lower correction values,
as illustrated in Figure 7.1.34.
The vertical interval reference (VIR) signal, as shown in Figure 7.1.34, provides references
for black level, luminance, and—in addition to burst—a reference for chroma amplitude and
phase. While originally developed to aid broadcasters, it has been employed in television receiv-
ers to correct for saturation and hue errors resulting from transmitter or path distortion errors.
7.1.2l Scanning Synchronization

The scan-synchronizing signal, consisting of the horizontal and vertical sync pulses, is removed
from the composite video signal by means of a sync-separator circuit. The classic approach has
been to ac-couple the video signal to an overdriven transistor amplifier (illustrated in Figure
7.1.35) that is biased off for signal levels smaller than VCO and saturates for signal levels greater
than V sat, a range of approximately 0.1 V.
p p
(a )
(b )
Figure 7.1.35 Sync separator: (a) typical circuit, (b) sync waveform.
Figure 7.1.36 Picture scanning section functional block diagram.

p p
Vertical synchronizing informa-

tion can be recovered from the output
of the sync separator by the technique
of integration. The classic two-section
RC integrator provides a smooth,
ramp waveform that corresponds to
the vertical sync block as shown in
Figure 7.1.36. The ramp is then sent
to a relaxation or blocking oscillator,
which is operating with a period
slightly longer than the vertical frame
period. The upper part of the ramp
will trigger the oscillator into conduc-
tion to achieve vertical synchroniza-
tion. The oscillator will then reset and
wait for the threshold level of the next
Figure 7.1.37 Modified integrator followed by a differ-
ramp. The vertical-hold potentiome-
entiator.
ter controls the free-running fre-
quency or period of the vertical
oscillator.
One modification to the integrator design is to reduce the integration (speed up the time con-
stants) and provide for some differentiation of the waveform prior to applying it to the vertical
oscillator (shown in Figure 7.1.37). Although degrading the noise performance of the system,
this technique provides a more certain and repeatable trigger level than the full integrator,
thereby leading to an improvement in interlace over a larger portion of the hold-in range. A sec-
ond benefit is to provide a more stable vertical lock when receiving signals that have distorted
vertical sync waveforms. These can be generated by video cassette recorders in nonstandard
playback modes, such as fast/slow forward, still, and reverse.
Vertical Countdown Synchronizing Systems

A vertical scan system using digital logic elements can be based upon the frequency relationship
of 525/2 that exists between the horizontal and vertical scans. Such a system can be considered
to be phase-locked for both horizontal scan and vertical scan, which will result in improved noise
immunity and picture stability. Vertical sync is derived from a pulse train having a frequency of
twice the horizontal rate. The sync is therefore precisely timed for both even and odd fields,
resulting in excellent interlace. This system requires no hold control.
A block diagram of one design is shown in Figure 7.1.38. The 31.5 kHz clock input is con-
verted to horizontal drive pulses by a divide-by-2 flip-flop. A two-mode counter, set for 525
counts for standard interlaced signals and 541 for noninterlaced signals, produces the vertical
output pulse. The choice of 541 allows the counter to continue until the arrival of vertical sync
from the composite video waveform. The actual vertical sync then resets the counter.
Other systems make use of clock frequencies of 16 and 32 times the horizontal frequency.
Low-cost crystal or ceramic resonator-controlled oscillators can be built to operate at these fre-
quencies. As in the first system, dual-mode operation is necessary in order to handle standard
interlaced and nonstandard sync waveforms. An exact 525/2 countdown is used with interlaced
p p
Figure 7.1.38 Vertical countdown sync processor, using a 2fH clock.
signals. For noninterlaced signals, the systems usually operate in a free-running mode with injec-
tion of the video-derived sync pulse causing lock.
A critical characteristic, and probably the most complex portion of any countdown system, is
the circuitry that properly adjusts for nonstandard sync waveforms. These can be simple 525
noninterlaced fields, distorted vertical sync blocks, blocks having no horizontal serrations, fields
with excessive or insufficient lines, and a combination of the above.
Impulse Noise Suppression

The simplest type suppression that will improve the basic circuit in a noise environment of
human origin consists of a parallel resistor and capacitor in series with the sync charging capaci-
tor. Variations of this simple circuit, which include diode clamps and switches, have been devel-
oped to speed up the noise suppression performance while avoiding excessive tilt in the sync
output during the vertical block.
More complex solutions to the noise problem include the noise canceler and noise gate. The
canceler monitors the video signal between the second detector and the sync separator. A noise
spike that exceeds the sync or pedestal level is inverted and added to the video after the isolation
resistor. This action cancels that part of the noise pulse which would otherwise produce an output
from the sync separator.
For proper operation, the canceler circuit must track the sync-tip amplitude from a strong RF
signal to the fringe level. The noise gate, in a similar manner, recognizes a noise pulse of large
amplitude and prevents the sync separator from conducting, either by applying reverse bias or by
opening a transistor switch in the emitter circuit.
Horizontal Automatic Phase Control

Horizontal scan synchronization is accomplished by means of an automatic phase control (APC)
loop, with theory and characteristics similar to those used in the chroma reference system. Input
signals to the horizontal APC loop are sync pulses from the sync separator and horizontal fly-
back pulses, which are integrated to form a sawtooth waveform. The phase detector compares the
p p
phase (time coincidence) of these two wave-

forms and sends a dc-coupled low-pass fil-
tered error signal to the voltage-controlled
oscillator to cause the frequency to shift in
the direction of minimal phase error between
the two input signals.
The recovery of the APC loop to a step
transient input involves the parameters of the
natural resonant frequency, as well as the
amount of overshoot permitted after the cor-
rect phase has been reached. Both of these
characteristics can be evaluated by use of a
jitter generator, which creates a time base
error between alternate fields. The resultant Figure 7.1.39 The pattern on a receiver screen
sync error and system dynamics can be seen displaying a vertical line with the even and odd
in the picture display of the receiver shown in fields offset by n ms.
Figure 7.1.39. This type of disturbance can
also occur when the receiver is operated from a signal that does not have horizontal slices in the
vertical sync block.
Phase detector circuitry in the form of classic double-diode bridge detectors, in either the bal-
anced or unbalanced configuration, can sometimes be found in low-cost receivers. Integrated cir-
cuits commonly utilize a gated differential amplifier (Figure 7.1.40.) in which the common feed
from the current source is driven by the sync pulse, and the sawtooth waveform derived from the
flyback pulse is applied to one side of the differential pair.
Vertical Scanning
Class-B vertical circuits consist essentially of an audio frequency amplifier with current feed-
back. This approach maintains linearity without the need for an adjustable linearity control. Yoke
impedance changes caused by temperature variations will not affect the yoke current, thus, a
thermistor is not required. The current-sensing resistor must, of course, be temperature stable.
An amplifier that uses a single NPN and a single PNP transistor in the form of a complementary
output stage is given in Figure 7.1.41. Quasi-complementary, Darlington outputs and other com-
mon audio output stage configurations also can be used.
Establishing proper dc bias for the output stages is critical. Too little quiescent current will
result in crossover distortion that will impose a faint horizontal white line in the center of the pic-
ture even though the distortion may not be detectable on an oscilloscope presentation of the yoke
current waveform. Too much quiescent current results in excessive power dissipation in the out-
put transistors.
Retrace Power Reduction

The voltage required to accomplish retrace results in a substantial portion of the power dissipa-
tion in the output devices. The supply voltage to the amplifier and corresponding power dissipa-
tion can be reduced by using a retrace switch or flyback generator circuit to provide additional
supply voltage during retrace. One version of a retrace switch is given in Figure 7.1.42. During
the trace time, the capacitor is charged to a voltage near the supply voltage. As retrace begins, the
p p
(a )
(b )
Figure 7.1.40 Gated phase detector in an IC implementation: (a) circuit diagram, (b) pulse-timing
waveforms.
voltage across the yoke goes positive, thus forcing QR into saturation. This places the cathode of
CR at the supply potential and the anode at a level of 1.5 to 2 times the supply, depending upon
the values of R R and CR.
Horizontal Scanning
The horizontal scan system has two primary functions:
• It provides a modified sawtooth-shaped current to the horizontal yoke coils to cause the elec-
tron beam to travel horizontally across the face of the CRT.
• It provides drive to the high-voltage or flyback transformer to create the voltage needed for
the CRT anode.
p p
Figure 7.1.41 A class-B vertical output stage.
Figure 7.1.42 Simplified schematic of a retrace switch.
Frequently, low-voltage supplies also are derived from the flyback transformer. The major
components of the horizontal-scan section consist of a driver stage, horizontal output device
(either bipolar transistor or SCR), yoke current damper diode, retrace capacitor, yoke coil, and
flyback transformer, as illustrated in Figure 7.1.43.
During the scan or retrace interval, the deflection yoke may be considered a pure inductance
with a dc voltage impressed across it. This creates a sawtooth waveform of current (see Figure
7.1.44). This current flows through the damper diode during the first half scan. It then reverses
direction and flows through the horizontal output transistor collector. This sawtooth-current
p p
Figure 7.1.43 Simplified horizontal scan circuit.
waveform deflects the electron beam across the face of the picture tube. A similarly shaped cur-
rent flows through the primary winding of the high-voltage output transformer.
At the beginning of the retrace interval, the transformer and yoke inductances transfer energy
to the retrace-tuning capacitor and the accompanying stray capacitances, thereby causing a half
sine wave of voltage to be generated. This high-energy pulse appears on the transistor collector
and is stepped up, via the flyback transformer, to become the high voltage for the picture tube
anode. Finally, at the end of the cycle, the damper diode conducts, and another horizontal scan is
started.
7.1.2m High Voltage Generation

High voltage in the range 8 to 16 kV is required to supply the anode of monochrome picture
tubes. Color tubes have anode requirements in the range 20 to 30 kV and focus voltage require-
ments of 3 to 12 kV. The horizontal flyback transformer is the common element in the generation
of these high voltages. There are three common variations of this design:
• The flyback with a half-wave rectifier
• Flyback driving a voltage multiplier
• Voltage-adding integrated flyback
Consider the simplified horizontal-scan circuit shown in Figure 7.1.43. The voltage at the top
of the high voltage (HV) winding consists of a series of pulses delivered during retrace from the
stored energy in the yoke field. The yoke voltage pulses are then multiplied by the turns ratio of
p p
Figure 7.1.44 Horizontal scan waveform timing diagrams.
the HV winding to primary winding. The peak voltage across the primary during retrace is given
by
T
V p ( pk ) = E in 0.8 ⎛⎝ 1.79 + 1.57 ------trace
-------------⎞
⎠
(7.1.2)
Tretrace
Where:
E in = supply voltage (B+)
0.8 = accounts for the pulse shape factor with third harmonic tuning
Ttrace = trace period, ≈ 52.0 μs
Tretrace = retrace period, ≈ 11.5 μs
p p
Flyback with Half-Wave Rectifier

The most common HV supply for small-screen monochrome and color television receivers uses
a direct half-wave rectifier circuit. The pulses at the top of the HV winding are rectified by the
single diode, or composite of several diodes in series. The charge is then stored in the capaci-
tance of the anode region of the picture tube. Considerable voltage step-up is required from the
primary to the HV winding. This results in a large value of leakage inductance for the HV wind-
ing, which decreases its efficiency as a step-up transformer.
Harmonic tuning of the HV winding improves the efficiency by making the total inductance
and the distributed capacitance of the winding plus the CRT anode capacitance resonate at an
odd harmonic of the flyback pulse frequency. For the single-rectifier circuit, usually the third
harmonic resonance is most easily implemented by proper choice of winding configuration,
which results in appropriate leakage inductance and distributed capacitance values. This will
result in HV pulse waveforms, which will give an improvement in the HV supply regulation
(internal impedance) as well as a reduction in the amplitude of ringing in the current and voltage
waveforms at the start of scan.
Voltage Multiplier Circuits

A voltage multiplier circuit consists of a combination of diodes and capacitors connected in such
a way that the dc output voltage is greater than the peak amplitude of the input pulse. The tripler
was a common HV system for color television receivers requiring anode voltages of 25 to 30 kV
dating back to the 1970s. The considerably reduced pulse voltage required from the HV winding
resulted in a flyback transformer with a more tightly coupled HV winding.
Integrated Flybacks
Most medium- and large-screen color receivers utilize an integrated flyback transformer in
which the HV winding is segmented into three or four parallel-wound sections. These sections
are series-connected with a diode between adjacent segments. These diodes are physically
mounted as part of the HV section. The transformer is then encapsulated in high-voltage polyes-
ter or epoxy.
Two HV winding construction configurations also have been used. One, the layer or solenoid-
wound type, has very tight coupling to the primary and operates well with no deliberate har-
monic tuning. Each winding (layer) must be designed to have balanced voltage and capacitance
with respect to the primary. The second, a bobbin or segmented-winding design, has high leak-
age inductance and usually requires tuning to an odd harmonic (e.g., the ninth). Regulation of
this construction is not quite as good as the solenoid-wound primary winding at a horizontal-fre-
quency rate. The +12 V, +24 V, +25 V, and –27 V supplies are scan-rectified. The +185 V, over-
voltage sensing, focus voltage, and 25 kV anode voltage are derived by retrace-rectified supplies.
The CRT filament is used directly in its ac mode.
Flyback-generated supplies provide a convenient means for isolation between different
ground systems, as required for an iso-hot chassis.
Power Supplies
Most receivers use the flyback pulse from the horizontal transformer as the source of power for
the various dc voltages required by the set. Using the pulse waveform at a duty cycle of 10 or 15
p p
Figure 7.1.45 Auxiliary power sources derived from the horizontal output transformer. (Courtesy
of Philips.)
percent, by proper winding direction and grounding of either end of the winding, several differ-
ent voltage sources can be created.
Scan rectified supplies are operated at a duty cycle of approximately 80 percent and are thus
better able to furnish higher current loads. Also, the diodes used in such supplies must be capable
of blocking voltages that are nine to ten times larger than the level they are producing. Diodes
having fast-recovery characteristics are used to keep the power dissipation at a minimum during
the turn-off interval because of the presence of this high reverse voltage.
A typical receiver system containing the various auxiliary power supplies derived from fly-
back transformer windings is shown in Figure 7.1.45. Transistor Q452 switches the primary
winding at a horizontal frequency rate. The +12 V, +24 V, +25 V, and –27 V supplies are scan-
rectified. The +185 V, overvoltage sensing, focus voltage, and 25 kV anode voltage are derived
by retrace-rectified supplies. The CRT filament is used directly in its ac mode.
As noted in the previous section, flyback-generated supplies provide a convenient means for
isolation between different ground systems. Figure 7.1.46 shows the block diagram of such a
television receiver power supply system [6, 7].
p p
Figure 7.1.46 Color receiver power supply using a switched-mode regulator system. (Courtesy of
Philips.)
7.1.3 References
1. FCC Regulations, 47 CFR, 15.65, Federal Communications Commission, Washington,
D.C.
2. DeVries, A.,et al: “Characteristics of Surface-Wave Integratable Filters (SWIFS),” IEEE
Trans., vol. BTR-17, no. 1, pg. 16.
3. Yamada and Uematsu: “New Color TV with Composite SAW IF Filter Separating Sound
and Picture Signals,” IEEE Trans., vol. CE-28, no. 3, pg. 193, 1982.
4. Neal, C. B., and S. Goyal: “Frequency and Amplitude Phase Effects in Television Broad-
cast Systems,” IEEE Trans., vol. CE-23, no. 3, pg. 241, August 1977.
5. Fockens, P., and C. G. Eilers: “Intercarrier Buzz Phenomena Analysis and Cures,” IEEE
Trans. Consumer Electronics, vol. CE-27, no. 3, pg. 381, August 1981.
6. IEEE Guide for Surge Withstand Capability, (SWC) Tests, ANSI C37.90a-1974/IEEE Std.
472-1974, IEEE, New York, N.Y., 1974.
p p
7. “Television Receivers and Video Products,” UL 1410, Sec. 71, Underwriters Laboratories,
Inc., New York, N.Y., 1981.
p p
g g
Chapter
7.2
ATSC DTV Receiver Systems
7.2.1 Introduction1
The introduction of a new television system must be viewed as a chain of elements that begins
with image and sound pickup and ends with image display and sound reproduction. The DTV
receiver is a vital link in this chain. By necessity, the ATSC system places considerable require-
ments upon the television receiver. The level of complexity of a DTV-compliant receiver is
unprecedented, and that complexity is made possible only through advancements in large-scale
integrated circuit design and fabrication.
The goal of any one-way broadcasting system, such as television, is to concentrate the hard-
ware requirements at the source as much as possible and to make the receivers—which greatly
outnumber the transmitters—as simple and inexpensive as possible. Despite the significant com-
plexity of a DTV receiver, this principal has been an important design objective from the start.
7.2.1a Noise Figure

One of the more important specifications that receiver designers must consider is the noise fig-
ure (NF). A number of factors enter into the ultimate carrier-to-noise ratio within a given
receiver. For example, the receiver planning factors applicable to UHF DTV service are shown in
Table 7.2.1 [1]. A consumer can enhance the noise performance of the installation by improving
any contributing factor; examples include installation of a low-noise amplifier (LNA) or a high-
gain antenna.
The assumptions made in the DTV system planning process were based on threshold-of-visi-
bility (TOV) measurements taken during specified tests and procedures [2]. These TOV numbers
1. This chapter is based on: ATSC, “Guide to the Use of the Digital Television Standard,”
Advanced Television Systems Committee, Washington, D.C., Doc. A/54A, December 4,
2003. Used with permission.
Editor’s note: This chapter provides an overview of the ATSC DTV receiver system based on
ATSC A/54A. Readers are encouraged to download the entire Recommended Practice from
the ATSC Web site (http://www.atsc.org). All ATSC Standards, Recommended Practices, and
7-61
y
Table 7.2.1 ATSC Receiver Planning Factors Used by the FCC (After [1].)
Planning Factors Low VHF High VHF UHF

Antenna impedance (ohms) 75 75 75
Bandwidth (MHz) 6 6 6
Thermal noise (dBm) –106.2 –106.2 –106.2
Noise figure (dB) 10 10 7
Frequency (MHz) 69 194 615
Antenna factor (dBm/dBμ) –111.7 –120.7 –130.7 (note 1)
Line loss (dB) 1 2 4
Antenna gain (dB) 4 6 10
Antenna F/B ratio (dB) 10 12 14
Note 1: See Appendix B of the Sixth Report and Order (MM 87-268), adopted April 3, 1997, for a
discussion of the dipole factor.
were correlated to bit error rate (BER)

Table 7.2.2 DTV Interference Criteria (After [2].)
results from the same tests. When character-
izing the Grand Alliance VSB modem hard- Co-channel DTV-into-NTSC 33.8 dB
ware in the lab and in field tests, only BER Co-channel NTSC-into-DTV 2.07 dB
measurements were taken. In Table 7.2.2, the Co-channel DTV-into-DTV 15.91 dB
results of these tests are expressed in equiva- Upper-adjacent DTV-into-NTSC –16.17 dB
Upper-adjacent NTSC-into-DTV –47.05 dB
lent TOV numbers derived from the BER
Upper-adjacent DTV-into-DTV –42.86 dB
measurements. It should be noted that the Lower-adjacent DTV-into-NTSC –17.95 dB
exact amount of adjacent-channel or co- Lower-adjacent NTSC-into-DTV –48.09 dB
channel interference entering the receiver Lower-adjacent DTV-into-DTV –42.16 dB
terminals is a function of the overall antenna
gain pattern used (not just the front/back
ratio), which is also a function of frequency.
7.2.2 Receiver System Overview

Figure 7.2.1 shows a general block diagram of the VSB terrestrial broadcast receiver [2]. The
major system elements include:
• Tuner
• Channel filtering and VSB carrier recovery
• Segment sync and symbol clock recovery
• Noncoherent and coherent automatic gain control (AGC)
• Data field synchronization
• Interference-rejection filter
• Channel equalizer
• Phase-tracking loop
• Trellis decoder
y
IF Filter
& NTSC Data Reed- Data
Phase Trellis
Tuner Syn- Rejection Equalizer De- Solomon De-
Tracker Decoder
chronous Filter Interleaver Decoder Randomizer
Detector
Sync
&
Timing
Figure 7.2.1 Simplified block diagram of the prototype VSB receiver. (From [2]. Used with permis-
sion.)
• Data de-interleaver
• Reed-Solomon decoder
• Data derandomizer
• Receiver loop acquisition sequencing
Descriptions of the major system elements of the prototype receiver are provided in the follow-
ing sections.
Current production designs usually differ somewhat from the prototype receiver with regard
to the order in which signal processing is performed [2]. Most recent designs digitize IF signals
and perform demodulation (such as synchronous detection) in the digital regime. In some
designs the arrangement of the NTSC rejection filter, equalizer, and phase-tracker (or symbol
synchronizer) portions of the receiver differ from those shown in Figure 7.2.1. Portions of the
sync and timing circuitry are likely to take their input signals after equalization, rather than
before, and bright-spectral-line symbol-clock-recovery circuits may respond to envelope detec-
tion of IF signals performed separately from synchronous detection. The cascading of trellis
decoder, data de-interleaver, Reed-Solomon decoder and data de-randomizer is characteristic of
many designs, however.
7.2.2a Tuner
The tuner as implemented in the prototype Grand Alliance system receives the 6 MHz signal
(UHF or VHF) from the antenna (Figure 7.2.2) [2]. The tuner is a high-side injection, double-
conversion type with a first IF frequency of 920 MHz. This places the image frequencies above 1
GHz, making them easy to reject by a fixed front-end filter. This first IF frequency was chosen
high enough that the input band-pass filter selectivity prevents first local oscillations (978–1723
MHz) leaking from the tuner front end and interfering with other UHF channels, but low enough
that second harmonics of UHF channels (470–806 MHz) fall above the first IF passband.
Although harmonics of cable-channel signals could occur in the first IF passband, they are not a
real problem because of the relatively flat spectrum (within 10 dB) and small signal levels (–28
dBm or less) used in cable systems.
y
AGC Band Pass

Band Pass Wide Band
IF Filter IF
Filter Amplifier SAW
1st LC Amplifier 2nd Amplifier
Filter 920MHz Filter
50-800 10dB Mixer Mixer
920MHz 44MHz
MHz 30 -1dB @
dB 6MHz
1st L.O. 2nd L.O.

Synthesizer 876MHz
.97-1.72GHz +0.3MHz
AFC
Channel Tuning Delayed AGC
Figure 7.2.2 Block diagram of the tuner subsystem in the prototype receiver. (From [2]. Used with
permission.)
The tuner input has a band-pass filter that limits the frequency range to 50–810 MHz, reject-
ing all other non-television signals that may fall within the image frequency range of the tuner
(beyond 920 MHz). In addition, a broadband tracking filter rejects other television signals, espe-
cially those much larger in signal power than the desired signal power. This tracking filter is not
narrow, nor is it critically tuned, and introduces very little channel tilt, if any. This contrasts with
the tracking filter used in some NTSC single-conversion tuner designs, in which the tracking fil-
ter is required to reject image signals only 90 MHz away from the desired channel.
A 10 dB gain, wideband RF amplifier increases the signal level into the first mixer, and is the
predominant factor determining the receiver noise figure (7–9 dB over the entire VHF, UHF, and
cable bands). The first mixer is a highly linear, double-balanced design to minimize even-har-
monic generation. The first mixer is a high-side-injection type, driven by a synthesized low-
phase-noise local oscillator (LO) operating at a frequency above those of the broadcast signal
selected for reception. Both the channel tuning (first LO) and broadband tracking filtering (input
band-pass filter) are controlled by microprocessor. The tuner is capable of tuning the entire VHF
and UHF broadcast bands as well as all standard, IRC, and HRC cable bands.
The mixer is followed by an LC filter in tandem with a narrow 920 MHz band-pass ceramic
resonator filter. The LC filter provides selectivity against the harmonic and sub-harmonic spuri-
ous responses of the ceramic resonators. The 920 MHz ceramic resonator band-pass filter has a -
1 dB bandwidth of about 6 MHz. A 920 MHz IF amplifier is placed between the two filters.
Delayed AGC of the first IF signal is applied immediately following the first LC filter. The 30-
dB-range AGC circuit protects the remaining active stages from large signal overload.
The second mixer is a low-side-injection type and is driven by the second LO, which is an 876
MHz voltage-controlled SAW oscillator. (Alternatively, the second mixer could be a high-side-
injection type. [3]) The second oscillator is controlled by the frequency-and-phase-lock-loop
(FPLL) synchronous detector. The second mixer, the output signal of which occupies a 41–47
MHz second IF frequency passband, drives a constant-gain 44 MHz amplifier. The output of the
tuner feeds the IF SAW filter and synchronous detection circuitry. The dual-conversion tuner
y
used in the Grand Alliance receiver is made out of standard consumer electronic components,
and is housed in a stamped metal enclosure.
Since the testing of the original Grand Alliance systems, alternative tuner designs have been
developed. Practical receivers with commercially acceptable performance are now manufactured
using both dual-conversion and single-conversion tuners.
Compared to an NTSC analog television receiver, a DTV receiver has two particularly impor-
tant additional design requirements in the front-end portion up to and including the demodula-
tor(s). One of these requirements is that the phase noise of the local oscillator(s) must be low
enough to permit digital demodulation that is reasonably free of symbol jitter. Symbol jitter gives
rise to intersymbol interference (ISI), which should be kept below levels likely to introduce data-
slicing errors. The Grand Alliance receiver could accommodate a total phase noise (transmitter
and receiver) of –78 dBc/Hz at 20 kHz offset from the carrier. (Note, all figures are measured at
20 kHz offset, and assume a noise-like phase distribution free of significant line frequencies that
may be caused by frequency synthesis.) By 2002, fully integrated demodulator designs typically
showed an improvement 5 dB or so over the Grand Alliance hardware.
The other particularly important additional design requirement of the front-end portion of the
DTV receiver is that the immunity to interference must be better in general than in an analog TV
receiver. Table 7.2.2 shows that such better performance was assumed in making the present
channel assignments. During the interim transition period in which both NTSC and DTV signals
are broadcast, immunity to NTSC interference from channels other than the one selected for
reception will be particularly important. This is because the variations in received levels from
stations to adjacent stations can be very large, with variations on the order of 40 dB not uncom-
mon. DTV-to-DTV variations can be the same order of magnitude. DTV receivers can use digital
filtering methods for rejecting NTSC co-channel interference; such methods are inapplicable to
NTSC receivers.
The principal NTSC interference problem for DTV reception is cross-modulation of a strong
out-of-channel NTSC signal with a desired-channel DTV signal in the RF amplifier and first
mixer stages of the DTV receiver. Actually, the desired-channel DTV signal is less adversely
affected by cross-modulation with a strong out-of-channel signal than a desired-channel analog
NTSC signal is. However, during the transition period the DTV signal will be subject to a more
demanding interference environment. Because each broadcaster was provided with a DTV chan-
nel, DTV power had to be reduced to avoid loss of NTSC service area. Furthermore, DTV
assignments could not be protected from the UHF taboos nor by adjacent channel restrictions.
Thus, the generally accepted conclusion is that DTV receivers should be capable of significantly
better interference rejection than NTSC receivers currently produced are.
7.2.2b Channel Filtering and VSB Carrier Recovery

The Grand Alliance prototype receiver generates in-phase and quadrature-phase analog carrier
signals at intermediate frequency, which are used to implement complex synchronous detection
of the VSB IF signal by analog methods [2]. A complex baseband DTV signal is reproduced that
has an I baseband DTV signal resulting from the in-phase synchronous detection as a real com-
ponent, and a Q baseband DTV signal resulting from the quadrature-phase synchronous detec-
tion as an imaginary component. The prototype receiver uses pilot carrier components in both
the I and Q baseband DTV signals for controlling carrier recovery [4]. All subsequent signal
recovery procedures involve only the I baseband DTV signal.
y
Tuner SAW IF
Filter Amplifier X
AFC
1st AGC 2nd AGC
Lowpass
L.O. L.O.
90 Filter
Reference Amplifer
Synthesizer Oscillator Limiter
3rd L.O.
APC
Channnel
Tuning
VCO Lowpass X
Filter
Q
X
Figure 7.2.3 Block diagram of the tuner, IF amplifier, and FPLL system in the prototype receiver.
Figure 7.2.3 illustrates portions of the prototype receiver, including the frequency- and phase-
lock loop (FPLL) circuit used for carrier recovery. The first LO is synthesized by a phase-lock
loop (PLL) and controlled by a microprocessor. The third LO is a fixed-frequency reference
oscillator. An automatic-phase-and-frequency control (AFPC) signal for the second LO comes
from the FPLL synchronous detector, which responds to the small pilot carrier in the received
DTV signal. The FPLL provides a pull-in range of ±100 kHz despite the automatic-phase-con-
trol low-pass filter having a cut-off frequency less than 2 kHz. When the second LO is in phase
lock, the relatively low cut-off frequency of this APC low-pass filter constrains operation of the
FPLL to a narrow enough bandwidth to be unaffected by most of the modulation of the digital
carrier by data and synchronizing signals. However, the FPLL bandwidth remains wide enough
to track out any phase noise on the signal (and, hence, on the pilot) of frequencies up to about 2
kHz. Tracking out low-frequency phase noise (as well as low frequency FM components) allows
the phase-tracking loop to be more effective.
The pilot beat signal in the Q baseband signal is apt to be at a frequency above the 2 kHz cut-
off frequency of the APC low-pass filter, so it cannot appear directly in the AFPC signal applied
to the VCO. The automatic-frequency-control (AFC) low-pass filter has a higher cut-off fre-
quency and selectively responds to any pilot beat signal in the I baseband signal that is below 100
kHz in frequency. The response of AFC filter exhibits positive phase shift (lag) of that pilot beat
signal that increases with frequency, from a 0º lag at zero-frequency to a 90º lag at a frequency
well above the 100 kHz cut-off frequency. The response to the pilot beat signal drives the ampli-
fier/limiter well into clipping, to generate a constant amplitude (±A) square wave that is used to
multiply the Q baseband DTV signal, to generate a switched-polarity Q baseband DTV signal as
a product output signal from the multiplier.
Without the AFC filter, the I channel beat note and the Q channel beat note would always be
phased at 90º relative to each other, and the direct component of the product output signal from
the multiplier would be zero because each half cycle of the ±A square wave would include equal
negative and positive quarter-cycles of the sinusoidal Q beat wave. However, the AFC filter lag
delays the square wave polarity transitions, so there is less than a quarter wave of the Q baseband
DTV signal before its sinusoidal zero-axis crossing, and more than a quarter wave of the Q base-
band signal after its sinusoidal zero-axis crossing. This results in a net dc value for the product
y
output signal from the multiplier. The polarity of the Q baseband DTV signal relative to the
square wave depends on the sense of the frequency offset between the IF pilot and the VCO
oscillations. Owing to the variable phase shift with frequency of the AFC filter, the amount of
phase shift between the square wave and the Q baseband beat note depends on the frequency dif-
ference between the VCO oscillation and the incoming pilot. The amount of this phase shift
determines the average value of the multiplied Q baseband DTV signal. This zero-frequency
component passes through the APC low-pass filter to provide a frequency-control signal that
reduces the difference between the VCO frequency and the carrier frequency of the incoming IF
signal. (An extended frequency sweep of the response of the FPLL will exhibit the traditional
bipolar S-curve AFC characteristic.)
When the frequency difference comes close to zero, the phase shift in the AFC filter
approaches zero, and so the AFC control voltage also approaches zero. The APC loop takes over
and phase-locks the incoming IF signal to the third LO. This is a normal phase-locked loop cir-
cuit, except for being bi-phase stable. However, the correct phase-lock polarity is determined by
forcing the polarity of the pilot to be the same as the known transmitted positive polarity [5–7].
The bi-phase stability arises in the following ways. When the VCO is in phase-lock, the
detected pilot signal in the real component I of the complex baseband DTV signal is at zero fre-
quency, causing the AFC filter response to be at a constant direct value. This constant direct
value conditions the amplifier/limiter output signal either to be continually at +A value or to be
continually at –A value. The Q baseband DTV signal resulting from the quadrature-phase syn-
chronous detection is multiplied by this value, and the APC low-pass filter response to the result-
ing product controls the frequency of the oscillations supplied by the second LO in the tuner. So,
there is no longer a zero-frequency AFC component generated by heterodyning higher-frequency
demodulation artifacts from the pilot. When the loop settles near zero degrees, the limited I value
is +A, and the Q baseband is used as in an ordinary PLL, locking the Q channel at 90º and the I
channel at 0º. However, if the loop happens to settle near 180º, the limited I value is –A, causing
the multiplied Q channel signal to be reversed in sense of polarity, and therefore driving the loop
to equilibrium at 180º.
The prototype receiver can acquire a signal and maintain lock at a signal-to-noise ratio of 0
dB or less, even in the presence of heavy interference. Because of its 100 kHz cut-off frequency,
the AFC low-pass filter rejects most of the spectral energy in the I baseband DTV signal with
5.38 MHz bandwidth. This includes most of the spectral energy in white noise, in randomized
data, and in the PN sequences in the DFS signal. The AFC low-pass filter also rejects most of the
demodulation artifacts in the I baseband DTV signal that arise from co-channel NTSC interfer-
ence, except those arising from the vestigial sideband of the co-channel NTSC interference.
Therefore, most of the energy from white noise, DTV symbols, and NTSC co-channel interfer-
ence is removed from the amplifier/limiter input signal. This makes it extremely likely that the
beat signal generated by demodulating the pilot is the largest component of the amplifier/limiter
input signal so that this beat signal “captures” the limiting and generates square wave essentially
unaffected by the other components. Any demodulation artifacts in the Q baseband DTV signal
that arise from co-channel NTSC interference that are down-converted to frequencies below the
2 kHz cut-off frequency of the APC low-pass filter when those artifacts are multiplied by the
amplifier/limiter output signal will be within 2 kHz of the digital carrier signal. Only a low-
energy portion of the NTSC vestigial sideband falls within this 4-kHz range of signal, so co-
channel NTSC interference will have little influence on AFC during pull-in of the second LO
towards phase-lock. When the second LO is phase-locked, the narrowband APC low-pass filter
y
7
5
3
1
-1
-3
-5
-7
Levels Before
Pilot Addition
(Pilot=1.25)
Sampling
Instants
1 2 3 4 5 6 7
Figure 7.2.4 Data segment sync waveforms. (From [2]. Used with permission.)
rejects most of the energy in white noise, in randomized data, in the DTV sync signals, and in co-
channel NTSC interference.
The synchronous detectors and the FPLL loop employed in the prototype VSB receiver were
constructed using analog electronic circuitry. Similar processing techniques can be employed in
an FPLL constructed in considerable portion using digital electronic circuitry, for use after syn-
chronous detectors constructed in digital circuitry and used for demodulating digitized IF DTV
signal.
7.2.2c Segment Sync and Symbol Clock Recovery

The repetitive data segment sync signals (Figure 7.2.4) are detected from among the synchro-
nously detected random data by a narrow bandwidth filter [2]. From the data segment sync, a
properly phased 10.76 MHz symbol clock is created along with a coherent AGC control signal.
A block diagram of this circuit is shown in Figure 7.2.5.
The synchronous detector supplies 10.76 Msymbols/s I-channel composite baseband data sig-
nal comprising data and sync sequences. An oscilloscope can be connected for viewing the tradi-
tional analog data eyes in this synchronous detection result. An A/D converter digitizes this
synchronous detection result for digital processing. Owing to the sampling process used in the
y
Synchronous To Data
Detector A/D Detection
Symbol
Clock Segment Sync
Pattern Reference
Data
PLL Segment
fd Sync
Detector
AGC to AGC
IF & Tuner Generator
Gain
Reference Confidence
Counter
Other
Decoder
Circuits
Figure 7.2.5 Segment sync and symbol clock recovery with AGC. (From [2]. Used with permis-
sion.)
digitization, the data eyes in the digitized synchronous detection result cannot be observed by the
conventional oscilloscope procedure.
A phase-lock loop derives a clean 10.76 MHz symbol clock for the receiver. With the PLL
free-running, the data segment sync detector containing a 4-symbol sync correlator searches for
the two-level data segment sync (DSS) sequences occurring at the specified repetition rate. The
repetitive segment sync is detected while the random data is not, enabling the PLL to lock on the
sampled sync from the A/D converter, and achieve data symbol clock synchronization. When a
confidence counter reaches a predefined level of confidence that the segment sync has been
found, subsequent receiver loops are enabled.
Both the segment sync detection and the clock recovery can work reliably at signal-to-noise
ratios of 0 dB or less, and in the presence of heavy interference.
7.2.2d Noncoherent and Coherent AGC

Prior to carrier and clock synchronization, non-coherent automatic gain control is performed
whenever any signal tends to exceed the dynamic range of the A/D converter [2]. This signal may
be a locked signal, an unlocked signal, or just noise/interference. The IF and RF gains are
reduced to keep the A/D converter within its dynamic range, with the appropriate AGC “delay”
being applied.
y
Minimum Data
+
Error Confidence Field
Integrator
Segment Counter Sync
- Detector #1
Received
I Channel Data
After NTSC Field Sync #1
Rejection Filter Reference
Data Symbols Data Segment Sync
Data
Field Sync #2
Reference
Minimum Data
-
Error Confidence Field
Integrator
Segment Counter Sync
+ Detector #2
Figure 7.2.6 The process of data field sync recovery in the prototype receiver. (From [2]. Used
with permission.)
When data segment syncs are detected, coherent AGC occurs using the measured segment
sync amplitudes [8]. The amplitude of the DSS sequences, relative to the discrete levels of the
random data, is determined in the transmitter. Once the DSS sequences are detected in the
receiver, they are compared to a reference value, with the difference (error) being integrated. The
integrator output then controls the IF and “delayed” RF gains, forcing them to whatever values
provide the correct DSS amplitudes.
7.2.2e Data Field Synchronization

Data field sync detection, shown in Figure 7.2.6, is achieved by comparing each received data
segment from the A/D converter (after interference rejection filtering to minimize co-channel
interference) with ideal field #1 and field #2 reference signals in the receiver [2]. Over-sampling
of the field sync is not necessary as a precision data segment and symbol clock has already been
reliably created by the clock recovery circuit. Therefore, the field sync recovery circuit can pre-
dict exactly where a valid field sync correlation should occur within each data segment, if it does
occur. The field sync recovery circuit performs a symbol-by-symbol comparison, using a confi-
dence counter to determine whether the valid field sync correlation does or does not occur in
each data segment [9]. When a predetermined level of confidence is reached that field syncs
have been detected on given data segments, the data field sync signal becomes available for use
by subsequent circuits. The polarity of the middle of the three alternating 63 bit pseudo random
(PN) sequences determines whether field 1 or field 2 is detected.
This procedure facilitates reliable and accurate field sync detection, even in heavy noise,
interference, or echo conditions. Field sync recovery can reliably occur at signal-to-noise ratios
of 0 dB or less, and in the presence of heavy interference.
y
v
C A
(a ) MHz
1.25 227½ f H= 3.579545

286f H = 4.5
19
6.
56 22 fH
(b )
228fH
57fH 285fH
342fH
Lower Baseband Edge (f=0)
D f H = NTSC Horizontal Line Rate
1.
.5 .5
P
(c ) 0 0
Expanded Views Below
D
1
618.88kHz = 39 3 fH
P
(d )
309.44 309.44
kHz kHz 618.88 kHz
45.8 kHz 618.88 kHz 45.8 kHz
381 13 f H = 6.
6. MHz
Figure 7.2.7 Receiver filter characteristics: (a) primary NTSC components, (b) comb filter
response, (c) filter band edge detail, (d) expanded view of band edge. (From [2]. Used with per-
mission.)
7.2.2f Interference-Rejection Filter

The interference rejection properties of the VSB transmission system are based on the frequency
location of the principal components of the NTSC co-channel interfering signal within the 6
MHz television channel and the periodic nulls of a VSB receiver baseband comb filter [2]. Fig-
ure 7.2.7a shows the location and approximate magnitude of the three principal NTSC compo-
nents:
• The video carrier (V) located 1.25 MHz from the lower-frequency edge of the channel alloca-
tion
• Chroma subcarrier (C) located 3.58 MHz above the video carrier frequency
y
Received Data
I Channel NTSC Rejection Filter MUX Output
Data Symbols +
-
D
Data Field Sync NTSC Rejection Filter

Reference Pattern +
+ - +
- D -
( )2 ( )2
Integrator Integrator
Minimum
Energy
Detector
Figure 7.2.8 Block diagram of the NTSC interference-rejection filter in the prototype receiver.
• Audio carrier (A) located 4.5 MHz above the video carrier frequency
If there is an expectation that the reception area for a DTV broadcasting station will suffer co-
channel interference from an NTSC broadcasting station, the DTV broadcasting station will gen-
erally shift its carrier to a frequency 28.615 kHz further from the lower frequency edge of the
channel allocation. As Figure 7.2.7 shows, this places the data carrier (pilot) 338.056 kHz from
the lower edge of the channel allocation, rather than the nominal 309.44 kHz.
The NTSC interference rejection filter (comb) is a single-tap linear feedforward filter, as
shown in Figure 7.2.8. Figure 7.2.7b shows the frequency response of the comb filter, which pro-
vides periodic spectral nulls spaced 57 × fH apart. That is, the nulls are 896.853 kHz apart,
which is (10.762 MHz / 12). There are 7 nulls within the 6 MHz channel. The demodulation arti-
fact of the video carrier falls 15.091 kHz above the second null in the comb filter response; the
demodulation artifact of the chroma subcarrier falls 7.224 kHz above the sixth null in the comb
filter response; and the demodulation artifact of the audio carrier falls 30.826 kHz above the sev-
enth null in the comb filter response. Owing to the audio carrier being attenuated by the Nyquist
roll-off of the DTV signal response, the demodulation artifact of the audio carrier in the comb
filter response is smaller than the demodulation artifact of the video carrier.
The comb filter, while providing rejection of steady-state signals located at the null frequen-
cies, has a finite response time of 12 symbols (1.115 μs). So, if the NTSC interfering signal has a
y
sudden step in carrier level (low-to-high, or high-to-low), one cycle of the zero-beat frequency
(offset) between the DTV and NTSC carrier frequencies will pass through the comb filter at an
amplitude proportional to the NTSC step size as instantaneous interference. Examples of such
steps of NTSC carrier are the leading and trailing edge of sync (40 IRE units). If the desired-to-
undesired (D/U) signal power ratio is large enough, data-slicing errors will occur. However,
interleaving will disperse the data-slicing errors caused by burst interference and will make it
easier for the Reed-Solomon code to correct them. (The Reed-Solomon error-correction cir-
cuitry employed in the Grand Alliance receiver locates as well as corrects byte errors and can
correct up to 10 byte errors/segment).
Although the comb filter reduces the NTSC interference, the data is also modified. The 7 data
eyes (8 levels) are converted to 14 data eyes (15 levels). The partial response process causes a
special case of intersymbol interference that doubles the number of like-amplitude eyes, but oth-
erwise leaves them open. The modified data signal can be properly decoded by the trellis
decoder. Note that, because of time sampling, only the maximum data eye value is seen after ana-
log-to-digital conversion.
NTSC interference can be automatically rejected by the circuit shown in Figure 7.2.8, which
determines whether or not the NTSC rejection filter is effective to reduce interference-plus-noise
in its response compared to interference-plus-noise in its input signal. The amount of interfer-
ence-plus-noise accompanying the received signal during the data field sync interval is measured
by comparing the received signal with a stored reference of the field sync. The interference-plus-
noise accompanying the NTSC rejection filter response during the data field sync interval is
measured by comparing that response with a combed version of the internally stored reference
field sync. The two interference-plus-noise measurements are each squared and integrated over a
few data field sync intervals. After a predetermined level of confidence is achieved, a multi-
plexer is conditioned to select the signal with the lower interference-plus-noise power for supply-
ing data to the remainder of the receiver.
There is a reason to not leave the rejection comb filter switched in all the time. The comb fil-
ter, while providing needed co-channel interference benefits, degrades white noise performance
by 3 dB. This is because the filter response is the difference of two full-gain paths, and as white
noise is un-correlated from symbol to symbol, the noise power doubles. There is an additional
0.3 dB degradation due to the 12 symbol differential coding. If little or no NTSC interference is
present, the comb filter is automatically switched out of the data path. When NTSC broadcasting
is discontinued, the comb filter can be omitted from digital television receivers.
7.2.2g Channel Equalizer

The equalizer/echo-suppressor compensates for linear channel distortions, such as tilt, and the
spectrum variations caused by echoes [2]. These distortions can come from the transmission
channel or from imperfect components within the receiver.
The equalizer in the prototype receiver uses a least-mean-square (LMS) algorithm and can
adapt on the transmitted binary training sequence as well as on the random data. The LMS algo-
rithm computes how to adjust the filter taps so that future symbol errors in the equalizer response
are likely to be reduced [10]. The algorithm does this by generating an estimate of the error cur-
rently present in the equalizer response. This error signal is used to compute a cross-correlation
with various delayed data signals. These correlations correspond to the adjustment that needs to
be made for each tap to reduce the error at the output.
y
The equalizer used in the prototype receiver was designed to achieve equalization by any of
three different methods:
• The equalizer can adapt on the prescribed binary training sequences in the DFS signals
• It can adapt on data symbols throughout the frame when the eyes are open
• Or it can adapt on data symbols throughout the frame when the eyes are closed (blind equal-
ization).
The principal differences among these three methods concern how the error estimate is gener-
ated.
The prototype receiver stores the prescribed binary training signal in read-only memory, and
the field sync recovery circuit determines the times that the prescribed binary training signal is
expected to be received. So, when adapting on the prescribed binary training sequences, the
receiver generates the exact reception error by subtracting the prescribed training sequence from
the equalizer response.
Tracking dynamic echoes requires tap adjustments more often than the training sequence is
transmitted. Therefore, the prototype Grand Alliance receiver was designed so that once equal-
ization is achieved, the equalizer switches to a decision-directed equalization mode that bases
adaptation on data symbols throughout the frame. In this decision-directed equalization mode,
reception errors are estimated by slicing the data with an 8-level slicer and subtracting it from the
equalizer response.
For fast dynamic echoes (e.g., airplane flutter) the prototype receiver used a blind equaliza-
tion mode to aid in acquisition of the signal. Blind equalization models the multi-level signal as
binary data signal plus noise, and the equalizer produces the error estimate by detecting the sign
of the output signal and subtracting a (scaled) binary signal from the output to generate the error
estimate.
To perform the LMS algorithm, the error estimate (produced using the training sequence, 8-
level slicer, or the binary slicer) is multiplied by delayed copies of the signal. The delay depends
upon which tap of the filter is being updated. This multiplication produces a cross-correlation
between the error signal and the data signal. The size of the correlation corresponds to the ampli-
tude of the residual echo present at the output of the equalizer and indicates how to adjust the tap
to reduce the error at the output.
A block diagram of the prototype system equalizer is shown in Figure 7.2.9. The dc bias of
the input signal is removed by subtraction as a preliminary step before equalization. The dc bias
is primarily the zero-frequency pilot component of the baseband DTV signal and is subject to
change when the receiver selects a different reception channel or when multipath conditions
change. The dc offset is tracked by measuring the dc value of the training signal.
The equalizer filter consists of two parts, a 64-tap feedforward transversal filter followed by a
192-tap decision-feedback filter. The equalizer operates at the 10.762 MHz symbol rate, as a T-
sampled (or synchronous) equalizer.
The output signals of the forward filter and feedback filter are summed to produce the equal-
izer response. The equalizer response is sliced by either an 8-level slicer (15-level slicer when the
comb filter is used) or by a binary slicer, depending upon whether the data eyes are open or
closed. (As pointed out in the previous section on interference filtering, the comb filter does not
close the data eyes, but generates in its response twice as many eyes of the same magnitude).
This sliced signal has the training signal and segment syncs reinserted. The resultant signal is fed
into the feedback filter, and subtracted from the output signal to produce the error estimate. The
y
I Channel Equalized
Input 64 Tap I Channel
Symbols + Output
10.76
Σ Forward
Filter
+
Σ
- Filter - 10.76
MS/sec. Taps MS/sec.
Tap Training
Measure
Σ Storage 192 Tap Sequence
Feedback M
Residual Filter U
μ X
DC Filter Slicer
Taps
ACC Tap
Delay 1 Storage Σ
D μ
Feedback
Forward ACC Tap Index
Tap Index
D
Error Delay 2
Estimate -
Σ
+
Figure 7.2.9 Simplified block diagram of the prototype VSB receiver equalizer. (From [2]. Used
with permission.)
error estimate is correlated with the input signal (for the forward filter), or by the output signal
(for the feedback filter). This correlation is scaled by a step-size parameter, and used to adjust
the value of the tap. The delay setting of the adjustable delays is controlled according to the index
of the filter tap that is being adjusted.
The complete prototype receiver demonstrated in field and laboratory tests showed cancella-
tion of –1 dB amplitude echoes under otherwise low-noise conditions, and –3 dB amplitude echo
ensemble with noise. In the latter case, the signal was 2.25 dB above the noise-only reception
threshold of the receiver.
7.2.2h Phase Tracker

The term “phase tracker” is a generic term applied to three servo loops-namely, a phase-tracking
loop, an amplitude-tracking loop, and an offset-tracking loop [2, 11]. The phase-tracking loop is
an additional decision-feedback loop which degenerates phase noise that is not degenerated by
the IF PLL operating on the pilot [12–14]. Accordingly, two concatenated loops, rather than just
one loop, degenerate phase noise in the demodulated DTV signal before its synchronous equal-
ization. Because the system is already frequency-locked to the pilot by the IF PLL (independent
of the data), the phase tracking loop bandwidth is maximized for phase tracking by using a first
order loop. Only first-order loops are used in the phase-tracker of the Grand Alliance receiver.
First-order loops were chosen for the phase-tracker because they follow fast changes in phase
better than simple higher-order loops do.
y
I Data Gain I Data

(From Equalizer) I’ I” (Slicer)
X Delay
Complex Q”
Multiplier
Accumulator Digital
Q’
Error
Limiter Filter
Decision
Phase Error
Gain Error
Sine Accumulator
Cosine
Table
Figure 7.2.10 The phase-tracking loop portion of the phase-tracker. (From [2]. Used with permis-
sion.)
The phase-tracking loop suppresses high-frequency phase noise that the IF PLL cannot sup-
press because of being relatively narrow-band. The IF PLL is relatively narrow-band in order to
lock on the pilot carrier with minimum data-induced jitter. The phase tracker takes advantage of
decision feedback to determine the proper phase and amplitude of the symbols and corrects a
wider bandwidth of phase perturbations. Because of the decision feedback, the phase tracker
does not introduce data-induced jitter. However, like any circuit that operates on data, the phase
tracker begins to insert excess noise as the S/N ratio drops to approach threshold and bad deci-
sions start to be made. Therefore, it may be desirable to vary phase tracker loop gain or switch it
out near threshold signal-to noise-ratio.
A block diagram of the phase-tracking loop is shown in Figure 7.2.10. The output of the real
equalizer operating on the I signal is first gain-controlled by a multiplier and then fed into a filter
that recreates an approximation of the Q signal. This is possible because of the VSB transmission
method, in which the I and Q components are related by a filter function that closely approxi-
mates a Hilbert transform. This filter is of simple construction, being a finite-impulse-response
(FIR) digital filter with fixed anti-symmetric coefficients and with every other coefficient equal
to zero. In addition, many filter coefficients are related by powers of two, thus simplifying the
hardware design.
These I and Q signals are then supplied to a complex multiplier operated as a de-rotator for
suppressing the phase noise. The amount of de-rotation is controlled by decision feedback of the
data taken from the output of the de-rotator. As the phase tracker is operating on the 10.76
Msymbol/s data, the bandwidth of the phase tracking loop is fairly large, approximately 60 kHz.
The gain multiplier is also controlled with decision feedback. Because the phase tracker adjusts
amplitude as well as phase, it can react to small AGC and equalizer errors. The phase tracker
time constants are relatively short, as reflected in its 60 kHz bandwidth.
The phase tracker provides an amplitude-tracking loop and an offset tracking loop as well as a
phase-tracking loop. The phase tracker provides symbol synchronization that facilitates the use
of a synchronous equalizer with taps at symbol-epoch intervals. Greater need for the phase
tracker is evidenced in the 16-VSB mode than in the 8-VSB mode, being essential for 16-VSB
reception with some tuners. The phase tracker also removes effects of AM and FM hum that may
be introduced in cable systems.
y
Trellis
Decoder
#0
Trellis
Decoder
#1
Trellis
Decoder
...
#2
...
Equalized ... ...
& ... Trellis
Phase- Corrected ... Decoded
... Data
Symbols ...
...
...
... ...
... ...
Trellis
Decoder
#10
Trellis
Decoder
#11
Figure 7.2.11 Functional diagram of the trellis code de-interleaver. (From [2]. Used with permis-
sion.)
7.2.2i Trellis Decoder

To help protect the trellis decoder against short burst interference, such as impulse noise or
NTSC co-channel interference, 12-symbol-code intrasegment interleaving is employed in the
transmitter [2]. As shown in Figure 7.2.11, the receiver uses 12 trellis decoders in parallel, where
each trellis decoder sees every 12th symbol [15, 16]. This code de-interleaving has all the same
burst noise benefits of a 12-symbol-code de-interleaver, but also minimizes the resulting code
expansion (and hardware) when the NTSC rejection comb filter is active.
Before the 8 VSB signal can be processed by the trellis decoder, it is necessary to remove the
segment sync. The segment sync is not trellis encoded at the transmitter. The Figure 7.2.12 cir-
cuit block diagram illustrates segment sync removal in the receiver.
The trellis decoder performs the task of slicing and convolutional decoding. It has two modes;
one when the NTSC rejection filter is used to minimize NTSC co-channel interference, and the
other when the NTSC rejection filter is not used. This is illustrated in Figure 7.2.13.
The insertion of the NTSC rejection filter is determined automatically (before the equalizer),
with this information passed to the trellis decoder. When there is little or no NTSC co-channel
interference, the NTSC rejection filter is not used, and an optimal trellis decoder is used to
decode the 4-state trellis-encoded data. Serial bits are regenerated in the same order in which
they were generated in the encoder.
In the presence of significant NTSC co-channel interference, when the NTSC rejection filter
(12 symbol, feed-forward subtractive comb) is employed, a trellis decoder optimized for this par-
tial-response channel is used. This optimal code requires 8 states. This is necessary because the
y
Comb Filter
Segment Sync Removal
Equalizer
Σ Phase Tracker
M
U
To
Trellis
D Σ X Decoder
Figure 7.2.12 Segment sync removal in prototype 8 VSB receiver. (From [2]. Used with permis-
sion.)
NTSC rejection filter, which has memory, represents another state machine seen at the input of
the trellis decoder. In order to minimize the expansion of trellis states, two measures are taken: 1)
special design of the trellis code, and 2) twelve-to-one interleaving of the trellis encoding. The
interleaving, which corresponds exactly to the 12 symbol delay in the NTSC rejection filter,
makes it so that each trellis decoder only sees a one-symbol delay NTSC rejection filter. By min-
imizing the delay stages seen by each trellis decoder, the expansion of states is also minimized. A
3.5 dB penalty in white noise performance is paid as the price for having good NTSC co-channel
performance. The additional 0.5 dB noise-threshold degradation beyond the 3 dB attributable to
comb filtering is due to the 12-symbol differential coding.
Because segment sync is not trellis encoded nor pre-coded, the presence of the segment sync
sequence in the data stream through the comb filter presents a complication that had to be dealt
with. Figure 7.2.12 shows the receiver processing that is performed when the comb filter is
present in the receiver. The multiplexer in the segment sync removal block is normally in the
upper position. This presents data that has been filtered by the comb to the trellis decoder. How-
ever, because of the presence of the sync character in the data stream, the multiplexer selects its
lower input during the four symbols that occur twelve symbols after the segment sync. The effect
of this sync removal is to present to the trellis decoder a signal that consists of only the difference
of two adjacent data symbols that come from the same trellis encoder, one transmitted before,
and one after the segment sync. The interference introduced by the segment sync symbol is
removed in this process, and the overall channel response seen by the trellis decoder is the single-
delay partial-response filter.
The complexity of the trellis decoder depends upon the number of states in the decoder trellis.
Since the trellis decoder operates on an 8-state decoder trellis when the comb filter is active, this
defines the amount of processing that is required of the trellis decoder. The decoder must per-
form an add-compare-select (ACS) operation for each state of the decoder. This means that the
decoder is performing 8 ACS operations per symbol time. When the comb filter is not activated,
the decoder operates on a 4-state trellis. The decoder hardware can be constructed such that the
same hardware that is decoding the 8-state comb-filter trellis can also decode the 4-state trellis
when the comb filter is disengaged, thus there is no need for separate decoders for the two
modes. The 8-state trellis decoder requires fewer than 5000 gates.
y
NTSC Rejection Filter 15 Level Symbols

+ Noise + Interference
Partial
+ Response
Trellis
- Decoder
D
Received Data
Symbols Out
Optimal
Trellis
Decoder
8 Level Symbols (D = 12 Symbols Delay)

+ Noise + Interference
Figure 7.2.13 Trellis decoding with and without the NTSC rejection filter. (From [2]. Used with per-
mission.)
After the transition period, when NTSC is no longer being transmitted, the NTSC rejection
filter and the 8-state trellis decoder can be omitted from digital television receivers.
7.2.2j Data De-Interleaver

The convolutional de-interleaver performs the exact inverse function of the transmitter convolu-
tional interleaver [2]. Its 1/6 data field depth, and inter-segment “dispersion” properties allow
noise bursts lasting as long as about 193 μs to be corrected by subsequent Reed-Solomon error-
correction circuitry of the type that locates byte errors as well as correcting them. Even strong
NTSC co-channel signals passing through the NTSC rejection filter, and creating short bursts
responsive to NTSC vertical edges, are reliably suppressed by this de-interleaving and RS error-
correction process. The de-interleaver uses data field sync for synchronizing to the first data byte
of the data field. The functional concept of the convolutional de-interleaver is shown in Figure
7.2.14
In actual practice the convolutional de-interleaver is usually constructed in dual-port random-
access memory (RAM) [17].
7.2.2k Reed-Solomon Decoder

The trellis-decoded byte data is sent to the (207,187) t =10 RS decoder [2]. The RS decoder in
the Grand Alliance receiver uses the 20 parity bytes to perform byte-error location and byte-error
correction on a segment-by-segment basis. Up to 10-byte errors per data segment are corrected
by the RS decoder. Isolated burst errors created by impulse noise, NTSC co-channel interfer-
ence, or trellis-decoding errors that are less than 193 μs duration are corrected by the combina-
tion of the de-interleaving and RS error correction.
y
(B-1)M
1
2 (B-2)M
From To
Trellis Reed-Solomon
Decoder Decoder
2M
50
M(=4 Bytes)
51
(B=)52
M=4, B=52, N=208, R-S Block =207, BXM=N
Figure 7.2.14 Functional diagram of the convolutional de-interleaver. (From [2]. Used with permis-
sion.)
7.2.2l Data De-Randomizer

The data is randomized at the transmitter by a pseudo random sequence (PRS). The de-random-
izer accepts the error-corrected data bytes from the RS decoder, and applies the same PRS ran-
domizing code to the data. The PRS code can be generated the same way as in the transmitter,
using the same PRS generator feedback and output taps. Since the PRS is locked to the reliably
recovered data field sync (and not to some code word embedded within the potentially noisy
data), it is exactly synchronized with the data, and performs reliably.
7.2.2m Receiver Loop Acquisition Sequencing

The Grand Alliance receiver incorporates a “universal reset” that initiates a number of “confi-
dence counters” and “confidence flags” involved in the lock-up process [2]. A universal reset
occurs, for example, when tuning to another station or turning on the receiver. The various loops
within the VSB receiver acquire and lock-up sequentially, with “earlier” loops being independent
from “later” loops. The order of loop acquisition is as follows:
• Tuner 1st LO synthesizer acquisition
• Non-coherent AGC reduces unlocked signal to within A/D range
• Carrier acquisition (FPLL)
• Data segment sync and clock acquisition
• Coherent AGC of signal (IF and RF gains properly set)
• Data field sync acquisition
• NTSC rejection filter insertion decision made
y
• Equalizer completes tap adjustment algorithm

• Trellis and RS data decoding begin
Most of the loops mentioned previously have confidence counters associated with them to
insure proper operation. However, the build-up or decay of confidence is not designed to be
equal. The confidence counters build confidence quickly for quick acquisition times, but lose
confidence slowly to maintain operation in noisy environments. The VSB receiver carrier, sync
and clock circuits will work in SNR conditions of 0 dB or less, as well as in severe interference
situations.
7.2.2n High Data Rate Mode

The VSB digital transmission system provides the basis for a family of DTV receivers suitable
for receiving data transmissions from a variety of media [2]. This family shares the same pilot,
symbol rate, data frame structure, interleaving, Reed-Solomon coding, and synchronization
pulses. The VSB system was originally designed to offer two modes: a simulcast terrestrial
broadcast mode, and a high data rate mode.
Most parts of the high-data-rate-mode VSB system are identical with or similar to corre-
sponding parts of the terrestrial system. A pilot, data segment sync, and data field sync are used
in both systems. The pilot in the high data rate mode also adds 0.3 dB to the data power. The
same symbol rate, the same segment rate, and the same field rate are used in the high-data-rate-
mode VSB system, allowing a VSB DTV receiver to lock up on either type of signal transmis-
sion. Also, the data frame definitions are identical. The primary difference is the number of
transmitted levels (8 versus 16) and the use of trellis coding and NTSC interference rejection fil-
tering in the terrestrial system.
The high-data-rate-mode receiver is similar to the VSB terrestrial receiver, except for the trel-
lis decoder being replaced by a slicer, which translates the multi-level symbols into data. Instead
of an 8-level slicer, a 16-level slicer is used. Also, note that the high-data-rate-mode receiver
does not employ an NTSC interference rejection filter.
7.2.3 Receiver Equalization Issues

The multipath distortion of a received DTV signal comprises an ensemble of “echoes” or
“ghosts” that accompany the principal component of the received DTV signal because of the sig-
nal being received via various paths [2]. Echoes received via paths that are shorter than that over
which the principal signal is received are termed “pre echoes.” Echoes received via paths that are
longer than that over which the principal signal is received are termed “post-echoes.” The DTV
receiver contains channel-equalization and echo-suppression filtering for selecting the principal
signal and suppressing any accompanying echoes that have sufficient strength to cause errors
during data-slicing. This filtering also corrects in some degree for the receiver not having opti-
mal frequency response (e.g., owing to tilt in the IF amplifier response).
The ATSC Digital Television Standard specifies no particular design for the channel-equal-
ization and echo-suppression filtering in the DTV receiver. When designing an adaptive equal-
izer, usually a designer considers and evaluates various possible ways of implementing this
filtering.
y
Echoes are considered to be of substantial strength if their presence causes a readily observ-
able increase in the number of data-slicing errors. There have been reports of pre-echoes of sub-
stantial strength being observed in the field that are advanced as much as 30 μs with respect to
the principal signal. There have been a significant number of observations of post-echoes of sub-
stantial strength that are delayed on the order of 45 μs with respect to the principal signal.
Equalizers capable of suppressing pre-echoes advanced no more than 3 μs and delayed no
more than about 45 μs were the general rule in the 1995 to 2000 time frame. Since 2000 there has
been increased awareness that channel equalizers should have capability for suppressing further-
advanced pre-echoes and products with over 90 μs of total equalization range have been pro-
duced.
A detailed discussion of DTV receiver equalization is beyond the scope of this chapter. Inter-
ested readers should consult ATSC Informational Document T3-600, “DTV Signal Reception
and Processing Considerations” [18]. Readers are also encouraged to review ATSC A/74, “Rec-
ommended Practice: Receiver Performance Guidelines” [19].
7.2.4 References
1. “Receiver Planning Factors Applicable to All ATV Systems,” Final Report of PS/WP3,
Advanced Television Systems Committee, Washington, D.C., December 1, 1994.
2. ATSC, “Guide to the Use of the ATSC Digital Television Standard,” Advanced Television
Systems Committee, Washington, D.C., Doc. A/54A, December 4, 2003.
3. Limberg, A. L. R: “Plural-Conversion TV Receiver Converting 1st IF to 2nd IF Using
Oscillations of Fixed Frequency Above 1st IF,” U. S. patent No. 6 307 595, 23 October
2001.
4. Citta, R.: “Automatic phase and frequency control system,” U. S. patent No. 4 072 909, 7
February 1978.
5. Krishnamurthy, G., T. G. Laud, and R. B. Lee: “Polarity Selection Circuit for Bi-phase Sta-
ble FPLL,” U. S. patent No. 5 621 483, 15 April 1997.
6. Sgrignoli, G. J.: “Pilot Recovery and Polarity Detection System,” U.S. patent No. 5 675
283, 7 October1997.
7. Mycynek, V., and G. J. Sgrignoli: “FPLL With Third Multiplier in an AC Path in the
FPLL,” U.S. patent No. 5 745 004, 28 April 1998.
8. Citta, R. W., D. M. Mutzabaugh, and G. J. Sgrignoli: “Digital Television Synchronization
System and Method,” U.S. patent No. 5 416 524, 16 May 1995.
9. Citta, R. W., G. J. Sgrignoli, and R. Turner: “Digital Signal with Multilevel Signals and
Sync Recognition,” U.S. patent No. 5 598 220, 28 January 1997.
10. Widrow, B., and M. E. J. Hoff: “Adaptive Switching Circuits,” IRE 1960 Wescon Conv.
Record, pp. 563–587.
11. Horwitz, T. P., R. B. Lee, and G. Krishnamurthy: “Error Tracking Loop,” U.S. patent No. 5
406 587, 11 April 1995.
y
12. Krishnamurthy, G., and R. B. Lee: “Error Tracking Loop Incorporating Simplified Cosine
Look-up Table,” U.S. patent No. 5 533 071, 2 July 1996.
13. Kim, J. G., K. S. Kim, and S. W. Jung: “Phase Error Corrector for HDTV Reception Sys-
tem,” U.S. patent No. 5 602 601, 11 February 1997.
14. Horwitz, T. P., R. B. Lee, and G. Krishnamurthy: “Error Tracking Loop,” U.S. patent No. 5
406 587, 11 April 1995.
15. Citta, R. W., and D. A. Wilming: “Receiver for a Trellis Coded Digital Television Signal,”
U.S. patent No. 5 636 251, 3 June 1997.
16. Wilming, D. A.: “Data Frame Structure and Synchronization System for Digital Television
Signal,” U.S. patent No. 5 629 958, 13 May 1997.
17. Fimoff, M., S. F. Halozan, and R. C. Hauge: “Convolutional Interleaver and Deinterleaver,”
U.S. patent No. 5 572 532, 5 November 1996.
18. ATSC: “Technology Group Report T3-600, DTV Signal Reception and Processing Consid-
erations,” Advanced Television Systems Committee, Washington, D.C., September 18,
2003.
19. ATSC: “Recommended Practice A/74, Receiver Performance Guidelines,” Advanced Tele-
vision Systems Committee, Washington, D.C., June 18, 2004
7.2.5 Bibliography
ACATS: “Field Test Results of the Grand Alliance Transmission Subsystem,” Document SS/
WP2-1354, September 16, 1994.
ACATS: “Final Report of the Spectrum Utilization and Alternatives Working Party of the Plan-
ning Subcommittee of ACATS.”
ATSC: “ATSC Transmission System, Equipment and Future Directions: Report of the ATSC
Task Force on RF System Performance,” Advanced Television Systems Committee, Wash-
ington, D.C., April 12, 2001.
Bendov, O.: “On the Validity of the Longley-Rice (50,90/10) Propagation Model for HDTV Cov-
erage and Interference Analysis,” Proceedings of the Broadcast Engineering Conference,
Bendov, O., J. F. X. Brown, C. W. Rhodes, Y. Wu, and P. Bouchard: “DTV Coverage and Service
Prediction, Measurement and Performance Indices,” IEEE Trans. on Broadcasting, vol. 47,
no. 3, pg. 207, September 2001.
Ciciora, Walter, et. al.: “A Tutorial on Ghost Canceling in Television Systems,” IEEE Transac-
tions on Consumer Electronics, IEEE, New York, N.Y., vol. CE-25, no. 1, pp. 9–44, Febru-
ary 1979.
Ghosh, M: “Blind Decision Feedback Equalization for Terrestrial Television Receivers,” Proc. of
the IEEE, vol. 86, no. 10, pp. 2070–2081, October, 1998.
y
Gibson, J. J., and R. M. Wilson: “The Mini-State—A Small Television Antenna,” RCA Engineer,
vol. 20, pp. 9–19, 1975.
Hufford, G. A: “A Characterization of the Multipath in the HDTV Channel,” IEEE Trans. on
Broadcasting, vol. 38, no. 4, December 1992.
Miyazawa, H.: “Evaluation and Measurement of Airplane Flutter Interference,” IEEE Trans. on
Broadcasting, vol. 35, no. 4, pp 362–367, December 1989.
O’Connor, Robert: “Understanding Television’s Grade A and Grade B Service Contours,” IEEE
Transactions on Broadcasting, vol. BC-14, no. 4, pp. 137–143, December 1968.
Plonka, Robert: “Can ATV Coverage be Improved with Circular, Elliptical or Vertical Polarized
Antennas?,” NAB Engineering Conference Proceedings, National Association of Broad-
Qureshi, Shahid U. H.: “Adaptive Equalization,” Proceedings of the IEEE, IEEE, New York,
N.Y., vol. 73, no. 9, pp. 1349–1387, September 1985.
Ungerboeck, Gottfried: “Fractional Tap-Spacing Equalizer and Consequences for Clock Recov-
ery in Data Modems,” IEEE Transactions on Communications, IEEE, New York, N.Y., vol.
COM-24, no. 8, pp. 856–864, August 1976.
g g
Chapter
7.3
Receiver Antenna Systems
K. Blair Benson
L. H. Hoke, Jr., L. E. Donovan, J. D. Knox, D. E. Manners, W. G. Miller,

R. J. Peffer, J. G. Zahnen, U. L. Rohde
7.3.1 Introduction
The antenna system is one of the most important circuit elements in the reception of television
signals. The ultimate quality of the picture and sound depends primarily on how well the antenna
system is able to capture the signal radiated by the transmitting antenna of the broadcast station
and to feed it, with minimum loss or distortion, through a transmission line to the tuner of the
receiver.
In urban and residential areas near the television station antenna, where strong signals are
present, a compact set-top telescoping rod or “rabbit ears” for VHF and a single loop for UHF
usually are quite adequate. The somewhat reduced signal strength in suburban and rural areas
generally requires a multielement roof-mounted antenna with a shielded 75Ω coaxial transmis-
sion line to feed the signal to the receiver.
Fringe areas, where the signal level is substantially lower, usually require a more complicated,
highly directional antenna, frequently on a tower, to produce an even marginal signal level. The
longer transmission line in such installations may dictate the use of an all-channel low-noise
preamplifier at the antenna to compensate for the loss of signal level in the line.
7.1.1a Basic Characteristics and Definitions of Terms

Antennas have a number of key characteristics that define their ability to receive energy from a
radiated field. These are as follows:
• Wavelength—the distance traveled by one cycle of a radiated electric signal. The frequency of
the signal is the number of cycles per second. It follows that the frequency f is inversely pro-
portional to the wavelength λ. Both wavelength and frequency are related to the speed of
light. Conversion between the two parameters can be made by dividing either into the quantity
of the speed of light c (i.e., λ × f = c). The velocity of radio waves in air is essentially the
7-85
y
same as that of light in free space, which is 2.9983 × 1010 cm/s, or for most calculations, 3 ×
1010 cm/s (186,000 mi/s).
• Radiation—the emission of coherent modulated electromagnetic waves in free space from a
single or a group of radiating antenna elements. Although the radiated energy, by definition,
is composed of mutually dependent electric and magnetic field vectors having specific mag-
nitude and direction, conventional practice in television engineering is to measure and specify
radiation characteristics in terms of the electric field.
• Polarization—the angle of the radiated electrical field vector in the direction of maximum
radiation. If the plane of the field is parallel to the ground, the signal is horizontally polarized;
if at right angles to the ground, it is vertically polarized. When the receiving antenna is
located in the same plane as the transmitting antenna, the received signal strength will be
maximum. If the radiated signal is rotated at the operating frequency by electrical means in
feeding the transmitting antenna, the radiated signal is circularly polarized. Circularly polar-
ized signals produce equal received signal levels with either horizontal or vertical polarized
receiving antennas. (See Figure 7.3.1.)
• Beamwidth. In the plane of the antenna, beamwidth is the angular width of the directivity pat-
tern where the power level of the received signal is down by 50 percent (3 dB) from the max-
imum signal in the desired direction of reception. Using Ohm’s law (P = E2/R) to convert
power to voltage, assuming the same impedance R for the measurements, this is equal to a
drop in voltage of 30 percent (Figure 7.3.2).
• Gain—the signal level produced (or radiated) relative to that of a standard reference dipole.
Gain is used frequently as a figure of merit. Gain is closely related to directivity, which in turn
is dependent upon the radiation pattern. High values of gain usually are obtained with a
reduction in beamwidth. Gain can be calculated only for simple antenna configurations. Con-
sequently, it usually is determined by measuring the performance of an antenna relative to a
reference dipole.
• Input impedance—the terminating resistance into which a receiving antenna will deliver max-
imum power. Similar to gain, input impedance can be calculated only for very simple formats,
and instead is usually determined by actual measurement.
• Radiation resistance. Defined in terms of transmission, using Ohm’s law, as the radiated
power P from an antenna divided by the square of the driving current I at the antenna termi-
nals.
• Bandwidth—a general classification of the frequency band over which an antenna is effective.
This requires a specification of tolerances on the uniformity of response over not only the
effective frequency spectrum but, in addition, over that of individual television channels. The
tolerances on each channel are important because they can have a significant effect on picture
quality and resolution, on color reproduction, and on sound quality. Thus, no single broad-
band response characteristic is an adequate definition because an antenna’s performance
depends to a large degree upon the individual channel requirements and gain tolerances, and
the deviation from flat response over any single channel. This further complicates the
antenna-design task because channel width relative to the channel frequency is greatly differ-
ent between low and high channels.
y
Figure 7.3.1 Polarization of the electric field of the transmitted wave
The foregoing problem is readily apparent by a comparison of low VHF and high UHF chan-
nel frequencies with the 6 MHz single-channel spectrum width specified in the FCC Rules. VHF
Channel 2 occupies 11 percent of the spectrum up to that channel, while UHF Channel 69 occu-
y
(a) (b)
Figure 7.3.2 Normalized E-field pattern of λ/2 and short dipole antennas: (a) plot representation,
(b) three-dimensional view of the radiation pattern.
pies only 0.75 percent of the spectrum up to Channel 69. In other words, a VHF antenna must
have a considerably greater broadband characteristic than that of a UHF antenna.
7.3.2 VHF/UHF Receiving Antennas

In strong-signal urban areas, a set-top antenna consisting of rabbit ears for VHF and loops for
UHF with a foot or so of 300 Ω twin-lead connecting to the receiver antenna terminals usually is
adequate if ghosts resulting from multipath reception are not present. In suburban and rural
areas, a roof-mounted antenna with extra elements to increase the signal level is required. In
weak-signal fringe areas, beyond the FCC-specified “Grade B” contour area of 2.5 mV/m cov-
ered by the transmitter, a high outdoor antenna and low-noise signal preamplifier may be neces-
sary at the antenna in order to boost the signal above that of any stray pickup from interference
signals such as auto ignition and CB radios, and to a level required by the receiver for stable
scanning and color synchronization.
7.1.2a Dipole Antenna

The dipole receiving antenna is not only the simplest type suitable for feeding a transmission line
but also is the basis for other more complex designs [1, 2]. It consists of two in-line rods or wires
with a total length equal to a half-wave at the primary band it is intended to cover. In Figure
y
7.3.3a, the two quarter-wave elements l are (a)

connected at their center to a two-wire
transmission line. Normally a 300 Ω bal-
anced transmission line is connected at the
center. However, because the radiation
resistance of a half-wave dipole is 73 Ω, (b)
use of the usual 300 Ω balanced transmis-
sion line will result in a mismatch and
some loss in transfer of received signal
power. In addition, if the receiving end of
the line is not a matching impedance of
approximately 73 Ω, there will be reflec- (c )
tions up and down the line. These will
show up in the received picture as horizon-
tal smear for short lines or ghosts dis-
placed to the right in the case of longer
Figure 7.3.3 Basic dipole antenna designs: (a)
lines. This picture distortion is avoided in
half-wave dipole, (b) conical dipole, ( c) cylindrical
receiver design, where the tuner input usu-
dipole.
ally is unbalanced 75 Ω coaxial cable, by
the use of a balanced-to-unbalanced
(balun) coupler. A balun consists of a compact transformer with closely coupled input and output
coils wound on a powdered-iron magnetic core. Figure 7.3.4 shows some common balun imple-
mentations [3].
Alternatively, a coaxial transmission line can be used to feed the tuner 75 Ω input directly. In
this case, a balun is mounted at the antenna to transform the balanced dipole signal to the unbal-
anced coaxial transmission line and the antenna’s balanced impedance of 73 Ω is closely matched
(a ) (b) (c)
Figure 7.3.4 Balun transformers: (a) balanced to unbalanced, 1/1, (b) balanced to unbalanced, 4/
1, (c) balanced to unbalanced on symmetrical balun core, 4/1. (Illustrations a and b after [4].)
y
to the 75 Ω transmission line. Dissimilar cir-

cuits can be matched by a resistive network,
but with a resultant loss in signal.
The bandwidth of a dipole antenna can be
increased by increasing the diameter of the
elements through the use of cones or cylin-
ders rather than wires or rods, as illustrated
in Figures 7.3.3b and c. This also increases
the impedance of the antenna so that it more
Figure 7.3.5 The folded dipole antenna.
nearly matches the balanced 300 Ω transmis-
sion line. The 3 dB, or half-power, beam
width, as shown in Figure 7.3.2, is 78º.
If the total length of the dipole is much less than a half wave, the characteristics become those
of a “short dipole” with a substantially lower resistance and high capacity.
Folded Dipole
An increase in bandwidth and an increase in impedance can be accomplished through the use of
the two-element folded-dipole configuration shown in Figure 7.3.5. The impedance can be
increased by a factor of as much as 10 by using rods of different diameter and varying the spac-
ing. The typical folded dipole with elements of equal diameter has a radiation impedance of 290
Ω, four times that of a standard dipole and closely matching the impedance of 300 Ω twin lead.
The quarter-wave dipole elements connected to the closely coupled half-wave single element
act as a matching stub between the transmission line and the single-piece half-wave element.
This broadbands the folded-dipole antenna to span a frequency range of nearly 2 to 1. For exam-
ple, the low VHF channels, 2 through 6 (54–88 MHz), can be covered with a single antenna.
V Dipole
The typical set-top rabbit-ears antenna on VHF Channel 2 is a “short dipole” with an impedance
of 3 or 4 Ω and a capacitive reactance of several hundred ohms. Bending the ears into a V in
effect tunes the antenna and will affect its response characteristic, as well as the polarization. A
change in polarization may be used in city areas to select or reject cross-polarized reflected sig-
nals to improve picture quality.
The fan dipole, based upon the V antenna, consists of two or more dipoles connected in paral-
lel at their driving point and spread at the ends, thus giving it a broadband characteristic. It can be
optimized for VHF reception by tilting the dipoles forward by 33º to reduce the beam splitting
(i.e., improve the gain on the high VHF channels where the length is longer than a half-wave).
A ground plane or flat reflecting sheet placed at a distance of 1/16 to 1/4 wave behind a dipole
antenna, as shown in Figure 7.3.6, can increase the gain in the forward direction by a factor of 2.
This design is often used for UHF antennas, e.g., a “bow-tie” or cylindrical dipole with reflector.
For outdoor applications, a screen or parallel rods can be used to reduce wind resistance. At 1/4-
wave to 1/2-wave spacing, the antenna impedance is 80 to 125 Ω, respectively.
y
(a )
(b )
(c )
(d )
Figure 7.3.6 The corner-reflector antenna: (a) mechanical configuration, (b) calculated pattern of
a square corner-reflector antenna with antenna-to-corner spacing of 0.5 wavelength, (c) pattern at
1.0 wavelength, (d) pattern at 1.5 wavelength. (Illustrations b–d from [5]).
y
7.1.2b Quarter-Wave Monopole

The quarter-wave monopole above a ground
plane, as shown in Figure 7.3.7, is another
antenna derived from the elementary dipole. It
is supplied with many personal portable televi-
sion receivers in which the set itself acts as a
ground plane. Although the monopole is
intended to receive vertically polarized signals,
it usually is moved in its swivel joint to either a
horizontal position or to some other angle to a
Figure 7.3.7 Vertical monopole above a
best compromise between direct and reflected ground plane.
signals. The theoretical resistance characteris-
tic of a monopole with an infinite ground plane
is 37 Ω.
7.1.2c Loop Antenna

The loop antenna set-top configuration—in effect, a form of the folded dipole—is used for UHF
reception. Analyzing it as a single-turn magnetic-field pickup loop, the radiation resistance can
be calculated and found to vary over a more than 10-to-1 range. The values for a typical 7-in
diameter wire loop for several UHF channels are given in Table 7.3.1.
7.1.2d Multielement Array

Multielement arrays can be used to achieve higher gain and directivity [2]. One popular design
for single-channel reception is the Yagi-Uda array, illustrated in Figure 7.3.8. The receiving ele-
ment is 1/2 wavelength for the center of the band covered. The single reflector element is slightly
longer, and the three director elements slightly shorter, all spaced approximately 1/4 wavelength
from each other. Typically, the bandwidth is only one or two channels. However, this can be
increased by trading-off of a slightly longer reflector and slightly shorter directors from theoreti-
cal for a slight loss in gain. On the other hand, broadbanding a single-channel five-element Yagi
to cover all the low-band VHF channels, for example, will reduce the gain figure from 9 dB to
between 4 and 6 dB. (See Figure 7.3.9.)
Another technique for broadening the bandwidth is to add shorter elements in groups between
the directors. For example, in a low-band VHF Yagi, elements can be added for high-band VHF
Table 7.3.1 Radiation Resistance for a Single-Conductor Loop Having a Diameter of 17.5 cm (6.9-
in) and a Wire Thickness of 0.2 mm (0.8-in)
Channel Frequency, MHz Radiation Resistance, Ω

14 473 108
42 641 367
69 803 954
83 887 1342
y
Figure 7.3.8 Five element Yagi-Uda array.
channels. The gain of the Yagi over a standard dipole for up to 15 elements and beamwidth direc-
tivity are listed in Table 7.3.2.
7.3.3 DBS Reception

Consumer satellite receiving antennas are based on several basic parabolic reflector designs. The
parabolic reflector exhibits high gain (referenced to an ideal isotropic antenna). As illustrated in
Figure 7.3.10, the basic configurations are:
• Prime focus, single parabolic reflector. This antenna places the pickup element and LNA in
front of the reflector precisely at the focal point of the parabola. Because the struts that sup-
port the LNA assembly are located directly within the received beam, every effort is made to
design these components with as little bulk as possible, yet physically strong enough to with-
stand adverse weather conditions.
• Double reflector. The primary reflector is parabolic in shape while the subreflector surface,
mounted in front of the focal point of the parabola, is hyperbolic in shape. One focus of the
hyperbolic reflector is located at the parabolic focal point, while the second focal point of the
subreflector defines the position for the pickup/LNA.
• Offset reflector. This design removes the pickup/LNA and its support from the radiated beam.
Although the reflector maintains the shape of a section of a parabola, the closed end of the
curve is not included. The LNA assembly, while still located at the focal point of the curve,
points at an angle from the vertex of the parabola shape.
Antenna type and size for any application is determined by the mode of transmission, band of
operation, location of the receiving station, typical weather in the receiving station locale, and
the required quality of the output signal. Digital transmissions allow a smaller main reflector to
be used because the decoding equipment is usually designed to provide error correction. The
data stream periodically includes information to check the accuracy of the data, and if errors are
found, to make corrections. Sophisticated error concealment techniques make it possible to hide
y
(a )
(b )
Figure 7.3.9 Gain-vs.-bandwidth for a Yagi antenna: (a) measured gain of a five-element single-
channel Yagi and a broadband Yagi, (b) measured gain of three twin-driven 10-element yagi
antennas and a single-channel 10-element Yagi. (From [6].)
y
Table 7.3.2 Typical Characteristics of Single-Channel Yagi-Uda Arrays
Number of Elements Gain, dB Beam Width, degrees

2 3–4 65
3 6–8 55
4 7–10 50
5 9–11 45
9 12–14 37
15 14–16 30
(a) (b)
errors that are too large for correction. For
entertainment programming, absolute cor-
rection is less critical and gives way pri-
marily to concealment techniques.
Phased Arrays
For some reception applications, the para-
bolic antenna is being superseded by a flat (c )
multielement antenna known as the planar
array. Ku-band antenna assemblies mea-
suring no more than a foot square have
been found to provide acceptable DBS
reception under a variety of weather condi-
tions.
The design principle is based upon that Figure 7.3.10 Basic types of satellite receive
of large phased arrays used for many years antennas: (a) prime focus, (b) double reflector, ( c)
in radar applications. Rather than collect- offset feed single reflector.
ing the radiated signal by focusing a para-
bolic reflector on a signal antenna, the
electric signal from each antenna is amplified and timed by a filter network so that all signals are
in phase at the summing point. The basic approach is shown in Figure 7.3.11. A more sophisti-
cated design employs a second smaller antenna to provide an error signal to operate automatic
circuits to align the antenna for maximum signal.
Printed circuit board manufacturing techniques are used to etch the antenna elements on a flat
panel, along with the coupling amplifier and phasing network for each antenna.
7.3.4 Transmission Lines

The basic transmission line equations are derived for distributed parameters R (series resistance),
G (shunt conductance), L (series inductance), and C (shunt capacitance). All are defined per unit
length of the line. Useful relationships are given in Table 7.3.3. Relationships between the volt-
age standing-wave ratio (VSWR) and wave transmission and reflection are given in Table 7.3.4.
The nomograph of Figure 7.3.12 provides a numerical solution for these equations. The Smith
chart (Figure 7.3.13) can be used to graphically determine the input impedance of a transmission
line of a known length with a known resistive or reactive load.
y
(a) (b )
(c)
Figure 7.3.11 Phased array antenna topologies: (a) end collector, (b) center collector, (c) series
phase shifter.
Coaxial shielded cable and several types of parallel line or balanced twin lead are commonly
used between the antenna and receiver. Figure 7.3.14 shows the insertion loss for five types of
line. Typically, the VSWR for the shielded coaxial lines is in the range of 1.3 to 1.5 in the VHF
band and 1.1 to 1.2 in the UHF band. The unshielded twin lead also has low VSWR (1.2 to 1.5)
in the VHF band, but increases to 2.0 in the UHF band.
Shielded twin-lead VSWR has values up to 2.5 in the VHF band and up to 3 at UHF frequen-
cies.
Although the twin-lead transmission line has lowest insertion loss under ideal conditions, the
presence of metal and moisture, as well as outdoor aging, will cause the electrical properties to
deteriorate, and—in fact—insertion loss may even exceed that of coaxial lines. Coaxial cable is
preferred in many installations because it can be routed along the antenna mast adjacent to
metallic structural members and directly through a hole in the house wall with no need for spac-
ers, which are required with twin lead.
7.3.5 Active Antenna Systems

An active antenna system consists of a preamplifier integrated with an antenna structure. When a
full-sized antenna is used, this combination will provide a signal level that is increased by the
gain factor of the amplifier over a similar passive antenna. The use of a preamplifier with a much
smaller antenna structure (L < λ/2) can also provide satisfactory reception to sites at which the
space for a full-sized VHF antenna is not available.
Five major parameters determine the system performance.
• Input and output impedance (300 or 75 Ω , or a combination of both) that must be selected to
match the antenna and transmission line characteristics.
y
Table 7.3.3 Summary of Transmission Line Equations
• Gain, defined as power output to the transmission line divided by power available from the
antenna source.
• Noise figure.
• VSWR caused by mismatch between the source (antenna) and preamplifier input.
• Distortion level or maximum undesired signal level capability, frequently stated as third-
order intercept point of the amplifier.
y
Table 7.3.4 Important Relations in Low-Loss Transmission Lines

y
Figure 7.3.12 Nomograph for transmission and deflection of power at high-voltage standing-wave
ratios (VSWR). (From: Feinstein, J.: “Passive Microwave Components,” Electronic Engineers’
Handbook, D. Fink and D. Christiansen (eds.), Handbook, McGraw-Hill, New York, N.Y., 1982 .)
y
Figure 7.3.13 The Smith Chart.

y
Figure 7.3.14 Insertion loss for common transmission lines.
7.3.6 CEA Antenna Control Interface Standard

EIA/CEA-909 was developed to facilitate automatic control of television receiving device
antenna parameters. Features provided for in the standard include:
• Electronic or mechanical steering of antenna directivity
• Control of preamp gain
• Control of antenna polarization
• Flexibility for future expansion and custom configuration
EIA/CEA-909 describes an antenna control physical interface and a control protocol using
serial bit stream at 8 kbits/s with [7]:
• Two bits for coarse direction control
• Two bits for fine direction control
• Two bits for preamp gain
• One bit for polarization control
• Seven bits for channel number
y
The state of each bit is signaled by using different

widths (42 μs and 83 μs) to enable the use of a D/S G P CH
simple RC detector at the antenna. The antenna
may use only the first seven bits, which allows the D/S Direction 4 bits
receiver to control the system. It may alternatively 2 bits (MSB) coarse direction or switching
use the second seven bits (channel number) to 2 bits (LSB) for fine direction
G Gain 2 bits
access control data stored in the antenna control- P Polarization 1 bit
ler, or it may use both. CH Channel number 7 bits
While these bit definitions were established to
Figure 7.3.15 Data structure specified in
facilitate receiver processing, there is no require-
EIA/CEA-909. (After [8].)
ment for antennas to actually use them as defined
For example, the preamp gain bits could be used
to switch in an additional antenna gain element instead of controlling an active amplifier. The
basic data structure is shown in Figure 7.3.15.
The standard does require an automatic programming procedure by the receiver. In this mode,
for each channel received, the first seven bits may be set for every possible combination (exhaus-
tive search) for a total of 128 combinations. For each state, the receiver measures and temporarily
stores a metric of the signal quality. The state that represents the best quality is then stored and
paired with the channel number. This automatic programming procedure works also if the
antenna control unit discards channel number information.
The standard defines a physical connector for the control interface, a modular telephone-type
connector with an offset latch using a six-conductor cable with the following assignments:
• Pin 2—12 volts @ 150 mA
• Pin 5—data
N
• Remaining pins grounded
EIA/CEA-909 provides a number of W E
antenna configurations that may be
used, and gives several example imple-
N S
mentations of the antenna control signal
processing circuitry. The standard
allows for a second (optional) mode
with many more control states and two-
E
way communication. Output to TV #1
+
Electronic steering of the antenna is
accomplished by using two or more This circuitry may be
duplicated for
antennas; four antennas are typically multiple TV's in
S the same manner
preferred, permitting North, East,
South, and West set positions. The
antennas are switched on and off as
appropriate or combined using a gain-
control network. The gain-control W
approach is preferred since it permits a

full 360° sweep of the composite
antenna pattern. As illustrated in Figure Figure 7.3.16 Using a gain-control scheme to pro-
7.3.16, if adjacent amplifiers are turned vide for 360° steering capability. (After [14].)
y
on, NE, SE, and so on are achieved. If amplifiers N

and E are adjusted in gain, a variable direction
Filter
Antenna Output
between N and E is achieved. Input to coax
Figure 7.3.17 shows an application of EIA/CEA-
909 in which a smart antenna preamplifier uses data Gain control bits
provided by the interface standard to adjust a prese- D/A
lector filter to a specific channel (or group of chan-
nels) and adjust the overall gain of the system, based ROM Channel #
upon input from the receiver.
Figure 7.3.17 Block diagram of a mast-
mounted smart antenna preamplifier.
7.3.7 References (After [14].)
1. Elliot, R. S.: Antenna Theory and Design, Prentice-Hall, Englewood Cliffs, N.J., pg. 64,
1981.
2. Lo, Y. T.: “TV Receiving Antennas,” in Antenna Engineering Handbook, H. Jasik (ed.),
McGraw-Hill, New York, N.Y., pp. 24–25, 1961.
3. Grossner, N.: Transformers for Electronic Circuits, 2nd ed., McGraw-Hill, New York, N.Y.,
pp. 344–358, 1983.
4. Radio Amateur’s Handbook, American Radio Relay League, Newington, Conn., 1983.
5. Kraus, J. D.: Antennas, McGraw-Hill, New York, N. Y., Chapter 12, 1950.
6. Jasik, H.: Antenna Engineering Handbook, McGraw-Hill, New York, N.Y., Chapter 24,
1961.
7. NAB: “Antenna Control Interface Standard and Implementation,” NAB TV TechCheck,
National Association of Broadcasters, Washington, D.C., April 22, 2002.
8. EIA/CEA-909, “Smart Antenna Interface,” Electronics Industries Alliance, Arlington, VA,
February 26, 2002.
y
g g
Chapter
7.4
Cable Television Systems
K. Blair Benson
D. Stevens McVoy, Joseph L. Stern, Charles A. Kase, Wilbur L. Pritchard
7.4.1 Introduction
Cable television systems use coaxial and fiber optic cable to distribute video, audio, and data sig-
nals to homes or other establishments that subscribe to the service. Systems with bidirectional
capability can also transmit signals from various points within the network to the central origi-
nating point (head end).
Cable distribution systems typically use leased space on utility poles owned by a telephone or
power distribution company. In areas with underground utilities, cable systems are normally
installed underground, either in conduits or buried directly, depending on local building codes
and soil conditions.
While current cable practice involves extensive use of fiber in new construction and
upgrades, it is important to understand cable techniques used prior to fiber’s introduction. This is
partly because a significant portion of cable systems have not yet upgraded to fiber and because
these older cable techniques illustrate important technical principles for the service.
7.4.1a Evolution of Cable Television

Cable television began in rural areas of the U.S. in the early 1950s as a means of bringing televi-
sion service to regions with no broadcast stations. Other early systems brought television recep-
tion into mountainous areas where the terrain blocked signals from individual home antennas.
These systems typically had limited capability. During the 1960s, cable moved into areas served
by broadcast stations but without a full complement of network channels. Twelve to 20 channels
were common and were used almost exclusively to carry broadcast television signals. By the
mid-1970s, satellite distribution of pay television programming made cable viable in urban areas
where a good selection of over-the-air programming already existed. The systems offered a
greater variety of viewing choices, including independent stations from distant cities and pay
television.
7-105
y
Systems constructed in the 1980s typically had a capacity of 50 to 100 channels and included
bidirectional transmission capability. Interactive programming, information retrieval, home mon-
itoring, and point-to-point data transmission were possible, although not widely implemented.
Many systems included an institutional network in addition to one or two subscriber networks.
The two main drivers of current progress in the cable industry are the hybrid fiber/coax
(HFC) architecture and digital video. These technologies have changed cable from a stable, well
understood, even stale technology into something with massive possibilities and choices [1].
In the mid 1990’s, the vision of what fiber in the cable plant could accomplish inspired the
cable industry to dramatically upgrade its physical facilities. The current vision driving progress
in the cable industry is digital video compression, which facilitates massive increases in the
amount of programming cable can deliver.
7.4.2 Coax-Based Cable Systems

The division between coax-based and fiber-based cable television systems essentially divide
conventional from state-of-the-art designs. While the implementation of fiber-based systems,
typically built around digital transmission techniques, represent the technologies of choice today,
many older systems still exist. It is instructive, therefore, to examine some of the key operational
considerations that went into designing these systems.
Cable television systems are generally of two types:
• Subscriber networks, which serve primarily residential subscribers. Services carried by sub-
scriber networks are mainly entertainment and information programming.
• Business networks, which serve commercial, educational, and governmental concerns. The
majority of channels on these systems carry informational programming or data.
7.4.2a Program Sources

Television picture and sound program material provided on cable systems may originate from
local or distant broadcast stations or may be relayed by microwave or satellite. A number of
advertiser-supported networks available only through cable are distributed by satellite, as are pay
television services. In addition, programs can be originated by the cable television operator. Such
local origination (LO) programming is produced in studio facilities. Bidirectional cable also
allows “remote” LO, using an upstream channel to transport the signal from the remote site to the
head end, where the signal is converted to a downstream channel for distribution to all subscrib-
ers.
7.4.2b Channel Assignments

Cable television systems primarily use VHF frequencies for distribution of content to subscrib-
ers. Bidirectional systems allow transmissions heading upstream on high frequencies (HF). UHF
signals may also be used. Some confusion arises in that channel numbers normally associated
with UHF channels are used in referring to the additional VHF channels. (See Table 7.4.1.) Busi-
ness networks might allocate 20 or 25 channels in the downstream direction and 15 or 20 chan-
nels to upstream services.
y
Table 7.4.1 Cable Television Channel Frequencies and Designations
Frequency, MHz Channel Designation

5–35 Reverse transmission Sub-band HF
54–88 2–6 Low-band VHF
120–170 14–22 Mid-band VHF
174–216 7–13 High-band VHF
216–300 23–36 Super-band VHF
300–464 37–64 Hyper-band VHF
468–644 65–94 Ultra-band VHF
7.4.2c Elements of a Cable System

Because cable television evolved from a specialized system for transmitting numerous television
channels in a sealed spectrum, rather than a general-purpose communications medium, the
topology or layout of the network was customized for maximum efficiency. The topology that
has evolved over the years is called tree-and-branch architecture. Many small and intermediate
sized systems fit this model. When analyzed, most large systems can be seen as having evolved
from this basic prototype. Typical cable television systems are comprised of four main elements
(Figure 7.4.1):
• A head end, the central originating point of all signals carried
• Trunk system, the main line carrying signals through a community
• Distribution system, which is bridged from the trunk system and carries signals into individ-
ual neighborhoods for distribution to subscribers
• Subscriber drops, the individual lines to subscribers’ television receivers, fed from taps on the
distribution system
In a subscriber’s home, the drop may terminate directly into the television receiver or into a
converter. Modern receivers and videocassette recorders are “cable-ready” and include the nec-
essary converters to access all available channels. Systems providing pay services may require an
access device (or other form of converter) to allow the viewer to receive these special services.
Newer cable systems use addressable converters or deseramblers, giving the cable operator con-
trol over the channels received by subscribers. Such control enables impulse viewing or pay-per-
view pay television service without a technician visiting the home to “install” the special service.
Interactive Service
While the main purpose of cable television is to deliver a greater variety of high-quality televi-
sion signals to subscribers, a growing interest is found in interactive communications, which
allow subscribers to interact with the program source and to request various types of informa-
tion. The growth in the capabilities of the Internet has led to renewed interest in this feature of
cable TV. An interactive system also can provide monitoring capability for special services such
as home security. Additional equipment is required in the subscriber’s home for such services.
Monitoring requires a home terminal, for example, whereas information retrieval requires a
modem (cable or conventional land-line) for data transmission.
y
Figure 7.4.1 Typical cable distribution network. (From [2]. Used with permission.)
Head End
The head end of a cable system is the origination point for all signals carried on the downstream
system. Signal sources include off-air stations, satellite services, and terrestrial microwave
relays. Programming may also originate at the head end. All signals are processed and then com-
bined for transmission via the cable system. In bidirectional systems, the head end also serves as
the collection (and turnaround) point for signals originated within the subscriber and business
networks.
The major elements of the head end are the antenna system, signal processing equipment,
pilot-carrier generators, combining networks, and equipment for any bidirectional and interactive
services. Figure 7.4.2 illustrates a typical analog system head end.
A cable television antenna system includes a tower and antennas for reception of local and
distant stations. The ideal antenna site is in an area of low ambient electrical noise, where it is
possible to receive the desired television channels with a minimum of interference and at a level
sufficient to obtain high-quality signals. For distant signals, tall towers and high-gain, directional
receiving antennas may be necessary to achieve sufficient gain of desired signals and to discrim-
inate against unwanted adjacent channel, co-channel, or reflected signals. For weak signals, low-
noise preamplifiers can be installed near the pickup antennas. Strong adjacent channels are atten-
uated with bandpass and bandstop filters.
Satellite earth receiving stations, or television receive only (TVRO) systems, are used by most
cable companies. Earth station sites are selected for minimum interference from terrestrial
y
Figure 7.4.2 A typical cable system head end. (From [2]. Used with permission.)
y
microwave transmissions and an unobstructed line-of-sight path to the desired satellites. Most
earth stations use parabolic receiving antennas of various sizes, a low-noise preamplifier at the
focal point of the antenna, and a waveguide to transmit the signal to receivers.
Signal Processing
Several types of signal processing are performed at the head end, including:
• Regulation of the signal-to-noise ratio at the highest practical value
• Automatic control of the output level of the signal to a close tolerance
• Reduction of the aural carrier level of television signals to avoid interference with adjacent
cable channels
• Suppression of undesired out-of-band signals
Received signals are converted as necessary to different channel frequencies for cable use.
Such processing of off-air signals is accomplished with heterodyne processors or demodulator/
modulator pairs. Signals from satellites are processed within a receiver and placed on a vacant
channel by a modulator. Similarly, locally originated signals are converted to cable channels with
modulators.
Pilot-carrier generators provide precise reference levels for automatic level control in trunk-
system amplifiers. Generally, two reference pilots are used, one near each end of the cable spec-
trum. Combining networks group the signals from individual processors and modulators into a
single output for connection to the cable network.
Interactive services require one or more data receivers and transmitters at the head end. Poll-
ing of home terminals and data collection are controlled by a computer system (server). Where
cable networks are used for point-to-point data transmission, modems are required at each end
location, with RF-turnaround converters used to redirect incoming upstream signals back down-
stream. Most cable television systems use computerized switching systems to program one or
more video channels from multiple-program sources.
Business networks on cable systems require switching, processing, and turnaround equipment
at the head end. Video, data, and audio signals that originate within the network must be routed
back out over either the business or the subscriber network. One method of accomplishing this is
to demodulate the signals to the base band, route them through a switching network, and then
remodulate them onto the desired network. Another method is to convert the signals to a com-
mon intermediate frequency, route them through an RF switching network, and then up-convert
them to the desired frequency with heterodyne processors.
In larger systems, different areas of the network may be tied together with a central head end
using supertrunks or multichannel microwave. Supertrunks or microwave systems are also used
where the pickup point for distant over-the-air stations is located away from the central head end
of the system. In this method, called the hub system, supertrunks are high-quality, low-loss
trunks, which often use FM transmission of video signals or feed-forward amplifier techniques
to reduce distortion. Multichannel microwave transmission systems may use either amplitude or
frequency modulation.
For new construction, fiber optic lines are used almost exclusively for trunking. Fiber also
typically replaces cable during system retrofit upgrading, required to support new user features
such as additional channels, cable modems, and Internet telephony.
y
Figure 7.4.3 Block diagram of a typical heterodyne-type head-end processor. (Courtesy of

Jerold.)
Heterodyne Processors
The heterodyne processor, shown in Figure 7.4.3, is the most common type of analog head-end
processor. The device converts the incoming signal to an intermediate frequency, where the aural
and visual carriers are separated. The signals are independently amplified and filtered with auto-
matic level control. The signals are then recombined and heterodyned to the desired output chan-
nel.
Prior to cable delivery systems, U.S. television licensing practices avoided placing two sta-
tions on adjacent channels. As a result, many television receivers were not designed to discrimi-
nate well between adjacent channel signals. When used with cable television, such receivers can
experience interference between the aural carrier of one channel and the visual carrier of the
adjacent channel. Channel processors alleviate this problem by changing the ratio between the
aural and visual carriers of each channel on the cable. The visual-to-aural carrier ratio is typi-
cally 15 to 16 dB on cable television systems. At one time, the aural carrier of broadcast stations
was only 10 dB lower than the visual carrier, but many stations now operate with the aural signal
20 dB below visual.
Channel conversion allows a signal received on one channel to be changed to a different chan-
nel for optimized transmission on the cable. Processors are generally modular, and with appro-
priate input and output modules, any input channel can be translated to any other channel of the
cable spectrum. Conversion is usually necessary for local broadcast stations carried on cable net-
works. If the local station were carried on its original carrier frequency, direct pickup of the sta-
tion off the air within the subscriber’s receiver would create interference with the signal from the
cable system.
The visual-signal intermediate-frequency (IF) passband of a typical heterodyne processor is
shown in Figure 7.4.4. The heterodyne processor is designed to reproduce the received signal
with a minimum of differential phase and amplitude distortion. This is best accomplished with a
y
Figure 7.4.4 Typical idealized video-IF response curve of a heterodyne processor. Note that the
visual carrier is located on the flat-loop portion of the curve in order to provide improved phase
response compared to television receivers. (From [2]. Used with permission.)
Figure 7.4.5 Block diagram of demodulator portion of a demodulator/modulator pair. (From [2].
y
Figure 7.4.6 The idealized video-IF response curve of a demodulator. The video carrier is located
6 dB below the maximum response. (From [2]. Used with permission.)
flat passband. Note that the visual carrier is positioned within the flatter portion of the passband
response curve. Television demodulators place the visual carrier at a point 6 dB down on the
response curve. For this reason, the heterodyne processor has better differential phase character-
istics than a modulator/demodulator pair.
An integral circuit of heterodyne processors is a standby carrier generator. If the incoming
signal fails for any reason, the standby signal is switched in to maintain a constant visual carrier
level, particularly on cable systems that use a television channel as a pilot carrier for trunk auto-
matic level control. Such standby carriers can include provisions for modulation from a local
source for emergency messages.
Demodulator/Modulator Pair
Demodulator/modulator pairs are commonly used to convert satellite or microwave signals to
cable channels. The demodulator is essentially a high-quality television receiver with base-band
video and audio outputs. The modulator is a low-power television transmitter. This approach to
channel processing provides increased selectivity, better level control, and more flexibility in the
switching of input and output signal paths, compared to other types of processors.
In the demodulator (Figure 7.4.5), an input amplifier, local oscillator, and mixer down-con-
vert the input signal to an intermediate frequency. Crystal control or phase-locked loops maintain
stable frequencies for the conversion. In the IF amplifier section, the aural and visual signals are
separated for detection. In comparing the IF response curve of a demodulator (Figure 7.4.6) with
that of the heterodyne processor (Figure 7.4.4), we find that the visual carrier is located on the
y
Figure 7.4.7 Block diagram of a conventional analog modulator. (From [2]. Used with permission.)
Figure 7.4.8 Block diagram of a single-channel amplifier with AGC. (From [2]. Used with permis-
sion.)
edge of the passband. The 4.5-MHz intercarrier sound is taken off before video detection, ampli-
fied, and limited to remove video components. The sound is often maintained on a 4.5-MHz
audio subcarrier for remodulation.
Demodulators are designed to minimize phase and amplitude distortion with linear detector
circuits. However, quadrature distortion is inherent in systems that use envelope detectors with
vestigial sideband signals. Such distortion can be corrected with video processing circuitry.
In the modulator (Figure 7.4.7), a composite input signal is applied to a separation section,
where video is separated from the sound subcarrier. The video is amplified, processed for dc res-
toration, and mixed with a carrier oscillator to an IF frequency. An amplifier boosts the signal to
the required power level, and a vestigial sideband filter is used to trim most of the lower sideband
for adjacent-channel operation.
y
Figure 7.4.9 FM radio heterodyne processor block diagram.
Audio, if derived from the composite input, is separated as a 4.5 MHz signal from the visual
information, filtered, amplified, and limited before being mixed with the visual carrier. If audio
is provided separately to the modulator, an FM modulation circuit generates the 4.5 MHz aural
subcarrier to be combined with the visual carrier. After the two are combined, they are mixed
upward to the desired output channel frequency and applied to a bandpass filter to remove spuri-
ous products of the upconversion before being applied to the cable system.
Single-Channel Amplifiers
The single-channel or strip amplifier is designed to amplify one channel and include bandpass
and bandstop filters to reject signals from other channels. In simplest form, the unit consists of
an amplifier, filter, and power supply. More elaborate designs include automatic gain control
(AGC) and independent control of the aural and visual carrier levels. (See Figure 7.4.8.)
Single-channel amplifiers are useful where the desired signal levels are relatively high and
undesired signals are low or absent. Selectivity is low compared with other processor types, and
the units generally lack the independent control of carrier levels, the ratio between carriers, inde-
pendent AGC of the carriers, or limiting for the aural carrier. They cannot be used for channel
conversion, and they usually present difficulties in adjacent-channel applications because of their
limited selectivity.
Supplementary Service Equipment

Most cable television systems have the ability to distribute other signals besides television. They
may offer additional services to subscribers, such as FM broadcast radio, data transmission,
interactive signaling, and specialized audio and video transmissions.
FM radio service on cable systems is commonly provided only in the standard FM broadcast
band from 88 to 108 MHz. Heterodyne processors are typically used to place an FM radio signal
into the cable system spectrum at a frequency where it will not interfere with direct pickup of
over-the-air signals. This process, in general terms, is illustrated in Figure 7.4.9.
Data transmission on a cable television network requires modems designed for the particular
bandwidth, frequency, and data transmission speed required. Many cable systems dedicate a por-
tion of the spectrum on a subscriber network or install a separate institutional network for data
y
Figure 7.4.10 A basic trunk system. (From [3]. Used with permission.)
transmission. Generally, these systems are bidirectional to allow interactive data transmission
and signaling. Modems for cable system applications are available for data speeds to 1.5 Mbits/s
and higher.
Head-End Signal Distribution

The trunk system (Figure 7.4.10) is designed for bulk transmission of multiple channels. The
trunk may connect numerous distribution points and may interconnect widely spaced sections of
a cable system or more than one cable system. A trunk system spanning long distances without
intermediate distribution is often referred to as a super trunk. Fiber has become the medium of
choice for trunk systems, displacing coaxial cable in new installations.
The coaxial cable used in trunk systems is made with a solid aluminum shield, a gas-injected
polyethylene, low-loss dielectric insulator, and a copper or copper-clad aluminum-center con-
ductor. Underground cable has an additional outer polyethylene jacket. Under the jacket is a
flooding compound to seal small punctures that might occur during or after installation. Trunk
cables are usually 3/4 to 1-in. in diameter and typically have a loss of about 1 dB/100 ft at 400
MHz.
For long-distance trunking with minimum distortion, frequency-modulation (FM) techniques
are preferred. While FM requires a larger portion of the available bandwidth than amplitude
modulation, the distortion in the delivered signal is considerably lower. Thus, FM techniques
permit longer trunk lengths.
y
Table 7.4.2 Distortion-vs.-Number of Amplifiers Cascaded in a Typical Cable System
Number of Amplifiers in Cascade Cross Modulation, dB Second Order, dB S/N, dB

1 –96 –86 60
2 –90 –81.5 57
4 –84 –77 54
8 –78 –725 51
16 –72 –68 48
32 –66 –63.5 45
64 –60 –59 42*
128 –54* –54.5* 39
* Lower limit of acceptable system performance.
Amplification
Amplifiers in the cable system are used to overcome losses in the coaxial cable and other devices
of the system. From the output of a given amplifier, through the span of coaxial cable (and an
equalizer to linearize the cable losses), and to the output of the next amplifier, unity gain is
required so that the same signal level is maintained on all channels at the output of each ampli-
fier unit.
Standard cable system design places repeater amplifiers from 1400 to 2000 ft apart, depend-
ing on the size and type of cable used. This represents an electrical loss of about 20 dB at the
highest frequency carried. Systems with amplifier cascades to 64 units are possible, depending
on the number of channels carried, the performance specifications chosen, the modulation
scheme used, and the distortion-correction techniques used. Table 7.4.2 tabulates representative
figures of distortion versus the number of amplifiers in cascade.
Bridging Amplifier
Signals from a trunk system are fed to the distribution system and eventually to subscriber drops.
A wide-band directional coupler extracts a portion of the signal from the trunk for use in the
bridge amplifier. The bridge acts as a buffer, isolating the distribution system from the trunk,
while providing the signal level required to drive distribution lines.
Subscriber-Area Distribution
As shown in Figure 7.4.11, distribution lines are fed from a bridging station. Distribution lines
are routed through the subscriber area with amplifiers and tap-off devices to meet the subscriber
density of the area. The cable for distribution lines is identical to that used for trunks, except that
its diameter is commonly 1/2 or 5/8-in. with a resulting increase in losses, typically to 1.5 dB/
100 ft at 400 MHz.
As the signal proceeds on a distribution line, attenuation of the cable and insertion losses of
tap devices reduce its level to a point where line-extender amplifiers may be required. The gain
of line extenders is relatively high (25 to 30 dB), so generally no more than two are cascaded,
because the high level of operation creates a significant amount of cross-modulation and inter-
modulation distortion. These amplifiers typically do not use automatic level-control circuitry to
compensate for variations in cable attenuation caused by temperature changes. By limiting the
system to no more than two line extenders per distribution leg, the variations are not usually suf-
ficient to affect picture quality.
y
Figure 7.4.11 Distribution and feeder system. (From [3]. Used with permission.)
Two types of tap-off devices are commonly used:

• Birectional-coupler multitap, which provides for connections from two to eight subscribers
per tap. Most manufacturers offer a series of taps, each given a value based on the amount of
signal taken from the distribution line. (See Figure 7.4.12.)
• Individual pressure taps, which fasten directly to the distribution cable. These devices find
limited use today. Pressure taps tend to create signal reflections, because their construction
and installation onto the cable simulate a distortion in the center conductor-to-shield distance.
A tap removes an appropriate amount of energy from the distribution cable through a direc-
tional coupler, then splits the energy into several paths. Each path proceeds through a drop into a
subscriber’s home. The value of each tap is selected to produce an adequate signal level for a
good S/N at the subscriber’s receiver, but not so high as to create intermodulation in the receiver.
Each tap introduces some attenuation into the distribution system, related to the amount of signal
tapped off for subscribers.
By extracting the signal through a directional coupler, a high attenuation from the tap outlets
back into the cable reduces the interference that originates in a subscriber drop cable from reach-
ing subscribers served by other multitaps. The directional characteristics of the taps also prevent
reflected signals from traveling along the distribution system and causing ghosting on subscrib-
ers’ receivers.
y
Figure 7.4.12 Multitap functional block diagram.
The splitter portion of a multitap is a hybrid design that introduces a substantial amount of
isolation from reflections or interference coming from a subscriber drop, both to the distribution
system and to other subscribers connected to the same tap.
The tap is the final service point immediately prior to a subscriber’s location, so it becomes a
convenient point where control of channels authorized to individuals can be applied. An addres-
sable tap, controlled from the system head end, can be used instead of the standard multitap to
enable access to basic cable services or to individual channels. When activated, the switch and
related components can enable or disable a channel, a group of channels, or the entire spectrum.
Trunk and Distribution Amplifiers

Because the trunk is the main artery of a cable television system, trunk amplifiers must provide a
minimum of degradation to the system. A typical amplifier station is of modular design with
plug-in units for trunk amplification, automatic level control, bridging, and signal splitting to
feed distribution cables. The modules receive power from the coaxial cable. A power supply in
each amplifier housing taps power from the cable to supply dc voltage to operate the other mod-
ules. The power is inserted onto the coaxial cable at points where a connection to the local power
utility is convenient. A step-down supply is used to insert 30 V ac or 60 V ac onto the coaxial
cable through a passive inductive network.
Figure 7.4.13 shows a block diagram of the typical trunk amplifier with provisions for auto-
matic level control and the distribution of gain. The first stage provides low-noise operation,
while remaining stages produce gain with low cross-modulation. Trunk amplifiers use hybrid
circuits, combining integrated circuits and discrete components in a single package. Such hybrid
packages are available with a variety of gain and performance characteristics.
y
Figure 7.4.13 Functional block diagram of a trunk-AGC amplifier and its gain distribution. ( From
Gain Control
Because the attenuation characteristics of coaxial cable change with ambient temperature, trunk
amplifiers include automatic gain-control (AGC) circuits to compensate for the changes and to
maintain the output level at a predetermined standard. Automatic control depends on the use of
pilot carriers. One is at the lower end of the spectrum, while the other is at the upper limit. Spe-
cial-purpose pilot carriers also can be placed on the system at the head end. Some system
designs use video channels as pilots. AGC circuits sample the trunk amplifier output at the pilot
frequencies and feed a control voltage back to diode attenuators at an intermediate point in the
amplifier circuit. The upper pilot adjusts the gain of the amplifier, while the lower pilot controls
the slope of an equalizer. This method is effective in maintaining the entire spectrum at a con-
stant output level.
A second design of trunk amplifier uses feed-forward distortion correction and permits car-
riage of a larger number of channels (for equivalent gain) than the standard trunk amplifier.
Cross-Modulation
The maximum output level permissible in cable television amplifiers depends primarily on
cross-modulation of the picture signals. Cross-modulation is most likely in the output stage of an
amplifier, where levels are high. The cross-modulation distortion products at the output of a typ-
ical amplifier (such as Figure 7.4.13) are 96 dB below the desired signal at an output level of +32
dB mV. In the system shown in Figure 7.4.13, gains of the third and fourth stages are each 8 dB.
With an amplifier output level of +32 dB mV, the output of the third stage would be +24 dB mV,
and that of the second stage, 16 dB mV.
Figure 7.4.14 expresses the relationship between cross-modulation distortion and amplifier
output level. A 1 dB decrease in output level corresponds to a 2 dB decrease in cross-modulation
distortion.
Cross-modulation increases by 6 dB each time the number of amplifiers in cascade doubles.
The distortion also increases by 6 dB each time the number of channels on the system is doubled.
Experience has shown that cable subscribers find a cross-modulation level of –60 dB objection-
able. Therefore, based on a cross-modulation component of –96 dB relative to normal visual car-
y
Figure 7.4.14 Cross-modulation and second-order beat distortion-vs.-output level of a trunk

amplifier. (From [2]. Used with permission.)
rier levels in typical trunk amplifiers, 64 such amplifiers can be cascaded before cross-
modulation becomes objectionable.
Frequency Response
The most important design consideration of a cable amplifier is frequency response. A response
that is flat within ± 0.25 dB from 40 MHz to 450 MHz is required of an amplifier carrying 50 or
more 6 MHz television channels to permit a cascade of 20 or more amplifiers. Circuit designs
must pay special attention to high-frequency parameters of the transistors and associated compo-
nents, as well as circuit layout and packaging techniques.
The circuit given in Figure 7.4.15 represents an amplifier designed for flat response. The
feedback network (C1, R1, L1) introduces sufficient negative feedback to maintain a nearly con-
stant output over a wide frequency range. Collector transformer T2 is bifilar-wound on a ferrite
core and presents a constant 75 Ω impedance over the entire frequency range. Splitting trans-
formers T1 and T3, of similar construction, have an additional function of introducing 180º phase
shift for the push-pull transistor pair Q1 and Q2, while maintaining 75 Ω input and output imped-
ances.
Second-Order Distortion
When cable television systems carried only 12 channels, those channels were spaced to avoid the
second harmonic of any channel from falling within any other channel. The expansion of sys-
tems to carry channels covering more than one octave of spectrum causes second-order beat dis-
tortion characteristics of amplifiers to become important. Push-pull amplifier circuits, as shown
y
in Figure 7.4.15, provide a common method of

reducing amplifier second-order distortion. Com-
bining several of these circuit building blocks into
an amplifier effectively cancels second-order dis-
tortion if the 180º phase relationship is main-
tained over the full amplifier bandwidth. The gain
of both halves of the amplifier must be equal and
the individual transistors must be optimized for
maximum second-order linearity.
Second-order distortion increases by approxi-
mately 1.5 dB for each 1 dB in amplifier output
level, as given in Figure 7.4.14. Referring to Table
7.4.2, each amplifier introduces a distortion level
of approximately –86 dB relative to the visual
carriers. Because second-order beats greater than
–55 dB become objectionable, approximately 64
amplifiers can be cascaded before that level is
attained.
Noise Figure
The typical trunk amplifier hybrid module has a
noise figure of approximately 7 dB. Equalizers,
band-splitting filters, and other components pre-
ceding the module add to the overall station noise
figure, producing a total of approximately 10 dB.
A single amplifier produces an S/N of approx-
imately 60 dB for a typical signal input level.
Subjective tests show that an S/N of approxi-
mately 43 dB is acceptable to most viewers. Figure 7.4.15 High-performance, push-
Again, based on Table 7.4.2, noise increases 3 dB pull, wide-band amplifier stage for a trunk
as the cascade of amplifiers is doubled. There- amplifier. (From [2]. Used with permission.)
fore, a cascade of 32 amplifiers, resulting in a sig-
nal-to-noise ratio of 45 dB, is a commonly used
cable television system design specification.
Hum Modulation
Because power for the amplifiers is carried on coaxial cable, some hum modulation of the RF
signals occurs. This modulation results from power-supply filtering and saturation of the induc-
tors used to bypass ac current around active and passive devices within the system. A hum mod-
ulation specification of –40 dB relative to visual carriers has been found to be acceptable in most
applications.
Subscriber-Premises Equipment
The output of the tap device feeds a 75 Ω coaxial drop cable into the home. The cable, generally
about 1/4-in. in diameter, is constructed with an outer polyethylene jacket, a shield of aluminum
y
foil surrounded by braid, a polyethylene dielectric insulator, and a copper-center conductor. A

loss at 400 MHz of 6 dB/100 ft is typical.
For reasons of safety and signal purity, isolation between the subscriber’s television receiver
and the distribution cable is required. In the multitap, blocking capacitors prevent the ac current
used to power amplifiers from reaching the subscriber drop. As the drop enters the subscriber’s
home, a grounding connection, required by local regulations, protects the television receiver
from power surges, such as those caused by lightning discharges, and protects the subscriber
from possible shock by voltages that might otherwise be present on the shield of the drop cable.
Finally, at the subscriber’s receiver, the signal from the unbalanced 75 Ω coaxial cable is con-
verted to a 300 Ω balanced output before connection to the tuner section of the receiver. This
matching transformer provides a further level of isolation to energy along the sheath of the cable.
Converters and Descramblers

Prior to the manufacture of cable-ready television receivers and VCRs, systems with more than
12 channels required the use of a converter at the subscriber’s location. A nontunable block con-
verter, which translates a group of channels to a different portion of the spectrum, can be used
for this purpose.
Most converters used today are tunable units, with microprocessor control of the frequency-
synthesizing circuit referenced to a crystal. The viewer selects a channel (usually with a hand-
held remote control unit), which is converted to a standard VHF frequency (usually channel 3 or
4). In such cases, the television receiver remains tuned to “channel 3,” and all tuning is done with
the converter.
Cable-ready tuners avoid the need for a converter by including additional tuning capabilities
in the receiving equipment. As the viewer selects the desired channel, digital switching sets the
input control lines of a phase-locked-loop frequency synthesizer. When selected, any of the addi-
tional channel signals are converted to the normal television IF amplifier frequency band.
Prior to cable-ready receivers, the use of nonstandard channels and converters provided an
easy method of limiting access to special channels. Now, some method of scrambling or other
control must be used to manage access. Early approaches included suppression of the horizontal
and vertical synchronizing pulses of the controlled program signal during transmission at the
head end. At the same time, a pilot signal that contained information about the missing sync was
also transmitted on the cable. In a descrambler, located in the subscriber’s home, the pilot infor-
mation provided the key to correctly re-create the synchronization.
A greater degree of security for scrambled signals can be attained through digital methods
and base-band scrambling. Numerous concepts have been implemented, including the inversion
of lines of video on a random basis, shuffling of lines of video, and conversion of the video into
a digital bit stream. These methods, requiring that signals be demodulated into audio and video,
are more expensive in terms of equipment costs, but they present a number of possible advan-
tages as well.
By including an addressable switch into a set-top descrambler (Figure 7.4.16), the cable oper-
ator can use computer control from the head end to restrict access to channels, depending on the
tiers of service a subscriber has chosen. In a bidirectional cable system, pay-per-view impulse
viewing is relatively easy to implement as well. A subscriber need only key in an appropriate
code to request reception of a special program. The head-end computer, constantly polling all
converter/descrambler units, detects the request, enables the special channel for the duration of
y
Figure 7.4.16 Addressable-converter block diagram. (Courtesy of Scientific Atlanta.)
the program request, and subsequently places an entry into the automated billing system for that
subscriber.
Off-Premises Systems
The off-premises approach is compatible with recent industry trends to become more consumer
electronics friendly and to remove security-sensitive electronics from the customer’s home [1].
This method controls the signals at the pole rather than at a decoder in the home. This increases
consumer electronics compatibility because authorized signals are present in a descrambled for-
mat on the customer’s drop. Customers with cable-compatible equipment can connect directly to
the cable drop without the need for converter/descramblers. This allows the use of all VCR and
TV features.
Signal security and control in the off-premises devices take different forms. Nearly all off-
premises devices are addressable. Specific channels or all channels are controlled remotely.
Interdiction technology is one approach to plant program security. In this format, the pay tele-
vision channels to be secured are transported through the cable plant in the clear (not scrambled).
The security is generated on the pole at the subscriber module by adding interference carrier(s)
to the unauthorized channels. An electronic switch is incorporated, allowing signals to be turned
off. In addition to being consumer electronics friendly, this method of security does not degrade
the picture quality on an authorized channel.
y
Bidirectional Cable Television Systems

Although a bidirectional cable system could be constructed using separate cables for the
upstream and downstream signal paths, the costs are usually prohibitive. More practical is a sin-
gle-cable system in which the direction of the signal is related to frequency, as outlined previ-
ously in this chapter. Using this approach, called subsplitting, appropriate circuits must be
integrated into all amplifiers in the system. Each amplifier station now has two amplifier sec-
tions, one for operation in each direction, with band-splitting and combining fi1ters used to
direct signals to the appropriate amplifier modules.
Bidirectional operation introduces a number of new problems into the cable system. Probably
the greatest difficulties encountered stem from transmitters external to the system operating on
frequencies between 5 and 35 MHz. That includes citizen’s band, amateur radio, and shortwave
broadcast, as well as some public-service applications. Ingress of signals from these transmitters
is difficult to prevent and interferes with reliable upstream transmission. Some of these problems
can be reduced by allowing the use of remotely controlled switchers at bridger amplifier loca-
tions. In effect, only certain portions of the distribution system can send signals upstream at any
given time, which places definite limitations on the effectiveness of bidirectional operation.
7.4.3 Fiber-Based Cable Systems

The last few years have brought exciting trends employing new technologies for cable television
[1]. Fiber is now installed to upgrade older systems and as part of rebuilds and new builds. New
amplifier technologies allow realization of cable’s inherent bandwidth, which exceeds 1 GHz.
A bandwidth of 1 GHz contains 160 slots of 6 MHz spectrum. These can be allocated to
NTSC, DTV, and to new services. The most exciting potential lies with utilizing video compres-
sion technology to squeeze four or more NTSC-like quality signals in a 6 MHz slot. This opens
the door for hundreds of channels. Near video on demand (NVOD) becomes practical; the most
popular movies could be repeated every few minutes to minimize the wait time before a movie
starts. The average wait time could be made shorter than the trip to the video store, and the sub-
scriber does not have to make a second trip to return the movie. A computer can keep track of
which channel to return to should the subscriber wish to take a break. It is possible to design sys-
tems that behave as if they are switched even though they remain more like a traditional cable
tree-and-branch structure.
Digital fiber links are used when video signals are transported in a digital common carrier
network. The video routes efficiently through common carrier points of presence and switching
networks.
7.4.3a Hybrid Fiber/Coax

The lasers that drive fiber optic links are expensive, and the receivers that convert the optical
energy into video signals also are relatively expensive [1]. To be practical, these components
must serve hundreds of subscribers each so that the costs can be shared. This is accomplished
with a layout in which fibers feed small coaxial systems, which in turn serve a few hundred to a
few thousand subscribers.
The hybrid fiber/cable (HFC) architecture has made it possible to cost effectively increase
bandwidth, signal quality, and reliability while reducing maintenance costs and retaining a craft-
y
Head End Hub Node

50,000 hp 500 hp
1310 nm Tx
Tx EDFA 10 dB Rx
5 dB Rx
Broadcast Primary
Analog
Redundant RF
EDFA Switch
Narrowcast Return Tx
Interactive
Services
CMTS
PSTN SDH SDH HDT
ISP Terminal Terminal QAM
Video
Server Return Rx
Figure 7.4.17 A traditional double hop HFC architecture. (After [4].)
friendly (easily serviced) plant. HFC makes two-way service practical. The bandwidth of coaxial
cable has no sharp cut off. In a coax-based system, it is the cascade of amplifiers that limits the
practical bandwidth. Twenty to forty amplifiers in cascade not only reduces bandwidth, but also
constitutes a considerable reliability hazard. Overlaying low-loss fiber over the trunk portion of
the plant eliminates the trunk amplifiers. This, in turn, leaves only the distribution portion of the
plant with its relatively short distances and only two or three amplifiers. Wider bandwidth is thus
facilitated.
Two way operation in a fiber system is quite practical for two reasons. First, the fiber itself is
not subject to ingress of interfering signals. Second, the cable system is broken up into a large
number of small cable systems, each isolated from the others by its own fiber link to the head-
end. If ingress should cause interference in one of these small cable systems, that interference
will not impair the performance of the other portions of the cable plant.
7.4.3b HFC Architectures

A traditional HFC architecture is shown in Figure 7.4.17 [4]. Broadcast analog video is transmit-
ted by an externally modulated 1550 nm transmitter from the head end to the hub. A redundant
transmitter and path are included for signal protection. Narrowcasting services (data, voice, and
possibly digital video) are typically transmitted via a synchronous digital hierarchy (SDH) link,
although other options include ATM, IP, or proprietary baseband digital transport. At the hub, the
digital narrowcast signals are processed for transmission over the HFC network by the appropri-
ate interface unit. A cable modem termination system (CMTS) converts the downstream data sig-
nal to QAM, and QPSK demodulates the upstream signal, as well as supplying media access
control. The CMTS and cable modems may be based on one of several protocols. The host digi-
tal terminal (HDT) performs the same tasks for cable telephony. Video servers are shown at the
head end, but could also be located in the hub to provide digital video and VOD service. After
processing and QAM modulation, the narrowcast signals are combined with the broadcast ana-
y
Head End Cabinet OSN Mini-Node

12,500 hp 800 hp 100 hp
Tx EDFA 1x16
10 dB EDFA 3 dB 1x8 Rx
Broadcast 2x1 1 dB
Primary
Analog
Redundant Optical
Narrowcast EDFA EDFA Switch
Interactive DWDM Tx 1x16 Return Tx
Services EDFA
DWDM
CMTS Primary
DWDM
HDT
Video Redundant Optical
Server EDFA Switch
MUX DEMUX
Figure 7.4.18 A hubless, DWDM, Deep Fiber Architecture. (After [4].)
log signal and fed to 1310 nm transmitters, which in turn feed multiple or individual optical
nodes. Each optical node typically serves 500 to 1000 subscribers, but nodes serving as few as
100 homes are practical. The fewer homes served per node, the more dedicated bandwidth is
available per subscriber. Because each hub serves on the order of 50,000 subs, the hubs must be
relatively large in order to house all of the equipment and fiber connections.
Several drawbacks exist for operators deploying traditional double-hop architectures for high-
bandwidth subscribers. Aside from the difficulty and expense of locating and building a large
hub site, there are the functional problems of redundancy and the sheer numbers of fibers neces-
sary to feed 500 or so nodes from each hub. For services such as telephony, most operators
require both equipment and fiber path redundancy down to at least the 500 to 1000 home level.
This is difficult to achieve in the double-hop architecture unless there are redundant receivers in
each node fed by redundant fiber rings and transmitters. This arrangement provides redundancy
down to the 100 home level, which might be considered overkill, and is definitely expensive. A
star architecture from the hub to the nodes would mitigate the redundancy problem, but would be
prohibitively expensive in terms of fiber cable, since there would be a unique fiber cable to each
node.
DWDM Architectures
A “hubless” architecture is shown in Figure 7.4.18 [4]. In this case, most of the equipment for-
merly located in the hub (CMTS, HDT, and video servers) has been pulled back to the head end.
Because the forward path signals are passively transmitted through the hub, it is easy to replace
the large hub site with a small, inexpensive cabinet or vault. In Figure 7.4.18, four cabinets serve
the 50,000 subs formerly served by a single large hub. Placing all of the equipment at the head
end has the additional advantage of lowering operational costs and permitting less total equip-
ment to be deployed in the initial stages. In the traditional architecture, it is necessary to locate at
least one CMTS, HDT, and possibly a video server at each hub, regardless of how limited the
demand. With the CMTS and HDT pulled back to the head end, this equipment can be shared
among multiple cabinets/hubs when demand is low.
y
The traditional architecture cost increases rapidly with increasing bandwidth demand, but the
dense wavelength division multiplexing (DWDM) architecture only requires additional, and rela-
tively inexpensive, DWDM transmitters. Similar to SDH, the DWDM architecture also provides
path and equipment protection via redundant optical amplifiers and optical switches. However, it
does not provide all of SDH’s performance monitoring capabilities. Therefore, for certain high-
priority services, such as telephony, a hybrid approach is possible. The low bandwidth, high-pri-
ority services like telephony can be transmitted to the hub/cabinet via SDH, while the high band-
width, non-lifeline services such as VOD can be transmitted via DWDM.
Deep Fiber-Node Segmentation

The number of subscribers served by an optical node for a typical HFC system has steadily
decreased over the last decade [4]. Formerly, 1000 to 2000 home nodes were the norm, but today
new builds and upgrades are more likely to be in the 100 to 500 homes per node range. Pushing
fiber deeper into the network results in improved bandwidth, performance, and reliability, partic-
ularly due to the reduction RF active elements in the coax plant. Completely eliminating RF
amplifiers from the system is an excellent goal, since the system would have a 33 to 50 percent
less active components resulting in greater reliability and reduced power consumption. Com-
pletely passive coax networks are possible for cases of very high subscriber density, such as mul-
tiple dwelling units (MDU’s) with +200 subscribers per mile. However, for more typical densities
of 100 subs/mile, it is necessary to segment down to approximately 35 homes per optical node in
order to create a completely passive coax network when transmitting a full complement of 60 to
80 analog channels. This assumes a four-output node with 51 to 53 dBmV outputs. Unfortu-
nately, for most cable plants, it is difficult or impossible to effectively use all four outputs from
the node, and there is not a tremendous cost difference between nodes utilizing 2, 3, or 4 ports.
Although it is generally better to push fiber as deep as possible, 35 homes per node is probably
not always cost-effective. A practical trade-off is to employ a four-output node passing 100 sub-
scribers and add one line extender to each node output. This provides plenty of RF level at the
home, and up to 50 Mbits/s of dedicated downstream bandwidth. With only a single line extender
in each path, high performance based primarily on the optical link(s) is easily obtained, and
higher order modulation formats like 256-QAM and 1024-QAM can be supported.
Application Considerations
Although DWDM has been applied in cable systems, most implementations demultiplex the sig-
nal at the hub and feed the individual wavelengths to large nodes [4]. This configuration saves
fiber between the head end and hub, but only provides one dedicated narrowcast wavelength per
1000 homes. In the DWDM deep fiber architecture shown in Figure 7.4.18, DWDM is extended
one level further into the network to an optically scalable node (OSN) that may be either strand-
or cabinet-mounted. The signal at the OSN is further split to feed 100 home mini-nodes. Starting
at the headend, the narrowcast signals are transmitted via directly modulated DWDM transmit-
ters. The wavelengths are multiplexed together, optically amplified by an erbium doped fiber
amplifier (EDFA), passed through the hub site, and are demultiplexed at the OSN. In this ulti-
mate configuration, a dedicated wavelength serves each mini-node, providing approximately 300
MHz of narrowcasting 256-QAM channels (~2 Gbits/s or 20 Mbits/s per subscriber). Eventually,
as the broadcast analog spectrum is gradually reduced and the narrowcast spectrum increased,
the entire 50 to 860 MHz band can be utilized to deliver <50 Mbits/s per subscriber. Initially,
y
however, the system can be scaled such that each wavelength is shared among all the mini-nodes
fed by a particular OSN. A simple optical splitter is deployed in the OSN instead of a DWDM
demultiplexer. As bandwidth demand increases, additional DWDM transmitters are added at the
head end, and DWDM demultiplexers replace the optical splitters in the OSN.
The broadcast analog video is redundantly transmitted via an externally modulated 1550 nm
transmitter from the head end to a hub, where it is amplified and split to feed the OSN’s. The
broadcast and demultiplexed narrowcast signals are combined and then transmitted over the
same fiber to a single receiver in the mini-node. The signals are combined via a simple 2:1 opti-
cal coupler, or by a 2:1 DWDM multiplexer, depending on the available loss budget. If a DWDM
mux is necessary because of link budget considerations, then another option is to leave the
broadcast and narrowcast signals separate, and transmit them over individual fibers to separate
receivers in the mini-node. The cost of a 2:1 DWDM mux and 10 km of fiber is greater than the
cost of an additional receiver. In addition, avoiding the 2:1 combiner more easily enables the
equipment associated with a 1:16 split to be packed into the OSN, as opposed to only serving 8
mini-nodes from each OSN. An additional advantage of this approach is that it eliminates poten-
tial problems associated with the CSO and CTB from the narrowcast signals interfering with the
broadcast signal when using a single receiver.
The plant costs of the deep fiber, DWDM architecture are greater than those of traditional
HFC architectures, due to the expanded number of nodes (one per 75 to 100 homes), but this cost
increase is mostly offset by the decrease in RF amplifiers and power supplies. The elimination or
minimization of SDH or ATM transport systems from the head end to hub actually makes the
hubless deep fiber system more cost-effective than the traditional network. The primary benefit
of these next generation architectures is the massively scalable bandwidth, along with the opera-
tional savings in power and RF plant and hub maintenance, and the increased network reliability
due to deeper redundant fiber and fewer total network elements. In addition, the hubless architec-
ture typically can be deployed much more quickly, an important advantage in today’s competitive
service provider environment.
7.4.3c Return Path

Traditionally, multiple return paths are simply combined such that the upstream signal from
thousands of subscribers is fed to a single CMTS or HDT [4]. This results in high noise levels,
and severely limits the available upstream bandwidth. As service penetration increases, operators
must be able to segment the upstream to serve smaller numbers of subscribers. Several methods
exist for segmenting the return path, thus providing dedicated upstream bandwidth to customers.
For the traditional double hop system, upstream DWDM provides basically the mirror image of
the DWDM downstream. Return path transmitters in the mini-node transmit back to the OSN,
where the signal is DWDM muxed with the signals from the other mini-nodes served by the
OSN. Because the signals are analog, it is often necessary to amplify with an EDFA before trans-
mitting back to the head end in order to maintain acceptable performance. At the head end, the
signals are demultiplexed and fed to individual return path receivers. The DWDM upstream
option provides excellent segmentation. But it is not scalable, since the required ITU lasers in
every mini-node and the cabinet EDFA’s combine to make the initial system deployment rela-
tively expensive. The system does provide excellent return path bandwidth of up to approxi-
mately 100 Mbits/s, assuming 16-QAM modulation. This corresponds to 1 Mbits/s peak rate per
subscriber, which may be more than necessary in the early stages of deployment.
y
Security
Module(s)
Operations
Internet
Content
Supporting
Video Headend Hardware & Consumer
Content Software Devices
Other
Content
OpenCable Device
Figure 7.4.19 Basic OpenCable architecture.
7.4.4 Open Cable

The Open Cable (OCAP) development effort began in September 1999 with the following pri-
mary objectives:
• Specify the next generation digital consumer device
• Encourage supplier competition
• Create a practical retail hardware platform
The initial authors included Liberate, OpenTV, Microsoft, PowerTV, Canal+, Sun, CableLabs,
various MSOs, and Excite@Home. The system was designed to provide a high degree of compli-
ance with DVB-MHP version 1 and a lesser amount with the ATSC DTV Application Software
Environment (DASE). Point-of-deployment (POD) security modules were also specified and
supported.
The OCAP specification of the network interface was standardized by the Society of Cable
and Telecommunications Engineers (SCTE) as DVS-313, with IEEE 1394 serving as the inter-
face to consumer devices. (See Figure 7.4.19.) Initial objectives included:
• Support retail product developers to extend OCAP to integrated receivers, personal video
recorders (PVRs), and other devices
• Support interactive service development through work with member companies
Business objectives include enabling service/application portability, preserving supplier diver-
sity, and encouraging innovation. The OCAP system has been designed to be hardware- and
operating system-independent. This was important because different cable companies have dif-
ferent vendor arrangements, which could otherwise limit system functionality.
Example applications for OCAP include the following
• Electronic program guide (EPG)
y
• Impulse pay per view (PPV)

• Video on demand (VOD)
• Interactive sports
• Game shows
Permissions and security issues are addressed in the OCAP specifications using application
authentication, encryption to protect private data, and permissions levels for various applica-
tions.
7.4.5 Regulatory Issues

Cable television (also called CATV or community antenna television) was developed in the late
1940's for communities unable to receive TV signals because of terrain or distance from TV sta-
tions [5]. Cable television system operators located antennas in areas with good reception,
picked up broadcast station signals, and then distributed them by coaxial cable to subscribers for
a fee.
In 1950, cable systems operated in only 70 communities in the U.S. These systems served
about 14,000 homes. Fifty years later there were more than 10,700 systems serving more than 65
million subscribers in more than 32,000 communities. Cable systems are operating in every state
of the U.S. and in many other countries, including Austria, Canada, Belgium, Germany, Great
Britain, Italy, Japan, Mexico, Spain, Sweden, and Switzerland.
Most cable systems are technically capable of offering between 36 and 60 channels. Channel
capacity in the industry has increased dramatically in recent years; some systems now offer in
excess of 100 channels. Most cable subscribers receive service from a system offering more than
54 channels.
The channel capacity of a cable system makes it possible for a cable television operator to
provide many services. In addition to over-the-air television broadcast signals, most systems also
offer diverse program services, including—for example—news, weather, business information,
movies, sports, special entertainment features, and programming designed for specific audiences
such as children, and specific ethnic and racial groups. Within the past few years, some cable
systems have begun offering a full-range of telecommunications services, including high-speed
Internet access and local telephone service.
7.4.5a Initial Jurisdiction and Rules

The Federal Communications Commission first established rules in 1965 for cable systems that
received signals by microwave antenna [5]. In March 1966, the Commission established rules for
all cable systems (whether or not served by microwave). The Supreme Court affirmed the Com-
mission's jurisdiction over cable in United States vs. Southwestern Cable Co., 392 U.S. 157
(1968). The Court ruled that “the Commission has reasonably concluded that regulatory author-
ity over CATV is imperative if it is to perform with appropriate effectiveness certain of its
responsibilities.” The Court found the Commission needed authority over cable systems to assure
the preservation of local broadcast service and to effect an equitable distribution of broadcast
services among the various regions of the country.
y
In March 1972, new rules regarding cable television became effective. These rules required
cable television operators to obtain a certificate of compliance from the Commission prior to
operating a cable television system or adding a television broadcast signal. The rules applicable
to cable operators fell into several broad subject areas:
• Franchise standards
• Signal carriage
• Network program nonduplication and syndicated program exclusivity
• Nonbroadcast or cablecasting services
• Cross-ownership
• Equal employment opportunity
• Technical standards
Cable television operators who originated programming were subject to equal time, Fairness
Doctrine, sponsorship identification, and other provisions similar to rules applicable to broad-
casters. Cable operators were also required to maintain certain records and to file annual reports
with the Commission concerning general statistics, employment, and finances.
In succeeding years, the Commission modified or eliminated many of the rules. Among the
more significant actions, the Commission deleted most of the franchise standards in 1977, sub-
stituted a registration process for the certificate of compliance application process in 1978, and
eliminated the distant signal carriage restrictions and syndicated program exclusivity rules in
1980. In 1983, the Commission deleted its requirement that cable operators file financial infor-
mation. In addition, court actions led to the deletion of the pay cable programming rules in 1977.
7.4.5b 1984 Congressional Policy and Rules

In October 1984, the U.S. Congress amended the Communications Act of 1934 by adopting the
Cable Communications Policy Act of 1984 [5]. The 1984 Cable Act established policies in the
areas of ownership, channel usage, franchise provisions and renewals, subscriber rates and pri-
vacy, obscenity and lockboxes, unauthorized reception of services, equal employment opportu-
nity, and pole attachments. The new law also defined jurisdictional boundaries among federal,
state and local authorities for regulating cable television systems.
7.4.5c 1992 Congressional Policy and Rules

Following the 1984 Cable Act, the number of households subscribing to cable television systems
increased, as did the channel capacity of many cable systems [5]. However, competition among
distributors of cable services did not increase, and, in many communities, the rates for cable ser-
vices far outpaced inflation. Responding to these concerns, Congress enacted the Cable Televi-
sion Consumer Protection and Competition Act of 1992. The 1992 Cable Act mandated a
number of changes in the manner in which cable television is regulated.
In adopting the 1992 Cable Act, Congress stated that it wanted to:
• Promote the availability of diverse views and information
• Rely on the marketplace to the maximum extent possible to achieve such availability
y
• Ensure cable operators continue to expand their capacity and program offerings
• Ensure cable operators do not have undue market power
• Ensure consumer interests are protected in the receipt of cable service
The Commission has adopted various regulations to implement these goals.
7.4.5d 1996 Congressional Policy and Rules

In adopting the Telecommunications Act of 1996, Congress noted that it wanted to provide a pro-
competitive, de-regulatory national policy framework designed to accelerate rapidly private sec-
tor deployment of advanced telecommunications and information technologies and services to
all Americans by opening all telecommunications markets to competition [5]. The Commission
has adopted regulations to implement the requirements of the 1996 Act and the intent of Con-
gress.
7.4.5e Cable Television Defined

Cable television is a video delivery service provided by a cable operator to subscribers via a
coaxial cable or fiber optic link [5]. Programming delivered without a wire via satellite or other
facilities is not “cable television” under the Commission's definitions.
A cable television system operator is any person or group of persons who provides cable ser-
vice over a cable system and directly or through one or more affiliates owns a significant interest
in such cable system, or who otherwise controls or is responsible for, through any arrangement,
the management and operation of such a cable system.
Cable service is the transmission to subscribers of video programming, or other programming
service. This definition includes any subscriber selection required in choosing video program-
ming or other programming service.
A cable system is a facility—consisting of a set of closed transmission paths and associated
signal generation, reception, and control equipment—that is designed to provide cable service
which includes video programming and which is provided to multiple subscribers within a com-
munity. This term does not include a facility that serves only to retransmit the television signals
of one or more television broadcast stations; a facility that serves subscribers without using any
public right-of-way; a facility of a common carrier which is subject in whole or in part, to the
provisions of Title II of the Communications Act, except that such facility shall be considered a
cable system to the extent such facility is used in the transmission of video programming directly
to subscribers, unless the extent of such use is solely to provide interactive on demand services;
an open video system, or any facilities of any electric utility used solely for operating its electric
utility system.
Cable services are often provided in tiers. A tier is a category of cable service or services pro-
vided by a cable operator for which a separate rate is charged by the cable operator. There are
three types of cable service:
• Basic service
• Cable programming service
• Per-channel or per-program (sometimes called pay-per-view) service
y
Basic service is the lowest level of cable service a subscriber can buy. It includes, at a mini-
mum, all over-the-air television broadcast signals carried pursuant to the must-carry require-
ments of the Communications Act, and any public, educational, or government access channels
required by the system's franchise agreement. It may include additional signals chosen by the
operator. Basic service is generally regulated by the local franchising authority (the local or state
entity empowered by federal, state, or local law to grant a franchise to a cable company to oper-
ate in a given area).
Cable programming service includes all program channels on the cable system that are not
included in basic service, but are not separately offered as per-channel or per-program services.
Pursuant to a 1996 federal law, the rates charged for cable programming services tiers provided
after March 31, 1999 are not regulated. There may be one or more tiers of cable programming
service.
Per-channel or per-program service includes those cable services that are provided as single-
channel tiers by the cable operator, and individual programs for which the cable operator charges
a separate rate. Neither of these services is regulated by the local franchising authorities or the
Commission.
7.4.5f Signal Carriage Requirements

The 1992 Cable Act established new standards for television broadcast station signal carriage on
cable systems [5]. Under these rules, each local commercial television broadcast station was
given the option of selecting mandatory carriage (must-carry) or retransmission consent (may
carry) for each cable system serving the same market as the commercial television station. The
market of a television station is established by its Area of Dominant Influence (ADI), as defined
by Arbitron and/or modified by the Commission. Every county in the country is assigned to an
ADI, and those cable systems and television stations in the same ADI are considered to be in the
same market. Upon the request of a television station or a cable system, the Commission has the
authority to change the ADI to which a station is assigned. As a result of Arbitron abandoning
the television research business, the Commission determined that, effective January 1, 2000, the
market of a television station shall be its Designated Market Area (DMA) as determined by
Nielsen Media Research.
Every three years, each local commercial television station has the right to elect either must-
carry or retransmission consent. The initial election was made on June 17, 1993, and was effec-
tive on October 6, 1993. All subsequent elections occurred (and will continue to occur) every
three years.
Generally, if a local commercial television station elects must-carry status, it is entitled to
insist on cable carriage in its local market. Each cable system with more than 12 channels must
set aside up to one-third of its channel capacity for must-carry stations. For example, if a cable
system has 60 channels, it must set aside 20 of those channels for must-carry stations. If there are
25 stations in the market which elected must-carry, the cable operator may choose 20 to carry.
On the other hand, if only 15 stations elected must-carry in the market, the cable system would
have to carry all 15 of these stations. A must-carry station has a statutory right to a channel posi-
tion, usually its over-the-air channel number, or another channel number on which it has histori-
cally been carried.
A cable system is not permitted to carry a commercial station without the station's consent.
Therefore, if the local commercial television station elects retransmission consent, the cable sys-
y
tem must obtain that station's consent prior to carrying or transmitting its signal. Except for
“superstations,” a cable system may not carry the signal of any television broadcast station that is
not located in the same market as the cable system without that broadcaster's consent. Supersta-
tions are transmitted via satellite, usually nationwide, and the cable system may carry such sta-
tions outside their local market without their consent. The negotiations between a television
station and a cable system are private agreements which may, but need not, include some form of
compensation to the television station such as money, advertising time, or additional channel
access.
Noncommercial Educational Television Stations

Every cable system across the country must carry at least one local noncommercial educational
(NCE) station. A noncommercial station which places a Grade B signal over a cable system's
principal headend, or whose city of license is within fifty miles of the cable system's principal
headend, is considered “local” for this purpose. Cable systems with more than 36 channels may
be required to carry all local noncommercial educational television stations that request carriage.
Any cable system operating in a market where no local NCE station is available is required to
import one NCE station.
Low Power Television Stations

The 1992 Cable Act provides mandatory carriage for “qualified” low power television stations
(LPTV) in certain situations. An LPTV station has to meet certain qualifications specified in the
Act and incorporated into the Commission's Rules, before it is qualified for the right to must-
carry. If an LPTV station is qualified, it may assert must-carry rights, and, provided the cable
operator has not met its mandatory carriage obligations, the LPTV station must be carried. Oth-
erwise, an LPTV station must negotiate for carriage under the retransmission consent provisions
or under the leased commercial access provisions.
Radio Programming
While the 1992 Cable Act's must-carry provisions only apply to local commercial and noncom-
mercial educational television stations, the Act's retransmission consent provisions apply to all
commercial broadcast stations. Many cable systems carry radio stations as an “all-band” offer-
ing, meaning that as with any standard radio receiver, all stations which deliver a signal to the
antenna are carried on the system. The Commission only requires consent from those radio sta-
tions within 57 miles of the cable system's receiving antenna. Thus, even though a cable opera-
tor's antenna may pick up a station's signal, operators are not required to obtain the consent of
stations outside of the 57 mile zone unless the station affirmatively seeks retransmission consent.
Manner of Carriage
Subject to the Commission's network nonduplication, syndicated exclusivity, and sports broad-
casting rules, cable systems must carry the entirety of the program schedule of every local televi-
sion station carried pursuant to the mandatory carriage provisions or the retransmission consent
provisions of the 1992 Cable Act. A broadcaster and a cable operator may negotiate for partial
carriage of the signal where the station is not eligible for must carry rights, either because of the
station's failure to meet the requisite definitions or because the cable system is outside the sta-
y
tion's market. In those situations where the carriage in the entirety rule applies, the primary video
and accompanying audio of all television broadcast stations must be carried in full, without alter-
ation or deletion of their content. Ancillary services such as closed captioning and program-
related material in the vertical blanking interval must be carried. However, other information
contained in the vertical blanking interval need not be carried.
7.4.5g Technical Standards

To ensure the delivery of satisfactory television signals to cable subscribers, the Commission has
adopted various technical performance standards that local franchising authorities are generally
authorized to enforce through their franchising process. Cable operators are required to establish
a complaint resolution process and to advise their subscribers about it annually.
In addition to establishing new standards for the delivery of color signals and of closed cap-
tioned data, the Commission generally preempted conflicting local standards. Cable systems
with fewer than 1,000 subscribers, as well as those serving rural areas, may negotiate with their
respective franchising authorities for certain lower standards.
7.4.6 References
1. Ciciora, Walter S.: “Cable Television,” in NAB Engineering Handbook, 9th ed., Jerry C.
Whitaker (ed.), National Association of Broadcasters, Washington, D.C., pp. 1339–1363,
1999.
2. Fink, D. G., and D. Christiansen (eds.): Electronic Engineer’s Handbook, 2nd ed.,
McGraw-Hill, New York, 1982.
3. Baldwin, T. F., and D. S. McVoy: Cable Communications, Prentice-Hall, Englewood Cliffs,
N.J., 1983.
4. Bonang, C., and C. Auvray-Kander: “Next Generation Broadband Networks for Interactive
Services,” in Proceedings of IBC 2000, International Broadcasting Convention, Amster-
dam, 2000.
5. “Cable Television Information Bulletin,” Federal Communications Commission, Washing-
ton, D.C., June 2000.
g g
Chapter
7.5
Satellite Delivery Systems
Carl Bentz, K. Blair Benson
7.5.1 Introduction
The first commercial satellite transmission occurred on July 10, 1962, when television pictures
were beamed across the Atlantic Ocean through Telstar 1. The launch vehicle, however, lacked
sufficient power to place the spacecraft into a stationary position. Three years later, after consid-
erable progress in the development of rocket motors, INTELSAT saw its initial craft, Early Bird
1, launched into a geostationary orbit, and a rapidly growing communications industry was born.
In the same year, the USSR inaugurated the Molnya series of satellites, which traveled in more
elliptical orbits, to better meet the needs of that nation and its more northerly position. The Mol-
nya satellites were placed in orbits inclined about 64° relative to the equator, with an orbital
period half that of the earth.
From these humble beginnings, satellite-based audio, video, and data communications have
emerged to become a powerful force in communications across all types of industries, from mil-
itary logistics and support to consumer television.
7.5.2 The Satellite Communications Link

A satellite relay system involves three basic elements. On the ground is an up-link transmitter
station, which beams signals to the satellite. The satellite, as the space segment of the system,
receives, amplifies, and retransmits the signals back to earth. The down-link receiving station
completes the system and with the up-link forms the earth segment. (See Figure 7.5.1.)
Two frequency bands are assigned for standard video satellite communications. C-band oper-
ates on 6 GHz frequencies for transmission of information to the satellite. The electronics pack-
age aboard the craft converts the signals to 4 GHz frequencies for the return to earth. At much
higher frequencies, the Ku-band up-link transmits in the region of 14 GHz, with conversion at
the satellite to down-link channels at 12 GHz. Satellites may provide service for either C- or Ku-
band, or both bands on the same spacecraft. For military and maritime applications, other fre-
quencies are typically used.
Because of the high frequencies used in the system, satellite communications are considered a
microwave radio service. This introduces some concerns with regard to satellite-link operations,
including the following:
7-137
y y
• A line-of-sight path between the transmit-

ter and the receiver is required.
• Attenuation of signals resulting from
meteorological conditions, such as rain
and fog, is problematic for Ku-band oper-
ation, but less troublesome for C-band
operation.
Arrangements for shielding the antennas
from terrestrial interference include siting
the antenna in a depression in the terrain,
either constructed or natural, or building a
wall of material that is opaque to microwave
energy around the antenna. Any foliage at
the antenna site threatens possible attenua-
tion effects if it extends into the beam path of
the antenna or through scattering of the sig-
nal energy by individual leaf surfaces.
Adequate received signal strength, based Figure 7.5.1 The satellite communications sys-
on the inverse-square law, requires highly tem consists of the up-link and down-link, which
directional transmitting and receiving anten- form the earth segment, and the satellite, which
nas, which, in turn, need significant accuracy forms the space segment.
of aiming. Most antennas are parabolic in
form and can be moved for best signal recep-
tion. The development of the phased array introduced a steerable radiation pattern from a rela-
tively fixed, flat-surface antenna primarily for Ku-band use.
The effects of galactic and thermal noise sources on low-level signals require electronics for
satellite service with exceptionally low noise characteristics.
7.5.2a The Satellite

It is convenient to discuss the satellite communications link by starting with its middle point, the
satellite itself. By understanding the relay station requisites for receiving and transmitting, it is
easier to realize the requirements placed on the up-link and down-link sections. Like all relay sta-
tions, the communications spacecraft contains antennas for receiving and retransmission. A low-
noise amplifier (LNA) boosts the signal from the antenna before frequency conversion for trans-
mission. A high-power amplifier (HPA) feeds the signal to a directional antenna with a radiation
pattern designed to cover a prescribed area of the earth. The electronics receives power generated
with solar cells and buffered with storage batteries, which are recharged by the array of solar
devices. The batteries are the only source of power when the satellite passes into the earth’s
shadow at certain times of the year. Finally, a guidance system stabilizes the attitude of the craft,
because as it rotates around the earth, the craft would otherwise begin to tumble and its antennas
would no longer point toward the earth. As a part of the guidance system, the craft also contains
rocket engines for station keeping—maintaining its position in the geostationary arc. The overall
mass and volume of all components of the craft must be limited, however, largely because of
launch considerations.
y y
Antennas
The antenna structure of a communications satellite consists of several different antenna sec-
tions. One receives signals from earth; another transmits those signals back to earth. The trans-
mitting antenna may be constructed in several sections to carry more than one signal beam.
Finally, a receive-transmit beacon antenna provides communication with a ground station to con-
trol the operation of the satellite. The control functions include turning parts of the electronics on
and off, adjusting the radiation pattern, and maintaining the satellite in its proper position.
The design of the complex antenna system for a satellite relies heavily on the horizontal and
vertical polarizations of signals to keep incoming and outgoing information separated. Multiple
layers in carefully controlled thicknesses of reflecting materials, which are sensitive to signal
polarizations, can be used for such purposes. Also, multiple-feed horns can develop more beams
to earth. Antennas for different requirements may combine several antenna designs, but nearly
all are based on the parabolic reflector, because of two fundamental properties of the parabolic
curve:
• Rays received by the structure that are parallel to the feed axis, are reflected and converged at
the focus
• Rays emitted from the focal point are reflected and emerge parallel to the feed axis
Special cases may involve spherical and elliptical reflectors, but the parabolic is most common.
7.5.2b Transponders
From the antenna, a signal is directed to a chain of electronic equipment known as a transponder.
This unit contains the circuits for receiving, frequency conversion, and transmission. Some of the
first satellites launched for INTELSAT and other systems contained only one or two transponder
units. However, initial demand for satellite link services forced satellite builders to develop more
economical systems with multiple-transponder designs. Such requirements meant refinements in
circuitry that would allow each receiver-transmitter chain to operate more efficiently (less power
draw). Multiple-transponder electronics also meant a condensation of the physical volume
needed for each transponder with as small as possible an increase in the overall size and mass of
the satellite to accommodate the added equipment.
General purpose satellites placed in orbit have 12 or 24 transponders for C-band units, each
with 36-MHz bandwidths. However, wide variations in the number of transponders and in their
operating bandwidths, exist. Some Ku-band systems use fewer transponders with wider band-
widths, but others have 40 or more transponders of 27 MHz bandwidth. Identification of the
transponders is described by an agreed-upon frequency plan for the band.
Users of satellite communication links are assigned to transponders generally on a lease basis.
Assignments usually leave one or more spare transponders aboard each craft, allowing for
standby capabilities in the event of a transponder failure.
By assigning transponders to users, the satellite operator simplifies the design of up-link and
down-link facilities. An earth station controller can be programmed for the transponders of inter-
est. For example, a television station may need access to several transponders from one satellite.
The operator enters only the transponder number (or carrier frequency) of interest. The receiver
handles retuning of the receiver and automatic switching of signals from the dual-polarity feed
horn on the antenna. Controllers can also be used to move an antenna from one satellite to
another automatically.
y y
Each transponder has a fixed center frequency and a specific signal polarization according to
the frequency plan. For example, all odd-numbered transponders might use horizontal polariza-
tion, while even-numbered ones might be vertically polarized. Excessive deviation from the cen-
ter carrier frequency by one signal does not produce objectionable interference between
transponders and signals because of the isolation provided by cross-polarization. This concept
extends to satellites in adjacent parking spaces in the geosynchronous orbit. Center frequencies
for transponders on adjacent satellites are offset in frequency from those on the first craft. An
angular offset of polarization affords additional isolation. The even and odd transponder assign-
ments are offset by 90° from one another. As spacing is decreased between satellites, the polar-
ization offset must be increased to reduce the potential for interference.
While this discussion has centered on the down-link frequency plan for a C-band satellite, up-
link facilities follow much the same plan, except that up-link frequencies are centered around 6
GHz. A similar plan for Ku-band equipment uses up-linking centered around 14 GHz and down-
link operation around 12 GHz.
7.5.2c Power Sources

Providing power to the electronics aboard the spacecraft is a major engineering challenge. Fortu-
nately, the constant evolution of solar cells and storage batteries have provided satellite systems
with various types of efficient power sources.
During the launch phase, control circuitry obtains power from rechargeable batteries on the
craft. After the final orbital position is achieved, the control station activates the deployment of
arrays of solar cells. When the solar panels are functional, all operational power is derived from
the panels, as well as current to recharge the batteries for backup during times of eclipse, caused
when the earth blocks illumination of the panels from direct sunlight.
Common solar cells include silicon with boron or arsenic doping to create pn junctions. The
light-to-electricity conversion efficiency of this material is approximately 10 percent. Aluminum
or gallium arsenide cells with higher efficiency are also used. The array is created by connecting
cells in series to achieve a larger voltage and a number of series groups in parallel for greater cur-
rent capability. An array of 2000 cells covering 1 m2 would produce more than 100 W.
With any type of photovoltaic material, the output of a cell is determined by a number of fac-
tors. Primary among them are:
• The surface area of the cell exposed to light
• The average solar flux per unit area (light intensity)
• The temperature of the cell material
• The variation of the light source from a position perpendicular to the cell surface
The initial capabilities, as well as the end-of-life capacity, of a solar power source must take
these factors into account and allow for them by designing in accordance with worst-case condi-
tions. One such consideration is the proper positioning of the solar array when operation of the
satellite is initiated. The variation of solar energy with the seasons varies approximately 8 per-
cent between the solar equinoxes and solstices, because of the varying angle of incidence of the
sunlight. This seasonal change produces a variation in temperature as well, and over time, the
bombardment by proton radiation from solar flares reduces the efficiency of electric current pro-
duction.
y y
Two types of solar panels are in general use. One is a flat configuration arranged as a pair of
deployable panel wings that extend to either side of the satellite body. In the other type, the solar
cells are attached to the barrellike body of the satellite.
Each approach has advantages and disadvantages. On the winglike panels, approximately
18,000 cells are used to maintain a 1200 W output capability at end-of-life. The other configura-
tion uses approximately 80,000 cells mounted on a 2 m or greater diameter cylinder that is 5 m or
more in height. The wing approach is usually the more efficient of the two.
The wing keeps all cells illuminated at all times, producing more electric current. However,
all cells are subject to a higher rate of bombardment by space debris and by solar proton radia-
tion, and to a higher average temperature (approximately 60°C) because of constant illumination.
The rotating-drum design places a small portion of all cells perpendicular to the sun’s rays at
any one time. Cells that are not perpendicular to the light source produce power as long as they
are illuminated, but the amount is reduced and depends on the angle of light incidence. The indi-
vidual cells are not all exposed in one direction simultaneously and are less likely to be struck by
space debris. Also, because cells on the drum are not exposed to sunlight at all times, their aver-
age temperature is significantly less, approximately 25°C, which extends their operating life-
time.
The output current, voltage, and power from the solar panels undergo variation, depending on
the operating conditions. Without a means of controlling or regulating the power, the operation
of the electronics package would also vary. An unregulated power bus is simpler and takes less of
the allowable mass, or mass budget, of the craft. System electronics circuitry must, however,
include on-board regulators to maintain consistent RF levels. The alternative is to use bus regula-
tion with more complex circuitry, a somewhat lower output power, and reduced reliability
because of the additional components in the system.
Some variation can be accommodated by the storage batteries in either case. The most serious
need for batteries following insertion into orbit of the geosynchronous satellite occurs approxi-
mately 84 days per year. The earth passes between the sun and the satellite, causing an eclipse.
The blackout period can last as long as 70 min within one day. During these periods, the need for
a rechargeable supply is essential. Nickel-cadmium batteries have played a major role in power-
ing such spacecraft.
Power to the electronics on the craft requires protective regulation to maintain consistent sig-
nal levels. Most of the equipment operates at low voltages, but the final stage of each transpon-
der chain ends in a high power amplifier. The HPA of C-band satellite channels may include a
traveling-wave tube (TWT) or a solid-state power amplifier (SSPA). Ku-band systems rely pri-
marily on TWT devices. Traveling-wave tubes, similar to their earth-bound counterparts,
klystrons, require multiple voltage levels. The filaments operate at 5 V, but beam-focus and elec-
tron-collection electrodes in the device require voltages in the hundreds and thousands of volts.
To develop such a range of voltages, the power supply must include voltage converters.
From these voltages, the TWT units produce output powers in the range of 8.5 to 20 W. Most
systems are operated at the lower end of the range for increased reliability and greater longevity
of the TWTs. In general, lifetime is assumed to be 7 years.
At the receiving end of the transponder, signals coming from the antenna are split into sepa-
rate bands through a channelizing network, which directs each input signal to its own receiver,
processing amplifier, and HPA. At the output, a combiner brings all the channels together again
into one signal to be fed to the transmitting antenna.
y y
7.5.2d Control of Satellites

The electronics and attitude control section of the satellite performs a number of key tasks,
including:
• Turning transponders on and off, as directed from the ground station
• Adjustments of antennas to keep the beams focused to form the footprint of signal levels des-
ignated for each satellite system
• Controlling the small package of rocket engines in the craft to maintain the correct parking
position in the geosynchronous orbit
Much of the attitude control is microprocessor-controlled, but the ground station can override
on-board equipment to perform necessary corrections.
Communications with the satellite from the ground often use other frequencies in addition to
those relayed through the transponders. Common arrangements include VHF and S-band links.
Control information and telemetry data reporting on the status of the satellite system to the
ground station are transmitted as encoded digital signals. Because of the proximity of satellites,
strict protocols are needed to prevent one satellite from responding when the neighboring device
is being interrogated. Protection against erroneous control signal reception is improved through
narrow-band reception, or wide-band reception combined with spread-spectrum transmission
and digitally encoded data.
Positioning and Attitude Control

Maintaining the position of the spacecraft in its orbit is called station keeping. With the altitude
of the craft at about 23,000 mi, some motion of the craft is possible without requiring major redi-
recting of earth station antennas. In most cases, a window of sides approximately 0.1° in length
as seen from earth (equivalent to 75 km) is considered acceptable. If the craft moves outside the
window, reception S/N is degraded.
Movement of a craft from its assigned location is caused by forces of the surrounding envi-
ronment. One significant factor is gravitational effects. It is the normal pull of gravity upon the
fast-moving craft (centrifugal force) that keeps the satellite in its orbit. The force of gravity has
been found to vary at sonic locations around the equator. As a result, satellites positioned near
the longitudes of gravitational anomalies become advanced or retarded from their correct loca-
tion, and corrections are periodically required.
Other factors to be accounted for are the gravitational attractions of the moon and sun. The
pull of the moon is approximately three times stronger than that of the sun. Most of the error to
the orbit caused by these bodies is perpendicular to the equatorial plane, that is, in the inclination
of the orbit with respect to the equator. Slight errors in the longitude of the craft result from solar
and lunar attractions, but these are compensated with a change in the dimensions of the orbit.
Lesser factors that disturb the position of the spacecraft and may require occasional correc-
tions include:
• The pressure on the solar power panels by the solar radiation flux
• The force caused by radio frequency power radiated from the satellite
• The magnetic field of the earth
y y
Figure 7.5.2 Attitude of the spacecraft is determined by pitch, roll, and yaw rotations around three
reference axes.
• The self-generated torques related to antenna displacements, the solar arrays, and changing
masses of on-board fuel supplies
The results of meteorite impacts on spacecraft are considered, but have not been experienced sig-
nificantly by any satellite to date.
The attitude of the spacecraft is defined as the combination of pitch, roll, and yaw. These are
motions that can occur without causing a change in the orbital position of the satellite. As illus-
trated in Figure 7.5.2:
• Pitch is the rotation of the craft about an axis perpendicular to the plane of the orbit
• Roll is rotation of the craft about an axis in the plane of the orbit
• Yaw is rotation about an axis that points directly toward the center of the earth
Satellite attitude is determined by the angular variation between satellite body axes and these
three reference axes.
Attitude control involves two functions. First, it is necessary to rotate that part of the satellite
that must point toward the earth around the pitch axis. Rotation must be precisely one revolution
per day or 0.25° per minute. Second, the satellite must be stabilized against torque from any other
source. To perform these functions, numerous detectors measure the sense of the satellite with
respect to earth horizons, the sun, specified stars, gyroscopes, and radio frequency signals. Atti-
tude correction systems may involve a control loop with the ground control station or may be
designed as a closed loop by storing all required attitude reference information within the satel-
lite itself.
Relative to the first function listed, satellites can be spin-stabilized gyroscopically by rotating
the body of the satellite in the range of 30 to 120 times per minute. This effectively stabilizes the
craft with respect to the axis of the spin and is used with cylindrical-type satellites. Although a
relatively simple approach, it has a major drawback—the antenna would not remain pointed
toward earth if it were mounted on the rotating portion of the craft. The solution is to de-spin the
section holding the antenna at one revolution per day about the pitch axis to maintain a correct
antenna heading.
y y
7.5.2e The Up-Link

The ground-based transmitting equipment of the satellite system consists of three sections: base-
band, intermediate frequency (IF), and radio frequency (RF). The baseband section interfaces
various forms of incoming signals with the design of the satellite being used. Signals provided to
the baseband section may already be in a modulated form with modulation characteristics
decided or determined by the terrestrial media bringing signals to the up-link site. Depending on
the nature of the incoming signal—voice, data, or video—some degree and type of processing
will be applied. In many cases, multiple incoming signals can be combined into a single up-link
signal through multiplexing.
When the incoming signals are in the correct format for passage through the satellite, they are
applied to a modulator, which converts the entire package upward to an intermediate frequency.
The use of an IF section has several advantages. First, a direct conversion between the baseband
and the output frequency presents difficulties in maintaining frequency stability of the output
signal. Second, any mixing or modulation step has the potential of introducing unwanted by-
products. Filtering in the IF may be used to remove spurious signals resulting from the mixing
process. Third, many terrestrial microwave systems include a 70 MHz IF section. If a signal is
brought into the up-link site by terrestrial microwave, it becomes a simple matter to connect the
signal directly into the IF section of the up-link system.
From the IF section, the signal is converted upward again, this time to the output frequency
before application to an HPA. Several amplifying devices are used in HPA designs, depending on
the power output and frequency requirements. For the highest power level in the C- or Ku-band,
klystrons are employed. Different devices are available with pulsed outputs ranging from 500 W
to 5 kW and a bandwidth capability of 40 MHz or greater.
Another type of vacuum power device for HPA designs is the traveling-wave tube While sim-
ilar in some aspects of operation to a klystron, the TWT is capable of amplifying a band of sig-
nals at least 10 times wider than the klystron. Thus, one TWT system can be used to amplify the
signals sent to several transponders on the satellite. With output powers from 100 W to 2.5 kW or
greater, the bandwidth capability of the TWT offers distinct advantages.
Solid-state amplifiers, based on field-effect transistor (FET) and other technologies, can be
used for both C- and Ku-band up-link HPA systems. These devices offer wide-band performance
and excellent reliability. They are also useful in amplifying signals directed at more than one
transponder.
7.5.2f The Down-Link

Satellite receiving stations, like up-link equipment, interface ground-based equipment to the sat-
ellite. To serve this function, the earth station—as a TVRO (television receive only) or a TR
(transmit-receive) facility—consists of the receiving antenna with a low-noise amplifier. A tuner,
an IF section, and a baseband output stage complete the basic equipment package.
Antennas for ground-based receiving stations fall into several categories. First, the antenna
used for transmitting to the satellite can also be used as a receiving antenna. Unlike terrestrial
two-way radio systems, where one station transmits and then the other, the satellite system par-
ticipates in both functions simultaneously. This feature is possible because of the offset between
up-link and down-link frequencies. To make operation possible, a circulator separates and
directs the transmitted and received signals to the correct sub-systems.
y y
Receiving antennas for commercial applications, such as radio/TV networks, CATV net-
works, and special services or teleconferencing centers, range from less than a meter to 7 m or
more in diameter. Antennas for consumer and business use are even smaller, depending on the
type of signal being received and the quality of the signal to be provided by the down-link instal-
lation. If the signals being received are strictly digital in nature, smaller sizes are usually suffi-
cient.
Antenna size for any application is determined primarily by the type of transmission, the band
of operation, the location of the receiving station, typical weather in the receiving station locale,
and the quality of the output signal required from the down-link. As indicated previously, digital
transmissions usually allow a smaller main reflector to be used because of the transmission char-
acteristics of digital systems. The data stream periodically includes information to check the
accuracy of the data and, if errors are found, to make corrections. If errors are too gross for error
correction, the digital circuitry may provide error concealment techniques to hide the errors.
Absolute correction is less critical for entertainment-type programming functions, such as televi-
sion and audio programming. The nature of the application also helps to determine if the antenna
must be strictly parabolic or if one of the spherical types will be sufficient.
Regarding the frequency band of operation, the lower the signal frequency, the larger the
antenna reflector must be for the same antenna gain. However, there are compensating factors
such as the required output signal quality, station location, and local environment. Generally, the
gain figure and directivity of larger reflectors are greater than those of smaller reflectors.
One of the most critical sections of the receiver is the low-noise amplifier or low-noise con-
verter, the first component to process the signal following the antenna. The amplifier must not
add noise to the signal, because once added, the noise cannot be removed. LNAs are rated by
their noise temperature. The cost of LNAs increases significantly as the temperature figure (and
noise) goes down.
Following the LNA, the receiver tuner converts signals to an IF frequency. As with the up-link
equipment, an output at 70 MHz is useful to connect to terrestrial microwave equipment if
desired. A second conversion takes video, audio, or data to baseband signals in a form most con-
venient for those using the communications link.
7.5.3 Performance Considerations

The performance of any communications link, irrespective of its bandwidth, is determined by the
carrier-to-noise ratio (C/N) [2]. In the case of a broadcast satellite system, C/N is determined by
the following elements:
• The effective isotropic radiated power (EIRP) of the satellite
• Frequency of operation
• Home terminal figure of merit (G/T)
• Degradations resulting from rain and other environmental factors
For a fixed G/T, improving the system performance requires increasing the EIRP, as illus-
trated in Figure 7.5.3. The curves show that the C/N depends on the strength of the transmitter
power and on the size of the reception area. Increasing the transmitter power can be costly,
whereas limiting the coverage reduces the service and the income derived from it.
y y
Figure 7.5.3 Carrier-to-noise ratio is determined by the spacecraft’s effective isotropic radiated
power (EIRP) and the home receiver figure of merit (G/T). For a fixed receiver quality, C/N can be
improved by increasing the EIRP. This may be accomplished by either increasing the satellite
transmitter power or reducing coverage area (increasing antenna gain). These curves assume an
operating frequency of 12 GHz. For a typical receiver noise figure of 3.0 dB, the antenna diame-
ters would range from 0.36 m for the 3 dB/K case to 2.5 m for the 20 dB/K case. (After [2].)
Because the picture quality is related to C/N, a common method of improving the picture is to
use higher figures of merit for the home terminal. This can be achieved with a larger antenna that
has a higher gain or with an LNA with lower noise-temperature characteristics.
The relationship between C/N and home-terminal antenna size is independent of carrier fre-
quency. It is independent because the antenna gain of the home terminal depends on the square
of the frequency in exactly the same way as the free-space loss. Therefore, in the equation for
transmission power, the two variables effectively cancel each other, all other things being equal.
On the other hand, for a given power level in a satellite and a given coverage area, the move to
higher carrier frequencies requires larger antennas to achieve the performance because of atmo-
spheric and rain losses. Thus, to maintain picture quality with atmospheric losses, high-power
signals have to be used in the satellite to accommodate the smaller home-receiver antennas.
y y
Figure 7.5.4 In the 12-GHlz band, analog video performance will degrade during rain. This exam-
ple shows performance for a system designed to have a S/N of 50 dB during clear weather. Cli-
mate types range from cold and dry areas such as Maine (B) to tropical areas such Florida and
Louisiana (E). (After [2].)
7.5.3a Home Terminal Variables

To deliver a quality picture to an analog or digital home terminal, two random variables must be
addressed:
• The subjective perception of picture quality
• Attenuation caused by rain
Both of these elements ultimately determine how successfully the system performs.
Studies of subjective evaluations of picture quality have been made worldwide, with the
results hinging on video S/N for analog systems and the video compression technology used for
digital systems, when operating above the “cliff effect”.
Rain Attenuation
Attenuation caused by rain can be a difficult problem to deal with. Both the rate and volume of
rain vary extensively around the U.S. and elsewhere. A rain attenuation model, drawn up by the
National Aeronautics and Space Administration for the 12 GHz broadcast band, summarizes the
different climates in the U.S. The hardest to deal with is that of the southeast, particularly in Lou-
isiana and Florida.
To illustrate how the climate curves reflect the video S/N, these parameters are plotted in Fig-
ure 7.5.4 against a percentage of time for each climate type. The picture quality in the worst cli-
mate would be usable until the C/N dropped below the FM threshold. This would happen about
0.2 percent of the time in the southeastern U.S. This amounts to almost 18 h a year.
y y
Figure 7.5.5 Receiving station G/T is a function of antenna diameter, receiver temperature, and
antenna temperature. In heavy rain, the antenna temperature increases from a clear-weather
value of perhaps 50K to a maximum value of 290K (ambient). (After [2].)
As mentioned previously, a critical characteristic for receiving high quality pictures from a
direct-broadcast satellite is the figure of merit (G/T) of the home terminal. At a particular fre-
quency, this figure simply depends on the size of the antenna and on the system noise tempera-
ture of the LNA/receiver system. The system noise temperature, for its part, is affected by rain.
Rain not only attenuates the signal but also increases the antenna temperature. This temperature
must be added to the LNA temperature to set the system noise temperature. Variation of the
antenna temperature is particularly distressing because the system temperature margin that can
account for a high G/T may be obliterated just when the margin is needed the most. System tem-
perature in clear weather and antenna size are, thus, not perfectly interchangeable.
For the same ratio, a larger antenna is better than a lower clear-weather system temperature,
because the increased gain of a larger antenna is not affected by rain.
A plot of the figure of merit versus antenna diameter for different low-noise amplifier tem-
peratures in both clear weather and heavy rain yields the results shown in Figure 7.5.5. A 1.8 m
antenna with a 480K temperature and a 1 m antenna with a 120K temperature both give the same
clear-weather figure of merit: 17 dB/K. However, in heavy rain the larger antenna G/T deterio-
rates to only 16 dB, whereas the smaller antenna system drops all the way to 13 dB.
y y
7.5.4 Digital Transmission

Many satellite systems are used for the transmission of digital information [1]. A typical digital
transmission up-link consists of a modem, upconverter, and power amplifier. The modem con-
verts digital information to and from a modulated carrier. The center-frequency of the modulated
carrier positions the signal within a satellite transponder. The upconverter translates the modu-
lated carrier to a satellite frequency and, thus, selects the transponder of the satellite.
The modem is the earth station component used to convert digital information to a format
suitable for transmission by satellite. The modem accepts a digital data input signal and outputs
an intermediate frequency, typically a range centered on either 70 or 140 MHz, containing the
modulated digital information.
Transmitted digital data is first applied to the encoder section of the modem for forward error
correction (FEC) encoding. This process appends additional bits to the original information to
provide error detection and correction capabilities. The resulting signal is scrambled using a
standard algorithm to ensure random data.
The aggregate data (i.e., original data plus error correction bits) is applied to the modulator
for frequency modulation onto an IF carrier. The IF carrier is selectable, typically in the range of
50 to 90 MHz for 70 MHz operation, or 100 to 180 MHz for 140 MHz operation. The center fre-
quency of the modem modulator is tuned to position the signal within the satellite transponder.
A satellite digital transmission system is characterized by the following parameters:
• Data rate. The data rate is the number of bits per second transmitted by the modem.
• Data interface. Modems normally support a number of data interfaces. The data interface
refers to the connector and signal levels.
• Code rate. The code rate refers to the FEC encoding scheme used. In some modems, the code
rate is selectable. The code rate configuration is referred to as “m/n.” In this context, m refers
to the number of original bits per block of transmitted bits; n refers to the number of original
bits plus error correction bits per block of transmitted bits. Thus, a code rate of 3/4 means that
for every three data bits, four data bits are transmitted. Thus a 1024 kbits/s modem operating
with a code rate of 3/4 would transmit 1365 kbits/s over the satellite channel.
• Modulation system. The modulation scheme refers to the method of indicating data bits.
Two common modulation schemes are employed in satellite transmission systems: bi-phase
shift keying (BPSK) and quadrature phase shift keying (QPSK). These modulation schemes
generate a periodic set of phase shifts referred to as symbols.
The symbol rate (i.e., the number of symbols per second) and data rate determines the amount of
bandwidth required in the channel.
In BPSK, two phase-shifts are used to represent two states. For this case, the symbol rate is
equal to the transmission rate. The QPSK scheme uses four phase shifts, thus transmitting two
bits per symbol. Therefore, QPSK uses a symbol rate that is half the transmission rate. QPSK
requires less bandwidth than BPSK but demands increased performance from the channel.
y y
7.5.5 References
1. Cook, James H., Jr., Gary Springer, Jorge B. Vespoli: “Satellite Earth Stations,” in NAB
Engineering Handbook, Jerry C. Whitaker (ed.), National Association of Broadcasters,
Washington, D.C., pp. 1285–1322, 1999.
2. Kase, C. A., and W. L. Pritchard: “Getting Set for Direct-Broadcast Satellites,” IEEE Spec-
trum, IEEE, New York, N.Y., vol. 18, no. 8, pp. 22–28, 1981.
g g
Chapter
7.6
Content Distribution
Chandy Nilakantan, Kishore Manghnani, Scott Johnson
7.6.1 Introduction
The Internet is not just ubiquitous, its use for conducting commerce has become pervasive. Great
strides have been made in the areas of providing widespread connectivity, increased capacity for
moving richer content and adequate security for maintaining privacy. One of the largest remain-
ing challenges is scalability—particularly for distributing and delivering mixed media content to
large numbers of locations and users in a predictable and efficient manner. The Internet has
proven to be a great transaction network but falls woefully short when it comes to delivering ser-
vices to even moderately sized communities or subscriber bases. On the other hand, present-day
broadcast infrastructures, satellite, cable, and terrestrial networks possess built-in scalability and
have the means to deliver broadband services to communities of any size, spread over any geog-
raphy, very effectively.
7.6.1a Internet and TV Convergence

Although the Internet and television are viewed on similar video display devices, the difference
between them remains profound [1]. The concept of convergence of these mediums is often mis-
understood or overplayed. The Internet remains an interactive (2-way) medium carrying largely
static content for point-to-point distribution. Meanwhile, television remains a primarily one-way
medium with dynamic content distributed in point-to-multipoint fashion.
Interactive television (ITV) is often considered the true convergence point between these sys-
tems, but in fact, it can be more accurately viewed as adapting certain features of each service to
create a new hybrid medium. ITV takes the rich, dynamic content of television and adds to it the
personalization and responsiveness of the Internet. ITV will likely be seen by consumers as nei-
ther TV nor the Internet, but a medium unique unto itself.
The first and most obvious method of merging these mediums is simply putting Internet con-
tent on a television screen, or conversely, viewing TV content on a computer. Neither of these
processes is as simple as it might seem, but the latter is generally more easily accomplished.
7-151
With the proper hardware, television programs can be fairly successfully viewed on a personal
computer screen. On the other hand, putting a Web page from a computer browser output directly
onto a TV screen is generally not a satisfying experience. This is because Web content is typi-
cally viewed on a computer by a single user positioned at close proximity to the screen, while
television is usually watched at far greater distances and often by groups of viewers. This distinc-
tion is referred to as the “one-foot vs ten-foot experience” or “lean forward vs. lean back” usage.
This implies that when viewing the typical Web page on a TV screen, fonts and graphics are gen-
erally too small to be comfortably viewed. Also, the selection of hyperlinks can be difficult with
an infrared remote control.
ITV systems have dealt with this issue in different ways. Some systems transcode the Web
page content in a specialized server so that it displays more appropriately on a TV screen. Other
systems perform this transcoding in the client receiver, or with a combination of client and
server-based steps. These transcodings typically involve font substitution to a larger typeface,
while retaining as much of the look and feel of the original page as possible. The overall success
of the process varies widely. Some systems do a better job than others, and some Web pages lend
themselves to such transcoding more readily than others.
In addition, because a mouse is generally not used with ITV receivers, Web-page links are
converted to another form of display. Common practice has been to highlight each link on the
screen one at a time, using the remote control navigation keys to sequence among the links. For
example, as a user pushes the “down arrow” key on the ITV remote, the next link down the page
will be highlighted. The user steps through the links on the page using the navigation keys until a
link that the user wishes to follow is highlighted. Then the user presses a “Go,” “OK” or “Enter”
key on the remote, and the display switches to the linked page, once again transcoded for TV dis-
play.
The ultimate solution to this problem involves the design of Web pages customized for TV
display by the content author. While this is rarely done at present, it may become standard prac-
tice as the ITV medium takes hold.
7.6.2 Trends in Internet Content

The growth of data traffic over the Internet is driven by a number of factors, including:
• The increase in the number of consumers accessing Internet on a more frequent basis
• Increase in the number of businesses using the Internet not only to reach customers, but also
to conduct business transactions with partners, vendors, and suppliers
• Proliferation of new wired and wireless Internet-enabled devices
• Increase in the availability and use of broadband Internet access services
• Proliferation of increasingly complex multimedia-rich content
• Emergence of new Internet-based applications designed to simultaneously reach large audi-
ences
Content on the Internet has evolved from simple, static Web pages to complex graphics and
audio/video streams. Delivery of this class of content requires significantly more bandwidth. In
addition, while the Internet has historically been used for one-to-one exchanges between a user
and a host server, new applications have emerged, such as online learning and live webcasts, that
require rich content to reach millions of users simultaneously. These applications require one-to-
many exchanges, and place considerable demands on the Internet infrastructure in terms of band-
width, scalability, and predictability of content delivery.
7.6.2a Limitations of Internet Infrastructure

The current Internet infrastructure is not designed to efficiently handle the delivery of rich, mul-
timedia content to a large number of simultaneous users. There are two fundamental reasons for
this:
• The Internet is a point-to-point network. The Internet is built to handle movement of data
from one point to another. When the same data needs to be sent to several locations, addi-
tional copies of the same data are sent, separately, on each path that connects the sender to the
individual receivers. As the number of simultaneous receiver locations increases to millions,
the Internet infrastructure hits a scalability wall.
• The Internet is a best-efforts network. The path between each user’s personal computer and
the computer that holds the requested content is a series of interconnected computer net-
works. Data moves through these networks in a hop-by-hop manner. Not all networks have the
same capacity to carry data and not all devices that interconnect these networks have the same
capacity to process the data at similar speeds. If any of the intermediate hops experience con-
gestion, as is the case when a large number of users try to access the same content, some of
the data packets can be dropped arbitrarily or delayed. These errors introduce a degree of
unpredictability and unreliability in content delivery. When the content includes audio and/or
video streams, this loss or delay of data packets can severely impact the listening and viewing
quality.
In recent years, large investments have been made to improve the physical infrastructure of
the Internet and to increase the amount of available bandwidth. Advances in networking technol-
ogies such as gigabit and terabit-speed routing, optical networking, broadband access, Web cach-
ing, and content replication have delivered solutions that greatly increase the Internet’s capacity
to carry high bandwidth content from point-to-point. For example:
• Last mile access: Users enjoy rich content by using access technologies like DSL and cable
modems, which have greatly advanced the carrying capacity of the “last mile” segment.
• High speed routers: Advances in ASIC technology and high-performance software enable
routers to handle complex network topologies and to process IP packets at gigabit and terabit
speeds. ISPs can deploy these routers and provision faster links to the core.
• Optical networking: Wave division multiplexing (WDM) and dense wave division multiplex-
ing (DWDM) have increased the available bandwidth in the core of the Internet by several
orders of magnitude.
In spite of increased last mile bandwidth—faster routers moving packets more efficiently and
quickly through the network and a high capacity core—there continues to be a lack of predict-
ability of response time for transactions on the Web. When the same content is being requested
by a large number of users, the congestion point is no longer the network—it is the data center
that serves the content. Various strategies have been employed to speed up the overall response
time of the Internet, by moving the content closer to the edge of the network and therefore requir-
ing fewer hops from the user. These strategies include the following:
• Web caching, which can be used to pre-fetch commonly requested content and store it in loca-
tions that are closer to the user, thus eliminating the need for every transaction to be serviced
by a data center. Web caching can be used to effectively decongest data centers, but it only
works when the content being served is relatively static. In other words, if the content changes
frequently, then the caches have to be replenished frequently as well, and this can result in
stale caches and unpredictable delays.
• Content replication, an effective method for improving response time to transactions, particu-
larly when one can predict which content will be in demand and at what locations. Content
that changes frequently, such as real-time stock updates, news highlights, and live Web events
can be replicated to many servers at locations close to users. Content serving occurs locally,
and therefore delivery is more predictable.
7.6.2b Content Distribution Network

The goal of the content distribution network (CDN) is to help Web site owners deliver their con-
tent to consumers and businesses with greater performance and higher reliability. The underlying
method used by content distribution networks is replication; by making copies of the content
available closer to where the users are, the overall Web access response time is reduced greatly.
There are two predominant modes of operation for CDNs:
• Data center based operation, where the content distribution network provider manages the
service out of private facilities or co-located facilities. Typically, in this mode, content provid-
ers or website owners are the direct beneficiaries of the service.
• ISP point-of-presence based operation, where the content distribution network provider’s ser-
vice is oriented towards improving the performance of the Web cache located at the ISP. In
this mode, the ISP is the direct beneficiary of the service.
The network infrastructures used by these service providers to move the content to the edge
locations vary greatly, from shared to dedicated, and from point-to-point to point-to-multipoint.
Content distribution networks are designed for specific types of content, and specific types of
services that are delivered based on that content. For example, some of these networks are
intended for streaming media services only and therefore transport streaming audio and video.
Others are for speeding up Web response and transporting Web objects. Still others are for deliv-
ering broadcast television and radio services and transporting digital audio and video programs.
The type of services associated with the content they carry dictates the underlying network archi-
tectures employed by these networks.
7.6.2c Broadcast Internet

Unlike the Internet, broadcast networks have been optimized for the transmission of rich content
to large numbers of users in a predictable, reliable, and scalable manner. The advantages they
bring to the infrastructure include:
• Broadcast networks are designed to carry rich, multimedia content. Traditional broadcast net-
works—including television, cable, and satellite networks—have been explicitly designed to
deliver high-quality, synchronized audio and video content to a large population of listeners
and viewers. In addition, many of these networks have migrated from analog to digital trans-
mission systems, thus greatly enhancing their ability to carry new types of digital content,
including Internet content.
• Broadcast networks are inherently scalable. By virtue of their point-to-multipoint transmis-
sion capability, it takes no more resources, bandwidth or other provisions, to send content to a
million locations as it does to one, as long all of the receiving locations are within the trans-
mission footprint of the broadcast network. In contrast, with the traditional Internet, each
location that is targeted to receive the content will add to the overall resources required to
complete the transmission.
• Broadcast networks offer predictable performance. Again, by virtue of the point-to-multipoint
nature of transmission on broadcast networks, there are no variances in the propagation delay
of data throughout the network, regardless of where a receiver is located. This inherent capa-
bility assures a uniform experience to all users within the broadcast network.
By enabling broadcast networks to connect with the Internet, the Broadcast Internet offers a
cost-effective, reliable, and seamless path for delivering multimedia rich content to large num-
bers of users and service providers simultaneously. Pan-continent broadcast networks, such as
digital satellite systems, can be used to distribute content to a very large, and potentially highly
dispersed, set of locations. Other types of broadcast networks with smaller footprints, such as
digital TV networks, can be used for local content distribution. The content is received and
cached locally at these locations, which will—in turn—serve the content to users that connect
either directly to that location or through a service provider that uses the location for content
serving. The content is served to the users using the traditional mechanisms employed on the
Internet. For example, instead of the ISPs replenishing their local caches on an on-demand basis,
they can adopt the strategy of pre-filling their caches based on analysis of the type of content
most frequently requested by the communities they serve. Content that is dynamic in nature,
such as audio/video streaming and stock tickers, will be updated on the local caches at the appro-
priate frequency, keeping the cached content current and providing better overall quality of ser-
vice to the users. Figure 7.6.1 illustrates the application of this principle to a cable-based delivery
system. Figure 7.6.2 expands the concept to a variety of distribution media.
The major design considerations for building Broadcast Internet-ready broadcast facilities
are:
• Multi-vendor interoperability. Any equipment that is added to the head-end to incorporate
Internet data must work seamlessly with all existing equipment and must fit within the opera-
tional framework used by the facility. This requires any new equipment to support standard
interfaces and protocols, and follow the guidelines set by standards organizations that cover
both the digital video and IP domains.
• Bandwidth optimization. Managing the available bandwidth in optimal fashion is a very
important design consideration. There is considerable variability in the use of bandwidth by
compression and multiplexing equipment, caused either by the inherent inefficiencies of the
equipment, the nature of the content, or the service mix. The head-end architecture must pro-
vide the ability to capitalize on this variability to maximize the use of the bandwidth in an
opportunistic manner. Two strategies are commonly applied for allocating bandwidth to data
services in an environment that includes legacy video services as well: 1) pre-allocated band-
Figure 7.6.1 A high-bandwidth data distribution system incorporating cable TV distribution to end-
users.
width, where the data services are treated just like another “channel” or “bundle of channels”
and the bandwidth is reserved for the exclusive use of these services; 2) opportunistic band-
width, where the data services are treated as “trickle services”, where there are no specific
time of delivery requirements for the service. In this case, bandwidth can be optimized by not
pre-allocating the bandwidth for the data services and taking advantage of the variability in
the bandwidth use by the legacy video programs to “fill the blanks” with data.
• Flexibility. The nature and type of Internet services that will be required to be supported in
any head-end is likely to change rapidly, as is evidenced by the rapid rate of technological and
application advancement in the Internet in general. The head-end architecture must easily
adapt to new applications requirements without having to undergo major design changes.
Currently, applications that incorporate data can be classified as following:
• Pure data services. These services use Internet or private data as the sole content. There are
no legacy, television-oriented digital audio/video elements in the content. This type of content
is typically targeted towards personal computers and integrated set-top boxes that have an
operating environment similar to personal computers and are enabled for interactive services
on the television. The content preparation, packaging, scheduling, and delivery systems used
for these services can be independent of those used for the legacy video services.
• Loosely co-related data services. These services use Internet or private data to enhance or
extend a legacy, television-oriented program, or to provide a parallel co-related content
Figure 7.6.2 High-speed data delivery using a variety of last-mile systems.
stream that is independent of the video stream, but scheduled for play-out simultaneously.
The parallel content stream can be viewed in real time or stored for later viewing. The content
preparation, packaging, scheduling, and delivery systems used for these services must have
the ability to interface and interact with those used for the legacy video services.
• Enhanced TV. These services use Internet or private data to enhance a legacy, television-ori-
ented program in a highly synchronized manner. The synchronization can be either presenta-
tion time-based or frame content-based. The content preparation, packaging, scheduling, and
delivery systems for these services must be fully integrated into those used for legacy video
services.
7.6.2d DTV Content Distribution Models

There are three basic models that can be employed for content distribution over digital television
systems:
• End-to-end DTV data broadcasting. In this model, the service provider creates services by
aggregating content from both the TV domain and the Internet domain, and distributes these
services over a satellite or land-based network to television stations in the targeted regions.
The television stations—either in real time or in a store-and-forward manner—then deliver
the services. The customers for such services would be consumers.
Table 7.6.1 Software-Only DTV Decoding Workload
DTV Hardware Over-the-air reception

VSB demodulation
System CPU ATSC demultiplex
Full MPEG-2 decoding
AC3 decoding
Applications overhead
Graphics Hardware Screen display
• Data distribution to Internet service providers. This model is suitable for local distribution
of content, particularly multicast services and streaming media, to Internet service providers
that then deliver that content to consumers over traditional land-based last-mile connections.
This model can be very effective for simulcast services—programs that are viewable on the
TV as well as on a PC, with enhancements specific to the PC.
• DTV as last mile. In this model, the DTV spectrum is used as a last mile in a content distri-
bution network. Content providers lease or acquire on demand last mile bandwidth for deliv-
ering Internet based services.
7.6.3 PC-Based DTV Receivers

All of the features common in cable or satellite set top box (STB) implementations are possible
in a DTV-PC solution. For example, the ability to pause a live broadcast while the viewer is oth-
erwise occupied is implemented with a time shift function. In time shift recording, the system
needs a fast-access large-storage medium to save the program for later display. Set top boxes may
include a hard drive in the system for this purpose. This function is easily implemented in the
DTVPC system because of the existing presence of the system hard drive.
DTVPC cards represent an ideal element in the interactive TV market. This is due to the fact
that most PCs already have the most important function required to implement interactive TV—
a mechanism for the viewer to provide feedback. The Internet connection found on most of
today’s PCs is the medium by which viewers will communicate back to the service provider with
information ranging from response to game show questions to political polling data.
7.6.3a Practical Implementation Considerations

There are three basic ways to bring the DTV experience to the PC platform:
• Software DTV decoding
• Hardware assisted software DTV decoding
• Hardware DTV decoding
Each of these approaches will be examined in the following sections.
Table 7.6.2 Hardware Assisted DTV Decoding Workload

VSB demodulation
System CPU ATSC demultiplex
Partial MPEG-2 decoding
AC3 decoding
Graphics Hardware Screen display
IDCT
Motion compensation
Software DTV Decoding

The software decoding approach forces all of the processing work on the system CPU. (See Table
7.6.1.) There are three components in the software-based solution: a receiver card, a high-speed
CPU, and the graphics chip. The receiver card, consisting of a DTV tuner and a vestigial side-
band (VSB) demodulator, receives the DTV broadcast signal and passes it to the CPU for further
processing. The CPU demultiplexes the transport signal and decodes the DTV data using soft-
ware algorithms.
Hardware Assisted Software DTV Decoding

This approach relies on a combination of CPU power and motion-compensation hardware found
in graphics chips. There are three components in the hardware assisted software-based solution:
a receiver card, a high-speed CPU, and the graphics chip. The receiver card, consisting of a DTV
tuner and 8-VSB demodulator, receives the DTV broadcast signal and passes it to the CPU for
further processing. The CPU demultiplexes the transport signal and decodes the DTV data using
software algorithms. The graphics chip assists the CPU with its on-chip motion compensation
and inverse discrete cosine transformation (IDCT) logic. This approach also requires a signifi-
cant amount of CPU bandwidth. (See Table 7.6.2.)
Hardware DTV Decoding

In this implementation, a dedicated hardware decoder is used for enabling DTV on a PC. A typi-
cal system is illustrated in Figure 7.6.3. The DTV decoder does not rely on the CPU or the graph-
ics chip for decoding the DTV stream. There is very little drain on CPU bandwidth as the add-in
card decodes the complete MP@HL MPEG stream. All 18 ATSC DTV formats can be supported
at full frame rates. The workload division for this approach is given in Table 7.6.3.
A DTVPC solution must meet the following criteria for it to be widely accepted among con-
sumers.
• Single-board solution for true HDTV and legacy analog TV reception
• Simple and easy installation
• Consumer price points for mass-market acceptance
• Support of PC screen formats
Figure 7.6.3 Single-board hardware-based DTV decoder.
Table 7.6.3 Hardware DTV Decoding Workload

VSB demodulation
ATSC demultiplex
Full MPEG-2 decoding
Screen display
System CPU AC3 decoding
The DTV card must also support multi-sync monitors and common PC display resolutions such
as 1024 × 768, 800 × 600, and 640 × 480. If the board can support DVD playback in addition to
ATSC and NTSC, the value proposition for the consumer is significantly enhanced.
7.6.4 Video Interconnection and Networking

One of the promises of DTV is a marriage of video and computer technologies in a wide range of
consumer devices [1]. To accomplish this goal, several fundamental technologies and standards
Table 7.6.4 Common Consumer Video Interfaces
Interface Functional Details Notes

Composite Uses a single 75 Ω coaxial cable, usually termi- Composite video inputs and/or out-
nated with RCA connectors. Composite video puts are present on almost all con-
signals can also be modulated onto an RF carrier temporary video equipment. New
along with audio and transmitted to other equip- video devices are moving away from
ment or systems, typically using 75 Ω RF cable this interface in favor of digital links,
terminated with F-connectors. which provide better performance
and include provisions for rights man-
agement (copy protection).
S-Video Uses a four-pin DIN connector termination. Also known as Y/C video, which is
(Although it may appear to be a single cable, more descriptive of the signal format.
internally the S-Video cable has two 75 Ω coaxial This interface is an improvement over
or twisted pair cables.) Conveys separate Y and composite video, however, it does not
C signals from one device to another. support advanced image formats or
digital rights management functions.
Widely used on consumer equip-
ment.
Analog component video Requires three 75 Ω coaxial cables, which carry Referred to as “color difference” video,
the Y, R-Y, and B-Y signals separately. Uses RCA Y, R-Y, B-Y, or Y, Pb, Pr. The term Y,
connector terminations. Cables may be color Cb, Cr denotes the same three video
coded and bundled together in a single sheath. signals but in digital form. In profes-
The color code is typically as follows: green for Y, sional video equipment the cables
red for R-Y, blue for B-Y. are usually terminated with imped-
ance-matched 75 Ω BNC connectors.
IEEE 1394 A digital interface supporting several tiers of per- A popular audio/video digital interface
formance. IEEE 1394 is a peer-to-peer interface that enjoys a large installed base of
that allows not only dubbing from one device to devices and systems.
another, but also allows multiple devices to share
a given source. Supports rights management
functions. Features an extensible architecture
that permits new functions to be implemented.
DVI A digital interface supporting several tiers of per- A preferred audio/video digital inter-
formance. Dual Link DVI supports 2 × 165 MHz face for new equipment.
channels of bandwidth. Incorporates High-Defini-
tion Copy Protection (HDCP); supports the
“Broadcast Flag.”
must be in place, and certain agreements must be reached among the many potential informa-
tion-providers.
A related challenge is the interface format itself, including the physical connecting device.
Table 7.6.4 summarizes the more common interfaces.
7.6.4a DVI
The Digital Visual Interface (DVI) grew out of a consortium of companies known as the Digital
Display Working Group, an open industry group lead by Intel, Compaq, Fujitsu, Hewlett Pack-
ard, IBM, NEC and Silicon Image [2]. The objective of the Group was to address industry
requirements for a digital connectivity specification for high-performance PCs and digital dis-
plays.
DVI has evolved to become a preferred interface between devices such as DVD players and
cable/satellite/terrestrial set-top DTV receivers. DVI is available in two basic flavors:
• DVI-I, which is a combination analog/digital interface for a certain degree of backward com-
patibility.
• DVI-D, which is 100 percent digital.
DVI accommodates two different bandwidths:
• Dual Link DVI, which supports 2 × 165 MHz channels of bandwidth.
• Single Link DVI, which supports a maximum bandwidth of 165 MHz on a single channel.
The consumer world adds another layer of complexity by incorporating High-Definition
Copy Protection (HDCP), which encrypts each pixel as it moves from the player or set-top box to
the display, using an asymmetric system that calls for periodic reauthorization from the source.
7.6.4b IEEE 1394

The interconnection of home information and entertainment devices is a key element in
advanced consumer products. IEEE 1394 is an international standard, low-cost digital interface
designed to integrate entertainment, communication, and computing electronics into consumer
multimedia products [3]. Originated by Apple Computer as a desktop LAN and developed by the
IEEE 1394 working group, IEEE 1394 has the following attributes:
• High data rate capabilities—the basic hardware/software standard can transport data at 100,
200, or 400 Mbits/s.
• Physically small—the serial cable can replace larger and more expensive interfaces.
• Ease of use—there is no need for terminators, device IDs, or elaborate setup procedures.
• Hot plug capable—users can add or remove 1394 devices with the bus active.
• Scaleable architecture—users can mix 100, 200, and 400 Mbits/s devices on a single bus.
• Flexible topology—support is provided for daisy chaining and branching, facilitating true
peer-to-peer communication.
• Nonproprietary—there are no significant licensing issues for implementation in products or
systems.
Serial bus management provides overall configuration control of the 1394 bus in the form of
optimizing arbitration timing, guarantee of adequate electrical power for all devices on the bus,
assignment of which compliant device is the cycle master, assignment of the isochronous chan-
nel ID, and notification of errors. Bus management is built upon the IEEE 1212 standard register
architecture.
There are two fundamental types of IEEE 1394 data transfer:
• Asynchronous: the traditional computer memory-mapped, load-and-store interface. Data
requests are sent to a specific address and an acknowledgment is returned.
• Isochronous: data channels provide guaranteed data transport at a predetermined rate. This is
especially important for time-critical multimedia data where just-in-time delivery eliminates
the need for costly buffering.
Much like LANs and WANs, IEEE 1394 is defined by the high level application interfaces
that use it, not a single physical implementation. Therefore, as new silicon technologies allow
higher speeds, longer distances, and alternate media, IEEE 1394 can scale to enable new applica-
tions.
Perhaps most important for use as a digital interface for consumer electronics is that IEEE
1394 is a peer-to-peer interface. This allows not only dubbing from one camcorder to another
without a computer, for example, but also allows multiple computers to share a given camcorder
without any special support in the camcorders or computers. All of these features of IEEE 1394
are key reasons why it is a preferred audio/video digital interface.
7.6.4c Digital Home Network

In the near-term future, data for audio, video, telephony, printing, and control functions are all
likely to be transported through the home over a digital network [3]. This network will allow the
connection of devices such as computers, digital TVs, digital VCRs, digital telephones, printers,
stereo systems, and remotely controlled appliances. To enable this scope of interoperability of
home network devices, standards for physical layers, network information, and control protocols
need to be generally agreed upon and accepted. While it would be preferable to have a single
stack of technology layers, no one selection is likely to satisfy all cost, bandwidth, and mobility
requirements for in-home devices.
From a broadcasting perspective, the ability to provide unrestricted entertainment services to
consumer devices in the home is a key point of interest. Also, ancillary data services directed to
various devices offer significant marketplace promise.
More generally, the principal physical layer interconnections at the DTV set are expected to
be RF (NTSC, VSB, QAM), baseband component (RGB, Y Pr Pb), and digital (1394, DVI).
Composite video and S-video will be around a long time as well. Nearly all of these are one-way
paths.
Other appliances (or sensors) in the home may use RF on the power lines, dedicated coax, and
dedicated twisted pair for control functions. From this plethora of physical layer choices, the sig-
naling defined in the set of 1394 standards, for example, is well suited to high-speed local con-
nections, and as the backbone for passing information around the home. This physical layer
comes in several versions.
On top of the physical layer, network and control layers are needed. This is the part of the
communications stack that was defined for the DTV interface (i.e., EIA-775), and for long dis-
tance 1394 by the Video Electronics Standards Association (VESA) Home Network group
(VHN). Their objective was to define a network layer approach to allow seamless operation
across different physical layers using Internet protocol (IP). HAVI (Home Audio/Video Interop-
erability) is a different network layer solution for home networking that has the same unifying
objective as VHN. Optimized for the 1394 physical layer, HAVI is focused on audio/video appli-
cations and requirements.
On top of the network layer, widely differing approaches exist to the application interfaces,
operating system, rendering engines, browsers, and the degree of linkage with the Internet.
HAVi Specifications
The Home Audio Video Interoperability Group (HAVi) was formed in 1999 by eight major con-
sumer electronics companies—Grundig, Hitachi, Matsushita, Philips, Sharp, Sony, Thomson,
and Toshiba. The goal of the group was to create common specifications for interconnecting
(networking) and providing interoperability between digital home entertainment products.
HAVI’s work focused on a multi-vendor architecture to manage home entertainment products
and to provide for seamless interoperability. Devices connected in a HAVi network can share
resources and functionality across the network. The standard is based on an underlying IEEE
1394 digital interface. In addition to the basic networking and control capabilities, HAVI pro-
vides a Java-based user interface that is optimized for television displays and further defines the
behavior of applications in the network.
7.6.4d PVRs
An important development in the television environment, and one that has strong resonance for
ITV applications, is the introduction of the personal video recorder (PVR). Because these
devices allow personalization of programming, they are attractive to consumers and offer a fun-
damental form of interactivity with the content environment for consumers.
These devices allow a substantial amount of television programming to be stored on a hard
disk and played back in non-linear fashion. Because a non-linear storage device (a hard drive) is
being used, these devices can begin to replay a program while it is still being recorded. This
allows a powerful feature of the PVR called live pause, by which the device is used as a buffer
for live programming. This permits users to pause a program that they are watching as it is cur-
rently being broadcast, and return to it a few minutes later without interruption.
PVRs also offer a range of program selection processes from simple searching of electronic
program guides to intelligent searching for keywords in program titles, subjects, content, talent
names, and so on. Advnced PVRs incorporate a browsing ability to seek and download stored
content from on-demand libraries, as well as access broadcast program schedules.
Some PVRs store ITV elements of a program so they can be selectively accessed upon later
playback. A PVR that can access both broadcast channels (via terrestrial, satellite, or cable) plus
the Internet via broadband connectivity will serve as a powerful media collection tool. When
interfaced to a home network system, such a storage device can act as the home media server.
PVRs can be standalone devices, part of a PC, a home server, or even a remote storage system
in which the actual physical storage is on a server at some service provider’s location, with
unique personal access by the user. In this respect, the PVR is essentially a user interface that
accesses storage locally or on various networks, with the actual storage location of a particular
program being transparent to the user. The ultimate vision of this process is the creation of vir-
tual channels, by which the users’s PVR builds the program stream that the user views from vari-
ous stored sources.
7.6.5 DTV Product Classification

In an effort to minimize consumer confusion regarding digital television, the Consumer Elec-
tronics Association (CEA) developed definitions and labels for DTV products. The definitions
were to be incorporated into manufacturers’ television marketing materials as DTV receivers

continued into the retail channels.
The CEA digital television definitions are as follows [4]:
• High-definition television (HDTV): A complete product/system with the following mini-
mum performance attributes: 1) receives ATSC terrestrial digital transmissions and decodes
all ATSC Table 3 video formats; 2) has active vertical scanning lines of 720P, 1080I, or
higher; 3) is capable of displaying a 16:9 image; and 4) receives and reproduces, and/or out-
puts Dolby Digital audio.
• High-definition television monitor: A monitor or display with the following minimum per-
formance attributes: 1) has active vertical scanning lines of 720P, 1080I, or higher; and 2) is
capable of displaying a 16:9 image. In specifications found on product literature and in
owner’s manuals, manufacturers were required to disclose the number of vertical scanning
lines in the 16:9 viewable area, which must be 720P, 1080I, or higher to meet the definition of
HDTV.
• High-definition television tuner: An RF receiver with the following minimum performance
attributes: 1) receives ATSC terrestrial digital transmissions and decodes all ATSC Table 3
video formats; 2) outputs the ATSC Table 3 720P and 1080P/I formats in the form of HD with
minimum active vertical scanning lines of 720P, 1080I, or higher; and 3) receives and repro-
duces, and/or outputs Dolby Digital audio. Additionally, this tuner may output HD formats
converted to other formats. The lower resolution ATSC Table 3 formats can be output at lower
resolution levels. Alternatively, the output can be a digital bitstream with the full resolution of
the broadcast signal.
• Enhanced definition television (EDTV): A complete product/system with the following
minimum performance attributes: 1) receives ATSC terrestrial digital transmissions and
decodes all ATSC Table 3 video formats; 2) has active vertical scanning lines of 480P or
higher; and 3) receives and reproduces, and/or outputs Dolby Digital audio. The aspect ratio
is not specified.
• Enhanced definition television monitor: A monitor or display with the following minimum
performance attributes: 1) has active vertical scanning lines of 480P or higher. No aspect ratio
is specified.
• Enhanced definition television tuner: An RF receiver with the following minimum perfor-
mance attributes: 1) receives ATSC terrestrial digital transmissions and decodes all ATSC
Table 3 video formats; 2) outputs the ATSC Table 3 720P and 1080I/P and 480P formats with
minimum active vertical scanning lines of 480P; and 3) receives and reproduces, and/or out-
puts Dolby Digital audio. Alternatively, the output can be a digital bitstream output capable of
transporting 480P, except the ATSC Table 3 480I format, which can be output at 480I.
• Standard definition television (SDTV): A complete product/system with the following per-
formance attributes: 1) receives ATSC terrestrial digital transmissions and decodes all ATSC
Table 3 video formats, and produces a useable picture; 2) has active vertical scanning lines
less than that of EDTV; and 3) receives and reproduces usable audio. No aspect ratio is speci-
fied.
• Standard definition television tuner: An RF receiver with the following minimum perfor-
mance attributes: 1) receives ATSC terrestrial digital transmissions and decodes all ATSC
Table 3 video formats; 2) outputs all ATSC table 3 formats in the form of NTSC output; and
3) receives and reproduces, and/or outputs Dolby Digital audio.
These industry standard definitions were intended to eliminate the confusion over product fea-
tures and capabilities of television sets and monitors intended for DTV applications. The agree-
ment promised to spur the sale of DTV-compliant sets by injecting a certain amount of logic into
the marketing efforts of TV set manufacturers.
7.6.6 Videoconferencing
With desktop computers nearly as ubiquitous in business these days as telephones, the time has
arrived for the next big push in telecommunications—interactive desktop videoconferencing.
Interaction via video has been used successfully for many years to permit groups of persons to
communicate from widely-distant locations. Such efforts have usually required some degree of
advance planning and specialized equipment ranging from custom-built fiber or coax services to
satellite links. These types of applications will certainly continue to grow, as the need to commu-
nicate on matters of business expands. The real explosion in videoconferencing, however, will
come when three criteria are met:
• Little—if any—advance planning is needed
• No special communications links need be installed to participate in a conference
• Participants can do it from their offices
The real promise of videoconferencing is to make it as convenient and accessible as a telephone
call.
7.6.6a Infrastructure Issues

The challenge of videoconferencing is not the video itself. Required equipment includes one or
more cameras, microphones, displays, and the equipment to control them. All of this is important
in any type of video conference, and strides are being made in this area all the time. Automatic
control of camera movements and intelligent audio switching are just two of the technologies that
have been developed to a high level of sophistication. The primary remaining challenge is getting
the signal to and from the video equipment on each end. To accomplish this, we need to look
toward computer networks and common carriers.
7.6.6b Full Motion Video

The ideal for any videoconferencing system is the highest possible video quality, in terms of both
frame rate and image quality. Despite the obvious focus on video, audio quality is usually the
most important element of a videoconferencing system because it is the audio that conveys most
of the information. The whole idea of a videoconference is to facilitate communication among
participants. Research has shown that if the audio is clear and audible, the next area on which
users focus is the video quality.
Image quality for a videoconferencing system is largely determined by the quality of the
codec being used and the bandwidth allocated to the session. Moreover, much of the codec qual-
ity is determined not simply by the hardware being used and how fast it is, but by the sophistica-
tion of the algorithms that run on the hardware.
7.6.6c Desktop Systems—Hardware and Software

Implementation of a videoconferencing system on the desktop is a party to which both hardware
and software must be invited. There are obvious hardware peripherals that must be present for
the system to work, such as a camera, microphone, speaker, and network connection (or modem).
The key element of the system is the codec, which can be implemented either in hardware or
software. Not surprisingly, the hardware approach is typically faster but more expensive. Soft-
ware-only codecs, however, have come a long way, and the continuing speed improvements in
personal computer CPUs capitalize upon these improvements.
The attractiveness of a software codec is that it can run on any machine, and there is no add-
on hardware to purchase. Software codecs also can be updated easily by changing the software
driver. Naturally, the performance of the software codec is a function of the type and speed of the
processor.
The software aspect of videoconferencing also includes considerations for general office
application programs. Interface capabilities with common desktop applications and “suites”
offer a host of user benefits. For example, a presentation can be viewed during the videoconfer-
ence and changes made in real-time.
Control and coordination of a videoconference has been made much easier thanks to smart
peripheral and supervisory systems. Items such as touch-screens, auto-tracking cameras, and
participant responses systems permit users to focus on communicating ideas rather than on pro-
ducing the videoconference. For example, students may electronically “raise their hand” by
pressing a button on a microphone. This notifies the instructor that someone has a question and
directs the camera to zoom-in on the student. Through such face-to-face interaction, students and
instructors can build relationships that make distance learning more responsive and effective.
Small Group-vs.-Large Group Systems

The nature of a desktop videoconferencing system is that it is usually restricted to one or perhaps
two individuals at each terminal. For larger groups, roll-around portable systems and theater-type
installations are options. Until recently, the theater-type or dedicated videoconferencing center
defined what a video conference was all about: bringing together large groups of people in dis-
tant locations. While these applications still exist in large numbers, the primary growth of video-
conferencing involves smaller groups.
Portable, roll-around systems offer the benefit of easy setup for small conference rooms and
group workspaces. Such packaged systems fill the growing need for group collaboration among
distant points.
7.6.6d System Implementations

Although any given videoconferencing session must be designed to meet the particular needs of
the subject matter, there are three basic topologies that can be applied to the application:
• Point-to-point, the simplest arrangement where two individuals are interconnected in real-
time (Figure 7.6.4a).
(a )
(b )
(c )
Figure 7.6.4 Primary videoconferencing modes: (a) point-to-point, (b) broadcast, (c) multicast.
• Broadcast, with a single origination point and multiple receiving points (Figure 7.6.4b)
• Multicast, where all (or at least some) of the participating individuals can communicate with
each other and/or with the group as a whole (Figure 7.6.4c).
7.6.7 References
1. Pizzi, Skip: “Internet and TV Convergence,” in Interactive TV Survival Guide, Jerry C.
Whitaker ed., McGraw-Hill, New York, N.Y., 2001.
2. Putman, Peter: “Square Pegs and Round Holes,” http://www.hdtvexpert.com.
3. Hoffman, Gary A.: “IEEE 1394: The A/V Digital Interface of Choice,” 1394 Technology
Association Technical Brief, 1394 Technology Association, Santa Clara, Calif., 1999.
4. NAB TV TechCheck: “CEA Establishes Definitions for Digital Television Products,”
National Association of Broadcasters, Washington, D.C., September 1, 2000.
7.6.8 Bibliography
NAB: “Consumer Electronic Consortium Publishes Updated Specifications for Home Audio
Video Interoperability,” TV TechCheck, National Association of Broadcasters, Washington,
D.C., May 21, 2001.
g g
Section
RF System Maintenance
8
Radio frequency (RF) equipment is unfamiliar to many persons entering the electronics industry.
Colleges do not routinely teach high-power RF principles, favoring instead digital technology.
Unlike other types of products, however, RF equipment often must receive preventive mainte-
nance to achieve its expected reliability. Maintaining RF gear is a predictable, necessary expense
that facilities must include in their operating budgets. Tubes (if used in the system) will have to
be replaced from time-to-time no matter what the engineer does; components fail every now and
then; and time must be allocated for cleaning and adjustments. By planning for these expenses
each month, unnecessary downtime can be avoided.
Although the reason generally given for minimum RF maintenance is a lack of time and/or
money, the cost of such a policy can be deceptively high. Problems that could be solved for a few
dollars may, if left unattended, result in considerable damage to the system and a large repair bill.
A standby system often can be a lifesaver, however its usefulness sometimes is overrated. The
best standby RF system is a main system in good working order.
In This Section:
Chapter 8.1: RF System Reliability Considerations 8-7

Introduction 8-7
The Maintenance Log 8-7
Case Study #1 8-8
Case Study #2 8-9
RF System Maintenance 8-9
Preventive Maintenance Routine 8-10
Resistors and Capacitors 8-11
Power Supply Components 8-11
Coils and RF Transformers 8-12
Relay Mechanisms 8-12
Connection Points 8-12
Klystron Devices 8-12
Power Grid Tubes 8-13
Tube Dissipation 8-13
Air-Handling System 8-14
8-1
y
8-2 Section Eight
Ambient Temperature 8-15

Thermal Cycling 8-17
Tube Changing Procedure 8-17
Extending Vacuum Tube Life 8-17
Power Tube Conditioning 8-17
Filament Voltage 8-18
Filament Voltage Management 8-19
PA Stage Tuning 8-19
When to Tune 8-20
Bibliography 8-21
Chapter 8.2: Preventing RF System Failures 8-23

Introduction 8-23
Common Mode Failures 8-23
Modifications and Updates 8-24
Spare Parts 8-24
Transmission Line/Antenna Problems 8-25
Effects of Modulation 8-25
Maintenance Considerations 8-26
UHF Transmission Systems 8-27
High Voltage Power Supply Problems 8-27
Power Supply Maintenance 8-27
Power Supply Metering 8-29
Overload Sensor 8-29
Transient Disturbances 8-30
Single Phasing 8-31
Temperature Control 8-32
Cooling System Maintenance 8-32
Air Cooling System Design 8-33
Layout Considerations 8-34
Case 1 8-34
Case 2 8-35
Case 3 8-35
Air Filters 8-36
Klystron Water Cooling Systems 8-36
Bibliography 8-37
Chapter 8.3: Troubleshooting RF Equipment 8-39

Introduction 8-39
Troubleshooting Procedure 8-39
The “Quick Fix” 8-40
Factory Service Assistance 8-40
Plate Overload Fault 8-41
Process of Elimination 8-43
RF System Faults 8-44
Component Substitution 8-46
y
Inside the PA Cavity 8-46

VSWR Overload 8-49
Power Control Faults 8-50
Thyristor Control System 8-51
Interlock Failures 8-52
Step-Start Faults 8-53
Protection Circuits 8-54
Chapter 8.4: Testing Coaxial Transmission Line 8-57

Introduction 8-57
Measuring VSWR 8-58
Calibration 8-59
Antenna Measurements 8-59
Testing Coaxial Lines 8-60
Crimps and Mismatches 8-61
Multiple Mismatches 8-61
Bibliography 8-62
Chapter 8.5: The Smith Chart 8-63

Introduction 8-63
Normalized Impedance 8-63
Propagation Constant 8-64
Applying the Smith Chart 8-65
Application Example 8-69
References 8-69
Bibliography 8-69
Chapter 8.6: Standby Power Systems 8-71

Introduction 8-71
Blackout Effects 8-71
Standby Power Options 8-72
Dual Feeder System 8-73
Peak Power Shaving 8-74
Advanced System Protection 8-74
Choosing a Generator 8-75
Generator Types 8-79
UPS Systems 8-80
Standby Power-System Noise 8-80
Batteries 8-81
Terms 8-82
Sealed Lead-Acid Battery 8-83
References 8-84
Bibliography 8-84

Adam, S. F.: Microwave Theory and Applications, Prentice-Hall, New York, N.Y., 1969.
y
8-4 Section Eight
Angevine, Eric: “Controlling Generator and UPS Noise,” Broadcast Engineering, PRIMEDIA
Intertec, Overland Park, Kan., March 1989.
Baietto, Ron: “How to Calculate the Proper Size of UPS Devices,” Microservice Management,
PRIMEDIA Intertec, Overland Park, Kan., March 1989.
Bowick, C.: RF Circuit Design, Howard W. Sams and Co., Indianapolis, IN, 1982.
Bryant, G. H.: Principles of Microwave Measurements, IEE Electrical Measurement Series,
Peter Peregrinus Ltd., London, 1988.
Buchmann, Isidor: “Batteries,” in The Electronics Handbook, Jerry C. Whitaker (ed.), pg. 1058,
CRC Press, Boca Raton, Fla., 1996.
“Cable Testing with Time Domain Reflectometry,” Application Note 67, Hewlett Packard, Palo
Alto, Calif., 1988.
Colin, R. E.: Foundations for Microwave Engineering, McGraw-Hill, New York, N.Y., 1966.
DeDad, John A.: “Auxiliary Power,” in Practical Guide to Power Distribution for Information
Technology Equipment, PRIMEDIA Intertec, Overland Park, Kan., pp. 31–39, 1997.
Federal Information Processing Standards Publication No. 94, Guideline on Electrical Power for
ington, D.C., 1983.
Gray, T. S.: Applied Electronics, Massachusetts Institute of Technology, 1954.
High Power Transmitting Tubes for Broadcasting and Research, Phillips Technical Publication,
Eindhoven, The Netherlands, 1988.
Highnote, Ronnie L.: The IFM Handbook of Practical Energy Management, Institute for Man-
agement, Old Saybrook, Conn., 1979.
“Improving Time Domain Network Analysis Measurements,” Application Note 62, Hewlett
Packard, Palo Alto, Calif., 1988.
Kaufhold, Gerry: “The Smith Chart, Parts 1–4,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., November 1989–-March 1990.
Kennedy, George: Electronic Communication Systems, 3rd ed., McGraw-Hill, New York, N.Y.,
1985.
Kolbert, Don: “Testing Coaxial Lines,” Broadcast Engineering, Intertec Publishing, Overland
Park, Kan., November 1991.
1988.
Montgomery, C. G., R. H. Dicke, and E. M. Purcell: Principles of Microwave Circuits,
Power Grid Tubes for Radio Broadcasting, Thomson-CSF publication #DTE-115, Thomson-
CSF, Dover, N.J., 1986.
Silence, Neal C.: “The Smith Chart and its Usage in RF Design,” RF Design, Intertec Publish-
ing, Overland Park, Kan., pp. 85–88, April 1992.
y
Smith, Morgan: “Planning for Standby AC Power,” Broadcast Engineering, PRIMEDIA Inter-
Smith, P. H.: Electronic Applications of the Smith Chart, McGraw-Hill, New York, N.Y., 1969.
Strickland, James A.: “Time Domain Reflectometry Measurements, Measurement Concepts
Series, Tektronix, Beaverton, Ore., 1970.
Stuart, Bud: “Maintaining an Antenna Ground System,” Broadcast Engineering, PRIMEDIA
Intertec, Overland Park, Kan., October 1986.
Svet, Frank A.: “Factors Affecting On-Air Reliability of Solid State Transmitters,” Proceedings
of the SBE Broadcast Engineering Conference, Society of Broadcast Engineers, Indianapo-
lis, IN, October 1989.
“TDR Fundamentals,” Application Note 62, Hewlett Packard, Palo Alto, CA., 1988.
The Care and Feeding of Power Grid Tubes, Varian EIMAC, San Carlos, Calif., 1984.
Whitaker, Jerry C.: Maintaining Electronic Systems, CRC Press, Boca Raton, Fla., 1991.
Whitaker, Jerry C.: RF Systems Handbook, CRC Press, Boca Raton, Fla., 2002.
y
g g
Chapter
8.1
RF System Reliability Considerations
8.1.1 Introduction
Most RF system failures can be prevented through regular cleaning and inspection, and close
observation. The history of the unit also is important in a thorough maintenance program so that
trends can be identified and analyzed.
8.1.1a The Maintenance Log

The control system front panel can tell a great deal about what is going on inside an RF genera-
tor. Develop a maintenance log and record all front-panel meter readings, as well as the positions
of critical tuning controls, on a regular basis, as illustrated in Figure 8.1.1. This information pro-
vides a history of the system and can be a valuable tool in noting problems at an early stage. The
most obvious application of such logging is to spot failing power tubes, but other changes occur-
ring in components can be identified.
Creating a history of the line and tank pressure for a pressurized transmission line can iden-
tify developing line or antenna problems. After the regulator is set for the desired line pressure,
record the tank and line readings each week and chart the data. If possible, make the observations
at the same time of day each week. Ambient temperature can have a significant effect on line
pressure; note any temperature extremes in the transmission line log when the pressure is
recorded. The transmission line pressure usually will change slightly between carrier-on and car-
rier-off conditions (depending on the power level). The presence of RF can heat the inner con-
ductor of the line, causing the pressure to increase. After a few months of charting the gradual
loss of tank pressure, a pattern should become obvious. Investigate any deviation from the nor-
mal amount of tank pressure loss over a given period.
Whenever a problem occurs with the RF system, make a complete entry describing the failure
in the maintenance log. Include a description of all maintenance activities required to return the
system to operational condition. Make all entries complete and clear. Include the following data:
• Description of the nature of the malfunction, including all observable symptoms and perfor-
mance characteristics.
• Description of the actions taken to return the system to a serviceable condition.
8-7
y y
8-8 RF System Maintenance
Figure 8.1.1 Example of a transmitter operating log that should be filled-out regularly by mainte-
nance personnel.
• Complete list of the components replaced or repaired, including the device schematic number
and part number.
• Total system downtime as a result of the failure.
• Name of the engineer who made the repairs.
The importance of regular, accurate logging can best be emphasized through the following exam-
ples
Case Study #1
Improper neutralization is detected on an AM broadcast transmitter IPA (intermediate power
amplifier), shown in Figure 8.1.2. The neutralization adjustment is made by moving taps on a
coil, and none have been changed. The history of the transmitter (as recorded in the maintenance
record) reveals, however, that the PA grid tuning adjustment has, over the past two years, been
moving slowly into the higher readings. An examination of the schematic diagram leads to the
conclusion that C-601 is the problem.
The tuning change of the stage was so gradual that it was not thought significant until an
examination of the transmitter history revealed that continual retuning in one direction only was
necessary to achieve maximum PA grid drive. Without a record of the history of the unit, time
could have been wasted in substituting expensive capacitors in the circuit, one at a time. Worse
yet, the engineer might have changed the tap on coil L-601 to achieve neutralization, further hid-
ing the real cause of the problem.
y y
Figure 8.1.2 AM transmitter IPA/PA stage exhibiting neutralization problem. A history of IPA retun-
ing (through adjustment of L-601) helped determine that loss of neutralization was the result of C-
601 changing in value.
Case Study #2
A UHF broadcast transmitter is found to exhibit decreasing klystron body-current. The typical
reading with average picture content is 50 mA, but over a 4-week period, the reading dropped to
30 mA. No other parameters show deviation from normal. Yet, the decrease in the reading indi-
cates an alternate path (besides the normal body-current circuitry) by which electrons return to
the beam power supply. A schematic diagram of the system is shown in Figure 8.1.3. Several fac-
tors could cause the body-current variation, including water leakage into the body-to-collector
insulation of the klystron. In time, this water can corrode the klystron envelope, possibly leading
to a loss of vacuum and klystron failure.
Water leakage can also cause partial bypassing of the body-current circuitry, an important
protection system in the transmitter. It is essential that the circuit functions normally at all times
and at full sensitivity in order to detect change when a fault condition occurs. Regular logging of
transmitter parameters ensures that developing problems such as this one are caught early.
8.1.2 RF System Maintenance

Cleaning is a large part of a proper maintenance routine. A shop vacuum and clean brush are
generally all that are required. Use isopropyl alcohol and a soft cloth for cleaning insulators on
high voltage components. Cleaning affords the opportunity to inspect each component in the sys-
tem and observe any changes.
Regular maintenance of insulators is important to the proper operation of RF final amplifier
stages because of the high voltages usually present. Pay particular attention to the insulators used
y y
Figure 8.1.3 Simplified high-voltage schematic of a klystron amplifier showing the parallel leakage
path that can cause a reduction in protection sensitivity of the body-current circuit.
in the PA tube socket. Because the supply of cooling air is passed through the socket, airborne
contaminants can be deposited on various sections of the assembly. These can create a high-volt-
age arc path across the socket insulators. Perform any cleaning work around the PA socket with
extreme care. Do not use compressed air to clean out a power tube socket. Blowing compressed
air into the PA or IPA stage of a transmitter will merely move the dirt from places where you can
see it to places where you cannot see it. Use a vacuum instead. When cleaning the socket assem-
bly, do not disturb any components in PA the circuit. Visually check the tube anode to see if dirt
is clogging any of the heat-radiating fins.
Cleaning is also important to proper cooling of solid-state components in the transmitter. A
layer of dust and dirt can create a thermal insulator effect and prevent proper heat exchange from
a device into the cabinet.
8.1.2a Preventive Maintenance Routine

A thorough inspection of the RF system on a regular basis is the key to minimizing equipment
downtime. Component problems often can be spotted at an early stage by regular inspection of
the system. Remember to discharge all capacitors in the circuit with a grounding stick before
touching any component in the high-voltage sections of the transmitter. Confirm that all primary
power has been removed before any maintenance work begins.
y y
Special precautions must be taken with systems that receive ac power from two independent
feeds. Typically, one ac line provides 208 V 3-phase service for the high-voltage section of the
system, and a separate ac line provides 120 V power for low-voltage circuits. Older transmitters
or high-power transmitters often utilize this arrangement. Check to see that all ac is removed
before any maintenance work begins.
Consider the following preventive maintenance procedures.
Resistors and Capacitors

• Inspect resistors and RF capacitors for signs of overheating.
• Inspect electrolytic or oil-filled capacitors for signs of leakage.
• Inspect feed-through capacitors and other high-voltage components for signs of arcing.
Transmitting capacitors—mica vacuum and doorknob types—should never run hot. They
may run warm, but usually as the result of thermal radiation from nearby components (such as
power tubes) in the circuit. An overheated transmitting capacitor is often a sign of incorrect tun-
ing. Vacuum capacitors present special requirements for the maintenance technician. Care in
handling is a prime requisite for maximum service life. Because the vacuum capacitor is evacu-
ated to a higher degree than most vacuum tubes, it is particularly susceptible to shock and rough
handling. Provide adequate protection to vacuum capacitors whenever maintenance is per-
formed. The most vulnerable parts of the capacitor are the glass-to-metal seals on each end of the
unit. Exercise particular care during removal or installation.
The current ratings of vacuum capacitors are limited by the glass-to-metal seal temperature
and the temperature of the solder used to secure the capacitor plates. Seal temperature is
increased by poor connecting clip pressure, excessive ambient temperatures, corrosion of the end
caps and/or connecting clip, excessive dust and dirt accumulation, or excessive currents. Dust
accumulation on sharp points in high-voltage circuitry near the vacuum capacitor can cause arcs
or corona that may burn a hole through the glass envelope.
Power Supply Components

• Inspect the mechanical operation of circuit breakers. Confirm that they provide a definite
snap to the off position (remove all ac power for this test), and that they firmly reseat when
restored. Replace any circuit breaker that is difficult to reset.
• Inspect power transformers and reactors for signs of overheating or arcing.
• Inspect oil-filled transformers for signs of leakage.
• Inspect transformers for dirt build-up, loose mounting brackets and rivets, and loose terminal
connections.
• Inspect high-voltage rectifiers and transient suppression devices for overheating and mechan-
ical problems.
Power transformers and reactors normally run hot to the touch. Check both the transformer
frame and the individual windings. On a 3-phase transformer, each winding should produce
about the same amount of heat. If one winding is found to run hotter than the other two, further
investigation is warranted. Dust, dirt, or moisture between terminals of a high-voltage trans-
y y
former may cause flashover failures. Insulating compound or oil around the base of a trans-
former can indicate overheating and/or leakage.
Coils and RF Transformers

• Inspect coils and RF transformers for indications of overheating.
• Inspect connection points for arcing or loose terminals.
Coils and RF transformers operating in a well-tuned system will rarely heat appreciably. If
discoloration is noticed on several loops of a coil, consult the factory service department to see if
the condition is normal. Pay particular attention to variable tap inductors, often found in AM
broadcast transmitters and phasors. Closely inspect the roller element and coil loops for over-
heating or signs of arcing.
Relay Mechanisms
• Inspect relay contacts, including high-voltage or high-power RF relays, for signs of pitting or
discoloration.
• Inspect the mechanical linkage to confirm proper operation. The contactor arm (if used)
should move freely, without undue mechanical resistance.
• Inspect vacuum contactors for free operation of the mechanical linkage (if appropriate), and
for indications of excessive dissipation at the contact points and metal-to-glass (or metal-to-
ceramic) seals.
Unless problems are experienced with an enclosed relay, do not attempt to clean it. More
harm than good can be done by disassembling properly working components for detailed inspec-
tion.
Connection Points
• Inspect connections and terminals that are subject to vibration. Tightness of connections is
critical to the proper operation of high-voltage and RF circuits.
• Inspect barrier strip and printed circuit board contacts for proper termination.
Although it is important that all connections are tight, be careful not to over tighten. The con-
nection points on some components, such doorknob capacitors, can be damaged by excessive
force. There is no section of an RF system where it is more important to keep connections tight
than in the power amplifier stage. Loose connections can result in arcing between components
and conductors that can lead to system failure. The cavity access door is a part of the outer con-
ductor of the coaxial transmission line circuit in FM and TV transmitters, and in many RF gener-
ators operating at VHF and above. High potential RF circulating currents flow along the inner
surface of the door, which must be fastened securely to prevent arcing.
8.1.3 Klystron Devices

Klystrons are expensive to buy and expensive to operate. Compared to tetrodes, they require
larger auxiliary components (such as power supplies and heat exchangers), and are physically
y y
larger. Yet, they are stable, provide high gain, and may be easily driven by solid state circuitry.
Klystrons are relatively simple to cool and are capable of long life with a minimum of mainte-
nance. Two different types of klystrons have found service in television applications:
• Integral cavity klystron, in which the resonant cavities are built into the body.
• External cavity klystron, in which the cavities are mechanically clamped onto the body and
are outside the vacuum envelope of the device.
This difference in construction requires different maintenance procedures. The klystron body
(the RF interaction region of the integral cavity device), is cooled by the same liquid that is fed to
the collector. Required maintenance involves checking for leaks and adequate coolant flow.
Although the cavities of the external cavity unit are air-cooled, the body may be water- or air-
cooled. Uncorrected leaks in a water-cooled body can lead to cavity and tuning mechanism dam-
age. Look inside the magnet frame with a flashlight once a week. Correct leaks immediately and
clean away coolant residues.
The air-cooled body requires only sufficient airflow. The proper supply of air can be moni-
tored with one or two adhesive temperature labels and a close visual inspection. Look for discol-
oration of metallic surfaces. The external cavities need a clean supply of cooling air. Dust
accumulation inside the cavities will cause RF arcing. Check air supply filters regularly. Some
cavities have a mesh filter at the inlet flange. Inspect this point as required.
It is possible to make a visual inspection of the cavities of an external-cavity device by remov-
ing the loading loops and/or air loops. This procedure is recommended only when unusual
behavior is experienced, and not as part of routine maintenance. Generally there is no need to
remove a klystron from its magnet frame and cavities during routine maintenance.
8.1.4 Power Grid Tubes

The best way to gain an understanding of the capabilities of a PA tube is to secure a copy of the
tube manufacturer's data sheet for each type of device, available either from the tube or transmit-
ter manufacturer. The primary value of the data sheets to the end-user is the listing of maximum
permissible values. These give the maintenance engineer a clear rundown of the maximum volt-
ages and currents that the tube can withstand under normal operation. Note these values and
avoid them.
An examination of the data sheet will show that a number of operating conditions are possi-
ble, depending upon the class of service required by the application. As long as the maximum
ratings of the tube are not exceeded, a wide choice of operating parameters, including plate volt-
age and current, screen voltage, and RF grid drive, are possible. When studying the characteristic
curves of each tube, remember that they represent the performance of a typical device. All elec-
tronic products have some tolerance among devices of a single type. Operation of a given device
in a particular system may be different than that specified on the data sheet or in the transmitter
instruction manual. This effect is more pronounced at VHF and above.
8.1.4a Tube Dissipation

Proper cooling of the tube envelope and seals is a critical parameter for long tube life. Deteriorat-
ing effects that result in shortened tube life and reduced performance increase as the temperature
y y
increases. Excessive dissipation is perhaps the single greatest cause of catastrophic failure in a
power tube. PA tubes used in broadcast, industrial, and research applications can be cooled using
one of three methods: forced-air, liquid, and vapor-phase cooling. In radio and VHF-TV trans-
mitters, forced-air cooling is by far the most common method used. Forced-air systems are sim-
ple to construct and easy to maintain.
The critical points of almost every PA tube type are the metal-to-ceramic junctions or seals.
At temperatures below 250°C these seals remain secure, but above that temperature, the bonding
in the seal may begin to disintegrate. Warping of grid structures also may occur at temperatures
above the maximum operating level of the tube. The result of prolonged overheating is shortened
tube life or catastrophic failure. Several precautions are usually taken to prevent damage to tube
seals under normal operating conditions. Air directors or sections of tubing may be used to pro-
vide spot-cooling to critical surface areas of the device. Airflow sensors prevent operation of the
system in the event of a cooling system failure.
Tubes that operate in the VHF and UHF bands are inherently subject to greater heating action
than devices operated at lower frequencies (such as AM service). This effect is the result of
larger RF charging currents into the tube capacitances, dielectric losses, and the tendency of
electrons to bombard parts of the tube structure other than the grid and plate in high-frequency
applications. Greater cooling is required at higher frequencies.
The technical data sheet for a given power tube will specify cooling requirements. The end-
user is not normally concerned with this information; it is the domain of the transmitter manu-
facturer. The end-user, however, is responsible for proper maintenance of the cooling system.
Air-Handling System
All modern air-cooled PA tubes use an air-system socket and matching chimney for cooling.
Never operate a PA stage unless the air-handling system provided by the manufacturer is com-
plete and in place. For example, the chimney for a PA tube often can be removed for inspection
of other components in the circuit. Operation without the chimney, however, may significantly
reduce airflow through the tube and result in excessive dissipation of the device. It also is possi-
ble that operation without the proper chimney could damage other components in the circuit
because of excessive radiated heat. Normally the tube socket is mounted in a pressurized com-
partment so that cooling air passes through the socket and then is guided to the anode cooling
fins, as illustrated in Figure 8.1.4. Do not defeat any portion of the air-handling system provided
by the manufacturer.
Cooling of the socket assembly is important for proper cooling of the tube base, and for cool-
ing of the contact rings of the tube itself. The contact fingers used in the collet assembly of a
socket typically are made of beryllium copper. If subjected to temperatures above 150°C for an
extended period of time, the beryllium copper will lose its temper (springy characteristic) and
will no longer make good contact with the base rings of the device. In extreme cases, this type of
socket problem can lead to arcing, which can burn through the metal portion of the tube base
ring. Such an occurrence can ultimately lead to catastrophic failure of the device because of a
loss of the vacuum envelope. Other failure modes for a tube socket include arcing between the
collet and tube ring that can weld a part of the socket and tube together. The end result is failure
of both the tube and the socket.
y y
Figure 8.1.4 Airflow system for an air-cooled power tube.
Ambient Temperature
The temperature of the intake air supply is a parameter that is usually under the control of the
maintenance engineer. The preferred cooling air temperature is no higher than 75°F, and no
lower than the room dew point. The air temperature should not vary because of an oversized air
conditioning system or because of the operation of other pieces of equipment at the transmission
facility. Monitoring the PA exhaust stack temperature is an effective method of evaluating overall
RF system performance. This can be easily accomplished. It also provides valuable data on the
cooling system and final stage tuning.
Another convenient method for checking the efficiency of the transmitter cooling system over
a period of time involves documenting the back pressure that exists within the PA cavity. This
measurement is made with a manometer, a simple device that is available from most heating,
ventilation, and air-conditioning (HVAC) suppliers. The connection of a simplified manometer
to a transmitter PA input compartment is illustrated in Figure 8.1.5.
When using the manometer, be careful that the water in the device is not allowed to backflow
into the PA compartment. Do not leave the manometer connected to the PA compartment when
the transmitter is on the air. Make the necessary measurement of PA compartment back pressure
and then disconnect the device. Seal the connection point with a subminiature plumbing cap or
other appropriate hardware.
By charting the manometer readings, it is possible to accurately measure the performance of
the transmitter cooling system over time. Changes resulting from the build-up of small dust par-
ticles (microdust) may be too gradual to be detected except through back-pressure charting. Be
y y
Figure 8.1.5 A manometer device used for measuring back pressure in the PA compartment of a
transmitter.
certain to take the manometer readings during periods of calm weather. Strong winds can result
in erroneous readings because of pressure or vacuum conditions at the transmitter air intake or
exhaust ports.
Deviations from the typical back-pressure value, either higher or lower, could signal a prob-
lem with the air-handling system. Decreased PA input compartment back pressure could indicate
a problem with the blower motor or a build-up of dust and dirt on the blades of the blower assem-
bly. Increased back pressure, on the other hand, could indicate dirty PA tube anode cooling fins
or a build-up of dirt on the PA exhaust ducting. Either condition is cause for concern. A system
suffering from reduced air pressure into the PA compartment must be serviced as soon as possi-
ble. Failure to restore the cooling system to proper operation may lead to premature failure of the
PA tube or other components in the input or output compartments. Cooling problems do not
improve. They always get worse.
Failure of the PA compartment air-interlock switch to close reliably may be an early indica-
tion of impending cooling system trouble. This situation could be caused by normal mechanical
wear or vibration of the switch assembly, or it may signal that the PA compartment air pressure
has dropped. In such a case, documentation of manometer readings will show whether the trou-
ble is caused by a failure of the air pressure switch or a decrease in the output of the air-handling
system.
y y
Thermal Cycling
Most power grid tube manufacturers recommend a warm-up period between application of fila-
ment-on and plate-on commands. Most RF equipment manufacturers specify a warm-up period
of about five minutes. The minimum warm-up time is two minutes. Some RF generators include
a time delay relay to prevent the application of a plate-on command until predetermined warm-
up cycle is completed. Do not defeat these protective circuits. They are designed to extend PA
tube life. Most manufacturers also specify a recommended cool-down period between the appli-
cation of the plate-off and filament-off commands. This cool-down, generally about 10 minutes is
designed to prevent excessive temperatures on the PA tube surfaces when the cooling air is shut
off. Large vacuum tubes contain a significant mass of metal, which stores heat effectively.
Unless cooling air is maintained at the base of the tube and through the anode cooling finds,
excessive temperature rise can occur. Again, the result can be shortened tube life, or even cata-
strophic failure because of seal cracks caused by thermal stress.
Most tube manufacturers suggest that cooling air continue to be directed toward the tube base
and anode cooling fins after filament voltage has been removed to further cool the device.
Unfortunately, however, not all control circuits are configured to permit this mode of operation.
8.1.4b Tube Changing Procedure

Plug-in power tubes must be seated firmly in their sockets and the connections to the anodes of
the tubes must be tight. Once in place, do not remove a tube assembly for routine inspection
unless it is malfunctioning. Whenever a tube is removed from its socket, carefully inspect the
fingerstock for signs of overheating or arcing. Keep the socket assembly clean and all connec-
tions tight. If any part of a PA tube socket is found to be damaged, replace the defective portion
immediately. In many cases, the damaged fingerstock ring can be ordered and replaced. In other
cases, however, the entire socket must be replaced. This type of work is a major undertaking,
requiring an experienced engineer.
8.1.4c Extending Vacuum Tube Life

RF power tubes are probably the most expensive replacement part that a transmitter or RF gener-
ator will need on a regular basis. With the cost of new and rebuilt tubes continually rising, main-
tenance engineers should do everything possible to extend tube life.
Power Tube Conditioning

Whenever a new tube is installed in a transmitter, inspect the device for cracks or loose connec-
tions (in the case of tubes that do not socket-mount). Also check for interelectrode short circuits
with an ohmmeter. Tubes must be seated firmly in their sockets to allow a good, low-resistance
contact between the fingerstock and contact rings. After a new tube, or one that has been on the
shelf for some time, is installed in the transmitter, run it with filaments only for at least 30 min-
utes, after which plate voltage may be applied. Next, slowly bring up the drive (modulation), in
the case of an AM or TV visual transmitter. Residual gas inside the tube may cause an interelec-
trode arc (usually indicated by the transmitter as a plate overload) unless it is burned off in such a
warm-up procedure.
y y
Keep an accurate record of performance for each tube. Shorter than normal tube life could
point to a problem in the RF amplifier itself. The average life that may be expected from a power
grid tube is a function of many parameters, including:
• Filament voltage
• Ambient operating temperature
• RF power output
• Operating frequency
• Operating efficiency
The best estimate of life expectancy for a given system at a particular location comes from
on-site experience. As a general rule of thumb, however, at least 12 months of service can be
expected from most power tubes. Possible causes of short tube life include:
• Improper transmitter tuning.
• Inaccurate panel meters or external wattmeter, resulting in more demand from the tube than is
actually required.
• Poor filament voltage regulation.
• Insufficient cooling system airflow.
• Improper stage neutralization.
Filament Voltage
A true reading RMS voltmeter is required to accurately measure filament voltage. Make the
measurement directly from the tube socket connections. Secure the voltmeter test leads to the
socket terminals and carefully route the cables outside the cabinet. Switch off the plate power
supply circuit breaker. Close all interlocks and apply a filament on command. Do not apply the
high voltage during filament voltage tests. Serious equipment damage and/or injury to the main-
tenance engineer may result.
A true-reading RMS meter, instead of the more common average responding RMS meter, is
suggested because the true-reading meter can accurately measure a voltage despite an input
waveform that is not a pure sine wave. Some filament voltage regulators use silicon-controlled
rectifiers (SCRs) to regulate the output voltage. Do not put too much faith in the front-panel fil-
ament voltage meter. It is seldom a true-reading RMS device; most are average-responding
meters (unless otherwise specified).
Long tube life requires filament voltage regulation. Many RF systems have regulators built
into the filament supply. Older units without such circuits often can be modified to provide a
well-regulated supply by adding a ferroresonant transformer or motor-driven auto-transformer to
the ac supply input. A tube whose filament voltage is allowed to vary along with the primary line
voltage will not achieve the life expectancy possible with a tightly regulated supply. This prob-
lem is particularly acute at mountain-top installations, where utility regulation is generally poor.
To extend tube life, some broadcast engineers leave the filaments on at all times, not shutting
down at sign-off. If the sign-off period is three hours or less, this practice can be beneficial. Fila-
ment voltage regulation is a must in such situations because the primary line voltages may vary
substantially from the carrier-on to carrier-off value. Do not leave voltage on the filaments of a
y y
klystron for a period of more than two hours if no beam voltage is applied. The net rate of evapo-
ration of emissive material from the cathode surface of a klystron is greater without beam volt-
age. Subsequent condensation of the material on gun components may lead to voltage hold-off
problems and an increase in body current.
Filament Voltage Management

By accurately managing the filament voltage of a thoriated tungsten power tube, the useful life of
the device can be extended considerably, sometimes to twice the normal life expectancy. The fol-
lowing procedure is recommended:
• Operate the filament at its full-rated voltage for the first 200 hours following installation.
• Following the burn-in period, reduce the filament voltage by 0.1 V per step until power output
begins to fall (for frequency modulated systems) or until modulating waveform distortion
begins to increase (for amplitude modulated systems).
• When the emissions floor has been reached, raise the filament voltage 0.2 V.
• Long-term operation at this voltage can result in a substantial extension in the usable life of
the tube, as illustrated in Figure 8.1.6.
Do not operate the tube with a filament voltage that is at or below 90 percent of its rated
value. At regular intervals, about every three months, check the filament voltage and increase it
if power output begins to fall or distortion begins to rise. Filament voltage should never be
increased to more than 105 percent of rated voltage. Some tube manufacturers place the mini-
mum operating point at 94 percent. Others recommend that the tube be set for 100 percent fila-
ment voltage and left there. The choice of which approach to follow is left to the user.
When it becomes necessary to boost filament voltage to more than 103 percent, order a new
tube. If the old device is replaced while it still has some life remaining, the facility will have a
standby tube that will perform well as a spare.
Check the filament current when the tube is first installed, and at annual intervals thereafter,
to assure that the filament draws the desired current. Tubes can fail early in life because of an
open filament bar that would have been discovered during the warranty period if a current check
had been made upon installation.
For one week of each year of tube operation, run the filament at full-rated voltage. This will
operate the getter and clean the tube of gas.
Filament voltage is an equally important factor in achieving long life in a klystron. The volt-
ages recommended by the manufacturer must be set and checked on a regular basis. Measure the
voltage at the filament terminals and calibrate the front-panel meter as needed.
8.1.5 PA Stage Tuning

There are probably as many ways to tune the PA stage of an RF generator or transmitter as there
are types of systems. Experience is the best teacher when it comes to adjusting for peak effi-
ciency and performance. Compromises often must be made among various operating parameters.
Whatever procedure is used, document the operating parameters and steps for future reference.
Do not rely on memory for a listing of the typical operating limits and tuning procedures for the
system. Write down the information and post it at the facility. The manufacturer's service depart-
y y
Figure 8.1.6 The effects of filament voltage management on the useful life of a thoriated tungsten
filament power tube. Note the dramatic increase in emission hours when filament voltage manage-
ment is practiced.
ment can be an excellent source for information about tuning a particular unit. Many times the
factory can provide pointers on how to simplify the tuning process, or what interaction of adjust-
ments may be expected. Whatever information is learned from such conversations, write it down.
8.1.5a When to Tune

Tuning can be affected by any number of changes in the PA stage. Replacing the final tube in an
AM transmitter usually does not significantly alter stage tuning. It is advisable, however, to run
through a touch-up tuning procedure just to be sure. Replacing a tube in an FM or TV transmit-
y y
ter, on the other hand, can significantly alter stage tuning. At high frequencies, normal tolerances
and variations in tube construction result in changes in element capacitance and inductance.
Likewise, replacing a component in the PA stage may cause tuning changes because of normal
device tolerances.
Stability is one of the primary objectives of transmitter tuning. Avoid tuning positions that do
not provide stable operation. Adjust for broad peaks or dips, as required. Tune so the system is
stable from a cold startup to normal operating temperature. Readings should not vary measurably
after the first minute of operation.
Adjust tuning not only for peak efficiency, but also for peak performance. These two ele-
ments of transmitter operations, unfortunately, do not always coincide. Trade-offs must some-
times be made in order to ensure proper operation of the system. For example, FM or TV aural
transmitter loading can be critical to wide system bandwidth and low synchronous AM. Loading
beyond the point required for peak efficiency must often be used to broaden cavity bandwidth.
Heavy loading lowers the PA plate impedance and cavity Q. A low Q also reduces RF circulating
currents in the cavity.
8.1.6 Bibliography
y y
g g
Chapter
8.2
Preventing RF System Failures
8.2.1 Introduction
The reliability and operating costs over the lifetime of an RF system can be significantly
impacted by the effectiveness of the preventive maintenance program designed and implemented
by the engineering staff. When dealing with a critical-system unit such as a broadcast transmitter
that must operate on a daily basis, maintenance can have a major impact—either positive or neg-
ative—on downtime and bottom-line profitability of the facility. The sections of a transmitter
most vulnerable to failure are those exposed to the outside world: the ac-to-dc power supplies
and RF output stage. These circuits are subject to high energy surges from lightning and other
sources.
The reliability of any communications system may be compromised by an enabling event
phenomenon. An enabling event phenomenon is an event which, while not causing a failure by
itself, sets up (or enables) a second event that can lead to failure of the communications system.
This phenomenon is insidious because the enabling event is often not self-revealing. Examples
include:
• A warning system that has failed or been disabled for maintenance.
• One or more controls set incorrectly so that false readouts are provided for operations person-
nel.
• Redundant hardware that is out of service for maintenance.
• Remote metering that is out of calibration.
8.2.2 Common Mode Failures

A common mode failure is one that can lead to the failure of all paths in a redundant configura-
tion. In the design of redundant systems, therefore, it is important to identify and eliminate
sources of common mode failures, or to increase their reliability to at least an order of magnitude
above the reliability of the redundant system. Common mode failure points in a transmission
system include the following:
8-23
g y
• Switching circuits that activate standby or redundant hardware.

• Sensors that detect a hardware failure.
• Indicators that alert personnel to a hardware failure.
• Software that is common to all paths in a redundant system.
The concept of software reliability in control and monitoring has limited meaning in that a
good program will always run, and copies of the program will always run. On the other hand, a
program with one or more errors will always fail, and so will copies of the program fail, given
the same input data. The reliability of software, unlike hardware, cannot be improved through
redundancy if the software in the parallel path is identical to the primary path.
8.2.2a Modifications and Updates

If problems are experienced with a system, examine what can be done to prevent the failure from
occurring again. A repeat of the problem can often be avoided by installing various protection
devices or consulting the factory for updates to the hardware. If the transmitter is several years
old, the factory service department can detail any changes that may have been made in the unit to
provide more reliable operation. Many of these modifications are minor and can be incorporated
into older models with little cost or effort. Modifications could include the following items:
• Changing a variable capacitor in a critical tuning stage to a vacuum variable for more stabil-
ity.
• Installing additional filtering in the high voltage power supply to improve AM noise perfor-
mance.
• Replacing older technology transistorized circuit boards with newer IC and power semicon-
ductor PWBs to improve reliability and performance.
• Improving the overload protection circuitry through the addition of new logic circuits.
• Adding transient protection devices at critical states of the transmitter.
8.2.2b Spare Parts

The spare parts inventory is a key aspect of any successful equipment maintenance program.
Having adequate replacement components on hand is important not only to correct equipment
failures, but in identifying those failures as well. Many parts—particularly in the high-voltage
power supply and RF chain—are difficult to test under static conditions. The only reliable way to
test the component may be to substitute one of known quality. If the system returns to normal
operation, then the substituted component is defective. Substitution is also a valuable tool in
troubleshooting intermittent failures caused by component breakdown under peak power condi-
tions.
g y
8.2.3 Transmission Line/Antenna Problems

The voltage standing wave ratio (VSWR) of an antenna and its transmission line is a vital param-
eter that has a considerable effect on the performance and reliability of a transmission system.
VSWR is a measure of the amount of power reflected back to the transmitter because of an
antenna and/or transmission line mismatch. A mismatched or defective transmission system will
result in a high degree of reflected power, or a higher VSWR.
The amount of reflected power that a given system can accept is a function of the application.
For example, it is common practice in FM broadcasting to maintain a VSWR of 1.1:1 as the
maximum level within the transmission channel that can be tolerated without degrading the qual-
ity of the on-air signal. For conventional TV broadcasting, a VSWR into the antenna feeder of
more than 1.04:1 will start to degrade picture quality, particularly on systems that have a long
transmission line. Reflections down the line from a mismatch at the antenna disrupt the perfor-
mance of the transmitter output stage. The reflections also cause multipath distortion within the
transmission line itself. When power is reflected back to the transmitter, it causes the RF output
stage to look into a mismatched load with unpredictable phase and impedance characteristics.
Because of the reflective nature of VSWR on a transmission system, the longer the transmission
line (assuming the reflection is originating at the antenna), the more severe the problem may be
for a given VSWR. A longer line means that reflected power seen at the RF output stage has
greater time (phase) delays, increasing the reactive nature of the load.
VSWR is affected not only by the rating of the antenna and transmission line as individual
units, but also by the combination of the two as a system. The worst-case system VSWR is equal
to the antenna VSWR multiplied by the transmission line VSWR. For example, if an antenna
with a VSWR of 1.05:1 is connected to a line with a VSWR of 1.05:1, the resulting worst-case
system VSWR would be 1.1025:1. Given the right set of conditions, an interesting phenomenon
can occur in which the VSWR of the antenna cancels the transmission line VSWR, resulting in a
perfect 1:1 match. The determining factors for this condition are the point of origin of the
antenna VSWR, the length of transmission line, and the observation point.
8.2.3a Effects of Modulation

The VSWR of a transmission system is a function of frequency and changes with carrier modu-
lation. This change may be large or small, but it will occur to some extent. The cause can be
traced to the frequency dependence of the VSWR of the antenna (and to a lesser extent, the trans-
mission line.) The effects of frequency on VSWR can be observed in Figure 8.2.1. Although this
plot of VSWR-versus-frequency for a common FM antenna is good, notice that with no modula-
tion the system VSWR is one figure. VSWR measurements are different with positive modula-
tion (carrier plus modulation) and negative modulation (carrier minus modulation).
VSWR is further complicated because power reflected back to the transmitter from the
antenna may not come from a single point, but from a number of different points. One reflection
might be caused by the antenna-matching unit, another by various flanges in the line, and a third
by a damaged part of the antenna itself. Because these reflection points are different lengths from
the transmitter PA plate, a variety of standing waves can be generated along the line, varying
with the modulating frequency.
Energy reflected back to the transmitter from the antenna is not lost. A small percentage of
the energy is turned into heat, but the majority of it is radiated by the antenna, delayed in time by
length of the transmission line.
g y
Figure 8.2.1 The measured performance of a single-channel FM antenna (tuned to 92.3 MHz).
The antenna provides a VSWR of less than 1.1:1 over a frequency range of 300 kHz.
8.2.3b Maintenance Considerations

In order to maintain low VSWR, the transmission line and antenna system should be serviced
regularly. Consider the following guidelines:
• Inspect the antenna elements, interconnecting cables, impedance transformers, and support
braces at least once each year. Falling ice can damage FM, TV, and communications antenna
elements if proper precautions are not taken. Icing on the elements of an antenna will degrade
the antenna VSWR because ice lowers the frequency of the electrical resonance of the
antenna. Two methods are commonly used to prevent a build-up of ice on high power trans-
mitting antennas: electrical de-icers and radomes.
• Check AM antennas regularly for structural integrity. Because the tower itself is the radiator,
bond together each section of the structure for good electrical contact.
• Clean base insulators and guy insulators (if used) as often as required.
• Keep lighting ball gaps or other protective devices clean and properly adjusted.
• Inspect the transmission line for signs of damage. Check supporting hardware and investigate
any indication of abnormal heating of the line.
• Keep a detailed record of VSWR in the facility maintenance log and investigate any increase
above the norm.
• Regularly check the RF system test load. Inspect the coolant filters and flow rate, as well as
the resistance of the load element.
g y
8.2.3c UHF Transmission Systems

The RF transmission system of high-power UHF stations may be externally diplexed after the
final RF amplifiers, using either coaxial or waveguide-type diplexers. Maintenance of the com-
bining sections is largely a matter of careful observation and record keeping:
• Monitor the reject loads on diplexers and power combiners regularly to ensure adequate cool-
ing.
• Check the temperature of the transmission line and related components, particularly coaxial
elements. Keep in mind that coax of the same size and carrying the same RF power runs
warmer in UHF systems than in VHF systems. This phenomenon is caused by the reduced
penetration depth (the skin effect) of UHF signal currents. Hot spots in the transmission line
can be caused by poor contact areas or by high VSWR. If they are the result of a VSWR con-
dition, the hot spots will be repeated every 1/2-wavelength toward the transmitter.
• Monitor the reverse power/VSWR meters closely. Some daily variation is not unusual, in
small amounts. Greater variations that are cyclical in nature are an indication of a long-line
problem—most likely at the antenna. Because transmission lines are usually long at UHF (a
taller tower allows greater coverage) and the wavelength is small leads to large phase changes
of a mismatch at the antenna. Mismatches inside the building do not cause the same cyclical
variation. If reverse power is observed to vary significantly, run the system with the test load
to see whether the problem disappears. If the variations are not present with a test load,
arrange for an RF sweep of the line.
A change in klystron output power is another effect of VSWR variation on a UHF transmitter.
The output coupler transforms the line characteristic impedance upward to approximately match
the beam impedance. This provides maximum power transfer from the cavity. Large VSWR
phase variations associated with long lines change the impedance that the output coupler sees.
This causes the output power to vary, sometimes more significantly than the reverse power
metering indicates. Ghosting on the output waveform is a common indication of antenna VSWR
problems in a long-line conventional (analog) TV system. If the input signal is clean and the out-
put has a ghost, arrange for an RF sweep.
8.2.4 High Voltage Power Supply Problems

The high-voltage plate supply is the first line of defense between external ac line disturbances
and the power amplifier stage(s). Next to the output circuit itself, the plate supply is the second
section of a transmitter most vulnerable to damage because of outside influences.
8.2.4a Power Supply Maintenance

Figure 8.2.2 shows a high-reliability power supply of the type common in transmission equip-
ment. Many transmitters use simpler designs, without some of the protection devices shown, but
the principles of preventive maintenance are the same:
• Thoroughly examine every component in the high-voltage power supply. Look for signs of
leakage on the main filter capacitor (C2).
g y
Figure 8.2.2 A common high-voltage transmitter power supply design.
• Check all current-carrying meter/overload shunt resistors (R1–R3) for signs of overheating.
• Carefully examine the wiring throughout the power supply for loose connections.
g y
• Examine the condition of the filter capacitor series resistors (R4 and R5), if used, for indica-
tions of overheating. Excessive current through these resistors could point to a pending failure
in the associated filter capacitor.
• Examine the condition of the bleeder resistors (R6–R8). A failure in one of the bleeder resis-
tors could result in a potentially dangerous situation for maintenance personnel by leaving the
main power supply filter capacitor (C2) charged after the removal of ac input power.
• Examine the plate voltage meter multiplier assembly (A1) for signs of resistor overheating.
Replace any resistors that are discolored with the factory-specified type.
When changing components in the transmitter high-voltage power supply, be certain to use
parts that meet with the approval of the manufacturer. Do not settle for a close match of a
replacement part. Use the exact replacement part. This ensures that the component will work as
intended and will fit in the space provided in the cabinet.
8.2.4b Power Supply Metering

Proper metering is one of the best ways to prevent failures in transmission equipment. Accurate
readings of plate voltage and current are fundamental to RF system maintenance. Check each
meter for proper mechanical and electrical operation. Replace any meter that sticks or will not
zero.
Overload Sensor
The plate supply overload sensor in some transmitters is arranged as shown in Figure 8.2.2. An
adjustable resistor—either a fixed resistor with a movable tap or a potentiometer—is used to set
the sensitivity of the plate overload relay. Check potentiometer-type adjustments periodically.
Fixed-resistor-type adjustments rarely require additional attention. Most manufacturers have a
chart or mathematical formula that may be used to determine the proper setting of the adjustment
resistor (R9) by measuring the voltage across the overload relay coil (K1) and observing the
operating plate current value. Clean the overload relay contacts periodically to ensure proper
operation. If mechanical problems are encountered with a relay, replace it.
Transmitter control logic for a high power UHF system is usually configured for two states of
operation:
• An operational level, which requires all the “life-support” systems to be present before the
HV command is enabled.
• An overload level, which removes HV when one or more fault conditions occur.
Inspect the logic ladder for correct operation at least once a month. At longer intervals, per-
haps annually, check the speed of the trip circuits. (A storage oscilloscope is useful for this mea-
surement.) Most klystrons require an HV removal time of less than 100 ms from the occurrence
of an overload. If the trip time is longer, damage may result to the klystron. Pay particular atten-
tion to the body-current overload circuits. Occasionally check the body current without applied
drive to ensure that the dc value is stable. A relatively small increase in dc body current can lead
to overheating problems.
g y
The RF arc detectors in a UHF transmitter also require periodic monitoring. External cavity
klystrons generally have one detector in each of the third and fourth cavities. Integral devices use
one detector at the output window. A number of factors can cause RF arcing, including:
• Overdriving the klystron
• Mistuning the cavities
• Poor cavity fit (external type only)
• Under coupling of the output
• High VSWR
Regardless of the cause, arcing can destroy the vacuum seal, if drive and/or HV are not
removed quickly. A lamp is usually included with each arc detector photocell for test purposes.
8.2.5 Transient Disturbances

Every electronic installation requires a steady supply of clean power in order to function prop-
erly. Recent advances in technology have made the question of ac power quality even more
important, as computers are integrated into transmission equipment. Different types and makes
of transmitters have varying degrees of transient overvoltage protection. Given the experience of
the computer industry, it is hard to overprotect electronic equipment from ac line disturbances.
Figure 8.2.2 shows surge suppression at two points in the power supply circuit. C1 and R4
make up an R/C snubber network that is effective in shunting high-energy, fast-rise time spikes
that may appear at the output of the rectifier assembly (CR1–CR6). Similar R/C snubber net-
works (R10–R12 and C3–C8) are placed across the secondary windings of each section of the 3-
phase power transformer. Any signs of resistor overheating or capacitor failure are an indication
of excessive transient activity on the ac power line. Transient disturbances should be suppressed
before the ac input point of the transmitter.
Assembly CR7 is a surge suppression device that should be given careful attention during
each maintenance session. CR7 is typically a selenium thyrector assembly that is essentially
inactive until the voltage across the device exceeds a predetermined level. At the trip point, the
device will break over into a conducting state, shunting the transient overvoltage. CR7 is placed
in parallel with L1 to prevent damage to other components in the transmitter in the event of a loss
of RF excitation to the final stage. A sudden drop in excitation will cause the stored energy of L1
to be discharged into the power supply and PA circuits in the form of a high-potential pulse. Such
a transient can damage or destroy filter, feedthrough, or bypass capacitors; damage wiring; or
cause PA tube arcing. CR7 prevents these problems by dissipating the stored energy in L1 as
heat. Investigate discoloration or other outward signs of damage to CR7. Such an occurrence
could indicate a problem in the exciter or IPA stage of the transmitter. Immediately replace CR7
if it appears to have been stressed.
Check spark gap surge suppressor X1 periodically for signs of overheating. X1 is designed to
prevent damage to circuit wiring in the event that one of the meter/overload shunt resistors (R1–
R3) opens. Because the spark gap device is nearly impossible to accurately test in the field and is
relatively inexpensive, it may be advisable to replace the component every few years.
g y
Figure 8.2.3 A protection circuit using relays for utility company ac phase-loss protection.
8.2.5a Single Phasing

Any transmitter using a 3-phase ac power supply is subject to the problem of single-phasing, the
loss of one of the three legs from the primary ac power distribution source. Single phasing is usu-
ally a utility company problem, caused by a downed line or a blown pole-mounted fuse. The loss
of one leg of a 3-phase line results in a particularly dangerous situation for 3-phase motors,
which will overheat and sometimes fail. AM transmitters utilizing pulse width modulation
(PWM) systems are also vulnerable to single-phasing faults. PWM AM transmitters can suffer
catastrophic failure of the plate power supply transformer as a result of the voltage regulation
characteristics of the modulation system. The PWM circuit will attempt to maintain carrier and
sideband power through the remaining legs of the 3-phase supply. This forces the active trans-
former section and its associated rectifier stack to carry as much as three times the normal load.
Figure 8.2.3 shows a simple protection scheme that has been used to protect transmission
equipment from damage caused by single phasing. Although at first glance the system looks as if
it would easily handle the job, operational problems can result. The loss of one leg of a 3-phase
line rarely results in zero (or near-zero) voltages in the legs associated with the problem line.
Instead, a combination of leakage currents caused by regeneration of the missing legs in induc-
tive loads and the system load distribution usually results in voltages of some sort on the fault
legs of the 3-phase supply. It is possible, for example, to have phase-to-phase voltages of 220 V,
185 V, and 95 V on the legs of a 3-phase, 208 Vac line experiencing a single-phasing problem.
These voltages often change, depending upon the equipment turned on at the transmitter site.
Integrated circuit technology has provided a cost-effective solution to this common design
problem in medium- and high-power RF equipment. Phase-loss protection modules are available
from several manufacturers that provide a contact closure when voltages of proper magnitude
and phase are present on the monitored line. The relay contacts can be wired into the logic con-
trol ladder of the transmitter to prevent the application of primary ac power during a single-phas-
ing condition. Figure 8.2.4 shows the recommended connection method. Note that the input to
the phase monitor module is taken from the final set of 3-phase blower motor fuses. In this way,
g y
Figure 8.2.4 A high-performance single-phasing protection circuit using a phase-loss module as

the sensor.
any failure inside the transmitter that might result in a single-phasing condition is taken into
account. Because 3-phase motors are particularly sensitive to single-phasing faults, the relay
interlock is tied into the filament circuit logic ladder. For AM transmitters utilizing PWM
schemes, the input of the phase loss protector is connected to the load side of the plate circuit
breaker. The phase-loss protector shown in the figure includes a sensitivity adjustment for vari-
ous nominal line voltages. The unit is small and relatively inexpensive. If your transmitter does
not have such a protection device, consider installing one. Contact the factory service depart-
ment for recommendations on the connection methods that should be used.
8.2.6 Temperature Control

The environment in which the transmitter is operated is a key factor in determining system reli-
ability. Proper temperature control must be provided for the transmitter to prevent thermal
fatigue in semiconductor components and shortened life in vacuum tubes. Problems can be
avoided if preventive maintenance work is performed on a regular basis.
8.2.6a Cooling System Maintenance

Each RF transmission system is unique, and requires an individual assessment of cooling system
needs. Still, a number of common preventive maintenance tasks apply to nearly all systems:
• Keep all fans and blowers clear of dirt, dust, and other foreign material that might restrict air-
flow. Check the fan blades and blower impellers for any imbalance conditions that could
result in undue bearing wear or damage. Inspect belts for proper tension, wear, and alignment.
• Regularly clean the blower motor. Motors are usually cooled by the passage of air over the
component. If the ambient air temperature is excessive or the air flow is restricted, the lubri-
cant will gradually be vaporized from the motor bearings and bearing failure will occur. If
g y
dirty air passes over the motor, the accumulation of dust and dirt must be blown out of the
device before the debris impairs cooling.
• Follow the manufacturer's recommendations for suggested frequency and type of lubrication.
Bearings and other moving parts may require periodic lubrication. Carefully follow any spe-
cial instructions on operation or maintenance of the cooling equipment.
• Inspect motor-mounting bolts periodically. Even well-balanced equipment experiences some
vibration, which can cause bolts to loosen over time.
• Inspect air filters weekly and replace or clean them as necessary. Replacement filters should
meet original specifications.
• Clean dampers and all ducting to avoid airflow restrictions. Lubricate movable and mechani-
cal linkages in dampers and other devices as recommended. Check actuating solenoids and
electromechanical components for proper operation. Movement of air throughout the trans-
mitter causes static electrical charges to develop. Static charges can result in a buildup of dust
and dirt in duct-work, dampers, and other components of the system. Filters should remove
the dust before it gets into the system, but no filter traps every dust particle.
• Check thermal sensors and temperature system control devices for proper operation.
8.2.6b Air Cooling System Design

Transmitter cooling system performance is not necessarily related to airflow volume. The cool-
ing capability of air is a function of its mass, not its volume. The designer must determine an
appropriate airflow rate within the equipment and establish the resulting resistance to air move-
ment. A specified static pressure that should be present within the ducting of the transmitter can
be a measure of airflow. For any given combination of ducting, filters, heat sinks, RFI honey-
comb shielding, tubes, tube sockets, and other elements in the transmitter, a specified system
resistance to airflow can be determined. It is important to realize that any changes in the position
or number of restricting elements within the system will change the system resistance and, there-
fore, the effectiveness of the cooling. The altitude of operation is also a consideration in cooling
system design. As altitude increases, the density (and cooling capability) of air decreases. A cal-
culated increase in airflow is required to maintain the cooling effectiveness that the system was
designed to achieve.
Figure 8.2.5 shows a typical 20 kW FM transmitter plant. The building is oriented so that the
cooling activity of the blowers is aided by normal wind currents during the summer months. Air
brought in from the outside for cooling is well-filtered in a hooded air intake assembly that holds
several filter panels. The building includes two air conditioners, one 15,000 BTU and the other
10,000 BTU. The thermostat for the smaller unit is set for slightly greater sensitivity than the
larger air conditioner, allowing small temperature increases to be handled more economically.
It is important to keep the transmitter room warm during the winter, as well as cool during the
summer. Install heaters and PA exhaust recycling blowers as needed. A transmitter that runs 24
hours a day normally will not need additional heating equipment, but stations that sign off for
several hours during the night should be equipped with electric room heaters (baseboard types,
for example) to keep the room temperature above 50°. PA exhaust recycling can be accomplished
by using a thermostat, relay logic circuit, and solenoid-operated register or electric blower. By
g y
Figure 8.2.5 A typical heating and cooling arrangement for a 20 kW FM transmitter installation.
Ducting of PA exhaust air should be arranged so that it offers minimum resistance to airflow.
controlling the room temperature to between 60°F and 70°F, tube and component life will be
improved substantially.
8.2.6c Layout Considerations

The layout of a transmitter room HVAC system can have a significant impact on the life of the
PA tube(s) and the ultimate reliability of the transmitter. Air intake and output ports must be
designed with care to avoid airflow restrictions and back-pressure problems. This process, how-
ever, is not as easy as it may seem. The science of airflow is complex and generally requires the
advice of a qualified HVAC consultant.
To help illustrate the importance of proper cooling system design and the real-world problems
that some facilities have experienced, consider the following case histories:
Case 1
A fully automatic building ventilation system (Figure 8.2.6) was installed to maintain room tem-
perature at 20°C during the fall, winter, and spring. During the summer, however, ambient room
temperature would increase to as much as 60°C. A field survey showed that the only building
exhaust route was through the transmitter. Therefore, air entering the room was heated by test
equipment, people, solar radiation on the building, and radiation from the transmitter itself
g y
before entering the transmitter. The

problem was solved through the addi-
tion of an exhaust fan (3,000 CFM). The
1 hp fan lowered room temperature by
20°C.
Case 2
A simple remote installation was con-
structed with a heat recirculating feature
for the winter (Figure 8.2.7). Outside
supply air was drawn by the transmitter Figure 8.2.6 Case study in which excessive sum-
cooling system blowers through a bank mertime hearing was eliminated through the addition
of air filters and hot air was exhausted of a 1 hp exhaust blower to the building. (Courtesy of
through the roof. A small blower and Harris.)
damper were installed near the roof exit
point. The damper allowed hot exhaust
air to blow back into the room through a
tee duct during winter months. For sum-
mer operation, the roof damper was
switched open and the room damper
closed. For winter operation, the
arrangement was reversed. The facility,
however, experienced short tube life
during winter operation, even though
the ambient room temperature during
winter was not excessive.
The solution involved moving the
roof damper 12 ft. down to just above Figure 8.2.7 Case study in which excessive back-
pressure to the PA cavity occurred during winter
the tee. This eliminated the stagnant “air
months, when the rooftop damper was closed. The
cushion” above the bottom heating duct
problem was eliminated by repositioning the damper
damper and significantly improved air as shown. (Courtesy of Harris.)
flow in the region. Cavity back pressure
was, therefore, reduced. With this rela-
tively simple modification, the problem
of short tube life disappeared.
Case 3
An inconsistency regarding test data was discovered within a transmitter manufacturer's plant.
Units tested in the engineering lab typically ran cooler than those at the manufacturing test facil-
ity. Figure 8.2.8 shows the test station difference, a 4-foot exhaust stack that was used in the engi-
neering lab. The addition of the stack increased airflow by up to 20 percent because of reduced
air turbulence at the output port, resulting in a 20°C decrease in tube temperature.
These examples point out how easily a cooling problem can be caused during HVAC system
design. All power delivered to the transmitter is either converted to RF energy and sent to the
g y
antenna or becomes heated air. Proper

design of a cooling system, therefore, is a
part of transmitter installation that should
not be taken lightly.
8.2.6d Air Filters

Once the transmitter is clean, keeping it
that way for long periods of time may
require improving the air filtering system.
Most filters are inadequate to keep out
very small dirt particles (microdust),
which can become a serious problem in an Figure 8.2.8 Case study in which air turbulence at
unusually dirty environment. Microdust the exhaust duct resulted in reduced airflow
also can become a problem in a relatively through the PA compartment. The problem was
clean environment after a number of years eliminated by adding a 4 ft extension to the output
of operation. In addition to providing a duct. (Courtesy of Harris.)
well-filtered air intake port for the trans-
mitter building, an additional air filter can
be placed in front of the normal transmitter filter assembly. With the extra filter in place, it gen-
erally is necessary only to replace or clean the outer filter panel. The transmitter's integral filter
assembly will stay clean, eliminating the work and problems associated with pulling the filter out
while the transmitter is operating. Be certain that the addition of supplemental filtering does not
restrict airflow into the transmitter.
8.2.6e Klystron Water Cooling Systems

The cooling system is vital to any transmitter. In a UHF unit, the cooling system may dissipate a
significant percentage of the input ac power in the form of waste heat in the klystron collector.
For vapor phase-cooled klystrons, pure (distilled or demineralized) water must be used. Because
the collector is only several volts above ground potential, it is not necessary to use de-ionized
water. The collector and its water jacket act like a distillery. Any impurities in the water will
eventually find their way into the water jacket and cause corrosion of the collector. It is essential
to use high-purity water with low conductivity, less than 10 mS/cm (millisiemens per centime-
ter), and to replace the water in the cooling jacket as needed. Efficient heat transfer from the col-
lector surface into the water is essential for long klystron life. Oil, grease, soldering flux residue
and pipe sealant containing silicone compounds must be excluded from the cooling system. This
applies to both vapor- and liquid-conduction cooling systems, although it is usually more critical
in the vapor-phase type. The sight glass in a vapor-phase water jacket provides a convenient
checkpoint for coolant condition. Look for unusual residues, oil on the surface, foaming and dis-
coloration. If any of these appear, contact the manufacturer for advice on how to flush the sys-
tem.
Water quality is essential to proper operation of a liquid-cooled klystron. In general, greater
flows and greater pressures are inherent in liquid-cooled-versus-vapor-phase systems, and when
a leak occurs, large quantities of coolant can be lost before the problem is discovered. Inspect the
condition of gaskets, seals, and fittings regularly. Most liquid-cooled klystrons use a distilled
g y
water and ethylene glycol mixture. Do not exceed a 50:50 mix by volume. The heat transfer of
the mixture is lower than that of pure water, requiring the flow to be increased, typically by 20–
25 percent. Greater coolant flow means higher pressure and suggests close observation of the
cooling system after adding the glycol. Allow the system to heat and cool several times. Then
check all plumbing fittings for tightness.
The action of heat and air on ethylene glycol causes the formation of acidic products. The
acidity of the coolant can be checked with litmus paper. Buffers can and should be added with
the glycol mixture. Buffers are alkaline salts that neutralize acid forms and prevent corrosion.
Because they are ionizable chemical salts, the buffers cause conductivity of the coolant to
increase. Measure the collector-to-ground resistance periodically. Coolant conductivity is typi-
cally acceptable if the resistance caused by the coolant is greater than 20 times the resistance of
the body-metering circuitry.
Experience has shown that a good way to ensure good coolant condition is to drain, flush, and
recharge the system every spring. The equipment manufacturer can provide advice on how this
procedure should be carried out and can recommend types of glycol to use. Maintain unrestricted
airflow over the heat exchanger coils and follow the manufacturers' instructions on pump and
motor maintenance.
8.2.7 Bibliography
g y
g g
Chapter
8.3
Troubleshooting RF Equipment
8.3.1 Introduction
Problems will occur from time to time with any piece of equipment. The best way to prepare for
a transmitter failure is to know the equipment well. Study the transmitter design and layout.
Know the schematic diagram and what each component does. Examine the history of the trans-
mitter by reviewing old maintenance logs to see what components have failed in the past.
8.3.2 Troubleshooting Procedure

When a problem occurs, the first task is to keep the transmitter on the air. If a standby transmitter
is available, the solution is obvious. If the facility does not have a standby, quick thinking will be
needed to minimize downtime and keep the unit running until repairs can be made. Most trans-
mitters have sufficient protective devices so that it is impossible to operate them with serious
problems. If the transmitter will not stay on the air on normal power, try lowering the power out-
put and see if the trip-offs are eliminated. Failing this, many transmitters have driver outputs that
can be connected to the antenna on a temporary basis, thereby bypassing the final amplifiers—
provided, of course, the failure is in one of the PA stages. Do not allow the transmitter to operate
at any power level if the meter readings are out of tolerance. Serious damage can result to the
system.
When presented with a problem, proceed in an orderly manner to trace it down. Many failures
are simple to repair if you stop and think about what's happening. Examine the last set of trans-
mitter readings and make a complete list of meter readings in the failure mode. Note which over-
load lamps are lit, and what other indicators are in an alarm state. With this information
assembled, the cause of the failure can often be identified. Looking over the available data and
the schematic diagram for 10 to 15 minutes can save hours of trial-and-error troubleshooting.
When checking inside the unit, look for changes in the physical appearance of components in
the problem area. An overheated resistor or leaky capacitor may be the cause of the problem, or
may point to the cause. Devices never fail without a reason. Try and piece together the sequence
of events that led to the problem. Then, the cause of the failure—not just the more obvious symp-
toms—will be corrected. When working with direct-coupled transistors, a failure in one device
8-39
g q p
will often cause a failure in another, so check all semiconductors associated with one found to be
defective.
In high power transmitters, look for signs of arcing in the RF compartments. Loose connec-
tions and clamps can cause failures that are hard to locate. Never rush through a troubleshooting
job. A thorough knowledge of the theory of operation and history of the transmitter is a great aid
in locating problems in the RF sections. Do not overlook the possibility of tube failure when
troubleshooting a transmitter. Tubes can fail in unusual ways; substitution may be the only prac-
tical test for power tubes used in modern transmitters.
Study the control ladder of the transmitter to identify interlock or fail-safe system problems.
Most newer transmitters have troubleshooting aids built in to help locate problems in the control
ladder. Older transmitters, however, often require a moderate amount of investigation before
repairs can be accomplished.
8.3.2a The “Quick Fix”

There is no such thing as a quick fix when it comes to transmission equipment. Carefully analyze
the problem and allow ample time for repair. It makes little sense to rush through a repair job in
order to get the system back on the air if another failure occurs right after the technician walks
out the door. Careful analysis of the cause and effects of the failure will ensure that the original
problem is solved, not just its obvious symptoms. If temporary repairs must be made in order to
return the transmitter to a serviceable condition, make them and then finish the job as soon as the
needed replacement parts are available.
8.3.2b Factory Service Assistance

Factory service engineers are available to aid in troubleshooting transmission equipment, but
such services have their limits. No factory engineer can fix a transmitter over the phone. The fac-
tory can suggest areas of the system to investigate and relate the solutions to similar failure
modes, but the facility engineer is the person who does the repair work. If the engineer knows the
equipment, and has done a good job of analyzing the problem, the factory can help. When calling
the factory service department, have the following basic items on hand:
• The type of transmitter and the exact failure mode. The service department will need to know
the meter readings before and after the problem, and whether any unusual circumstances pre-
ceded the failure. For example, it would be important for the factory to know that the failure
occurred after a brief power outage, or during an ice storm.
• A list of what has already been done in an effort to correct the problem. All too often the fac-
tory is called before any repair efforts are made. The service engineer will need to know what
happens when the high voltage is applied, and what overloads may occur.
• A copy of the transmitter diagram and component layout drawings on hand. A thorough
knowledge of the transmitter design and construction allows the maintenance engineer to
intelligently converse with the factory service representative.
g q p
8.3.3 Plate Overload Fault

Of all the problems that can occur in a transmitter, probably the best known—and most feared—
is the plate supply overload. Occasional plate trip-offs (one or two a month) are not generally
cause for concern. Most of these occurrences can be attributed to power line transients. More fre-
quent trip-offs require a closer inspection of the transmission system. For the purposes of our
discussion, we will assume that the plate supply overload occurs frequently enough to make con-
tinued operation of the transmitter difficult.
8.3.3a Troubleshooting Procedure

The first step in any transmitter troubleshooting procedure is to switch the system to local con-
trol so that the maintenance technician, not an off-site operator, has control over the unit. This is
important for safety reasons. Next switch the transmitter automatic recycle circuit off. While
troubleshooting, the transmitter should not be allowed to cycle through an overload any more
times than absolutely necessary. Such action only increases the possibility of additional compo-
nent damage. Use a logical, methodical approach to finding the problem. The following proce-
dure is recommended:
• Determine the fault condition. When the maintenance engineer arrives at the transmitter site,
the unit will probably be down. The carrier will be off, but the filaments will still be lit. Check
all multimeter readings on the transmitter and exciter. If they indicate a problem in a low volt-
age stage, troubleshoot that failure before bringing the high voltage up.
• Assuming that all low voltage systems are operating normally, switch the filaments off and
make a quick visual check inside the transmitter cabinet. Determine whether there is any
obvious problem. Pay particular attention to the condition of power transformers and high
voltage capacitors. Check for signs of arcing in the PA compartment. Look on the floor of the
transmitter and in the RF compartments to see if there are any pieces of components laying
around. Sniff inside the cabinet for hints of smoke. Check the circuit breakers and fuses to see
what failures might be indicated.
• After running through these preliminary steps, restart the filaments. After the appropriate
warm-up period, bring up the high voltage. Watch the front panel meters to see how they
react. Observe what happens and listen for any sound of arcing. If the transmitter will come
up, quickly run through the PA and IPA meter readings. Check the VSWR meter for excessive
reflected power.
• Assuming that problems persist, determine whether the plate supply overload is RF- or dc-
based. With the plate off, switch the exciter off. Bring up the high voltage (plate supply). If
the overload problem remains, the failure is based in the dc high voltage power supply. If the
problem disappears, the failure is centered in the transmitter RF chain. Proper bias must be
present on all vacuum tube stages of the RF system when this test is performed. The PA tube
bias supply is usually switched on with the filaments, and can generally be read from the front
panel of the transmitter. Confirm proper bias before applying high voltage with no excitation.
It is also important that the exciter is switched off while the high voltage is off. Removing
excitation from a transmitter while it is on the air can result in a large transient overvoltage,
which can cause arcing or component damage.
g q p
• If the overload is based in the high voltage dc power supply, shut down the transmitter and
check the schematic diagram for the location in the circuit of the plate overload sensor relay
(or comparator circuit). This will indicate within what limits component checking will be
required. The plate overload sensor is usually found in one of two locations: the PA cathode
dc return, or the high voltage power supply negative connection to ground. Transmitters using
a cathode overload sensor generally have a separate high voltage dc overload sensor in the
plate power supply.
• A sensor in the cathode circuit will substantially reduce the area of component checking
required. A plate overload with no excitation in such an arrangement would almost certainly
indicate a PA tube failure, because of either an inter-electrode short circuit or a loss of vac-
uum. Do not operate the transmitter when the PA tube is out of its socket. This is not an
acceptable method of determining whether a problem exits with the PA tube. Substitute a
spare tube instead. Operating a transmitter with the PA tube removed can result in damage to
other tubes in the transmitter when the filaments are on, and damage to the driver tubes and
driver output/PA input circuit components when the high voltage is on.
• If circuit analysis indicates a problem in the high voltage power supply itself, use an ohmme-
ter to check for short circuits. Remove all power from the transmitter and discharge all filter
capacitors before beginning any troubleshooting work inside the unit. When checking for
short circuits with an ohmmeter, take into account the effects that bleeder resistors and high
voltage meter multiplier assemblies can have on resistance readings. Most access panels on
transmitters use an interlock system that will remove the high voltage and ground the high
voltage supplies when a panel is removed. For the purposes of ohmmeter tests, these inter-
locks may have to be temporarily defeated. Never defeat the interlocks unless all ac power has
been removed from the transmitter and all filter capacitors have been discharged using the
grounding stick supplied with the transmitter.
Following the preliminary ohmmeter tests, check the following components in the dc plate
supply:
• Oil-filled capacitors for signs of overheating or leakage.
• Feed-through capacitors for signs of arcing or other damage.
• The dc plate blocking capacitor for indications of insulation breakdown or arcing.
• All transformers and chokes for signs of overheating or winding failure.
• Transient suppression devices for indications of overheating or failure.
• Bleeder resistors for signs of overheating.
• Any surge-limiting resistors placed in series with filter capacitors in the power supply for
indications of overheating or failure. A series resistor that shows signs of overheating can be
an indication that the associated filter capacitor has failed.
If the plate overload trip-off occurs only at elevated voltage levels, ohmmeter checks will not
reveal the cause of the problem. It may be necessary, therefore, to troubleshoot the problem using
the process of elimination.
g q p
8.3.3b Process of Elimination

Troubleshooting through the process of elimination involves isolating various portions of the cir-
cuit—one section at a time—until the defective component is found. Special precautions are
required before performing such work:
• Never touch anything inside the transmitter without first removing all ac power and then dis-
charging all filter capacitors with the grounding stick.
• Never perform troubleshooting work alone; another person should always be present.
• Whenever a wire is disconnected, temporarily wrap it with electrical tape and secure the con-
nector so it will not arc-over to ground or another component when power is applied.
• Analyze each planned test before it is conducted. Every test in the troubleshooting process
requires time, and so steps should be arranged to provide the greatest amount of information
about the problem.
• Check with the transmitter manufacturer to find out what testing procedures the company rec-
ommends. Ask what special precautions should be taken.
Troubleshooting the high voltage plate supply is usually done under the following conditions:
• Exciter off
• Plate and screen IPA voltages off
• PA screen voltage off
Individual transmitters may require different procedures. Check with the manufacturer first.
Figure 8.3.1 shows a typical transmitter high voltage power supply. Begin the troubleshooting
process by breaking the circuit at point A. If the overload condition persists, the failure is caused
by a problem in the power supply itself, not in the PA compartment. If, on the other hand, the
overload condition disappears, a failure in the feedthrough capacitor (C1), decoupling capacitors
(C2, C3), or blocking capacitor (C4) is indicated.
If a problem is indicated in the PA compartment, reconnect the high voltage supply line at
point A and break the circuit at point B. A return of the overload problem would indicate a failure
in one of the decoupling capacitors or feedthrough capacitor.
In order to avoid unnecessary effort and time in troubleshooting, use the process of elimina-
tion to identify sections of the circuit to be examined. If, for example, the test at point A had indi-
cated the problem was not in the load, but in the power supply, a logical spot to perform the next
test would be at point C (for long high voltage cable runs). This test would identify or eliminate
the interconnecting cable as a cause of the fault condition. If the cable run from the high voltage
supply to the PA compartment is short and direct, point D might be the next logical point to
check. Breaking the connection at the input to the power supply filter allows the rectifiers and
interconnecting cables to be checked. Components protected by transient suppression devices (as
shown above L1) should be considered a part of the component they are designed to protect. If a
choke is removed from the circuit for testing, its protective device must also be removed. Failure
to remove both connections will usually result in damage to the protective device.
To avoid creating a new problem while trying to correct the original failure, break the circuit
in only one point at a time. Also, study the possible adverse effects of each planned step in the
process. Disconnecting certain components from the circuit may cause overvoltages or power
g q p
Figure 8.3.1 A typical transmitter three-phase high-voltage power supply circuit.
supply ripple that may damage other components in the transmitter. Consult the manufacturer to
be sure.
Perform any troubleshooting work on a transmitter with extreme care. Transmitter high volt-
ages can be lethal. Work inside the transmitter only after all ac power has been removed and after
all capacitors have been discharged using the grounding stick provided with the transmitter.
Remove primary power from the unit by tripping the appropriate power distribution circuit
breakers in the transmitter building. Do not rely on internal contractors or SCRs to remove all
dangerous ac. Do not defeat protective interlock circuits. Although defeating an access panel
interlock switch may save work time, the consequences can be tragic.
8.3.4 RF System Faults

Although RF troubleshooting may seem intimidating, there is no secret to it. Patient examination
of the circuit and careful study of the schematic diagram will go a long way toward locating the
problem. The first step in troubleshooting an RF problem is to determine whether the fault is RF-
based or dc-based.
8.3.4a Troubleshooting Procedure

Check the load by examining the transmitter overload indicators. Most transmitters monitor
reflected power from the antenna and will trip-off if excessive VSWR is detected. If the VSWR
g q p
Figure 8.3.2 The PA grid input circuit of a grounded-screen FM transmitter.
fault indicator is not lit, the load is not likely the cause of the problem. A definitive check of the
load can be made by switching the transmitter output to a dummy load and bringing up the high
voltage. The PA tube may be checked by substituting one of known quality. When the tube is
changed, carefully inspect the contact fingerstock for signs of overheating or arcing. Be careful
to protect the socket from damage when removing and inserting the PA tube. Do not change the
tube unless there is good reason to believe that it may be defective.
If problems with the PA stage persist, examine the grid circuit of the tube. Figure 8.3.2 shows
the input stage of a grounded screen, FM transmitter. A short circuit in any of the capacitors in
the grid circuit (C1–C5) will effectively ground the PA grid. This will cause a dramatic increase
in plate current, because the PA bias supply will be shorted to ground along with the RF signal
from the IPA stage.
The process of finding a defective capacitor in the grid circuit begins with a visual inspection
of the suspected components. Look for signs of discoloration because of overheating, loose con-
nections, and evidence of package rupture. The voltage and current levels found in a transmitter
PA stage are often sufficient to rupture a capacitor if an internal short circuit occurs. Check for
component overheating right after shutting the transmitter down. (As mentioned previously,
remove all ac power and discharge all capacitors first.) A defective capacitor will often overheat.
Such heating can also occur, however, because of improper tuning of the PA or IPA stage, or a
defective component elsewhere in the circuit.
Before replacing any components, study the transmitter schematic diagram to determine
which parts in the circuit could cause the failure condition that exists. By knowing how the trans-
mitter works, many hours can be saved in checking components that an examination of the fault
condition and the transmitter design would show to be an unlikely cause of the problem.
Check blocking capacitors C6 and C7. A breakdown in either component would have serious
consequences. The PA tube would be driven into full conduction, and could arc internally. The
working voltages of capacitors C1–C5 could also be exceeded, damaging one or more of the
components. Because most of the wiring in the grid circuit of a PA stage consists of wide metal
straps (required because of the skin effect), it is not possible to view stress points in the circuit to
narrow the scope of the troubleshooting work. Areas of the system that are interconnected using
components that have low power dissipation capabilities, however, should be closely examined.
g q p
For example, the grid bias decoupling components shown in Figure 8.3.2 (R1, L3, and C5)
include a low wattage (2 W) resistor and a small RF choke. Because of the limited power dissipa-
tion ability of these two devices, a failure in decoupling capacitor C5 would likely cause R1 and
possibly L3 to burn out. The failure of C5 in a short circuit would pull the PA grid to near ground
potential, causing the plate current to increase and trip off the transmitter high voltage. Depend-
ing on the sensitivity and speed of the plate overload sensor, L3 could be damaged or destroyed
by the increased current it would carry to C5, and therefore, to ground.
If L3 were able to survive the surge currents that resulted in PA plate overload, the choke
would continue to keep the plate supply off until C5 was replaced. Bias supply resistor R1, how-
ever, would likely burn out because the bias power supply is generally switched on with the
transmitter filament supply. Therefore, unless the PA bias power supply line fuse opened, R1
would overheat and probably fail.
Because of the close spacing of components in the input circuit of a PA stage, carefully check
for signs of arcing between components or sections of the tube socket. Keep all components and
the socket itself clean at all times. Inspect all interconnecting wiring for signs of damage, arcing
to ground, or loose connections.
8.3.4b Component Substitution

Substituting a new component for a suspected part can save valuable time when troubleshooting.
With some components, it is cost-effective to replace a group of parts that may include one
defective component because of the time involved in gaining access to the damaged device. For
example, the grid circuit of the PA stage shown in Figure 8.3.2 includes three doorknob capaci-
tors (C2–C4) formed into a single assembly. If one device was found to be defective, it might be
advantageous to simply replace all three capacitors. These types of components are often inte-
grated into a single unit that may be difficult to reach. Because doorknob capacitors are rela-
tively inexpensive, it would probably be best to replace the three as a group. This way, the entire
assembly is eliminated as a potential cause of the fault condition.
A good supply of spare parts is a valuable troubleshooting tool. In high-power transmitting
equipment, substitution is sometimes the only practical means of finding the cause of a problem.
The manufacturer's factory service department can usually recommend a minimum spare parts
inventory. Obvious candidates for inventory include components that are not available locally,
such as high voltage fixed-value capacitors, vacuum variable capacitors, and specialized semi-
conductors.
8.3.4c Inside the PA Cavity

One of the things that makes troubleshooting a cavity-type power amplifier difficult is the nature
of the major component elements. The capacitors do not necessarily look like capacitors, and the
inductors do not necessarily look like inductors. It is often difficult to relate the electrical sche-
matic diagram to the mechanical assembly that exists within the transmitter output stage. At
VHF and UHF frequencies—the domain of cavity PA designs—inductors and capacitors can
take on unusual mechanical forms.
Consider the PA cavity schematic diagram shown in Figure 8.3.3. The grounded-screen stage
is of conventional design. Decoupling of the high-voltage power supply is accomplished by C1,
C2, C3, and L1. Capacitor C3 is located inside the PA chimney (cavity inner conductor). The RF
g q p
Figure 8.3.3 An FM transmitter PA output stage built around a 1/4-wavelength cavity with capaci-
tive coupling to the load.
Figure 8.3.4 The equivalent electrical circuit of the PA stage shown in Figure 8.3.3.
sample lines provide two low-power RF outputs for a modulation monitor or other test instru-
ments. Neutralization inductors L3 and L4 consist of adjustable grounding bars on the screen
grid ring assembly. The combination of L2 and C6 prevents spurious oscillations within the cav-
ity.
Figure 8.3.4 shows the electrical equivalent of the PA cavity schematic diagram. The 1/4-
wavelength cavity acts as the resonant tank for the PA. Coarse tuning of the cavity is accom-
plished by adjustment of the shorting plane. Fine tuning is performed by the PA tuning control,
which acts as a variable capacitor to bring the cavity into resonance. The PA loading control con-
sists of a variable capacitor that matches the cavity to the load. There is one value of plate load-
ing that will yield optimum output power, efficiency, and PA tube dissipation. This value is
dictated by the cavity design and values of the various dc and RF voltages and currents supplied
to the stage.
g q p
Figure 8.3.5 The mechanical equivalent of the PA stage shown in Figure 8.3.4.
The logic of a PA stage often disappears when the maintenance engineer is confronted with
the actual physical design of the system. As illustrated in Figure 8.3.5, many of the components
take on an unfamiliar form. Blocking capacitor C4 is constructed of a roll of Kapton insulating
material sandwiched between two circular sections of aluminum. (Kapton is a registered trade-
mark of DuPont.) PA plate tuning control C5 consists of an aluminum plate of large surface area
that can be moved in or out of the cavity to reach resonance. PA loading control C7 is constructed
much the same as the PA tuning assembly, with a large-area paddle feeding the harmonic filter,
located external to the cavity. The loading paddle may be moved toward the PA tube or away
from it to achieve the required loading. The L2–C6 damper assembly actually consists of a 50 Ω
noninductive resistor mounted on the side of the cavity wall. Component L2 is formed by the
inductance of the connecting strap between the plate tuning paddle and the resistor. Component
C6 is the equivalent stray capacitance between the resistor and the surrounding cavity box.
From this example it can be seen that many of the troubleshooting techniques that work well
with low-frequency RF and dc do not necessarily apply in cavity stages. It is, therefore, critically
important to understand how the system operates and what each component does. Because many
of the cavity components (primarily inductors and capacitors) are mechanical elements more
than electrical ones, troubleshooting a cavity stage generally focuses on checking the mechanical
integrity of the box.
Most failures resulting from problems within a cavity are the result of poor mechanical con-
nections. All screws and connections must be kept tight. Every nut and bolt in a PA cavity was
included for a reason. There are no insignificant screws that do not need to be tight. However, do
not over tighten either. Stripped threads and broken component connection lugs will only cause
additional grief.
When a problem occurs in a PA cavity, it is usually difficult to determine which individual
element (neutralization inductor, plate tuning capacitor, loading capacitor, etc.) is defective from
the symptoms the failure will display. A fault within the cavity is usually a catastrophic event that
g q p
will take the transmitter off the air. It is often impossible to bring the transmitter up for even a
few seconds to assess the fault situation. The only way to get at the problem is to shut the trans-
mitter down and take a look inside.
Closely inspect every connection, using a trouble light and magnifying glass. Look for signs
of arcing or discoloration of components or metal connections. Check the mechanical integrity
of each element in the circuit. Be certain the tuning and loading adjustments are solid, without
excessive mechanical play. Look for signs of change in the cavity. Check areas of the cavity that
may not seem like vital parts of the output stage, such as the maintenance access door finger-
stock and screws. Any failure in the integrity of the cavity, whether at the base of the PA tube or
on part of the access door, will cause high circulating currents to develop and may prevent proper
operation of the stage. If a problem is found that involves damaged fingerstock, replace the
affected sections. Failure to do so will likely result in future problems because of the high cur-
rents that can develop at any discontinuity in the cavity inner or outer conductor.
8.3.4d VSWR Overload

A VSWR overload in transmission equipment can result from a number of different problems.
The following checklist presents some common problems and solutions:
1. VSWR overloads are usually caused by an improper impedance match external to the trans-
mitter. The first step in the troubleshooting procedure is to substitute a dummy load for the
entire antenna and transmission line system. Connect the dummy load at the transmitter
output port, thereby eliminating all coax, external filters and other RF hardware that might
be present in the system.
2. If the VSWR trip fault is eliminated in step 1, the problem lies somewhere in the transmis-
sion line or antenna. The dummy load can next be moved to the point at which the transmis-
sion line leaves the building and heads for the tower (if different than the point checked in
step 1). This test will check any RF plumbing, switches, or filter assemblies. If the VSWR
overload condition is still absent, the problem is centered in the transmission line or the
antenna.
3. If a standby antenna is not available, operation may still be possible at reduced power on a
temporary basis. For example, if arcing occurs in the antenna or line at full power, emer-
gency operation may be possible at half power. Inspect the antenna and line for signs of
trouble. Repair work beyond this point normally requires specialized equipment and a
tower crew. This discussion assumes that the problem is not caused by ice buildup on the
antenna, which can be alleviated by reducing the transmitter power output until VSWR
trips do not occur.
4. If step 1 shows the VSWR overload to be internal to the transmitter, determine whether the
problem is caused by an actual VSWR overload or by a failure in the VSWR control cir-
cuitry. Check for this condition by disabling the transmitter exciter and bringing up the high
voltage. Under these conditions, RF energy will not be generated. (It is assumed that the
transmitter has proper bias on all stages and is properly neutralized.) If a VSWR overload is
indicated, the problem is centered in the VSWR control circuitry and not the RF chain. Pos-
sible explanations for control circuitry failure include loose connections, dirty switch con-
tacts, dirty calibration potentiometers, poor PC board edge connector contacts, defective IC
amplifiers or logic gates, and intermittent electrolytic capacitors.
g q p
5. If step 4 shows the VSWR overload is real, and not the result of faulty control circuitry,
check all connections in the output and coupling sections of the final stage. Look for signs
of arcing or loose hardware, particularly on any movable tuning components. Inspect high-
voltage capacitors for signs of overheating, which might indicate failure; and check coils
for signs of dust build-up, which might cause a flash-over. In some transmitters, VSWR
overloads can be caused by improper final stage tuning or loading. Consult the equipment
instruction book for this possibility. Also, certain transmitters include glass-enclosed
spark-gap lightning protection devices (gas-gaps) that can be disconnected for testing.
6. If VSWR overload conditions resulting from problems external to the transmitter are expe-
rienced at an AM radio station, check the following items:
• Component dielectric breakdown. If a normal (near zero) reflected power reading is indicated
at the transmitter under carrier-only conditions, but VSWR overloads occur during modula-
tion, component dielectric breakdown may be the problem. A voltage breakdown could be
occurring within one of the capacitors or inductors at the antenna tuning unit (ATU) or pha-
sor. Check all components for signs of damage. Clean insulators as required. Carefully check
any open-air coils or transformers for dust buildup or loose connections.
• Narrowband antenna. If the overload occurs with any modulating frequency, the probable
cause of the fault is dielectric breakdown. If, on the other hand, the overload seems particu-
larly sensitive to high-frequency modulation, then narrow antenna bandwidth is indicated.
Note the action of the transmitter forward/reflected power meter. An upward deflection of
reflected power with modulation is a symptom of limited antenna bandwidth. The greater the
upward deflection, the more limited the bandwidth. If these symptoms are observed, conduct
an antenna impedance sweep of the system.
• Static buildup. Tower static buildup is characterized by a gradual increase in reflected power
as shown on the transmitter front panel. The static buildup, which usually occurs prior to or
during thunderstorms and other bad weather conditions, continues until the tower base ball
gaps arc-over and neutralize the charge. The reflected power reading then falls to zero. A
static drain choke at the tower base to ground will generally prevent this problem.
• Guy wire arc-over. Static buildup on guy wires is similar to a nearby lightning strike in that no
charge is registered during the buildup of potential on the reflected power meter. Instead, the
static charge builds on the guys until it is of sufficient potential to arc across the insulators to
the tower. The charge is then removed by the static drain choke and/or ball gaps at the base of
the tower. Static buildup on guy wires can be prevented by placing RF chokes across the insu-
lators, or by using non-metallic guys. Arcing across the insulators may also be reduced or
eliminated by regular cleaning.
8.3.5 Power Control Faults

A failure in the transmitter ac power control circuitry can result in problems ranging from zero
RF output to a fire inside the unit. Careful, logical troubleshooting of the control system is man-
datory. Two basic types of primary ac control are used in transmitters today:
• SCR (thyristor) system
g q p
Figure 8.3.6 Block diagram of a three-phase thyristor-control system.
• Relay logic
8.3.5a Thyristor Control System

A failure in the thyristor power control system of a transmitter is not easy to overlook. In the
worst case, no high voltage at all will be produced by the transmitter. In the best case, power con-
trol may be erratic or uneven when using the continuously variable power adjustment mode.
Understanding how the servo circuit works and how it is interconnected is the first step in cor-
recting such a problem.
Figure 8.3.6 shows a block diagram of a typical thyristor control circuit. Three gating cards
are used to drive back-to-back SCR pairs, which feed the high-voltage power transformer pri-
mary windings. Although the applied voltage is three phase, the thyristor power control configu-
ration simulates a single-phase design for each phase-to-phase leg. This allows implementation
of a control circuit that consists basically of a single-phase gating card duplicated three times
(one for each load phase). This approach has advantages from the standpoint of design simplic-
ity, and also from the standpoint of field troubleshooting. In essence, each power control circuit
is identical, allowing test voltages and waveforms from one gating card to be directly compared
with a gating card experiencing problems.
If the high-voltage supply will not come up at all, the problem involves more than a failure in
just one of the three gating cards. The failure of any one gating board would result in reduced
power output (and other side effects), but not in zero output. Begin the search with the interlock
system.
g q p
8.3.5b Interlock Failures

Modern transmitters provide the engineer with built-in diagnostic readouts on the status of the
transmitter interlock circuits. These may involve discrete LEDs or a microcomputer-driven
visual display of some type. If the transmitter has an advanced status display, the process of
locating an interlock fault is relatively simple. For an older transmitter that is not so equipped,
substantially more investigation will be needed.
Make a close observation of the status of all fuses, circuit breakers, transmitter cabinet doors,
and access panels. Confirm that all doors are fully closed and secured. Switch the transmitter
from remote to local control (if operated remotely) to eliminate the remote control system as a
possible cause of the problem. Observe the status of all control-panel indicator lamps. Some
transmitters include an interlocks open lamp; other units provide an indication of an open inter-
lock through the filament on or plate off push-button lamps. These indicators can save valuable
minutes or even hours of troubleshooting, so pay attention to them. Replace any burned-out indi-
cator lamps as soon as they are found to be defective.
If the front-panel indicators point to an interlock problem, pull out the schematic diagram of
the transmitter, get out the DMM, and shut down the transmitter. If the transmitter interlock cir-
cuit operates from a low-voltage power supply, such as 24 V dc, use a voltmeter to check for the
loss of continuity. Remove ac power from all sections of the transmitter except the low-voltage
supply by tripping the appropriate front-panel circuit breakers. Be extremely careful when work-
ing on the transmitter to avoid any line voltage ac. If the layout of the transmitter does not permit
safe troubleshooting with only the low-voltage power supply active, remove all ac from the unit
by tripping the wall-mounted main breaker. Then, use an ohmmeter to check for the loss of con-
tinuity.
If the transmitter interlock circuit operates from 120 V ac or 220 V ac, remove all power from
the transmitter by tripping the wall-mounted main breaker. Use an ohmmeter to locate the prob-
lem. Many older transmitters use line voltages in the interlock system. Do not try to troubleshoot
such transmitters with ac power applied.
Finding a problem such as an open control circuit interlock is basically a simple procedure, in
spite of the time involved. Do not rush through such work. When searching for a break in the
interlock system, use a methodical approach to solving the problem. Consider the circuit config-
uration shown in Figure 8.3.7. The most logical approach to finding a break in the control ladder
is to begin at the source of the 24 V dc input and, step by step, work toward the input of the power
controller. Although this approach may be logical, it can also be time consuming. Instead, elimi-
nate stages of the interlock circuit. For example, make the first test at terminal A. A correct volt-
age reading at this point in the circuit will confirm that all of the interlock door switches are
operating properly.
With the knowledge that the problem is after terminal A, move on to terminal B. If the 24 V
supply voltage disappears, check the fault circuit overload relays to find where the control signal
is lost. Often, such interlock problems can be attributed to dirty contacts in one of the overload
relays. If a problem is found with one set of relay contacts, clean all of the other contacts in the
overload interlock string for good measure. Be sure to use the proper relay-contact cleaning
tools. If sealed relays are used, do not attempt to clean them. Instead, replace the defective unit.
g q p
Figure 8.3.7 A typical transmitter interlock circuit. Terminals A and B are test points used for trou-
bleshooting the system in the event of an interlock system failure.
8.3.5c Step-Start Faults

The high-voltage power supply of any medium- or high-power transmitter must include provi-
sions for in-rush current-limiting upon the application of a plate-on command. The filter capaci-
tor(s) in the power supply will appear as a virtual short circuit during a sudden increase in
voltage from the rectifier stacks. To avoid excessive current surges through the rectifiers, capac-
itor(s), choke, and power transformer, nearly all transmitters use some form of step-start
arrangement. Such circuits are designed to limit the in-rush current to a predictable level. This
can be accomplished in a variety of ways.
For transmitters using a thyristor power control system, the step-start function can be easily
designed into the gate firing control circuit. An R-C network at the input point of the gating
cards is used to ramp the thyristor pairs from a zero conduction angle to full conduction (or a
conduction angle preset by the user). This system provides an elegant solution to the step-start
requirement, allowing plate voltage to be increased from zero to full value within a period of
approximately five seconds.
Transmitters employing a conventional ac power control system usually incorporate a step-
start circuit consisting of two sets of contactors: the start contactor and the run contactor, as
illustrated in Figure 8.3.8. Surge-limiting resistors provide sufficient voltage drop upon applica-
tion of a plate-on command to limit the surge current to a safe level. Auxiliary contacts on the
start contactor cause the run contactor to close as soon as the start contacts have seated.
A fault in the step-start circuit of a transmitter is often evidenced—initially at least—by ran-
dom tripping of the plate supply circuit breaker at high-voltage turn-on. If left uncorrected, this
condition can lead to problems such as failed power rectifiers or filter capacitors.
Troubleshooting a step-start fault in a system employing thyristor power control should begin
at the R-C ramp network. Check the capacitor to see if it has opened. Monitor the control voltage
to the thyristor gating cards to confirm that the output voltage of the controller slowly increases
g q p
Figure 8.3.8 A typical three-phase step-start power control system.
to full value. If it does and the turn-on problem persists, the failure likely involves one or more of
the gating cards.
When troubleshooting a step-start fault in a transmitter employing the dual contactor arrange-
ment, begin with a close inspection of all contact points on both contactors. Pay careful attention
to the auxiliary relay contacts of the start contactor. If the contacts fail to properly close, the full
load of the high-voltage power supply will be carried through the resistors and start contactor.
These devices are normally sized only for intermittent duty. They are not intended to carry the
full load current for any length of time. Look for signs of arcing or overheating of the contact
pairs and current-carrying connector bars. Check the current-limiting resistors for excessive dis-
sipation and continuity.
8.3.5d Protection Circuits

Many engineers enjoy a false sense of security with transmission equipment because of the pro-
tection devices included in most designs. Although conventional overload circuits provide pro-
tection against most common failure modes, they are not foolproof. The first line of defense in
the transmitter, the ac power system circuit breakers, can allow potentially disastrous currents to
flow under certain fault conditions.
Consider the thyristor ac power servo system shown in Figure 8.3.9. This common type of
voltage-regulator circuit adjusts the condition angle of the SCR pairs to achieve the desired dc
output from the high-voltage power supply. An alternative configuration could have the output
voltage sample derived from a transmission line RF pickup and amplifier/detector. In this way,
g q p
Figure 8.3.9 A common three-phase thyristor servo ac power control system.
the primary power control is adjusted to match the desired RF output from the transmitter. If one
of the high-voltage rectifier stacks of this system failed in a short circuit condition, the output
voltage (and RF output) would fall, causing the thyristor circuit to increase the conduction period
of the SCR pairs. Depending on the series resistance of the failed rectifier stack and the rating of
the primary side circuit breaker, the breaker may or may not trip. Remember that the circuit
breaker was chosen to allow operation at full transmitter power with the necessary headroom to
prevent random tripping. The primary power system, therefore, can dissipate a significant
amount of heat under reduced power conditions, such as those that would be experienced with a
drop in the high-voltage supply output. The difference between the maximum designed power
output of the supply (and, therefore, the transmitter) and the failure-induced power demand of the
system can be dissipated as heat without tripping the main breaker.
Operation under such fault conditions, even for 20 seconds or less, can cause considerable
damage to power-supply components, such as the power transformer, rectifier stack, thyristors,
or system wiring. Damage can range from additional component failures to a fire in the affected
section of the transmitter.
g q p
g g
Chapter
8.4
Testing Coaxial Transmission Line
8.4.1 Introduction
Antenna and transmission line performance measurements are often neglected until a problem
occurs. Many facilities do not have the equipment necessary to perform useful measurements.
Experience is essential, because much of the knowledge obtained from such tests is derived by
interpreting the raw data. In general, transmission systems measurements should be made:
• Before and during installation of the antenna and transmission line. Barring unforeseen oper-
ational problems, this will be the only time that the antenna is at ground level. Ready access
to the antenna allows a variety of key measurements to be performed without climbing the
tower.
• During system troubleshooting when attempting to locate a problem. Following installation,
these measurements usually concern the transmission line itself.
• To ensure that the transmission line is operating normally, many facilities check the transmis-
sion line and antenna system on a regular basis. A quick sweep of the line with a network ana-
lyzer and a time-domain reflectometer (TDR) may disclose developing problems before they
can cause a transmission line failure.
Ideally, the measurements should be used to confirm a good impedance match, which can be
interpreted as minimum VSWR or maximum return loss. Return loss is related to the level of
signal that is returned to the input connector after the signal has been applied to the transmission
line and reflected from the load. A line perfectly matched to the load would transfer all energy to
the load. No energy would be returned, resulting in an infinite return loss, or an ideal VSWR of
1:1. The benefits of matching the transmission line system for minimum VSWR include:
• Most efficient power transfer from the transmitter to the antenna system.
• Best performance with regard to overall bandwidth.
• Improved transmitter stability with tuning following accepted procedures more closely.
• Minimum transmitted signal distortions.
The network analyzer allows the maintenance engineer to perform a number of critical mea-
surements in a short period of time. The result is an antenna system that is tuned as close as prac-
8-57
g
Figure 8.4.1 Network analyzer plot of an FM broadcast antenna operating at 94.3 MHz.
tical for uniform impedance across the operating bandwidth. A well-matched system increases
operating efficiency by properly coupling the signal from the transmitter to the antenna. Figure
8.4.1 shows a network analyzer plot of an FM broadcast antenna.
8.4.2 Measuring VSWR

Historically, a slotted line device has been commonly used to measure VSWR on a transmission
line. A slotted line includes a probe that penetrates the outer conductor of the line through a slot.
The probe, in close proximity with the inner conductor, measures the voltage or samples the field
along the center conductor. The sample is detected, which results in a voltage proportional to the
actual signal on the center conductor. It is an accurate, reliable instrument. However, the slotted
line procedure takes a considerable amount of time to accurately sweep a transmission line over a
wide bandwidth, and then to plot the resulting data.
A network analyzer incorporating a return loss bridge performs antenna measurements more
quickly. The analog network analyzer consists of a sweep generator coupled to a tracking
receiver. A sample of the signal applied to the transmission line is compared to the return signal
through a return loss bridge or directional coupler. By adding a storage or normalizer device to
store the signal digitally, the instrument can provide a stable display while sweeping the line at
low speed to find all irregularities that may exist at discrete frequencies.
Digital designs do not use a sweep generator, but instead an integral synthesizer. In this way,
the return loss is measured at discrete frequencies. Software-calibration procedures correct each
measurement for system frequency and phase response errors, delay irregularities, and directiv-
ity errors in the return loss bridge or directional coupler. By calibrating a software-controlled
g
unit at the top of the transmission line, measurements will accurately show antenna characteris-
tics without effects of the transmission line. Results are plotted on an X-Y plotter or defined and
stored for later printout.
One particularly desirable feature of a network analyzer is its capability to display either a
Smith chart or a more simple Cartesian X-Y presentation of return loss-versus-frequency. (Some
units may provide both displays simultaneously.) The Smith chart is useful, but interpretation can
be confusing. The Cartesian presentation, while technically not better, is usually easier to inter-
pret.
8.4.2a Calibration
Calibration methods vary for different instruments. For one method, a short circuit is placed
across the network analyzer terminals, producing a return loss of zero (the short reflects all sig-
nals applied to it). The instrument is then checked with a known termination. This step often
causes the inexperienced technician to go astray. The termination should have known character-
istics and full documentation. It is acceptable procedure to check the equipment by examining
more than one termination, where the operator knows the characteristics of the devices used. Sig-
nificant changes from the known characteristics suggest that additional tests should be per-
formed. After the test unit is operating correctly, check to ensure that the adapters and connectors
to be used in the measurement do not introduce errors of their own. An accepted practice for this
task involves the use of a piece of transmission line of known quality. A 20-ft section of line
should sufficiently separate the input and output connectors. The results of any adjustment at
either end will be noticeable on the analyzer. Also, the length allows adjustments to be made
fairly easy. The section of line used should include tuning stubs or tuners to permit the connec-
tors to be matched to the line across the operating channel.
The facility's dummy load must next be matched to the transmission line. Do not assume that
the dummy load is an appropriate termination by itself, or a station reference. The primary func-
tion of a dummy load is to dissipate power in a manner that allows easy measurement. It is nei-
ther a calibration standard nor a reference. Experience proves it is necessary to match dummy
and transmission line sections to maintain a good reference. The load is matched by looking into
the transmission line at the patch panel (or other appropriate point). Measurements are then taken
at locations progressively closer to the transmitter, until the last measurement is made at the out-
put connection of the transmitter. After the dummy load is checked, it serves as a termination.
8.4.2b Antenna Measurements

An antenna should be properly tuned before it is placed into service. Any minor tuning adjust-
ments to the antenna should be made at its base, not by compensation at the input to the transmis-
sion line. Impedance adjustments are typically made with tuning rings on the center conductor or
with an impedance-matching section. Adjustments are performed while observing the return loss
on the network analyzer. Transmission line rings are less convenient, but less expensive than an
impedance-matching section. The rings can be used for matching short runs or the overall line
between the transmitting equipment and the antenna. Either tuning method may be used at the
antenna.
Both tuning systems operate by introducing a discontinuity into the transmission line. The
ring effectively changes the diameter of the center conductor, causing an impedance change at
g
that point on the line. This introduces a reflection into the line, the magnitude of which is a func-
tion of the size of the ring. The phase of the reflection is a function of the location of the ring
along the length of the center conductor.
Installing the ring is usually a cut-and-try process. It may be necessary to open, adjust, close,
and test the line several times. However, after a few cuts, the effect of the ring will become appar-
ent. It is not uncommon to need more than one ring on a given piece of transmission line for a
good match over the required bandwidth. When a match is obtained, the ring normally is sol-
dered into place.
Impedance-matching hardware also is available for use with waveguide. A piece of material is
placed into the waveguide and its location is adjusted to create the desired mismatch. For any
type of line, the goal is to create a mismatch equal in magnitude, but opposite in phase, to the
existing undesirable mismatch. The overall result is a minimum mismatch and minimum VSWR.
A tuner alters the line characteristic impedance at a given point by changing the distance
between the center and outer conductors by effectively moving the outer conductor. In reality, it
increases the capacity between the center and outer conductors to produce a change in the imped-
ance, and introduce a reflection at that point.
8.4.3 Testing Coaxial Lines

When dealing with transmission lines, a few electrical parameters have great influence on the
ability of the cable to transfer energy from one place to another. Among these are surge imped-
ance (ZO), the dielectric constants (k) of the insulating materials, and the velocity of propagation
(VP) of electromagnetic waves, both in free space and within the transmission lines.
In a coaxial cable, with a single center conductor and an outer conductor, the surge impedance
is determined by the relationship
ZO = 138
×
D
log 10 ---
--------
k d 8.4.1
Where
k = the dielectric constant of the insulating material
D = the inside diameter of the outer conductor
d = the outside diameter of the center conductor
In 2-conductor transmission lines, the surge impedance is determined by
ZO 276
×
2 S
log 10 -----
= --------
k s 8.4.2
Where:
k = the dielectric constant of the insulating material
S = the spacing between the centers of the two conductors
s = the diameter of the conductors
(Diameters and spacings can be measured in either centimeters or inches.)
g
If the inductance and capacitance per foot (or meter) of the transmission lines are given by the
manufacturer, the surge impedance can be calculated by
ZO = L
--- 8.4.3
C
Where:
L = quoted inductance per unit length
C = quoted capacitance per unit length
Note that the actual length of the line is not a factor in any of these formulas. The surge
impedance of the line is independent of cable length, and wholly dependant on cable type.
One common variety of 1/2-inch foam dielectric coaxial cable has an inductance per foot of
0.058 microhenries (μH) and a capacitance per foot of 23.1 picofarads (pF). Its surge impedance
is therefore
–6 – 12
ZO = ( 0.058 × 10 / 23.1 × 10 ) = 50 Ω 8.4.4
Most transmitters use 50 Ω or 52 Ω coaxial transmission lines to feed energy to an antenna.

Receivers traditionally employ 75 Ω or 300 Ω impedances.
8.4.3a Crimps and Mismatches

Because the surge impedance of a transmission line is based upon its cross-sectional characteris-
tics, its impedance can be changed if the cable is compressed or bent past its recommended bend
radius. The inner and outer conductors would no longer exhibit the same spacing between them
at that point, causing a change in impedance there (Equation 8.4.1.) This change of impedance
causes some of the energy in the line to be reflected in both directions. Some of the energy never
reaches the load. Standing waves are created, which in turn cause energy to be radiated by the
transmission line. (A perfectly matched transmitter, line, and load system has no standing waves,
no lost energy except for dielectric losses, and no radiation from the transmission line itself.)
A crimp in a line will cause an increase of capacitance at that point. This will cause the
impedance to be reduced. Mismatches can also occur between connectors and sections of trans-
mission line. Manufacturers will often specify what type of cable a connector is to be used with
instead of listing its surge impedance. Although this may be adequate for ordering connectors for
a given installation, consider the spare parts stock issue.
Multiple Mismatches
Unfortunately, if there is more than one discontinuity in the cable system, it becomes more diffi-
cult to determine the impedance of the second and subsequent mismatches.
Some TDR manufacturers have attempted to provide correction factors to account for multi-
ple mismatch errors, but they may not work for all situations. In general, the larger the first mis-
match, the greater the error in calculating the second mismatch. When the values of the
mismatched cable sections and/or loads are known, the technician can work backward and calcu-
g
late what the values should have been. When the cable values and/or loads are unknown, only the
first mismatch can be calculated accurately.
8.4.4 Bibliography
“Cable Testing with Time Domain Reflectometry,” Application Note 67, Hewlett Packard, Palo
Alto, Calif., 1988.
Kennedy, George: Electronic Communication Systems, 3rd ed., McGraw-Hill, New York, N.Y.,
1985.
Kolbert, Don: “Testing Coaxial Lines,” Broadcast Engineering, Intertec Publishing, Overland
Park, Kan., November 1991.
“Improving Time Domain Network Analysis Measurements,” Application Note 62, Hewlett
Packard, Palo Alto, Calif., 1988.
Strickland, James A.: “Time Domain Reflectometry Measurements, Measurement Concepts
Series, Tektronix, Beaverton, Ore., 1970.
“TDR Fundamentals,” Application Note 62, Hewlett Packard, Palo Alto, CA., 1988.
g g
Chapter
8.5
The Smith Chart
8.5.1 Introduction
The Smith chart is generally acknowledged to be the most universal tool for RF design work. RF
problems can be represented by complicated mathematics, such as hyperbolic or quadratic func-
tions, or by polynomial functions and differential equations. These equations involve complex
numbers, and must be solved for each specific frequency at each characteristic impedance, a
time-consuming and potentially error-prone process. The Smith chart simplifies these equations
into a chart format.
Several complications must be overcome to achieve a general solution for RF design prob-
lems. First is the variation of basic parameters caused by frequency; each frequency reacts differ-
ently to a given impedance. Second is the wide range of possible fundamental impedances. For
example, no two antennas of different design have exactly the same characteristic impedance,
although most fall into the range between 30 Ω and 120 Ω. The Smith chart provides a method to
“normalize” the frequency and characteristic impedance of an RF problem to permit solution on
a standardized chart.
Two quantities must be determined before using a Smith Chart: normalized impedance and
propagation constant. In addition, the expected or measured characteristic impedance of at least
one part of the RF circuit must be defined.
8.5.2 Normalized Impedance

Whenever an RF signal enters a transmission line or RF amplifier, there is a brief time during
which energy flows only forward into the line (or circuit.) When the RF wave has travelled to the
end or termination of the transmission line, any mismatch, no matter how small, will cause some
energy to be reflected. During the brief time when RF energy is only flowing forward into the cir-
cuit, the signal “sees” the characteristic impedance (ideal impedance) of the transmission line
circuit. For transmission lines, this number is specified by the manufacturer over a range of fre-
quencies. Four properties are used to define the characteristic impedance:
• The pure wire resistance of the transmission line per linear unit (R).
• The inductive reactance of the wire at the frequency of interest (L).
8-63
• The capacitive susceptance of the wire/shield combination at the frequency of interest (C).
• The conductance per unit length of the transmission line (G).
For applications where the length of the transmission line is short, the resistance of the wire is
negligible and so is the conductance. Even for long lengths of transmission line, such as antenna
feed lines, the resistance can be factored into the solution after the rest of the graphical work has
been done. Thus, the general equation for characteristic impedance (Z) of a typical transmission
line is
Z = L⁄ C (8.5.1)
Because the inductive reactance and capacitive susceptance are calculated at the frequency of
circuit operation, the frequency-related terms drop out, simplifying the process. For RF applica-
tions, the frequency-dependent values of inductance and capacitance of a transmission line result
from:
• Physical dimensions of the cable, such as diameter and thickness of the inner conductor.
• Distance between the inner conductor and the outer shield.
• Dielectric constant between the inner and outer conductors.
For waveguide, the physical dimensions are the width and breadth of a cross-section of the
guide expressed in fractions of a wavelength of the operating frequency. Note that these parame-
ters relate to operation at a single frequency. Modulated signals present further complications.
8.5.2a Propagation Constant

The propagation constant of a transmission line is defined as the natural logarithm of the ratio
between the input current and output current for the forward-travelling portion of the wave
(assuming that the length of the transmission line is exactly one wavelength at the frequency of
interest). The propagation constant is a complex number, with the real portion representing the
resistive attenuation of the transmission line, and the complex portion representing the phase
constant. The phase constant includes both the wavelength within the transmission line and the
velocity of propagation. The characteristic impedance and propagation constant of a transmis-
sion line are usually provided by the manufacturer. They represent information that defines the
operating efficiency of the transmission line in and of itself, disregarding the ultimate circuit
application.
While the characteristic impedance and propagation constant are given by the manufacturer,
and define the ideal performance of forward-travelling waves in a unit length of transmission
line, the normalized characteristic impedance provides the actual specifics that result from using
the transmission line in an RF application. By connecting a generator and a receiver to the trans-
mission line, several circuit mismatches are introduced. The generator does not have much
impact on the problem, as long as the output impedance is similar to the characteristic impedance
of the selected transmission line. The receiver, or load that is placed on the end of the transmis-
sion line, and the length of the transmission line in wavelengths, exert a much greater influence
on the circuit. For example, if the receiver is a totally open circuit, Kirchoff's Law can be applied
to determine that almost 100 percent of the energy injected into the transmission line will be
Figure 8.5.1 Resistance circles of the Smith chart coordinate system. (After [1].)
reflected. The Smith chart can be used to identify the best way to provide maximum energy
transfer from the transmission line into the load. In addition to transmission line problems, Smith
chart techniques apply to RF amplifiers, attenuators, directional couplers, antennas, antenna-
matching networks, and signal distribution networks.
8.5.3 Applying the Smith Chart

The Smith chart is built around two families of curves. The first is the resistance circle, as illus-
trated in Figure 8.5.1. Resistance circles are centered on the only straight line on the chart, the
vertical line (called the resistance axis). Notice that the concentric circles are given values start-
ing with zero at the top, and ending with infinity at the bottom. The circles represent ratios of
whatever value is assigned to the center point, called the prime center, which has a value of 1.0.
In operation, the prime center usually is set equal to the characteristic impedance of the transmis-
sion line, and the ratios of each circle serve as multipliers. This normalizes the chart for any
value assigned the prime center. All points along a resistance circle have the same resistance.
The second family of curves is the reactance curves, shown in Figure 8.5.2. Only segments of
these curves are plotted. As with the resistance family, these curves are labeled with normalizing
multipliers, with respect to prime center. Also, all points along any one curve have the same reac-
tance value. Combining the two curves yields a precursor to the Smith chart, illustrated in Figure
8.5.3.
Figure 8.5.2 Reactance curves of the Smith chart coordinate system. (After [1].)
Figure 8.5.3 Complete Smith chart coordinate system, with sample values plotted. (After [1].)
Figure 8.5.4 SWR circles drawn on a Smith chart. (After [1].)
One other family of useful circles, usually not printed on the chart, is plotted with a compass,
as needed. These are the standing wave ratio (SWR) circles. They occur as concentric circles
centered on prime center, as shown in Figure 8.5.4. In typical use, Smith charts are drawn with
the resistance component line horizontal. The SWR and attenuation scales are radial scales.
They are used by aligning a straightedge at right angles to the scale of interest, finding the
desired value, and tracing a line up the Smith chart. The bottom transmission coefficient scale is
a magnitude scale, which must be applied differently. A compass or divider is used to transfer
linear distances from portions of the Smith chart.
The Smith chart can be used to efficiently design a filter and/or impedance-matching net-
work. Each part of each problem can be solved using the chart, taking advantage of the graphical
nature of the tool to obtain possible solutions quickly. Because power components for RF work
come in limited size and value ranges, impedance-matching problems often must be reworked to
obtain final component values that can be realized with off-the-shelf parts. The flexibility of the
Smith chart in solving a complex problem is a great help in this respect.
When using a Smith Chart, perform the following steps (in sequence):
• Write down the pertinent actual resistance and reactance values.
• Choose a convenient denominator that brings the normalized resistance of the characteristic
impedance close to 1.0. Normalize all other values by dividing by this number. Note which
values are impedances for series circuits and which are admittances for shunt circuits.
• Plot the normalized values onto the Smith chart.
Figure 8.5.5 Example of Smith chart for lumped element match. (From [1]. Used with permission.)
• Resolve the problem by applying the rules of Smith chart.

• De-normalize circuit values by multiplying.
• Solve the reactance equations at the frequency of interest to obtain correct final component
values for the circuit.
Engineers sometimes ignore the Smith Chart as a design tool because they believe commer-
cial CAD programs are all they need. However, a quick look at these programs reveals that they
make extensive use of the Smith Chart. To use these programs effectively, an understanding of
the chart is needed.
The Smith Chart is particularly valuable in that phase of design when modification of the
impedance characteristics of the circuit is required. Even when purely numerical methods can be
used, the Smith Chart’s graphical representation can add an intuitive insight that saves consider-
able design time.
8.5.3a Application Example

As an example of the usefulness of the Smith chart, consider the case of lumped element match-
ing. A 75 + j 25 Ω load is to be matched to a 50 Ω transmission line. On a chart normalized to 50
Ω, the normalized impedance of the load is 1.5 + j 0.5, as plotted in Figure 8.5.5. The elements
required are determined by following the constant reactance lines in both the impedance and
admittance portions of the chart. From the load, we can follow the constant susceptance line to
the point that lies at 1 + j 0.8 (normalized) on the impedance chart. This route “travels” from val-
ues of –0.2 to –0.49 for a net change of 0.29 in the direction corresponding to a shunt inductor.
On the impedance chart, the path from 1 + j 0.8 to 1 + j 0.0 corresponds to a series capacitor with
a normalized impedance of – j 0.8. De-normalizing to a 50 Ω system, we see that a shunt induc-
tor of 0.29 × 50 = 14.5 Ω is placed across the load, followed by a series capacitor of 0.8 × 50 =
40 Ω to the 50 Ω line.
The Smith chart is a useful design and analysis tool that has withstood the test of time. This
chart, which is nothing more than a mapping of constant reactance and resistance of the imped-
ance plane onto the complex reflection coefficient plane, is one that every RF engineer should be
familiar with.
8.5.4 References
1. Silence, Neal C.: “The Smith Chart and its Usage in RF Design,” RF Design, Intertec Pub-
lishing, Overland Park, Kan., pp. 85–88, April 1992.
8.5.5 Bibliography
Adam, S. F.: Microwave Theory and Applications, Prentice-Hall, New York, N.Y., 1969.
Bowick, C.: RF Circuit Design, Howard W. Sams and Co., Indianapolis, IN, 1982.
Bryant, G. H.: Principles of Microwave Measurements, IEE Electrical Measurement Series,
Peter Peregrinus Ltd., London, 1988.
Colin, R. E.: Foundations for Microwave Engineering, McGraw-Hill, New York, N.Y., 1966.
Kaufhold, Gerry: “The Smith Chart, Parts 1–4,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., November 1989–-March 1990.
Montgomery, C. G., R. H. Dicke, and E. M. Purcell: Principles of Microwave Circuits,
Smith, P. H.: Electronic Applications of the Smith Chart, McGraw-Hill, New York, N.Y., 1969.
g g
Chapter
8.6
Standby Power Systems
8.6.1 Introduction
When utility company power problems are discussed, most people immediately think of black-
outs. The lights go out, and everything stops. With the facility down and in the dark, there is
nothing to do but sit and wait until the utility company finds the problem and corrects it. This
process generally takes only a few minutes. There are times, however, when it can take hours. In
some remote locations, it can even take days.
Blackouts are, without a doubt, the most troublesome utility company problem that a facility
will have to deal with. Statistics show that power failures are, generally speaking, a rare occur-
rence in most areas of the country. They are also short in duration. Studies have shown that 50
percent of blackouts last 6 s or less, and 35 percent are less than 11 min long. These failure rates
usually are not cause for concern to commercial users, except where computer-based operations,
transportation control systems, medical facilities, and communications sites are concerned.
When continuity of operation is critical, redundancy must be carried throughout the system.
The site never should depend upon one critical path for ac power. For example, if the facility is
fed by a single step-down transformer, a lightning flash or other catastrophic event could result in
a transformer failure that would bring down the entire site. A replacement could take days or
even weeks.
8.6.1a Blackout Effects

A facility that is down for even 5 min can suffer a significant loss of productivity that may take
hours or days to rebuild. A blackout affecting a transportation or medical center could be life-
threatening. Coupled with this threat is the possibility of extended power-service loss due to
severe storm conditions. Many broadcast and communications relay sites are located in remote,
rural areas or on mountaintops. Neither of these kinds of locations are well-known for their
power reliability. It is not uncommon in mountainous areas for utility company service to be out
for extended periods after a major storm. Few operators are willing to take such risks with their
business. Most choose to install standby power systems at appropriate points in the equipment
chain.
8-71
y y
Figure 8.6.1 The classic standby power system using an engine-generator set. This system pro-
tects a facility from prolonged utility company power failures.
The cost of standby power for a facility can be substantial, and an examination of the possible
alternatives should be conducted before any decision on equipment is made. Management must
clearly define the direct and indirect costs and weigh them appropriately. Include the following
items in the cost-vs.-risk analysis:
• Standby power-system equipment purchase and installation cost.
• Exposure of the system to utility company power failure.
• Alternative operating methods available to the facility.
• Direct and indirect costs of lost uptime because of blackout conditions.
A distinction must be made between emergency and standby power sources. Strictly speaking,
emergency systems supply circuits legally designated as being essential for safety to life and
property. Standby power systems are used to protect a facility against the loss of productivity
resulting from a utility company power outage.
8.6.2 Standby Power Options

To ensure the continuity of ac power, many commercial/industrial facilities depend upon either
two separate utility services or one utility service plus on-site generation. Because of the grow-
ing complexity of electrical systems, attention must be given to power-supply reliability.
The engine-generator shown in Figure 8.6.1 is the classic standby power system. An auto-
matic transfer switch monitors the ac voltage coming from the utility company line for power
failure conditions. Upon detection of an outage for a predetermined period of time (generally 1
to 10 s), the standby generator is started; after the generator is up to speed, the load is transferred
from the utility to the local generator. Upon return of the utility feed, the load is switched back,
y y
and the generator is stopped. This basic

type of system is used widely in indus-
try and provides economical protection
against prolonged power outages (5 min
or more).
The transfer device shown in Figure
8.6.1 is a contactor-type, break-before-
make unit. By replacing the simple
transfer device shown with an automatic
overlap (static) transfer switch, as
shown in Figure 8.6.2, additional func-
tionality can be gained. The overlap
transfer switch permits the on-site gen-
erator to be synchronized with the load,
making a clean switch from one energy Figure 8.6.2 The use of a static transfer switch to
source to another. This functionality transfer the load from the utility company to the on-
offers the following benefits: site generator.
• Switching back to the utility feed

from the generator can be accomplished without interruption in service.
• The load can be cleanly switched from the utility to the generator in anticipation of utility line
problems (such as an approaching severe storm).
• The load can be switched to and from the generator to accomplish load shedding objectives.
8.6.2a Dual Feeder System

In some areas, usually metropolitan centers, two utility company power drops can be brought
into a facility as a means of providing a source of standby power. As shown in Figure 8.6.3, two
separate utility service drops—from separate power-distribution systems—are brought into the
plant, and an automatic transfer switch changes the load to the backup line in the event of a
main-line failure. The dual feeder system provides an advantage over the auxiliary diesel
arrangement in that power transfer from main to standby can be made in a fraction of a second if
a static transfer switch is used. Time delays are involved in the diesel generator system that limit
its usefulness to power failures lasting more than several minutes.
The dual feeder system of protection is based on the assumption that each of the service drops
brought into the facility is routed via different paths. This being the case, the likelihood of a fail-
ure on both power lines simultaneously is remote. The dual feeder system will not, however, pro-
tect against area-wide power failures, which can occur from time to time.
The dual feeder system is limited primarily to urban areas. Rural or mountainous regions gen-
erally are not equipped for dual redundant utility company operation. Even in urban areas, the
cost of bringing a second power line into a facility can be high, particularly if special lines must
be installed for the feed. If two separate utility services are available at or near the site, redundant
feeds generally will be less expensive than engine-driven generators of equivalent capacity.
Figure 8.6.4 illustrates a dual feeder system that utilizes both utility inputs simultaneously at
the facility. Notice that during normal operation, both ac lines feed loads, and the “tie” circuit
breaker is open. In the event of a loss of either line, the circuit-breaker switches reconfigure the
y y
Figure 8.6.3 The dual utility feeder system of ac power loss protection. An automatic transfer
switch changes the load from the main utility line to the standby line in the event of a power inter-
ruption.
load to place the entire facility on the single

remaining ac feed. Switching is performed auto-
matically; manual control is provided in the event
of a planned shutdown on one of the lines.
8.6.2b Peak Power Shaving

Figure 8.6.5 illustrates the use of a backup diesel
generator for both standby power and peak power
shaving applications. Commercial power custom-
ers often can realize substantial savings on utility
company bills by reducing their energy demand
during certain hours of the day. An automatic
overlap transfer switch is used to change the load Figure 8.6.4 A dual utility feeder system
from the utility company system to the local die- with interlocked circuit breakers.
sel generator. The changeover is accomplished by
a static transfer switch that does not disturb the
operation of load equipment. This application of a standby generator can provide financial return
to the facility, whether or not the unit is ever needed to carry the load through a commercial
power failure.
8.6.2c Advanced System Protection

A more sophisticated power-control system is shown in Figure 8.6.6, where a dual feeder supply
is coupled with a motor-generator set to provide clean, undisturbed ac power to the load. The m-
y y
Figure 8.6.5 The use of a diesel generator for standby power and peak power shaving applica-
tions.
g set will smooth over the transition from the main utility feed to the standby, often making a
commercial power failure unnoticed by on-site personnel. A conventional m-g typically will give
up to 0.5 s of power fail ride-through, more than enough to accomplish a transfer from one utility
feed to the other. This standby power system is further refined in the application illustrated in
Figure 8.6.7, where a diesel generator has been added to the system. With the automatic overlap
transfer switch shown at the generator output, this arrangement also can be used for peak
demand power shaving.
Figure 8.6.8 shows a simplified schematic diagram of a 220 kW UPS system utilizing dual
utility company feed lines, a 750 kVA gas-engine generator, and five dc-driven motor-generator
sets with a 20-min battery supply at full load. The five m-g sets operate in parallel. Each is rated
for 100 kW output. Only three are needed to power the load, but four are on-line at any given
time. The fifth machine provides redundancy in the event of a failure or for scheduled mainte-
nance work. The batteries are always on-line under a slight charge across the 270 V dc bus. Two
separate natural-gas lines, buried along different land routes, supply the gas engine. Local gas
storage capacity also is provided.
8.6.2d Choosing a Generator

Engine-generator sets are available for power levels ranging from less than 1 kVA to several
thousand kVA or more. Machines also can be paralleled to provide greater capacity. Engine-gen-
erator sets typically are classified by the type of power plant used:
• Diesel. Advantages: rugged and dependable, low fuel costs, low fire and/or explosion hazard.
Disadvantages: somewhat more costly than other engines, heavier in smaller sizes.
y y
Figure 8.6.6 A dual feeder standby power system using a motor-generator set to provide power
fail ride-through and transient-disturbance protection. Switching circuits allow the m-g set to be
bypassed, if necessary.
Figure 8.6.7 A premium power-supply backup and conditioning system using dual utility feeds, a
diesel generator, and a motor-generator set.
y y
Figure 8.6.8 Simplified installation diagram of a high-reliability power system incorporating dual
utility feeds, a standby gas-engine generator, and five battery-backed dc m-g sets. (After [1].)
• Natural and liquefied petroleum gas. Advantages: quick starting after long shutdown periods,
long life, low maintenance. Disadvantage: availability of natural gas during area-wide power
failure subject to question.
• Gasoline. Advantages: rapid starting, low initial cost. Disadvantages: greater hazard associ-
ated with storing and handling gasoline, generally shorter mean time between overhaul.
• Gas turbine. Advantages: smaller and lighter than piston engines of comparable horsepower,
rooftop installations practical, rapid response to load changes. Disadvantages: longer time
required to start and reach operating speed, sensitive to high input air temperature.
The type of power plant chosen usually is determined primarily by the environment in which
the system will be operated and by the cost of ownership. For example, a standby generator
located in an urban area office complex may be best suited to the use of an engine powered by
natural gas, because of the problems inherent in storing large amounts of fuel. State or local
building codes can place expensive restrictions on fuel-storage tanks and make the use of a gaso-
line- or diesel-powered engine impractical. The use of propane usually is restricted to rural areas.
y y
Figure 8.6.9 Typical configuration of an engine-generator set. (From [2]. Used with permmission.)
The availability of propane during periods of bad weather (when most power failures occur) also
must be considered.
The generator rating for a standby power system should be chosen carefully and should take
into consideration the anticipated future growth of the plant. It is good practice to install a
standby power system rated for at least 25 percent greater output than the current peak facility
load. This headroom gives a margin of safety for the standby equipment and allows for future
expansion of the facility without overloading the system.
An engine-driven standby generator typically incorporates automatic starting controls, a bat-
tery charger, and automatic transfer switch. (See Figure 8.6.9.) Control circuits monitor the util-
ity supply and start the engine when there is a failure or a sustained voltage drop on the ac
supply. The switch transfers the load as soon as the generator reaches operating voltage and fre-
quency. Upon restoration of the utility supply, the switch returns the load and initiates engine
shutdown. The automatic transfer switch must meet demanding requirements, including:
• Carrying the full rated current continuously
• Withstanding fault currents without contact separation
• Handling high inrush currents
• Withstanding many interruptions at full load without damage
y y
The nature of most power outages requires a sophisticated monitoring system for the engine-
generator set. Most power failures occur during periods of bad weather. Most standby generators
are unattended. More often than not, the standby system will start, run, and shut down without
any human intervention or supervision. For reliable operation, the monitoring system must check
the status of the machine continually to ensure that all parameters are within normal limits.
Time-delay periods usually are provided by the controller that require an outage to last from 5 to
10 s before the generator is started and the load is transferred. This prevents false starts that
needlessly exercise the system. A time delay of 5 to 30 min usually is allowed between the resto-
ration of utility power and return of the load. This delay permits the utility ac lines to stabilize
before the load is reapplied.
The transfer of motor loads may require special consideration, depending upon the size and
type of motors used at a plant. If the residual voltage of the motor is out of phase with the power
source to which the motor is being transferred, serious damage can result to the motor. Excessive
current draw also may trip overcurrent protective devices. Motors above 50 hp with relatively
high load inertia in relation to torque requirements, such as flywheels and fans, may require spe-
cial controls. Restart time delays are a common solution.
Automatic starting and synchronizing controls are used for multiple-engine-generator instal-
lations. The output of two or three smaller units can be combined to feed the load. This capability
offers additional protection for the facility in the event of a failure in any one machine. As the
load at the facility increases, additional engine-generator systems can be installed on the standby
power bus.
Generator Types
Generators for standby power applications can be induction or synchronous machines. Most
engine-generator systems in use today are of the synchronous type because of the versatility, reli-
ability, and capability of operating independently that this approach provides [2]. Most modern
synchronous generators are of the revolving field alternator design. Essentially, this means that
the armature windings are held stationary and the field is rotated. Therefore, generated power
can be taken directly from the stationary armature windings. Revolving armature alternators are
less popular because the generated output power must be derived via slip rings and brushes.
The exact value of the ac voltage produced by a synchronous machine is controlled by vary-
ing the current in the dc field windings, while frequency is controlled by the speed of rotation.
Power output is controlled by the torque applied to the generator shaft by the driving engine. In
this manner, the synchronous generator offers precise control over the power it can produce.
Practically all modern synchronous generators use a brushless exciter. The exciter is a small
ac generator on the main shaft; the ac voltage produced is rectified by a 3-phase rotating rectifier
assembly also on the shaft. The dc voltage thus obtained is applied to the main generator field,
which is also on the main shaft. A voltage regulator is provided to control the exciter field cur-
rent, and in this manner, the field voltage can be precisely controlled, resulting in a stable output
voltage.
The frequency of the ac current produced is dependent on two factors: the number of poles
built into the machine, and the speed of rotation (rpm). Because the output frequency must nor-
mally be maintained within strict limits (60 Hz or 50 Hz), control of the generator speed is essen-
tial. This is accomplished by providing precise rpm control of the prime mover, which is
performed by a governor.
y y
There are many types of governors; however, for auxiliary power applications, the isochro-
nous governor is normally selected. The isochronous governor controls the speed of the engine
so that it remains constant from no-load to full load, assuring a constant ac power output fre-
quency from the generator. A modern system consists of two primary components: an electronic
speed control and an actuator that adjusts the speed of the engine. The electronic speed control
senses the speed of the machine and provides a feedback signal to the mechanical/hydraulic actu-
ator, which in turn positions the engine throttle or fuel control to maintain accurate engine rpm.
The National Electrical Code provides guidance for safe and proper installation of on-site
engine-generator systems. Local codes may vary and must be reviewed during early design
stages.
8.6.2e UPS Systems

An uninterruptible power system is an elegant solution to power outage concerns. The output of
the UPS inverter can be a sine wave or pseudo sine wave. When shopping for a UPS system, con-
sider the following:
• Power reserve capacity for future growth of the facility.
• Inverter current surge capability (if the system will be driving inductive loads, such as
motors).
• Output voltage and frequency stability over time and with varying loads.
• Required battery supply voltage and current. Battery costs vary greatly, depending upon the
type of units needed.
• Type of UPS system (forward-transfer type or reverse-transfer type) required by the particular
application. Some sensitive loads may not tolerate even brief interruptions of the ac power
source.
• Inverter efficiency at typical load levels. Some inverters have good efficiency ratings when
loaded at 90 percent of capacity, but poor efficiency when lightly loaded.
• Size and environmental requirements of the UPS system. High-power UPS equipment
requires a large amount of space for the inverter/control equipment and batteries. Battery
banks often require special ventilation and ambient temperature control.
8.6.2f Standby Power-System Noise

Noise produced by backup power systems can be a serious problem if not addressed properly.
Standby generators, motor-generator sets, and UPS systems produce noise that can disturb build-
ing occupants and irritate neighbors and/or landlords.
The noise associated with electrical generation usually is related to the drive mechanism,
most commonly an internal combustion engine. The amplitude of the noise produced is directly
related to the size of the engine-generator set. First consider whether noise reduction is a neces-
sity. Many building owners have elected to tolerate the noise produced by a standby power gener-
ator because its use is limited to emergency situations. During a crisis, when the normal source
of power is unavailable, most people will tolerate noise associated with a standby generator.
y y
If the decision is made that building occupants can live with the noise of the generator, care
must be taken in scheduling the required testing and exercising of the unit. Whether testing
occurs monthly or weekly, it should be done on a regular schedule.
If it has been determined that the noise should be controlled, or at least minimized, the easiest
way to achieve this objective is to physically separate the machine from occupied areas. This may
be easier said than done. Because engine noise is predominantly low-frequency in character,
walls and floor/ceiling construction used to contain the noise must be massive. Lightweight con-
struction, even though it may involve several layers of resiliently mounted drywall, is ineffective
in reducing low-frequency noise. Exhaust noise is a major component of engine noise but, fortu-
nately, it is easier to control. When selecting an engine-generator set, select the highest-quality
exhaust muffler available. Such units often are identified as “hospital-grade” mufflers.
Engine-generator sets also produce significant vibration. The machine should be mounted
securely to a slab-on-grade or an isolated basement floor, or it should be installed on vibration
isolation mounts. Such mounts usually are specified by the manufacturer.
Because a UPS system or motor-generator set is a source of continuous power, it must run
continuously. Noise must be adequately controlled. Physical separation is the easiest and most
effective method of shielding occupied areas from noise. Enclosure of UPS equipment usually is
required, but noise control is significantly easier than for an engine-generator because of the
lower noise levels involved. Nevertheless, the low-frequency 120 Hz fundamental of a UPS sys-
tem is difficult to contain adequately; massive constructions may be necessary. Vibration control
also is required for most UPS and m-g gear.
8.6.3 Batteries
Batteries are the lifeblood of most UPS systems. Important characteristics include the following:
• Charge capacity—how long the battery will operate the UPS
• Weight
• Charging characteristics
• Durability/ruggedness
Additional features that add to the utility of the battery include:
• Built-in status/temperature/charge indicator and/or data output port
• Built-in over-temperature/over-current protection with auto-reset capabilities
• Environmental friendliness
The last point deserves some attention. Many battery types must be recycled or disposed of
through some prescribed means. Proper disposal of a battery at the end of its useful life is, thus,
an important consideration. Be sure to check the original packaging for disposal instructions.
Failure to follow the proper procedures could have serious consequences.
Research has brought about a number of different battery chemistries, each offering distinct
advantages. Today’s most common and promising rechargeable chemistries include the follow-
ing:
y y
• Nickel cadmium (NiCd)—used for portable radios, cellular phones, video cameras, laptop
computers, and power tools. NiCds have good load characteristics, are economically priced,
and are simple to use.
• Lithium ion (Li-Ion)—now commonly available and typically used for video cameras and lap-
top computers. This battery promises to replace some NiCds for high energy-density applica-
tions.
• Sealed lead acid (SLA)—used for uninterruptible power systems, video cameras, and other
demanding applications where the energy-to-weight ratio is not critical and low battery cost is
desirable.
• Nickel metal hydride (NiMH)—used for cellular phones, video cameras, and laptop comput-
ers where high energy is of importance and cost is secondary.
• Lithium polymer (Li-Polymer)—this battery has the highest energy density and lowest self-
discharge of common battery types, but its load characteristics typically suit only low current
applications.
• Reusable alkaline—used for light duty applications. Because of its low self-discharge, this
battery is suitable for portable entertainment devices and other non-critical appliances that are
used occasionally.
No single battery offers all the answers; rather, each chemistry is based on a number of com-
promises.
A battery, of course, is only as good as its charger. Common attributes for the current genera-
tion of charging systems include quick-charge capability and automatic battery condition analy-
sis and subsequent intelligent charging.
8.6.3a Terms
The following terms are commonly used to specify and characterize batteries:
• Energy density. The storage capacity of a battery measured in watt-hours per kilogram (Wh/
kg).
• Cycle life. The typical number of charge-discharge cycles for a given battery before the
capacity decreases from the nominal 100 percent to approximately 80 percent, depending
upon the application.
• Fast-charge time. The time required to fully charge an empty battery.
• Self-discharge. The discharge rate when the battery is not in use.
• Cell voltage. The output voltage of the basic battery element. The cell voltage multiplied by
the number of cells provides the battery terminal voltage
• Load current. The maximum recommended current the battery can provide.
• Current rate. The C-rate is a unit by which charge and discharge times are scaled. If dis-
charged at 1C, a 100 Ah battery provides a current of 100 A; if discharged at 0.5C, the avail-
able current is 50 A.
y y
Figure 8.6.10 The charge states of an SLA battery. (From [3]. Used with permission.)
• Exercise requirement. This parameter indicates the frequency that the battery needs to be
exercised to achieve maximum service life.
8.6.3b Sealed Lead-Acid Battery

The lead-acid battery is a commonly used chemistry. The flooded version is found in automo-
biles and large UPS battery banks. Most smaller, portable systems use the sealed version, also
referred to as SLA or gelcell.
The lead-acid chemistry is commonly used when high power is required, weight is not critical,
and cost must be kept low [3]. The typical current range of a medium-sized SLA device is 2 Ah
to 50 Ah. Because of its minimal maintenance requirements and predictable storage characteris-
tics, the SLA has found wide acceptance in the UPS industry, especially for point-of-application
systems.
The SLA is not subject to memory. No harm is done by leaving the battery on float charge for
a prolonged time. On the negative side, the SLA does not lend itself well to fast charging. Typical
charge times are 8 to 16 hours. The SLA must always be stored in a charged state because a dis-
charged SLA will sulphate. If left discharged, a recharge may be difficult or even impossible.
Unlike the common NiCd, the SLA prefers a shallow discharge. A full discharge reduces the
number of times the battery can be recharged, similar to a mechanical device that wears down
when placed under stress. In fact, each discharge-charge cycle reduces (slightly) the storage
capacity of the battery. This wear-down characteristic also applies to other chemistries, including
the NiMH.
The charge algorithm of the SLA differs from that of other batteries in that a voltage-limit
rather than current-limit is used. Typically, a multi-stage charger applies three charge stages con-
sisting of a constant-current charge, topping-charge, and float-charge. (See Figure 8.6.10.) Dur-
ing the constant-current stage, the battery charges to 70 percent in about five hours; the
remaining 30 percent is completed by the topping-charge. The slow topping-charge, lasting
another five hours, is essential for the performance of the battery. If not provided, the SLA even-
tually loses the ability to accept a full charge and the storage capacity of the battery is reduced.
y y
The third stage is the float-charge that compensates for self-discharge after the battery has been
fully charged.
During the “constant current charge,” the SLA battery is charged at a high current, limited by
the charger itself. After the voltage limit is reached, the topping charge begins and the current
starts to gradually decrease. Full-charge is reached when the current drops to a preset level or
reaches a low-end plateau.
The proper setting of the cell voltage limit is critical and is related to the conditions under
which the battery is charged. A typical voltage limit range is from 2.30 V to 2.45 V. If a slow
charge is acceptable, or if the room temperature can exceed 30°C (86°F), the recommended volt-
age limit is 2.35 V/cell. If a faster charge is required and the room temperature remains below
30°C, 2.40 or 2.45 V/cell can be used.
8.6.4 References
1. Lawrie, Robert: Electrical Systems for Computer Installations, McGraw-Hill, New York,
N.Y., 1988.
2. DeDad, John A.: “Auxiliary Power,” in Practical Guide to Power Distribution for Informa-
tion Technology Equipment, PRIMEDIA Intertec, Overland Park, Kan., pp. 31–39, 1997.
3. Buchmann, Isidor: “Batteries,” in The Electronics Handbook, Jerry C. Whitaker (ed.), pg.
1058, CRC Press, Boca Raton, Fla., 1996.
8.6.5 Bibliography
Angevine, Eric: “Controlling Generator and UPS Noise,” Broadcast Engineering, PRIMEDIA
Intertec, Overland Park, Kan., March 1989.
Baietto, Ron: “How to Calculate the Proper Size of UPS Devices,” Microservice Management,
PRIMEDIA Intertec, Overland Park, Kan., March 1989.
Federal Information Processing Standards Publication No. 94, Guideline on Electrical Power for
ington, D.C., 1983.
Highnote, Ronnie L.: The IFM Handbook of Practical Energy Management, Institute for Man-
agement, Old Saybrook, Conn., 1979.
Smith, Morgan: “Planning for Standby AC Power,” Broadcast Engineering, PRIMEDIA Intertec,
Overland Park, Kan., March 1989.
Stuart, Bud: “Maintaining an Antenna Ground System,” Broadcast Engineering, PRIMEDIA
Intertec, Overland Park, Kan., October 1986.
g g
Section
Test Equipment
9
The audio/video business in one in which there is a constant flow of new products. This rapid
advancement of technology requires an equally rapid advancement in training technical person-
nel. Many large manufacturers offer entry-level maintenance training on their products, but these
programs may be restricted to authorized resellers of the product line. Hands-on training is
unquestionably the best way to learn how to service a given piece of equipment. When a student
is allowed to practice a new technique with the supervision of a skilled instructor, the highest
level of learning and retention occurs.
Modern computer-based test instruments provide the ability to rapidly transmit data from one
location to another. Instruments, such as oscilloscopes and spectrum analyzers, are available that
can output a waveform or other data to a modem for transmission to a central service facility for
analysis. Instead of a single person grappling with a difficult problem, the on-site technician can
call on the resources and experience of the service center. Field service, thus, becomes a team
effort that includes the technician in the field and the often more experienced service center
engineers and staff. Software programs are available to permit test equipment at a remote loca-
tion to be configured as required, capture data, and transmit that data to the service center for
analysis. Teleservicing also makes possible the creation of reference libraries of key waveforms
and data patterns. These data facilitate troubleshooting, and are useful in documenting the per-
formance characteristics of the equipment being maintained. Over time, such documentation can
become a valuable addition to the service record of a given piece of equipment.
Service sites are sometimes in less than ideal locations, such as a mountaintop or a remote
microwave relay site. Teleservicing in such cases provides numerous benefits. If the problem
involves intermittent failures, the test instrument can be set up in a “babysitting mode” and left to
capture a critical signal when it occurs.
In This Section:
Chapter 9.1: Troubleshooting Digital Systems 9-9

Introduction 9-9
Defining Key Terms 9-9
Elements of a Computer 9-10
Approach to Troubleshooting 9-12
9-1
q p
9-2 Section Nine
Servicing Obsolete Hardware 9-13

Shipping Electronic Hardware 9-14
Isolating the Defective Subassembly 9-15
Power Supply Failures 9-15
Cable Considerations 9-17
Bibliography 9-17
Chapter 9.2: Digital Test Instruments 9-19

Introduction 9-19
Digital Multimeters 9-19
Voltage Conversion 9-21
Conventional Test Instruments 9-23
Frequency Counter 9-23
Multi-Function Hand-held Instrument 9-23
Logic Instruments 9-23
Logic Current Tracer 9-24
Logic Test Clip 9-24
Logic Analyzer 9-25
Signature Analyzer 9-28
Manual Probe Diagnosis 9-29
Emulative Tester 9-29
Protocol Analyzer 9-31
Automated Test Instruments 9-31
Software Considerations 9-32
Applications 9-32
Bibliography 9-32
Chapter 9.3: Oscilloscopes 9-35

Introduction 9-35
Specifications 9-35
Bandwidth 9-35
Rise Time 9-36
Making Measurements 9-36
Advanced Features 9-37
Digital Oscilloscope 9-37
Operating Principles 9-38
A/D Conversion 9-39
DSO Architecture 9-40
DSO Features 9-42
Capturing Transient Waveforms 9-42
Triggering 9-43
Advanced Features 9-43
Bibliography 9-44
Chapter 9.4: Spectrum Analysis 9-45

Introduction 9-45
Principles of Operation 9-45
q p
Test Equipment 9-3
Applications 9-47
On-Air Measurements 9-48
Spurious Harmonic Distortion 9-49
Selective-Tuned Filter Alignment 9-50
Small-Signal Troubleshooting 9-52
Defining Terms 9-52
References 9-57
Bibliography 9-58
Chapter 9.5: Reliability Engineering 9-59

Introduction 9-59
Objectives 9-59
Terminology 9-60
Quality Assurance 9-61
Inspection Process 9-62
Reliability Evaluation 9-62
Failure Analysis 9-63
Standardization 9-63
Reliability Analysis 9-65
Statistical Reliability 9-65
Roller-Coaster Hazard Rate 9-67
Environmental Stress Screening 9-68
Latent Defects 9-71
Field Experience 9-73
U. S. Navy Research 9-73
Collins Avionics Research 9-73
General Electric Research 9-74
Hughes Aircraft Research 9-74
Parts Count Design Tradeoffs 9-75
Operating Environment 9-75
Maintenance Actions 9-76
Maintenance Procedures 9-76
Impact of New Technologies 9-77
Reliability-Centered Maintenance 9-77
Maintenance Classifications 9-78
Software Reliability 9-79
Hardware Versus Software 9-79
Causal Analysis 9-81
Failure Mode, Effect, and Criticality Analysis (FMECA) 9-81
Improving Software Reliability 9-81
Software Maintainability 9-82
Software System Safety 9-82
References 9-83
Bibliography 9-84

Albright, John R.: “Waveform Analysis with Professional-Grade Oscilloscopes,” Electronic Ser-
vicing and Technology, Intertec Publishing, Overland Park, Kan., April 1989.
q p
9-4 Section Nine
Allen, Tom: “Components of Microprocessor-Based Systems,” Electronic Servicing and Tech-

nology, Intertec Publishing, Overland Park, Kan., January 1988.
Allen, Tom: “Troubleshooting Microprocessor-Based Circuits: Part 1,” Electronic Servicing and
Technology, Intertec Publishing, Overland Park, Kan., January 1988.
Technology, Intertec Publishing, Overland Park, Kan., February 1988.
Babel, Philip S.: “Software Development Integrity Program.” Briefing paper for the Aeronauti-
cal Systems Division, Air Force Systems Command. From Yates, W., and Shaller, D.:
“Reliability Engineering as Applied to Software,” Proceedings IEEE Reliability and Main-
tainability Symposium, IEEE, New York, N.Y., 1990.
Brauer, D. C., and G. D. Brauer: “Reliability-Centered Maintenance,” Proceedings IEEE Reli-
ability and Maintainability Symposium, IEEE, New York, N.Y., 1987.
Breya, Marge: “New Scopes Make Faster Measurements,” Mobile Radio Technology Magazine,
Intertec Publishing, Overland Park, Kan., November 1988.
Bychowski, Phil: “Reverse Engineering,” Microservice Management, Intertec Publishing, Over-
land Park, Kan., March 1989.
Cameron, D., and R. Walker: “Run-In Strategy for Electronic Assemblies,” Proceedings IEEE
Reliability and Maintainability Symposium, IEEE, New York, N.Y., 1986.
Capitano, J., and J. Feinstein: “Environmental Stress Screening Demonstrates its Value in the
Field,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New York,
N.Y., 1986.
Carey, Gregory D.: “Automated Test Instruments,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., November 1989.
Carey, Gregory D.: “Isolating Microprocessor-Related Problems,” Electronic Servicing and
Technology, Intertec Publishing, Overland Park, Kan., June 1988.
Clark, Richard J.: “Electronic Packaging and Interconnect Technology Working Group Report
(IDA/OSD R&M Study),” IDA Record Document D-39, August 1983.
Clodfelter, Jim: “Troubleshooting the Microprocessor,” Microservice Management, Intertec
Publishing, Overland Park, Kan., November 1985.
Crop, Roland: “A Hierarchical Approach to Servicing,” Microservice Management, Intertec
Publishing, Overland Park, Kan., October 1988.
DesPlas, Edward: “Reliability in the Manufacturing Cycle,” Proceedings IEEE Reliability and
Maintainability Symposium, IEEE, New York, N.Y., 1986.
Detwiler, William L.: “Troubleshooting with a Digital Multimeter,' Mobile Radio Technology,
Intertec Publishing, Overland Park, Kan., September 1988.
Devaney, John: “Piece Parts ESS in Lieu of Destructive Physical Analysis,” Proceedings IEEE
Doyle, Edgar, Jr.: “How Parts Fail,” IEEE Spectrum, IEEE, New York, N.Y., October 1981.
q p
Test Equipment 9-5
Ferrara, K. C., S. J. Keene, and C. Lane: “Software Reliability from a System Perspective,” Pro-
ceedings IEEE Reliability and Maintainability Symposium, IEEE, New York, N.Y., 1988.
Fink, D., and D. Christiansen (eds.): Electronics Engineers' Handbook, 3rd ed., McGraw-Hill,
Fortna, H., R. Zavada, and T. Warren: “An Integrated Analytic Approach for Reliability Improve-
ment,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New York,
N.Y., 1990.
Gore, George: “Choosing a Hand-Held Test Instrument,” Electronic Servicing and Technology,
Intertec Publishing, Overland Park, Kan., December 1988.
Greenberg, Bob: “Repairing Microprocessor-Based Equipment,” Sound and Video Contractor,
Intertec Publishing, Overland Park, Kan., February 1988.
Griffin, P.: “Analysis of the F/A-18 Hornet Flight Control Computer Field Mean Time Between
Failure,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New York,
N.Y., 1985.
Hall, F., R. A. Paul, and W. E. Snow: “R&M Engineering for Off-the-Shelf Critical Software,”
Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New York, N.Y.,
1988.
Hansen, M. D., and R. L. Watts: “Software System Safety and Reliability,” Proceedings IEEE
Harju, Rey: “Hands-Free Operation: A New Generation of DMMs,” Microservice Management,
Intertec Publishing, Overland Park, Kan., August 1988.
Harris, Brad: “The Digital Storage Oscilloscope: Providing the Competitive Edge,” Electronic
Servicing and Technology, Intertec Publishing, Overland Park, Kan., June 1988.
Harris, Brad: “Understanding DSO Accuracy and Measurement Performance,” Electronic Ser-
Hermason, Sue E., Major, USAF: letter dated December 2, 1988. From Yates, W., and Shaller,
D.: “Reliability Engineering as Applied to Software,” Proceedings IEEE Reliability and
Hobbs, Gregg K.: “Development of Stress Screens,” Proceedings IEEE Reliability and Main-
Horn, R., and F. Hall: “Maintenance Centered Reliability,” Proceedings IEEE Reliability and
Hoyer, Mike: “Bandwidth and Rise Time: Two Keys to Selecting the Right Oscilloscope,” Elec-
tronic Servicing and Technology, Intertec Publishing, Overland Park, Kan., April 1990.
Irland, Edwin A.: “Assuring Quality and Reliability of Complex Electronic Systems: Hardware
and Software,” Proceedings of the IEEE, vol. 76, no. 1, IEEE, New York, N.Y., January
1988.
Kenett, R., and M. Pollak: “A Semi-Parametric Approach to Testing for Reliability Growth, with
Application to Software Systems,” IEEE Transactions on Reliability, IEEE, New York,
N.Y., August 1986.
q p
9-6 Section Nine
Kinley, Harold: “Using Service Monitor/Spectrum Analyzer Combos,” Mobile Radio Technol-
ogy, Intertec Publishing, Overland Park, Kan., July 1987.
Maynard, Eqbert, OUSDRE Working Group Chairman: “VHSIC Technology Working Group
Report” (IDA/OSD R&M Study), Document D-42, Institute of Defense Analysis, Novem-
ber 1983.
Montgomery, Steve: “Advanced in Digital Oscilloscopes,” Broadcast Engineering, Intertec Pub-
lishing, Overland Park, Kan., November 1989.
Neubauer, R. E., and W. C. Laird: “Impact of New Technology on Repair,” Proceedings IEEE
Pepple, Carl: “How to Use a Spectrum Analyzer at the Cell Site,” Cellular Business, Intertec
Publishing, Overland Park, Kan., Marcy 1989.
Persson, Conrad: “Oscilloscope Special Report,” Electronic Servicing and Technology, Intertec
Persson, Conrad: “Oscilloscope: The Eyes of the Technician,” Electronic Servicing and Technol-
ogy, Intertec Publishing, Overland Park, Kan., April 1987.
Persson, Conrad: “Test Equipment for Personal Computers,” Electronic Servicing and Technol-
Persson, Conrad:, “The New Breed of Test Instruments,” Broadcast Engineering, Intertec Pub-
Powell, Richard: “Temperature Cycling vs. Steady-State Burn-In,” Circuits Manufacturing, Ben-
will Publishing, September 1976.
Ogden, Leonard: “Choosing the Best DMM for Computer Service,” Microservice Management,
Robinson, D., and S. Sauve: “Analysis of Failed Parts on Naval Avionic Systems,” Report No.
D180-22840-1, Boeing Company, Seattle, Wash., October 1977.
Siner, T. Ann: “Guided Probe Diagnosis: Affordable Automation,” Microservice Management,
Smeltzer, Dennis: “Packing and Shipping Equipment Properly,” Microservice Management,
Smith, A., R. Vasudevan, R. Matteson, and J. Gaertner: “Enhancing Plant Preventative Mainte-
nance via RCM,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE,
Smith, William B.: “Integrated Product and Process Design to Achieve High Reliability in Both
Early and Useful Life of the Product,” Proceedings IEEE Reliability and Maintainability
Symposium, IEEE, New York, N.Y., 1987.
Sokol, Frank: “Specialized Test Equipment,” Microservice Management, Intertec Publishing,
Overland Park, Kan., August 1989.
Spradlin, B. C.: “Reliability Growth Measurement Applied to ESS,” Proceedings IEEE Reliabil-
ity and Maintainability Symposium, IEEE, New York, N.Y., 1986.
q p
Test Equipment 9-7
Toorens, Hans: “Oscilloscopes: From Looking Glass to High-Tech,” Electronic Servicing and
Technology, Intertec Publishing, Overland Park, Kan., April 1990.
Tustin, Wayne: “Recipe for Reliability: Shake and Bake,” IEEE Spectrum, IEEE, New York,
N.Y., December 1986.
Ware, Peter: “Servicing Obsolete Equipment,” Microservice Management, Intertec Publishing,
Whitaker, Jerry C.: Electronic Systems Maintenance Handbook, CRC Press, Boca Raton, Fla.,
2001.
Wickstead, Mike: “Signature Analyzers,” Microservice Management, Intertec Publishing, Over-
land Park, Kan., October, 1985.
Wilson, M. F., and M. L. Woodruff: “Economic Benefits of Parts Quality Knowledge,” Proceed-
ings IEEE Reliability and Maintainability Symposium, IEEE, New York, N.Y., 1985.
Wolf, Richard J.: “Spectrum Analyzer Uses for Two-Way Technicians,” Mobile Radio Technol-
Wong, Kam L.: “Demonstrating Reliability and Reliability Growth with Environmental Stress
Screening Data,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE,
Wong, K. L., I. Quart, L. Kallis, and A. H. Burkhard: “Culprits Causing Avionics Equipment
Failures,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New York,
N.Y., 1987.
Worm, Charles M.: “The Real World: A Maintainer's View,” Proceedings IEEE Reliability and
Yates, W., and D. Shaller: “Reliability Engineering as Applied to Software,” Proceedings IEEE
q p
g g
Chapter
9.1
Troubleshooting Digital Systems
9.1.1 Introduction
Repair of computer-based hardware requires a different approach to maintenance than conven-
tional analog circuits. Successful troubleshooting depends on four key elements:
• Understanding how the system operates
• Using the right test equipment
• Performing troubleshooting steps in the right sequence
• Paying attention to what the unit-under-test (UUT) is telling you
9.1.1a Defining Key Terms

Before proceeding to discuss troubleshooting techniques, it is appropriate to review the follow-
ing basic terms.
basic input/output system (BIOS) Software instructions contained in ROM that configure and
initialize a computer system.
direct memory access (DMA) A method of gaining direct access to the computer system mem-
ory for data transfer without involving the microprocessor directly.
dynamic random-access memory (DRAM) A type of memory device that is faster than static
memory and uses less power. Dynamic memory must, however, be refreshed periodically to
maintain its data.
emulation A technique using software and hardware in which one device is made to behave like
another device.
interrupt A signal that suspends the normal execution of a program and forces the microproces-
sor to begin executing a special interrupt handler routine. Interrupts are used to signal key-
strokes, critical inputs, or other information to which the microprocessor must respond
immediately.
kernel The core of a computer system: the microprocessor, clock, and bus buffers.
9-9
g g y
9-10 Test Equipment
Figure 9.1.1 Basic elements of a computer system.
machine cycle The shortest amount of time necessary for a microprocessor to complete an oper-
ation or a process.
memory map The assignment of various devices to portions of the microprocessor memory
space.
9.1.2 Elements of a Computer

A wide variety of computers are available to both consumers and industrial users. Although each
manufacturer has its own approach to system design, the basic concepts vary little from one unit
to another. Figure 9.1.1 shows a block diagram of a typical computer architecture. The major ele-
ments of the system include:
• Microprocessor
• Clock generator
• Read-only memory
• Random-access memory
• Data/address buses and decoders
• DMA controller
• Input/output interface
The microprocessor is the central control element of the system. Everything that happens in
the circuit is initiated by the microprocessor. The microprocessor executes one operation at a
g g y
time, usually consisting of data read, data write, or arithmetic functions. In response to instruc-
tions contained in the control program, the microprocessor orders data to be placed on the data
bus, reads it, performs calculations as required, and writes the resulting data to specific circuit
elements over the same bus.
The system clock determines the rate at which the microprocessor executes instructions. All
data movement within the kernel is synchronized with the master clock. A key function of the
clock is to make sure that data placed on the bus has time to stabilize before a read or write oper-
ation occurs. The clock pulses define valid data windows for the various buses. The clock also
provides refresh signals for the DRAM chips.
ROM contains built-in instructions for the microprocessor. These instructions are specific to
a given type of device, known as the instruction set. All microprocessor-based systems require
ROM instructions to get them started when the system is first powered up (booting).
RAM consists of one or more banks of memory devices that serve as a temporary storage area
for data. Data contained in RAM is volatile; it will be lost when power is removed from the sys-
tem. Program instructions read from the storage media are stored in and executed from RAM.
Buses are communications paths that conduct information from one place to another. Three
types of buses are fundamental to computer systems:
• Data bus—carries information to and from the microprocessor, connecting it with every part
of the system that handles data. The number of data bus lines is equal to the number of bits
the microprocessor can handle at one time; each bit requires its own line. Data bus lines are
bidirectional.
• Address bus—carries information that identifies the location of a particular piece of data.
Each device in the system has a specific range of addresses unique to that device. The data
and address buses work together to respond to the read and write commands of the micropro-
cessor.
• Control and status bus—the control and status lines of the microprocessor are connected to
the other kernel devices to effect control of system operations. These lines allow the micro-
processor to specify whether it wants to read data from a particular device, or write data to it.
They also provide a means for the device to notify the microprocessor when data is available
for transmission.
Both the data and address buses employ buffers located at the outer boundary of the kernel to
isolate the buses from other circuits on the system board.
Address decoders on the address bus notify each remote device when the address placed on
the bus by the microprocessor is within the address range of that particular device. This notifica-
tion is performed by turning the chip select (CS) pin of the device on or off. A device can
respond to requests from the microprocessor only when the CS pin is enabled. This prevents the
wrong device from responding to commands from the microprocessor.
The direct memory access controller permits the movement of data to and from memory
without going through the microprocessor. The DMA controller is designed to perform data
transfers involving peripheral devices, such as mass storage disks. Like other devices in the com-
puter system, the DMA controller operates only in response to instructions from the micropro-
cessor. When executing instructions, the DMA circuit assumes control over the data, address,
and control lines from the microprocessor. This process is know as cycle stealing. For one or
more clock cycles, the DMA chip takes over the communications buses, and the microprocessor
restricts itself to internal functions (such as arithmetic calculations). Because the microprocessor
g g y
9-12 Test Equipment
will not relinquish control over the buses for more than a few cycles at a time, the DMA control-
ler and microprocessor may pass control of the buses back and forth several times before the
DMA controller completes a given task.
I/O circuitry allows the microprocessor kernel to communication with peripheral devices,
such as the keyboard, monitor, storage media, and communications ports. Unlike the components
of the microprocessor kernel, I/O devices are not always governed by the microprocessor clock.
Peripherals may operate asynchronously with the system clock. The I/O circuitry typically con-
tains buffers that serve to isolate the kernel from the peripherals, and function as holding points
for data on its way to or from the kernel. For data going to the kernel, the buffers hold the data
until the microprocessor is ready to accept it. For data that the microprocessor is sending to a
peripheral device, the buffers hold the data until the device is ready to accept it.
9.1.3 Approach to Troubleshooting

The failure modes that may be experienced with a computer system are widely varied. Some
general rules, however, may be applied to nearly all troubleshooting efforts. First, follow the nat-
ural signal flow as it moves from the microprocessor through the system. Begin with the kernel
components and move outward toward the peripheral devices. When a failure node is found,
trace backward through the circuit until a point is located where the input is good, but the output
is bad.
This approach may be used to locate failures to the board level or to the component level.
Board swapping (or board float) is a commonly used repair technique today. Computer systems
are designed in a modular fashion to facilitate easy service and upgrade capability. The high cost
of skilled technical labor has led many service organizations to avoid component-level trouble-
shooting unless absolutely necessary. Such detailed work is left to service depots that have
sophisticated diagnostic instruments. Still, the cost of board float is significant. It represents
money sitting unproductively in the form of unused printed wiring boards (PWBs). The techni-
cian may also replace more than one board in an effort to locate the problem, and then simply
return all boards replaced to the service depot for analysis. Studies by various organizations have
found that as many as 35 percent of all boards returned for rework have been found to be in
proper working order. Also, there is no guarantee that the replacement board, unless handled
properly, will perform correctly when installed. The more a PWB is handled and shuffled in
inventory, the more likely it may become damaged before being placed in a system. The use of
board swapping versus repair to the component level is a function of many factors, including:
• Complexity of the hardware. As the level of complexity increases, the likeliness of cost-effec-
tive component-level repair decreases.
• Technical documentation available. Service is always made more difficult by a lack of
detailed documentation.
• Experience level of maintenance personnel. Solving complex problems usually requires
highly-skilled technicians.
• Construction of the hardware. A system built in a modular fashion is a natural candidate for
board swapping. A PWB using socket-mounted devices is a candidate for component-level
repair.
g g y
• Types of devices used. An PWB populated with DIP integrated circuits is much easier to
repair than one populated with surface-mounted ICs.
• Nature of the malfunction. Some failures can be diagnosed with little more than a DMM and
a logic probe. Other failures require complex and expensive test instruments.
• Mean time to repair (MTTR) considerations. Hardware used in critical applications where
downtime cannot be tolerated is best repaired using the board swapping technique.
Because of the complexity of computer equipment today, many service departments have
established a hierarchical approach to troubleshooting. This technique involves assigning one of
three levels to each service problem:
• First-tier problems—obvious failures that usually can be solved without extensive trouble-
shooting or expensive test equipment. First-tier problems are handled by technology general-
ists, often less experienced or entry-level personnel. In a sizable maintenance organization,
this group is the largest of the three. It handles the majority of service work.
• Second-tier problems—failures that are harder to diagnose than first-tier faults. Test equip-
ment is required to troubleshoot the system. Complex problems often call for sophisticated
instruments.
• Third-tier problems—the most difficult to troubleshoot failures. Complex and expensive test
equipment is required, as well as extensive experience on the unit being serviced. In a large
maintenance organization, this is the smallest of the three groups.
Such a tiered approach to service results in the most cost-effective use of technical talent. It
also offers fast turn-around of products to the customer.
9.1.3a Servicing Obsolete Hardware

Maintaining equipment that is no longer manufactured or supported by the original equipment
manufacturer (OEM) is a growing problem in computer service. Just because a piece of equip-
ment is “obsolete” does not mean that it is worthless. The typical marketing life cycle for com-
puter-based hardware lasts from 1 to 2 yrs. The equipment will, however, continue to provide
useful service many years after it is considered obsolete by the OEM. A product usually experi-
ences four major stages in its useful lifetime, as illustrated in Figure 9.1.2.
The decision to keep or replace a given piece of hardware is based primarily on economic fac-
tors. New equipment requires an initial outlay of capital. The central question is whether it will
have lower life-cycle costs than the older equipment that it would replace. When all factors are
considered, the conclusion is often made that the old equipment should be retained, at least until
maintenance becomes prohibitively expensive.
The acquisition of replacement parts in sufficient quantities, and at reasonable prices, is the
major challenge in maintaining after-market hardware. Some add-in cards are generic and can be
acquired from one of several sources. Parts for obsolete equipment may also be obtained by
demanufacturing (cannibalizing) other units. Buying whole units for demanufacture may be a
good solution because assembled equipment is almost always less expensive than the sum of the
field replaceable units (FRUs).
Perhaps the most difficult aspect of servicing obsolete equipment is the lack of technical doc-
umentation. Maintenance and service manuals may no longer be available from the OEM. The
g g y
9-14 Test Equipment
Figure 9.1.2 The four major cycles of product lifetime.
service technician, therefore, may have to practice some degree of reverse engineering to repair
the system. By definition, reverse engineering means working backward from a finished product
to develop schematics, parts lists, operating standards, and test procedures. Even if this data is
unavailable from the OEM, it may be available from other sources, including:
• The equipment owner. Large contract sales often specify that the buyer receives technical
documentation as part of the purchase.
• Parts suppliers to the OEM. A wealth of technical information is available from semiconduc-
tor manufacturers. By ordering application notes from the maker of the microprocessor chip,
for example, insight may be gained into how the system operates. Most newer computer-
based products are built around chip sets, and the chip set manufacturer may be able to supply
extensive documentation on how the products operate.
• Documentation from similar systems. Product lines tend to share many design concepts. If the
maintenance manual for a given piece of hardware is not available, but documentation for an
earlier or later system from the same manufacturer is available, sufficient information may be
gleaned from the available data to solve the problem at hand.
9.1.3b Shipping Electronic Hardware

It is inevitable that some circuit boards and other computer components will fail during their use-
ful lifetime. Although inoperative, these parts represent a significant investment for the user.
Proper shipping is an important element in the maintenance chain. PWBs must be protected from
bouncing, jostling, and static discharge. Careless handling during shipping can magnify what
may originally have been a minor problem. In some cases, the board may be completely
destroyed by improper shipping techniques. Protecting sensitive parts from rough handling is
g g y
easily accomplished with the proper packing materials, such as bubble wrap and anti-static foam
sheets. Furthermore, all electronic assemblies should be shipped in anti-static bags to prevent
ESD damage.
9.1.4 Isolating the Defective Subassembly

Troubleshooting computer hardware is not as difficult as most technicians might think. With the
proper technical documentation, a well-built system should not present significant repair prob-
lems. Virtually all computer systems conduct a power-up self-test (POST) routine when power is
applied. This routine checks the operational condition of each major component of the computer,
and initializes system controllers. The initialization and self-test functions of the POST process
are tightly interwoven. Functions can be divided into two basic categories:
• Checking and initializing central system hardware. A failure during this process may result in
a fatal error that will prevent the system from booting. The error may be identified either by a
message on the display, or by a series of beeps from the internal speaker (if the video display
cannot be initialized). If the system detects a fatal error, the ROM monitor will typically place
the processor into a hard loop with the error code in a specified register, and all interrupts dis-
abled. The effect is a processor halt without going into a true halt state. Under this condition,
no further checks will be made on the motherboard.
• Checking and initializing peripheral hardware on the I/O expansion bus. A failure during this
process usually does not result in a fatal error, unless a defective expansion PWB holds a data
or address line high or low. The booting process can also be halted by an interrupt or DMA
request signal from a peripheral device that will not clear when the acknowledge signal is
transmitted by the CPU.
Central hardware is checked first during the POST. The system then tests peripheral devices.
The POST routine varies from one type of system to another, however, computers built with a
particular type of microprocessor tend to initialize the same.
Any troubleshooting job is made easier by fully documenting a working system. By under-
standing how the system is supposed to work, repair is simplified.
9.1.4a Power Supply Failures

Power supply failures are usually obvious, and catastrophic to operation of the system. Most
switching power supplies used in computer equipment today incorporate protective circuitry to
shut down the system if an over-voltage or over-current condition is detected. To prevent damage
to the computer system board, remove the supply from the chassis before beginning work. The
supply usually must be loaded before troubleshooting. Some supplies will not start-up unless
loaded; others may not operate within normal specifications if no load is applied. Check the
technical documentation for the supply and connect a suitable resistive load. Most specification
sheets list minimum and maximum current levels for the various outputs. If a minimum spec is
not listed, a load is usually not required. For those specs that list a minimum load, connect a resis-
tor of the proper value and wattage. When loaded to specifications, voltages should fall within
about 0.2 V of the correct level.
g g y
9-16 Test Equipment
Figure 9.1.3 Block diagram of a switching power supply.
If intermittent disruptions are noted in the computer system, check for noise or ripple on the
power supply output pins. Noise can often enter the processor and related devices through the
supply, causing erratic operation. Less than 0.1 V ripple should be observed on an oscilloscope.
Check the reference manual for a ripple specification. 60 Hz ac line ripple may occur from a
defective (open or marginal) filter capacitor or a shorted filter choke. High frequency noise may
also be observed as a result of insufficient filtering of the switching stage.
Most computer equipment requires +5 V, +12 V, and –12 V. A simplified block diagram of a
switching supply is shown in Figure 9.1.3. Major elements of the example system include:
• Input and energy storage—provides input line conditioning (RFI filtering and current limit-
ing) and ac-to-dc rectification and filtering.
• Startup and reference—supplies unregulated dc to the startup circuit, supplies a +5 V refer-
ence for output voltage control, and generates trigger signals to the drive pulse generator cir-
cuit (which triggers the power supply into operation).
• Control and drive—manages power pulse generation during normal operation and during cur-
rent-limiting conditions, and directs power control pulses to the drive transformer and into the
primary inverter circuit.
• Primary inverter—provides drive signals for the power switching transistors and supplies cur-
rent-level status information for the shut-down circuit.
g g y
• Protection—detects and limits high current on the primary winding of the primary inverter
transformer, directs shutdown in the event of an overvoltage condition on the +5 V output,
and directs shutdown in the event of excessive temperature within the supply chassis.
• Secondary output—provides rectification, filtering, and regulation of the required operating
supply voltages.
When troubleshooting a catastrophic power supply failure, first check the primary line fuse.
Consult the technical documentation for the location of internal, PWB-mounted fuses, if used.
Check rectifiers and input-side filter capacitors. Both are vulnerable to failure from transient dis-
turbances on the ac line. Check the switching transistors, which are subjected to large switching
currents during normal operation. Check the startup circuit for the presence of control pulses.
Because of the availability of inexpensive stock power supplies, replacement of a defective
unit is typically the best option.
9.1.4b Cable Considerations

Cabling is a frequent source of failures for computer-based equipment. Connectors are subject to
rough handling and abuse that may result in bent or broken pins, or open or shorted internal con-
nections. Many times a defective cable is at the root of what appears to be a computer system
problem. Use of excessive lengths of cable can also result in marginal performance. Capacitive
loading of high frequency pulse trains can reduce the amplitude of data signals, and distort the
waveform shape.
In addition to capacitive loading, crosstalk becomes a problem with long cable runs. Because
of the low current levels used, the resistance of the cable is typically a minor concern. Long cable
runs also expose the signal-carrying lines to interference from radio frequency or 60 Hz ac line
energy. The likelihood of electromagnetic interference (EMI) problems increase as the cable
length increases. The use of shielded cable will reduce the effects of EMI.
9.1.5 Bibliography
Allen, Tom: “Components of Microprocessor-Based Systems,” Electronic Servicing and Tech-
nology, Intertec Publishing, Overland Park, Kan., January 1988.
Technology, Intertec Publishing, Overland Park, Kan., January 1988.
Technology, Intertec Publishing, Overland Park, Kan., February 1988.
Bychowski, Phil: “Reverse Engineering,” Microservice Management, Intertec Publishing, Over-
land Park, Kan., March 1989.
Carey, Gregory D.: “Isolating Microprocessor-Related Problems,” Electronic Servicing and
Technology, Intertec Publishing, Overland Park, Kan., June 1988.
Clodfelter, Jim: “Troubleshooting the Microprocessor,” Microservice Management, Intertec
Publishing, Overland Park, Kan., November 1985.
g g y
9-18 Test Equipment
Crop, Roland: “A Hierarchical Approach to Servicing,” Microservice Management, Intertec

Publishing, Overland Park, Kan., October 1988.
Greenberg, Bob: “Repairing Microprocessor-Based Equipment,” Sound and Video Contractor,
Intertec Publishing, Overland Park, Kan., February 1988.
Smeltzer, Dennis: “Packing and Shipping Equipment Properly,” Microservice Management,
Ware, Peter: “Servicing Obsolete Equipment,” Microservice Management, Intertec Publishing,
Whitaker, Jerry C.: Electronic Systems Maintenance, CRC Press, Boca Raton, Fla., 1991.
g g
Chapter
9.2
Digital Test Instruments
9.2.1 Introduction
As the equipment used by consumers and industry becomes more complex, the requirements for
highly-skilled maintenance technicians also increases. Maintenance personnel today require
advanced test equipment and must think in a “systems mode” to troubleshoot much of the hard-
ware now in the field. New technologies and changing economic conditions have reshaped the
way maintenance professionals view their jobs. As technology drives equipment design forward,
maintenance difficulties will continue to increase. Such problems can be met only through
improved test equipment and increased technician training.
Servicing computer-based professional equipment typically involves isolating the problem to
the board level and then replacing the defective PWB. Taken on a case-by-case basis, this
approach seems efficient. The inefficiency in the approach, however (which is readily apparent),
is the investment required to keep a stock of spare boards on hand. Furthermore, because of the
complex interrelation of circuits today, a PWB that appeared to be faulty may actually turn out to
be perfect. The ideal solution is to troubleshoot down to the component level and replace the
faulty device instead of swapping boards. In many cases, this approach requires sophisticated
and expensive test equipment. In other cases, however, simple test instruments will do the job.
Although the cost of most professional equipment has been going up in recent years, mainte-
nance technicians have seen a buyer's market in test instruments. The semiconductor revolution
has done more than given consumers low-cost computers and disposable devices. It has also
helped to spawn a broad variety of inexpensive test instruments with impressive measurement
capabilities.
9.2.2 Digital Multimeters

Beyond the basic voltage-resistance-current measurements, sophisticated digital multimeters
(DMMs) provide a number of unique functions. Large-scale integration has made it possible to
pack more features into the limited space inside a DMM, and to do so without significantly
increasing the cost of the instrument. In addition, because of competition among manufacturers,
the functionality of DMMs is continually improving. The result: increasingly sophisticated prod-
ucts.
9-19
g
9-20 Test Equipment
Digital multimeters can be divided into two basic classifications:

• Portable. Hand-held instruments intended for field applications offer extensive capabilities
in a portable, rugged package. Some hand-held meters include true-RMS measurements, sub-
millisecond response times, capacitance and frequency modes, and recording capabilities.
• Benchtop. Instruments intended for fixed applications usually offer a wide range of features
and interface to other instruments. Options include display devices with multiple readouts,
and built-in communications capabilities. A communications interface permits the meter to
be controlled over standard computer ports for data storage, analysis, and remote operation.
Evolutionary developments in DMM design have provided a number of helpful features to
maintenance technicians, including:
• Audible continuity check. The instrument emits a beep when the measured resistance value is
less than 5 Ω or so. This permits point-to-point continuity tests to be performed without actu-
ally looking at the DMM.
• Audible high voltage warning. The instruments emits a characteristic beep when the mea-
sured voltage is dangerously high (100 V or more). This safety feature alerts the technician to
use extra care when making measurements.
• Sample-and-hold. The instrument measures and stores the sampled value after it has stabi-
lized. Completion of the sample function is signaled by an audible beep. This feature permits
the technician to concentrate on probe placement, not on trying to read the instrument.
• Continuously variable tone. The instrument produces a tone that varies in frequency with the
measured parameter. This feature permits a circuit to be adjusted for a peak or null value
without actually viewing the DMM. It also may be used to check for intermittent continuity.
• Ruggedized case. Durable cases for portable instruments permit normal physical abuse to be
tolerated. Many DMMs in use today may be dropped or exposed to water without failure.
• Input overload hardening. Protection devices are placed at the input circuitry to prevent dam-
age in the event of an accidental overvoltage. Fuses and semiconductor-based protection com-
ponents are used.
• Logic probe functions. Signal high/low and pulse detection circuitry built into a DMM probe
permits the instrument to function as both a conventional multimeter and a logic probe. This
eliminates the necessity to change instruments in the middle of a troubleshooting sequence.
• Automatic shutoff. A built-in timer removes power to the DMM after a preset period of non-
operation. This feature extends the life of the internal batteries.
DMM functions are further expanded by the wide variety of probes and accessories available
for many instruments. Specialized probes include:
• High-voltage—used for measuring voltages above 500 V. Many probes permit voltages of 50
kV or more to be measured.
• Clamp-type current—permits current in a conductor to be measured without breaking the cir-
cuit. Probes are commonly available in current ranges of 1–2 A and up.
• Demodulator— converts a radio frequency signal into a dc voltage for display on the DMM.
g
Figure 9.2.1 Functional block diagram of a basic DMM. LSI chip technology has reduced such
systems to a single device.
• Temperature—permits the DMM to be used for display of temperature of a probe device.

Probes are available for surface measurement of solids and for immersion into liquid or gas.
Advances in DMM technology have resulted in increased reliability. DMMs are engineered to
include a minimum of moving parts. Auto-ranging eliminates much of the mechanical switch
contacts and movement common with older technology instruments. This approach reduces the
number of potential sources of failure, such as corrosion and intermittent switch contacts. Gold
plating of the remaining contacts in the DMM further reduces the chance of failure.
9.2.2a Voltage Conversion

Figure 9.2.1 shows a generalized block diagram of a DMM. Conversion of an input voltage to a
digital equivalent can be accomplished in one of several ways. The dual-slope conversion
method (also know as double-integration) is one of the more popular techniques (Figure 9.2.2).
Vin represents the input voltage (the voltage to be measured). VREF is a reference voltage with a
polarity opposite that of the measured voltage, supplied by the digital conversion circuit. Capaci-
tor C1 and operational amplifier U1 constitute an integrator. S1 is an electronic switch that is ini-
tially in the position shown. When the meter probes are first connected to the circuit, the sample
voltage is applied to the input of the integrator for a specified period of time, called the integra-
tion period. The integration period is usually related to the 60 Hz line frequency; integration
periods of 1/60th of a second and 1/10th of a second are common. The output signal of the inte-
grator is a voltage determined by the RC time constant of R1 and C1. Because of the nature of
the integrator, the maximum voltage (the voltage at the end of the integration period) is propor-
tional to the voltage being measured. At the end of the integration period, switch S1 is moved to
the other position (VREF) and a voltage of opposite polarity to the measured voltage is applied.
g
9-22 Test Equipment
(a)
(b)
Figure 9.2.2 DMM analog-to-digital conversion process: (a) functional block diagram, (b) graph of
Vc vs. time.
The capacitor is then discharged to zero. As shown in the figure, the discharge interval is directly
proportional to the maximum voltage, which in turn is proportional to the applied voltage.
At the same time that the integration interval ends and the discharge interval begins, a counter
in the meter begins counting pulses generated by a clock circuit. When the voltage reaches zero,
the counter stops. The number of pulses counted is, therefore, proportional to the discharge
period. This count is converted to a digital number and displayed as the measured voltage.
Although this method works well, it is somewhat slow, so many microcomputer-based meters use
a variation called multislope integration.
When a DMM is used in the resistance testing mode, it places a low voltage with a constant-
current characteristic across the test leads. After the leads are connected, the voltage across the
measurement points is determined, which provides the resistance value (voltage divided by the
known constant current = resistance). The meter converts the voltage reading into an equivalent
resistance for display. The voltage placed across the circuit under test is kept low to protect semi-
conductor devices that might be in the loop. The voltage at the ohmmeter probe is about 0.1 V or
less, too low to turn on silicon junctions.
The ultimate performance of a DMM in the field is determined by its inherent accuracy. Key
specifications include:
• Accuracy—how closely the meter indicates the actual value of a measured signal, specified
in percent of error. Zero percent indicates a perfect meter.
• Frequency response—the range of frequencies that can be measured by the meter without
exceeding a specified amount of error.
g
• Input impedance—the combined ac and dc resistance at the input terminals of the multime-
ter. The impedance, if too low, can load critical circuits and result in measurement errors. An
input impedance of 10 MΩ or greater will prevent loading.
• Precision—the degree to which a measurement is carried out. As a general rule, the more
digits a DMM can display, the more precise the measurement will be.
9.2.3 Conventional Test Instruments

Troubleshooting a digital system requires a wide range of test instruments, from simple to com-
plex. Some of the more common conventional test instruments include frequency counters and
logic instruments of various types.
9.2.3a Frequency Counter

A frequency counter provides an accurate measure of signal cycles or pulses over a standard
period of time. It is used to totalize or to measure frequency, period, frequency ratio, and time
intervals. Key specifications for a counter include:
• Frequency range—the maximum frequency that the counter can resolve.
• Resolution—the smallest increment of change that can be displayed. The degree of resolution
is selectable on many counters. Higher resolution usually requires a longer acquisition time.
• Sensitivity—the lowest amplitude signal that the instrument will count (measured in fractions
of a volt).
• Timebase accuracy—a measure of the stability of the timebase. Stability is measured in parts
per million (ppm) while the instrument is subjected to temperature and operating voltage vari-
ations.
9.2.3b Multi-Function Hand-held Instrument

Although specialized instruments such as a separate frequency counter or capacitance meter may
be the best choice for a specific range of tasks, combinations of instruments are available for the
convenience of maintenance personnel. Separate analyzers usually offer a broader range of mea-
surement capabilities, but a combined instrument often makes field servicing easier.
Logic Instruments
The simplest of all logic instruments is the logic probe. A basic logic probe tells the technician
whether the logic state of the point being checked is high, low, or pulsed. Most probes include a
pulse stretcher that will show the presence of a 1-shot pulse. The indicators are usually LEDs. In
some cases, a single LED is used to indicate any of the conditions (high, low, or pulsing); gener-
ally, individual LEDs are used to indicate each state. Probes usually include a switch to match the
level sense circuitry of the unit to the type of logic being checked (TTL, CMOS, ECL, etc.). The
probe receives its power from the circuit under test. Connecting the probe in this manner sets the
approximate value of signal voltage that constitutes a logic low or high. For example, if the
g
9-24 Test Equipment
Figure 9.2.3 Use of a logic pulser and current tracer to locate a short-circuit on a PWB.
power supply voltage for a CMOS logic circuit is 18 V, a logic low would be about 30 percent of
18 V (5.4 V), and a logic high would be about 70 percent of 18 V (12.6 V).
The logic pulser is the active counterpart to the logic probe. The logic pulser generates a sin-
gle pulse, or train of pulses into a circuit to check the response.
Logic Current Tracer

The close spacing of pins and conductors on a crowded PWB invites short circuits. On a large
board, locating a short can be a difficult and time-consuming task. The logic current tracer can
be a valuable tool in troubleshooting such failures.
Even in an extremely thin printed circuit conductor, a change in current induces a magnetic
field. The probe of a logic current tracer is designed to sense changes in the minute magnetic
field of a PWB trace. An LED indicates the presence of the field. To locate a short circuit, the
current tracer is used in conjunction with the logic pulser. As illustrated in Figure 9.2.3, pulses
are injected at an identifiable failure point. The tracer is then moved slowly along the circuit
trace until the LED goes off, indicating that the point of the fault has been passed.
Logic Test Clip

A logic test clip is a device that clips directly over an IC package to display the logic status of all
pins on the chip. Each probe of the clip contacts one pin of the IC. The probes are connected to
LEDs that indicate whether that point in the circuit is at a logic high or low, or pulsed.
The IC comparator is an extension of the logic test clip. The comparator allows the operation
of an in-circuit IC to be compared against a known-good IC of the same type placed in the com-
g
Figure 9.2.4 Typical display of a timing analyzer.
parator's “reference” socket. While the circuit containing the IC under test is operating, the com-
parator matches the responses of the reference IC to the test device.
9.2.4 Logic Analyzer

The logic analyzer is really two instruments in one: a timing analyzer and a state analyzer. The
timing analyzer is analogous to an oscilloscope. It displays information in the same general form
as a scope, with the horizontal axis representing time and the vertical axis representing voltage
amplitude. The timing analyzer samples the input waveform to determine whether it is high or
low. The instrument cares about only one voltage threshold. If the signal is above the threshold
when it is sampled, it will be displayed as a 1 (or high); any signal below the threshold is dis-
played as a 0 (or low). From these sample points, a list of ones and zeros is generated to represent
a picture of the input waveform. This data is stored in memory and used to reconstruct the input
waveform, as shown in Figure 9.2.4. A block diagram of a typical timing analyzer is shown in
Figure 9.2.5.
Sample points for the timing analyzer are developed by an internal clock. The period of the
sample can be selected by the user. Because the analyzer samples asynchronously to the unit-
under-test under the direction of the internal clock, a long sample period results in a more accu-
rate picture of the data bus. A sample period should be selected that permits an accurate view of
data activity, while not filling up the instrument's memory with unnecessary data.
g
9-26 Test Equipment
Figure 9.2.5 Block diagram of a timing analyzer.
Figure 9.2.6 Use of a delay period between the trigger point and the trace point of a timing ana-
lyzer.
Accurate sampling of data lines requires a trigger source to begin data acquisition. A fixed
delay may be inserted between the trigger point and the trace point to allow for bus settling. This
concept is illustrated in Figure 9.2.6. A variety of trigger modes are commonly available on a
timing analyzer, including:
• Level triggering—data acquisition and/or display begins when a logic high or low is detected.
• Edge triggering—data acquisition/display begins on the rising or falling edge of a selected
signal. Although many logic devices are level-dependent, clock and control signals are often
edge-sensitive.
• Bus state triggering—data acquisition/display is triggered when a specific code is detected
(specified in binary or hexadecimal).
The timing analyzer is constantly taking data from the monitored bus. Triggering, and subse-
quent generation of the trace point, controls the data window displayed to the technician. It is
g
Figure 9.2.7 Data-capturing options available from a timing analyzer.
Figure 9.2.8 Use of a timing analyzer for detecting glitches on a monitored line.
possible, therefore, to configure the analyzer to display data that precedes the trace point. (See
Figure 9.2.7.) This feature can be a powerful troubleshooting and developmental tool.
The timing analyzer is perhaps the best method of detecting glitches in computer-based
equipment. (A glitch being defined as any transition that crosses a logic threshold more than
once between clock periods. See Figure 9.2.8.) The triggering input of the analyzer is set to the
bus line that is experiencing random glitches. When the analyzer detects a glitch, it displays the
bus state preceding, during, or after occurrence of the disturbance.
The state analyzer is the second half of a logic analyzer. It is used most often to trace the exe-
cution of instructions through a microprocessor system. Data, address, and status codes are cap-
tured and displayed as they occur on the microprocessor bus. A state is a sample of the bus when
the data are valid. A state is usually displayed in a tabular format of hexadecimal, binary, octal,
or assembly language. Because some microprocessors multiplex data and addresses on the same
lines, the analyzer must be able to clock-in information at different clock rates. The analyzer, in
essence, acts as a demultiplexer to capture an address at the proper time, and then to capture data
present on the same bus at a different point in time. A state analyzer also gives the operator the
ability to qualify the data stored. Operation of the instrument may be triggered by a specific logic
pattern on the bus. State analyzers usually offer a sequence term feature that aids in triggering. A
sequence term allows the operator to qualify data storage more accurately than would be possible
with a single trigger point. A sequence term usually takes the following form:
g
9-28 Test Equipment
Figure 9.2.9 Illustration of the selective storage capability of a state analyzer.
• find xxxx
• then find yyyy
• start on zzzz
A sequence term is useful for probing a subroutine from a specific point in a program. It also
makes possible selective storage of data, as shown in Figure 9.2.9.
To make the acquired data easier to understand, most state analyzers include software that
interpret the information. Such disassemblers (also known as inverse assemblers) translate hex,
binary, or octal codes into assembly code (or some other format) to make then easier to read.
The logic analyzer is used routinely by design engineers to gain an in-depth look at signals
within digital circuits. A logic analyzer can operate at high speeds, making the instrument ideal
for detecting glitches resulting from timing problems. Such faults are usually associated with
design flaws, not with manufacturing defects or failures in the field.
While the logic analyzer has benefits for the service technician, its use is limited. Designers
require the ability to verify hardware and software implementations with test equipment; service
technicians simply need to quickly isolate a fault. It is difficult and costly to automate test proce-
dures for a given PWB using a logic analyzer. The technician must examine a long data stream
and decide if the data are good or bad. Writing programs to validate state analysis data is possi-
ble, but—again—is time-consuming.
9.2.4a Signature Analyzer

Signature analysis is a common troubleshooting tool based on the old analog method of signal
tracing. In analog troubleshooting, the technician followed an annotated schematic that depicted
waveforms and voltages that should be observed with an oscilloscope and voltmeter. The signa-
ture analyzer captures a complex data stream from a test point on the PWB and converts it into a
hexadecimal signature. These hexadecimal signatures are easy to annotate on schematics, per-
mitting their use as references for comparison against signatures obtained from a unit under test.
The signature by itself means nothing; the value of the signature analyzer comes from comparing
a known-good signature with the UUT.
A block diagram of a simplified signature analyzer is shown in Figure 9.2.10. As shown, a
digital bit stream, accepted through a data probe during a specified measurement window, is
g
Figure 9.2.10 Simplified block diagram of a signature analyzer.
marched through a 16-bit linear feedback shift register. Whatever data is left over in the register
after the specified measurement window has closed is converted into a 4-bit hexadecimal read-
out. The feedback portion of the shift register allows just one faulty bit in a digital bit stream to
create an entirely different signature than would be expected in a properly operating system. This
tool allows the maintenance technician to identify a single bit error in a digital bit stream with a
high degree of certainty, even when picking up errors that are timing-related.
To function properly, the signature analyzer requires start and stop signals, a clock input, and
the data input. The start and stop inputs are derived from the circuit being tested and are used to
bracket the beginning and end of the measurement window. During this gate period, data is input
through the data probe. The clock input controls the sample rate of data entering the analyzer.
The clock is most often taken from the clock input pin of the microprocessor. The start, stop, and
clock inputs may be configured by the technician to trigger on the rising or falling edges of the
input signals.
9.2.4b Manual Probe Diagnosis

Manual probe diagnosis integrates the logic probe, logic analyzer, and signature analyzer to trou-
bleshoot computer-based hardware. The manual probe technique employs a database consisting
of nodal-level signature measurements. Each node of a known-good unit-under-test is probed
and the signature information saved. This known-good data is then compared to the data from a
faulty UUT. Efficient use of the manual probe technique requires a skilled operator to determine
the proper order for probing circuit nodes.
9.2.4c Emulative Tester

Guided diagnostics can be used on certain types of hardware to facilitate semi-automated circuit
testing. The test connector typically taps into one of the system board input/output ports and,
armed with the proper software, the instrument goes through a step-by-step test routine that exer-
cises the computer circuits. When the system encounters a fault condition, it outputs an appropri-
ate message.
Emulative testers are a variation on the general guided diagnostics theme. Such an instrument
plugs into the microprocessor socket of the computer and runs a series of diagnostic tests to iden-
g
9-30 Test Equipment
tify the cause of a fault condition. An emulative tester emulates the board's microprocessor while
verifying circuit operation. Testing the board “inside-out” allows easy access to all parts of the
circuit; synchronization with the various data and address cycles of the PWB is automatic. Test
procedures, such as read/write cycles from the microprocessor are generic, so that high quality
functional tests can be quickly created for any board. Signature analysis is used to verify that cir-
cuits are operating correctly. Even with long streams of data, there is no maximum memory
depth.
Several different types of emulative testers are available. Some are designed to check only one
brand of computer, or only computers based on one type of microprocessor. Other instruments
can check a variety of computer systems using so-called personality modules that adapt the
instrument to the system being serviced.
Guided-fault isolation (GFI) is practical with an emulative tester because the instrument
maintains control over the entire system. Automated tests can isolate faults to the node level. All
board information and test procedures are resident within the emulative test instrument, includ-
ing prompts to the operator on what to do next. Using the microprocessor test connection com-
bined with movable probes or clips allows a closed loop test of virtually any part of a circuit.
Input/output capabilities of emulative testers range from single point probes to well over a hun-
dred I/O lines. These lines can provide stimulus as well as measurement capabilities.
The principal benefit of the guided probe over the manual probe is derived from the creation
of a topology database for the UUT. The database describes devices on the UUT, their internal
fault-propagating characteristics, and their inter-device connections. In this way, the tester can
guide the operator down the proper logic path to isolate the fault. A guided probe accomplishes
the same analysis as a conventional fault tree, but differs in that troubleshooting is based on a
generic algorithm that uses the circuit model as its input. First, the system compares measure-
ments taken from the failed UUT against the expected results. Next, it searches for a database
representation of the UUT to determine the next logical place to take a measurement. The guided
probe system automatically determines which nodes impact other nodes. The algorithm tracks its
way through the logic system of the board to the source of the fault. Programming of a guided
probe system requires the following steps:
• Development of a stimulus routine to exercise and verify the operating condition of the UUT
(go/no-go status).
• Acquisition and storage of measurements for each node affected by the stimulus routine. The
programmer divides the system or board into measurement sets (MSETs) of nodes having a
common timebase. In this way, the user can take maximum advantage of the time domain fea-
ture of the algorithm.
• Programming a representation of the UUT that depicts the interconnection between devices,
and the manner in which signals can propagate through the system. Common connectivity
libraries are developed and reused from one application to another.
• Programming a measurement set cross reference database. This allows the guided probe
instrument to cross MSET boundaries automatically without operator intervention.
• Implementation of a test program control routine that executes the stimulus routines. When a
failure is detected, the guided probe database is invoked to determine the next step in the trou-
bleshooting process.
g
9.2.4d Protocol Analyzer

Testing a local area network (LAN)—whether copper, fiber optic, or wireless—presents special
challenges to the design engineer or maintenance technician. The protocol analyzer is commonly
used to service LANs and other communications systems. The protocol analyzer performs data
monitoring, terminal simulation, and bit error rate tests (BERTs). Sophisticated analyzers pro-
vide high-level decide capability. This ability refers to the open system interconnection (OSI)
network seven layer model. These functions enable the engineer to observe activity of a commu-
nications link and to exercise the facility to verify proper operation. Simulation capability is usu-
ally available to emulate almost any data terminal or data communications equipment.
LAN test equipment functions include monitoring of protocols and emulation of a network
node, bridge, or terminal. Statistical information is provided to isolate problems to particular
devices, or to high periods of activity. Statistical data includes use, packet rates, error rates, and
collision rates.
9.2.5 Automated Test Instruments

Computers can function as powerful troubleshooting tools. Advanced technology systems, cou-
pled with add-on interface cards and applications software, provide a wide range of testing capa-
bilities. Computers can also be connected to many different instruments to provide for automated
testing. There are two basic types of stand-alone automated test instruments (ATEs): functional
and in-circuit.
Functional testing exercises the unit under test to identify faults. Individual PWBs or subas-
semblies may be checked using this approach. Functional testing provides a fast go/no-go quali-
fication check. Dynamic faults are best discovered through this approach. Functional test
instrument are well suited to high volume testing of PWB subassemblies, however, programming
is a major undertaking. An intimate knowledge of the subassembly is needed to generate the
required test patterns. An in-circuit emulator is one type of functional tester.
In-circuit testing is primarily a diagnostic tool. It verifies the functionality of the individual
components of a subassembly. Each device is checked and failing parts are identified. In-circuit
testing, while valuable for detailed component checking, does not operate at the clock rate of the
subassembly. Propagation delays, race conditions, and other abnormalities may go undetected.
Access to key points on the PWB may be accomplished in one of two ways:
• Bed of nails. The subassembly is placed on a dedicated test fixture and held in place by a vac-
uum or mechanical means. Probes access key electrical traces on the subassembly to check
individual devices. Through a technique known as backdriving, the inputs of each device are
isolated from the associated circuitry and the component is tested for functionality. This type
of testing is expensive and is only practical for high-volume subassembly qualification testing
and rework.
• PWB clips. Intended for lower volume applications than a bed of nails instrument, PWB clips
replace the dedicated test fixture. The operator places the clips on the board as directed by the
instrument. As access to all components simultaneously is not required, the tester is less hard-
ware-intensive and programming is simplified. Many systems include a library of software
routines designed to test various classes of devices. Test instruments using PWB clips tend to
g
9-32 Test Equipment
be slow. Test times for an average PWB may range from 8–20 minutes versus one minute or
less for a bed of nails.
9.2.5a Software Considerations

Most automated test instrument packages include software that permits the user to customize test
procedures to meet the required task. The software, in effect, writes software. The user enters the
codes needed by each automatic instrument, selects the measurements to be performed, and tells
the computer where to store the test data. The program then puts the final software together after
the user answers key configuration questions. The software then looks up the operational codes
for each instrument and compiles the software to perform the required tests. Automatic genera-
tion of programming greatly reduces the need to have experienced programmers on staff. Once
installed in the computer, programming becomes as simple as fitting graphic symbols together
on the screen.
Applications
The most common applications for computer-controlled testing are data gathering, product go/
no-go qualification, and troubleshooting. All depend on software to control the instruments.
Acquiring data can often be accomplished with a computer and a single instrument. The com-
puter collects dozens or hundreds of readings until the occurrence of some event. The event may
be a preset elapsed time or the occurrence of some condition at a test point, such as exceeding a
preset voltage or dropping below a preset voltage. Readings from a variety of test points are then
stored in the computer. Under computer direction, test instruments can also run checks that
might be difficult or time-consuming to perform manually. The computer may control several
test instruments such as power supplies, signal generators, and frequency counters. The data is
stored for later analysis.
9.2.6 Bibliography
Carey, Gregory D.: “Automated Test Instruments,” Broadcast Engineering, Intertec Publishing,
Detwiler, William L.: “Troubleshooting with a Digital Multimeter,' Mobile Radio Technology,
Intertec Publishing, Overland Park, Kan., September 1988.
Gore, George: “Choosing a Hand-Held Test Instrument,” Electronic Servicing and Technology,
Intertec Publishing, Overland Park, Kan., December 1988.
Harju, Rey: “Hands-Free Operation: A New Generation of DMMs,” Microservice Management,
Persson, Conrad: “Test Equipment for Personal Computers,” Electronic Servicing and Technol-
Persson, Conrad:, “The New Breed of Test Instruments,” Broadcast Engineering, Intertec Pub-
g
Ogden, Leonard: “Choosing the Best DMM for Computer Service,” Microservice Management,
Siner, T. Ann: “Guided Probe Diagnosis: Affordable Automation,” Microservice Management,
Sokol, Frank: “Specialized Test Equipment,” Microservice Management, Intertec Publishing,
Overland Park, Kan., August 1989.
2001.
Wickstead, Mike: “Signature Analyzers,” Microservice Management, Intertec Publishing, Over-
land Park, Kan., October, 1985.
g
g g
Chapter
9.3
Oscilloscopes
9.3.1 Introduction
The oscilloscope is one of the most general-purpose of all test instruments. A scope can be used
to measure voltages, examine waveshapes, check phase relationships, examine clock pulses, and
countless other functions. There has been considerable progress in scope technology within the
decade or so, most noticeably the move to portable digital systems. Some instruments offer logic
tracing functions, network connectivity, and hard copy printout.
9.3.2 Specifications
A number of parameters are used to characterize the performance of an oscilloscope. Key param-
eters include bandwidth and risetime.
9.3.2a Bandwidth
Oscilloscope bandwidth is defined as the frequency at which a sinusoidal signal will be attenu-
ated by a factor of 0.707 (or reduced to 70.7 percent of its maximum value). This is referred to as
the –3 dB point. Bandwidth considerations for sine waves are straightforward. For square waves
and other complex waveforms, however, bandwidth considerations become substantially more
involved.
Square waves are made up of an infinite number of sine waves: the fundamental frequency
plus mostly odd harmonics. Fourier analysis shows that a square wave consists of the fundamen-
tal sine wave plus sine waves that are odd multiples of the fundamental. Figure 9.3.1 illustrates
the mechanisms involved. The fundamental frequency contributes about 81.7 percent of the
square wave. The third harmonic contributes about 9.02 percent, and the fifth harmonic about
3.24 percent. Higher harmonics contribute less to the shape of the square wave. As outlined here,
approximately 94 percent of the square wave is derived from the fundamental and the third and
fifth harmonics. Inaccuracies introduced by the instrument are typically 2 percent or less. The
user's ability to read a conventional CRT display may contribute another four percent error. Thus,
a total of six percent error may be introduced by the instrument and the operator. A 96 percent
9-35
p
9-36 Test Equipment
(a ) (b) (c )
Figure 9.3.1 The mechanisms involved in square waves: (a) individual waveforms involved, (b)
waveshape resulting from combining the fundamental and the first odd harmonic (the third har-
monic, (c) waveshape resulting from combining the fundamental and the first, third, and fifth har-
monics.
accurate reproduction of a square wave should, therefore, be sufficient for all but the most criti-
cal applications. It follows that scope bandwidth is important up to and including the fifth har-
monic of the desired signal.
9.3.2b Rise Time

The rise time performance required of an oscilloscope depends on the degree of accuracy needed
in measuring the input signals. A 2–3 percent rise time accuracy can be obtained from an instru-
ment that is specified to have approximately five times the rise time of the signal being mea-
sured. Rise time can be determined from the bandwidth of the instrument
0.35
Tr = ---------- (9.3.1)
BW
Where:
Tr = instrument rise time
BW = instrument bandwidth
9.3.2c Making Measurements

The oscilloscope is a measurement instrument, displaying voltage waveforms within powered
circuits. To conduct certain types of waveform analysis, a separate signal generator is required to
inject test signals. With a conventional analog oscilloscope, actual measurement of the waveform
must be done visually by the operator. The graticule divisions along the trace are counted and
multiplied by the time base setting to determine the time interval. The TIME/DIV control deter-
mines the value of each increment on the X-axis of the display. Typical sweep ranges vary from
0.05 μs to 0.5 μs per division. To determine voltage, the operator must estimate the height of a
portion of the trace and multiply it by the voltage range selected on the instrument. The VOLT/
p
Oscilloscopes 9-37
DIV control determines the value of each increment on the Y-axis of the display. On a dual chan-
nel scope, two controls are provided, one for each source. Typical vertical sensitivity ranges vary
from 5 mV to 5 V per division.
The need for interpolation introduces inaccuracies in the measurement process. Determina-
tion of exact values is especially difficult in the microsecond ranges, where even small incre-
ments can make a big difference in the measurement. It is also difficult with a conventional
scope to make a close-up examination of specific portions of a waveform. For this purpose, an
oscilloscope with two independently adjustable sweep generators is recommended. A small sec-
tion of the waveform can be selected for closer examination by adjusting the B channel sweep-
time and delay-time controls while observing an intensified portion of the waveform. The inten-
sified section can then be expanded across the screen. This type of close-up examination is use-
ful for measuring pulse rise times and viewing details in complex signals, such as video
horizontal sync or color burst.
9.3.2d Advanced Features

Modern oscilloscopes remove the potential inaccuracies of previous instruments by incorporat-
ing advanced features, including:
• Numeric measurement and display. On-screen readouts provide key measurement parameters
for the operator. Readout options include peak waveform voltage, frequency, and sweep and
sensitivity settings for the instrument. On-screen readout makes waveform photography or
screen capture more valuable because the photograph displays not only the waveform, but
instrument settings as well.
• Triggering. A wide selection of trigger modes provides operational flexibility. The trigger
mode determines what condition starts the horizontal sweep to display the voltage waveform.
The trigger signal can be derived from an external source, or from the input channel(s). The
trigger is usually continuously adjustable through both positive and negative slopes of the sig-
nal. Selectable trigger coupling includes special settings needed in particular applications. A
variable hold-off control can be used to obtain a stable display when measuring periodic or
complex signals.
• Control and monitoring. Microprocessors have been integrated into scopes to reduce the num-
ber of mechanical switch contacts, which are prone to failure, and to permit front-panel set-
ting recall capabilities. Auxiliary signal outputs are available to feed other instruments, such
as a distortion analyzer or frequency counter.
9.3.3 Digital Oscilloscope

The digital storage oscilloscope (DSO) offers a number of significant advantages beyond the
capabilities of analog instruments. A DSO can store in memory the signal being observed, per-
mitting in-depth analysis impossible with previous technology. Because the waveform resides in
memory, the data associated with the waveform can be transferred to a computer for real-time
processing, or for processing at a later time.
p
9-38 Test Equipment
Figure 9.3.2 Simplified block diagram of a digital storage oscilloscope.
Figure 9.3.3 The quantization process.
9.3.3a Operating Principles

Before the introduction of the DSO, the term storage oscilloscope referred to a scope that used a
storage CRT. This type of instrument stored the display waveform as a trace on the scope face.
The digital storage oscilloscope operates on a completely different premise. Figure 9.3.2 shows a
block diagram of a DSO. Instead of being amplified and directly applied to the deflection plates
of a CRT, the waveform is first converted into a digital form and stored in memory. To reproduce
the waveform on the CRT, the data is sequentially read and converted back into an analog signal
for display.
The analog-to-digital (A/D) converter transforms the input analog signal into a sequence of
digital bits. The amplitude of the analog signal, which varies continuously in time, is sampled at
preset intervals. The analog value of the signal at each sample point is converted into a binary
number. This quantization process is illustrated in Figure 9.3.3. The sample rate of a digital
oscilloscope must be greater than two times the highest frequency to be measured (Nyquist
limit). The higher the sampling rate relative to the input signal, the greater the measurement
accuracy. A sample rate 10 times the input signal is sufficient for most applications. This rule-of-
thumb applies for single-shot signals, or signals that are constantly changing. The sample rate
may also be expressed as the sampling interval, or the period of time between samples. The sam-
ple interval is the inverse of the sampling frequency.
p
Oscilloscopes 9-39
Although a DSO is specified by its maximum sampling rate, the actual rate used in acquiring
a given waveform is usually dependent on the time-per-division setting of the oscilloscope. The
record length (samples recorded over a given period of time) defines a finite number of sample
points available for a given acquisition. The DSO must, therefore, adjust its sampling rate to fill
a given record over the period set by the sweep control. To determine the sampling rate for a
given sweep speed, the number of displayed points per division is divided into the sweep rate per
division. Two additional features can modify the actual sampling rate:
• Use of an external clock for pacing the digitizing rate. With the internal digitizing clock dis-
abled, the digitizer will be paced at a rate defined by the operator.
• Use of a peak detection (or glitch capture) mode. Peak detection allows the digitizer to sam-
ple at the full digitizing rate of the DSO, regardless of the time base setting. The minimum
and maximum values found between each normal sample interval are retained in memory.
These minimum and maximum values are used to reconstruct the waveform display with the
help of an algorithm that recreates a smooth trace along with any captured glitches. Peak
detection allows the DSO to capture glitches even at its slowest sweep speed. For higher per-
formance, a technique known as peak-accumulation (or envelope mode) may be used. With
this approach, the instrument accumulates and displays the maximum and minimum excur-
sions of a waveform for a given point in time. This builds an envelope of activity that can
reveal infrequent noise spikes, long-term amplitude or time drift, and pulse jitter extremes.
9.3.3b A/D Conversion

The A/D conversion process is critical to the overall accuracy of the oscilloscope. Three types of
converters are commonly used:
• Flash converter
• Successive approximation converter
• Charge-coupled device (CCD) converter
The CCD is an analog storage array. The DSO can capture high-speed events in real time using a
CCD, then route the captured information to be digitized to low cost A/D converters.
Conversion resolution is determined by the number of bits into which the analog signal is
transformed. The higher the number of bits available to the A/D converter, the more discrete lev-
els the digital signal is able to describe. For example, an 8-bit digitizer has 256 levels, while a 10-
bit digitizer has 1,024 levels. Although a higher number of bits allows greater discrimination
between voltage values, the overall accuracy of the DSO may be limited by other factors, includ-
ing the accuracy of the analog amplifier(s) feeding the digitizer(s). The usefulness of a high con-
version digitizer may also be limited by the resolution of the display screen.
While most common DSOs provide 8-bit digitizing, there may be significant differences
between individual 8-bit instruments. First among these is the useful storage bandwidth. This
specification is further divided into single-shot bandwidth and equivalent-time bandwidth.
Achieving adequate samples for a high frequency waveform places stringent requirements on
the sampling rate. High frequency waveforms require high sampling rates. Equivalent-time sam-
pling is often used to provide high-bandwidth capture. This technique relies on sampling a repet-
itive waveform at a low rate to build up sample density. This concept is illustrated in Figure
9.3.4. When the waveform to be acquired triggers the DSO, several samples are taken over the
p
9-40 Test Equipment
Figure 9.3.4 Increasing sample density through equivalent-time-sampling.
duration of the waveform. The next repetition of the waveform triggers the instrument again, and
more samples are taken at different points on the waveform. Over many repetitions, the number
of stored samples can be built up to the equivalent of a high sampling rate.
System operating speed also has an effect on the display update rate. Update performance is
critical when measuring waveform or voltage changes. If the display does not track the changes,
adjustment of the circuit may be difficult.
9.3.3c DSO Architecture

A detailed block diagram of a basic DSO is shown in Figure 9.3.5. The features included in this
design are typical of a digital scope utilizing a CRT display device. The major circuit blocks are:
• Data acquisition—converts the analog input signal into a digital equivalent for storage. Data
control signals are routed to a microprocessor through an I/O port.
• Data processing—provides interpolation of the stored data for reproduction of the analog
signal on the CRT or other display device. The processed data is transferred to the display
memory. Except during transfer periods, the contents of the display memory are constantly
being sent to the digital-to-analog (D/A) converter and displayed on the CRT.
• Vertical deflection—generates the vertical deflection voltages for the CRT. The input signal
is impedance-matched and amplified by the Channel 1 and Channel 2 preamplifiers to a level
suitable for driving subsequent stages. Channel selection switches operate under control of
the microprocessor.
• Horizontal deflection—generates the necessary voltages to deflect the beam spot horizon-
tally across the screen. The stage includes trigger and sweep circuits. The horizontal deflec-
tion stage also acts as an X-axis amplifier for X-Y operation. Channel selection switches
operate under control of the microprocessor.
• CRT driver—provides the high voltages necessary to drive the CRT, and controls the Z-axis
of the CRT. The Z-axis circuit sets the intensity and focus of the CRT beam spot. For other
types of display devices, such as LCD or plasma units, the appropriate functions are per-
formed in this block.
• Power supplies—provides the necessary low voltages for other stages of the instrument.
p
Figure 9.3.5 Block diagram of a DSO showing the conversion, I/O memory, and display circuits (Courtesy of Kikusui.)
Oscilloscopes 9-41
p
9-42 Test Equipment
• Calibration circuit—provides a reference signal for calibration of the oscilloscope probe,

and for operation of the amplifiers in the non-calibrated mode.
9.3.3d DSO Features

The digital oscilloscope has become a valuable tool in troubleshooting both analog and digital
products. Advanced components and construction techniques have led to lower costs for digital
scopes and higher performance. Digital scopes can capture and analyze transient signals, such as
race conditions, clock jitter, glitches, dropouts, and intermittent faults. Automated features
reduce testing and troubleshooting costs through the use of recallable instrument setups, direct
parameter readout, and unattended monitoring. Benefits include:
• Increased resolution (determined by the quality of the analog-to-digital converter).
• Memory storage of digitized waveforms.
• Automatic setup for repetitive signal analysis. For complex multichannel configurations that
are used often, front-panel storage/recall features can save dozens of manual selections and
adjustments. When multiple memory locations are available, multiple front-panel setups can
be stored to save even more time.
• Auto-ranging. Many instruments will automatically adjust for optimum sweep, input sensitiv-
ity, and triggering. The instrument's microprocessor automatically configures the front panel
for optimum display. Such features permit the operator to concentrate on making measure-
ments, not on adjusting the scope.
• Instant hardcopy output from printers and plotters.
• Remote programmability for automated test applications, including access via various types
of networks.
• Trigger flexibility. Single-shot digitizing oscilloscopes capture transient signals and allow the
user to view the waveform which preceded the trigger point.
• Signal analysis. Intelligent scopes can make key measurements and comparisons. Display
capabilities include voltage peak, mean voltage, RMS value, rise time, fall time, and fre-
quency.
• Cursor measurement. Advanced oscilloscopes permit the operator to take measurements or
perform comparative operations on data appearing on the display. A measurement cursor con-
sists of a pair of lines or dots that can be moved around the screen as needed.
• Trace quality. Eye fatigue is reduced noticeably with a DSO when viewing low-repetition sig-
nals. For example, a 60 Hz waveform can be difficult to view for extended periods of time on
a conventional scope because the display tends to flicker. A DSO overcomes this problem by
writing waveforms to the screen at the same rate regardless of the input signal.
9.3.3e Capturing Transient Waveforms

Single-shot digitizing makes it possible to capture and clearly display transient and intermittent
signals. Waveforms such as signal glitches, dropouts, logic race conditions, intermittent faults,
clock jitter, and power-up sequences can be examined with the help of a digital oscilloscope.
p
Oscilloscopes 9-43
With single-shot digitizing, the waveform is captured the first time it occurs, on the first trigger.
It can then be displayed immediately or held in memory for analysis at a later date.
Triggering
Basic triggering modes available on a digital oscilloscope permit the user to select the desired
source, its coupling, level, and slope. More advanced digital scopes contain triggering circuitry
similar to that found in a logic analyzer. These powerful features let the user trigger on elusive
conditions, such as pulse widths less than or greater than expected, intervals less than or greater
than expected, and specified logic conditions. The logic triggering can include digital pattern,
state qualified, and time/event qualified conditions. Many trigger modes are further enhanced by
allowing the user to hold-off the trigger by a selectable time or number of events. Hold-off is
especially useful when the input signal contains bursts of data or follows a repetitive pattern.
These features are summarized as follows:
• Pulse-width triggering lets the operator quickly check for pulses narrower than expected or
wider than expected. The pulse-width trigger circuit checks the time from the trigger source
transition of a given slope (typically the rising edge) to the next transition of opposite slope
(typically the falling edge). The operator can interactively set the pulse-width threshold for
the trigger. For example, a glitch can be considered any signal narrower than 1/2 of a clock
period. Conditions preceding the trigger can be displayed to show what events led up to the
glitch.
• Interval triggering lets the operator quickly check for intervals narrower than expected or
wider than expected. Typical applications include monitoring for transmission phase changes
in the output of a modem or for signal dropouts, such as missing bits in a transport stream.
• Pattern triggering lets the user trigger on the logic state (high, low, or either) of several inputs.
The inputs can be external triggers or the input channels themselves. The trigger can be gen-
erated either upon entering or exiting the pattern. Applications include triggering on a partic-
ular address select or data bus condition. Once the pattern trigger is established, the operator
can probe throughout the circuit, taking measurement synchronous with the trigger.
• State qualified triggering enables the oscilloscope to trigger on one source, such as the input
signal itself, only after the occurrence of a specified logic pattern. The pattern acts as an
enable or disable for the source.
Advanced Features
Some digital oscilloscopes provide enhanced triggering modes that permit the user to select the
level and slope for each input. This flexibility makes it easy to look for odd pulse shapes in the
pulse-width trigger mode and for subtle dropouts in the interval trigger mode. It also simplifies
testing different logic types (TTL, CMOS, and ECL), and testing analog/digital combinational
circuits with dual logic triggering. Additional flexibility is available on multiple channel scopes.
Once a trigger has been sensed, multiple simultaneously sampled inputs permit the user to moni-
tor conditions at several places in the unit under test, with each channel synchronized to the trig-
ger. Additional useful trigger features for monitoring jitter or drift on a repetitive signal include:
• Enveloping
• Extremes
p
9-44 Test Equipment
• Waveform delta
• Roof/floor
Each of these functions are related, and in some cases describe the same general operating mode.
Various scope manufacturers use different nomenclature to describe proprietary triggering meth-
ods. Generally speaking, as the waveshape changes with respect to the trigger, the scope gener-
ates upper and lower traces. For every nth sample point with respect to the trigger, the maximum
and minimum values are saved. Thus, any jitter or drift is displayed in the envelope.
Advanced triggering features provide the greatest benefit in conjunction with single-shot
sampling. Repetitive sampling scopes can only capture and display signals that repeat precisely
from cycle to cycle. Several cycles of the waveform are required to create a digitally-recon-
structed representation of the input. If the signal varies from cycle to cycle, such scopes can be
less effective than a standard analog oscilloscope for accurately viewing a waveform.
9.3.4 Bibliography
Albright, John R.: “Waveform Analysis with Professional-Grade Oscilloscopes,” Electronic Ser-
Breya, Marge: “New Scopes Make Faster Measurements,” Mobile Radio Technology Magazine,
Intertec Publishing, Overland Park, Kan., November 1988.
Harris, Brad: “The Digital Storage Oscilloscope: Providing the Competitive Edge,” Electronic
Servicing and Technology, Intertec Publishing, Overland Park, Kan., June 1988.
Harris, Brad: “Understanding DSO Accuracy and Measurement Performance,” Electronic Ser-
Hoyer, Mike: “Bandwidth and Rise Time: Two Keys to Selecting the Right Oscilloscope,” Elec-
tronic Servicing and Technology, Intertec Publishing, Overland Park, Kan., April 1990.
Montgomery, Steve: “Advanced in Digital Oscilloscopes,” Broadcast Engineering, Intertec Pub-
Persson, Conrad: “Oscilloscope Special Report,” Electronic Servicing and Technology, Intertec
Persson, Conrad: “Oscilloscope: The Eyes of the Technician,” Electronic Servicing and Technol-
ogy, Intertec Publishing, Overland Park, Kan., April 1987.
Toorens, Hans: “Oscilloscopes: From Looking Glass to High-Tech,” Electronic Servicing and
Technology, Intertec Publishing, Overland Park, Kan., April 1990.
2001.
g g
Chapter
9.4
Spectrum Analysis
9.4.1 Introduction
An oscilloscope-type instrument displays voltage levels referenced to time and a spectrum ana-
lyzer displays signal levels referenced to frequency. The frequency components of the signal
applied to the input of the analyzer are detected and separated for display against a frequency-
related time base. Spectrum analyzers are available in a variety of ranges with some models
designed for use with audio or video frequencies, and others intended for use with RF frequen-
cies.
The primary application of a spectrum analyzer is the measurement and identification of RF
signals. When connected to a small receiving antenna, the analyzer can measure carrier and side-
band power levels. By expanding the sweep width of the display, offset or multiple carriers can
be observed. By increasing the vertical sensitivity of the analyzer and adjusting the center fre-
quency and sweep width, it is possible to observe the occupied bandwidth of the RF signal. Con-
vention dictates that the vertical axis displays amplitude, and the horizontal axis displays
frequency. This frequency-domain presentation allows the user to glean more information about
the characteristics of an input signal than is possible from an oscilloscope. Figure 9.4.1 compares
the oscilloscope and spectrum analyzer display formats.
9.4.2 Principles of Operation

A spectrum analyzer intended for use at RF frequencies is shown in block diagram form in Fig-
ure 9.4.2. The instrument includes a superheterodyne receiver with a swept-tuned local oscillator
(LO) that feeds a CRT or other type of display. The tuning control determines the center fre-
quency (Fc) of the spectrum analyzer, and the scan-width selector determines how much of the
frequency spectrum around the center frequency will be covered. Full-feature spectrum analyz-
ers also provide front-panel controls for scan-rate selection and bandpass filter selection. Key
specifications for a spectrum analyzer include:
• Resolution: the frequency separation required between two signals so that they may be
resolved into two distinct and separate displays on the screen. Resolution is usually specified
9-45
p y
9-46 Test Equipment
for equal-level signals. When two sig- (a)

nals differ greatly in amplitude and are
close together in frequency, greater res-
olution is required to separate them on
the display.
• Scan width: the amount of frequency
spectrum that can be scanned and
shown on the display. Scan width is
usually stated in kilohertz or megahertz (b)
per division. The minimum scan width
available usually is equal to the resolu-
tion of the instrument.
• Dynamic range: the maximum ampli-
tude difference that two signals can
have and still be viewed on the display.
Dynamic range is usually stated in
Figure 9.4.1 Comparison of waveform displays:
decibels.
(a) oscilloscope, ( b) spectrum analyzer. (After [1].)
• Sensitivity: the minimum signal level
required to produce a usable display on
the screen. If low-level signal tracing is planned, as in receiver or off-air monitoring, the sen-
sitivity of the spectrum analyzer is important.
Figure 9.4.2 Block diagram of a spectrum analyzer. (After [1].)

p y
When using the spectrum analyzer, care must be taken to not overload the front-end with a
strong input signal. Overloading can cause “false” signals to appear on the display. These false
signals are the result of non-linear mixing in the front-end of the instrument. False signals may
be identified by changing the RF attenuator setting to a higher level. The amplitude of false sig-
nals (caused by overloading) will drop much more than the amount of increased attenuation.
The spectrum analyzer is useful in troubleshooting receivers as well as transmitters. As a
tuned signal tracer, it is well adapted to stage-gain measurements and other tests. There is one
serious drawback, however. The 50 Ω spectrum analyzer input can load many receiver circuits
too heavily, especially high impedance circuits such as FET amplifiers. Isolation probes are
available to overcome loading problems. Such probes, however, also attenuate the input signal,
and unless the spectrum analyzer has enough reserve gain to overcome the loss caused by the iso-
lation probe, the instrument will fail to provide useful readings. Isolation probes with 20 dB to
40 dB attenuation are typical. As a rule of thumb, probe impedance should be at least 10 times
the impedance of the circuit to which it is connected.
9.4.3 Applications
The primary application for a spectrum analyzer centers around measuring the occupied band-
width of an input signal. Harmonics and spurious signals can be checked and potential causes
investigated. Figure 9.4.3 shows a typical test setup for making transmitter measurements.
The spectrum analyzer is also well-suited to making accurate transmitter FM deviation mea-
surements. This is accomplished using the Bessel null method. The Bessel null is a mathematical
function that describes the relationship between spectral lines in frequency modulation. The
Bessel null technique is highly accurate; it forms the basis for modulation monitor calibration.
The concept behind the Bessel null method is to drive the carrier spectral line to zero by chang-
ing the modulating frequency. When the carrier amplitude is zero, the modulation index is given
by a Bessel function. Deviation may be calculated from
Δ fc = MI × f (9.4.1)
Where:
Δfc = deviation frequency
MI = modulation index
fm = modulating frequency
The carrier frequency “disappears” at the Bessel null point, with all power remaining in the
FM sidebands.
A tracking generator may be used in conjunction with the spectrum analyzer to check the
dynamic response of frequency-sensitive devices, such as transmitter isolators, cavities, ring
combiners, duplexers, and antenna systems. A tracking generator is a frequency source that is
locked in step with the spectrum analyzer horizontal trace rate. The resulting display shows the
relationship of the amplitude-versus-frequency response of the device under test. The spectrum
analyzer may also be used to perform gain-stage measurements. The combination of a spectrum
analyzer and a tracking generator makes filter passband measurements possible. As measure-
p y
9-48 Test Equipment
Figure 9.4.3 Test setup for measuring the harmonic and spurious output of a transmitter. The
notch filter is used to remove the fundamental frequency to prevent overdriving the spectrum ana-
lyzer input. (After [2].)
ments are made along the IF chain of a receiver, the filter passbands become increasingly narrow,
as illustrated in Figure 9.4.4.
9.4.3a On-Air Measurements

The spectrum analyzer is used for three primary on-air tests:
• Measuring unknown signal frequencies. A spectrum analyzer may be coupled to the output of
a transmitter to determine its exact operating frequency.
• Intermod and interference signal tracking. A directional Yagi antenna may be used to identify
the source of an interfering signal. If the interference is on-frequency, but carries little intelli-
gence, chances are good that it is an intermod being produced by another transmitter. Use the
wide dispersion display mode of the analyzer and note signal spikes that appear simulta-
neously with the interference. A troubleshooter, armed with a spectrum analyzer, a directional
antenna, and the knowledge of how intermod signals are generated can usually locate a sus-
pected transmitter rapidly. Each suspected unit then can be tested individually by inserting an
isolator between the transmitter PA and duplexer (or PA and antenna). When the intermod sig-
p y
nal amplitude drops the equivalent of

the reverse insert loss of the isolator,
the offending transmitter has been
located.
• Field strength measurements. With an
external antenna, a spectrum analyzer
can be used for remote field strength
measurements. Broadband (usually
omnidirectional) antennas are the most
versatile because they allow several fre-
quency bands to be checked simulta-
neously. Obviously, the greater the
external antenna height, the greater the
testing range.
9.4.3b Spurious Harmonic Distortion

An incorrectly-tuned or malfunctioning
transmitter can produce spurious harmon-
ics. The spectrum analyzer provides the
best way to check for spectral purity. Cou-
ple a sample of the transmitter output sig-
nal to the analyzer input, either by loop
coupling or RF sampling, as illustrated in
Figure 9.4.5. For low power levels from
portable and mobile units, transmit into a Figure 9.4.4 Using a spectrum analyzer to mea-
dummy load and use a small antenna on sure the tunnel effect of bandpass filters in a
the analyzer input. For maximum accuracy receiver. (After [2].)
when measuring larger amounts of power,
use an RF sampler to control the input level to the analyzer. Use maximum RF attenuation ini-
tially on the analyzer front end to prevent overload damage and internal intermod. False signals
on the display may also be observed if covers or shields are not in place on the radio under test.
The oscillator, doubler, or tripper levels may radiate sufficient signals to register on the analyzer
display.
Energize the transmitter and observe any spurious harmonics on the analyzer display. After
centering the main signal of interest and adjusting the input attenuation for maximum display
amplitude, calculate the spurious radiation. Spurious signal attenuation is measured in decibels,
referenced to the amplitude of the transmitter fundamental. When a radio transmitter has a har-
monic distortion problem, the defective stage usually can be isolated by tuning each stage and
observing the spurious harmonics on the analyzer display. When the defective stage is tuned, the
harmonics either shift frequency or change in amplitude. Signal tracing, with a probe and heavy
input attenuation, also helps to determine where the distortion first occurs. Overdriven stages are
prime culprits.
p y
9-50 Test Equipment
Figure 9.4.5 Common test setup to measure transmitter harmonic distortion with a spectrum ana-
lyzer. (After [3].)
9.4.3c Selective-Tuned Filter Alignment

The high input sensitivity and visual, frequency-selective display of the spectrum analyzer pro-
vide the technician with an efficient analog tuning instrument. Injection circuits can be peaked
quickly by loop-coupling the analyzer input to the mixer section of a receiver. This is most help-
ful in tuning older radios when a service manual is not available. Radios without test points, or
that require elaborate, specialized test sets, can be tuned in a similar manner. Portable radio
transmitters can be tuned directly into their antennas for maximum signal strength. This reveals
problems such as improper signal transfer to the antenna or a defective antenna.
Cavities, combiners, and duplexers are all selective-tuned filters. They are composed of three
types of passive filters:
• Bandpass
• Band-reject
• Combination pass/reject
The spectrum analyzer is well-suited to tuning these types of filters, particularly duplexers.
Couple the cavity between the tracking generator and the spectrum analyzer. Calibrate the tuning
frequency in the center of the output display, and adjust the generator output and analyzer input
level for an optimal display trace. Then tune the cavity onto frequency. Measure the various filter
characteristics and compare them to the manufacturer's specifications. Figure 9.4.6 shows a typ-
ical spectrum analyzer display of a single cavity device. Important filter characteristics include
frequency, bandwidth, insertion loss, and selectivity. Measure the bandwidth at the points 3 dB
down from maximum amplitude. Insertion loss is equal to the cavity's attenuation of the center-
tune frequency. Calculate filter Q (quality factor), which is directly proportional to selectivity, by
dividing the filter center frequency by the bandwidth of the filter.
Duplexer tuning is similar to cavity tuning but is complicated by interaction between the mul-
tiple cavities. Using the test setup shown in Figure 9.4.7, couple the tracking generator output to
the duplexer antenna input. Couple the receive and transmit ports alternately to the spectrum
analyzer while tuning the pass or reject performance of each cavity. When adjusting a combina-
tion cavity, tune the reject last because it will tend to follow the pass tuning. Alternate tuning
between the receive and transmit sides of the duplexer until no further improvement can be
gained. To assure proper alignment, terminate all open ports into a 50 Ω load while tuning.
p y
Figure 9.4.6 Typical spectrum analyzer display of a single-cavity filter. Bandwidth (BW) = 458 MHz
to 465 MHz = 7 MHz. Filter quality factory (Q) = fCT / BW = 463 MHz / 7 MHz = 66. The trace
shows 1 Db insertion loss. (After [3].)
Figure 9.4.7 Test setup for duplexer tuning using a spectrum analyzer. (After [3].)
The duplexer insertion loss of each port can be measured as the difference between a refer-
ence amplitude and the pass frequency amplitude. The reference signal is measured by shorting
the generator output cable to the analyzer input cable. This reference level nulls any cable losses
to prevent them from being included in duplexer insertion loss measurements.
p y
9-52 Test Equipment
9.4.3d Small-Signal Troubleshooting

The spectrum analyzer is well-suited for use in small-signal RF troubleshooting. Stage gain,
injection level, and signal loss can be easily checked.
Receiver sensitivity loss of less than 6 dB is usually a difficult problem to trace with most test
instruments. Such losses usually are associated with the front-end of the radio, either as a stage
gain or injection amplifier deficiency. Before the spectrum analyzer was commonly available,
there was little to aid a technician in tracing such a problem. An RF millivolt meter, with typical
usable sensitivity of 1 mV, was of little help in troubleshooting such microvolt-level signals.
Oscilloscopes are restricted in the same way, as well as having inadequate bandwidth for the high
frequencies used in the front-end circuits of most receivers. A spectrum analyzer, with a typical 1
μV sensitivity and broad RF bandwidth, is well adapted for RF troubleshooting. Microvolt-level
signals can be accurately traced, allowing signal losses to be examined at each individual stage.
To conduct measurements, inject a test signal generated by a communications monitor into
the antenna input of the radio under test. Use an appropriate probe to trace the signal from stage-
to-stage. Stage gain, mixer output, injection level, and filter loss can be checked from the
antenna to the discriminator. Compare measurements taken to those of a correctly functioning
unit to isolate the problem. A standard test signal level of 100 μV usually works well. It is large
enough to allow testing of the RF amplifier without causing overload. If the radio includes an
AGC circuit, disable it temporarily during testing to permit measurement of the true stage gain.
When tracing a loss of sensitivity, start with the RF mixer. The mixer provides a junction
point for narrowing the direction of troubleshooting. If the output of the mixer is correct, the
defective stage lies beyond the mixer, probably in the IF amplifier. If the mixer output is low,
check the input signal to the mixer. Check the RF amplifier, input cavities, and receive/transmit
(RX/TX) switching circuit (if used). If the mixer IF carrier injection is low, check the multipliers
and oscillators along that path.
The bandwidth of IF filters can be measured in-circuit by varying the receiver input fre-
quency and measuring the bandwidth of the 3 dB roll-off points. A crystal bandpass filter can be
tested for flaws by injecting an over-deviated signal into the receiver. A cracked filter crystal will
produce a sharp, deep notch on the analyzer display. The key to comparative troubleshooting is
accurate recording of test signal levels of a radio known to be working correctly prior to trouble-
shooting.
Remember to protect the front-end of the analyzer from dc voltages and overload when trac-
ing signals with a probe. When testing unknown signal levels, use maximum input attenuation
and external attenuators as needed to avoid analyzer damage. Isolate the analyzer input from dc
voltages by using a capacitively-coupled probe.
9.4.4 Defining Terms1

1 dB compression point The point approaching saturation at which the output is 1 dB less than
it should be if the output linearly followed input.
antenna sweep A technique for measuring the return loss of an antenna to determine the antenna
tuning.
1. Definitions courtesy of Tektronix.

p y
average detection A detection scheme wherein the average (mean) amplitude of a signal is mea-
sured and displayed.
B-SAVE A (or B, C MINUS A) Waveform subtraction mode wherein a waveform in memory is
subtracted from a second, active waveform and the result displayed on screen.
band switching Technique for changing the total range of frequencies (band) to which a spec-
trum analyzer can be tuned.
baseband The lowest frequency band in which a signal normally exists (often the frequency
range of a receiver’s output or a modulator’s input); the band from dc to a designated fre-
quency.
baseline clipper A means of blanking the bright baseline portion of the analyzer display.
Bessel functions Solutions to a particular type of differential equation; predicts the amplitudes
of FM signal components.
Bessel null method A technique most often used to calibrate FM deviation meters. A modulat-
ing frequency is chosen such that some frequency component of the FM signal nulls at a
specified peak deviation.
calibrator A signal generator producing a specified output used for calibration purposes.
carrier-to-noise ratio (C/N) The ratio of carrier signal power to average noise power in a given
bandwidth surrounding the carrier; usually expressed in decibels.
center frequency The frequency at the center of a given spectrum analyzer display.
coax bands The range of frequencies that can be satisfactorily passed via coaxial cable.
comb generator A source producing a fundamental frequency component and multiple compo-
nents at harmonics of the fundamental.
component In spectrum analysis, usually denotes one of the constituent sine waves making up
electrical signals.
decibel (dB) Ten times the logarithm of the ratio of one electrical power to another.
delta F (ΔF) A mode of operation on a spectrum analyzer wherein a difference in frequency may
be read out directly.
distortion Degradation of a signal, often a result of nonlinear operations, resulting in unwanted
signal components. Harmonic and intermodulation distortion are common types.
dynamic range The maximum ratio of two simultaneously present signals which can be mea-
sured to a specified accuracy.
emphasis Deliberate shaping of a signal spectrum or some portion thereof, often used as a
means of overcoming system noise. Preemphasis is often used before signal transmission
and deemphasis after reception.
envelope The limits of an electrical signal or its parameters. For instance, the modulation enve-
lope limits the amplitude of an AM carrier.
equivalent noise bandwidth The width of a rectangular filter that produces the same noise
power at its output as an actual filter when subjected to a spectrally flat input noise signal.
p y
9-54 Test Equipment
Real filters pass different noise power than implied by their nominal bandwidths because
their skirts are not infinitely sharp.
external mixers A mixer, often in a waveguide format, that is used external to a spectrum ana-
lyzer.
filter A circuit that separates electrical signals or signal components based on their frequencies
filter loss The insertion loss of a filter: the minimum difference in dB between the input signal
level and the output level.
first mixer input level Signal amplitude at the input to the first mixer stage of a spectrum ana-
lyzer. An optimum value is usually specified by the manufacturer.
flatness Unwanted variations in signal amplitude over a specified bandwidth, usually expressed
in dB.
Fourier analysis A mathematical technique for transforming a signal from the time domain to
the frequency domain and vice versa.
frequency band A range of frequencies that can be covered without switching.
frequency deviation The maximum difference between the instantaneous frequency and the car-
rier frequency of an FM signal.
frequency domain representation The portrayal of a signal in the frequency domain; represent-
ing a signal by displaying its sine wave components; the signal spectrum.
frequency marker An intensified or otherwise distinguished spot on a spectrum analyzer dis-
play indicating a specified frequency point.
frequency range That range of frequencies over which the performance of the instrument is
specified.
fundamental frequency The basic rate at which a signal repeats itself.
grass Noise or a noise-like signal giving the ragged, hashy appearance of grass seen close-up at
eye level.
graticule The calibrated grid overlaying the display screen of spectrum analyzers, oscilloscopes,
and other test instruments.
harmonic distortion The distortion that results when a signal interacts with itself, often because
of non linearities in the equipment, to produce sidebands at multiples, or harmonics, of the
frequency components of the original signal.
harmonic mixing A technique wherein harmonics of the local oscillator signal are deliberately
mixed with the input signal to achieve a large total input bandwidth. Enables a spectrum
analyzer to function at higher frequencies than would otherwise be possible.
harmonics Frequency components of a signal occurring at multiples of the signal’s fundamental
frequency.
heterodyne spectrum analyzer A type of spectrum analyzer which scans the input signal by
sweeping the incoming frequency band past one of a set of fixed resolution bandwidth fil-
ters and measuring the signal level at the output of the filter.
p y
intermediate frequency (IF) In a heterodyne process, the sum or difference frequency at the
output of a mixer stage which will be used for further signal processing.
IF gain The gain of an amplifier stage operating at the IF frequency.
instantaneous frequency The rate of change of the phase of a sinusoidal signal at a particular
instant.
intermodulation distortion The distortion that results when two or more signals interact, usu-
ally because of non-linearities in the equipment, to produce new signals.
linear scale A scale wherein each increment represents a fixed difference between signal levels.
LO output A port on a spectrum analyzer where a signal from the local oscillator is made avail-
able; used for tracking generators and external mixing.
local oscillator An oscillator that produces the internal signal that is mixed with an incoming
signal in a mixer to produce the IF signal.
logarithmic scale A scale wherein each scale increment represents a fixed ratio between signal
levels.
magnitude-only measurement A measurement that responds only to the magnitude of a signal
and is insensitive to its phase.
MAX HOLD A spectrum analyzer feature which captures the maximum signal amplitude at all
displayed frequencies over a series of sweeps.
max span The maximum frequency span that can be swept and displayed by a spectrum ana-
lyzer.
MAX/MIN A display mode on some spectrum analyzers that shows the maximum and mini-
mum signal levels at alternate frequency points; its advantage is its resemblance to an ana-
log display.
maximum input level The maximum input signal amplitude which can be safely handled by a
particular instrument.
MIN HOLD A spectrum analyzer feature which captures the minimum signal amplitude at all
displayed frequencies over a series of sweeps.
mixing The process whereby two or more signals are combined to produce sum and difference
frequencies of the signals and their harmonics.
noise Unwanted random disturbances superimposed on a signal which tend to obscure it.
noise bandwidth The frequency range of a noise-like signal. For white noise, the noise power is
directly proportional to the bandwidth of the noise.
noise floor The self-noise of an instrument or system that represents the minimum limit at which
input signals can be observed. The spectrum analyzer noise floor appears as a “grassy”
baseline in the display even when no signal is present.
noise sideband Undesired response caused by noise internal to the spectrum analyzer appearing
on the display immediately around a desired response, often having a pedestal-like appear-
ance.
p y
9-56 Test Equipment
peak/average cursor A manually controllable function which enables the user to set the thresh-
old at which the type of signal processing changes prior to display in a digital storage sys-
tem.
peak detection A detection scheme wherein the peak amplitude of a signal is measured and dis-
played. In spectrum analysis, 20 log (peak) is often displayed.
peaking The process of adjusting a circuit for maximum amplitude of a signal by aligning inter-
nal filters. In spectrum analysis, peaking is used to align preselectors.
period The time interval at which a process recurs; the inverse of the fundamental frequency.
phase lock The control of an oscillator or signal generator so as to operate at a constant phase
angle relative to a stable reference signal source. Used to ensure frequency stability in spec-
trum analyzers.
preselector A tracking filter located ahead of the first mixer which allows only a narrow band of
frequencies to pass into the mixer.
products Signal components resulting from mixing or from passing signals through other non-
linear operations such as modulation.
pulse stretcher Pulse shaper that produces an output pulse whose duration is greater than that of
the input pulse and whose amplitude is proportional to that of the peak amplitude of the
input pulse.
pulse repetition frequency (PRF) The frequency at which a pulsing signal recurs; equal to the
fundamental frequency of the pulse train.
reference level The signal level required to deflect the CRT display to the top graticule line.
reference level control The control used to vary the reference level on a spectrum analyzer.
resolution bandwidth (RBW) The width of the narrowest filter in the IF stages of a spectrum
analyzer. The RBW determines how well the analyzer can resolve or separate two or more
closely spaced signal components.
return loss The ratio of power sent to a system to that returned by the system. In the case of
antennas, the return loss can be used to find the SWR.
ring A transient response wherein a signal initially performs a damped oscillation about its
steady-state value.
SAVE function A feature of spectrum analyzers incorporating display storage which enables
them to store displayed spectra.
sensitivity Measure of a spectrum analyzer’s ability to display minimum level signals at a given
IF bandwidth, display mode, and any other influencing factors.
shape factor In spectrum analysis, the ratio of an RBW filter’s 60 dB bandwidth to its 3 dB or 6
dB width (depending on manufacturer).
sideband Signal components observable on either or both sides of a carrier as a result of modu-
lation or distortion processes.
sideband suppression An amplitude modulation technique wherein one of the AM sidebands is
deliberately suppressed, usually to conserve bandwidth.
p y
single sweep Operating mode in which the sweep generator must be reset for each sweep. Espe-
cially useful for obtaining single examples of a signal spectrum.
span per division, Span/Div Frequency difference represented by each major horizontal divi-
sion of the graticule.
spectrum The frequency domain representation of a signal wherein it is represented by display-
ing its frequency distribution.
spurious response An undesired extraneous signal produced by mixing, amplification, or other
signal processing technique.
stability The property of retaining defined electrical characteristics for a prescribed time and in
specified environments.
starting frequency The frequency at the left hand edge of the spectrum analyzer display.
sweep speed, sweep rate The speed or rate, expressed in time per horizontal divisions, at which
the electron beam of the CRT sweeps the screen.
time domain representation Representation of signals by displaying the signal amplitude as a
function of time; typical of oscilloscope and waveform monitor displays.
time-varying signal A signal whose amplitude changes with time.
total span The total width of the displayed spectrum; the span/div times the number of divisions.
tracking generator A signal generator whose output frequency is synchronized to the frequency
being analyzed by the spectrum analyzer.
ultimate rejection The ratio, in dB, between a filter’s passband response and its response beyond
the slopes of its skirts.
vertical scale factor, vertical display factor The number of dB, volts, etc., represented by one
vertical division of a spectrum analyzer display screen.
waveform memory Memory dedicated to storing a digital replica of a spectrum.
waveform subtraction A process wherein a saved waveform can be subtracted from a second,
active waveform.
zero hertz peak A fictitious signal peak occurring at zero Hertz that conveniently marks zero
frequency. The peak is present regardless of whether or not there is an input signal.
zero span A spectrum analyzer mode of operation in which the RBW filter is stationary at the
center frequency; the display is essentially a time domain representation of the signal prop-
agated through the RBW filter.
9.4.5 References
1. Kinley, Harold: “Using Service Monitor/Spectrum Analyzer Combos,” Mobile Radio Tech-
nology, Intertec Publishing, Overland Park, Kan., July 1987.
2. Pepple, Carl: “How to Use a Spectrum Analyzer at the Cell Site,” Cellular Business, Inter-
tec Publishing, Overland Park, Kan., Marcy 1989.
p y
9-58 Test Equipment
3. Wolf, Richard J.: “Spectrum Analyzer Uses for Two-Way Technicians,” Mobile Radio Tech-
nology, Intertec Publishing, Overland Park, Kan., July 1987.
9.4.6 Bibliography
Whitaker, Jerry C. (ed.): Electronic Systems Maintenance Handbook, 2nd ed., CRC Press, Boca
Raton, Fla., 2002.
g g
Chapter
9.5
Reliability Engineering
9.5.1 Introduction
Before the advent of semiconductors, reliability of electronic equipment was an elusive goal.
Racks full of vacuum tubes did not lend themselves to long-term stability and dependable opera-
tion. In actual circuit complexity, the systems were generally simple compared to today's com-
puter-based hardware. However, the primary active devices that made the equipment of 30–40
years ago operate—tubes—were fragile components with a more-or-less limited lifetime.
Enter solid state electronics and the wonders that it has provided. We have been blessed with
increased reliability, increased performance, reduced space requirements, reduced heat genera-
tion, and practical affordable systems that were little more than dreams 30 years ago, or even 20
years ago. However, in our march ahead with technology, engineers have acquired new problems
in the areas of preventive maintenance and troubleshooting.
The electronics industry has grown up in a technical sense. And the point we have reached in
equipment sophistication requires new approaches to reliability and maintainability.
Reliability is the primary concern of any equipment user. The best computer system in the
world is useless if it does not work. The most powerful and sophisticated transmitter is of no
value if it will not stay on the air. Maintainability ranks right behind reliability in concern to pro-
fessional users. When equipment fails, the user needs to know that it can be returned to service
within a reasonable length of time.
The science of reliability and maintainability is not just an esoteric concept developed by the
aerospace industry to satisfy government dictates. It is a science that has fostered the continued
improvements in components and system that we enjoy today.
9.5.1a Objectives
The ultimate goal of any maintenance department is zero downtime. This is an elusive goal, but
one that can be approximated by examining the vulnerable areas of plant operation and taking
steps to prevent a sequence of events that could result in system failure. In cases where failure
prevention is not practical, a reliability assessment should encompass the stocking of spare parts,
circuit boards, or even entire systems. A large facility can often cost-justify the purchase of back-
9-59
y g g
9-60 Test Equipment
up gear that can be used as spares for the entire complex. The cost of backup hardware is expen-
sive, but so is downtime.
Failures can, and do, occur in electronic systems. The goal of product quality assurance at
every step in the manufacturing and operating chain is to ensure that failures do not produce a
systematic or repeatable pattern. The ideal is to eliminate failures altogether. Short of that, the
secondary goal is to end up with a random distribution of failure modes. This situation indicates
that the design of the system is fundamentally correct and that failures are caused by random
events that cannot be predicted. In an imperfect world, this is often the best we can hope for. Reli-
ability and maintainability must be built into products or systems at every step in the design, con-
struction, and maintenance process. It cannot be treated as an afterthought.
9.5.1b Terminology
In order to understand the principles of reliability engineering, the following basic terms must be
defined:
availability Probability of a system subject to repair operating satisfactorily on demand.
average life The mean value for a normal distribution of product or component lives; generally
applied to mechanical failures resulting from “wear-out.”
burn-in Initially high failure rate encountered when first placing a component on test. Burn-in
failures are usually associated with manufacturing defects and the debugging phase of early
service.
defect Any deviation of a unit or product from specified requirements. A unit or product may
contain more than one defect.
degradation failure A failure that results from a gradual change in performance characteristics
of a system or part with time.
downtime Time during which equipment is not capable of doing useful work because of mal-
function. This does not include preventive maintenance time. In other words, downtime is
measured from the occurrence of a malfunction to the correction of that malfunction.
failure A detected cessation of ability to perform a specified function or functions within previ-
ously established limits. It is beyond adjustment by the operator by means of controls nor-
mally accessible during routine operation of the system. (This requires that measurable
limits be established to define “satisfactory performance.”)
failure mode and effects analysis (FMEA) An iterative documented process performed to iden-
tify basic faults at the component level and determine their effects at higher levels of
assembly.
failure rate The rate at which failure occurs during an interval of time as a function of the total
interval length.
fault tree analysis (FTA) An iterative documented process of a systematic nature performed to
identify basic faults, determine their causes and effects, and establish their probabilities of
occurrence.
lot size A specific quantity of similar material or collection of similar units from a common
source; in inspection work, the quantity offered for inspection and acceptance at any one
y g g
time. It may be a collection of raw material, parts, subassemblies inspected during produc-
tion, or a consignment of finished products to be sent out for service.
maintainability The probability that a failure will be repaired within a specified time after a
failure occurs.
mean time to failure (MTTF) The measured operating time of a single piece of equipment
divided by the total number of failures during the measured period of time. This measure-
ment is normally made during that period between early life and wear-out failures.
mean time to repair (MTTR) The measured repair time divided by the total number of failures
of the equipment.
mode of failure The physical description of the manner in which a failure occurs and the operat-
ing condition of the equipment or part at the time of the failure.
part failure rate The rate at which a part fails to perform its intended function.
quality assurance (QA) All those activities, including surveillance, inspection, control, and doc-
umentation aimed at ensuring that the product will meet its performance specifications.
reliability The probability that an item will perform satisfactorily for a specified period of time
under a stated set of use conditions.
reliability growth Action taken to move a hardware item toward its reliability potential, during
development or subsequent manufacturing or operation.
reliability prediction Compiled failure rates for parts, components, subassemblies, assemblies,
and systems. These generic failure rates are used as basic data to predict a value for reli-
ability.
sample One or more units selected at random from a quantity of product to represent that prod-
uct for inspection purposes.
sequential sampling Sampling inspection in which, after each unit is inspected, the decision is
made to accept, reject, or inspect another unit. (Sequential sampling as defined here is
sometimes called unit sequential sampling or multiple sampling.)
system A combination of parts, assemblies and sets joined together to perform a specific opera-
tional function or functions.
test to failure Testing conducted on one or more items until a predetermined number of failures
have been observed. Failures are induced by increasing electrical, mechanical, and/or envi-
ronmental stress levels, usually in contrast to life tests in which failures occur after
extended exposure to predetermined stress levels. A life test can be considered a test-to-
failure using age as the stress.
9.5.2 Quality Assurance

The manufacturers of electronic components and systems design and implement quality assur-
ance procedures for one fundamental reason: nothing is perfect. The goal of a QA program is to
ensure to both the manufacturer and the customer that all but some small mutually acceptable
percentage of devices or systems shipped will be as close to perfection as economics and the
y g g
9-62 Test Equipment
state-of-the-art allow. There are trade-offs in this process. It would be unrealistic, for example, to
perform extensive testing to identify potential failures if the cost of that testing exceeded the cost
savings generated by not having to replace the devices later in the field.
The focus of any QA effort is quality and reliability. These terms are not synonymous. They
are related, but do not provide the same measure of a product:
• Quality is the measure of a product's performance relative to some established criteria.
• Reliability is the measure of a product's life expectancy.
Stated from a different perspective, quality answers the question of whether the product meets
applicable specifications now; reliability answers the question of how long the product will con-
tinue to meet its specifications.
9.5.2a Inspection Process

Quality assurance for components is normally performed through sampling rather than through
100 percent inspection. The primary means used by QA departments for controlling product
quality at the various processing steps include:
• Gates—a mandatory sampling of every lot passing through a critical production stage. Mate-
rial cannot move on to the next operation until QA has inspected and accepted the lot.
• Monitor points—a periodic sampling of some attribute of the component. QA personnel sam-
ple devices at a predetermined frequency to verify that machines and operators are producing
material that meets pre-established criteria.
• Quality audit—an audit carried out by a separate group within the QA department. This
group is charged with ensuring that all production steps throughout the manufacturer's facility
are in accordance with current specifications.
• Statistical quality control—a technique based on advanced computer modeling that incorpo-
rates data accumulated at each gate and monitor point to construct statistical profiles for each
product, operation, and piece of equipment within the plant. Analysis of this data over time
allows quality engineers to assess trends in product performance and failure rates.
Quality assurance for a finished sub-assembly or system may range from a simple go-no test
to a thorough operational check-out that may take days or even weeks to complete. Most manu-
facturers test individual printed wiring boards (PWBs) before they are integrated into a larger
system. Test fixtures of various types provide manual or automated testing of the cards and
alignment of circuits, if required.
9.5.2b Reliability Evaluation

Reliability prediction is the process of quantitatively assessing the reliability of a component or
system during development, before large-scale fabrication and field operation. During product
development, predictions serve as a guide by which design alternatives can be judged for reliabil-
ity. To be effective, the prediction technique must relate engineering variables to reliability vari-
ables.
A prediction of reliability is obtained by determining the reliability of each critical item at the
lowest system level and proceeding through intermediate levels until an estimate of overall reli-
y g g
ability is obtained. This prediction method depends on the availability of accurate evaluation
models that reflect the reliability of lower-level components. Various formal prediction proce-
dures are used, based on theoretical and statistical concepts, including:
• Parts-count method. The parts-count approach to reliability prediction provides an estimate
of reliability based on a count by part type (ICs, transistors, resistors, capacitors, and other
components). This method is useful during the early design stage of a product, when the
amount of available detail is limited. It involves counting the number of components of each
type, multiplying that number by a generic failure rate for each part type, and summing the
products to obtain the failure rate of each functional circuit, subassembly assembly, and/or
block depicted in the system block diagram. This method is useful in the design phase
because it provides rapid estimates of reliability, permitting assessment of the feasibility of a
given concept.
• Stress-analysis method The stress-analysis technique is similar to the parts-count method,
but utilizes a detailed parts model plus calculation of circuit stress values for each part before
determining the failure rate. Each part is evaluated in its electric-circuit and mechanical-
assembly application based on an electrical and thermal stress analysis. After part failure
rates have been established, a combined failure rate for each functional block is determined.
9.5.2c Failure Analysis

Failure mode and effects analysis can be performed with data taken from actual failure modes
observed in the field, or from hypothetical failure modes derived from one of the following:
• Design analysis
• Reliability-prediction activities
• Experience with how specific parts fail
In the most complete form of FMEA, failure modes are identified at the component level.
Failures are analytically induced into each component, and failure effects are evaluated and
noted, including the severity and frequency (or probability) of occurrence. Using this approach,
the probability of various system failures can be calculated, based on the probability of lower-
level failure modes.
Fault-tree analysis is a tool commonly used to analyze failure modes found during design,
factory test, or field operations. The approach involves several steps, including the development
of a detailed logic diagram, which depicts basic faults and events that can lead to system failure
and/or safety hazards. This data is used to formulate corrective suggestions which, when imple-
mented, will eliminate or minimize faults considered critical. An example FTA chart is shown in
Figure 9.5.1.
9.5.2d Standardization
Standardization and reliability go hand-in-hand. Standardization of electronic components began
with military applications in mind; the first recorded work was performed by the U. S. Navy with
vacuum tubes. The Navy recognized that some control at the component level was essential to
the successful incorporation of electronics into Naval systems. The scope of the increase in solid
state electronics within defense hardware is demonstrated dramatically in the evolution of the
y g g
9-64 Test Equipment
Figure 9.5.1 Example fault-tree analysis diagram. (After [1].)
Navy destroyer. In 1937, a destroyer carried less than 100 active devices, almost all of them vac-
uum tubes. A typical destroyer today carries 10 million or more active devices, most of which are
complex integrated circuits (ICs). Standardization is essential to controlling this level of com-
plexity.
Standardization and reliability are closely related, although there are many aspects of stan-
dardization whose reliability implications are subtle. The primary advantages of standardization
include:
• Total product interchangeability. Standardization ensures that products of the same part
number provide the same physical and electrical characteristics. There have been innumerable
instances of replacement devices bearing the same part number as failed devices, but not
functioning identically to them. In many cases, the differences in performance were so great
that the system would not function with the new device.
• Consistency and configuration control. Component manufacturers constantly redefine their
products in order to improve yields and performance. Consistency and configuration control
provide the user with assurance that product changes will not affect the interchangeability of
the part.
• Efficiencies of volume production. Standardization programs usually result in production
efficiencies that reduce the costs of parts, relative to components with the same level of reli-
ability screening and control.
• Effective spares management. Use of standardized components makes the stocking of spare
parts a much easier task. This aspect of standardization is not a minor consideration.
y g g
• Multiple product sources. Standardiza-

tion encourages second-sourcing. Multi-
ple sources helps hold down product
costs, and encourages manufacturers to
strive for better product performance.
9.5.3 Reliability Analysis

The science of reliability and maintainability
matured during the 1960s with the develop-
ment of sophisticated computer systems and Figure 9.5.2 The statistical distribution of equip-
complex military and spacecraft electronics. ment or component failures as a function of time
for electronic devices and systems.
Components and systems never fail without
a reason. That reason may be difficult to
find, however, determination of failure modes and weak areas in system design or installation are
fundamental to increasing the reliability of any component or system, whether it is an integrated
circuit, aircraft autopilot, or broadcast transmitter.
All equipment failures are logical—some are predictable. When a system fails, it is usually
related either to poor quality components or abuse of the system or a part within—either because
of underrating or environmental stress. Even the best-designed components can be badly manu-
factured. A process can go awry, or a step involving operator intervention can result in an occa-
sional device that is substandard or likely to fail under normal stress. Hence, the process of
screening and/or burn-in to weed out problem parts is a universally accepted quality control tool
for achieving high reliability.
9.5.3a Statistical Reliability

Figure 9.5.2 illustrates what is commonly known as the bathtub curve. It divides the expected
lifetime of a class of parts into three segments: infant mortality, useful life, and wear out. A typi-
cal burn-in procedure consists of the following steps:
• The parts are electrically biased and loaded (that is, they are connected in a circuit represent-
ing a typical application).
• The parts are cycled on and off (power applied and then removed) for a pre-determined num-
ber of times. The number of cycles can range from 10 to several thousand during the burn-in
period, depending on the component under test.
• The components under load are exposed to a high temperature (typically 125° to 150°C) for a
selected time (typically 72 to 168 hours). This represents an accelerated life test for the part.
An alternative approach involves temperature shock testing in which the component product
is subjected to temperature extremes, with rapid changes between the hot-soak and cold-soak
conditions. After the stress period, the components are tested for adherence to specifications.
Parts meeting the established specs are accepted for shipment to customers. Parts that fail to
meet specs are discarded.
Figure 9.5.3 illustrates the benefits of temperature cycling to product reliability. The charts
compare the distribution of component failures identified through steady-state high temperature
y g g
9-66 Test Equipment
(a ) (b )
Figure 9.5.3 Distribution of component failures identified through burn-in testing: (a) steady-state
high-temperature burn-in, (b) temperature cycling. (After [2].)
burn-in versus temperature cycling. Note that cycling screened out a significant number of IC
failures. The distribution of failures under temperature cycling usually resembles the distribution
of field failures. Temperature cycling more closely simulates real-world conditions than steady-
state burn-in. The goal of burn-in testing is to ensure that the component lot is advanced past the
infant mortality stage (T-1 on the bathtub curve). This process is used not only for individual
components, but for entire systems.
Such a systems approach to reliability is effective, but not foolproof. The burn-in period is a
function of statistical analysis; there are no absolute guarantees. The natural enemies of elec-
tronic parts are heat, vibration, and excessive voltage. Figure 9.5.4 documents failures-versus-
hours in the field for a piece of avionics equipment. The conclusion is made that a burn-in period
of 200 hours or more will eliminate 60 percent of the expected failures. But, the burn-in period
for another system using different components may well be another number of hours.
So, what does this mean to you, the end-user? Simply that infant mortality is a statistical fact
of life in the operation of any piece of hardware. Most engineers can relate to the problems of
“working the bugs” out of a new piece of equipment.
Because the goal of burn-in testing is to catch system problems and potential faults before the
device or unit leaves the manufacturer, the longer the burn-in period, the greater the likelihood of
catching additional failures. The problems involved with extended burn-in, however, are time
and money. Longer burn-in translates to longer delivery delays and additional costs for the
equipment manufacturer, which are likely to be passed on to the user. The point at which a prod-
uct is shipped is based largely on experience with similar components or systems and the finan-
cial requirement to move products to users.
y g g
Figure 9.5.4 The failure history of a piece of avionics equipment as a function of time. Note that 60
percent of the failures occurred within the first 200 hours of service. (After [3].)
Roller-Coaster Hazard Rate

The bathtub curve has been used for
decades to represent the failure rate of
an electronic system. More recent data,
however, has raised questions regarding
the accuracy of the curve shape. A num-
ber of reliability scientists have sug-
gested that the probability of failure,
known in the trade as the hazard rate, is
more accurately represented as a roller-
coaster track, as illustrated in Figure Figure 9.5.5 The roller-coaster hazard rate curve for
9.5.5. Hazard rate calculations require electronic systems. (After [4].)
analysis of both the number of failures
of the system under test, and the number
of survivors. Experts say that previous estimating processes and averaging tended to smooth the
roller-coaster curve so that the humps were less pronounced, leading to an incorrect conclusion
insofar as the hazard rate is concerned. The testing environment also has a significant effect on
the shape of the hazard curve, as illustrated in Figure 9.5.6. Note that at the higher operating tem-
perature (greater environmental stress), the roller-coaster hump has moved to an earlier age.
y g g
9-68 Test Equipment
Figure 9.5.6 The effects of environmental conditions on the roller-coaster hazard rate curve. (After
[4].)
9.5.3b Environmental Stress Screening

The science of reliability analysis is rooted in the understanding that there is no such thing as a
random failure; every failure has a cause. For reasonably designed and constructed electronic
equipment, failures result from built-in flaws or latent defects. Because different flaws are sensi-
tive to different stresses, a variety of environmental forces must be applied to a unit under test to
identify any latent defects. This is the underlying concept behind environmental stress screening
(ESS).
ESS takes the burn-in process a step further by combining two of the major environmental
factors that cause parts or units to fail: heat and vibration. Qualification testing for products at a
factory practicing ESS involves a carefully planned series of checks for each unit off the assem-
bly line. Units are subjected to random vibration and temperature cycling during production (for
sub-assemblies and discrete components) and upon completion (for systems). The goal is to
catch product defects at the earliest possible stage of production. ESS can also lead to product
improvements in design and manufacture if feedback from the qualification stage to the design
and manufacturing stages is implemented. Figure 9.5.7 details the failure modes typically uncov-
ered by ESS, and Figure 9.5.8 illustrates the improvement in reliability that can typically be
achieved through ESS over simple burn-in screening, and through ESS with feedback to earlier
production stages. Significant reductions in equipment failures in the field can be gained. Table
9.5.1 compares the merits of conventional reliability testing and ESS.
Designing an ESS procedure for a given product is no easy task. The environmental stresses
imposed on the product must be great enough to cause fall-out of marginal components during
qualification testing. The stresses must not be so great, however, as to cause failures in good
products. Any unit that is stressed beyond its design limits will eventually fail. The proper selec-
tion of stress parameters, generally random vibration on a specially designed vibration generator,
and temperature cycling in an environmental chamber, can uncover product flaws in minutes that
y g g
Figure 9.5.7 System failure modes typically uncovered by environmental stress screening. (After
[5].)
Figure 9.5.8 The effects of environmental stress screening on the reliability bathtub curve. (After
[5].)
might take weeks or months to occur in the field. The end result is greater product reliability for
the user.
The ESS concept requires that every single product undergo qualification testing before
implementation into a larger system for shipment to an end-user. The flaws uncovered by ESS
vary from one unit to the next, but types of failures tend to respond to particular environmental
y g g
9-70 Test Equipment
Table 9.5.1 Comparison of Conventional Reliability Testing and Environmental Stress Screening
(After [4].)
Parameter Conventional Testing Environmental Stress Screening

Test condition Simulates operational equipment profile Accelerated stress condition
Test sample size Small 100 percent of production
Total test time Limited Very large
Number of failures Small Large
Reliability growth Potential for gathering useful data small Potential for gathering useful data
very good.
Figure 9.5.9 Integrated circuit screening process and potential results. (After [6].)
stresses. Figure 9.5.9 lists the effects of environmental stresses on integrated circuit elements and
parameters. These data clearly demonstrate that the burn-in screens must match the flaws sought,
or the flaws will probably not be found.
The concept of flaw-stimulus relationships can also be shown in Venn diagram form. Figure
9.5.10 shows a Venn diagram for a hypothetical, but specific product. The required screen would
be different for a different product. For clarity, not all stimuli are shown. Note that there are many
latent defects that will not be uncovered by any one stimulus. For example, a solder splash that is
just barely clinging to a circuit board would probably not be broken loose by high temperature
burn-in or voltage cycling, but vibration or thermal cycling would probably break the particle
loose. Remember also that the defect may only be observable during a stimulation and not
observable during a static bench test.
The levels of stress imposed on a product during ESS should be greater than the stress to
which the product will be subjected during its operational lifetime, while still below the maxi-
mum design parameters. This rule of thumb is pushed to the limits under an enhanced screening
process. Enhanced screening places the component or system at well above the expected field
y g g
Figure 9.5.10 Venn diagram representation of the relationship between flaw precipitation and
applied environmental stress. (After [6].)
environmental levels. This process has been found to be useful and cost effective on many pro-
grams and products. Enhanced screening, however, requires the development of screens that are
carefully identified during product design development so that the product can survive the quali-
fication tests. Enhanced screening techniques are often required for cost effective products on a
cradle to grave basis. That is, early design changes for screenability save tremendous costs over
the lifetime of the product.
The types of products that can be economically checked through ESS break down into two
categories: high-dollar items and mass-produced items. Units that are physically large in size,
such as RF generators or mainframe computers, are difficult to test in the finished state. Still,
qualification tests using more primitive methods, such as cam-driven truck bed shakers, are prac-
tical. Because most large systems generate a large about of heat, subjecting the equipment of
temperature extremes may also be accomplished. Sophisticated ESS for large systems, however,
must rely on qualification testing at the sub-assembly stage.
The basic hardware compliment for an ESS test station includes a thermal chamber shaker
and controller/monitor. A typical test sequence includes 10 minutes exposure to random vibra-
tion, followed by 10 cycles between temperature minimum and maximum. To save time, the two
tests may be performed simultaneously.
9.5.3c Latent Defects

The cumulative failure rate observed during early life of an electronic system is dominated by
the latent defect content of the product, not its inherent failure rate. Product design is the major
determinant of inherent failure rate. A product design will show a higher than expected inherent
rate if the system contains components that are marginally over-stressed, have inadequate func-
tional margin, or contain a sub-population of components that exhibit a wear-out life shorter than
the useful life of the product. Industry has grown to expect the high instantaneous failure rate
observed when a product is first placed into service. The burn-in process, whether ESS or con-
y g g
9-72 Test Equipment
ventional, is aimed at shielding customers from

the detrimental effects of infant mortality. The
key to reducing early product-life failures lies
in reducing the number of latent defects.
A latent defect is some abnormal character-
istic of the product or its parts that will likely
result in failure at some future time depending
on:
• The degree of abnormality
• The magnitude of applied stress
• The duration of applied stress Figure 9.5.11 Estimation of the probable
For example, consider a solder joint on a dou- time-to-failure caused by an abnormal solder
ble-sided PWB. One characteristics of the joint joint. (After [7].)
is the degree to which the plated-through hole
is filled with solder, characterized as “percent fill.” All other characteristics being acceptable, a
joint that is 100 percent filled offers the maximum mechanical strength, minimum resistance,
and greatest current carrying capacity. Conversely, a joint that is zero percent filled has no
mechanical strength, and only if the lead is touching the barrel or pad does it have any significant
electrical properties. Between these two extreme cases are degrees of abnormality. For a fixed
magnitude of applied stress:
• A grossly abnormal solder joint will probably fail in a short time.
• A moderately abnormal solder joint will probably fail in a longer period of time than a grossly
abnormal joint.
• A mildly abnormal solder joint will probably fail in a much longer period of time than either
of the preceding conditions.
Figure 9.5.11 illustrates this concept. A similar time-stress relationship holds for a fixed
degree of abnormality and variable applied stress.
A latent defect will eventually advance to a patent defect when exposed to environmental or
other stimuli. A patent defect is a flaw that has advanced to the point where an abnormality actu-
ally exists. To carry-on the solder example, a cold solder joint represents a flaw (latent defect).
After vibration and/or thermal cycling, the joint will (we assume) crack. The joint would now
have become a detectable (patent) defect. Some latent defects can be stimulated into patent
defects by thermal cycling, some by vibration, and some by voltage cycling. Not all flaws
respond to all stimuli.
There is strong correlation between the total number of physical and functional defects found
per unit of product during the manufacturing process, and the average latent defect content of
shipping product. Product and process design changes aimed at reducing latent defects not only
improve the reliability of shipping product, but also result in substantial manufacturing cost sav-
ings.
y g g
9.5.3d Field Experience

A wealth of research has been conducted by
the aerospace industry into what types of
parts are likely to fail in a given application.
These data provide an important, although at
times contradictory, perspective on what to
look for when a system fails. It is important
to remember that such studies provide only a
snapshot in time for a specific set of circum-
stances. Still, the data yield useful insight
into why systems fail.
U. S. Navy Research
In [8] research conducted by the U.S. Navy
Figure 9.5.12 The distribution of failure modes
cited connectors and cabling as the main cul-
in the field for an avionics system. (After [8].)
prits in system failures (see Figure 9.5.12).
The study concluded that:
• Approximately 60 percent of all avionics failures are the result of high-level components, pri-
marily connectors and cabling.
• 25 percent of all failures are caused by maintenance and test procedures.
• 13 percent are the result of over stress and abuse of system components.
• Less than 2 percent are the result of (IC) failures.
The high percentage of connector and cable related problems will come as no surprise to most
maintenance technicians. Wiring harnesses may be subjected to significant physical abuse under
the best operating conditions.
Collins Avionics Research

In [9] research conducted by Collins Avion-
ics of removal rates for components indi-
cated that the majority of failures were the
result of semiconductors (see Figure 9.5.13).
Specifically, the study showed:
• Microcircuit failures account for 32 per-
cent of device removals.
• Capacitors represent 19 percent of device
removals.
• Transistor failures account for 15 percent
of device removals.
• All other parts account for 34 percent of
device removals. Figure 9.5.13 Component replacement distribu-
tion for avionics hardware. (After [9].)
y g g
9-74 Test Equipment
On the face of it, this data would

contradict the Navy research quoted
previously. Several important points,
however, must be considered:
• Data based on “parts removed” may
or may not provide a true picture of
failure rates. In the course of servic-
ing a piece of equipment, several
parts may be replaced before the sys-
tem is returned to service. Simply
because a component is replaced
does not mean that the component
was defective.
Figure 9.5.14 Failure rate analysis from a General
• Data contained in various technical Electric reliability study. (After [10].)
papers reflects the parts mix of the
system being studied. For example, a
system consisting of several chassis, each including a large number of PWBs, will be more
vulnerable to connector-related failures than a self-contained system with only a few input/
output (I/O) connectors and a single PWB.
• The number of adjustable components may change the failure rates for a class of devices. For
example, a piece of analog equipment with a large number of trimmer pots will likely experi-
ence a greater number of “resistor” failures than a digital system with no trimmer pots.
• The scope of the research must be considered. When a manufacturer examines failure rates
for a given piece of equipment, the research usually will not include cables or assemblies
exterior to the manufacturer's box.
General Electric Research

In [10] research conducted by General Electric of field failures cited PWB-related problems as
the major cause of system failures (see Figure 9.5.14). Conclusions from the research were:
• PWBs and solder joints account for 38 percent of failures.
• Integrated circuits account for 23 percent of failures.
• Resistors account for 11 percent of failures.
• Semiconductors (other than ICs) account for 10 percent of failures.
• All other parts account for 18 percent of failures.
The high failure rate for PWBs was, according to the report, the result of a smear problem in
the production process of the system being examined. A failure rate study for the system after the
production process was corrected would, presumably, result in a different distribution of failures.
Hughes Aircraft Research

In [11] research conducted by Hughes Aircraft showed that field failures were caused primarily
by semiconductors, including discrete diodes, discrete transistors, ICs, and hybrid microcircuits.
y g g
Figure 9.5.15 shows the failure distribu-

tion. General conclusions of the study
were:
• Integrated circuits account for 30 per-
cent of failures.
• Resistors account for 28 percent of fail-
ures.
• Semiconductors (other than ICs)
account for 20 percent of failures.
• Capacitors account for 13 percent of
failures.
• All other components account for 9
percent of failures.
The Hughes report states that solder
joints and PWBs are believed to represent
Figure 9.5.15 Avionics hardware parts replace-
a significant number of failures, although
ment distribution. (After [11].)
hard data were not available to verify the
assumption.
9.5.3e Parts Count Design Tradeoffs

Designing for reliability and circuit efficiency must be applied early in the development stage if
parts count reductions are to result in any significant reduction in system failure rates.
For example, in [6] the authors report that if VLSI chips were used in the F-16 airborne radar
to implement the existing logic without reducing number of circuit boards, the parts count would
be reduced by as much as 85 percent, but reliability would increase by a factor of only 2 percent.
However, if the design was optimized by inserting high-density VLSI chips into a logic imple-
mentation reconfigured to minimize the board count, and correspondingly the connector count,
reliability would be increased by a factor of 7 percent over the baseline.
Semicustom ICs can readily achieve significant reductions in component counts. Reducing
the parts count is important because of fewer explicit components in the design—ICs, resistors,
capacitors, and inductors—but also because of the smaller number of implicit components—
wires, solder joints, and connectors. Every resistor or capacitor adds two solder joints to a cir-
cuit; integrated circuits can add as few as 8 or as many as 200. These joints are prone to manu-
facturing defects, corrosion, and failure from thermal mismatch.
9.5.3f Operating Environment

The operating environment of an electronic system—resulting either from external environmen-
tal conditions or unintentional component underrating—may be significantly more stressful than
expected by either the system manufacturer or the component supplier. Unintentional component
underrating represents a design fault, but unexpected environmental conditions are possible for
many applications, particularly in remote locations.
y g g
9-76 Test Equipment
Conditions of extreme low or high temperatures, high humidity, and vibration during trans-
portation may have a significant impact on long-term reliability of the system. It is, for example,
possible—and even probable—that the vibration stress of the truck ride to a remote transmitting
site will represent the worst-case vibration exposure of the transmitter and all components within
during the lifetime of the product.
Manufacturers report that most of the significant vibration and shock problems for land-
based products arise from the shipping and handling environment. Shipping tends to be an order
of magnitude more severe than the operating environment with respect to vibration and shock.
Early testing for these problems involved simulation of actual shipping and handling events, such
as end-drops, truck trips, side impacts, and rolling over curbs and cobblestones. Although unso-
phisticated by today's standards, these tests were capable of improving product resistance to ship-
ping-induced damage. End-users must realize that the field is the final test environment and
burn-in chamber for the equipment that we use.
9.5.4 Maintenance Actions

The goal of preventive maintenance is to improve reliability. The goal of maintenance is to return
a failed system to operation. Unfortunately, maintenance activities also carry the possibility of
introducing failures into electronic systems. In fact, studies have shown that maintenance-
induced failures are a major factor contributing to the maintenance burden of manhours, materi-
als, and support costs. One analysis [13] showed that on certain electronic equipment, mainte-
nance-induced failures generally average between 20 and 30 percent of all failures, and could run
as high as 48 percent. Another analysis [14] attributed approximately 30 percent of the failed
parts in an aircraft weapons assembly to electrical over stress or other abuse. Electrical over
stress was defined in the report as the inadvertent application of an overvoltage during mainte-
nance or through an improper troubleshooting procedure.
Reliability has an inverse relationship with frequency of maintenance actions. If the MTBF
increases, maintenance actions, including preventive tasks, will be reduced. If reliability can be
improved without significantly increasing the complexity or difficulty of performing mainte-
nance, the overall maintenance support effort will be enhanced.
9.5.4a Maintenance Procedures

Simplification of maintenance procedures has been a goal of both commercial and military hard-
ware for decades. There are many published papers attesting to the improvements in maintain-
ability achieved in modern systems. Unfortunately, any shortcomings in design must be
compensated for by maintenance technicians. Modern equipment packs more parts of greater
complexity into smaller packages. Most problems related to maintainability are discovered after
the product is in the field, and because product improvement programs are expensive, techni-
cians must usually live with the inherent faults of the hardware. Furthermore, most product
enhancement work is designed to improve performance, not maintainability.
Examples abound of systems designed with little regard to maintainability. The best docu-
mented cases can be found with military hardware. One example [15], is particularly interesting:
Each engine on the B-1B bomber has an accessory drive gearbox (ADG). A hinged access
door with four thumb latches is provided on each compartment panel for servicing. The access
y g g
door permits checking of the ADG oil without having to remove the compartment panel. How-
ever, the oil level sight gauge requires line-of-sight reading. The way it is installed, the gauge
cannot be read through the access door, even with an inspection mirror. The entire compartment
panel, secured with 63 fasteners, must be removed just to see if oil servicing is needed.
It is clear that maintainability must play a larger role in the design formula of equipment.
With additional hardware complexity, designers must consider up-front the need to fix a piece of
equipment after it has left the factory.
Impact of New Technologies

Complex integrated circuits dominate the current generation of PWBs, and common repair tech-
niques are inadequate to troubleshoot and repair such boards. New design features, however,
promise to render fault isolation, testing, and component replacement less time consuming. Fea-
tures that will have a favorable impact on the repair of advanced technology PWBs include:
• Design for testability
• Built-in test
• Manufacturing processes such as pin grid array components, surface-mounted devices, and
socketed parts
The use of advanced devices offers tremendous advantages for circuit card assembly (CCA)
design, however, it significantly complicates the ability to test, isolate a fault, and repair the
PWB. The test problems associated with the use of advanced technologies have begun to result in
components and circuits designed for testing by an external analyzer and/or self-test features.
This work is welcomed by technicians.
9.5.4b Reliability-Centered Maintenance

Reliability-centered maintenance (RCM) is a tool used by reliability, safety, and/or maintenance
engineers for developing an optimum maintenance program. RCM defines the requirements and
tasks to be performed in achieving, restoring, or maintaining the operational capability of an
individual piece of equipment or an entire system. The RCM concept was developed during the
1970s by the commercial airline industry. While originally structured to meet a set of specific
needs, RCM may be adapted to a wide range of maintenance-important operations, including
avionics hardware, transportation control systems, and other equipment whose failure could
result in significant property damage or injury.
Implementing RCM requires analysis of failure mode, rate, and criticality data to determine
the most effective maintenance procedures for a given situation. The RCM process is initially
applied during the design/development phase and reapplied, as necessary, during deployment to
sustain an optimum maintenance program based on field experience. RCM identifies two basic
types of components:
• Non safety-critical components. Scheduled maintenance tasks should be performed only
when the task will reduce the life-cycle cost of ownership.
• Safety-critical components. Scheduled maintenance tasks should be performed only when the
tasks will prevent a decrease in reliability and/or deterioration of safety to unacceptable lev-
els.
y g g
9-78 Test Equipment
RCM is based on the premise that reliability is a design characteristic to be realized and pre-
served during the operational life of the system. This concept is further extended to assert that
efficient and cost-effective lifetime maintenance and logistic support programs can be developed
using decision logic that focuses on the consequences of failure. Such a program provides the
following benefits:
• The probability of failure is reduced.
• Incipient failures are detected and corrected either before they occur, or before they develop
into major defects.
• Greater cost-effectiveness of the maintenance program is achieved.
Maintenance Classifications
Under RCM, maintenance tasks are classified into the following areas:
• Hard-time maintenance—failure modes that require scheduled maintenance at predeter-
mined fixed intervals of age or usage.
• On-condition maintenance—failure modes that require scheduled inspections or tests
designed to measure deterioration of an item so that corrective maintenance can be per-
formed.
• Condition monitoring—failure modes that require unscheduled tests or inspection of com-
ponents where failure can be tolerated during operation of the system, or where impending
failure can be detected through routine monitoring during normal operations.
Each step in the RCM process is critical. When properly executed, RCM results in a preven-
tive maintenance program that ensures the system will remain in good working condition. RCM
analysis involves the following key steps:
• Determine the maintenance-important items in the system
• Acquire failure data on the system
• Develop fault-tree analysis data
• Apply decision logic to critical failure modes
• Compile and record maintenance classifications
• Implement the RCM program
• Apply sustaining-engineering based on field experience
The RCM process has a life-cycle perspective. The driving force is reduction of the scheduled
maintenance burden and support cost while maintaining the necessary readiness state. As with
other reliability tasks, the RCM process is reapplied as available data move from a predicted state
to operational experience.
y g g
9.5.5 Software Reliability

The cost of software in most sophisticated
products easily exceeds the cost of the hard-
ware on which it is designed to operate. Soft-
ware costs consist of:
• Identifying the requirements of the pro-
gram
• Writing the program
• Debugging the program prior to delivery
• Follow-up maintenance and program
updates
Figure 9.5.16 An estimation of the percentages
• Customer technical support of software errors that occur in each develop-
ment cycle for certain military electronics sys-
Any discussion of software reliability
tems. (After [16].)
must include a definition of what constitutes
a failure. Software is reliable if it does what
the customer intended it to do whenever it is called upon to perform. The software fails if it does
not deliver this promise. Statistics reported in [16] show that, for military systems, the largest
problem in writing software is identifying exactly what the software is supposed to do. As shown
in Figure 9.5.16, “requirements analysis” constitutes more than half of the software errors that
occur during the development cycle of a major project. It follows that the up-front design stage
offers the greatest opportunity to impact software reliability. The areas of software development
presenting the greatest problems can be divided into the nine categories shown in Figure 9.5.17.
Here again, translation of customer requirements is the source of most issues.
Even though the major software errors are made prior to integration testing, statistics from the
Air Force [17] show that at least 76 percent of those errors are not discovered and corrected until
integration and beyond. Figure 9.5.18 illustrates the impact of this delayed discovery on project
costs. The relative cost of error detection
and resolution during system integration
and beyond can be 10–15 times greater
that the cost of early resolution.
It is fair to point out that the specific
conclusions of studies such as those cited
in this section will change depending on
the technologies involved and the ultimate
applications. Still, the important observa-
tion to be drawn from these studies is the
wide variety of elements that go into the
general category of software reliability.
9.5.5a Hardware Versus Software

Figure 9.5.17 An estimation of the root causes of
The distinctions between hardware and software errors discovered in certain military elec-
software are numerous and obvious. The tronic systems. (After [16].)
y g g
9-80 Test Equipment
Figure 9.5.18 The costs of correcting software errors at key points in the development cycle.
(After [16].)
differences arise from the fact that a software program is essentially an abstraction; it is immune
from mechanical defects, electrical noise, and physical degradation that affect hardware. As both
technologies advance, however, the distinctions between hardware and software are beginning to
lessen. For example, system designers freely trade hardware complexity for software simplicity,
and vice-versa
Hardware reliability is concerned with latent defects, physical design, and the operating envi-
ronment. Software reliability is pure logic; it has no problems analogous with wear out. Software
is also used to mask hardware failures. Fault-tolerant systems are charged with maintaining sys-
tem integrity in the face of a hardware malfunction. Software failures for fault-tolerant systems
can be categorized into one or more of the following categories:
• Failure to perform functions required under normal operating conditions.
• Failure to perform functions required under abnormal conditions. Such conditions could
include overload, invalid input, or another abnormal operating state.
• Failure to recover from a hardware failure.
• Failure to diagnose a hardware failure.
Fault-tolerant control systems with duplicate critical parts are expected to function despite
any single hardware problem. Enumeration of all possible faults is a staggering task. Systems
that operate in real-time, and subject to quasi-random inputs can be driven into internal states
beyond classification. Software design and reliability testing for a fault-tolerant system is no
easy matter.
Because software failures come with greatly varying levels of importance, it is necessary to
grade them in steps ranging from fatal flaws to cosmetic problems. The density of software bugs
must also be related to the likely rate at which the defects will create actual failures under field
conditions. A quality standard for software balances the engineer's desire to remove all possible
defects with the need to deliver products within a reasonable timeframe, and at a reasonable price
y g g
point. Under conditions in which the test environment is a good model of the actual field envi-
ronment, and the software code is exercised completely, the “find-and-fix” process can be used
as a reliable guide to software quality. Under these conditions, the rate at which bugs are found is
roughly proportional to the density of bugs in the code.
Software run on critical systems is often checked by purposely overloading the system. Rea-
son tells us that problems can be expected to occur under the stresses that are present when
queues reach their limits, registers go into overflow, and transactions are delayed. Such abnormal
behavior is likely to exercise “contingency software” not used under normal conditions.
Distinguishing between hardware and software problems is often a difficult task. Problems
manifest themselves at the system level and sometimes may be remedied by changes in either the
software or the hardware. Because software is more flexible, the easiest and most economical
way to fix a problem is usually through a software change. With this point in mind, it is possible
that the perceived contribution of software to system unreliability may be artificially high.
Causal Analysis
In the general application, causal analysis attempts to determine the cause of some observed
phenomena. In software development, causal analysis seeks to explain how faults have been
injected into the program, and how they have led to system or software failures. Once a fault is
identified, programmers must then determine at what point in the process the fault was injected,
and how it can be prevented in future program development.
Failure Mode, Effect, and Criticality Analysis (FMECA)

FMECA is an analysis tool that examines the programming code from a contingent perspective
to determine how the software will behave when presented with unusual input or environmental
conditions. By examining operation of the software under errant conditions, the robustness of the
code can be greatly enhanced. This process also leads the designer to mitigate the possible
effects of failure that the system can be reasonably expected to encounter. FMECA must be
applied at the beginning of a software program design effort. Considerations include the possi-
bility of operator errors or hardware errors, such as the failure of a communications system if the
device is linked to other hardware.
Fault injection is the physical realization of FMECA. Faults are purposely introduced into a
system to see how it reacts and handles the condition. This stresses the fault-tolerant and fault-
recovery aspects of the design. The faults can either be put into the software or the hardware.
This process forces designers to look at how their systems will behave under harsh environmen-
tal conditions.
9.5.5b Improving Software Reliability

Historically, the way to improve software reliability has been to extensively test the program
prior to delivery to the customer. Errors found during the testing process are then corrected.
There is, however, a better way. The key lesson learned from the science of hardware reliability is
also applicable to software: reliability must be designed into the product. It is not cost effective
to test reliability into a program. Every time an error is corrected during testing, the possibility
exists for introducing new errors into the program. In addition, it is much more difficult to find,
correct and document errors in later stages of the development cycle.
y g g
9-82 Test Equipment
Attempts to predict software reliability have been made by mathematicians utilizing probabil-
ity and statistics tools. However, most models have fallen short of the goal of developing a soft-
ware equivalent of common hardware models. Still, the basic concepts employed in hardware
reliability assurance can successfully be applied to software. Reliability engineering concepts
and tools may, if applied within the software development process, directly improve the reliabil-
ity of the end-product.
Software Maintainability
The maintainability of software is determined by those characteristics of the code and computer
support resources that affect the ability of software programmers and analysts to change instruc-
tions. Such changes are made to:
• Correct program errors
• Expand system capabilities
• Add or remove specific features from the program
• Modify software to make it compatible with hardware changes
Modern software programs are usually constructed of modules that perform specific func-
tions. Modular construction is achieved by dividing total program functions into a series of ele-
ments that each have a single input (entry point) and a single output (exit point). The
development of a large system may permit the use of a given module a number of times, simpli-
fying software design and de-bugging. Modules performing a specific function, such as I/O
management or display control, may also be used on other product lines.
The top-level modules of a software program provide functional control over the lower-level
modules on which they call. The top-level software, therefore, is exercised more frequently than
lower-level modules. Various estimates of run-time distributions show that just 4 to 10 percent of
the total software package operates for 50 to 90 percent of the time in a real-time program.
9.5.5c Software System Safety

Software is not intrinsically hazardous. However, when software is integrated with hardware that
performs critical functions, the potential for catastrophic effects is present. For a hazard to exist,
there must be some undesirable or unexpected action possible from the system. A failure of the
hardware or hardware/software interaction may cause the hazard to occur. Normally only the
software that exercises direct control over the hardware, or software that monitors the hardware,
is considered safety critical. Software hazards may be categorized as follows:
• An inadvertent or unwanted event occurrence
• An out-of-sequence event occurrence
• Failure of an event to occur
• The magnitude or direction of the event is lacking or nonexistent
Reliability and safety engineering play a major role in the system design phase of critical soft-
ware. Fault tree analysis is used to ensure that failure modes are identified and are designed-out
of the hardware/software. Fault tree analysis may also be used to assess the potential system haz-
y g g
ards. Even under the most thorough testing program, it is virtually impossible to foresee and test
every conceivable software condition in a major system. The goal of software safety, therefore, is
to identify all potential hazards and ensure that they have either been eliminated or reduced to an
acceptable level of risk.
9.5.6 References
1. Fink, D., and D. Christiansen (eds.): Electronics Engineers' Handbook, 3rd ed., McGraw-
2. Powell, Richard: “Temperature Cycling vs. Steady-State Burn-In,” Circuits Manufacturing,
Benwill Publishing, September 1976.
3. Capitano, J., and J. Feinstein: “Environmental Stress Screening Demonstrates its Value in
the Field,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New York,
N.Y., 1986.
4. Wong, Kam L.: “Demonstrating Reliability and Reliability Growth with Environmental
Stress Screening Data,” Proceedings IEEE Reliability and Maintainability Symposium,
IEEE, New York, N.Y., 1990.
5. Tustin, Wayne: “Recipe for Reliability: Shake and Bake,” IEEE Spectrum, IEEE, New
York, N.Y., December 1986.
6. Hobbs, Gregg K.: “Development of Stress Screens,” Proceedings IEEE Reliability and
7. Smith, William B.: “Integrated Product and Process Design to Achieve High Reliability in
Both Early and Useful Life of the Product,” Proceedings IEEE Reliability and Maintain-
ability Symposium, IEEE, New York, N.Y., 1987.
8. Clark, Richard J.: “Electronic Packaging and Interconnect Technology Working Group
Report (IDA/OSD R&M Study),” IDA Record Document D-39, August 1983.
9. Wilson, M. F., and M. L. Woodruff: “Economic Benefits of Parts Quality Knowledge,”
1985.
10. Griffin, P.: “Analysis of the F/A-18 Hornet Flight Control Computer Field Mean Time
Between Failure,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE,
11. Wong, K. L., I. Quart, L. Kallis, and A. H. Burkhard: “Culprits Causing Avionics Equip-
ment Failures,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New
York, N.Y., 1987.
12. Horn, R., and F. Hall: “Maintenance Centered Reliability,” Proceedings IEEE Reliability
and Maintainability Symposium, IEEE, New York, N.Y., 1983.
13. Maynard, Eqbert, OUSDRE Working Group Chairman: “VHSIC Technology Working
Group Report” (IDA/OSD R&M Study), Document D-42, Institute of Defense Analysis,
November 1983.
y g g
9-84 Test Equipment
14. Robinson, D., and S. Sauve: “Analysis of Failed Parts on Naval Avionic Systems,” Report
No. D180-22840-1, Boeing Company, Seattle, Wash., October 1977.
15. Worm, Charles M.: “The Real World: A Maintainer's View,” Proceedings IEEE Reliability
and Maintainability Symposium, IEEE, New York, N.Y., 1987.
16. Babel, Philip S.: “Software Development Integrity Program.” Briefing paper for the Aero-
nautical Systems Division, Air Force Systems Command. From Yates, W., and Shaller, D.:
“Reliability Engineering as Applied to Software,” Proceedings IEEE Reliability and Main-
17. Hermason, Sue E., Major, USAF: letter dated December 2, 1988. From Yates, W., and
Shaller, D.: “Reliability Engineering as Applied to Software,” Proceedings IEEE Reliabil-
9.5.7 Bibliography
Brauer, D. C., and G. D. Brauer: “Reliability-Centered Maintenance,” Proceedings IEEE Reli-
ability and Maintainability Symposium, IEEE, New York, N.Y., 1987.
Cameron, D., and R. Walker: “Run-In Strategy for Electronic Assemblies,” Proceedings IEEE
DesPlas, Edward: “Reliability in the Manufacturing Cycle,” Proceedings IEEE Reliability and
Devaney, John: “Piece Parts ESS in Lieu of Destructive Physical Analysis,” Proceedings IEEE
Doyle, Edgar, Jr.: “How Parts Fail,” IEEE Spectrum, IEEE, New York, N.Y., October 1981.
Ferrara, K. C., S. J. Keene, and C. Lane: “Software Reliability from a System Perspective,” Pro-
ceedings IEEE Reliability and Maintainability Symposium, IEEE, New York, N.Y., 1988.
Fortna, H., R. Zavada, and T. Warren: “An Integrated Analytic Approach for Reliability Improve-
ment,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE, New York,
N.Y., 1990.
Hall, F., R. A. Paul, and W. E. Snow: “R&M Engineering for Off-the-Shelf Critical Software,”
1988.
Hansen, M. D., and R. L. Watts: “Software System Safety and Reliability,” Proceedings IEEE
Irland, Edwin A.: “Assuring Quality and Reliability of Complex Electronic Systems: Hardware
and Software,” Proceedings of the IEEE, vol. 76, no. 1, IEEE, New York, N.Y., January
1988.
Kenett, R., and M. Pollak: “A Semi-Parametric Approach to Testing for Reliability Growth, with
Application to Software Systems,” IEEE Transactions on Reliability, IEEE, New York,
N.Y., August 1986.
y g g
Neubauer, R. E., and W. C. Laird: “Impact of New Technology on Repair,” Proceedings IEEE
Smith, A., R. Vasudevan, R. Matteson, and J. Gaertner: “Enhancing Plant Preventative Mainte-
nance via RCM,” Proceedings IEEE Reliability and Maintainability Symposium, IEEE,
Spradlin, B. C.: “Reliability Growth Measurement Applied to ESS,” Proceedings IEEE Reliabil-
Whitaker, Jerry C.: Maintaining Electronic Systems Handbook, CRC Press, Boca Raton, Fla.,
2001.
Yates, W., and D. Shaller: “Reliability Engineering as Applied to Software,” Proceedings IEEE
y g g
g g
Section
Audio Test and Measurement

10
Measurements are made on audio equipment to check performance under specified conditions
and to assess suitability for use in a particular application. They may be used to verify specified
system performance or as a way of comparing several pieces of equipment for use in a system.
Measurements may also be used to identify components in need of adjustment or repair. What-
ever the application, audio measurements are an important part of audio engineering.
This section focuses on the basic audio test procedures. Digital technology has added a new
layer of complexity with regard to audio test and measurement. Still, for many applications, ana-
log methods provide familiar, reliable tools for confirming system performance.
In This Section:
Chapter 10.1: Audio Measurement and Analysis 10-5

Introduction 10-5
Level Measurements 10-6
Root-Mean-Square Technique 10-6
Average-Response Measurements 10-7
Peak-Response Measurements 10-7
Types of Meters 10-10
Decibel (dB) Measurements 10-12
Referenced Measurements 10-13
Crosstalk and Separation Measurements 10-13
Noise Measurements 10-15
Noise Analysis 10-15
Bibliography 10-17
Chapter 10.2: Audio Phase and Frequency Measurement 10-19

Introduction 10-19
Phase Measurement 10-19
Relation to Frequency 10-20
Frequency Measurement 10-21
Measurement Techniques 10-22
10-1
10-2 Section Ten

Fast Fourier Transform Measurement 10-26
References 10-30
Bibliography 10-30
Chapter 10.3: Nonlinear Audio Distortion 10-33

Introduction 10-33
Harmonic Distortion 10-33
Intermodulation Distortion 10-38
Specialized IM Tests 10-41
Distortion-Measurement Hardware 10-42
Addition and Cancellation of Distortion 10-43
Signal-Source Effects in Measurements 10-44
References 10-46
Bibliography 10-47
Chapter 10.4: Time Domain Audio Measurements 10-49

Introduction 10-49
Time Domain Characterization 10-49
Level-Gated Burst Measurements 10-52
Impulse Response 10-54
References 10-54
Bibliography 10-55

Bauman, P., S. Lipshitz, and J. Vanderkooy: “Cepstral Techniques for Transducer Measurement:
Part II,” Audio Engineering Society preprint 2302, AES, New York, N.Y., October 1985.
Berman, J. M., and L. R. Fincham: “The Application of Digital Techniques to the Measurement
of Loudspeakers,” J. Audio Eng. Soc., AES, New York, N.Y., vol. 25, June 1977.
Cabot, R. C.: “Limiters, Compressors, and Expanders,” Sound & Video Contractor, Intertec Pub-
lishing, Overland Park, Kan., vol. 26, November 1985.
Cabot, R. C.: “Measurement of Audio Signal Slew Rate,” Audio Engineering Society preprint
1414, AES, New York, N.Y., November 1978.
Engelson, M., and F. Telewski: Spectrum Analyzer Theory and Applications, Artech House, Nor-
wood, Mass., 1974.
Hofer, B.: “Practical Extended Range DIM Measurements,” Audio Engineering preprint 2334,
AES, New York, N.Y., March 1986.
Ladegaard, P.: “Swept Distortion Measurements—An Effective Tool for Revealing TIM in
Amplifiers with Good Subjective Correlation to Subjective Evaluation,” B&K Application
Note 17-234, B&K Instruments, Naerum, Denmark, 1977.
Leinonen, F., M. Otala, and J. Curl: “Method for Measuring Transient Intermodulation Distor-
tion,” Audio Engineering Society preprint 1185, AES, New York, N.Y., October 1976.
Audio Test and Measurement 10-3
Linkwitz, S.: “Narrow Band Testing of Acoustical Systems,” Audio Engineering Society pre-
print 1342, AES, New York, N.Y., May 1978.
Lipshitz, S., T. Scott, and J. Vanderkooy: “Increasing the Audio Measurement Capability of FET
Analyzers by Microcomputer Post-Processing,” Audio Engineering Society preprint 2050,
AES, New York, N.Y., October 1983.
Metzler, R. E.: “Automated Audio Testing,” Studio Sound, August 1985.
Metzler, R. E., and B. Hofer: “Wow and Flutter Measurements,” Audio Precision 1 Users' Man-
ual, Audio Precision, Beaverton, Ore., July 1986.
Moller, H.: “Electroacoustic Measurements,” B&K Application Note 16-035, B&K Instruments,
Naerum, Denmark.
Moller, H., and C. Thompsen: “Swept Electroacoustic Measurements of Harmonic Difference
Frequency and Intermodulation Distortion,” B&K Application Note 15-098, B&K Instru-
ments, Naerum, Denmark.
Otala, M., and E. Leinonen: “The Theory of Transient Intermodulation Distortion,” IEEE Trans.
Acoust. Speech Signal Process., ASSP-25(1), February 1977.
Preis, D.: “A Catalog of Frequency and Transient Responses,” J. Audio Eng. Soc., AES, New
York, N.Y., vol. 24, June 1976.
Ramirez, R.: “The FFT—Fundamentals and Concepts,” Tektronix, Inc., Beaverton, Ore., 1975.
Randall, R. B.: Application of B&K Equipment to Frequency Analysis, 2nd ed., B&K Instru-
ments, Naerum, Denmark, 1977.
Schrock, C.: “The Tektronix Cookbook of Standard Audio Measurements,” Tektronix Inc., Bea-
verton, Ore., 1975.
Skritek, P.: “Simplified Measurement of Squarewave/Sinewave and Related Distortion Test
Methods,” Audio Engineering Society preprint 2195, AES, New York, N.Y., 1985.
Small, R.: “Total Difference Frequency Distortion: Practical Measurements,” J. Audio Eng. Soc.,
AES, New York, N.Y., vol. 34, no. 6, pg. 427, June 1986.
Theile, A. N.: “Measurement of Nonlinear Distortion in a Bandlimited System,” J. Audio Eng.
Soc., AES, New York, N.Y., vol. 31, pp. 443–445, June 1983.
Tremaine, H. W.: Audio Cyclopedia, Howard W. Sams, Indianapolis, Ind., 1975.
Vanderkooy, J.: “Another Approach to Time Delay Spectrometry,” Audio Engineering Society
preprint 2285, AES, New York, N.Y., October 1985.
g g
Chapter
10.1
Audio Measurement and Analysis
Dr. Richard C. Cabot, P.E.
Bruce Hofer, Robert Metzler
10.1.1 Introduction
Many parameters are important in audio devices and merit attention in the measurement process.
Some common audio measurements are frequency response, gain or loss, harmonic distortion,
intermodulation distortion, noise level, phase response, and transient response. There are other,
equally important tests too numerous to list here. This chapter will explain the basics of these
measurements, describe how they are made, and give some examples of their application.
Most measurements in audio (and other fields) are composed of measurements of fundamen-
tal parameters. These parameters include signal level, phase, and frequency. Most other measure-
ments consist of measuring these fundamental parameters and displaying the results in
combination by using some convenient format. For example, signal-to-noise ratio (S/N) is a pair
of level measurements made under different conditions expressed as a logarithmic, or decibel
(dB), ratio.
When characterizing an audio device, it is common to view it as a box with input terminals
and output terminals. In normal use an audio signal is applied to the input, and the audio signal,
modified in some way, appears at the output. In the case of an equalizer the modification to the
signal is an intentional change in the gain with frequency (frequency response). Often it is
desired to know or verify the details of this gain change. This is accomplished by measurement.
Real-world behavior being what it is, audio devices will also modify other parameters of the
audio signal which should have been left alone. To quantify these unintentional changes to the
signal we again turn to measurements. Using the earlier example of an equalizer, changes to the
amplitude-versus-frequency response of the signal inevitably bring changes in phase versus fre-
quency. Some measurements are what are known as one-port measurements such as impedance
or noise level. These are not concerned with both input and output signals, only with one or the
other.
Measurement of level is fundamental to most audio specifications. Level can be measured
either in absolute terms or in relative terms. Power output is an example of an absolute level mea-
surement; it does not require any reference. S/R and gain or loss are examples of relative, or ratio
10-5
y
10-6 Audio Test and Measurement
measurements; the result is expressed as a ratio of two measurements. Though it may not appear
so at first, frequency response is also a relative measurement. It expresses the gain of the device
under test as a function of frequency, with the midband gain as a reference.
Distortion measurements are a way of quantifying the amount of unwanted components
added to a signal by a piece of equipment. The most common technique is total harmonic distor-
tion (THD), but others are often used. Distortion measurements express the amount of unwanted
signal components relative to the desired signal, usually as a percentage or decibel value. This is
also an example of multiple level measurements that are combined to give a new measurement
figure.
10.1.2 Level Measurements

The simplest definition of a level measurement is the alternating-current (ac) amplitude at a par-
ticular place in a system. However, in contrast to direct-current (dc) measurements, there are
many ways of specifying ac voltage. The most common methods are average response, root-
mean-square (rms) response, and peak response. Strictly speaking, the term level refers to a log-
arithmic, or decibel, measurement. However, common parlance employs the term for any ac
amplitude measurement, and that convention will be followed here.
10.1.2a Root-Mean-Square Technique

The rms technique measures the effective power of the ac signal. It specifies the value of the dc
equivalent which would dissipate the same power if either were applied to a load resistor. This
process is illustrated in Figure 10.1.1 for voltage measurements. The input signal is squared, and
the average value is found. This is equivalent to finding the average power. The square root of
this value is taken to get the signal from a power value back to a voltage. For the case of a sine
wave, the rms value is 0.707 of its maximum value.
Suppose that the signal is no longer a sine wave but rather a sine wave and several of its har-
monics. If the rms amplitude of each harmonic is measured individually and added, the resulting
value will be the same as an rms measurement on the signals together. Because rms voltages can-
not be added directly, it is necessary to perform an rms addition. This is illustrated in Equation
(10.1.1). Each voltage is squared, and the squared values are added. The square root of the result-
ing value is taken, yielding the rms voltage of their combination.
V r ms = V2 +V 2…
+V 2
(10.1.1)
total
rms 1 rms 2 rms n
Note that the result is not dependent on the phase relationship of the signal and its harmonics.
The rms value is determined completely by the amplitude of the components. This mathematical
predictability is very powerful in practical applications of level measurement, enabling measure-
ments made at different places in a system to be correlated. It is also extremely important in cor-
relating measurements with theoretical calculations.
An interesting result for gaussian random noise is that the rms value equals the standard devi-
ation of the amplitude distribution. In fact, if the amplitude distribution of any signal is plotted,
the standard deviation of the distribution is, by definition, the rms value.
y
(a)
(b)
Figure 10.1.1 Root-mean-square (rms) measurements: (a) relationship of rms and average val-
ues, (b) the rms measurement circuit.
10.1.2b Average-Response Measurements

Historically, average-responding voltmeters were common in audio, mainly because of their low
cost. They measure ac voltage by rectifying it and filtering the resulting waveform to its average
value as shown in Figure 10.1.2. This results in a dc voltage that can be read on a standard dc
voltmeter. As shown in the figure, the average value of a sine wave is 0.637 of its maximum
amplitude. Average-responding meters are usually calibrated to read the same as an rms meter
for the case of a single-sine-wave signal. This results in the measurement being scaled by a con-
stant K of 0.707/0.637, or 1.11. Meters of this type are called average-responding, rms cali-
brated. For signals other than sine waves, the response will be different and hard to predict. If
multiple sine waves are applied, the reading will depend on the phase shift between the compo-
nents and will no longer match the rms measurement. A comparison of rms and average-
response measurements is made in Figure 10.1.3 for various waveforms. If the average readings
are adjusted as described above to make the average and rms values equal for a sine wave, all the
numbers in the average column should be increased by 11.1 percent, while the rms-average num-
bers should be reduced by 11.1 percent.
10.1.2c Peak-Response Measurements

Peak-responding meters measure the maximum value that the ac signal reaches as a function of
time (Figure 10.1.4). The signal is full-wave-rectified to find its absolute value and passed
through a diode to a storage capacitor. When the absolute value of the voltage rises above the
value stored on the capacitor, the diode will conduct and increase the stored voltage. When the
voltage decreases, the capacitor will maintain the old value. Some means for discharging the
capacitor is required to allow measuring a new peak value. In a true peak detector, this is a
y
(a)
(b)
Figure 10.1.2 Average measurements: (a) illustration of average detection, (b) average-measure-
ment circuit.
switch. Common peak detectors usually use a large resistor to discharge the capacitor gradually
after the user has had a chance to read the meter.
The ratio of the true peak value to the rms value is called the crest factor. As can be seen from
Figure 10.1.5, this is a measure of the peakedness of the signal. For any signal but an ideal square
wave the crest factor will be greater than 1. A comparison of crest-factor values for various
waveforms is shown in Figure 10.1.5. As the signals become more peaked, the crest factor will
increase. The q parameter in the gaussian-noise example is a measure of the percent of time dur-
ing which the signal is greater than the rms value. True gaussian noise has infinitely large peaks
and therefore an infinite crest factor.
By introducing a controlled charge and discharge time with resistors, a quasi-peak detector is
achieved. These charge and discharge times are selected to simulate the ear's sensitivity to impul-
sive peaks. International standards define these response times and set requirements for reading
accuracy on pulses and sine-wave bursts of various durations. Quasi-peak detectors normally
have their gain adjusted so that they read the same as an rms detector for sine waves.
Another method of specifying signal amplitude is called peak-equivalent sine. It is the rms
level of a sine wave having the same peak-to-peak amplitude as the signal under consideration.
This is the peak value of the waveform scaled by the correction factor 1.414, corresponding to
the peak-to-rms ratio of a sine wave. This is useful when specifying test levels of waveforms in
distortion measurements. If the distortion of a device is measured as a function of amplitude, a
point will be reached where the output level cannot increase any further. At this point the peaks
of the waveform will be clipped, and the distortion will rise rapidly with further increases in
level. If another signal is used for distortion testing on the same device, it is desirable that the
levels at which clipping is reached correspond. Signal generators are normally calibrated in this
way to allow changing between waveforms without clipping or readjusting levels.
y
Figure 10.1.3 Comparison of rms and average characteristics. (Courtesy of EDN, January 20, 1982,)
y
(a)
(b)
Figure 10.1.4 Peak measurements: (a) illustration of peak detection, (b) peak-measurement cir-
cuit.
10.1.2d Types of Meters

Most meters measure the true rms value of
the waveform by using integrated circuits
designed for this task. This allows accurate
measurements of voltage for all signals, not
just sine waves. As mentioned earlier, rms
measurements accurately reflect the heating
power of the waveform in a resistor or a
loudspeaker. This is critical to many mea-
surements, such as specification of power in Figure 10.1.5 Definition of the crest factor.
amplifiers. However, many noise specifica-
tions were developed in the days of average-
responding meters, and verifying these requires an average-responding unit. Therefore, good-
quality audio test equipment allows selection between these two responses, giving compatibility
with old and new techniques.
One obvious difference between various audio voltmeters is the type of display, analog or dig-
ital. Each type has its advantages. Analog meters are not easy to use for exact measurements
because of their multiple scales and the need to interpolate numbers from the printed scale. In
contrast, a digital meter gives a direct readout of the value to more digits of precision than could
ever be obtained from an analog-meter scale. This enables very precise measurement of gain and
output power with little chance for error.
However, nothing is perfect. The digital meter is suited only for measuring relatively stable
signals. When trying to monitor program level to determine system operating levels under actual
use, it becomes very difficult to extract a single number from the mass of flashing digits. The
y
Figure 10.1.6 Effect of bandwidth of measurements.
analog meter can handle this job with ease. Another application for analog meters is monitoring
the results of an adjustment for a peak or a null. Some manufacturers have put both analog and
digital displays on the same instrument to enable the best of both worlds. The analog scale on
such meters typically does not have very fine graduations on it and is intended only for approxi-
mate measurements of rapidly changing signals. Other digital instruments provide a simulated
analog display using bar graphs or other means.
The bandwidth of a voltmeter can have a significant effect on the accuracy of the reading. For
a meter with a single-pole rolloff (i.e., one bandwidth-limiting component in the signal path),
significant errors can occur in measurements. For such a meter with a specified bandwidth of
100 kHz, there will be a 10 percent error in measurements of signal at 50 kHz. To obtain 1 per-
cent accurate measurements (with other error sources in the meter ignored), the signal frequency
must be less than 10 kHz.
Another problem with limited-bandwidth measuring devices is shown in Figure 10.1.6. Here,
is a distorted sine wave is being measured by two meters with different bandwidths. The meter
with the narrower bandwidth does not respond to all the harmonics and gives a lower reading.
The severity of this effect varies with the frequency being measured and the bandwidth of the
meter; it can be especially severe when measuring wideband noise. Most audio requirements are
adequately served by a meter with a 500-kHz bandwidth. This allows reasonably accurate mea-
surement of signals to about 100 kHz. Peak measurements are even more sensitive to bandwidth
effects. Systems with restricted low-frequency bandwidth will produce tilt in a square wave, and
bumps in the high-frequency response will produce an overshoot. The effect of either situation
will be an increase in the peak reading.
Accuracy is a measure of how well a meter measures a signal at a midband frequency, usually
1 kHz. This sets a basic limit on the performance of the meter in establishing the absolute ampli-
tude of a signal. It is also important to look at the flatness specification to see how well this per-
formance is maintained with changes in frequency. The flatness specification describes how well
the measurements at any other frequency will track those at 1 kHz. If a meter has an accuracy of
2 percent at 1 kHz and a flatness of 1 dB (10 percent) from 20 Hz to 20 kHz, the inaccuracy can
be as wide as 12 percent at 20 kHz.
Meters often have a specification on accuracy that changes with voltage range, being most
accurate only in the range in which they were calibrated. A meter with 1 percent accuracy on the
2-V range and 1 percent accuracy per step would be 3 percent accurate on the 200-V scale. By
using the flatness specification given previously, the overall accuracy for a 100-V 20-kHz sine
y
wave is 14 percent. In many meters an additional accuracy derating is given for readings as a per-
centage of full scale, making readings at less than full scale less accurate.
However, the accuracy specification is not normally as important as the flatness. When per-
forming frequency response or gain measurements, the results are relative and are not affected by
the absolute voltage used. When measuring gain, however, the attenuator accuracy of the instru-
ment is a direct error source. Similar comments apply to the accuracy and flatness specifications
for signal generators. Most are specified in the same manner as voltmeters, with the inaccuracies
adding in much the same way.
10.1.3 Decibel (dB) Measurements

Measurements in audio work are often expressed in decibels. Audio signals span a wide range of
level. The sound pressure of a rock-and-roll band is about 1 million times that of rustling leaves.
This range is too wide to be accommodated on a linear scale. The decibel is a logarithmic unit
which compresses this wide range down to a more easily handled range. Order-of-magnitude
(factor-of-10) changes result in equal increments on a decibel scale. Furthermore, the human ear
perceives changes in amplitude on a logarithmic basis, making measurements with the decibel
scale reflect audibility more accurately.
A decibel may be defined as the logarithmic ratio of two power measurements or as the loga-
rithmic ratio of two voltages. Equations (10.1.2) and (10.1.3) define the decibel for both power
and voltage measurements.
E
dB = 20 log ----1- (10.1.2)
E2
P
dB = 10 log ----1- (10.1.3)
P2
There is no difference between decibel values from power measurements and decibel values
from voltage measurements if the impedances are equal. In both equations the denominator vari-
able is usually a stated reference. This is illustrated with an example in Figure 10.1.7. Whether
the decibel value is computed from the
power-based equation or from the voltage-
based equation, the same result is obtained.
A doubling of voltage will yield a value
of 6.02 dB, while a doubling of power will
yield 3.01 dB. This is true because doubling
voltage results in a factor-of-4 increase in
power. Table 10.1.1 shows the decibel values
for some common voltage and power ratios.
These are handy to commit to memory, and
they make quick comparisons of readings
especially easy. Figure 10.1.7 Equivalence of voltage and power
decibels.
y
10.1.3a Referenced Measurements Table 10.1.1 Common dB Values
The previous example showed the decibel value dB Value Voltage Ratio Power Ratio
obtained from two measured quantities. Often 0 1 1
+1 1.122 1.259
audio engineers express the decibel value of a sig-
+2 1.259 1.586
nal relative to some standard reference instead of +3 1.412 1.995
another signal. The reference for decibel mea- +6 1.995 3.981
surements may be predefined as a power level, as +10 3.162 10
in dBm (decibels above 1 mW), or it may be a +20 10 100
voltage reference. When measuring dBm or any +40 100 10,000
power-based decibel value, the reference imped- –1 0.891 0.794
–2 0.794 0.631
ance must be specified or understood. For exam-
–3 0.707 0.501
ple, 0 dBm (600 Ω) would be the correct way to –6 0.501 0.251
specify level. Both 600 and 150 Ω are common –10 0.3163 0.1
reference impedances in audio work. –20 0.1 0.01
The equations assume that the circuit being –40 0.01 0.0001
measured is terminated in the reference imped-
ance used in the decibel calculation. However,
most voltmeters are high-impedance devices and
are calibrated in decibels relative to the voltage required to reach 1 mW in the reference imped-
ance. This voltage is 0.775 V in the 600-Ω case. Termination of the line in 600 Ω is left to the
user. If the line is not terminated, it is not correct to speak of a dBm measurement. The case of
decibels in an unloaded line is referred to as dBu (or sometimes dBv) to denote that it is refer-
enced to a 0.775-V level without regard to impedance.
Another common decibel reference in voltage measurements is 1 V. When using this refer-
ence, measurements are presented as dBV. Often it is desirable to specify levels in terms of a ref-
erence transmission level somewhere in the system under test. These measurements are
designated dBr where the reference point or level must be separately conveyed.
10.1.3b Crosstalk and Separation Measurements

One application of decibel measurements using two measured quantities is crosstalk. Crosstalk is
the leakage of signal from one audio channel into another. In the general case of crosstalk, the
channels are not necessarily related. For the special case of crosstalk between the two channels of
a stereo system the term separation is used. As a general rule, if the two channels under consid-
eration carry two channels of the same audio program, the term separation is used. If the two
channels are unrelated, the term crosstalk is used.
Crosstalk or separation is defined as the difference in decibels between the interfering-signal
level in the source channel and the receiving channel. Crosstalk and separation specifications are
usually expressed as a positive decibel value. Measuring crosstalk or separation consists of
applying a sine wave to one channel of the device under test and measuring the level of the signal
in the other channel. The amount of leakage is likely to depend on the frequency used, so these
measurements are normally made as a function of frequency and expressed in graphical form as
with an amplitude-frequency-response curve.
Take the case of a 10-V, l-kHz sine wave in the left channel of a stereo system. If the right
channel is not driven with signal, there should be no 1 kHz present at its output. However, if a
y
Figure 10.1.8 Crosstalk measurement with a bandpass filter.
level measurement on the right channel yields 10 mV of signal, the separation is defined to be 60
dB.
One possible problem with this procedure is that the 10 mV may not be 1 kHz leaking from
the other channel but may represent the noise floor of the system under test. If this is true, the
separation measurement is inaccurate. The solution is to use a bandpass filter tuned to the fre-
quency of the test tone, thereby rejecting system noise and other interfering components. If the
measurements are to be made as a function of frequency, the bandpass-filter frequency should be
slaved to the generator frequency. This approach is illustrated in Figure 10.1.8. One channel is
driven from a sine-wave generator at its nominal operating level and terminated in its normal
load impedance. The other channel input is terminated by the normal source impedance, and the
output is terminated by the normal load impedance. The level at the output of the driven channel
is measured with a voltmeter, and the output of the undriven channel is measured by a voltmeter
with a bandpass filter centered on the test-signal frequency. The level difference between these
measurements, expressed in decibels, is the separation. When the two channels have different
gains, it is common to correct the measurements for the gain difference so as to present the
crosstalk referred to the channel inputs.
Crosstalk between two audio channels is sometimes nonlinear. The presence of a signal in one
channel will sometimes yield tones of other frequencies in the receiving channel. This is espe-
cially true in transmission systems where cross-modulation can occur between carriers. These
tones disappear when the source signal is removed, clearly indicating that they are due to the sus-
pected source. Measuring them can be tedious if sine waves are used because the frequency of
the received interference may not be easily predictable. There may also be some test frequencies
which cause the interference products to appear outside the channel bandwidth, hiding the effect
being tested. Therefore, this test is often performed with a random noise source so that all possi-
ble interference frequencies are tested.
y
10.1.4 Noise Measurements

Noise measurements are simply specialized level mea-
surements. It has long been recognized that the ear's
sensitivity varies with frequency, especially at low lev-
els. This effect was studied in detail by Fletcher and
Munson and later by Robinson and Dadson. The
Flecher-Munson hearing-sensitivity curve for the
threshold of hearing is given in Figure 10.1.9. The ear
is most sensitive in the region of 2 to 4 kHz, with
Figure 10.1.9 Fletcher-Munson
rolloffs above and below these frequencies. This effect curves of hearing sensitivity versus
is responsible for the apparent loss of bass when the frequency.
volume is reduced on a hi-fi system. To predict how
loud something will sound it is necessary to use a fil-
ter which duplicates this nonflat behavior electrically. The filter weights the signal level on the
basis of its frequency, thus earning the name weighting filter.
Various attempts have been made to do this, resulting in several standards for noise measure-
ment. Some of these weighting filters are shown overlaid on the hearing-threshold curve in Fig-
ure 10.1.10 The most common filter in use in the United States is the A-weighting curve. This
filter is placed in front of a high-sensitivity voltmeter, and the amplitude of the noise is measured
10.1.4a Noise Analysis

Power line (mains) hum is low-frequency interference at the power line frequency or its multi-
ples. This interference is generally grouped with noise when making measurements. For North
American lines, the components occur at 60 Hz, 120 Hz, 180 Hz, 240 Hz, etc. For 50-Hz power
lines, the corresponding frequencies are 50 Hz, 100 Hz, 150 Hz, 200 Hz, etc. The dominant com-
ponents are usually fundamental, second, and third harmonic, which can be removed with a high-
pass filter at approximately 400 Hz. With a three-pole or sharper filter, 1-dB-accurate measure-
ments can be made down to 1 kHz. By making unweighted noise measurements, both with and
Figure 10.1.10 Common weighting filters for audio measurements.

y
Figure 10.1.11 Weighted voltmeter architecture.
without the high-pass filter, the effects of hum can be estimated. Subtracting the two decibel
measurements results in a level ratio measurement called the hum-to-hiss ratio. Good-quality
equipment will have hum-to-hiss ratios of less than 1 dB.
Even more information about the underlying sources of noise may be obtained with spectral
analysis. This may be accomplished with any of the common spectrum-analysis techniques,
including:
• Fast-Fourier-transform (FFT) analyzers
• Heterodyne analyzers
• Real-time analyzers (RTAs)
Each offers advantages for particular applications, but all provide considerable insight to the
measurement being made. Figure 10.1.12 shows the spectrum of output noise from a profes-
sional equalizer. The measuring equipment uses a sweeping one-third-octave filter and an rms
detector. Note the presence of power-line-related signals, both 120-Hz and 180-Hz components.
The 120-Hz product is due to asymmetrical charging currents in the power supply, while the 180-
Hz product is the result of transformer field leakage.
Another approach which is often useful in noise analysis is to view the noise on an oscillo-
scope and trigger the oscilloscope with an appropriate synchronization signal. The components
of the noise related to the synchronization signal will remain stationary on the screen, and the
unrelated energy will produce fuzz on the trace. For example, to investigate line-related compo-
nents, the oscilloscope should be triggered on the power line. If interference is suspected from a
nearby television signal, the oscilloscope can be triggered on the television vertical and/or hori-
zontal sync signals.
Noise may be expressed as an absolute level (usually in dBm or dBu) by simply measuring
the weighted voltage (proper termination being assumed in the case of dBm) at the desired point
in the system. However, this is often not very meaningful. A 1-mV noise voltage at the output of
a power amplifier may be quite good, while 1 mV of noise at the output of a microphone would
render it useless for anything but measuring jet planes. A better way to express noise perfor-
mance is the signal-to-noise ratio. S/N is a decibel measure of the noise level using the signal
level measured at the same point as a reference. This makes measurements at different points in a
system or in different systems directly comparable. A signal with a given S/N can be amplified
with a perfect amplifier or attenuated with no change in the S/N. Any degradation in S/N at later
points in the system is due to limitations of the equipment that follows.
y
Figure 10.1.12 Typical spectrum of noise and hum measured with a sweeping one-third-octave fil-
ter.
10.1.5 Bibliography
Naerum, Denmark.
y
verton, Ore., 1975.
g g
Chapter
10.2
Audio Phase and Frequency
Measurement
10.2.1 Introduction
When a signal is applied to the input of a device, the output will appear at a later point in time.
For a sine-wave excitation this delay between input and output may be expressed as a proportion
of the sine-wave cycle, usually in degrees. One cycle is 360°, one half-cycle is 180°, etc. This
measurement is illustrated in Figure 10.2.1. The phasemeter input signal no. 2 is delayed from, or
is said to be lagging, input no.1 by 45°. Most audio measuring gear measures phase directly by
measuring the proportion of one signal cycle between zero crossings of the signals. This can be
done with an edge-triggered set-reset flip-flop as shown in Figure 10.2.1. The output of this flip-
flop will be a signal which goes high during the time between zero crossings of the signals. By
averaging the amplitude of this pulse over one cycle (i.e., measuring its duty cycle) a measure-
ment of phase results.
10.2.2 Phase Measurement

Phase is typically measured and recorded as a function of frequency over the audio range. For
most audio devices, phase and amplitude responses are closely coupled. Any change in ampli-
tude that varies with frequency will produce a corresponding phase shift. A device that has no
more phase shift than what is required by the amplitude-response variation with frequency is
described as having minimum-phase characteristics. A typical phase-and-amplitude- versus-fre-
quency plot of a graphic equalizer is shown in Figure 10.2.2.
A fixed time delay will introduce a phase shift that is a linear function of frequency. This time
delay can introduce large values of phase shift at high frequencies which are of no significance in
practical applications. The time delay will not distort the waveshape of complex signals and will
10-19
q y
Figure 10.2.1 Basic measurement of a phase shift between two signals.
not be audible. There can be problems with time delay when the delayed signal will be used in
conjunction with an undelayed signal. This would be the case if one channel of a stereo signal
was delayed and the other was not. If we subtract out the absolute time delay from a phase plot,
the remainder will truly represent the audible portions of the phase response. In instances where
absolute delay is a concern, the original phase curve is more relevant.
10.2.2a Relation to Frequency

When dealing with complex signals, the meaning of phase becomes unclear. Viewing the signal
as the sum of its components according to Fourier theory, we find a different value of phase shift
at each frequency. With a different phase value on each component, which one is to be used? If
the signal is periodic and the waveshape is unchanged passing through the device under test, a
phase value may still be defined. This may be done by using the shift of the zero crossings as a
fraction of the waveform period. Indeed, most commercial phasemeters will display this value.
However, if there is differential phase shift with frequency, the waveshape will be changed. It is
then not possible to define any phase-shift value, and phase must be expressed as a function of
frequency.
Another useful expression of the phase characteristics of an audio device is group delay.
Group delay is the slope of the phase response. It expresses the relative delay of the spectral com-
q y
Figure 10.2.2 Typical phase-and-amplitude-versus-frequency plot of a graphic equalizer.
ponents of a complex waveform. This describes the delay in the harmonics of a musical tone rel-
ative to the fundamental. If the group delay is flat, all components will arrive together. A peak or
rise in the group delay indicates that those components will arrive later by the amount of the peak
or rise. It is computed by taking the derivative of the phase response versus frequency. Mathe-
matically
Group delay = – ( phase at f2 – phase at f1 ) / ( f2 – f 1 ) (10.2.4)
This requires that phase be measured over a range of frequencies to give a curve that can be
differentiated. It also requires that the phase measurements be performed at frequencies which
are close enough together to provide a smooth and accurate derivative.
10.2.3 Frequency Measurement

Frequency is a fundamental characteristic of periodic signals. It is simply the number of times
per second that the signal being measured repeats its pattern. An alternative way to specify this
parameter is the period of the signal; i.e., the time taken for one cycle of the pattern to occur.
Care should be taken not to confuse pitch and frequency. Pitch is essentially the perceived fre-
quency. Indeed, for complex waveforms such as narrowband noise of frequency-modulated sine
waves, frequency is difficult to define. For example, what is the “frequency” of a signal consist-
ing of 2-kHz, 3-kHz, 4-kHz, and 5-kHz sine waves? When this signal is heard, the brain will
“insert” the missing l-kHz fundamental and perceive a 1-kHz pitch. Pitch, though not always
obvious from electrical measurements, is readily apparent to a listener.
q y
Figure 10.2.3 High-resolution frequency measurement.
10.2.3a Measurement Techniques

Early methods of frequency measurement employed vibrating reeds or frequency-to-voltage con-
version circuits. These were not precision measurement techniques, but the frequency-to-volt-
age-converter approach was easy to use and became popular. Frequency measurement has
advanced greatly since the development of digital logic. Early designs used digital counters to
count the number of zero crossings during a fixed time window. For ease of design, these time
windows (gates) were decimal fractions or multiples of 1 s.
Modern designs take advantage of the computing power of microprocessors and measure
period, reciprocating the result to obtain the frequency. To perform this measurement both a
high-frequency reference clock and the input signal are counted during the gate interval, as illus-
trated in Figure 10.2.3. The frequency of the input signal may then be computed by the formula
F = clock frequency × no. of signal cycles /count (10.2.2)
Note that the gate interval does not enter into the calculation and may be chosen on the basis
of the speed of measurements desired. Longer gate intervals and higher-frequency clocks will
result in higher-resolution measurements. However, the gate interval must be an integer multiple
of the input-signal period. This is easy to ensure with appropriate logic circuitry. For the fairly
typical case of a 10-Hz signal, a 0.1-s gate, and a 10-MHz clock, we would have a 1-cycle gate
and a count of
10MHz × 1 cycle/10 Hz = 1 million (10.2.3)
giving a resolution of 6 digits. A 1-s gate would allow 10 cycles of input signal, giving a count of
10 million.
q y
Another scheme is sometimes used for

measuring low frequencies quickly and to
high resolution. This involves locking a
voltage-controlled oscillator (VCO) to a
multiple of the input frequency, usually
100, with a phase-locked loop (PLL). The
counter then counts the VCO output and
obtains a factor-of-l00 improvement in
resolution for the same gate time. This
scheme requires several cycles of input
signal for the PLL to acquire and lock to
the input. This time must be included in Figure 10.2.4 Sweeping spectrum analyzer
the measurement time, reducing the response.
improvement.
All these measurement techniques are limited as to achievable accuracy by the accuracy of the
reference. Typical quality crystal oscillators provide an accuracy of several parts per million at
room temperature at a cost of a few dollars. By temperature-compensating the crystal oscillator,
the ambient-temperature effects may be removed.
10.2.3b Spectrum Analysis

A spectrum analyzer, generally speaking, is a device that displays a signal in the frequency
domain. The most common approach is the sweeping-filter technique. It may be visualized as a
bandpass filter that sweeps in frequency as shown in Figure 10.2.4. The output level from the fil-
ter drives the vertical axis of a display, while the signal frequency is the horizontal-axis variable.
This gives a graph of the frequency content of the signal. A single sine wave will give a display
of one peak at a horizontal point corresponding to the frequency of the input.
Because of the problems in tuning a bandpass filter electronically, instruments have typically
been implemented along the lines of the block diagram shown in Figure 10.2.5. A sweeping
oscillator (local oscillator, LO) is mixed with the input signal in a multiplier, shifting the fre-
quency of the signal to that of the fixed-frequency bandpass filter. This process is called hetero-
dyning and is used in most radio receivers. This has also prompted some manufacturers of these
units to call them heterodyne analyzers. Two important characteristics identify this type of ana-
lyzer. First, the bandwidth of the analysis filter is fixed by the characteristics of the fixed-fre-
quency filter. The bandwidth of the
analysis will not depend on the analysis
frequency. Second, the analyzer must
sweep through the frequency range of
interest. If something happens at one fre-
quency when the analyzer is tuned to
another frequency, it will be missed.
The fixed-frequency-filter stage, called
the intermediate frequency (IF) amplifier,
is typically a multiple-stage filter with
amplifiers between each stage. This allows Figure 10.2.5 Block diagram of a heterodyne
a sharp rolloff characteristic and very spectrum analyzer.
q y
Figure 10.2.6 Error and delay (bias errors) in writing out peaks and valleys in a spectrum. (After
[1].)
small bandwidths. Means are normally provided for setting the IF bandwidth, allowing the reso-
lution of the spectrum analyzer to be adjusted. The frequency-analysis range, or span, of the ana-
lyzer is set by the tuning of the LO. A minimum bandwidth is required for any value of span and
sweep speed. This requirement allows the IF filter to settle to its steady-state response on the
input signal. If the sweep is too fast or the bandwidth too small, the filter output will give an
incorrect reading. The shape of the response seen on the analyzer screen for different sweep rates
is shown in Figure 10.2.6. As the sweep rate is increased above the optimum value, the peak will
start to drop and its frequency will shift in the direction of the sweep. The optimum bandwidth B
for a particular sweep time T and dispersion or total frequency sweep range D is given in Figure
10.2.7.
Another approach to spectrum analysis is using a real-time analyzer (RTA). A parallel bank of
bandpass filters is driven with the signal to be analyzed. The outputs of the filters are rectified
and displayed in bar-graph form on a cathode-ray tube (CRT) or other suitable display as shown
in Figure 10.2.8. The resulting display is shown in Figure 10.2.9. The filters are at fixed frequen-
cies, usually spaced every one-third octave or full octave from 20 Hz to 20 kHz. This results in
30 filters for the one-third-octave case and 10 filters for octave-band units. These frequencies
have been standardized by the IEC and are given in Table 10.2.1. Some units have been built with
12 filters per octave, or a total of 120 filters, for even higher resolution. Because RTAs are nor-
mally made with these fractional-octave filters, they are constant-percentage-bandwidth devices.
This means that the bandwidth of the filters is always a fixed percentage of the center frequency.
The advantage of parallel-filter analyzers is their instantaneous display and their ability to see
transient events. Since all filters are constantly monitoring the signal, all transients will be seen.
Disadvantages include the low-resolution display and the inability to trade resolution for fre-
quency range after the unit is manufactured.
q y
Figure 10.2.7 Optimum resolution setting for spectrum analyzers. Read Boptimum for a given dis-
persion and sweep time. (After [2].)
Figure 10.2.8 Real-time-analyzer block diagram.
RTAs are commonly used with a random-noise or multitone test signal to measure a device or
system response quickly. A random-noise signal is a signal whose instantaneous amplitude is a
random, usually gaussian variable. Random noise has a spectrum (amplitude versus frequency)
made up of all frequencies over the bandwidth of the noise. Various terms are used to describe
q y
the spectral distribution of noise; the most

common are pink noise and white noise.
White noise has an equal energy per hertz of
bandwidth. If it is measured by using a filter
whose bandwidth is constant with frequency
(such as a heterodyne analyzer), the spec-
trum will be flat. If it is measured with a
fractional-octave filter, as are RTAs, the
spectrum will rise at 3 dB per octave. Pink
noise is random noise that has equal energy Figure 10.2.9 Real-time (parallel-filter) analyzer
per octave. If it is displayed on an RTA, the response.
response will be flat. A heterodyne analyzer
will show a rolloff of 3 dB per octave.
For most audio testing work with RTAs,
the system or device under test is excited by Table 10.2.1 ANSI-ISO Preferred Frequencies.
the pink-noise test signal. The output of the 25 Hz 250 Hz 2.5 kHz
system is measured with the RTA, and the 31.5 Hz 315 Hz 3.15 kHz
frequency response is displayed without 40 Hz 400 Hz 4.0 kHz
waiting for a sweep. Because of the random 50 Hz 500 Hz 5.0 kHz
63 Hz 630 Hz 6.3 kHz
nature of the noise signal, there is significant
80 Hz 800 Hz 8.0 kHz
fluctuation in the amplitude of each filter 100 Hz 1.0 kHz 10.0 kHz
output. The average level will be correct, but 125 Hz 1.25 kHz 12.5 kHz
the instantaneous value will change consid- 160 Hz 1.6 kHz 16.0 kHz
erably with time. This puts significant uncer- 200 Hz 2.0 kHz 20.0 kHz
tainty in the response being measured. This
uncertainty may be reduced, but not elimi-
nated, by increasing the averaging time in the
ac-to-dc converters after the filters.
Multitone test signals, which consist of a sine wave at the center frequency of each filter used
in the analysis, may be constructed. When this signal is displayed on an RTA, the response will
be shown without the random fluctuations common to pink noise. However, because the signal
consists of a small number of sine waves, the response shown on the RTA is the steady-state
response at each filter's center frequency. Because multitone signals measure at a single fre-
quency in each filter, they will be greatly affected by sharp peaks and dips in the response being
measured. When pink noise is used to make this same measurement, the RTA's response will be
the average over the frequency range of each filter. This becomes a distinct advantage of random
noise as a source for some acoustic measurements. It will average the response over the range of
the analyzing filter, resulting in a measurement less affected by room modes and resonances in
transducers.
10.2.4 Fast Fourier Transform Measurement

Time and frequency domains are alternate ways of looking at a signal. The French mathemati-
cian Fourier proved that any signal may be represented as a series of sine waves summed
together. Indeed, this is the justification for analyzing signals with a narrowband-tunable-filter
q y
spectrum analyzer to examine these components. Advances in technology have made it possible
to implement directly Fourier's theory with hardware and software. Instruments that perform
Fourier analysis digitize the signal, sampling the waveform at a rate faster than the highest-fre-
quency input signal, and convert these samples into a numerical representation of the signal's
instantaneous value. Fourier series provides a way to convert these signal samples into samples
of the signal spectrum. This transforms the data from the time domain to the frequency domain.
The FFT is merely a technique for efficiently computing the Fourier series by eliminating redun-
dant mathematical operations.
The FFT operates on a piece of the signal that has been acquired and stored in memory for the
calculation. Take, for example, the section of a sine wave shown in Figure 10.2.10. This is a
piece of a sine wave which continues in time on both sides of the selected segment. The FFT
algorithm does not know anything about the waveform outside the piece that it is using for calcu-
lations. It therefore assumes that the signal repeats itself outside the “window” it has of the sig-
nal. This is important because an incorrectly selected piece of the signal may lead to very strange
results.
Consider the sine wave of Figure 10.2.10 and the possible pieces of it that have been selected
for analysis. In the first example, the beginning and end of the window have been chosen to coin-
cide with zero crossings of the signal. If the selected segment is repeated, an accurate representa-
tion of the signal is obtained. The FFT of this signal will give the correct spectrum, a single-
frequency component. If the window is chosen incorrectly, as in the second example, there is a
noninteger number of cycles in the waveform. When this segment is repeated, the resulting wave-
form will not look like the original sine wave. The computed spectrum will also be in error; it
will be the spectrum of the discontinuous waveform. Transients that start at zero and decay to
zero before the end of the sample segment will not suffer from any discontinuities. Continuous
random or periodic signals will, however, be affected in a manner analogous to the effects on sine
waves.
Clearly, then, the choice of windows for a signal that is to be transformed is critical to obtain-
ing correct results. It is often difficult to select the correct end-points of the window, and even
more difficult without operator involvement. The window function may be thought of as a rectan-
gular-shaped function which multiplies the signal. Intuitively it seems that the sharp discontinui-
ties introduced by the endpoints of the window are at fault for the spurious components in the
FFT. This may be proved theoretically but is beyond the scope of this discussion. A simple solu-
tion to the windowing problem is to use nonrectangular windows. Multiplying the signal by a
window that decreases to zero gradually at its end-points eliminates the discontinuities. Using
these windows modifies the data and results in a widening of the spectral peaks in the frequency
domain. This is illustrated in Figure 10.2.11. This tradeoff is unavoidable and has results in the
development of many different windowing functions that emphasize the spectral widening or the
rejection of spurious components. Perhaps the most common window function the Hamming
window, illustrated in Figure 10.2.12, which is a cosine function raised above zero. It rejects spu-
rious signals by at least 42 dB but spreads the spectral peaks by only 40 percent.
The FFT algorithm always assumes that the signal being analyzed is continuous. If transient
signals are being analyzed, the algorithm assumes that they repeat at the end of each window. If a
transient may be guaranteed to have decayed to zero before the end of the window time, a rectan-
gular window may be used. If not, a shaped window must still be used. However, if there is sig-
nificant transient energy at the end of the window, the computed spectrum will not include the
frequency contribution of that data.
q y
(a)
(b)
(c)
Figure 10.2.10 The FFT assumes that all signals are periodic and that they duplicate what is cap-
tured inside the FFT window: (a) periodic signal, integral number of cycles in the measurement
window; (b) nonperiodic signal transient, (c) nonperiodic signal, random. (After [3].)
Fourier-transform algorithms can be written for any number of data points. However, it is eas-
iest if the number of points can be expressed as the product of two smaller numbers. In this case,
the transform may be broken into the product of two smaller transforms, each of a length equal to
the smaller numbers. The most convenient lengths for transforms, based on this scheme, are
powers of 2. Common transform lengths are 512 points and 1024 points. These would ordinarily
provide a spectrum with 256 and 512 components, respectively. However, because of errors at
high frequencies due to aliasing and the rolloff introduced by windowing only 200 or 400 lines
are displayed.
The transformation from the time domain to the frequency domain by using the FFT may be
reversed to go from the frequency domain to the time domain. This allows signal spectra to be
analyzed and filtered and the resulting effect on the time-domain response to be assessed. Other
transformations may be applied to the data, yielding greater ability to separate the signal into its
components.
q y
(a )
(b )
(c )
Figure 10.2.11 Window shapes trade off major-lobe bandwidth and side-lobe rejection. (a) An
almost periodic waveform in the rectangular acquisition window. The FFT magnitude (expanded 4
times for detail) shows closely adjacent, nearly equal components; one has substantial leakage.
Also, a small wrinkle at two divisions from center hints at a possible third component. (b) Multiply-
ing the waveform by a Hamming window reduces side-lobe leakage and reveals a third low-fre-
quency component in the FFT magnitude (expanded 4 times for detail). (c) A Parzen window
offers more side-lobe reduction, but the increased bandwidth of the major lobe causes the two
nearly equal components to merge completely into each other. ( After [3].)
q y
Figure 10.2.12 Some common FFT data windows and their frequency-domain parameters. (After
[3].)
10.2.5 References
1. Randall, R. B.: Application of B&K Equipment to Frequency Analysis, 2nd ed., B&K
Instruments, Naerum, Denmark, 1977.
2. Engelson, M., and F. Telewski: Spectrum Analyzer Theory and Applications, Artech House,
Norwood, Mass., 1974.
3. Ramirez, R.: “The FFT—Fundamentals and Concepts,” Tektronix, Inc., Beaverton, Ore.,
1975.
10.2.6 Bibliography
q y
Naerum, Denmark.
verton, Ore., 1975.
q y
g g
Chapter
10.3
Nonlinear Audio Distortion
10.3.1 Introduction
Distortion is a measure of signal impurity. It is usually expressed as a percentage or decibel ratio
of the undesired components to the desired components of a signal. Distortion of a device is
measured by feeding it one or more sine waves of various amplitudes and frequencies. In sim-
plistic terms, any frequencies at the output which were not present at the input are distortion.
However, strictly speaking, components due to power line interference or other spurious signals
are not distortion. There are many methods of measuring distortion in common use: harmonic
distortion and at least three different types of intermodulation distortion. These are different test
procedures rather than different forms of distortion in the device under test.
10.3.2 Harmonic Distortion

The transfer characteristic of a typical device is shown in Figure 10.3.1. This represents the out-
put voltage at any point in the signal waveform for a given input voltage; ideally this is a straight
line. The output waveform is the projection of the input sine wave on the device transfer charac-
teristic. A change in the input produces a proportional change in the output. Because the actual
transfer characteristic is nonlinear, a distorted version of the input waveshape appears at the out-
put.
Harmonic distortion measurements excite the device under test with a sine wave and measure
the spectrum of the output. Because of the nonlinearity of the transfer characteristic, the output is
not sinusoidal. By using Fourier series, it can be shown that the output waveform consists of the
original input sine wave plus sine waves at integer multiples (harmonics) of the input frequency.
The spectrum of the distorted signal is shown in Figure 10.3.2. For a l-kHz input, the output con-
sists of 1 kHz, 2 kHz, 3 kHz, etc. The harmonic amplitudes are proportional to the amount of dis-
tortion in the device under test. The percentage harmonic distortion is the rms sum of the
harmonic amplitudes divided by the rms amplitude of the fundamental.
10-33
Figure 10.3.1 Total harmonic distortion (THD) test of transfer characteristic.
Figure 10.3.2 Reading THD from a spectrum analyzer.
To measure harmonic distortion with a spectrum analyzer the procedure illustrated in Figure
10.3.2 is used. The fundamental amplitude is adjusted to the 0-dB mark on the display. The
amplitudes of the harmonics are then read and converted to linear scale. The rms sum of these
values is taken and represents the THD. This procedure is time-consuming and difficult for an
unskilled operator. Even skilled operators have trouble in obtaining accuracies better than 2 dB
in the final result because of equipment limitations and the problems inherent in reading num-
bers off a trace on the screen of an analyzer.
A simpler approach to the measurement of harmonic distortion is the notch-filter distortion
analyzer. This device, commonly referred to as simply a distortion analyzer, removes the funda-
mental of the signal to be investigated and measures the remainder. A block diagram of such a
unit is shown in Figure 10.3.3. The fundamental is removed with a notch filter, and its output is
then measured with an ac voltmeter. Since distortion is normally presented as a percentage of the
fundamental level, this level must be measured or set equal to a predetermined reference value.
Additional circuitry (not shown) is required to set the level to the reference value for calibrated
Figure 10.3.3 Basic block diagram of an harmonic distortion analyzer.
Figure 10.3.4 Conversion graph for indicated distortion and true distortion. (From [1]. Used with
permission.)
measurements. Some analyzers use a series of step attenuators and a variable control for setting
the input level to the reference value.
More sophisticated units eliminate the variable control by using an electronic gain control.
Others employ a second ac-to-dc converter to measure the input level and compute the percent-
age using a microprocessor. Completely automatic units also provide autoranging logic to set the
attenuators and ranges. This provision significantly reduces the effort and skill required to make
a measurement.
The correct method of representing percentage distortion is to express the level of the har-
monics as a fraction of the fundamental level. However, commercial distortion analyzers use the
total signal level as the reference voltage. For small amounts of distortion these two quantities
are equivalent. At large values of distortion the total signal level will be greater than the funda-
mental level. This makes distortion measurements on these units lower than the actual value. The
relationship between the measured distortion and the true distortion is given in Figure 10.3.4.
The errors are negligible below 10 percent measured distortion and are not significant until 20
percent measured distortion.
Figure 10.3.5 Interference in distortion and noise measurements.
The need to tune the notch filter to the correct frequency can also make this a very tedious
measurement. Some manufacturers have circumvented this problem by including the measure-
ment oscillator and analyzer in one package, placing the analyzer and oscillator frequency con-
trols on the same knob or button. This eliminates the problem only when the signal source used
for the test is the internal oscillator. A better approach, used by some manufacturers, is to mea-
sure the input frequency and tune the filter to the measured frequency. This eliminates any need
to adjust the analyzer frequency.
Because of the notch-filter response, any signal other than the fundamental will influence the
results, not just harmonics. Some of these interfering signals are illustrated in Figure 10.3.5. Any
practical signal contains some hum and noise, and the distortion analyzer will include these in
the reading. Because of these added components, the correct term for this measurement is total
harmonic distortion and noise (THD + N). Although this factor does limit the readings of equip-
ment for very low distortion, it is not necessarily bad. Indeed it can be argued that the ear hears
all components present in the signal, not just the harmonics. Some interfering signals, such as the
19-kHz pilot tone used in frequency-modulation (FM) stereo, may be outside the range of audi-
bility and therefore totally undesirable.
Additional filters are included on most distortion analyzers to reduce unwanted hum and
noise as illustrated in Figure 10.3.6. These usually consist of one or more high-pass filters (400
Hz is almost universal) and several low-pass filters. Common low-pass-filter frequencies are
22.4 kHz, 30 kHz, and 80 kHz. Better equipment will include filters at all these frequencies to
ease the tradeoff between limiting bandwidth to reduce noise and the reduction in reading accu-
racy from removing desired components of the signal. When used in conjunction with a good
differential input on the analyzer, these filters can solve most practical measurement noise prob-
lems.
The use of a sine-wave test signal and a notch-type distortion analyzer has the distinct advan-
tage of simplicity in both instrumentation and use. This simplicity has an additional benefit in
ease of interpretation. The shape of the output waveform from a notch-type analyzer indicates
the slope of the nonlinearity. Displaying the residual components on the vertical axis of an oscil-
loscope and the input signal on the horizontal gives a plot of the transfer characteristics' devia-
tion from a best-fit straight line. This technique is diagramed in Figure 10.3.7. The trace will be
Figure 10.3.6 Use of filters to reduce interference in THD measurement.
Figure 10.3.7 Transfer function monitoring.
a horizontal line for a perfectly linear device. If the transfer characteristic curves upward on pos-
itive input voltages, the trace will bend upward at the right-hand side. Examination of the distor-
tion components in real time on an oscilloscope will show such characteristics as oscillation on
the peaks of the signal, crossover distortion, and clipping. This is a valuable tool in the design
and development of audio circuits and one which no other distortion measurement method can
fully match. Viewing the residual components in the frequency domain using a spectrum ana-
lyzer also yields considerable information about the distortion mechanism inside the device
under test.
Both the frequency and the amplitude of the sine-wave stimulus are adjustable parameters in
harmonic-distortion testing. This often proves to be of great value in investigating the nature of a
distortion mechanism. By measuring at low frequencies, thermal distortion and related effects
may be examined in detail. Using frequencies near the power line frequency and looking for
beats in the distortion products can reveal power supply limitations and line-related interference.
Measurements at high frequencies can reveal the presence of nonlinear capacitances or slew-rate
limiting. By examining the slope of the distortion change with frequency, several mechanisms
which are active in the same frequency range may be isolated.
Limitations when measuring distortion at high frequencies are the major problem with THD
testing, as illustrated in Figure 10.3.8. Because the components being measured are harmonics of
the input frequency, they may fall outside the pass-band of the device under test. An audio
recorder with a cutoff frequency of 22 kHz

(typical for a good machine) will allow mea-
surement only up to the third harmonic of a 7-
kHz input signal. THD measurements on a 20-
kHz input are impossible because all the distor-
tion components are filtered out by the
recorder.
Figure 10.3.8 Problems in measuring har-

10.3.3 Intermodulation Distortion monic distortion in band-limited systems.
The commonly accepted solution to the limita-

tions of THD analysis is to measure distortion
by the interaction or intermodulation (IM) of
two or more signals passing through a device
simultaneously. Many methods have been
devised to measure this interaction. The most
common of these is SMPTE IM, named after
the Society of Motion Picture and Television
Engineers, which first standardized its use. IM
measurements according to the SMPTE Figure 10.3.9 Measuring intermodulation in
method have been in use since the 1930s. The band-limited systems.
test signal is a low-frequency tone (usually 60
Hz) and a high-frequency tone (usually 7 kHz)
mixed in a 4:1 amplitude ratio. Other amplitude ratios and frequencies are used occasionally. The
signal is applied to the device under test, and the output signal is examined for modulation of the
upper frequency by the low-frequency tone. The amount by which the low-frequency tone modu-
lates the high-frequency tone indicates the degree of nonlinearity. As with harmonic-distortion
measurement, this may be done with a spectrum analyzer or with a dedicated distortion-analysis
instrument. The modulation components of the upper signal appear as sidebands spaced at multi-
ples of the lower-frequency tone as illustrated in Figure 10.3.9. The amplitudes of the sidebands
are rms-summed and expressed as a percentage of the upper-frequency level.
The most direct way to measure SMPTE IM distortion is to measure each component with a
spectrum analyzer and rms-sum them together. The spectrum-analyzer approach has a drawback
in that it is sensitive to frequency modulation of the carrier as well as amplitude modulation.
Because Doppler effects cause frequency modulation, this approach cannot be used on loud-
speakers.
A distortion analyzer for SMPTE testing is quite straightforward. The signal to be analyzed is
passed through a high-pass filter to remove the low-frequency tone as shown in Figure 10.3.10.
The high-frequency tone is then demodulated as if it were an amplitude-modulation (AM) radio
signal to obtain the sidebands, which are low-pass-filtered to remove any remaining high-fre-
quency energy. The resultant demodulated low-frequency signal will follow the envelope of the
high-frequency tone. This low-frequency fluctuation is the distortion and is displayed as a per-
centage of the high-frequency tone's amplitude. Since this low-pass filtering sets the measure-
ment bandwidth, noise has little effect on SMPTE measurements. The analyzer is very tolerant
of harmonics of the two input signals, allowing fairly simple oscillators to be used to generate the
Figure 10.3.10 Basic block diagram of an SMPTE intermodulation analyzer.
Figure 10.3.11 SMPTE intermodulation test of transfer characteristic.
test signal. Indeed, early analyzers used a filtered version of the power line for the low-frequency
tone: hence the 60-Hz low tone frequency in the SMPTE standard. It is important, however, that
no harmonics of the low-frequency signal generator extend into the measurement range of the
high-frequency tone. The analyzer will be unable to distinguish these from sidebands. After the
first stage of filtering in the analyzer, there is little low-frequency energy left to create IM in the
analyzer. This considerably simplifies the remaining circuitry.
As shown in Figure 10.3.11, when this composite signal is applied to the test device, the out-
put waveform is distorted. As the high-frequency tone is moved along the transfer characteristic
by the low-frequency tone, its amplitude changes.

The high-frequency tone is being used to measure
the gain at each point of the transfer characteris-
tic. This results in low-frequency amplitude mod-
ulation of the high-frequency tone. This
modulation results in sidebands around the high-
frequency lone as described previously. This test
is therefore quite sensitive to such conditions as
crossover distortion and clipping. High-order
Figure 10.3.12 CCIT intermodulation in
nonlinearities create bumps in the transfer charac-
band-limited systems.
teristic which produce large amounts of SMPTE
IM.
SMPTE testing is also useful for exciting low-frequency thermal distortion. As the low-fre-
quency signal moves around, exciting thermal effects, the gain of the device changes, creating
modulation distortion. Another excellent application is the testing of output inductance-capaci-
tance (LC) stabilization networks in power amplifiers. Low-frequency signals may saturate the
output inductor, causing it to become nonlinear. Because the frequency is low, very little voltage
is dropped across the inductor, and there is little low-frequency harmonic distortion. The high-
frequency tone will develop a signal across the inductor because of the rising impedance with
frequency. When the low-frequency tone creates a nonlinear inductance, the high-frequency tone
becomes distorted. This illustrates an important advantage of SMPTE IM testing. The sensitivity
may be quite high to low-frequency distortion mechanisms in that the components occur at high
frequencies. In most electronic audio circuits there is feedback which decreases at high frequen-
cies. This lower feedback allows the distortion components to escape much of the reduction oth-
erwise achieved at low frequencies by such feedback.
The inherent insensitivity of SMPTE IM to wow and flutter fostered widespread use of the
SMPTE test in applications which involve recording audio signals. Much use is made of SMPTE
IM in the film industry, for example.
It is often claimed that because the distortion components in an SMPTE test are not harmoni-
cally related to either input they will be more objectionable to the ear. Musical instruments are
rich in harmonics but contain few if any components which are inharmonic. With the typical 60-
Hz low-frequency tone used in SMPTE measurements, it is doubtful that the sidebands could be
outside the masking range of the ear. However, it is quite possible for the test to be indicative of
audible performance even if the test-signal distortion is not audible.
A number of studies have compared SMPTE IM readings with harmonic-distortion readings.
For most classic transfer-function nonlinearities, the SMPTE test is approximately 12 dB more
sensitive than THD. However, when heavy feedback is employed or when dynamic distortion
effects are present, the difference becomes considerably less predictable.
Other IMD test methods include the twin-tone intermodulation or CCIT difference-frequency
distortion test. The test signal consists of two closely spaced high-frequency tones as shown in
Figure 10.3.12. When these are passed through a nonlinear device, IM products are generated at
frequencies related to the difference in frequency between the original tones. The distortion
products in this test are usually far removed from the input signal. This positions them outside
the range of the auditory system's masking effects.
10.3.3a Specialized IM Tests

The use of swept-frequency two-tone IM tests to study transient intermodulation (TIM), or
slope-induced distortion, has numerous advantages including the ability to adjust the test fre-
quency to enable qualitative study of the underlying distortion mechanism [2]. Factors such as
the steepness of the change in distortion with frequency and the frequency at which the change
begins are useful in separating static from dynamic distortion mechanisms. The dominant order
of the nonlinearity also gives useful information about its origins.
The totally in-band character of the test is one of its most attractive attributes. The difference
product at f1 – f2 is widely separated in frequency and is easy to measure with a low-pass filter.
This provides a simple implementation of the measurement and is the one most commonly used.
However, the difference-frequency product is responsive only to even-order, or asymmetrical,
nonlinearities in the device under test. It is possible to measure the odd-order products without a
spectrum analyzer by using a precision squarer and bandpass filters. This approach has limita-
tions in dynamic range because of the multiplier or squarer and its attendant inaccuracies.
Another approach, detailed by Thiele [3] and Small [4], involves a clever choice of test fre-
quencies to place both even- and odd-order components near each other. The idea is to choose
the two test frequencies in almost a 3:2 frequency ratio. If they were exactly a 3:2 ratio, the sec-
ond-order and third-order products would both fall at the 1 position in frequency; i.e., for tones
of 10kHz and 15 kHz, both distortion components would fall at 5 kHz. Both components may
then be measured with a single multipole bandpass filter tuned to this frequency. To prevent pos-
sible cancellation of the two distortion products if they should happen to be of opposite phase,
the frequencies are offset slightly. The distortion components will then be at slightly different
frequencies, yielding an unambiguous measurement.
The sine-wave-square-wave IM test was originally proposed by Leinonen, Otala, and Curl [5].
It consists of summing a square wave at 3.18 kHz and a sine wave at 15 kHz (of one-fourth of the
square-wave amplitude) and applying them to the device under test. The signal is low-pass-fil-
tered at either 30 kHz or 100 kHz and is described as DIM (dynamic intermodulation) 30 or DIM
100, respectively. As the amplifier slews on the corners of the square wave, it becomes nonlinear
and distorts the sine wave. The resulting IM components are measured with a spectrum analyzer,
rms-summed, and expressed as a percent of the 15-kHz sine-wave amplitude.
The test-signal spectrum is shown in Figure 10.3.13. Leinonen and associates [5] define nine
IM components in the audio band to be summed when making the measurement. If the square
wave is not exactly 50 percent duty cycle, there will be even-order harmonics of 3.18 kHz in the
test signal. These even-order-harmonic components of the square wave, when introduced by the
device under test, are not included in the measurement. However, these components may be as
large as the IM components with some test devices and do provide valuable insight into the
device nonlinearities.
Owing to its complexity in measurement, the test was slow to be adopted by the audio indus-
try despite widespread interest in transient forms of distortion. Skritek [6] and later Hofer [7]
developed a procedure that simplifies this measurement considerably. The basic approach is
illustrated in Figure 10.3.14. Both an even-order and an odd-order IM product fold down to the
2-kHz region from interaction of the square wave and sine wave. By using a square wave of 3.15
kHz and a sine wave at 15 kHz, these appear at 750 Hz and 2.4 kHz. The 3.18-kHz square wave
will produce components at 900 Hz and 2.28 kHz. The amplitude of these two components may
be measured with a high-order elliptic low-pass filter at approximately 2.5 kHz. Their amplitude
is expressed as a percentage of the 15-kHz tone amplitude. Owing to the practical difficulties of
Figure 10.3.13 Spectrum of sine-wave-square-wave TIM test.
Figure 10.3.14 Simplified TIM test-measurement system.
measuring the 15-kHz signal directly, the rms amplitude of the total signal is measured and a cor-
rection factor is applied. Additional high-pass filtering at approximately 400 Hz may be used to
eliminate the effects of hum on the measurement. A block diagram of a DIM distortion analyzer
using this measurement approach is shown in Figure 10.3.15.
10.3.4 Distortion-Measurement Hardware

Distortion measurements should be performed with an rms-responding meter in the distortion
analyzer. This is necessary to make the reading represent the true power in the distortion. With
most practical distortion measurements, the rms response will read about 2 dB higher than the
average response. Some arguments can be made for measuring distortion with peak or quasi-
peak detectors. The averaging time of the ear is between 50 and 250 ms, depending on test condi-
Figure 10.3.15 Block diagram of a simplified DIM distortion meter.
tions and procedures. This has led to the definition of the quasi-peak meter for measuring tele-
phone noise as described previously.
Accuracy of most distortion analyzers is specified at better than 1 dB, but this can be mislead-
ing. Separate specifications are often put on the bandwidth and ranges, as is common for voltme-
ters. A more important specification for distortion measurements is the residual distortion of the
measurement system. Manufacturers of distortion analyzers often specify the oscillator and the
distortion analyzer separately. A system in which the oscillator and the analyzer are each speci-
fied at 0.002 percent THD can have a system residual distortion of 0.004 percent. If the noise of
the analyzer and/or the oscillator is specified separately, this must be added to the residual speci-
fication to find the residual THD + N of the system. It is not uncommon to find this limiting sys-
tem residual at most input voltages. For example, an analyzer specified at 0.002 percent
distortion and 20-μV input noise will have a 0.003 percent residual at 1-V input and 0.02 percent
at 0.1-V input. These voltages are common when measuring analog mixing consoles and pream-
plifiers, resulting in a serious practical limitation with some distortion analyzers.
Many commercial units specify the residual distortion at only one input voltage or at the full
scale of one range. The performance may degrade by as much as 10 dB when the signal is at the
bottom of an input range. This is true because THD + N measurements are a ratio of the distor-
tion components and noise to the signal level. At the full-scale input voltage, the voltage in the
notch filter is a maximum and the filter's noise contribution will be minimized. As the level
drops, the residual noise in the notch filter becomes a larger percentage of the reading. When the
next input range occurs, the residuals will improve again. This limitation is in addition to the
input-noise problem discussed previously because it results from noise in a later portion of the
instrument.
10.3.4a Addition and Cancellation of Distortion

Another often-overlooked problem is that of distortion addition and cancellation in the test
equipment or the device under test. Consider the examples in Figures 10.3.16 and 10.3.17. Sup-
pose one device under test has a transfer characteristic similar to that diagramed at the top of Fig-
ure 10.3.16a and another has the characteristic diagramed at the bottom. If they are cascaded, the
resulting transfer-characteristic nonlinearity will be magnified as shown. The effect on sine
waves in the time domain is illustrated in Figure 10.3.16b. The distortion component generated
(a )
(b )
Figure 10.3.16 Addition of distortion: (a) addition of transfer-function nonlinearities, (b) addition of
distortion components.
by each nonlinearity can be seen to be in phase and will sum to a component of twice the magni-
tude. However, if the second device under test has a complementary transfer characteristic as
shown in Figure 10.3.17, we obtain quite a different result. When the devices are cascaded, the
effects of the two curves will cancel, yielding a straight line for the transfer characteristic. The
corresponding distortion products are out of phase with each other, resulting in no distortion
components in the final output.
It is quite common for this to occur at low levels of distortion, especially between the test
equipment and the device under test. For example, if the test equipment has a residual of 0.002
percent when connected to itself and readings of 0.001 percent are obtained from the circuit
under test, cancellations are occurring. It is also possible for cancellations to occur in the test
equipment itself, with the combined analyzer and signal generator system giving readings lower
than the sum of their individual residuals. If the distortion is from an even-order (asymmetrical)
nonlinearity, reversing the phase of the signal between the offending devices will change a can-
cellation to an addition. If the distortion is from an odd-order (symmetrical) nonlinearity, phase
inversions will not affect the cancellation.
10.3.5 Signal-Source Effects in Measurements

The signal source is an often-overlooked factor in the accuracy of level measurements. Let us
consider the measurement of a low-pass filter's gain above its –3-dB frequency. This is shown in
(a )
(b )
Figure 10.3.17 Cancellation of distortion: (a) cancellation of distortion waveform, (b) cancellation
of transfer-characteristic nonlinearity.
Figure 10.3.18 for a three-pole low-pass filter. If the frequency of the generator is off by 3 per-
cent, the gain measurement will be off by 1 dB. A higher-order filter or a less accurate generator
will produce more error.
Figure 10.3.19 shows the effect of generator distortion on the gain measurement of a multi-
pole high-pass filter. The harmonics of the generator are not attenuated as much by the filter as is
the fundamental. If the signal-source distortion is high, as with function generators, the gain
measurement will be in error. This effect is most important when measuring notch filters in an
equalizer, where the distortion will appear as inadequate notch depth.
Figure 10.3.19 suggests that the effect of these errors on distortion measurements is more
severe. The gain introduced by a filter on the harmonics of the generator can make them exceed
the distortion of the filter itself. A three-pole high-pass filter will introduce 18 dB of gain at the
second harmonic and 29 dB of gain at the third. Under these conditions an oscillator which has
0.001 percent second- and 0.001 percent third-harmonic distortion will read 0.03 percent when
measuring a distortion-free filter. These errors necessitate the use of an oscillator with very low
distortion.
Another source of error in measurements is the output impedance of the generator. The
amplitude and phase response of a device under test will often be affected by its input impedance
interacting with the source impedance of the generator. These devices form a resistive divider in
which the shunt leg is the nonconstant impedance of the device under test. This causes a varia-
tion of the voltage at the input to the test device, thus corrupting the measurements. Low-output-
impedance generators will suffer less variation with load than high-impedance generators. How-
ever, if the system response is being measured, the generator impedance should be equal to the
Figure 10.3.18 Frequency-accuracy effects in measuring filters.
Figure 10.3.19 Source-distribution effects in high-pass filters.
source impedance of the device normally driving that input. Transformer input stages often
require a specific source impedance to provide correct damping for optimum high-frequency
response. Too large a source impedance will cause excessive rolloff while too low a source
impedance will produce an underdamped or peaked response. For example, most microphone
inputs are designed to be driven from a 150-Ω source, a value close to the typical microphone
source impedance.
10.3.6 References
1. Tremaine, H. W.: Audio Cyclopedia, Howard W. Sams, Indianapolis, Ind., 1975.
2. Ladegaard, P.: “Swept Distortion Measurements—An Effective Tool for Revealing TIM in
Amplifiers with Good Subjective Correlation to Subjective Evaluation,” B&K Application
Note 17-234, B&K Instruments, Naerum, Denmark, 1977.
3. Theile, A. N.: “Measurement of Nonlinear Distortion in a Bandlimited System,” J. Audio
Eng. Soc., AES, New York, N.Y., vol. 31, pp. 443–445, June 1983.
4. Small, R.: “Total Difference Frequency Distortion: Practical Measurements,” J. Audio Eng.
Soc., AES, New York, N.Y., vol. 34, no. 6, pg. 427, June 1986.
5. Leinonen, F., M. Otala, and J. Curl: “Method for Measuring Transient Intermodulation Dis-
tortion,” Audio Engineering Society preprint 1185, AES, New York, N.Y., October 1976.
6. Skritek, P.: “Simplified Measurement of Squarewave/Sinewave and Related Distortion Test
Methods,” Audio Engineering Society preprint 2195, AES, New York, N.Y., 1985.
7. Hofer, B.: “Practical Extended Range DIM Measurements,” Audio Engineering preprint
2334, AES, New York, N.Y., March 1986.
10.3.7 Bibliography
Naerum, Denmark.
verton, Ore., 1975.
g g
Chapter
10.4
Time Domain Audio Measurements
10.4.1 Introduction
In addition to characterizing the frequency-domain behavior of a device, it is also informative to
examine the time-domain behavior of audio components. The most common signals for this pur-
pose are sine waves, triangle waves, square waves, and tone bursts.
10.4.2 Time Domain Characterization

Sine waves are used in the time domain to measure clipping behavior. The top of a sine wave nor-
mally has a rounded appearance, but when a device is driven into saturation or clipping with a
sine wave, the top, bottom, or both will flatten. If the device is well behaved, the end of the flat
region should change smoothly to a sine wave. There should be no glitches or ringing on the sig-
nal as the device comes out of saturation. Triangle waves are sometimes used instead of sine
waves for this application because their peaks are sharper, making the flattening easier to see.
Square waves are used to determine the time-domain equivalents of amplitude and phase ver-
sus frequency measurements. The appearance of the output waveform from the device under test
gives much qualitative information about its behavior. A selection of square-wave responses is
given in Figure 10.4.1. Quantitative measurements can also be obtained from the square-wave
response. The most common of these are rise time, overshoot, ringing, tilt, and slew rate. The
first four are measurements of the linear behavior of the device and may be related to the fre-
quency-domain behavior. These are illustrated in Figure 10.4.2.
The rise time is the time required for the signal amplitude to change from 10 to 90 percent of
the total square-wave amplitude. For a single-pole low-pass device the rise time (RT) and the
bandwidth (BW) are related by
RT = 0.35 / BW (10.4.1)
10-49
Figure 10.4.1 Effects of amplitude and phase response on square-wave characteristics. (From [1].
Time Domain Measurements 10-51
Figure 10.4.2 Definition of rise time, overshoot, ringing, and droop. (From [1]. Used with permis-
sion.)
For multipole systems this equation will be only approximate.

The overshoot of a device is the amount by which the peak of the square wave exceeds steady-
state positive or negative amplitude, as shown. Large overshoots are indicative of peaking or
excess phase shifts in the high-frequency response of the device under test. The maximum over-
shoot is normally the parameter specified, but the subjective appearance of the overshoot is also
of interest. Well-behaved devices will have smooth overshoot, symmetrical on both positive and
negative peaks. Overshoot normally occurs only on the leading edges of a square wave. However,
if the device is linear-phase, as are many digital filters, there will be symmetrical overshoot on
both leading and trailing edges.
Ringing is the tendency of band-limited square waves to oscillate on the peaks. This is largely
a subjective measurement with the results stated as the severity of ringing. The greater the high-
frequency peaking, the greater will be the ringing.
Tilt is a measure of low-frequency behavior. As low frequencies are filtered, phase shifts are
introduced which cause the leading edge of the square wave to rise and the trailing edge to fall.
This produces a tilt to the top and bottom of the square wave. The tilt is usually expressed as a
percentage of the peak amplitude of the square wave.
Slew rate is a measure of how fast a signal changes from one instantaneous value to another.
The ideal square wave changes from one amplitude extreme to the other instantly. Practical
devices cannot keep up with this transition and will often have a maximum speed with which
they can change. This speed limitation will result in a tilted straight-line portion of the square-
wave edge. This should not be confused with the exponential rounding of square-wave edges
which results from a high-frequency rolloff (the effect measured by rise time). The effects of
bandwidth limiting are linear; the effects of slew-rate limiting are not. A simple way to deter-
mine whether a signal is slew-limited is to increase the signal amplitude. A bandwidth-limited
signal will make the transition in the same time; a slew-limited signal will take longer.
Tone bursts are another technique for evaluating the response of audio devices to transients.
They are created by gating a sine wave on and off at its zero crossings. A tone burst concentrates
the energy of the waveform closer to a particular frequency, enabling evaluation of individual
sections of the audio-frequency range. However, they still contain substantial high-frequency
energy as shown in Figure 10.4.3. The number of cycles on and off significantly affects the fre-
quency spread of the energy. This frequency spreading can yield anomalous results, requiring
extreme care in interpretation. Linkwitz [2] proposed the use of shaped bursts, employing a win-
dowing function much the same as the Hamming window employed in FFT analysis. The shap-
ing of the burst rise and fall reduces the spread in frequency, concentrating the energy near the
frequency of the sine wave.
Tone-burst testing is common with loudspeakers, yielding qualitative information on the
damping characteristics of the drivers at a glance. As with square waves, the common parameters
specified in tone-burst measurements are overshoot and ringing. The overshoot or undershoot on
a tone burst is usually taken to be the amount by which the burst envelope goes above or below
the steady-state on level. Ringing on a tone burst refers to the tendency to continue oscillating
after the burst has stopped. This gives the appearance of a tail which continues after the body of
the burst.
10.4.2a Level-Gated Burst Measurements

The bursts considered so far have been sine waves gated completely off. It is often useful to gate
a sine wave between two different levels. The most common applications of this procedure are
testing amplifiers and compressors. When an amplifier is driven into clipping, its circuitry loses
feedback and portions of the circuitry become saturated. As the amplifier comes out of this over-
load condition, it may do unpredictable things. These are often easier to see when a signal is
present. A waveform such as a dual-level tone burst can overload the amplifier periodically while
maintaining signal between overloads, enabling examination of this behavior. Compressors,
expanders, limiters, and similar devices are designed to change the dynamic range of audio sig-
nals. To evaluate the dynamic, or time, behavior of these devices it is necessary to have a signal
which has significant changes in level with time. The two-level tone burst is ideal for this appli-
cation, enabling observation of the transition of the compressor from one gain to another. An
example of typical compressor behavior is shown in Figure 10.4.4.
Tone bursts may be used to make response measurements in the presence of interfering reflec-
tions such as occur in rooms. The basic principle is to stimulate the device under test with a burst
and perform measurements on the resulting signal with a gate (electronic switch) that eliminates
later-arriving reflections. The amplitude of the impulse is sampled during the body of the burst
after any ringing on the leading edge has died away. The sample gate is turned off before any
reflections have arrived, yielding the free-field response. Linkwitz [2] modified this procedure to
use a shaped tone burst that concentrates the energy much more within the frequency band of
interest. This allows the maximum amplitude of the burst to be measured, eliminating the need to
gate the measurement during the center portion of the burst. The reflections may usually be
ignored because they are lower in level than the direct signal.
(a)
(b)
Figure 10.4.3 Tone-burst spectra: (a) single-cycle sine-wave burst, (b) five-cycle sine-wave burst.
(From [2]. Used with permission.
Figure 10.4.4 Typical tone-burst response of a compressor, illustrating the time response of the
compressor. (From [3]. Used with permission.)
10.4.2b Impulse Response

Another method for checking the time-domain behavior of devices is the impulse response. The
device under test is excited by a very narrow pulse whose amplitude is set to the maximum
allowed for linear operation. The response may be viewed on an oscilloscope for qualitative judg-
ments. However, the power of the technique lies in capturing the response with a digital oscillo-
scope or FFT analyzer. Performing an FFT will yield the frequency-domain behavior of the
device, both amplitude and phase. Under most conditions this information is identical to what
would be obtained by a sine-wave-based sweep of frequency. If linear operation is maintained by
proper choice of levels, the only differences will be in terms of heating effects (as occur in loud-
speaker voice coils) and dynamic level effects (as occur in compressors).
10.4.3 References
1. Tremaine, H. M.: Audio Cyclopedia, Howard W. Sams, Indianapolis, IN, 1975.
2. Linkwitz, S.: “Narrow Band Testing of Acoustical Systems,” Audio Engineering Society
preprint 1342, AES, New York, N.Y., May 1978.
3. Cabot, R. C.: “Limiters, Compressors, and Expanders,” Sound & Video Contractor, Inter-
tec Publishing, Overland Park, Kan., vol. 26, November 1985.
10.4.4 Bibliography
Naerum, Denmark.
verton, Ore., 1975.
Section
Video Signal Measurement and Analysis

11
As video technology steams full speed ahead into the digital domain, the shortcomings and deg-
radations associated with analog technology—which video engineers have come to accept and
deal with—are rapidly disappearing. In their place, however, are new problems. Digital devices
and systems bring their own unique mix of issues that must be addressed, including:
• Quantization. The quantization process, by design, discards information. It takes an analog
waveform with infinite variability and blocks it into a collection of bits, the number of which
is determined by the bit length of the system.
• Concatenation. Defined as the connection of elements end-to-end, concatenation for video
and audio describes the effects of chaining compression and decompression systems.
• Video Processing. It is commonly assumed that as long as a video clip is manipulated in the
digital domain, it will not be degraded. In a general sense this is true, however, certain opera-
tions will discard information that can not be recreated downstream. Changes in sizing,
adjustment of color hue and saturation, and adjustment of luminance values are just some of
the operations that can result in degradation of the signal unless proper precautions are taken.
Something as simple as improper gamma setup on monitors can result in a host of problems
as the signal meanders through the production process. Once picture information is discarded,
it cannot be completely recreated.
• Transmission. In order for a digital video signal to be useful, it usually must be moved from
one location to another. This almost always involves codecs and a transmission medium. This
medium may be coax, fiber, or a radio frequency link. With any of these systems, degrada-
tions are possible; some are more vulnerable than others. An RF link usually has the greatest
level of exposure to interfering signals. Coax, on the other hand, is basically closed to outside
influences but has a finite cable length over which reliable communications can take place.
The important message here is that digital is not always perfect and that the need for test equip-
ment and quality control does not disappear simply because a room full of analog boxes is
replaced with a computer workstation. Furthermore, just because the picture “looks good” on a
local monitor does not mean that it will look good (or at least look the same) at the end of a ter-
restrial or satellite link.
11-1
-
11-2 Section Eleven
New test instruments are rising to the challenge posed by the new technologies being intro-
duced to the video production process. As the equipment used by broadcasters and video profes-
sionals becomes more complex, the requirements for advanced, specialized maintenance tools
also increases. These instruments range from simply go/no-go status indicators to automated test
routines with preprogrammed pass/fail limits. Video quality control efforts must focus on the
overall system, not just a particular island.
The attribute that makes a good test instrument is really quite straightforward: accurate mea-
surement of the signal under test. The attributes important to the user, however, usually involve
the following:
• Affordability
• Ease of use
• Performance
Depending upon the application, the order of these criteria may be inverted (performance, ease
of use, then affordability). Suffice it to say, however, that all elements of these specifications
combine to translate into the user’s definition of the ideal instrument.
Computer-based video test instruments provide the maintenance engineer with the ability to
rapidly measure a number of parameters with exceptional accuracy. Automated instruments offer
a number of benefits, including reduced setup time, test repeatability, waveform storage and
transmission capability, and remote control of instrument/measurement functions.
The memory functions of the new breed of instruments provide important new capabilities,
including archiving test setups and reference waveforms for ongoing projects and comparative
tests. Hundreds of files can be saved for later use. With automatic measurement capabilities,
even a novice technician can perform detail-oriented measurements quickly and accurately.
In the rush to embrace advanced, specific-purpose test instruments, it is easy to overlook the
grandparents of all video test devices—the waveform monitor and vectorscope. Just because
they are not new to the scene does not mean that they have outlived their usefulness.
The waveform monitor and vectorscope still fill valuable roles in the test and measurement
world. Both, of course, have their roots in the general-purpose oscilloscope. This heritage
imparts some important benefits. The scope is the most universal of all instruments, combining
the best abilities of the human user and the machine. Electronic instruments are well equipped to
quickly and accurately measure a given amplitude, frequency, or phase difference; they perform
calculation-based tasks with great speed. The human user, however, is far superior to any
machine in interpreting and analyzing an image. The waveform monitor and vectorscope—in an
instant—presents to the user a wealth of information that allows rapid characterization and
understanding of the signal under consideration.
In This Section:
Chapter 11.1: Video Information Concepts 11-7

Introduction 11-7
Information Content of Video Signals 11-8
Picture Element Characteristics 11-10
Vertical and Horizontal Resolution 11-12
The Video Spectrum 11-13
Minimum Video Frequency 11-15
-
Video Signal Measurement and Analysis 11-3
Maximum Video Frequency 11-15

Horizontal Resolution 11-16
Video Frequencies Arising From Scanning 11-16
References 11-18
Chapter 11.2: Measurement of Color Displays 11-19

Introduction 11-19
Assessment of Color Reproduction 11-21
Chromatic Adaptation and White Balance 11-22
Overall Gamma Requirements 11-22
Perception of Color Differences 11-23
Display Resolution and Pixel Format 11-24
Contrast Ratio 11-24
Display Measurement Techniques 11-25
Subjective CRT Measurements 11-25
Shrinking Raster Method 11-26
Line-Width Method 11-27
TV Limiting Resolution Method 11-27
Application Considerations 11-28
Objective CRT Measurements 11-28
Half-Power-Width Method 11-28
Fourier Transform Methods 11-29
Discrete Frequency Method 11-30
Viewing Environment Considerations 11-31
Picture Monitor Alignment 11-31
References 11-33
Chapter 11.3: Camera Performance Verification 11-35

Introduction 11-35
Visual Inspection and Mechanical Check 11-35
Electrical Parameters and Measurements 11-36
Confirmation of the Camera Encoder 11-37
Confirmation of Auto Black 11-37
Lens Back-Focus 11-37
Black Shading 11-38
Detail Circuit 11-38
Optional Tests 11-38
White Shading 11-38
Flare 11-39
Linear Matrix 11-39
Color Reference Pattern 11-40
References 11-40
Bibliography 11-41
Chapter 11.4: Conventional Video Measurements 11-43

Introduction 11-43
Color Bar Test Patterns 11-43
HD/SD Video Test Signal 11-46
Basic Waveform Measurements 11-47
-
11-4 Section Eleven
Sync Pulses 11-49

Basic Vectorscope Measurements 11-49
Line Select Features 11-49
Distortion Mechanisms 11-50
Differential Phase 11-52
Differential Gain 11-55
ICPM 11-55
Tilt and Rounding 11-56
Group Delay 11-58
Classification of Distortion Mechanisms 11-59
Automated Video Signal Measurement 11-63
References 11-66
Bibliography 11-67
Chapter 11.5: Application of the Zone Plate Signal 11-69

Introduction 11-69
Simple Zone Plate Patterns 11-70
Producing the Zone Plate Signal 11-73
Complex Patterns 11-73
The Time (Motion) Dimension 11-74
References 11-77
Chapter 11.6: Picture Quality Measurement 11-79

Introduction 11-79
Signal/Picture Quality 11-81
Automated Picture-Quality Measurement 11-82
In-Service Measurements 11-84
References 11-85
Chapter 11.7: DTV Transmission Performance Issues 11-87

Introduction 11-87
Characterizing the Transmission System 11-87
Linear Errors 11-88
Nonlinear Errors 11-88
Phase Noise 11-89
Bit Error Rate 11-89
Transmission System Measurements 11-90
Emissions Mask 11-90
Sideband Splatter 11-90
In-Band Signal Characterization 11-92
Power Specification and Measurement 11-92
Measurement Routine 11-93
DTV Operating Frequency Considerations 11-93
Frequency Tolerances 11-93
Hardware Options for Tight Frequency Control 11-94
Additional Considerations 11-94
References 11-95
-
Video Signal Measurement and Analysis 11-5

“Alignment of NTSC Color Picture Monitors,” SMPTE Recommended Practice RP 167,
SMPTE, White Plains, N.Y., 1995.
ANSI Standard T1.801.03-1996, “Digital Transport of One-Way Video Signals: Parameters for
Objective Performance Assessment,” ANSI, Washington, D.C., 1996.
ATSC: “Guide to the Use of the ATSC Digital Television Standard,” Advanced Television Sys-
tems Committee, Washington, D.C., Doc. A/54A, 2 December 2004.
ATSC: “Transmission Measurement and Compliance for Digital Television,” Advanced Televi-
sion Systems Committee, Washington, D.C., Doc. A/64-Rev A, May 30, 2000.
Bender, Walter, and Alan Blount: “The Role of Colorimetry and Context in Color Displays,”
Human Vision, Visual Processing, and Digital Display III, Bernice E. Rogowitz (ed.), Proc.
SPIE 1666, SPIE, Bellingham, Wash., pp. 343–348, 1992.
Bentz, Carl: “Inside the Visual PA, Part 2,” Broadcast Engineering, Intertec Publishing, Over-
land Park, Kan., November 1988.
Bentz, Carl: “Inside the Visual PA, Part 3,” Broadcast Engineering, Intertec Publishing, Over-
land Park, Kan., December 1988.
Bentz, Carl, and Jerry C. Whitaker: “Video Transmission Measurements,” in Maintaining Elec-
tronic Systems, CRC Press, Boca Raton, Fla., pp. 328–346, 1991.
Bishop, Donald M.: “Practical Applications of Picture Quality Measurements,” Proceedings of
Digital Television '98, Intertec Publishing, Overland Park, Kan., 1998.
“Broadening the Applications of Zone Plate Generators,” Application Note 20W7056, Tektronix,
Beaverton, Ore., 1992.
“Critical Viewing Conditions for Evaluation of Color Television Pictures,” SMPTE Recom-
mended Practice RP 166, SMPTE, White Plains, N.Y., 1995.
DTV Express Training Manual on Terrestrial DTV Broadcasting, Harris Corporation, Quincy,
Ill., September 1998.
“Eye Diagrams and Sampling Oscilloscopes,” Hewlett-Packard Journal, Hewlett-Packard, Palo
Alto, Calif., pp. 8–9, December 1996.
Fibush, David K.: “Picture Quality Measurements for Digital Television,” Proceedings of the
Digital Television ’97 Summit, Intertec Publishing, Overland Park, Kan., December 1997.
Fibush, David K.: “Practical Application of Objective Picture Quality Measurements,” Proceed-
ings IBC ’97, IEE, pp. 123–135, Sept. 16, 1997.
Finck, Konrad: “Digital Video Signal Analysis for Real-World Problems,” in NAB 1994 Broad-
cast Engineering Conference Proceedings, National Association of Broadcasters, Washing-
ton, D.C., pg. 257, 1994.
Gloeggler, Peter: “Video Pickup Devices and Systems,” in NAB Engineering Handbook,” 9th
Ed., Jerry C. Whitaker (ed.), National Association of Broadcasters, Washington, D.C.,
1999.
-
11-6 Section Eleven
Hamada, T., S. Miyaji, and S. Matsumoto: “Picture Quality Assessment System by Three-Lay-
ered Bottom-Up Noise Weighting Considering Human Visual Perception,” SMPTE Jour-
nal, SMPTE, White Plains, N.Y., pp. 20–26, January 1999.
MacAdam, D. L.: “Visual Sensitivities to Color Differences in Daylight,” J. Opt. Soc. Am., vol.
32, pp. 247–274, 1942.
Mertz, P.: “Television and the Scanning Process,” Proc. IRE, vol. 29, pp. 529–537, October 1941.
Pank, Bob (ed.): The Digital Fact Book, 9th ed., Quantel Ltd, Newbury, England, 1998.
Reed-Nickerson, Linc: “Understanding and Testing the 8-VSB Signal,” Broadcast Engineering,
Intertec Publishing, Overland Park, Kan., pp. 62–69, November 1997.
Robin, M., and M. Poulin: Digital Television Fundamentals, 2nd ed., McGraw-Hill, New York,
N.Y., 2001.
SMPTE Engineering Guideline EG 1, “Alignment Color Bar Test Signal for Television Picture
Monitors,” SMPTE, White Plains, N.Y., 1990.
“SMPTE Standard: For Television—Color Reference Pattern,” SMPTE 303M, SMPTE, White
Plains, N.Y., 1999.
SMPTE Standard: SMPTE 259M, “Serial Digital Interface for 10-bit 4:2:2 Components and
4Fsc NTSC Composite Digital Signals,” SMPTE, White Plains, N.Y., 1997.
Standards and Definitions Committee, Society for Information Display.
Stremler, Ferrel G.: “Introduction to Communications Systems,” Addison-Wesley Series in Elec-
trical Engineering, Addison-Wesley, New York, December 1982.
Tannas, Lawrence E., Jr.: Flat Panel Displays and CRTs, Van Nostrand Reinhold, New York, pg.
18, 1985.
Uchida, Tadayuki, Yasuaki Nishida, and Yukihiro Nishida: “Picture Quality in Cascaded Video-
Compression Systems for Digital Broadcasting,” SMPTE Journal, SMPTE, White Plains,
N.Y., pp. 27–38, January 1999.
Verona, Robert: “Comparison of CRT Display Measurement Techniques,” Helmet-Mounted Dis-
plays III, Thomas M. Lippert (ed.), Proc. SPIE 1695, SPIE, Bellingham, Wash., pp. 117–
127, 1992.
-
Chapter
11.1
Video Information Concepts
Donald G. Fink, Bernard D. Loughlin, Donald C. Livingston, William F. Bailey
11.1.1 Introduction1
For the purpose of this handbook, video information may be defined as data conveying a descrip-
tion of a picture that can be displayed by an appropriate picture-reproducing system through the
use of television signals. When the information is that possessed by the electrical signals occur-
ring at a given location in a television system, the character of the information can be interpreted
in terms of the picture which would be displayed by some reference reproducer when driven by
the given signals. For purposes of this handbook, the reference reproducer is here defined as a
television signal-processing and display system which correctly displays the full information
content of any set of signals that have been properly formed and have been properly inserted into
the signal-processing circuits of the system. All the signal-processing circuits are assumed to be
free of distortion, noise, and interference. The only nonlinear element in the reference reproducer
is taken to be its picture display device, whose photoelectric transfer characteristics are specified
to be the same as those of the program director’s studio monitor.
In conventional monochrome and color systems, the information in the associated pictures
and signals can be broadly classified as:
• Monochrome information: the information carried by the monochrome signal in both mono-
chrome and color systems
• Coloring information: the information carried by all signals other than the monochrome sig-
nal in a color television system
Monochrome information by itself describes a monochrome picture without reference to its
chromaticity. Monochrome and coloring information together describe a color picture.
When it is desired to discuss the types of information carried by the signals in the primary-
color signal channels of a color television transmitter or receiver, the following classifications
are pertinent:
• Red-primary information: the information carried by a red-primary signal
1. Portions of this chapter are adapted from: D. G. Fink (ed.): Television Engineering Handbook,
McGraw-Hill, New York, N.Y., 1957. Used with permission.
11-7
-
11-8 Video Measurement Techniques
• Green-primary information: the information carried by a green-primary signal

• Blue-primary information: the information carried by a blue-primary signal
If a picture or a scene is under discussion without reference to any electrical signals that may
be associated with it, a useful classification for the types of information contained in it is:
• Luminance information: the information that describes the distribution of luminance levels in
the picture
• Chromaticity information: the information that describes the distribution of chromaticity val-
ues in the picture if it is a color picture
There are two methods for classifying the information in a monochrome picture: line infor-
mation in the picture associated with a monochrome television system can be called mono-
chrome information in general and luminance information in particular. It is monochrome
information because it is brought to the picture by a monochrome signal; it is luminance infor-
mation because it describes the distribution of luminance levels in the picture.
In the same way, the information content of a color picture produced by a compatible color
television system can be described in at least three ways:
• It can be said to consist of a combination of monochrome information and coloring informa-
tion because it is brought to the picture by the monochrome signal and by an auxiliary color-
ing signal (the chrominance signal).
• It can be said to consist of a combination of luminance and chromaticity information because
it describes the distribution of luminance levels and chromaticity values in the picture.
• Finally, it can be said to consist of red-, green-, and blue-primary information on analogous
grounds.
11.1.2 Information Content of Video Signals

The information content of video signals can be described in terms of the picture which they
would elicit from a reference reproducer. Various systematic procedures for carrying out such
descriptions can be developed through use of the following: 1) luminance level, 2) monochrome
level, 3) chromaticity value, 4) raster coordinates, 5) point image, 6) line image, and 7) picture
element.
The luminance level at a given point in a picture is the ratio of the luminance at that point to
the maximum luminance occurring anywhere in the picture during some appropriate interval of
time. Its value can lie anywhere between zero and unity. The reciprocal of the lowest luminance
level in the scene is called the contrast range of that scene.
The monochrome level at a given point in a color television picture produced through use of
the transmission signals of a compatible color television system is the luminance level in the
monochrome picture that appears in the absence of the chrominance signal.
The chromaticity value at a given point in a color television picture is the quantitative specifi-
cation of the chromaticity at that point in terms of a suitable set of chromaticity coordinates.
The raster coordinates of a given point in a picture are a pair of numbers specifying the posi-
tion of that point with respect to a suitable set of reference axes. The horizontal and vertical axes
may be chosen along the upper edge and left-hand edge of the raster, respectively, and a distance
-
equal to the raster height may be designated as unity along both axes. Thus, raster coordinates of
a point in a raster with a 4:3 aspect ratio would range between zero and unity for the vertical
coordinate and between zero and 1.33 for the horizontal one.
A point image is the pattern of light formed on a television display in response to light
received by a television camera from a single object point in its field of view. The position of any
given point image on a reproducing raster is determined by the position of the corresponding
object point together with whatever geometric distortions (e.g., pin-cushioning) may exist in the
raster. Its position is also influenced by the line structure in the scanning raster unless its size is
large in comparison with the distance between adjacent scanning lines, as happens when the
object point lies appreciably outside the camera field of focus.
A line image is the pattern of light formed on a television display in response to light received
by a television camera from a uniformly illuminated ideally narrow line in its field of view. The
position of any given line image on a reproducing raster is determined by the position of the cor-
responding object line in a manner analogous to that for a point image. A line image oriented
perpendicular to the scanning path may be called a transverse line image, and one parallel to this
path may be called a longitudinal line image.
A picture element is the smallest area of a television picture capable of being delineated by an
electrical signal passed through the system or part thereof. The number of picture elements (pix-
els) in a complete picture, and their geometric characteristics of vertical height and horizontal
width, provide information on the total amount of detail which the raster can display and on the
sharpness of that detail, respectively.
The description of the information content of a picture through use of the above quantities
may be carried out in varying degrees of exactness. In the case of a monochrome picture, the
method used would be expected to lead to the specification of definite numerical values for the
displayed luminance level at points having given raster coordinates. If the monochrome picture
concerned is that resulting from withdrawal of the chrominance signal from a color display dur-
ing a color transmission, the observed luminance levels represent monochrome levels of the
color picture. In the case of a color picture, the method would lead to the specification of numer-
ical values for both luminance levels and chromaticity values at points having given raster coor-
dinates.
The most direct rigorous approach involves specification of numerical values at all points in
the picture. This requires associating the pertinent colorimetric values with the numerical values
of explicit mathematical functions of the raster coordinates. In general, these are continuous
functions because of the continuous nature of the distribution functions representing typical
aperture transmittances.
It is meaningful to specify only as much information as can be recognized to consist of sepa-
rate and independent elements. To carry out this type of description, the raster area can be subdi-
vided into picture elements, the total number of elements being determined by the spectrum
characteristics of the monochrome signal and by the percentage of the total frame scanning time
that is consigned to blanking intervals. The video signal waveform for one frame period can be
regarded as a portion of a repetitive wave corresponding to a stationary picture, and its spectrum
then turns out to have a finite number of discrete frequency components. The total information
capacity of this spectrum is twice the number of components, since each component can be
adjusted in both amplitude and phase.
The waveform for a single frame has the same total information capacity, and the waveform
for any fractional part of a frame has a corresponding fraction of this total information capacity.
Consequently, the portion of the waveform used for active scanning of the raster during one
-
frame period has an information capacity equal to twice the number of frequency components in
the spectrum for the signal multiplied by the number representing the fractional portion of the
frame period used in active scanning. This information capacity is a number that may be inter-
preted as the number of picture elements in the raster area. The elements may be regarded as
being uniformly distributed over the raster in the sense that their separations along the scanning
lines represent distances traversed by the scanning spot in equal times.
11.1.2a Picture Element Characteristics

An array of picture elements can be considered to be distributed over the television raster.
Because the information content of a signal can be described in terms of the picture that the sig-
nal might elicit from a reference reproducer, it follows that picture elements can be associated
with signals as well as with pictures.
The specification for the dimensions of the picture element leads to a particularly simple
result in a television system wherein the camera scanning aperture is a mathematical point, all
signal channels have sharp cutoffs at some frequency below which transmission is distortionless,
and all transducers including the picture display device are linear. It is clear that this signal itself
has the form of an ideal impulse function. Its frequency expression is given by
∞
E 0 ( t ) = -1- ∫ cos 2 π ft df (11.1.1)
π 0
when its amplitude has been chosen so that
∞
∫ – ∞ E ( t ) dt = 1 (11.1.2)
f and t here denote frequency and time, respectively. After band limiting, the frequency expan-
sion of the resulting signal becomes
fc
E ( t ) = -1- ∫ a ( f ) cos [ 2 πft – α( f ) ] df (11.1.3)
π 0
in which a (f) and α (f) are the attenuation and phase shift, respectively, of the channel at fre-
quency f, and fc is the frequency above which no appreciable transmission occurs.
The luminance distribution across the associated transverse line image of a signal in this sys-
tem has the form of the impulse response of the signal channel, which, in turn, is a curve of the
form sin x / x. Specifically, it is given by Equation (11.1.3) with a (f) constant and α (f) equal to
zero. This integrates to
f sin 2 πf t
E ( t ) = ⎛⎝ --c-⎞⎠ ⎛⎝ -----------------c---⎞⎠ (11.1.4)
π 2 πf t
c
where fc is the cutoff frequency of the channel.
-
Note that the function E(t) has its maximum value at t = 0 and that it extends over the time
range t = ± 1/4 fc. The time interval Δ t = 1/2 fc is then representative of the maximum distance
through which the line image can be moved along the scanning path so that the luminance at a
given point on the raster reaches the value E(0) at an internal point in the range and is down to
2E(0) / π at either end point. Specifically, this distance is the distance traversed by the scanning
spot in time Δ t. The length of the picture element is defined to be this distance.
The number of picture elements in the raster is twice the maximum number of frequency
components that the given signal can possess, multiplied by the factor representing the fractional
portion of the total frame scanning interval used in active raster scanning. The elements are spec-
ified to be distributed uniformly over the raster, the distance between centers of adjacent ele-
ments on a given scanning path being fixed. That this distance is precisely equal to the picture-
element length in the television system under discussion can be shown as follows.
The spectrum for a signal representing a stationary picture can be shown to contain only fre-
quencies that are integral multiples of the picture repetition frequency f. Thus, if the channel cut-
off frequency is fc, the signal can contain not more than fc / f frequency components. This
corresponds to 2 fc / f information elements, and these are to be conceived as distributed uni-
formly throughout the raster scanning and blanking intervals. Picture elements are simply those
information elements that occur during the raster scanning intervals. The time interval between
successive elements of either type is evidently given by the frame scanning interval 1/f divided
by the total number 2 fc / f of information elements occurring in that time. This value turns out to
be 1/fc, which is precisely the time interval Δ t found just previously to be indicative of the pic-
ture-element length. It follows that in the system under discussion, adjacent picture elements on
a scanning path are in contact with each other, although they do not overlap. It is in order to bring
about this particular result that the picture-element length is defined in terms of luminance ratios
of 2/π.
The definition for picture-element width leads to a similar result. In the system under discus-
sion, displacing a longitudinal line image from exact coincidence with a given scanning path
results in an abrupt drop of the displayed luminance from the value characterizing the line image
to zero. Hence, the added condition appended to the definition becomes applicable and pre-
scribes that the picture-element width in this case is equal to the raster pitch distance.
It is thus seen that in a television system having a mathematical-point scanning aperture in the
camera, having sharp-cutoff signal transmission channels with distortionless passbands, and hav-
ing no nonlinear transducers, the picture elements exactly fill the raster area without overlapping.
Each such picture element may be called an ultimate picture element for the given television
system. It has the smallest possible dimensions that the picture element in a linear system with
given cutoff frequency can possess. Enlarging the camera scanning aperture from a point to a
configuration with finite dimensions immediately brings about an increase in the picture-ele-
ment length; a considerably enlarged aperture also causes the picture-element width to increase
so that overlap occurs both longitudinally and transversely. Replacement of the sharp-cutoff fre-
quency characteristic of the signal channel by a gradual one with the same cutoff frequency
changes the impulse response from that in Equation (11.1.4) to the more general form in Equa-
tion (11.1.3). The latter invariably yields a greater interval between points for which E(t) has a
value equal to a fraction 2/π of its maximum value than does the former. Lowering the cutoff fre-
quency of the channel causes the total number of picture elements to decrease and causes each
element to become longer.
In a nonlinear system, the actual picture element can be smaller than the ultimate element.
This can occur, for example, when the picture display device has a gamma exponent greater than
-
Table 11.1.1 Data on Picture Elements in Various Analog Television Systems
Parameter CCIR Standard Designation

M A B/G L E
U.S. U.K CCIR France France
NTSC MONO PAL SECAM MONO
(note 2) (note 3)
Information elements per frame 280,000 240,000 400,000 480,000 800,000
Picture elements per raster 210,000 180,000 300,000 360,000 590,000
Picture elements per line 440 480 520 620 790
Picture element length 30 28 26 21 17
See note (1).
Picture element width 21 27 17 17 13
See note (1).
Notes:
(1) Based on raster height = 10,000 units
(2) Discontinued service
unity. This is significant since the gamma exponent is usually found to lie in the range γ = 2.2
± 0.2 for conventional cathode-ray picture tubes. In practice, the picture-element length is never-
theless usually larger than that of the ultimate element because of the effects of camera aperture
size and the gradual cutoff in the frequency characteristic of the signal channel. The width, how-
ever, is ordinarily that of the ultimate element.
The total number of ultimate picture elements in the raster of a given television system is the
product of the number 2 fc / f of information elements in the frame interval and the percentage of
the frame interval consumed in active scanning of the raster. Thus, if a percentage PH of each
line and a percentage PV of the total number of lines are used in active scanning of the raster,
then there are 2PH PV fc / f picture elements in the raster. Moreover, the number of picture ele-
ments on each line of the raster is given by this number divided by the number of lines in the ras-
ter. The number of lines in a frame is fH / f, where fH is the line scanning frequency, and the
number of lines in the raster is therefore PV fH / f. Consequently, there are 2PH fc / fH picture ele-
ments on each raster line.
A list of picture-element statistics for various television systems is given in Table 11.1.1.
11.1.2a Vertical and Horizontal Resolution

The total amount of detail that can be resolved simultaneously by a given signal is specified by
the ratio of the raster area to the resolution-element area. The vertical resolution is measured by
the ratio of the raster dimension across the scanning path to the resolution-element width. This
number is the maximum number of uniform longitudinal stripes that can be fully resolved by the
signal when the edge of at least one stripe coincides with a scanning path. The horizontal resolu-
tion is measured by the ratio of the raster dimension across (not parallel to) the scanning path to
the resolution-element length. The resolution-element length would correspond to half a cycle at
the channel cutoff frequency if the channel had a sharp cutoff characteristic and if the camera
scanning aperture was a mathematical point. Under these conditions, the length of the picture
-
Table 11.1.2 Data on Resolution Elements in Various Analog Television Systems
Parameter CCIR Standard Designation

M A B/G L E
U.S. U.K CCIR France France
NTSC MONO PAL SECAM MONO
(note 1) (note 2)
Resolution elements per raster 150,000 130,000 210,000 250,000 410,000
Vertical resolution 340 260 400 400 530
Horizontal resolution 330 360 390 470 590
Notes:
element would also have this value. The resolution-element width, however, would be equal to
the raster pitch distance divided by the Kell factor, or to 1.4 times the raster pitch distance. At the
same time, the picture-element width would be exactly equal to the raster pitch distance. Just as
the picture element under this special condition may be called the ultimate picture element, the
resolution element under this same condition can be called the ultimate resolution element. Its
dimensions indicate the maximum resolving power of which the given television system is capa-
ble.
When the signal-channel cutoff frequency is unchanged although cutoff is made gradual
rather than sharp, both picture element and resolution element become longer. Their widths
remain unchanged as long as the camera scanning aperture remains a mathematical point. When
the camera aperture is enlarged beyond a diameter comparable with the raster pitch distance,
both picture-element and resolution-element width become greater than their ultimate values,
and it cannot be tacitly assumed that the ratio of their widths remains equal to the Kell factor.
Actual dimensions of the resolution element in a given television system cannot be calculated
on theoretical grounds but must be established by subjective experimental observation. Statistics
on the ultimate resolution element, however, are readily calculated on the basis of the specifica-
tions given previously. They are presented for various television systems in Table 11.1.2, a value
of 0.70 being assumed for the Kell factor.
11.1.3 The Video Spectrum

The spectrum of the video signal arising from the scanning process in a television camera
extends from a lower limit determined by the timed rate of change of the average luminance of
the scene to an upper limit determined by the time during which the scanning spots cross the
sharpest vertical boundary in the scene as focused within the camera. This concept is illustrated
in Figure 11.1.1. The distribution of spectral components within these limits is determined by the
following:
• The distribution of energy in the camera scanning system
• Number of lines scanned per second
• Percentage of line-scan time consumed by horizontal blanking
-
(a )
(b )
Figure 11.1.1 Video signal spectra: (a) camera scanning spot, shown with a Gaussian distribution,
passing over a luminance boundary on a scanning line; (b) corresponding camera output signal
resulting from the convolution of the spot and luminance distributions.
• Number of fields or frames scanned per second

• Rates at which the luminance and chrominance values of the scene change in size, position,
and boundary sharpness
To the extent that the contents and dynamic properties of the scene cannot be predicted, the
spectrum limits and energy distribution are not defined. However, the spectra associated with
certain static and dynamic test charts and waveform generators may be used as the basis for
video system design and testing. Among the configurations of interest are:
• Flat fields of uniform luminance and/or chrominance
• Fields divided into two or more segments of different luminance by sharp vertical, horizontal,
or oblique boundaries
The case of the divided fields includes the horizontal and vertical wedges of test charts and the
concentric circles of zone plate charts, illustrated in Figure 11.1.2. The reproductions of such
patterns typically display diffuse boundaries and other degradations that may be introduced by
the camera scanning process, the amplitude and phase responses of the transmission system, the
receiver scanning system, and other artifacts associated with scanning, encoding, and transmis-
sion.
-
(a) (b) (c)
(d) (e) (f)
(g) (h) ( i)
Figure 11.1.2 Scanning patterns of interest in analyzing conventional video signals: (a), (b), (c)
flat fields useful for determining color purity and transfer gradient (gamma); ( d) horizontal half-field
pattern for measuring low-frequency performance; (e) vertical half field for examining high-fre-
quency transient performance; (f) display of oblique bars; (g) in monochrome, a tonal wedge for
determining contrast and luminance transfer characteristics; in color, a display used for hue mea-
surements and adjustments; (h) wedge for measuring horizontal resolution; (i) wedge for measur-
ing vertical resolution.
11.1.3a Minimum Video Frequency

To reproduce a uniform value of luminance from top to bottom of an image scanned in the con-
ventional interlaced fashion, the video signal spectrum must extend downward to include the
field-scanning frequency. This frequency represents the lower limit of the spectrum resulting
from scanning an image whose luminance does not change. Changes in the average luminance
are reproduced by extending the video spectrum to a lower frequency equal to the reciprocal of
the duration of the luminance change. Because a given average luminance may persist for many
minutes, the spectrum extends essentially to zero frequency (dc). Various techniques of preserv-
ing or restoring the dc component are employed in conventional television to extend the spec-
trum from the field frequency down to dc.
11.1.3a Maximum Video Frequency

In the analysis of the maximum operating frequency for a conventional video system, three val-
ues must be distinguished:
-
• The maximum output signal frequency generated by the camera or other pickup/generating
device
• Maximum modulating frequency corresponding to: 1) the fully transmitted (radiated) side-
band, or 2) the system used to convey the video signal from the source to the display
• Maximum video frequency present at the picture-tube (display) control electrodes
The maximum camera frequency is determined by the design and implementation of the
imaging element. The maximum modulating frequency is determined by the extent of the video
channel reserved for the fully transmitted sideband. The channel width, in turn, is chosen to pro-
vide a value of horizontal resolution approximately equal to the vertical resolution implicit in the
scanning pattern. The maximum video frequency at the display is determined by the device and
support circuitry of the display system.
11.1.3a Horizontal Resolution

The horizontal resolution factor is the proportionality factor between horizontal resolution and
video frequency. It may be expressed as
Rh
Hr = -----
α
× ι (11.1.5)
Where:
Hr = horizontal resolution factor in lines per megahertz
Rh = lines of horizontal resolution per hertz of the video waveform
α = aspect ratio of the display
ι = active line period in microseconds
For NTSC, the horizontal resolution factor is:
2
78.8 = ------------ × 52.5 (11.1.6)
4⁄ 3
11.1.3a Video Frequencies Arising From Scanning

The signal spectrum arising from the scanning process comprises a number of discrete compo-
nents at multiples of the scanning frequencies. Each spectrum component is identified by two
numbers, m and n, which describe the pattern that would be produced if that component alone
were present in the signal. The value of m represents the number of sinusoidal cycles of bright-
ness measured horizontally (in the width of the picture) and n the number of cycles measured
vertically (in the picture height). The 0, 0 pattern is the dc component of the signal, the 0, 1 pat-
tern is produced by the field-scanning frequency, and the 1, 0 pattern is produced by the line-
scanning frequency. Typical patterns for various values of m and n are shown in Figure 11.1.3.
By combining a number of such patterns (including m and n values up to several hundred), in the
appropriate amplitudes and phases, any image capable of being represented by the scanning pat-
-
p
tern may be built up. This is a 2-dimensional -4

form of the Fourier series.
The amplitudes of the spectrum components
-3
decrease as the values of m and n increase.
Because m represents the order of the harmonic
of the line-scanning frequency, the correspond- -2
ing amplitudes are those of the left-to-right
variations in brightness. A typical spectrum -1
resulting from scanning a static scene is shown
in Figure 11.1.4. The components of major n 0
magnitude include:
• The dc component 1
• Field-frequency component
2
• Components of the line frequency and its
harmonics
3
Surrounding each line-frequency harmonic is a
cluster of components, each separated from the
4
next by an interval equal to the field-scanning
frequency.
0 1 2 3 4
It is possible for the clusters surrounding
m
adjacent line-frequency harmonics to overlap
one another. As shown in Figure 11.1.4, two Figure 11.1.3 An array of image patterns cor-
patterns situated on adjacent vertical columns responding to indicated values of m and n.
produce the same value of video frequency (After [1].)
when scanned. Such “intercomponent confu-
sion” of spectral energy is fundamental to the
scanning process. Its effects are visible when a heavily striated pattern (such as that of a fabric
with an accented weave) is scanned with the striations approximately parallel to the scanning
lines. In the NTSC and PAL color systems, in which the luminance and chrominance signals
occupy the same spectral region (one being interlaced in frequency with the other), such inter-
component confusion may produce prominent color fringes. Precise filters, which sharply sepa-
rate the luminance and chrominance signals (comb filters), can remove this effect, except in the
diagonal direction.
In static and slowly moving scenes, the clusters surrounding each line-frequency harmonic
are compact, seldom extending further than 1 or 2 kHz on either side of the line-harmonic fre-
quency. The space remaining in the signal spectrum is unoccupied and may be used to accommo-
date the spectral components of another signal having the same structure and frequency spacing.
For scenes in which the motion is sufficiently slow for the eye to perceive the detail of moving
objects, it may be safely assumed that less than half the spectral space between line-frequency
harmonics is occupied by energy of significant magnitude. It is on this principle that the NTSC-
and PAL-compatible color television systems are based. The SECAM system uses frequency-
modulated chrominance signals, which are not frequency interlaced with the luminance signal.
Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.co )

p
DC component
Frequency of image repetition
Frequency of line scanning
2X frequency of line scanning
3X frequency of line scanning
4X frequency
Amplitude
of line scanning
0, +1
0, 0
+1, -1
+1, 0
+1, +1
+1, -4
+1, -3
+1, -2
+2, -1
0, +2
0, +3
0, +4
+4, +1
+2, +1
+2, -3
+2, -2
+2, +2
+2, +3
+3, -1
+4, -1
+2, 0
+1, +2
+1, +3
+1, +4
+3, +1
+3, -3
+4, -3
+4, -2
m, n
+4, +2
+3, +2
+3, +3
+4, +3
+3, 0
+4, 0
Frequency +3, -2
Figure 11.1.4 The typical spectrum of a video signal, showing the harmonics of the line-scanning
frequency surrounded by clusters of components separated at intervals equal to the field-scanning
frequency. (After [1].)
11.1.4 References
1. Mertz, P.: “Television and the Scanning Process,” Proc. IRE, vol. 29, pp. 529–537, October
1941.
Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)

g g
Chapter
11.2
Measurement of Color Displays
11.2.1 Introduction
The chromaticity and luminance of a portion of a color display device may be measured in sev-
eral ways. The most fundamental approach involves a complete spectroradiometric measurement
followed by computation using tables of color-matching functions. Spectroradiometers are avail-
able for this purpose. Another method, somewhat faster but less accurate, involves the use of a
photoelectric colorimeter. Because these devices have spectral sensitivities approximately equal
to the CIE color-matching functions, they provide direct readings of tristimulus values.
For setting up the reference white, it is often simplest to use a split-field visual comparator
and to adjust the display device until it matches the reference field (usually D65) of the compar-
ator. However, because a large spectral difference (large metamerism) usually exists between the
display and the reference, different observers may make different settings by this method. Conse-
quently, settings by one observer—or a group of observers—with normal color vision often are
used simply to provide a reference point for subsequent photoelectric measurements.
An alternative method of determining the luminance and chromaticity coordinates of any area
of a display involves measuring the output of each phosphor separately, then combining the mea-
surements using the center of gravity law, by which the total tristimulus output of each phosphor
is considered as an equivalent weight located at the chromaticity coordinates of the phosphor.
Consider the CIE chromaticity diagram shown in Figure 11.2.1 to be a uniform flat surface
positioned in a horizontal plane. For the case illustrated, the center of gravity of the three weights
(Tr , Tg, Tb), or the balance point, will be at the point Co. This point determines the chromaticity
of the mixture color. The luminance of the color Co will be the linear sum of the luminance out-
puts of the red, green, and blue phosphors. The chromaticity coordinates of the display primaries
may be obtained from the manufacturer. The total tristimulus output of one phosphor may be
determined by turning off the other two CRT guns, measuring the luminance of the specified
area, and dividing this value by the y chromaticity coordinate of the energized phosphor. This
procedure then is repeated for the other two phosphors. From this data, the color resulting from
given excitations of the three phosphors may be calculated as follows:
• Chromaticity coordinates of red phosphor = xr , yr
• Chromaticity coordinates of green phosphor = xg, yg
11-19
D l d df Di i l E i i Lib @M G Hill ( di i l i i lib )
Measurement of Color Displays
1.0
520 nm
0.8
510 nm 540 nm
T c1
Tg G 560 nm
0.6
d2
500 nm
y 580 nm
0.4 d3 d1
C1 600 nm
490 nm Co
Tr
R 780 nm
d
4
0.2
480 nm
Tb
B
0.00 380 nm
Figure 11.2.1 The CIE 0.01931 chromaticity

0.2 0.4
diagram 0.6
illustrating use 0.8 1.0of gravity law (T d
of the center r 1
x
= Tg d2, Tc1 = Tr + Tg, Tc1 d3 = Tb d4).
• Chromaticity coordinates of blue phosphor = xb , yb

• Luminance of red phosphor = Yr
• Luminance of green phosphor = Yg
• Luminance of blue phosphor = Yb
Yr
Total tristimulus value of red phosphor = Xr + Y r + Zr = ----- = T (11.2.1)
yr r
Yg
Total tristimulus value of green phosphor = X g + Y g + Z g = ----- = Tg (11.2.2)
yg
Y
Total tristimulus value of blue phosphor = X b + Y b + Zb = ----b- = T b (11.2.3)
yb
Consider Tr as a weight located at the chromaticity coordinates of the red phosphor and Tg as a
weight located at the chromaticity coordinates of the green phosphor. The location of the chro-
p y
maticity coordinates of color C1 (blue gun of color CRT turned off) can be determined by taking
moments along line RG to determine the center of gravity of weights Tr and Tg:
Tr × d1 = T g × d2 (11.2.4)
The total tristimulus value of C1 is equal to T r + T g = T c1 (11.2.5)
Taking moments along line C1B will locate the chromaticity coordinates of the mixture color Co:
Tc1 × d3 = T b × d4 (11.2.6)
The luminance of the color Co is equal to Y r + Y g + Y b (11.2.7)
11.2.1a Assessment of Color Reproduction

A number of factors may contribute to poor color rendition in a display system. To assess the
effect of these factors, it is necessary to define system objectives, then establish a method of
measuring departures from the objectives. Visual image display may be categorized as follows:
• Spectral color reproduction: The exact reproduction of the spectral power distributions of the
original stimuli. Clearly, this is not possible in a video system with three primaries.
• Exact color reproduction: The exact reproduction of tristimulus values. The reproduction is
then a metameric match to the original. Exact color reproduction will result in equality of
appearance only if the viewing conditions for the picture and the original scene are identical.
These conditions include the angular subtense of the picture, the luminance and chromaticity
of the surround, and glare. In practice, exact color reproduction often cannot be achieved
because of limitations on the maximum luminance that can be produced on a color monitor.
• Colorimetric color reproduction: A variant of exact color reproduction in which the tristimu-
lus values are proportional to those in the original scene. In other words, the chromaticity
coordinates are reproduced exactly, but the luminances all are reduced by a constant factor.
Traditionally, color video systems have been designed and evaluated for colorimetric color
reproduction. If the original and the reproduced reference whites have the same chromaticity,
if the viewing conditions are the same, and if the system has an overall gamma of unity, colo-
rimetric color reproduction is indeed a useful criterion. These conditions, however, often do
not hold; then, colorimetric color reproduction is inadequate.
• Equivalent color reproduction: The reproduction of the original color appearance. This might
be considered as the ultimate objective, but it cannot be achieved because of the limited lumi-
nance generated in a display system.
• Corresponding color reproduction: A compromise by which colors in the reproduction have
the same appearance that colors in the original would have had if they had been illuminated to
produce the same average luminance level and the same reference white chromaticity as that
of the reproduction. For most purposes, corresponding color reproduction is the most suitable
objective of a color video system.

p y
• Preferred color reproduction: A departure from the preceding categories that recognizes the
preferences of the viewer. It is sometimes argued that corresponding color reproduction is not
the ultimate aim for some display systems, such as color television, and that it should be taken
into account that people prefer some colors to be different from their actual appearance. For
example, suntanned skin color is preferred to average real skin color, and sky is preferred
bluer and foliage greener than they really are.
Even if corresponding color reproduction is accepted as the target, some colors are more impor-
tant than others. For example, flesh tones must be acceptable—not obviously reddish, greenish,
purplish, or otherwise incorrectly rendered. Likewise, the sky must be blue and the clouds white,
within the viewer's range of acceptance. Similar conditions apply to other well-known colors of
common experience.
11.2.1b Chromatic Adaptation and White Balance

With properly adjusted cameras and displays, whites and neutral grays are reproduced with the
chromaticity of D65. Tests have shown that such whites (and grays) appear satisfactory in home
viewing situations even if the ambient light is of quite different color temperature. Problems
occur, however, when the white balance is slightly different from one camera to the next or when
the scene shifts from studio to daylight or vice versa. In the first case, unwanted shifts of the dis-
played white occur, whereas in the other, no shift occurs even though the viewer subconsciously
expects a shift.
By always reproducing a white surface with the same chromaticity, the system is mimicking
the human visual system, which adapts so that white surfaces always appear the same, whatever
the chromaticity of the illuminant (at least within the range of common light sources). The effect
on other colors, however, is more complicated. In video cameras, the white balance adjustment
usually is made by gain controls on the R, G, and B channels. This is similar to the von Kries
model of human chromatic adaptation, although the R, G, and B primaries of the model are not
the same as the video primaries. It is known that the von Kries model does not accurately account
for the appearance of colors after chromatic adaptation, and so it follows that making simple gain
changes in a video camera is not the ideal approach. Nevertheless, this approach seems to work
well in practice, and the viewer does not object to the fact, for example, that the relative increase
in the luminances of reddish objects in tungsten light is lost.
11.2.1c Overall Gamma Requirements

Colorimetric color reproduction requires that the overall gamma of the system—including the
camera, the display, and any gamma-adjusting electronics—be unity. This simple criterion is the
one most often used in the design of a video color rendition system. However, the more sophisti-
cated criterion of corresponding color reproduction takes into account the effect of the viewing
conditions. In particular, several studies have shown that the luminance of the surround is impor-
tant. For example, a dim surround requires a gamma of about 1.2, and a dark surround requires a
gamma of about 1.5 for optimum color reproduction.

1.0
510 nm 520 nm
0.8
540 nm
500 nm
560 nm
0.6
y 580 nm
490 nm
0.4
600 nm
480 nm 780 nm
0.2
0.00 380 nm
0.0 0.2 0.4 0.6 0.8 1.0

x
Figure 11.2.2 Ellipses of equally perceptible color differences (as identified by MacAdam).
11.2.1d Perception of Color Differences

The CIE 1931 chromaticity diagram does not map chromaticity on a uniform-perceptibility
basis. A just-perceptible change of chromaticity is not represented by the same distance in differ-
ent parts of the diagram. Many investigators have explored the manner in which perceptibility
varies over the diagram. The study most often quoted is that of MacAdam, who identified a set
of ellipses that are contours of equal perceptibility about a given color, as shown in Figure 11.2.2
[1].
From this and similar studies it is apparent, for example, that large distances represent rela-
tively small perceptible changes in the green sector of the diagram. In the blue region, much
smaller changes in the chromaticity coordinates are readily perceived.
Furthermore, viewing identical images on dissimilar displays can result in observed differ-
ences in the appearance of the image [2]. Several factors affect the appearance of the image,
including:
• Physical factors, including display gamut, illumination level, and black point
• Psychophysical factors, including chromatic induction and color constancy
Each of these factors interact in such a way that prediction of the appearance of an image on a
given display becomes difficult. System designers have experimented with colorimetry to facili-
tate the sharing of image data among display devices that vary in manufacture, calibration, and
location. Of particular interest today is the application of colorimetry to imaging in a networked
window system environment, where it is often necessary to ensure that an image displayed
remotely looks like the image displayed locally.
Downloaded from Digital Engineering Library @ McGraw Hill (www digitalengineeringlibrary com)
p y
Studies have indicated that image context and image content are also factors that affect color
appearance. The use of highly chromatic backgrounds in a windowed display system is popular,
but it usually will affect the appearance of the colors in the foreground.
11.2.1e Display Resolution and Pixel Format

The pixel represents the smallest resolvable element of a display. The size of the pixel varies
from one type of display to another. In a monochrome CRT, pixel size is determined primarily by
the following factors:
• Spot size of the electron beam (the current density distribution)
• Phosphor particle size
• Thickness of the phosphor layer
The term pixel was developed in the era of monochrome television, and the definition was—at
that time—straightforward. With the advent of color-triad-based CRTs and solid-state display
systems, the definition is not nearly so clear.
For a color CRT, a single triad of red, green, and blue phosphor dots constitutes a single pixel.
This definition assumes that the mechanical and electrical parameters of the CRT will permit
each triad to be addressed without illuminating other elements on the face of the tube. Most dis-
play systems, however, will not meet this criterion. Depending on the design, a number of triads
may constitute a single pixel in a CRT display. A more all-inclusive definition for the pixel is: the
smallest spatial-information element as seen by the viewer [3].
Dot pitch is one of the principal mechanical criteria of a CRT that determines, to a large
extent, the resolution of the display. Dot pitch is defined as the center-to-center distance between
adjacent green phosphor dots of the red, green, blue triad.
The pixel format is the arrangement of pixels into horizontal rows and vertical columns. For
example, an arrangement of 640 horizontal pixels by 480 vertical pixels results in a 640 × 480
pixel format. This description is not a resolution parameter in itself, simply the arrangement of
pixel elements on the screen. Resolution is the measure of the ability to delineate picture detail; it
is the smallest discernible and measurable detail in a visual presentation [4].
Pixel density is a parameter closely related to resolution, stated in terms of pixels per linear
distance. Pixel density specifies how closely the pixel elements are spaced on a given display. It
follows that a display with a given pixel format will not provide the same pixel density, or resolu-
tion, on a large-size screen (such as 27-in diagonal) as on a small-size screen (such as 13-in diag-
onal).
Television lines is another term used to describe resolution. The term refers to the number of
discernible lines on a standard test chart. As before, the specification of television lines is not, in
itself, a description of display resolution. A 525-line display on a 17-in monitor will appear to a
viewer to have greater resolution than a 525-line display on a 30-inch monitor. Pixel density is
the preferred resolution parameter.
11.2.1f Contrast Ratio

The purpose of a video display is to convey information by controlling the illumination of phos-
phor dots on a screen, or by controlling the reflectance or transmittance of a light source. The

p y
contrast ratio specifies the observable difference between a pixel that is switched on and one that
is in its corresponding off state:
L
Cr = on
-------- (11.2.8)
L of f
Where:
Cr = contrast ratio of the display
Lon = luminance of a pixel in the on state
Loff = luminance of a pixel in the off state
The area encompassed by the contrast ratio is an important parameter in assessing the perfor-
mance of a display. Two contrast ratio divisions typically are specified:
• Small area: comparison of the on and off states of a pixel-sized area
• Large area: comparison of the on and off states of a group of pixels
For most display applications, the small-area contrast ratio is the more critical parameter.
11.2.2 Display Measurement Techniques

A number of different techniques have evolved for measuring the static performance of picture
display devices and systems [5]. Most express the measured device performance in a unique fig-
ure of merit or metric. Although each approach provides useful information, the lack of standard-
ization in measurement techniques may make it difficult to directly compare the performance of
a given class of devices.
Regardless of the method used to measure performance, the operating parameters must be set
for the anticipated operating environment. Key parameters include:
• Input signal level
• System/display line rate
• Luminance (brightness)
• Contrast
• Image size
• Aspect ratio
If the display is used in more than one environmental condition—such as under day and night
conditions—a set of measurements is appropriate for each application.
11.2.2a Subjective CRT Measurements

Three common subjective measurement techniques are used to assess the performance of a CRT
[7]:
• Shrinking raster

p y
• Line width
• TV limiting resolution
Predictably, subjective measurements tend to exhibit more variability than objective measure-
ments. Although they generally are not used for acceptance testing or quality control, subjective
CRT measurements provide a fast and relatively simple means of performance assessment.
Results usually are consistent when performed by the same observer. Results for different
observers often vary, however, because different observers use different visual criteria to make
their judgments.
The shrinking raster and line-width techniques are used to estimate the vertical dimension of
the display (CRT) beam spot size (footprint). Several underlying assumptions accompany this
approach:
• The spot is assumed to be symmetrical and Gaussian in the vertical and horizontal planes.
• The display modulation transfer function (MTF) calculated from the spot-size measurement
results in the best performance envelope that can be expected from the device.
• The modulating electronics are designed with sufficient bandwidth so that spot size is the
limiting performance parameter.
• The modulation contrast at low spatial frequencies approaches 100 percent.
Depending on the application, not all of these assumptions are valid:
• Assumption 1. Verona [5] has reported that the symmetry assumption is generally not true.
The vertical spot profile is only an approximation to the horizontal spot profile; most spot
profiles exhibit some degree of astigmatism. However, significant deviations from the sym-
metry and Gaussian assumptions result in only minor deviations from the projected perfor-
mance when the assumptions are correct.
• Assumption 2. The optimum performance envelope assumption infers that other types of
measurements will result in the same or lower modulation contrast at each spatial frequency.
The MTF calculations based on a beam footprint in the vertical axis indicate the optimum
performance that can be obtained from the display because finer detail (higher spatial fre-
quency information) cannot be written onto the screen smaller than the spot size.
• Assumption 3. The modulation circuit bandwidth must be sufficient to pass the full incoming
video signal. Typically, the video circuit bandwidth is not a problem with current technology
circuits, which usually are designed to provide significantly more bandwidth than the display
is capable of reproducing. However, in cases where this assumption is not true, the calculated
MTF based purely on the vertical beam profile will be incorrect. The calculated performance
will be better than the actual performance of the display.
• Assumption 4. The calculated MTF is normalized to 100 percent modulation contrast at zero
spatial frequency and ignores the light scatter and other factors that degrade the actual mea-
sured MTF. Independent modulation-contrast measurements at a low spatial frequency can be
used to adjust the MTF curve to correct for the normalization effects.
Shrinking Raster Method

The shrinking raster measurement procedure is relatively simple. Steps include the following [5]:

p y
• The brightness and contrast controls are set for the desired peak luminance with an active ras-
ter background luminance (1 percent of peak luminance) using a stair-step video signal.
• While displaying a flat-field video signal input corresponding to the peak luminance, the ver-
tical gain/size is reduced until the raster lines are barely distinguishable.
• The raster height is measured and divided by the number of active scan lines to estimate the
average height of each scan line. The number of active scan lines is typically 92 percent of the
line rate. (For example, a 525-line display has 480 active lines, an 875-line display has 817,
and a 1025-line display has 957 active lines.)
The calculated average line height typically is used as a stand-alone metric of display perfor-
mance.
The most significant shortcoming of the shrinking raster method is the variability introduced
through the determination of when the scan lines are barely distinct to the observer. Blinking and
other eye movements often enhance the distinctness of the scan lines; lines that were indistinct
become distinct again.
Line-Width Method
The line-width measurement technique requires a microscope with a calibrated graticule [5]. The
focused raster is set to a 4:3 aspect ratio, and the brightness and contrast controls are set for the
desired peak luminance with an active raster background luminance (1 percent of peak lumi-
nance) using a stair-step video signal. A single horizontal line at the anticipated peak operating
luminance is presented in the center of the display. The spot is measured by comparing its lumi-
nous profile with the graticule markings. As with the shrinking raster technique, determination
of the line edge is subjective.
TV Limiting Resolution Method

This technique involves the display of 2-dimensional high-contrast bar patterns or lines of vari-
ous size, spacing, and angular orientation [5]. The observer subjectively determines the limiting
resolution of the image by the smallest set of bars that can be resolved. Following are descrip-
tions of potential errors with this technique:
• A phenomenon called spurious resolution can occur that leads the observer to overestimate
the limiting resolution. Spurious resolution occurs beyond the actual resolution limits of the
display. It appears as fine structures that can be perceived as line spacings closer than the
spacing at which the bar pattern first completely blurs. This situation arises when the fre-
quency-response characteristics fall to zero, then go negative, and perhaps oscillate as they
die out. At the bottom of the negative trough, contrast is restored, but in reverse phase (white
becomes black, and black becomes white).
• The use of test charts imaged with a video camera can lead to incorrect results because of the
addition of camera resolution considerations to the measurement. Electronically generated
test patterns are more reliable image sources.
The proper setting of brightness and contrast is required for this measurement. Brightness and
contrast controls are adjusted for the desired peak luminance with an active raster background
luminance (1 percent of peak luminance) using a stair-step video signal. Too much contrast will

p y
result in an inflated limiting resolution measurement; too little contrast will result in a degraded
limiting resolution measurement.
Electronic resolution pattern generators typically provide a variety of resolution signals from
100 to 1000 TV lines/picture height (TVL/ph) or more in a given multiple (such as 100). Figure
11.2.3 illustrates an electronically generated resolution test pattern for high-definition video
applications.
Application Considerations
The subjective techniques discussed in this section, with the exception of TV limiting resolution,
measure the resolution of the display [5]. The TV pattern test measures image resolution, which
is quite different.
Consider as an example a video display in which the scan lines can just be perceived—about
480 scan lines per picture height. This indicates a display resolution of at least 960 TV lines,
counting light and dark lines, per the convention. If a pattern from an electronic generator is dis-
played, observation will show the image beginning to deteriorate at about 340 TV lines. This
characteristic is the result of beats between the image pattern and the raster, with the beat fre-
quency decreasing as the pattern spatial frequency approaches the raster spatial frequency. This
ratio of 340/480 = 0.7 (approximately) is known as the Kell factor. Although debated at length,
the factor does not change appreciably in subjective observations.
11.2.2b Objective CRT Measurements

Four common types of objective measurements may be performed to assess the capabilities of a
CRT display device:
• Half-power width
• Impulse Fourier transform
• Knife-edge Fourier transform
• Discrete frequency
Although more difficult to perform than the subjective measurements discussed so far, objective
CRT measurement techniques offer greater accuracy and better repeatability. Some of the proce-
dures require specialized hardware and/or software.
Half-Power-Width Method
Under the half-power-width technique, a single horizontal line is activated with the brightness
and contrast controls set to a typical operating level. The line luminance is equivalent to the high-
light luminance (maximum signal level). The central portion of the line is imaged with a micro-
scope in the plane of a variable-width slit. The open slit allows all the light from the line to pass
through to a photodetector. The output of the photodetector is displayed on an oscilloscope. As
the slit is gradually closed, the peak amplitude of the photodetector signal decreases. When the
signal drops to 50 percent of its initial value, the slit width is recorded. The width measurement
divided by the microscope magnification represents the half-power width of the horizontal scan
line.

p y
Figure 11.2.3 Wide aspect ratio resolution test chart produced by an electronic signal generator.
(Courtesy of Tektronix.)
The half-power width is defined as the distance between symmetrical integration limits, cen-
tered about the maximum intensity point, which encompasses half of the total power under the
intensity curve. The half-power width is not the same as the half-intensity width measured
between the half-intensity points. The half-intensity width is theoretically 1.75 times greater than
the half-power width for a Gaussian spot luminance distribution.
It should be noted that the half-power line-width technique relies on line width to predict the
performance of the CRT. Many of the precautions outlined previously apply here also. The pri-
mary difference, however, is that line width is measured under this technique objectively, rather
than subjectively.
Fourier Transform Methods

The impulse Fourier transform technique involves measuring the luminance profile of the spot
and then taking the Fourier transform of the distribution to obtain the MTF. The MTF, by defini-
tion, is the Fourier transform of the line spread function. Commercially available software may
be used to perform these measurements using either an impulse or knife edge as the input wave-
form. Using the vertical spot profile as an approximation to the horizontal spot profile is not
always appropriate, and the same reservations expressed in the previous section apply in this case
as well.
The measurement is made by generating a single horizontal line on the display with the
brightness and contrast set as discussed previously. A microphotometer with an effective slit-
aperture width approximately one-tenth the width of the scan line is moved across the scan line
(the long slit axis parallel to the scan line). The data taken is stored in an array, which represents

p y
the luminance profile of the CRT spot, distance vs. luminance. The microphotometer is cali-
brated for luminance measures and for distance measures in the object plane. Each micron step
of the microphotometer represents a known increment in the object plane. The software then cal-
culates the MTF of the display based on its line spread from the calibrated luminance and dis-
tance measurements. Finite slit-width corrections also may be made to the MTF curve by
dividing it by a measurement-system MTF curve obtained from the luminance profile of an ideal
knife-edge aperture or a standard source.
The knife-edge Fourier transform measurement may be conducted using a low-spatial-fre-
quency vertical bar pattern (5 to 10 cycles) across the display with the brightness and contrast
controls set as discussed previously. The frequency response of the square wave pattern genera-
tor and video pattern generator should be greater than the frequency response of the display sys-
tem. The microphotometer scans from the center of a bright bar to the center of a dark bar (left to
right), measuring the width of the boundary and comparing it to a knife edge. The microphotom-
eter slit is oriented vertically, with its long axis parallel to the bars. The scan usually is made
from a light bar to a dark bar in the direction of spot movement. This procedure is preferred
because waveforms from scans in the opposite direction may contain certain anomalies. When
the beam is turned on in a square wave pattern, it tends to overshoot and oscillate. This behavior
produces artifacts in the luminance profile of the bar edge as the beam moves from an off to an
on state. In the on-to-off direction, however, the effects are minimal and the measured waveform
does not exhibit the same anomalies that can corrupt the MTF calculations.
The bar-edge (knife-edge) measurement, unlike the other techniques discussed so far, uses the
horizontal spot profile to predict display performance. All of the other techniques use the vertical
profile as an approximation of the more critical horizontal spot profile. The bar-edge measure-
ment will yield a more accurate assessment of display performance because the displayed image
is being generated with a spot scanned in the horizontal direction.
Discrete Frequency Method

The discrete sine wave frequency-response measurement technique provides the most accurate
representation of display performance. With this approach, there are no assumptions implied
about the shape of the spot, the electronics bandwidth, or low-frequency light scatter. Discrete
spatial frequency sine wave patterns are used to obtain a discrete spatial frequency MTF curve
that represents the signal-in to luminance-out performance of the display.
The measurement is begun by setting the brightness and contrast as discussed previously, with
black-level luminance at 1 percent of the highlight luminance. A sine wave signal is produced by
a function generator and fed to a pedestal generator where it is converted to an RS-170-A or RS-
343 signal, then applied to the CRT. The modulation and resulting spatial frequency pattern are
measured with a scanning microphotometer. The highlight and black-level measurements are
used to calculate the modulation constant for each spatial frequency from 5 cycles/display width
to the point that the modulation constant falls to less than 1 percent. The modulation constant
values then are plotted as a function of spatial frequency, generating a discrete spatial frequency
MTF curve.

p y
11.2.3 Viewing Environment Considerations

The environment in which a video display device is viewed is an important criterion for critical
viewing situations. Applications in which color purity and adherence to set standards are impor-
tant require a standardized (or at least consistent) viewing environment. For example, textile col-
ors viewed on a display with a white surround will appear different than the same colors viewed
with a black surround. By the same token, different types of ambient lighting will make identical
colors appear different on a given display.
The SMPTE has addressed this issue with RP 166, which specifies the environmental and
surround conditions that are required in television or video program review areas for the “consis-
tent and critical evaluation” of conventional television signals [6]. Additionally, the practice is
designed to provide for repeatable color grading or correction. A number of important parame-
ters are specified in RP 166, including the following:
• The distance of the observer from the monitor screen should be 4 to 6 picture heights for
SDTV displays.
• The observer should view the monitor screen at a preferred angle in both the horizontal and
vertical planes of 0º ± 5º and, in any event, no greater than ±15º from the perpendicular to the
midpoint of the screen.
• The viewing area decor should have a generally neutral matte impression, without dominant
colors.
• Surface reflectances should be nonspecular and should not exceed 10 percent of the peak
luminance value of the monitor white.
The Recommended Practice suggests placing the monitor in a freestanding environment 2.5
to 5 screen heights in front of the wall providing the visual surround. Another acceptable
approach is to mount the monitor in a wall with its face approximately flush with the surface of
the wall. It is further recommended that all light sources in use during picture assessment or
adjustment have a color quality closely matching the monitor screen at reference white (i.e.,
D65).
It is often necessary to have black-and-white monitors surrounding one or more color moni-
tors in a studio control room. According to RP 166, the black-and-white monitors should be the
same color temperature as the properly adjusted color monitor(s), 6500 K. Black-and-white
monitors are normally equipped with P4 phosphors, at about 9300 K. This cooler color tempera-
ture prevents the background surrounding the color monitors from remaining neutral. Most
black-and-white monitors can be ordered with 6500 K phosphors.
11.2.3a Picture Monitor Alignment

The proper adjustment and alignment of studio picture monitors is basic to video quality control.
Uniform alignment throughout the production chain also ensures consistency in color adjust-
ment, which facilitates the matching of different scenes within a program that may be processed
at different times and in different facilities. The SMPTE has addressed this requirement for con-
ventional video through RP 167. The Recommended Practice offers a step-by-step process by
which color monitors can be set. Key elements of RP 167 include the following [7]:

p y
• Initial conditions. Setup includes allowing the monitor to warm up and stabilize for 20 to 30
minutes. The room ambient lighting should be the same as it is when the monitor is in normal
service, and several minutes must be allowed for visual adaptation to the operating environ-
ment.
• Initial screen adjustments. The monitor is switched to the setup position, in which the red,
green, and blue screen controls are adjusted individually so that the signals are barely visible.
• Purity. Purity, the ability of the gun to excite only its designated phosphor, is checked by
applying a low-level flat-field signal and activating only one of the three guns at a time. The
display should have no noticeable discolorations across the face.
• Scan size. The color picture monitor application establishes whether the overscan or under-
scan presentation of the display will be selected. An underscanned display is one in which the
active video (picture) area, including the corners of the raster, is visible within the screen
mask. Normal scan brings the edges of the picture tangent to the mask position. Overscan
should be no more than 5 percent.
• Geometry and aspect ratio. Display geometry and aspect ratio are adjusted with the cross-
hatch signal by scanning the display device with the green beam only. Correct geometry and
linearity are obtained by adjusting the pincushion and scan-linearity controls so that the pic-
ture appears without evident distortions from the normal viewing distance.
• Focus. An ideal focus target is available from some test signal generators; if it is unavailable,
multiburst, crosshatch, or white noise can be used as tools to optimize the focus of the dis-
played picture.
• Convergence. Convergence is adjusted with a crosshatch signal; it should be optimized for
either normal scan or underscan, depending upon the application.
• Aperture correction. If aperture correction is used, the amount of correction can be esti-
mated visually by ensuring that the 2T sin2 pulse has the same brightness as the luminance
bar or the multiburst signal when the 3 and 4.2 MHz bursts have the same sharpness and con-
trast.
• Chrominance amplitude and phase. The chrominance amplitude and phase are adjusted
using the SMPTE color bar test signal and viewing only the blue channel. Switching off the
comb filter, if it is present, provides a clear blue channel display. Periodically, the red and
green channels should be checked individually in a similar manner to verify that the decoders
are working properly. A detailed description of this procedure is given in [7].
• Brightness, color temperature, and gray scale tracking. The 100-IRE window signal is
used to supply the reference white. Because of typical luminance shading limitations, a cen-
trally placed PLUGE [8] signal is recommended for setting the monitor brightness control.
The black set signal provided in the SMPTE color bars also can be used for this purpose.
• Monitor matching. When color matching two or more color monitors, the same alignment
steps should be performed on each monitor in turn. Remember, however, that monitors cannot
be matched without the same phosphor sets, similar display uniformity characteristics, and
similar sharpness. The most noticeable deviations on color monitors are the lack of uniform
color presentations and brightness shading. Color matching of monitors for these parameters

p y
can be most easily assessed by observing flat-field uniformity of the picture at low, medium,
and high amplitudes.
For complete monitor-alignment procedures, see [7].
11.2.4 References
1. MacAdam, D. L.: “Visual Sensitivities to Color Differences in Daylight,” J. Opt. Soc. Am.,
vol. 32, pp. 247–274, 1942.
2. Bender, Walter, and Alan Blount: “The Role of Colorimetry and Context in Color Dis-
plays,” Human Vision, Visual Processing, and Digital Display III, Bernice E. Rogowitz
(ed.), Proc. SPIE 1666, SPIE, Bellingham, Wash., pp. 343–348, 1992.
3. Tannas, Lawrence E., Jr.: Flat Panel Displays and CRTs, Van Nostrand Reinhold, New York,
N.Y., pg. 18, 1985.
4. Standards and Definitions Committee, Society for Information Display.
5. Verona, Robert: “Comparison of CRT Display Measurement Techniques,” Helmet-Mounted
Displays III, Thomas M. Lippert (ed.), Proc. SPIE 1695, SPIE, Bellingham, Wash., pp.
117–127, 1992.
6. “Critical Viewing Conditions for Evaluation of Color Television Pictures,” SMPTE Recom-
mended Practice RP 166, SMPTE, White Plains, N.Y., 1995.
7. “Alignment of NTSC Color Picture Monitors,” SMPTE Recommended Practice RP 167,
SMPTE, White Plains, N.Y., 1995.
8. Quinn, S. F., and C. A. Siocos: “PLUGE Method of Adjusting Picture Monitors in Televi-
sion Studios—A Technical Note,” SMPTE Journal, SMPTE, White Plains, N.Y., vol. 76,
pg. 925, September 1967.

p y

g g
Chapter
11.3
Camera Performance Verification
Peter Gloeggler
11.3.1 Introduction
With tube-type cameras, it was almost mandatory to fine-tune the device before a major shoot to
achieve optimum performance. The intrinsic stability of the CCD imager and the stability of the
circuitry used in modern CCD cameras now make it possible to operate the camera for several
months without internal readjustment. Physical damage to the camera or lens, in use or transport,
is probably the most frequent cause of a loss in performance in a CCD-based device. With care-
ful handling, the probability of malfunction is very small. It is nevertheless prudent to schedule a
quick check-out of the camera before the start of a major shoot, when the high cost of talent and
other aspects of the production are considered.
The following items are appropriate for inclusion in such a check-out procedure. If the test
results show a significant deviation from the manufacturers specifications or from the data previ-
ously obtained for the same test, a more thorough examination of the camera, as prescribed in the
camera service manual, is then indicated.
11.3.2 Visual Inspection and Mechanical Check

Visually inspect the camera and lens for evidence of physical damage as a clue to the possibility
of more serious internal damage. Carefully operate the lens adjustments—manual and servo
zoom, focus, and manual iris. If there is evidence of binding or a rough spot in any of these
adjustments, physical damage that may affect the optical performance of the lens must be sus-
pected. Inspect the front and rear lens elements; clean if necessary using pure alcohol and soft,
lint-free wipes. Fingerprints, in particular, should be removed as quickly as possible to avoid
harm to the optical coating of the lens elements. Note that lens manufacturers generally discour-
age the use of silicon-impregnated wipes for cleaning high quality optics.
11-35
11.3.3 Electrical Parameters and Measurements

Top performance from a camera requires that all elements of the system be in good working
order. Key issues include the following:
• Colorimetry. The colorimetry measurement evaluates the ability of the camera to accurately
reproduce a given scene. For in-depth analysis, detailed information can be obtained from a
graphical display of camera performance plotted on the CIE 1976 chromaticity diagram,
accompanied by tabular data of CIELUV parameters.
• CCD defects. Defective pixels, or clusters of pixels, in the CCD create “holes” in the scene-
capture mechanism. Test equipment is available to perform this measurement and to then plot
the defective pixels on a representation of the CCD imager surface.
• Fixed pattern noise. This measurement characterizes persistent current variations that corre-
late to CCD pixels. Pixel deviation from a nominal current level is plotted on a histogram.
• Frequency response and aliasing. Degradation in waveform amplitude as a function of fre-
quency is characterized by the frequency response measurement. Any multiburst squarewave
or sinewave chart designed for frequency response testing with six to 12 packets can be used.
Aliasing quantifies the aliasing phenomenon of the CCD image sampling process. The
aliased components can be measured in both amplitude and actual frequency represented. A
multiburst sinewave chart can be used for this measurement.
• Detail. To enhance an image, cameras include circuits that artificially sharpen the brightness
transitions by adding a correction signal. This measurement quantifies the amount of correc-
tion and expresses it as a percentage of the transition amplitude. Both horizontal and vertical
detail should be measured and the corresponding waveform graphed. Any chart that contains
a white patch, such as a window chart or a gray scale chart with a white chip, can be used to
evaluate detail.
• Gamma. The input-to-output light transfer characteristic of the camera is plotted and the
exponent of the curve (gamma) is the measured result. Shading errors should be accounted
for during the measurement process and eliminated from the measured results.
• Geometry and registration. Geometry measures, on the green channel, the geometric distor-
tions that exist in the camera. How well the red and blue channels track the green channel is
then expressed as registration error. Both geometry and registration can be measured using a
standard or custom registration chart. The error results are expressed as a percentage of pic-
ture height. Because errors closer to the center of the image are far more noticeable than
errors towards the edges, different zones can be plotted over the display to better evaluate sys-
tem performance.
• Shading. Both lens distortions and camera circuitry can cause the camera output level to vary
from one area to another within a field while the input light level is held constant. The signal
level variation is measured both across lines (horizontal shading) and down fields (vertical
shading). Errors are expressed as a percentage of peak white. A median filter is sometimes
used to eliminate high frequency amplitude errors that may skew the measured results.
• Vertical smear. This measurement evaluates the ability of the CCD camera to handle exces-
sive amounts of light in the input image. A black chart with a centered, small aperture is

required for this measurement. Vertical smear monitors three regions of the chart: the aper-
ture (for white referencing), and above and below the aperture (for smearing).
11.3.3a Confirmation of the Camera Encoder

A properly adjusted encoder is particularly useful because it provides a convenient window to
look inside the camera and confirm proper operation of the remaining circuitry. Encoder set-up
is easy to confirm because the color bar generator, normally provided in a professional camera,
offers a convenient self test of the encoder. To confirm proper operation of the camera encoder,
perform the following steps:
• Apply the camera (encoder) output signal to a waveform monitor (WFM), vectorscope, and
picture monitor (a high-resolution black and white monitor with 800 TVL or higher resolu-
tion is recommended).
• Terminate the WFM, vectorscope, and picture monitor using a discrete 75 Ω termination. The
preferred tolerance for this termination is ±0.1 percent, and in any event, no greater than ±1
percent. Internal terminations should not be used unless they have been tested and, if neces-
sary replaced with terminations within the recommended tolerance.
• Select the color bar mode on the camera. Confirm on the vectorscope that the burst and I and
Q vectors are of the correct phase and amplitude. Confirm that all of the color bar vectors fall
within the tolerance boxes on the vectorscope.
If all of the foregoing vectors are within tolerance, correct operation and adjustment of the
encoder is confirmed.
11.3.3b Confirmation of Auto Black

To confirm proper operation of the Auto Black circuit:
• Activate the Auto Black circuitry of the camera.
• Confirm that the lens caps during this adjustment. Confirm the character display in the view-
finder indicates the Auto Black adjustment has been successfully executed.
• Select the 0 dB, +9 dB, and +18 dB gain settings in sequence and confirm the black level
adjustment is correct for all three gain settings. This is most easily confirmed with the vector-
scope; the output signal should be a dot at the center of the display with no shift in position as
gain is switched. The only change should be an increase in noise at the higher gain settings.
11.3.3c Lens Back-Focus

The lens back-focus adjustment trims the lens to the specific optical dimensions of the camera.
Whenever a new lens is put on a camera, it is necessary to make this adjustment. Some lenses use
a screwdriver lock, while others use a knurled knob to lock the lens back-focus adjustment in
place. Accidental misadjustment in use has been known to occur, and it is therefore recom-
mended to confirm this adjustment. To set the lens back-focus:
• Place a Siemens Star Chart (Figure 11.3.1) at least 10 ft from the camera. Place the chart in a
location with low lighting such that the lens iris is wide open.

• Using a high-resolution picture monitor: 1)

Adjust the lens zoom for full close-up and
adjust for best focus using the focus ring on
the front of the lens. 2) Adjust the zoom for
maximum wide angle position and adjust
for optimum focus using the lens back-focus
adjustment. Repeat both steps several times.
• Securely lock the lens back-focus adjust-
ment in place.
• Confirm the lens stays in focus over the full
zoom range. Figure 11.3.1 Siemens star chart.
11.3.3d Black Shading

To confirm black shading, perform the following steps:
• Cap the lens. Raise the master black level to about 10–12 IRE to avoid clipping.
• Observe the waveform monitor in the vertical display mode, and then in the horizontal display
mode. If the black level is a thin horizontal line, there is no black shading in any of the three
color components.
• Restore the black level to its proper position.
11.3.3e Detail Circuit

Using the 11-step gray scale chart, confirm the amplitude of the detail signal as required for the
application. A stronger detail signal typically is required for a lower performance recorder and
less for a higher performance recorder.
11.3.3f Optional Tests

If the camera system is capable of resolving fine detail close to the limiting resolution specified
by the manufacturer, there is a strong assurance that the lens, camera optics, and overall camera
signal processing circuits are working correctly. Use a suitable chart with resolution wedges or a
chart with a multiburst pattern and confirm that the overall resolution of the camera system is
close to the manufacturers specification (Figure 11.3.2).
White Shading
To confirm white shading:
• Set up a uniformly lit white test chart.
• Using the waveform monitor, open the lens to obtain about 70 IRE units of video (confirm
that the iris is in the range of f / 4.0 to f / 5.6; adjust the lighting if necessary), and confirm
there is a minimum of vertical, then horizontal, shading.
• Adjust as necessary using the camera horizontal and vertical white shading controls.

Figure 11.3.2 Multiburst chart (10 MHz).
Flare
The camera flare correction circuitry provides an approximate correction for flare or scattering
of peripheral rays in the various parts of the optical system. To confirm the adjustment of the
flare correction circuit:
• Frame an 11-step gray scale chart that includes a very low reflectance strip of black velvet
added to the chart.
• Adjust the iris from fully closed until the white chip is at 100 IRE units of video. The flare
compensation circuitry is adjusted correctly if there is almost no rise in the black level of the
velvet strip as the white level is increased to 100 IRE units and only a small rise in the black
level, with no change in hue, when the iris is opened one more f-stop beyond the 100 IRE
units point.
• Adjust the R, G, and B flare controls as defined in the camera service manual if the flare cor-
rection is not adjusted correctly.
Linear Matrix
When it is necessary to use two dissimilar cameras models in a multicamera shoot, and either of
the two models provides an adjustable linear matrix, it is possible to use the variable matrix to
obtain a better colorimetry match between cameras. Specific matrix parameters and adjustments
(if any) will be found in the camera service manuals.

Figure 11.3.3 Color reference pattern layout specified in SMPTE 303M. (From [1]. Used with per-
mission.)
11.3.3g Color Reference Pattern

SMPTE 303M defines the electrical and physical representation of a television color reference
pattern [1]. It also specifies colorimetry, geometry, and related parameters.
The color reference pattern is made up of 24 sample colors whose colorimetric designations
are distributed throughout the color television gamut. As shown in Figure 11.3.3, these samples
are square and are arranged in four rows of six samples per row. The first two rows consist of
colors that are designed to simulate the color appearance of natural objects. The third row con-
sists of colors that represent subtractive primaries (cyan, magenta, and yellow) as well as the
binary combinations of these colors (red, green, and blue). The last row consists of a six-step
neutral gray scale.
All video signal levels are calibrated using SMPTE RP 177, the standard television system
D65 white point, and each of the following standard television primary colorimetry definitions:
• ITR-R BT.709
• SMPTE 170M/SMPTE 240M
• PAL (EBU)
• NTSC (1953 original values)
The standard further specifies illumination and camera positioning details.
11.3.4 References
1. “SMPTE Standard: For Television—Color Reference Pattern,” SMPTE 303M, SMPTE,
White Plains, N.Y., 1999.

11.3.5 Bibliography
Gloeggler, Peter: “Video Pickup Devices and Systems,” in NAB Engineering Handbook,” 9th
Ed., Jerry C. Whitaker (ed.), National Association of Broadcasters, Washington, D.C.,
1999.

g g
Chapter
11.4
Conventional Video Measurements
Carl Bentz
11.4.1 Introduction
Although there are a number of computer-based television signal monitors capable of measuring
video, sync, chroma, and burst levels (as well as pulse widths and other timing factors), some-
times the best way to see what a signal is doing is to monitor it visually. The waveform monitor
and vectorscope are oscilloscopes specially adapted for the video environment. The waveform
monitor, like a traditional oscilloscope, operates in a voltage-versus-time mode. While an oscil-
loscope timebase can be set over a wide range of intervals, the waveform monitor timebase trig-
gers automatically on sync pulses in the conventional TV signal, producing line- and field-rate
sweeps, as well as multiple lines, multiple fields, and shorter time intervals. Filters, clamps, and
other circuits process the video signal for specific monitoring needs. The vectorscope operates in
an X-Y voltage-versus-voltage mode to display chrominance information. It decodes the signal in
much the same way as a television receiver or a video monitor to extract color information and to
display phase relationships. These two instruments serve separate, distinct purposes. Some mod-
els combine the functions of both types of monitors in one chassis with a single CRT display.
Others include a communications link between two separate instruments.
Beyond basic signal monitoring, the waveform monitor and vectorscope provide a means to
identify and analyze signal aberrations. If the signal is distorted, these instruments allow a tech-
nician to learn the extent of the problem and to locate the offending equipment.
Although designed for analog video measurement and quality control, the waveform monitor
and vectorscope still serve valuable roles in the digital video facility, and will continue to do so
for some time.
11.4.1a Color Bar Test Patterns

Color bars are the most common pattern for testing encoders, decoders, and other video devices.
The test pattern typically contains several bars filled with primary and complementary colors.
There are many variants, which differ in color sequence, orientation, saturation and intensity.
The standard color bar sequence is white, yellow, cyan, green, magenta, red, blue, and black.
This sequence can be produced in an RGB format by a simple 3-bit counter. The typical specifi-
11-43
(a) WYCGMRB WYCGMRB WYCGMRB

0.7V
Red
0V
-0.3V
0.7V
Green
0V
-0.3V
0.7V
Blue
0V
-0.3V
0.943V
0.7V
PAL Composite
0V
-0.3V
100/0/100/0 100/0/100/25 100/0/75/0
Figure 11.4.1 Examples of color test bars: (a) ITU nomenclature, (b, next page) SMPTE EG 1
color bar test signal, (c) EIA RS-189-A test signal, (d) multi-format color bar test signal. [Drawing
(b) after SMPTE EG 1, (c) after EIA RS-189-A, (d) after SMPTE RP 219.]
cation of color bar levels (ITU-R Rec. BT.471-1, “Nomenclature and Description of Color Bar
Signals”) is a set of four numbers separated by slashes or dots and giving RGB levels as a per-
centage of reference white in the following sequence:
white bar / black bar / max colored bars / min colored bars
For example, 100/0/100/25 means 100 percent R, G, and B on the white bar; 0 percent R, G, and
B on the black bar; 100 percent maximum of R, G, and B on colored bars; and 25 percent mini-
mum of R, G, and B on colored bars. (See Figure 11.4.1a.)
Some color bar patterns merit special names. For example, 100/0/75/0 bars are often called
“EBU bars” or “75 percent bars”, and 100/0/100/25 bars are known as “BBC bars.” Neverthe-
less, the ITU four number nomenclature remains the only reliable specification system to desig-
nate the exact levels for color bar test patterns.
The SMPTE variation is a matrix test pattern according to SMPTE engineering guideline EG
1, “Alignment Color Bar Test Signal for Television Picture Monitors.” The guideline specifies
that 67 percent of the field contains seven (without black) 75 percent color bars, plus 8 percent of
the field with the “new chroma set” bars (blue/black/magenta/black/cyan/black/gray), and the

(b )
(c)
remaining 25 percent contains a combination of –I, white, Q, black, and the black set signal (a
version of PLUGE). This arrangement is illustrated in Figure 11.4.1b.
Chroma gain and chroma phase for picture monitors are usually adjusted by observing the
standard encoded color bar signal with the red and green CRT guns switched off. The four visi-
ble blue bars are set for equal brightness. The use of the chroma set feature greatly increases the
accuracy of this adjustment because it provides a signal with the blue bars to be matched verti-

Figure 11.4.1d
cally adjacent to each other. Because the bars are adjacent, the eye can easily perceive differences
in brightness. This also eliminates effects resulting from shading or purity from one part of the
screen to another.
The EIA color bar signal is a matrix test pattern according to RS-189-A, which consists of 75
percent of the field containing seven (without black) 75 percent color bars (same as SMPTE) and
the remaining 25 percent containing a combination of –I, white, Q, and black. (See Figure
11.4.1c.)
HD/SD Video Test Signal

The SMPTE, in recognition of the need for convenient measurement of both high-definition
(HD) and standard-definition (SD) video signals within a given plant, developed a recommended
practice designed to serve both domains [1]. SMPTE RP-219 specifies a color bar pattern com-
patible with both HD and SD environments. The multi-format color bar signal is originated as an
HD signal with an aspect ratio of 16:9 and may be down converted to an SD color bar signal with
an aspect ratio of either 4:3 or 16:9.
The color bar signal is generated with unconventionally slow rise and fall time value, and is
therefore only intended to facilitate video level control and monitor color adjustments of HD and
SD equipment. It can be applied to HD video productions, especially in a multi-format environ-

Figure 11.4.2 Waveform monitor display of a color-bar signal at the two-line rate. (Courtesy of
Tektronix.)
ment where HD video sources are frequently converted and used as SD video content either in
525- or 625-line environment with same frame frequencies as in the original HDTV signal.
The multi-format color bar signal is composed of four specific patterns, shown in Figure
11.4.1d. The first part of the color bar signal represents a signal for the 4:3 aspect ratio; a second
part of the total signal adds side panels for the 16:9 aspect ratio. A third part adds black and
white ramps and additional color information, and the last part completes the total signal by add-
ing white and black bars, in addition to a set of near-black-level steps for monitor black level
adjustment.
11.4.1b Basic Waveform Measurements

Waveform monitors are used to evaluate the amplitude and timing of video signals and to show
timing relationships between two or more signals [2]. The familiar color-bar pattern is the only
signal required for these basic tests. Figure 11.4.2 shows a typical waveform display of color
bars. It is important to realize that all color bars are not created equal. Some generators offer a
choice of 75 percent or 100 percent amplitude bars. Sync, burst, and setup amplitudes remain the
same in the two color-bar signals, but the peak-to-peak amplitudes of high-frequency chromi-
nance information and low-frequency luminance levels change. The saturation of color, a func-
tion of chrominance and luminance amplitudes, remains constant at 100 percent in both modes.
The 75 percent bar signal has 75 percent amplitude with 100 percent saturation. In 100 percent
bars, amplitude and saturation are both 100 percent.
Chrominance amplitudes in 100 percent bars exceed the maximum amplitude that should be
transmitted via NTSC. Therefore, 75 percent amplitude color bars, with no chrominance infor-
mation exceeding 100 IRE, are the standard amplitude bars for conventional television. In the 75

(a)
(b)
Figure 11.4.3 EIA RS-189-A color-bar displays: (a) color displays of gray and color bars, (b) wave-
form display of reference gray and primary/complementary colors, plus sync and burst.
percent mode, a choice of 100 IRE or 75 IRE white reference level may be offered. Figure 11.4.3
shows 75 percent amplitude bars with a 100 IRE white level. Either white level can be used to set
levels, but operators must be aware of which signal has been selected. SMPTE bars have a white
level of 75 IRE as well as a 100 IRE white flag.
The vertical response of a waveform monitor depends upon filters that process the signal in
order to display certain components. The flat response mode displays all components of the sig-
nal. A chroma response filter removes luminance and displays only chrominance. The low-pass
filter removes chrominance, leaving only low-frequency luminance levels in the display. Some
monitors include an IRE filter, designed to average-out high-level, fine-detail peaks on a mono-
chrome video signal. The IRE filter aids the operator in setting brightness levels. The IRE
response removes most, but not all, of the chrominance.

If the waveform monitor has a dual filter mode, the operator can observe luminance levels
and overall amplitudes at the same time. The instrument switches between the flat and low-pass
filters. The line select mode is another useful feature for monitoring live signals.
Sync Pulses
Most waveform monitors include 0.5 μs or 1 μs per division magnification (MAG) modes, which
can be used to verify H-sync width between approximately 4.4 μs and 5.1 μs. The width is mea-
sured at the – 4 IRE point. On waveform monitors with good MAG registration, sync appearing
in the middle of the screen in the 2-line mode remains centered when the sweep is magnified.
Check the rise and fall times of sync, and the widths of the front porch and entire blanking inter-
val. Examine burst and verify there are between 8 and 11 cycles of subcarrier.
Check the vertical intervals for correct format, and measure the timing of the equalizing
pulses and vertical sync pulses. The acceptable limits for these parameters are shown in Figure
11.4.4.
11.4.1c Basic Vectorscope Measurements

The vectorscope displays chrominance amplitudes, aids hue adjustments, and simplifies the
matching of burst phases of multiple signals. These functions require only the color-bar test sig-
nal as an input stimulus. To evaluate and adjust chrominance in the TV signal, observe color bars
on the vectorscope. The instrument should be in its calibrated gain position. Adjust the vector-
scope phase control to place the burst vector at the 9 o'clock position. Note the vector dot posi-
tions with respect to the six boxes marked on the graticule. If everything is correctly adjusted,
each dot will fall on the crosshairs of its corresponding box, as shown in Figure 11.4.5.
The chrominance amplitude of a video signal determines the intensity or brightness of color.
If the amplitudes are correct, the color dots fall on the crosshairs in the corresponding graticule
boxes. If vectors overshoot the boxes, chrominance amplitude is too high; if they undershoot, it is
too low. The boxes at each color location can be used to quantify the error. In the radial direction,
the small boxes indicate a ± 2.5 IRE error from the standard amplitude. The large boxes indicate
a ± 20 percent error.
Other test signals, including a modulated staircase or multiburst, are used for more advanced
tests. It is important to closely observe how these signals appear at the output of the waveform
generator. Knowing what the undistorted signal looks like simplifies the subsequent identifica-
tion of distortions.
11.4.1d Line Select Features

Some waveform monitors and vectorscopes have line select capability. They can display one or
two lines out of the entire video frame of 525 lines. (In the normal display, all of the lines are
overlaid on top of one another.) The principal use of the single line feature is to monitor VITS
(vertical interval test signals). VITS allows in-service testing of the transmission system. A typ-
ical VITS waveform is shown in Figure 11.4.6. A full-field line selector drives a picture monitor
output with an intensifying pulse. The pulse causes a single horizontal line on a picture monitor
to be highlighted. This indicates where the line selector is within the frame to correlate the wave-
form monitor display with the picture.

Figure 11.4.4 Sync pulse widths for NTSC.
11.4.1e Distortion Mechanisms

The video system should respond uniformly to signal components of different frequencies. In an
analog system, this parameter is generally evaluated with a waveform monitor. Different signals
are required to check the various parts of the frequency spectrum. If the signals are all faithfully
reproduced on the waveform monitor screen after passing through the video system, it is safe to
assume that there are no serious frequency response problems.
At very low frequencies, look for externally introduced distortions, such as power-line hum or
power supply ripple, and distortions resulting from inadequacies in the video processing equip-
ment. Low-frequency distortions usually appear on the video image as flickering or slowly vary-
ing brightness. Low-frequency interference can be seen on a waveform monitor when the dc
restorer is set to the slow mode and a 2-field sweep is selected. Sine wave distortion from ac
power-line hum may be observed in Figure 11.4.7. A bouncing APL signal can be used to detect
distortion in the video chain. Vertical shifts in the blanking and sync levels indicate the possibil-
ity of low-frequency distortion of the signal.

Figure 11.4.5 Vectorscope display of a color-bar signal. (Courtesy of Tektronix.)
Figure 11.4.6 Composite vertical-interval test signal (VITS) inserted in field 1, line 18. The video
level in IRE units is shown on the left; the radiated carrier signal is shown on the right.
Field-rate distortions appear as a difference in shading from the top to the bottom of the pic-
ture. A field-rate 60 Hz square wave is best suited for measuring field-rate distortion. Distortion
of this type is observed as a tilt in the waveform in the 2-field mode with the dc restorer off.

Figure 11.4.7 Waveform monitor display showing additive 60 Hz degradation. (Courtesy of Tek-
tronix.)
Line-rate distortions appear as streaking, shading, or poor picture stability. To detect such
errors, look for tilt in the bar portion of a pulse-and-bar signal. The waveform monitor should be
in the 1H or 2H mode with the fast dc restorer selected for the measurement.
The multiburst signal is used to test the high-frequency response of a system. The multiburst
includes packets of discrete frequencies within the television passband, with the higher frequen-
cies toward the right of each line. The highest frequency packet is at about 4.2 MHz, the upper
frequency limit of the NTSC system. The next packet to the left is near the color subcarrier fre-
quency (3.58 MHz, approximately) for checking the chrominance transfer characteristics. Other
packets are included at intervals down to 500 kHz. The most common distortion is high-fre-
quency rolloff, seen on the waveform monitor as reduced amplitude packets at higher frequen-
cies. This type of problem is shown in Figure 11.4.8. The television picture exhibits loss of fine
detail and color intensity when such impairments are present. High frequency peaking, appear-
ing on the waveform as higher amplitude packets at the higher frequencies, causes ghosting on
the picture.
Differential Phase
Differential phase (dφ) distortion occurs if a change in luminance level produces a change in the
chrominance phase. If the distortion is severe, the hue of an object will change as its brightness
changes. A modulated staircase or ramp is used to quantify the problem. Either signal places
chrominance of uniform amplitude and phase at different luminance levels. Figure 11.4.9 shows
a 100 IRE modulated ramp. Because dφ can change with changes in APL, measurements at the
center and at the two extremes of the APL range are necessary.
To measure dφ with a vectorscope, increase the gain control until the vector dot is on the edge
of the graticule circle. Use the phase shifter to set the vector to the 9 o'clock position. Phase error

Figure 11.4.8 Waveform monitor display of a multiburst signal showing poor high-frequency
response. (Courtesy of Tektronix.)
Figure 11.4.9 Waveform monitor display of a modulated ramp signal. (Courtesy of Tektronix.)
appears as circumferential elongation of the dot. The vectorscope graticule has a scale marked
with degrees of dφ error. Figure 11.4.10 shows a dφ error of 5º.
More information can be obtained from a swept R–Y display, which is a common feature of
waveform monitor and vectorscope instruments. If one or two lines of demodulated video from
the vectorscope are displayed on a waveform monitor, differential phase appears as tilt across the
line. In this mode, the phase control can be adjusted to place the demodulated video on the base-
line, which is equivalent in phase to the 9 o'clock position of the vectorscope. Figure 11.4.11
shows a dφ error of approximately 6º with the amount of tilt measured against a vertical scale.

Figure 11.4.10 Vectorscope display showing 5° differential phase error. (Courtesy of Tektronix.)
Figure 11.4.11 Vectorscope monitor display showing a differential phase error of 5.95° as a tilt on
the vertical scale. (Courtesy of Tektronix.)
This mode is useful in troubleshooting applications. By noting where along the line the tilt
begins, it is possible to determine at what dc level the problem starts to occur. In addition, field-
rate sweeps enable the operator to look at dφ over the field.
A variation of the swept R–Y display may be available in some instruments for precise mea-
surement of differential phase. Highly accurate measurements can be made with a vectorscope
that includes a precision phase shifter and a double-trace mode. This method involves nulling the
lowest part of the waveform with the phase shifter, and then using a separate calibrated phase
control to null the highest end of the waveform. A readout in tenths of a degree is possible.

Figure 11.4.12 Vectorscope display of a 10 percent differential gain error. (Courtesy of Tektronix.)
Differential Gain
Differential gain (dG) distortion refers to a change in chrominance amplitude with changes in
luminance level. The vividness of a colored object changes with variations in scene brightness.
The modulated ramp or staircase is used to evaluate this impairment with the measurement taken
on signals at different APL points.
To measure differential gain with a vectorscope, set the vector to the 9 o'clock position and
use the variable gain control to bring it to the edge of the graticule circle. Differential gain error
appears as a lengthening of the vector dot in the radial direction. The dG scale at the left side of
the graticule can be used to quantify the error. Figure 11.4.12 shows a dG error of 10 percent.
Differential gain can be evaluated on a waveform monitor by using the chroma filter and
examining the amplitude of the chrominance from a modulated staircase or ramp. With the wave-
form monitor in 1H sweep, use the variable gain to set the amplitude of the chrominance to 100
IRE. If the chrominance amplitude is not uniform across the line, there is dG error. With the gain
normalized to 100 IRE, the error can be expressed as a percentage. Finally, dG can be precisely
evaluated with a swept display of demodulated video. This is similar to the single trace R–Y
methods for differential phase. The B–Y signal is examined for tilt when the phase is set so that
the B–Y signal is at its maximum amplitude. The tilt can be quantified against a vertical scale.
ICPM
Television receivers may use a method known as intercarrier sound to reproduce audio informa-
tion. Sound is recovered by beating the audio carrier against the video carrier, producing a 4.5
MHz IF signal, which is demodulated to produce the sound portion of the transmission. From the
interaction between audio and video portions of the signal, certain distortions in the video at the
transmitter can produce audio buzz at the receiver. Distortions of this type are referred to as inci-
dental carrier phase modulation or ICPM. The widespread use of stereo audio for television
increased the importance of measuring this parameter at the transmitter, because the buzz is

Figure 11.4.13 Waveform monitor display using the ICPM graticule of a five-level stairstep signal
with no distortion. (Courtesy of Tektronix.)
more objectionable in stereo broadcasts. It is generally suggested that less than 3° of ICPM be
present in the radiated signal.
ICPM is measured using a high-quality demodulator with a synchronous detector mode and
an oscilloscope operated in a high-gain X-Y mode. Waveform and vector monitors usually have
such a mode as well. Video from the demodulator is fed to the Y input of the scope and the
quadrature output is fed to the X input terminal. Low-pass filters make the display easier to
resolve.
An unmodulated 5-step staircase signal produces a polar display, shown in Figure 11.4.13, on
a graticule developed for this purpose. Notice that the bumps all rest in a straight vertical line, if
there is no ICPM in the system. Tilt indicates an error, as shown in Figure 11.4.14. The graticule
is calibrated in degrees per radial division for differential gain settings. Adjustment, but not mea-
surement, can be performed without a graticule.
Tilt and Rounding

Good picture quality requires that the characteristics of the visual transmitter be as linear as pos-
sible. The transmitted channel bandwidth of conventional television (6 MHz for NTSC) intro-
duces various obstacles to achieving a high degree of linearity. The wideband amplifiers of
television transmission systems must have a flat response over the entire frequency range of
interest. The design process considers solutions for the problems of spurious harmonics resulting
from stray component parameters, and distortions to signals from all sources. Frequency distor-
tion in television systems, ranging from 15 kHz to several hundred kilohertz, is more visible on
test equipment if a line-frequency square-wave signal (with a rise time of approximately 200 ns)
modulates the visual transmitter. Distortion characteristics of this range generally fall into the
category of rounding and tilt.

Figure 11.4.14 Waveform monitor display using the ICPM graticule of a five-level stairstep signal
with 5° distortion. (Courtesy of Tektronix.)
It is possible to run tests on a transmission system using a number of specific frequencies,

monitoring the response to each at various points in the system. However, a more efficient
method involves using square-wave test signals with rise times selected to avoid overshoots in
transmission. During monitoring of the signals at various test points, tilt and rounding of the
square-wave corners may be observed as impaired frequency response. Two square-wave fre-
quencies prove to be useful for this work. The lower frequency of 60 Hz (50 Hz in PAL) aids in
detecting response errors in the frequency range below approximately 15 kHz. A 15 kHz square
wave serves to determine response to several hundred kilohertz. These two test signals identify
low frequency response problems when observed on an oscilloscope as a tilt of the flat part of the
square wave (top or bottom) and a rounding of the corners. Ideal response would produce flat
tops and bottoms without any rounding at the corners.
Tilt can be determined for both test signal ranges. The measurement is made on a portion of
the square wave derived from a demodulator. Particular start and end points of the measurement
are suggested to avoid ringing as a result of pulse rise times. To ensure the best measurements,
the sweep range and vertical sensitivity of the oscilloscope are set as wide as possible to accom-
modate the entire length of trace segment of interest. Tops and bottoms of the square waves
should be measured separately. (See Figure 11.4.15.)
Rounding is determined from the 15 kHz signal. It is realistic to delay the start point and to
define a duration along the scope trace for the measurement. Separate observations should be
made on rising and falling sides of the square wave, as shown in Figure 11.4.16.
The effects of tilt and rounding on a video signal in the transmission system sometimes can be
seen in the demodulated image. Unless it is excessive, tilt is least obvious. It appears partly as a
flickering in the picture and partly as a variation in luminance level from top-to-bottom and/or
side-to-side of the image. Rounding, in this case, manifests itself as a reduction of detail in
medium-frequency parts of the image. In cases of excessive tilt or rounding error, it may be diffi-
cult for some TV receivers to lock on to the degraded signal.

Figure 11.4.15 Measurement of tilt on a visual signal with 50 Hz and 15 kHz signals. (After [3].)
Figure 11.4.16 Measurement of rounding of a 15 kHz square-wave test signal to identify high-fre-
quency-response problems in a television transmitter. (After [3].)
Group Delay
The transients of higher-frequency components (15 kHz to 250 kHz), produce overshoot on the
leading edge of transitions and, possibly, ringing at the top and bottom of the waveform area. The
shorter duration of the “flat” top and bottom lowers concern about tilt and rounding. A square
wave contains a fundamental frequency and numerous odd harmonics. The number of the har-
monics determines the rise or fall times of the pulses. Square waves with rise times of 100 ns (T)
and 200 ns (2T) are particularly useful for television measurements. In the 2T pulse, significant
harmonics approach 5 MHz; a T pulse includes components approaching 10 MHz. Because the
TV system should carry as much information as possible, and its response should be as flat as
possible, the T pulse is a common test signal for determination of group delay.

Figure 11.4.17 Tolerance graticule mask for measuring transient response of a visual transmitter.
A complete oscillation is first displayed on the screen. The timebase then is expanded, and the
signal X-Y position controls are adjusted to shift the trace into the tolerance mask. (After [4].)
Group delay is the effect of time- and frequency-sensitive circuitry on a range of frequencies.
Time, frequency, phase shift, and signal delay all are related and can be determined from circuit
component values. Excessive group delay in a video signal appears as a loss of image definition.
Group delay is a fact of life that cannot be avoided, but its effect can be reduced through predis-
tortion of the video signal. Group delay adjustments can be made before the modulator stage of
the transmitter, while monitoring of the signal from a feedline test port and adjusting for best
performance.
Group delay can be monitored using a special purpose scope or waveform graticule. The goal
in making adjustments is to fit all excursions of the signal between the smallest tolerance mark-
ings of the graticule, as illustrated in Figure 11.4.17. Because quadrature phase errors are caused
by the vestigial sideband transmission system, synchronous detection is needed to develop the
display.
Classification of Distortion Mechanisms

Distortions affecting video systems can be generally divided into three broad categories: 1) lin-
ear distortion, 2) nonlinear distortion, and 3) noise. Table 11.4.1 summarizes the significant lin-
ear distortions and test methods, and Table 11.4.2 lists the common causes and effects of linear
distortions. Nonlinear distortion mechanisms and causes are detailed in Tables 11.4.3 and 11.4.4,
respectively. The significant types of noise are given in Table 11.4.5.

Table 11.4.1 Summary of Significant Linear Distortions and Test Methods (After [5])
Parameter Definition Test Signal Test Method

Gain versus frequency Gain variation over a band Multiburst or sweep. Measure the signal level at
response from 500 kHz to the maxi- selected frequencies in dB
mum frequency range. with respect to the level at
500 kHz.
Long-time distortion Damped signal-level oscilla- Any test signal allowing a Measure overshoot peak as
tions resulting from sud- controlled change in the a percentage of the peak
den picture-level changes. average picture level. luminance level.
Field-time distortion Change in the shape of a Field-rate square wave. Measure the peak-to-peak
field-time square wave. tilt of top bar in percent with
respect to the center of bar
excluding the first and last
last 250 μs.
Line-time distortion Change in the shape of a Line-rate square wave. Measure the peak-to-peak
line-time square wave. tilt of top of bar in percent
with respect to the center
of bar excluding the first
and last 1 μs.
K factor Qualified subjective impair- 2T sine-squared pulse Read the K rating in percent
ment rating. on the appropriate grati-
cule.
Chrominance-to-luminance Change in the ratio of Modulated sine-squared Measure the amplitude of
gain chrominance versus lumi- pulse. the chrominance compo-
nance amplitude. nent in percent with
respect to the luminance
component.
Chrominance-to-luminance Change in timing of the Modulated sine-squared Measure the delay of
delay chrominance versus lumi- pulse. chrominance component in
nance. ns with respect to the lumi-
nance component.
Table 11.4.2 Causes and Effects of Linear Distortions (After [5].)
Cause Test Signal Distortion Picture Impairment

Poor gain/frequency Poor C/L ratio. Loss of luminance detail and color
response above 500 kHz saturation.
High-frequency group delay Delay of chrominance with Color fringing.
respect to luminance.
Limited bandwidth with Poor K rating. Echo-like disturbances.
sharp cutoff
Poor response in 20-Hz to Field-time waveform distortion. Vertical smearing.
2-kHz band
Poor response in 5- to 500- Line-time waveform distortion. Horizontal smearing.
kHz band

Table 11.4.3 Summary of Nonlinear Distortions and Test Methods (After [5].)

Luminance nonlinearity Output/input amplitude propor- Luminance staircase. Measure the percentage of
tionally change of a small unit the largest step amplitude
step function as the step level variation.
is shifted from blanking level to
white level. The average pic-
ture level is kept constant
Chrominance gain nonlin- Output/input chrominance sub- Modulated pedestal. Measure the percentage of
earity carrier amplitude proportional- steps 3 and 1 chrominance
ity change as the subcarrier amplitude change with ref-
amplitude is varied from a min- erence to step 2.
imum to a maximum specified
value. Luminance and aver-
age picture levels are kept con-
stant.
Chrominance phase nonlin- Chrominance subcarrier phase Modulated pedestal. Measure the largest phase
earity variation when the subcarrier difference (in degrees) of
amplitude is varied from a min- steps 3 or 2 with reference
imum to a maximum specified to step 1.
value. Luminance and aver-
age picture levels are kept con-
stant.
Chrominance to luminance Variation of the luminance sig- Modulated pedestal. Measure the percentage of
intermodulation nal amplitude resulting from the largest luminance level
the superimposition of a speci- change due to chromi-
fied amplitude chrominance nance level.
signal. The average picture
level is kept constant.
Differential gain Amplitude change of a constant Modulated ramp or stair- Measure the percentage of
small-amplitude chrominance case. the largest chrominance
subcarrier superimposed on a subcarrier level change
luminance signal level that with reference to the burst
changes from blanking to amplitude.
white. The average picture
Differential phase Change in phase of a constant Modulated staircase or Measure the largest chroma
small-amplitude chrominance ramp. phase change in degrees
subcarrier without phase mod- with reference to the burst
ulation superimposed on a phase.
luminance signal level that
changes from blanking to
white. The average picture

Table 11.4.4 Visible Effects of Nonlinear Distortions (After [5].)
Distortion Picture Impairment

Luminance nonlinearity Distortion of picture brightness values.
Chrominance/luminance intermodulation Chrominance-level-related picture brightness
changes.
Chrominance gain nonlinearity Distortion of color saturation values.
Differential gain Luminance-level-related color saturation changes.
Chrominance phase nonlinearity Chrominance-level-related color hue changes.
Differential phase Luminance-level-related color hue changes.
Table 11.4.5 Summary of Significant Types of Noise and Test Methods (After [5].)

Continuous random The ratio, expressed in dB, of the Flat field. Feed the test signal to the input of
noise nominal amplitude of the lumi- the equipment under test and
nance signal (714.3 or 700 mV) measure the S/N at the output
to the RMS amplitude of the with an RMS-reading instru-
noise in a frequency band ment. Alternately, use an oscillo-
extending from 10 kHz to the scope preceded by a band-
upper frequency of the video limiting filter and synchronized to
band (4.2, 5.0, 5.5, or 6.0 MHz). display a single line.
Hum The ratio, expressed in dB, of the Flat field. Feed a signal to the input of the
nominal amplitude of the lumi- equipment under test and mea-
nance signal (714.3 or 700 mV) sure the S/N at the output with a
to the peak-to-peak amplitude of peak-to-peak reading instrument.
the noise after band-limiting to 10 Alternately, use an oscilloscope
kHz. synchronized to display a single
field and measure the peak-to-
peak hum amplitude.
Other periodic noise The ratio, expressed in dB, of the Flat field. Feed a test signal to the equip-
nominal amplitude of the lumi- ment under test and use an oscil-
nance signal (714.3 or 700 mV) loscope to isolate, identify, and
to the peak-to-peak amplitude of measure the peak-to-peak ampli-
the noise in a frequency band tude of the interfering signal in
extending from 1 kHz to the the specified frequency band.
upper frequency of the video
band (4.2, 5.0, 5.5, or 6.0 Mhz).
Crosstalk The ratio, expressed in dB, of the Sweep signal and Feed a sweep signal to the
nominal amplitude of the lumi- flat field. “unwanted” channel and a flat
nance signal (714.3 or 700 mV) field to the “wanted” channel.
to the peak-to-peak amplitude of Normalize the gains and mea-
the interfering signal. sure the peak-to-peak value of
the interfering signal at the
wanted channel output.

The effects of noise vary depending on the type:

• Random, resulting in colored and monochrome specks in the displayed image.
• Hum, resulting in slow-moving horizontal black bars.
• Other periodic noise, resulting in a steady or slow-moving vertical or diagonal pattern or
group of patterns affecting the luminance and/or chrominance.
• Signal crosstalk, resulting in faint steady or slow-moving pictures of the interfering signals on
the display.
11.4.2 Automated Video Signal Measurement

Video test instruments based on computer systems provide the maintenance engineer with the
ability to rapidly measure a number of parameters with exceptional accuracy. Automated instru-
ments offer a number of benefits, including reduced setup time, test repeatability, waveform stor-
age and transmission capability, and remote control of instrument/measurement functions.
Typical features of this class of instrument include the following:
• Waveform monitor functions
• Vectorscope monitor functions
• Picture display capability
• Automatic analysis of an input signals
• RS-232 and/or GPIB I/O ports
Figure 11.4.18 shows a block diagram of a representative automated video test instrument. Sam-
ple output waveforms are shown in Figure 11.4.19.
Automated test instruments offer the end-user the ability to observe signal parameters and to
make detailed measurements on them. The instrument depicted in Figure 11.4.18 offers a “Mea-
sure Mode,” in which a captured waveform can be expanded and individual elements of the
waveform examined. The instrument provides interactive control of measurement parameters, as
well as graphical displays and digital readouts of the measurement results. Figure 11.4.20 illus-
trates measurement of sync parameters on an NTSC waveform.
Another feature of many automated video measurement instruments is the ability to set oper-
ational limits and parameters that—if exceeded—are brought to the attention of an operator. In
this way, operator involvement is required only if the monitored signal varies from certain preset
limits. Figure 11.4.21 shows an example error log.

Figure 11.4.18 Block diagram of an automated video test instrument. (Courtesy of Tektronix.)

(a )
(b )
Figure 11.4.19 Automated video test instrument output charts: (a) waveform monitor mode dis-
play of color bars, (b) vectorscope mode display of color bars. (Courtesy of Tektronix.)

Figure 11.4.20 Expanded view of a sync waveform with measured parameters. (Courtesy of Tek-
tronix.)
Figure 11.4.21 Example error log for a monitored signal. (Courtesy of Tektronix.)
11.4.3 References
1. SMPTE Recommended Practice RP 219: “High-Definition, Standard-Definition Compati-
ble Color Bar Signal,” SMPTE, White Plains, N.Y., 2002.

2. Bentz, Carl, and Jerry C. Whitaker: “Video Transmission Measurements,” in Maintaining

Electronic Systems, CRC Press, Boca Raton, Fla., pp. 328–346, 1991.
2. Bentz, Carl: “Inside the Visual PA, Part 2,” Broadcast Engineering, Intertec Publishing,
3. Bentz, Carl: “Inside the Visual PA, Part 3,” Broadcast Engineering, Intertec Publishing,
Overland Park, Kan., December 1988.
4. Robin, M., and M. Poulin: Digital Television Fundamentals, 2nd ed., McGraw-Hill, New
York, N.Y., 2001.
11.4.4 Bibliography
SMPTE Engineering Guideline EG 1, “Alignment Color Bar Test Signal for Television Picture
Monitors,” SMPTE, White Plains, N.Y., 1990.
Pank, Bob (ed.): The Digital Fact Book, 9th ed., Quantel Ltd, Newbury, England, 1998.

g g
Chapter
11.5
Application of the Zone Plate Signal
11.5.1 Introduction
The increased information content of advanced high-definition display systems requires sophis-
ticated processing to make recording and transmission practical [1]. This processing uses various
forms of bandwidth compression, scan-rate changes, motion-detection and motion-compensa-
tion algorithms, and other techniques. Zone plate patterns are well suited to exercising a complex
video system in the three dimensions of its signal spectrum: horizontal, vertical, and temporal.
Zone plate signals, unlike most conventional test signals, can be complex and dynamic. Because
of this, they are capable of simulating much of the detail and movement of actual video, exercis-
ing the system under test with signals representative of the intended application. These digitally
generated and controlled signals also have other important characteristics needed in test wave-
forms for video systems.
A signal intended for meaningful testing of a video system must be carefully controlled, so
that any departure from a known parameter of the signal is attributable to a distortion or other
change in the system under test. The test signal also must be predictable, so that it can be accu-
rately reproduced at other times or places. These constraints usually have led to test signals that
are electronically generated. In a few special cases, a standardized picture has been televised by a
camera or monoscope—usually for a subjective, but more detailed, evaluation of overall perfor-
mance of the video system.
A zone plate is a physical optical pattern, which was first used by televising it in this way.
Now that electronic generators are capable of producing similar patterns, the label “zone plate”
is applied to the wide variety of patterns created by video test instruments.
Conventional test signals, for the most part limited by the practical considerations of elec-
tronic generation, have represented relatively simple images. Each signal is capable of testing a
narrow range of possible distortions; several test signals are needed for a more complete evalua-
tion. Even with several signals, this method may not reveal all possible distortions or allow study
of all pertinent characteristics. This is true especially in video systems employing new forms of
sophisticated signal processing.
11-69
pp g
Figure 11.5.1 Multiburst video test waveform: (a, left) picture display, (b, right) multiburst signal as
viewed on a waveform monitor (1H). (Courtesy of Tektronix.)
11.5.1a Simple Zone Plate Patterns

The basic testing of a video communication channel historically has involved the application of
several single frequencies—in effect, spot-checking the spectrum of interest [1]. A well-known
and quite practical adaptation of this idea is the multiburst signal, shown in Figure 11.5.11. This
test waveform has been in use since the earliest days of video. The multiburst signal provides
several discrete frequencies along a TV line.
The frequency-sweep signal is an improvement on multiburst. Although harder to implement
in earlier generators, it was easier to use. The frequency-sweep signal, illustrated in Figure
11.5.2, varies the applied signal frequency continuously along the TV line. In some cases, the
signal is swept as a function of the vertical position (field time). Even in these cases, the signal
being swept is appropriate for testing the spectrum of the horizontal dimension of the picture.
Figure 11.5.3 shows the output of a zone plate generator configured to produce a horizontal
single-frequency output. Figure 11.5.4 shows a zone plate generator configured to produce a fre-
quency-sweep signal. Electronic test patterns, such as these, may be used to evaluate the follow-
ing system characteristics:
• Channel frequency response
• Horizontal resolution
• Moiré effects in recorders and displays
• Other impairments
Traditionally, patterns that test vertical (field) response have been less frequently used. As
new technologies implement conversion from interlaced to progressive scan, line-doubling dis-
play techniques, vertical antialiasing filters, scan conversion, motion detection, or other pro-
1. Figure 11.5.1(a) and other photographs in this chapter show the “beat” effects introduced by
the screening process used for photographic printing. This is largely unavoidable. The screen-
ing process is quite similar to the scanning or sampling of a television image—the patterns
are designed to identify this type of problem.

pp g
Figure 11.5.2 Conventional sweep-frequency test waveform: (a, left) picture display, (b, right)
waveform monitor display, with markers (1H). (Courtesy of Tektronix.)
Figure 11.5.3 Single horizontal frequency test signal from a zone plate generator: (a, left) picture
display, (b, right) waveform monitor display (1H). (Courtesy of Tektronix.)
Figure 11.5.4 Horizontal frequency-sweep test signal from a zone plate generator: ( a, left) picture
display, (b, right) waveform monitor display (1H). (Courtesy of Tektronix.)

pp g
Figure 11.5.5 Single vertical frequency test signal: (a, left) picture display, (b, right) magnified ver-
tical-rate waveform, showing the effects of scan sampling. (Courtesy of Tektronix.)
Figure 11.5.6 Vertical frequency-sweep picture display. (Courtesy of Tektronix.)
cesses that combine information from line to line, vertical testing patterns are increasingly
important.
In the vertical dimension, as well as the horizontal, tests can be done at a single frequency or
with a frequency-sweep signal. Figure 11.5.5 illustrates a magnified vertical-rate waveform dis-
play. Each “dash” in the photo represents one horizontal scan line. Sampling of vertical frequen-
cies is inherent in the scanning process, and the photo shows the effects on the signal waveform.
Note also that the signal voltage remains constant during each line, changing only from line to
line in accord with the vertical-dimension sine function of the signal. Figure 11.5.6 shows a ver-
tical frequency-sweep picture display.
The horizontal and vertical sine waves and sweeps are quite useful, but they do not use the
full potential of a zone plate signal source.

pp g
Figure 11.5.7 Combined horizontal and vertical frequency-sweep picture display. (Courtesy of
Tektronix.)
Producing the Zone Plate Signal

A zone plate signal is created in real time by a test signal generator [1]. The value of the signal at
any instant is represented by a number in the digital hardware. This number is incremented as the
scan progresses through the three dimensions that define a point in the video image: horizontal
position, vertical position, and time.
The exact method by which these dimensions alter the number is controlled by a set of coeffi-
cients. These coefficients determine the initial value of the number and control the size of the
increments as the scan progresses along each horizontal line, from line to line vertically, and
from field to field temporally. A set of coefficients uniquely determines a pattern, or a sequence
of patterns, when the time dimension is active.
This process produces a sawtooth waveform; overflow in the accumulator holding the signal
number effectively resets the value to zero at the end of each cycle of the waveform. Usually, it is
desirable to minimize the harmonic energy content of the output signal; in this case, the actual
output is a sine function of the number generated by the incrementing process.
11.5.1b Complex Patterns

A pattern of sine waves or sweeps in multiple dimensions may be produced, using the unique
architecture of the zone plate generator [1]. The pattern shown in Figure 11.5.7, for example, is a
single signal sweeping both horizontally and vertically. Figure 11.5.8 shows the waveform of a
single selected line (line 263, SMPTE 240M). Note that the horizontal waveform is identical to
the one shown in Figure 11.5.4(b), even though the vertical-dimension sweep is now also active.
Actually, different lines will give slightly different waveforms. The horizontal frequency and
sweep characteristics will be identical, but the starting phase must be different from line to line
to construct the vertical signal.
Figure 11.5.9 shows a 2-axis sweep pattern that is most often identified with zone plate gener-
ators, perhaps because it quite closely resembles the original optical pattern. In this circle pat-
tern, both horizontal and vertical frequencies start high, sweep to zero (in the center of the
screen), and sweep up again to the end of their respective scans. The concept of 2-axis sweeps
actually is more powerful than it might first appear. In addition to purely horizontal or vertical

pp g
Figure 11.5.8 Combined horizontal and vertical frequency sweeps, selected line waveform display
(1H). This figure shows the maintenance of horizontal structure in the presence of vertical sweep.
(Courtesy of Tektronix.)
Figure 11.5.9 The best-known zone plate pattern, combined horizontal and vertical frequency
sweeps with zero frequency in the center screen. (Courtesy of Tektronix.)
effects, there are possible distortions or artifacts that are apparent only with simultaneous excita-
tion in both axes. In other words, the response of a system to diagonal detail may not be predict-
able from information taken from the horizontal and vertical responses.
11.5.1c The Time (Motion) Dimension

Incrementing the number in the accumulator of the zone plate generator from frame to frame (or
field to field in an interlaced system) creates a predictably different pattern for each vertical scan
[1]. This, in turn, creates apparent motion and exercises the signal spectrum in the temporal
dimension. Analogous to the single-frequency and frequency-sweep examples given previously,

pp g
Figure 11.5.10 Vertical frequency-sweep picture display. (Courtesy of Tektronix.)
appropriate setting of the time-related coefficients will create constant motion or motion sweep
(acceleration).
Specific motion-detection and interpolation algorithms in a system under test may be exer-
cised by determining the coefficients of a critical sequence of patterns. These patterns then may
be saved for subsequent testing during development or adjustment. In an operational environ-
ment, appropriate response to a critical sequence could ensure expected operation of the equip-
ment or facilitate fault detection.
Although motion artifacts are difficult to portray in the still-image constraints of a printed
book, the following example gives some idea of the potential of a versatile generator. In Figure
11.5.10, the vertical sweep maximum frequency has been increased to the point where it is zero-
beating with the scan at the bottom of the screen. (Note that the cycles/ph of the pattern matches
the lines/ph per field of the scan.) Actually, in direct viewing, there is another noticeable artifact
in the vertical center of the screen: a harmonic beat related to the gamma of the display CRT.
Because of interlace, this beat flickers at the field rate. The photograph integrates the interfield
flicker, thereby hiding the artifact, which is readily apparent when viewed in real time.
Figure 11.5.11 is identical to the previous photo, except for one important difference—
upward motion of 1/2-cycle per field has been added to the pattern. Now the sweep pattern itself
is integrated out, as is the first-order beat at the bottom. The harmonic effects in center screen no
longer flicker, because the change of scan vertical position from field to field is compensated by
a change in position of the image. The resulting beat pattern does not flicker and is easily photo-
graphed or, perhaps, scanned to determine depth of modulation.
A change in coefficients produces hyperbolic, rather than circular 2-axis patterns, as shown in
Figure 11.5.12. Another interesting pattern, which has been used for checking complex codecs,
is shown in Figure 11.5.13. This is also a moving pattern, which was altered slightly to freeze
some aspects of the movement for the purpose of taking the photograph.

pp g
Figure 11.5.11 The same vertical sweep as shown in Figure 11.5.10, except that appropriate pat-
tern motion has been added to “freeze” the beat pattern in the center screen for photography or
other analysis. (Courtesy of Tektronix.)
Figure 11.5.12 A hyperbolic variation of the 2-axis zone plate frequency sweep. (Courtesy of Tek-
tronix.)
Figure 11.5.13 A 2-axis frequency sweep in which the range of frequencies is swept several times
in each axis. Complex patterns such as this may be created for specific test requirements. ( Cour-
tesy of Tektronix.)

pp g
11.5.2 References
1. “Broadening the Applications of Zone Plate Generators,” Application Note 20W7056, Tek-
tronix, Beaverton, Ore., 1992.

pp g

g g
Chapter
11.6
Picture Quality Measurement
11.6.1 Introduction
Picture-quality measurement methods include subjective testing, which is always used—at least
in an informal manner—and objective testing, which is most suitable for system performance
specification and evaluation [1]. A number of types of objective measurement methods are pos-
sible for digital television pictures, but those using a human visual system model are the most
powerful.
As illustrated in Figure 11.6.1, three key testing layers can be defined for the modern televi-
sion system:
• Video quality. This consists of signal quality and picture quality
• Protocol analysis. Protocol testing is required because the data formatting can be quite com-
plex and is relatively independent of the nature of the uncompressed signals or the eventual
conversion to interfacility transmission formats. Protocol test equipment can be both a source
of signals and an analyzer that locates errors with respect to a defined standard and deter-
mines the value of various operational parameters for the stream of data.
• Transmission system analysis. To send the video data to a remote location, one of many pos-
sible digital data transmission methods can be used, each of which imposes its own analysis
issues.
Table 11.6.1 lists several dimensions of video-quality measurement methods. Key definitions
include the following:
• Subjective measurements: the result of human observers providing their opinions of the video
quality.
• Objective measurements: performed with the aid of instrumentation, manually with humans
reading a calibrated scale or automatically using a mathematical algorithm.
• Direct measurements: performed on the material of interest, in this case, pictures (also known
as picture-quality measurements).
• Indirect measurements: made by processing specially designed test signals in the same man-
ner as the pictures (also known as signal-quality measurements). Subjective measurements
11-79
y
Figure 11.6.1 Video testing layers for digital television. (After [1].)
Table 11.6.1 Video Quality Definitions (After [1].)
Parameter In-Service Out-of-Service

Indirect measurement
Objective signal quality Vertical interval test signals Full-field test signals
Direct measurement
Subjective picture quality Program material Test scenes
Objective picture quality Program material Test scenes
are performed only in a direct manner because the human opinion of test signal picture qual-
ity is not particularly meaningful.
• In-service measurements: made while the program is being displayed, directly by evaluating
the program material or indirectly by including test signals with the program material.

y
Figure 11.6.2 Functional environment for traditional video measurements. (After [1].)
• Out-of-service: appropriate test scenes are used for direct measurements, and full-field test
signals are used for indirect measurements.
In the mixed environment of compressed and uncompressed signals, video-quality measure-
ments consist of two parts: signal quality and picture quality.
11.6.2 Signal/Picture Quality

Signal quality measurements are made with a suite of test signals as short as one line in the ver-
tical interval [1]. In a completely uncompressed system, such testing will give a good character-
ization of picture quality. This is not true, however, for a system with compression encoding/
decoding because picture quality will change based on the data rate, complexity, and encoding
algorithm. Picture-quality measurements, instead, require natural scenes (or some equivalent
thereof) that are much more complex than traditional test signals. These complex scenes stress
the capabilities of the encoder, resulting in nonlinear distortions that are a function of the picture
content.
Out-of-service picture-quality measurements are similar to indirect signal-quality measure-
ments in one aspect: the determination of the system response to a high-quality reference. How-
ever, they actually measure the degradation in reference picture quality rather than that of a
synthetic test signal. In service, such signals determine the response to program material and its
degradation through the system. If the program material is “easy,” the measurement may not
have a great deal of practical value.
Objective indirect signal-quality measurements are a reasonably good way to determine the
picture quality for uncompressed systems. That is, there is a good correlation between subjective
measurements made on pictures from the system and objective measurements made on a suite of
test signals using the same system. (See Figure 11.6.2.) The correlation is not perfect for all tests,
however. There are distortions in composite systems, such as false color signals caused by poorly
filtered high-frequency luminance information being detected as chroma. These distortions are
not easily measured by objective means. Also, there are objective measurements that are so sen-
sitive they do not directly relate to subjective results. However, such objective results often are

y
Figure 11.6.3 Functional environment for digital video measurements. (After [1].)
useful because their effect will be seen by a human observer if the pictures are processed in the
same way a number of times.
The use of digital compression has expanded the types of distortions that can occur in the
modern television system. Because signal-quality measurements will not do the job, objective
picture-quality measurements are needed, as illustrated in Figure 11.6.3. The total picture-quality
measurement space has increased because of subjective measurements that now include mul-
timinute test scenes with varying program material and variable picture quality. The new objec-
tive measurement methods must have strong correlation with subjective measurements and cover
a broad range of applications.
Even with all the objective testing methods available for analog and digital video, it is impor-
tant to have human observation of the pictures. Some impairments are not easily measured, yet
are obvious to a human observer. This situation certainly has not changed with the addition of
digital compression. Therefore, casual or informal subjective testing by a reasonably expert
viewer remains an important part of system evaluation and/or monitoring.
11.6.3 Automated Picture-Quality Measurement

Objective measurements can be made automatically with an instrument that determines picture
degradation through the system [1]. Two somewhat mutually exclusive ways are available for
classifying objective picture-quality measurement systems. Although there are several practical
methods with a variety of algorithmic approaches, they may be generally divided into the follow-
ing classes:
• Feature extraction, which does an essentially independent analysis of input and output pic-
tures.

y
Figure 11.6.4 Block diagram of the feature-extraction method of picture-quality analysis. (After
[1].)
• Picture processing, where the complete input and output pictures are directly compared in
some manner and must be available at the measurement instrument (also known as picture
differencing).
From a system standpoint, the glass box approach utilizes knowledge of the compression system
to measure degradation. An example would be looking for blockiness in a DCT system. The
black box approach makes no assumptions about operation of the system.
In feature extraction, analysis time is reduced by comparing only a limited set of picture char-
acteristics. These characteristics are calculated, and the modest amount of data is embedded into
the picture stream. Examples of compression impairments would be block distortion, blurring/
smearing, or “mosquito noise.” At the receiver, the same features of the degraded picture are cal-
culated, providing a measure of the differences for each feature (Figure 11.6.4). The major weak-
ness of this approach is that it does not provide correlation between subjective and objective
measurements across a wide variety of compression systems or source material.
For the picture-processing scheme, the reference and degraded pictures are filtered in an
appropriate manner, resulting in a data set that may be as large as the original pictures. The dif-
ference between the two data sets is a measure of picture degradation, as illustrated in Figure
11.6.5.
A number of approaches to objective picture quality measurement have been proposed by
researchers using the human visual system (HVS) model as a basis. Such a model provides an

y
Reference Processing Degraded

Picture System Picture
Image Picture Data Image

Processing Difference Processing
Picture Picture
Data Data
Results
Figure 11.6.5 Block diagram of the picture-differencing method of picture-quality analysis. (After
[1].)
image-quality metric that is independent of video material, specific types of impairments, and
the compression system used. The study of the HVS has been going on for some time, investigat-
ing such properties as contrast sensitivity, spatio-temporal response, and color perception [2].
Derivatives of this work include the JNDmetrix (Sarnoff/Tektronix) [3] and the Three-Layered
Bottom-Up Noise Weighting Model [4].
11.6.3a In-Service Measurements

In-service measurements are made simultaneously with regular program transmissions [5]. For
traditional analog signal quality measurements, this is easily accomplished using vertical blank-
ing interval test signals. In-service measurements also are important for systems incorporating
bit-rate reduction and other forms of digital processing. Picture quality analysis must use the
actual program material analyzed at the source and the processed output for the measurement.
Both picture differencing and feature extraction can be used for in-service measurements.
The availability of the source sequence may not be a limitation for in-service monitoring in most
applications, given that picture quality is limited primarily by the (transmission) compression
process. Presuming this allows a defect monitoring solution to be designed to look for trace arti-
facts, rather than a loss of image fidelity.
Although this approach would seemingly relegate double-ended picture quality testing to the
systems evaluations lab, it is conceivable that program producers may want to have absolute
image quality confirmed, or even guaranteed.
Real-time measurements are desirable for in-service operation. In this context, real-time
means that measurements are continuously available as the sequence passes through the system
under test. There may be a nominal delay for processing, much as is done with the averaging
function in sophisticated analog measurement systems. Generally speaking, every video field
would be evaluated.
Non-real-time is the condition where the source and processed sequence information are cap-
tured, with results available after some nominal time for computation, perhaps a minute or two.

y
Figure 11.6.6 In-service picture quality measurement system. (After [5].)
In-service picture quality measurement using the picture differencing method can be accom-
plished as shown in Figure 11.6.6. At the program source, a system-specific decoder is used to
provide the processed sequence to a picture quality evaluation system. The results would be
available in non-real-time, sampled real-time, or real-time, depending on the computational
power of the measurement instrument. The system-specific decoder provides the same picture
quality output as the decoder at the receive end of the transmission system.
Practical transmission systems for broadcast applications operate either virtually error-free
(resulting in no picture degradation) or with an unrecoverable number of errors (degradation so
severe that quality measurements are not particularly useful). Therefore, bit-error-rate evaluation
based on calculations in the receiver may provide sufficient monitoring of the transmission sys-
tem.
The foregoing is based on the assumption that the transmission path is essentially fixed.
Depending upon the application, this may not always be the case. Consider a terrestrial video
link, which may be switched to a variety of paths before reaching its ultimate destination. The
path also can change at random intervals depending on the network arrangement. The situation
gives rise to compression concatenation issues for which there are no easy solutions.
As identified by Uchida [6], video quality deterioration accumulates with each additional
coding device added to a cascade connection. This implies that, although picture deterioration
caused by a single codec may be imperceptible, cascaded codecs, as a whole, can cause serious
deterioration that cannot be overlooked. Detailed measurements of concatenation issues are doc-
umented in [6].
11.6.4 References
1. Fibush, David K.: “Picture Quality Measurements for Digital Television,” Proceedings of
the Digital Television ’97 Summit, Intertec Publishing, Overland Park, Kan., December
1997.

y
2. ANSI Standard T1.801.03-1996, “Digital Transport of One-Way Video Signals: Parame-

ters for Objective Performance Assessment,” ANSI, Washington, D.C., 1996.
3. Fibush, David K.: “Practical Application of Objective Picture Quality Measurements,” Pro-
ceedings IBC ’97, IEE, pp. 123–135, Sept. 16, 1997.
4. Hamada, T., S. Miyaji, and S. Matsumoto: “Picture Quality Assessment System by Three-
Layered Bottom-Up Noise Weighting Considering Human Visual Perception,” SMPTE
Journal, SMPTE, White Plains, N.Y., pp. 20–26, January 1999.
5. Bishop, Donald M.: “Practical Applications of Picture Quality Measurements,” Proceed-
ings of Digital Television '98, Intertec Publishing, Overland Park, Kan., 1998.
6. Uchida, Tadayuki, Yasuaki Nishida, and Yukihiro Nishida: “Picture Quality in Cascaded
Video-Compression Systems for Digital Broadcasting,” SMPTE Journal, SMPTE, White
Plains, N.Y., pp. 27–38, January 1999.

g g
Chapter
11.7
DTV Transmission Performance Issues
11.7.1 Introduction
For analog signals, some transmission impairments are tolerable because the effect at the
receiver is often negligible, even for some fairly significant faults [1]. With digital television,
however, an improperly adjusted transmitter could mean the loss of viewers in the Grade B cov-
erage area. DTV reception, of course, does not degrade gracefully; it simply disappears. Atten-
tion to several parameters is required for satisfactory operation of the 8-VSB system. First
among these is the basic FCC requirement against creating interference to other over-the-air ser-
vices. To verify that there is no leakage into adjacent channels, out-of-band emission testing is
required.
As with analog transmitters, flat frequency response across the channel passband is necessary.
A properly aligned DTV transmitter exhibits many of the same characteristics as a properly
aligned analog unit: flat frequency response and group delay, with no leakage into adjacent chan-
nels. In the analog domain, group delay results in chrominance/luminance delay, which degrades
the displayed picture but still leaves it viewable. Group delay problems in DTV transmitters,
however, result in intersymbol interference (ISI) and a rise in the bit error rate (BER), possibly
causing the receiver to drop in and out of lock. Even low levels of ISI can cause receivers operat-
ing near the edge of the digital cliff to lose the picture completely. Amplitude and phase errors
can cause similar problems, again, resulting in reduced coverage area.
11.7.2 Characterizing the Transmission System

While it may be important to measure a number of specific parameters to confirm proper opera-
tion of a particular 8-VSB transmitter, four measurement can typically be used to characterize
the transmitted signal performance compared to a theoretical ideal. These measurements are as
follows:
• Signal to noise. S/N is the simple ratio of desired signal to undesired signal power. Noise is
defined as anything that degrades or impairs the signal, including distortion products,
intersymbol interference caused by frequency response or group delay errors, or ordinary
white noise. With regard to VSB transmission, S/N is the average power of ideal symbol val-
11-87
ues divided by the noise power. This is the difference between the idea signal and the actual
signal as demodulated along the real (in-phase) axis.
• Modulus error ratio. MER is a complex form of the S/N measurement that is made by
including the quadrature channel information in the ideal and error signal power computa-
tions. MER and S/N will be approximately equal unless their is an imbalance between the I
and Q channels. If the value of MER is significantly less than the S/N, Q-axis clipping is
likely to occur. This results because the Q-axis contains most of the amplitude peaks.
• Pilot amplitude error. PAE measurements quantify the pilot carrier deviation from the idea
case.
• Error vector magnitude. EVM is the rms value of the magnitudes of the symbol errors
along the real (in-phase) axis, divided by the magnitude of the real (in-phase) part of the out-
ermost constellation state. Because EVM includes both I and Q channels, it will indicate
transmitter clipping slightly before S/N. EVM is the magnitude of error induced by noise and
distortions compared to an ideal version of the signal.
Several types of transmitter impairments can degrade the overall performance of the system.
These impairments divide roughly into linear and nonlinear errors.
11.7.2a Linear Errors

Transmitter response imperfections and small impedance mismatches can cause linear distor-
tions in the radiated signal. Group delay and frequency response problems can result in noise
emphasis and inter-symbol interference at the receiver. This may indicate a need for adjustment
of the final RF filter, or for a pre-correction system.
Frequency response error is a frequency domain (spectral) function. Specifically, it is the dif-
ference between the spectral response of an ideal 8-VSB signal based on the root-raised cosine
frequency response to that of the actual signal.
Group delay error is also a frequency domain function. Group delay is the delay that a spe-
cific portion of the spectrum experiences through the transmission path. The ideal curve is a con-
stant (flat) response across the channel.
11.7.2b Nonlinear Errors

Nonlinear errors cause spectral spreading outside the channel band, raising the adjacent channel
“shoulders.” Amplitude and phase nonlinearity will decrease the transmitter S/N, reducing the
coverage area. It follows that nonlinear error measurements can be used as a guide for adjusting
8-VSB transmitter linearity.
Amplitude errors are a result of gain errors caused by instantaneous signal amplitude excur-
sions. Beyond a certain point, transmitter output power decreases with increasing amplitude,
which gives rise to clipping. This is one of the causes of out-of-channel emissions (spectral
regrowth).
Phase errors exist in the amplitude domain. When the phase shift of a signal passing through
an amplifier varies with the input signal amplitude, phase errors will result.

(a )
(b)
Figure 11.7.1 An 8-VSB constellation diagram: (a) a near-perfect condition, (b) constellation dia-
gram with noise and phase shift (the spreading of the pattern is caused by noise; the slant is
caused by phase shift). (After [1].)
11.7.2c Phase Noise

Integrated phase noise is a single figure of merit describing the phase variations that the trans-
mitter frequency synthesizer adds to the digital modulation process. This variation causes phase
and frequency deviations and rotates the decision points away from their ideal phase values.
11.7.2d Bit Error Rate

Eye patterns and BER have become well-known parameters of digital signal measurement,
although they may not always be the best parameters to monitor for 8-VSB transmission. The
constellation diagram and MER, on the other hand, provide insight into the overall system health.
RF constellations are displayed on the I and Q axes. Constellations of tight vertical dot patterns
with no slanting or bending indicate proper operation, as illustrated in Figure 11.7.1a. The 8-
VSB levels are the in-phase signal, so they are displayed left to right.
An 8-VSB signal is a single sideband waveform with a pilot carrier added. In a single side-
band signal, phase does not remain constant. Therefore, the constellation points (dots) occur in a
vertical pattern. As long as the dot pattern is vertical and the points form narrow lines of equal
height, the signal is considered good and can be decoded.
Figure 11.7.1b shows an 8-VSB signal that has noise and phase shift. Noise is indicated by
the spreading of the dot pattern. Phase problems are indicated by the slant along the Q-axis.
Although BER is a valid measurement for 8-VSB, a better approach involves monitoring the
MER, which usually will reveal problems before the BER is affected. In many cases, MER pro-
vides enough warning time to correct problems that would result in an increase in the BER. MER
provides an indication of how far the points in the constellation have migrated from the ideal. A

considerable amount of migration might occur before boundary limits are exceeded. Degradation
in the BER is apparent only when those limits have been exceeded.
11.7.2e Transmission System Measurements

The key measurements required for a DTV transmitter proof of performance include output
power, pilot frequency, intermodulation products (IMD), error vector magnitude, and harmonic
power levels [2, 3]. The average or RMS output power can be measured with a calorimeter, or a
precision probe and average-reading power meter. The calorimeter measures flow rate and tem-
perature rise of the transmitter test load liquid coolant, from which average power may be com-
puted. Because this is the quantity of interest for DTV applications, no other conversion factors
are necessary. Note that there is no need to measure peak power so long as the average power is
known and EVM and spectral regrowth requirements are met.
Emissions Mask
In practice, it is difficult to measure compliance with the RF emissions mask directly. For near-in
spectral components, within the adjacent channel, a common procedure is as follows [2]:
• Measure the transmitter IMD level with a resolution bandwidth of 30 kHz throughout the fre-
quency range of interest. This results in an adjustment to the standard FCC mask by 10.3 dB.
Under this test condition, the measured shoulder breakpoint levels should be at least –36.7 dB
from the mid band level.
• Measure the transmitter output spectrum without the filter using a spectrum analyzer.
• Measure the filter rejection vs. frequency using a network analyzer.
• Add the filter rejection to the measured transmitter spectrum. The sum should equal the trans-
mitter spectrum with the filter.
Output harmonics may be determined in the same manner as the rest of the output spectrum.
They should be at least –99.7 dB below the mid band power level. EVM is checked with a vector
signal analyzer or similar instrument.
The output power, pilot frequency, inband frequency response, and adjacent channel spectrum
should be measured periodically to assure proper transmitter operation. These parameters can be
measured while the transmitter is in service with normal programming.
Sideband Splatter
Interference from a DTV signal on either adjacent channel into NTSC or another DTV signal
will be primarily due to sideband splatter from the DTV channel into the adjacent channel [2].
The limits to this out-of-channel emission are defined by the RF mask as described in the “Mem-
orandum Opinion and Order on Reconsideration of the Sixth Report and Order,” adopted Febru-
ary 17, 1998, and released February 23, 1998.
For all practical purposes, high-power television transmitters will invariably generate some
amount of intermodulation products as a result of non-linear distortion mechanisms. Intermodu-
lation products appear as spurious sidebands that fall outside the 6 MHz channel at the output of
the transmitter. (See Figure 11.7.2.) Intermodulation products appear as noise in receivers tuned

Figure 11.7.2 Measured DTV transmitter output and sideband splatter. (After [2]. Courtesy of Har-
ris.)
to either first adjacent channel, and this noise adds to whatever noise is already present. The
overall specifications for the FCC mask are given in Figure 11.7.3.
Salient features of the RF mask include the following [2]:
• The shoulder level, at which the sideband splatter first appears outside the DTV channel, is
specified to be 47 dB below the effective radiated power (ERP) of the radiated DTV signal.
When this signal is displayed on a spectrum analyzer whose resolution bandwidth is small
compared to the bandwidth of the signal to be measured, it is displayed at a lower level than
would be the case in monitoring an unmodulated carrier (one having no sidebands). If the
analyzer resolution bandwidth is 0.5 MHz, and the signal power density is uniform over 5.38
MHz (as is the case for DTV), then the analyzer would display the DTV spectrum within the
DTV channel 10.3 dB below its true power. The correction factor is 10 log (0.5/5.38) = 10.3
dB. Thus, the reference line for the in-band signal shown across the DTV channel is at –10.3
dB, relative to the ERP of the radiated signal.
• The shoulder level is specified as – 47 dB, relative to the ERP. The shoulder level is at 36.7
dB below the reference line at –10.3 dB.
• The RF mask is flat for the first 0.5 MHz from the DTV channel edges, at –47 dB relative to
the ERP, and is shown to be 36.7 dB below the reference level, which is 10.3 dB below the
ERP.
• The RF mask from 0.5 MHz outside the DTV channel descends in a straight line from a value
of –47 dB to –110 dB at 6.0 MHz from the DTV channel edges.
• Outside of the first adjacent channels, the RF mask limits emissions to –110 dB below the
ERP of the DTV signal. No frequency limits are given for this RF mask. This limit on out-of-
channel emissions extends to 1.8 GHz in order to protect 1.575 MHz GPS signals.

Figure 11.7.3 RF spectrum mask limits for DTV transmission. The mask is a contour that illus-
trates the maximum levels of out-of-band radiation from a transmitted signal permitted by the FCC.
This graph is based on a measurement bandwidth of 500 kHz. (After [3]. Used with permission.)
• The total power in either first adjacent channel permitted by the RF mask is 45.75 dB below
the ERP of the DTV signal within its channel.
• The total NTSC weighted noise power in the lower adjacent channel is 59 dB below the ERP.
• The total NTSC weighted noise power in the upper adjacent channel is 58 dB below the ERP.
In-Band Signal Characterization

The quality of the in-band emitted signal of a DTV transmitter can be specified and measured by
determining the departure from the 100 percent eye opening [3]. This departure, or error, has
three components:
• Circuit noise (white noise)
• Intermodulation noise caused by various nonlinearities
• Intersymbol interference
The combination of all of these effects can be specified and measured by the error vector magni-
tude. This measurement is described in [3].
Power Specification and Measurement

Conventional NTSC broadcast service allows a power variation ranging from 80 to 110 percent
of authorized power [3]. These values correspond to –0.97 and +0.41 dB, respectively. Because

Table 11.7.1 DTV Pilot Carrier Frequencies for Two Stations (Normal offset above lower channel
edge: 309.440559 kHz; after [4])
Channel Relationship DTV Pilot Carrier Frequency Above Lower Channel Edge
NTSC Station NTSC Station NTSC Station DTV Station
Zero Offset + 10 kHz Offset – 10 kHz Offset No Offset
DTV with lower adjacent 332.138 kHz 342.138 kHz 322.138 kHz
NTSC ± 3 Hz ± 3 Hz ± 3 Hz
DTV co-channel with 338.056 kHz 348.056 kHz 328.056 kHz
NTSC ± 1 kHz ± 1 kHz ± 1 kHz
DTV co-channel with DTV + 19.403 kHz above + 19.403 kHz above + 19.403 kHz above 328.8436 kHz
DTV DTV DTV ± 10 Hz
of the cliff effect at the fringes of the service coverage area for a DTV signal, the allowable lower
power value will have a direct impact on the DTV reception threshold. A reduction of 0.97 dB in
transmitted power will change the DTV threshold of 14.9 dB (which has been determined to
yield a 3 × 10-6 error rate) to 15.87 dB, or approximately a 1-mile reduction in coverage distance
from the transmitter. Therefore, the average operating power of the DTV transmitted signal is of
significant importance.
The ATSC in [3] recommends a lower allowed power value of 95 percent of authorized power
and an upper allowed power value of 105 percent of authorized power.
Measurement Routine
The foregoing list of 8-VSB measurements should be performed during commissioning of the
RF system, and at regular intervals over the life of the transmitter. Sophisticated test instruments
are available from a number of manufacturers to aid in this work. Among the features provided
on many instruments are automatic measurement of selected parameters and storage of results
for comparison with future measurements. The storage feature can be a powerful tool in identify-
ing developing problems at an early stage.
11.7.3 DTV Operating Frequency Considerations

Table 11.7.1 summarizes the various pilot carrier offsets for different interference situations in a
6 MHz system (NTSC environment) [4]. Note that if more than two stations are involved the
number of potential frequencies will increase. For example, if one DTV station operates at an
offset because of a lower-adjacent-channel NTSC station, a co-channel DTV station may have to
adjust its frequency to maintain a 19.403 kHz pilot offset. If the NTSC analog station operates at
an offset of plus or minus 10 kHz, both DTV stations should compensate for that. Cooperation
between stations will be essential in order to reduce interference.
11.7.3a Frequency Tolerances

The tightest specification is for a DTV station with a lower adjacent NTSC analog station [4]. If
both NTSC and DTV stations are at the same location, they may simply be locked to the same

reference. The co-located DTV station carrier should be 5.082138 MHz above the NTSC visual
carrier (22.697 kHz above the normal pilot frequency). The co-channel DTV station should set
its carrier 19.403 kHz above the co-located DTV carrier.
If there is interference with a co-channel DTV station, the analog station is expected to be sta-
ble within 10 Hz of its assigned frequency.
While it is possible to lock the frequency of the DTV station to the relevant NTSC station,
this may not be the best option if the two stations are not at the same location. It will likely be
easier to maintain the frequency of each station within the necessary tolerances. Where co-chan-
nel interference is a problem, that will be the only option.
In cases where no type of interference is expected, a pilot carrier-frequency tolerance of ±1
kHz is acceptable, but in all cases, good practice is to use a tighter tolerance if practicable.
11.7.3b Hardware Options for Tight Frequency Control

Oscillator frequency tolerance is generally expressed as a fraction—in parts per billion, for
example [4]. Given a worst case example of a UHF station (NTSC or DTV) on channel 69, the
frequency stability required to meet the ± 3 Hz tolerance will be 3.7 × 10–9 or 3.7 parts per bil-
lion. If two stations are involved and the oscillators are not locked to a common reference, each
must maintain its frequency within half that range, or 1.8 × 10–9.
A high-quality oven-controlled crystal oscillator (OXCO) can achieve stability of 5 × 10–11
per day or 3 × 10–10 per year (30 parts per billion) at oscillator frequencies up to 30 MHz. This is
good enough for the short term, but would have to be monitored and adjusted throughout the
year. Long term it would not meet the ±10 Hz or 12.5 × 10–9 requirement for co-channel DTV
stations.
A rubidium standard is much better and will meet the worst-case requirements. Commercial
equipment may have stability of better than 5 × 10–10 (0.5 parts per billion) per year and/or 1 ×
10–10 per month with a life of 15 years.
The Global Positioning System (GPS) satellites provide a more stable reference. Commer-
cially available GPS time-reference receivers provide a signal with an accuracy of 1 × 10–12
(0.001 parts per billion). A disciplined clock oscillator with a long time constant is required to
eliminate short-term fluctuations in the GPS reference. In the event the GPS signal is lost, these
devices can maintain an accuracy of better than 1 × 10–10 per day (0.1 part per billion) for short
periods of time until the signal returns, well within the requirements.
11.7.3c Additional Considerations

Stations may have to modify their NTSC exciters for greater stability [4]. A rubidium or GPS
standard combined with a direct digital synthesizer may be the easiest way to do this.
To minimize interference in the case of a DTV transmitter on an upper adjacent channel to an
analog transmitter, the chroma subcarrier frequency of the analog signal must be precisely con-
trolled, in addition to the DTV pilot frequency and the analog carrier frequency.
Phase noise specifications should be checked when considering a frequency source. It is rec-
ommended that the phase noise of the transmitted signal at 20 kHz be at least –104 dBc/Hz.
Some frequency synthesizers may not meet this requirement.

11.7.4 References
1. Reed-Nickerson, Linc: “Understanding and Testing the 8-VSB Signal,” Broadcast Engi-
neering, Intertec Publishing, Overland Park, Kan., pp. 62–69, November 1997.
2. DTV Express Training Manual on Terrestrial DTV Broadcasting, Harris Corporation,
Quincy, Ill., September 1998.
3. ATSC: “Transmission Measurement and Compliance for Digital Television,” Advanced
Television Systems Committee, Washington, D.C., Doc. A/64-Rev A, May 30, 2000.
4. ATSC: “Guide to the Use of the ATSC Digital Television Standard,” Advanced Television
Systems Committee, Washington, D.C., Doc. A/54A, 4 December 2003.

DTV Transmission Performance Issues 12_1
Subject Index
A
A/B sync 3-19
absolute delay 10-20
ac line disturbance 5-133
ac power control 8-50
accuracy 10-11
active antenna system 7-96
active mixer 6-49
ac-to-RF efficiency 3-94
adaptive equalizer 3-97
add-compare-select 7-78
additive white Gaussian noise 2-70, 2-80
address bus 9-11
addressable converters 7-107
addressable tap 7-119
Advance Television Technology Center 2-70, 2-80
advection 1-40
AGC control range 6-22
aggregate event information table 3-118
air filtering system 8-36
airflow rate 8-33
air-handling system 8-14
air-interlock switch 8-16
airplane flutter 7-32
all-digital IBOC 2-80
allotment plan (DTV) 3-69
all-pass filter 4-18
all-pole filter 4-19
AM stereo 2-72, 6-68
AM-FM receivers 6-39
amplitude discriminator 6-62
amplitude limiting 6-64
amplitude modulation 1-68, 2-5, 3-9
amplitude shift keying 2-60
amplitude-tracking loop 7-76
analog compatibility tests 2-85
analog multiplier 7-35
ancillary data 3-43
annular control electrode (ACE) pulsing 3-77
antenna bandwidth 5-1
antenna efficiency 1-67
antenna elevation pattern 5-47
antenna gain 5-40
antenna height above average terrain 3-9
antenna matching 6-7
antenna power gain 3-9
antenna sweep 9-52
antenna tower 5-81, 5-82
antenna VSWR 5-53
antennas 5-33
aperiodic oscillator 6-57
aperture correction 11-32
arc detector 8-30
arcing 8-40
array antenna 5-22
aspect ratio 3-9
asynchronous counter 6-54
asynchronous data transfer 7-162
atmospheric charge energy 5-122
atmospheric electrical energy 5-116
atmospheric icing 5-83
atmospheric noise 6-41
ATSC A/80 3-111
ATSC GCR Standard 3-37
attenuation band 4-18
audio measurements 10-5
aural 3-75
aural amplifier 3-79
auto black circuit 11-37
automated test instrument 9-31
automatic chroma control 7-23
automatic fine tuning 7-24, 7-33
automatic frequency control 7-33
automatic gain control 6-47, 7-62
automatic gain control system 7-31
automatic level-control 7-120
automatic phase control 7-41
automatic-frequency-control 7-66
automatic-phase-and-frequency control 7-66
average AGC 7-31
average detection 9-53
average life 9-60
average luminance 11-15
average power 3-81

average-responding voltmeter 10-7
A-weighting curve 10-15
axial ratio 5-38
azimuthal pattern shaping 5-48
B
back lobe 5-25
backdriving 9-31
balance point 11-19
balanced peak detector 7-35
balun 7-89
band switching 9-53
band-reject filter 4-18
bandwidth 5-1, 7-86, 9-35, 10-11
bandwidth modification 1-67
bare rock grounding 5-106
Barkhausen effect 6-9
Barlow-Wadley principle 1-61
baseband 9-53
baseband equalization DOCR 3-103
baseline clipper 9-53
basic input/output system 9-9
bathtub curve 9-65
batteries 8-81
battery action (on tower elements) 5-101
batwing panel antenna 5-70
beam perveance 3-80
beam pulsing 3-77
beam tilt 5-13, 5-47, 5-74
beamwidth 5-12, 7-86
Bessel function 1-75, 2-12, 9-53
Bessel null method 9-53
best-efforts network 7-153
bidirectional cable system 7-125
binary frequency-shift keying 1-84
binary on-off keying 1-84
binary phase-shift keying 1-85, 2-60
binary training sequence 7-74
bi-phase shift keying 7-149
bit error rate 2-59, 6-28, 7-62, 9-31
bit-error-rate evaluation 11-85
black body radiation 1-11
blackout 8-71
blend to analog 2-80
blend to mono 2-80
blind equalization 7-74
block error rate 2-70, 2-80
blocking oscillator 7-49
Bode diagram 1-54
body-current overload 8-29
bonding 5-106
bouncing APL signal 11-50
bouquet association table (BAT) 3-123
breaking radiation 1-17
bridge amplifier 7-117
brightness 3-12, 11-32
brightness control 7-36
broadband antenna 5-74
broadband panel radiator 5-74
broadband tracking filter 7-64
Broadcast Internet 7-155
broadcast videoconference 7-168
bulkhead panel 5-136
bulkhead panel grounding 5-137
buried ground system 5-102
burn-in 9-60, 9-65
burst-referenced ACC system 7-46
business networks 7-106
butterfly antenna 5-70
C
C band 7-137
cable modem termination system 7-126
cable service 7-133
cable system 7-133
cable television 7-105
cable television system 7-133
cable-ready tuner 7-123
cadence signal 3-107
calibrator 9-53
camera black shading 11-38
camera detail circuit 11-38
camera encoder measurement 11-37
camera flare correction 11-39
camera performance mesurement 11-35
camera scanning aperture 11-10
camera scanning system 11-13
camera white shading 11-38
candelabra structure 5-65
candelabra tower 5-73
capture ratio 6-22
carrier wave 1-70
carrier-to-noise ratio 7-145, 9-53
causal analysis 9-81

cavity amplifier 2-16
cell voltage 8-82
center frequency 9-53
center of gravity law 11-19
centimeter waves 1-13
channel equalizer 7-62
channel equalizing 6-32
channel filtering 7-62
channel impulse response 6-28
channel-coding domain 3-44
characteristic impedance 4-8, 8-63
charge-coupled device converter 9-39
chemically activated ground system 5-94
chirp pulse 3-37
chirp spread spectrum 1-86
chroma response filter 11-48
chroma subcarrier 7-46
chromatic adaptation 11-22
chromaticity coordinates 11-20
chromaticity information 3-14, 11-8
chromaticity value 11-8
chrominance 3-10
chrominance amplitude and phase 11-32
CIE 1931 chromaticity diagram 11-23
circuit model 9-30
circular polarization 5-12, 5-29, 5-36, 5-69, 7-86
circular waveguide 4-14
circulators 4-28
city grade coverage 7-11
clear channel station 2-6
clipping 10-37, 10-40
cloud discharge 5-121
cloud-to-air flash 5-121
cloud-to-ground lightning 5-119
coax and waveguide grounding 5-108
coax bands 9-53
coaxial surge arrestor 5-141
coaxial transmission line 4-5
code de-interleaving 7-77
code rate 3-113, 7-149
codec 2-25
coded orthogonal frequency division multiplexing 2-62, 3-125
coherence bandwidth 6-35
coherence time 6-36
coherent AGC 7-70
coherent demodulator 6-7
cold-soak 9-65
color adjustment 11-31
color bar pattern 11-46
color bars 11-43
color reference pattern 11-40
color subcarrier reference 7-41
color temperature 11-31, 11-32
color-bar pattern 11-47
color-difference signals 3-14
color-difference system 7-43
colorimetric color reproduction 11-21
comb generator 9-53
commercial satellite 7-137
common amplification 3-76
common mode failure 8-23
companding 2-26
compatibility 2-70, 2-80
compatibility tests 2-85
complex number 8-64
complex signals 10-20
composite bandwidth 5-1
composite basehand 6-66
composite stereo baseband signal 2-23
concatenation 11-1, 11-85
confidence counter 7-80
confidence flag 7-80
connectivity library 9-30
constant efficiency amplifier 3-91
constant-current charge 8-83
constellation diagram 11-89
content distribution network 7-154
content replication 7-154
contrast control 7-36
contrast range 11-8
contrast ratio 11-25
contribution service 3-112
control and status bus 9-11
control ladder 8-40
control system 8-7
convergence 7-151, 11-32
conversion resolution 9-39
convolutional coding 2-62
convolutional de-interleaver 7-79
cooling system 8-14, 8-36
coring 7-40
corner frequency 4-18
corner-reflector antenna 5-20

corresponding color reproduction 11-21
Cosine Law 5-17
cosmic rays 1-14
coulomb 5-119
coupling-coefficient factor 5-39
C-QUAM 6-69
crest factor 10-8
critical array 5-27
critical frequency 4-9
critical value 3-72
cross modulation 6-13, 6-14, 7-65, 7-120
crossover distortion 10-37, 10-40
cross-polar energy 4-14
cross-polarization 2-38
crosstalk 10-13
crystal oscillator 1-50
cumulative distribution function 3-58, 3-64
current rate 8-82
current-limit battery charger 8-83
current-limiting 8-53
cutoff frequency 4-11, 4-18
cycle life 8-82
cycle stealing 9-11
D
D/U 2-81
data bus 9-11
data de-interleaver 7-63
data derandomizer 7-63
data field sync 7-70
data field synchronization 7-62
data frames 3-48
data rate 7-149
data segment sync 3-47, 7-68, 7-69
data sink 3-112
data source 3-112
data transmission 7-115
data-slicing 7-65
data-stream syntax 3-122
dc restoration 7-37
dc to light spectrum 1-10
decibel 10-12
decision feedback 7-76
decode time stamp (DTS) 3-44
deemphasis 1-79
degradation failure 9-60
degree of modulation 1-69
de-icing 5-85
delay element 3-11
delay spread 2-59, 3-98, 3-99, 6-31
delayed AGC 7-64
delta modulation 1-82
demodulation 6-57
demodulator 6-40, 7-23
demultiplexing 1-68
dense wavelength division multiplexing 7-128
derating factor 5-34
desensitization 6-13, 6-14
Designated Market Area 7-134
desired to undesired field strength 3-72
detector 6-57, 7-23
Diacrode 3-84
dielectric breakdown 8-50
dielectric constant 8-60
diesel generator 8-75
difference-frequency distortion 10-40
differential gain distortion 11-55
differential peak detector 7-35
differential phase distortion 11-52
diffraction 1-30
digital audio broadcasting 2-15, 2-57
digital data system 1-79
Digital Display Working Group 7-161
digital multimeter 9-19
digital on-channel repeater 3-100
digital performance tests 2-85
digital storage oscilloscope 9-37
Digital Visual Interface 7-161
diplexer 4-17, 5-76
dipole antenna 5-1, 5-20, 7-88
dipole factor 3-67
Dirac pulse 6-28
direct broadcast satellite 3-121
direct digital frequency synthesis 1-61
direct lightning strike 5-106
direct memory access 9-9
direct method 5-28
direct modulation 2-14
direct table lookup 1-61
directional antenna 5-36
directional antenna system 5-23
directional array 5-24
directional-coupler multitap 7-118
directivity 5-12
direct-to-home satellite broadcast 3-114
disassembler 9-28
discrete sine wave frequency-response measurement 11-30
discriminator 6-62
display measurement 11-25
dissipator electrode 5-123
distortion 9-53, 10-33
distortion analyzer 10-34
distributed constant circulator 4-30
distributed transmission adapter 3-107
distributed transmission network 3-104
distributed transmission packet 3-107
distribution service 3-112
distribution system 7-107
DMM 9-21
DOCR 3-100
dominant wavelength 3-12
Doppler effect 6-27, 10-38
Doppler shift 3-109
dot interlace 3-13
dot pitch 11-24
double reflector antenna 7-93
double-balanced mixer 6-51
double-sideband amplitude modulation 1-68
doubly truncated waveguide 4-15
down-link 7-137
downtime 8-39, 9-59, 9-60
driven electrode 5-89
DTV peak envelope power (PEP) 3-82
DTV planning factors 3-67
DTV receiver 7-61
DTV RF envelope 3-82
dual feeder system 8-73
Dual Link DVI 7-162
dual-mode channel combiner 5-77
dual-modulus prescaling 1-58
dual-polarity feed horn 7-139
dual-polarized transition 4-12
dual-slope conversion 9-21
dummy load 8-59
duty cycle 10-19
DVB applications 3-122
DVB-C 3-124
DVB-MC 3-125
DVB-MS 3-125
DVB-S 3-124
DVB-T 3-125
DVI-D 7-162
DVI-I 7-162
DWDM architecture 7-128
dynamic echo 7-74
dynamic intermodulation 10-41
dynamic random-access memory 9-9
dynamic range 6-12, 9-46, 9-53
E
E vector 4-11
earth 5-88
earth current 5-122
earth electrode resistance 5-88
earthing 5-88
earth-to-electrode resistance 5-89
effective isotropic radiated power 7-145
effective radiated power 2-16, 3-10, 3-69, 3-98, 3-101, 5-12, 5-35
EIA color bar signal 11-46
EIA/CEA 909 7-102
electric field 5-65, 5-129
electromagnetic focusing 3-80
electromagnetic generation 5-116
electromagnetic interference 9-17
electromagnetic pulse (EMP) 5-127
electromagnetic radiation 1-1
electromagnetic spectrum 1-9, 1-19
electromagnetic waves 5-65, 7-86
electron-bunching 3-76
electronic program guide (EPG) 3-123
electronic speed control 8-80
electrostatic discharge (ESD) 5-127
elevation gain 5-40
elliptical polarization 5-36
elliptical reflector 7-139
EM spectrum 1-9
Emerald Book 5-88
emergency power 8-72
emission multiplex 3-115
EMP event 5-129
EMP radiation 5-129
emulation 9-9
emulative tester 9-29
enabling event phenomenon 8-23
encoder 1-67
end-fire array 5-21

energy density 8-82
engine-generator set 8-72, 8-75
enhanced screening 9-70
envelope 9-53
envelope detector 6-60, 7-27, 7-114
envelope mode 9-39
Environmental Impact Statement 5-82
environmental stress screening 9-68
equiband chroma system 7-40
equipment safety 5-88
equipotential area 5-106
equivalent color reproduction 11-21
equivalent noise bandwidth 9-53
equivalent-time bandwidth 9-39
erbium doped fiber amplifier 7-128
error concealment 7-93
error vector magnitude 6-18, 11-88, 11-90, 11-92
error-correction 1-79
ethylene glycol 8-37
Eureka 147 2-62
event information table (EIT) 3-123
exact color reproduction 11-21
exothermic connection 5-99
extended-hybrid IBOC 2-81
external cavity klystron 8-13
external-cavity klystron 3-79
eye pattern 6-23
F
facility 5-87
fade margin 2-37, 3-97
fading 6-27
fading simulator 6-33
fail-safe system 8-40
failure 9-60
failure mode 8-54
failure mode and effects analysis 9-60
failure modes 9-63
failure rate 9-60
fan dipole 7-90
far-field region 5-14
fast-charge time 8-82
fast-Fourier-transform 10-16
fault injection 9-81
fault isolation 5-88
fault tree 9-30
fault tree analysis 9-60
fault-tolerant systems 9-80
FCC propagation curves 3-68
feature extraction 11-82
feedback balanced diode detector 7-28
ferrite isolator 4-28
ferrite loop antenna 6-43
fiber-based cable television 7-106
field 3-10
field rate side channel 3-109
field replaceable unit 9-13
field strength 3-68, 9-49
figure of merit 7-86
filament voltage 8-19
filament voltage regulation 8-18
filter 4-17, 6-45, 9-54
filter alignment 4-18
filter loss 9-54
filter order 4-19
finite-impulse-response filter 7-76
first detector 6-6, 6-40
first mixer 7-64
first mixer input level 9-54
first order loop 7-75
flange buildup 4-9
flash converter 9-39
flash energy 5-122
flat response mode 11-48
flatness 9-54
flaw-stimulus relationship 9-70
flesh-tone correction 7-46
flicker effect 6-9
float-charge 8-83
flyback generator 7-51
flyback transformer 7-52
FM propagation 5-29
FM stereo 6-66
focus 11-32
folded dipole 5-20
folded unipole antenna 5-28
forward error correction 3-102, 3-112, 3-124, 7-149
Foster-Seeley discriminator 6-63
Fourier analysis 9-54
Fourier transform 11-29
FPLL 7-64
fractional-N-division synthesizer 1-56
frame 3-10
free space loss 2-42, 7-146
free-space azimuthal pattern 5-60
frequency 10-21
frequency bands 1-9, 9-54
frequency changer 6-6
frequency counter 9-23
frequency demodulator 6-62
frequency detector 6-62
frequency deviation 9-54
frequency diversity 2-62
frequency divider 6-54
frequency domain representation 9-54
frequency hopping 1-86
frequency marker 9-54
frequency modulation 1-74, 2-12, 3-10
frequency plan 7-139
frequency pulling 1-52
frequency pushing 1-51
frequency range 9-54
frequency reference 1-45
frequency response 2-18, 11-88
frequency segments 1-9
frequency shift keying 2-61
frequency spectrums 1-9
frequency translation 1-67, 3-100
frequency-and-phase-lock-loop 7-64
frequency-division multiplexing 2-60
frequency-domain multiplexing 1-68
frequency-sweep signal 11-70
frequeney-division multiplexing 6-66
Fresnel zone 1-34, 2-39
functional testing 9-31
fundamental frequency 9-54
fundamental impedance 8-63
G
gain 5-12, 7-86
gain compression 6-14
gain control 6-19
galvanized wire 5-102
gamma 11-11
gamma correction 3-13
gamma ray band 1-17
gas turbine generator 8-77
gasoline generator 8-77
gate 9-62
gate pulse generator 7-41
gated AGC 7-31
gated coincidence detector 7-35
gelcell battery 8-83
generator rating 8-78
geometry and aspect ratio 11-32
geostationary arc 7-138
geosynchronous orbit 7-142
ghost canceling reference 3-35
ghost cancellation 3-35
glaze ice 5-83
glitch 9-27
glitch capture 9-39
Global Positioning System 11-94
Grade A coverage 7-11
Grade B contour 3-67
Grade B coverage 7-11
Grand Alliance receiver 7-80
graticule 9-54
gray scale tracking 11-32
grazing incidence 1-34
Green Book 5-88
ground 5-88
ground resistance test meter 5-92
ground ring 5-89
ground rod 5-90, 5-104, 5-133
ground system 5-133
ground system checklist 5-144
ground wave propagation 6-41
grounding 5-133
grounding electrode 5-89
grounding electrode conductor 5-88
grounding elements 5-133
ground-strap connection 5-108
group delay 10-20, 11-59, 11-88
guard interval 3-126, 3-127
guided-fault isolation 9-30
guy anchor grounding 5-101
guy galloping 5-84
guy wire 8-50
guy-anchor point 5-103
guy-anchor point grounding 5-98
guyed tower grounding 5-101
H
H vector 4-11
half-intensity width 11-29
half-power frequency 4-18

half-power points 5-1
half-power-width 11-28
half-wave dipole 1-21
Hamming window 10-27
Hanover bars 3-12
hard gamma rays 1-17
hard limiting 1-78
hard rime ice 5-84
hard X rays 1-17
harmonic distortion 9-54, 10-33
harmonic mixing 9-54
harmonic oscillation 5-84
harmonic output power 1-52
HAVI 7-163, 7-164
hazard rate 9-67
head end 7-105, 7-107, 7-108
heterodyne analyzer 10-16
heterodyne principle 6-5, 6-39
heterodyne processor 7-111
heterodyne spectrum analyzer 9-54
heterodyning 10-23
high voltage power supply 7-54
High-Definition Copy Protection 7-162
high-gain receive antenna 7-61
high-power amplifier 7-138
high-power isolator 4-30
high-voltage plate supply 8-27
high-voltage supply 8-51
Hilbert transform 7-76
Home Audio Video Interoperability Group 7-164
horizontal frequency interlace 3-13
horizontal polarization 5-12, 5-36
horizontal resolution 11-12
horizontal resolution factor 11-16
horizontal scan synchronization 7-50
horizontal scan system 7-52
horizontal shading 11-36
horizontally polarized 7-86
host digital terminal 7-126
hot carrier diode 6-50
hot switching 4-26
hot-soak 9-65
H-sync width 11-49
hub system 7-110
hue 3-12
hue-correcting system 7-46
hum 10-15
human visual system model 11-79
hum-to-hiss ratio 10-16
hybrid combiner 4-20
hybrid fiber/cable (HFC) architecture 7-125
hybrid fiber/coax (HFC) 7-106
hybrid IBOC 2-71, 2-81
I
iBiquity Digital 2-66
iBiquity FM IBOC system 2-80
IBOC 2-69, 2-71, 2-79, 2-81
ICPM 11-55
ideal impedance 8-63
IEEE 1212 7-162
IEEE 1394 7-130, 7-162
IF amplifier 7-23
IF gain 9-55
IF processing DOCR 3-100
image frequency 6-5, 6-39, 7-64
image signal 7-22
impedance 8-61
impedance matching 5-14, 5-58
impulse Fourier transform 11-28
impulse response 10-54
in-band on channel (radio) 2-15, 2-57, 2-65, 2-69, 2-79
incident wave 5-52
incidental carrier phase modulation 7-34, 11-55
incidental phase modulation 2-19
in-circuit testing 9-31
in-cloud icing 5-83
index of refraction 1-39
inductive output tube 3-77
infrared band 1-10
input impedance 7-86
in-rush current 8-53
in-service measurement 11-84
instantaneous frequency 9-55
instantaneous sampling 1-79
instruction set 9-11
integral cavity klystron 8-13, 3-79
integrated flyback transformer 7-56
integrated receiver decoder 3-115, 3-116, 3-123
integration period 9-21
interactive cable system 7-107
interactive television 3-123
intercarrier sound system 7-33, 11-55

intercept point 6-16
intercity relay 2-21
intercomponent confusion 11-17
interdiction technology 7-124
interference 6-15
interference rejection 3-100
interference-rejection filter 7-62
interlaced scanning 3-10
interlock circuit 8-52
intermediate amplifier 7-12
intermediate frequency 6-39, 9-55
intermediate power amplifier 3-79
intermittent service area 2-5
intermod 9-48
intermodulation 6-15
intermodulation distortion 2-19, 3-92, 9-55, 10-38
intermodulation noise 11-92
intermodulation products 3-100, 11-90
Internet 7-151
interpolation filter 3-132
interrupt 9-9
interrupt handler 9-9
inter-symbol interference 3-127
intersymbol interference 2-59, 6-18, 7-65, 7-73, 11-87, 11-92
interval trigger 9-43
inverse assembler 9-28
inverse fast Fourier transform 3-127
inverse-square law 7-138
ionization 5-128
IOT (inductive output tube) 3-84
IR band 1-11
IRE response 11-48
IRE standard scale 3-10
isochronous data transfer 7-163
isotropic antenna 1-20
isotropic radiator 1-20
J
jitter 6-23
JNDmetrix 11-84
Johnson counter 6-70
Johnson noise 6-8
K
Kell factor 11-28
kernel 9-9
keyed AGC 7-31
Klystrode 3-77
klystron 3-76, 3-84, 3-88, 7-144
knife-edge Fourier transform 11-28
knife-edge Fourier transform measurement 11-30
Ku band 7-137
L
L network 5-14
latent defect 9-68, 9-72
LC filter 6-46
LDMOS 3-83, 3-85
lead-acid battery 8-83
least-mean-square 7-73
lens back-focus 11-37
level 10-6
level measurement 10-6
light to gamma ray spectrum 1-10
lighting requirements 5-82
lightning 5-133
lightning arrestor 5-141
lightning effect 5-115
lightning prevention 5-122
lightning rod 5-126
lightning strike 5-118
line bounce 7-38
line image 11-8
line select 11-49
line stretcher 5-18
linear array antenna 5-22
linear channel distortion 7-73
linear distortion 11-59, 11-88
linear matrix 3-14
linear scale 9-55
linear-beam device 3-88
line-width measurement 11-27
liquefied-petroleum gas generator 8-77
lithium ion battery 8-82
lithium polymer battery 8-82
live pause 7-164
LO output 9-55
load current 8-82
local origination programming 7-106
local oscillator 6-39, 7-64, 9-55
logic analyzer 9-25
logic current tracer 9-24
logic probe 9-23

logic test clip 9-24
log-periodic antenna 5-21
log-periodic dipole 5-21
longitudinal line image 11-9
Longley-Rice 2-71, 2-81, 3-67
loop antenna 7-92
lower-sideband 2-5
lower-sideband signal 1-68
low-noise amplifier 7-138, 7-145
low-noise amplifier (LNA) 7-61
low-noise converter 7-145
low-noise preamplifier 7-108
luminance 3-10
luminance information 11-8
luminance level 11-8
lump constant circulator 4-30
M
machine cycle 9-10
magnetic declination 5-25
magnetic doublet 1-20
magnetic north 5-25
magnitude scale 8-67
main facility ground point 5-136
maintainability 9-59, 9-61
maintenance 9-76
maintenance log 8-7
maintenance program 8-7
major lobe 5-24
manometer 8-15
manual gain control 6-14
master guide table 3-117
matching network 5-14
mean time between failure (MTBF) 3-94, 9-61
mean time to repair (MTTR) 3-94, 9-61
meander gate 3-21
measurement set 9-30
medium wave broadcasting 2-5
memory map 9-10
message corruption 1-86
metal-oxide-semiconductor field-effect transistor (MOSFET) 3-83
meter readings 8-7
m-g set 8-74
microwave band 1-12
microwave multipoint distribution system (MMDS) 3-125
microwave radio service 7-137
microwave tower grounding 5-133
millimeter waves 1-13
minimum detectable signal 6-10
minimum discernible signal 6-12
minimum-phase 10-19
mixed-highs principle 3-12
mixer 6-39, 7-22
mixer circuit 6-49
mixing 9-55
mixing products 2-19
mod-anode pulsing 3-77
mode of failure 9-61
modulated staircase 11-52
modulated waveform 1-67
modulation index 1-74, 2-12
modulation system 1-67
modulation transfer function 11-26
modulus error ratio 11-88
moiré effects 11-70
moiré pattern 7-43
monitor point 5-28, 9-62
monochrome level 11-8
motion-detection and interpolation 11-75
mountaintop grounding 5-106
MPEG Layer II 3-122
MPEG-2 AAC 2-71, 2-81
MPEG-2 compliance point 3-122
MPEG-2 data packets 3-122
MSDC (multistage depressed collector) klystron 3-84
multiburst 11-49
multiburst signal 11-70
multicast videoconference 7-168
multichannel microwave 7-110
multi-format color bar signal 11-47
multiloop synthesizer 1-59
multipath 2-71, 2-81, 6-42
multipath distortion 8-25
multipath echo 3-36
multipath immunity 3-126
multipath interference 2-59
multiple frequency network 3-105, 3-109
multiple sampling 9-61
multiple-tower systems 5-60
multiplexed analog component (MAC) 3-121
multiplexer 5-76
multiplexing 1-67
multiresonator filter 6-46
multislope integration 9-22
multislot antenna 5-69
multistage depressed collector (MSDC) klystron 3-77
must-carry 7-134
N
National Electrical Code 5-87
National Radio Systems Committee 2-69, 2-71, 2-81
National Television System Committee 3-11
natural electrode 5-89
natural gas generator 8-77
natural radioactive decay 5-116
natural sampling 1-79
near video-on-demand 3-123, 7-125
negative sign convention 5-14
negative transmission 3-10
network analyzer 8-57
network information table (NIT) 3-123
nickel cadmium battery 8-82
nickel metal hydride battery 8-82
noise 9-55, 11-59
noise bandwidth 9-55
noise canceler 7-50
noise factor 6-6, 6-8
noise figure 6-8, 6-40, 7-61, 7-64
noise floor 9-55
noise gate 7-50
noise limited contour 3-106
noise measurement 10-15
noise sideband 9-55
noise temperature 7-145
nonlinear distortion 11-59, 11-88, 11-90
nonlinearity 10-44
nonsoil grounding 5-107
normalized characteristic impedance 8-64
normalized impedance 8-63
notch filter 4-18
notch-filter distortion analyzer 10-34
NRSC 2-81
NTSC 3-11, 3-37
NTSC composite video 7-44
NTSC interference rejection filter 7-72
NTSC rejection filter 7-77
null-fill 5-47
Nyquist criteria 1-79
O
objective testing 2-71, 2-81, 11-79
OCAP 7-130
occupied bandwidth 9-47
odd-order IM products 6-13
OFDM 2-82, 3-126
offset tracking loop 7-76
one-port measurements 10-5
Open Cable 7-130
open service information system 3-123
operating parameters 8-19
opportunistic 7-156
opportunistic bandwidth 7-156
optical spectrum 1-10
optical splitter 7-129
optically scalable node 7-128
orthogonal energy 4-14
orthogonal frequency division multiplexing 2-72, 2-82, 3-126, 3-127
orthogonal frequency-division multiplexing (OFDM) 3-125
oscillator 1-51
oscilloscope 9-35, 11-43
outer code 3-124
out-of-band emission testing 11-87
out-of-channel emission 11-90
out-of-service picture-quality measurements 11-81
oven-controlled crystal oscillator 1-46, 11-94
overscan 11-32
overshoot 10-51
P
P4 phosphor 11-31
PA compartment 8-16
PA tube 8-14
PA tuning 8-19
PAC 2-71, 2-81
PAL color system 3-11, 3-18
PAL sync 3-19
panel antenna 5-30, 5-49
parabolic reflector 7-139
parallel amplification 3-84
parallel resonant oscillator 6-57
parasitic director elements 5-21
part failure rate 9-61
parts-count method 9-63
passive filter 4-17
passive mixer 6-49

patent defect 9-72
path engineering 2-38
pattern null 5-24
pattern optimization 5-13
pattern spatial frequency 11-28
pattern trigger 9-43
payload 3-124
pay-per-view pay television service 7-107, 7-133
peak AGC 7-31
peak detection 9-39, 9-56
peak of sync 3-75
peak power shaving 8-74
peak-accumulation 9-39
peak-equivalent sine 10-8
peaking 9-56
peak-responding meter 10-7
peak-to-average power ratio 3-58
percentage of modulation 1-69
Perceptual Audio Coding 2-66, 2-71, 2-81, 2-82
perceptual coding 2-65
performance tests 2-85
perimeter ground 5-103
perimeter ground system 5-140
personal video recorder 7-164
personnel ground mat 5-106
personnel protection 5-105
personnel safety 5-88
phantom tee network 5-17
phase 10-19
Phase Alternation Line 3-11
phase constant 8-64
phase discriminator 6-63
phase lock 9-56
phase modulation 1-78
phase noise 11-89
phase shift 10-19
phase tracker 7-75, 7-76
phase/frequency detector 1-53
phased array 7-138
phase-locked loop 1-51, 2-15, 6-40, 6-53, 7-66, 10-23
phase-locked loop detector 7-35
phase-shift keying 2-60
phase-tracking loop 7-76
phasor common point 5-28
physical layer 3-44
pi network 5-17
picture differencing 11-83
picture element 11-8
picture monitor alignment 11-31
picture-quality measurement 11-79
piezoelectric effect 1-46
piezoelectric quartz crystal 1-45
pilot amplitude error 11-88
pilot carrier 7-120
pilot carrier frequency 3-61
pilot-carrier generator 7-110
pink noise 10-26
pitch 7-143, 10-21
pixel 11-9, 11-24
pixel density 11-24
pixel format 11-24
planar array 7-95
planar array antenna 5-22
plate supply overload 8-29, 8-41
point discharge theory 5-122
point image 11-8
point-of-presence 7-154
point-to-point network 7-153
point-to-point videoconference 7-167
polarization 3-10, 5-12, 5-36, 7-86
polarization ratio 5-38
polarized transition 4-12
post-delay ghost 3-36
post-tuning drift 1-52
power factor 4-8
pre-allocated bandwidth 7-155
precipitational icing 5-83
precorrection 3-79
predistortion 11-59
pre-echo 3-100
preemphasis 1-79, 2-13, 3-23
pre-equalizer 3-59
preferred color reproduction 11-22
pre-ghost 3-36
premises 5-87
premises wiring 5-87
prescaling 6-54
preselector 9-56
preselector filter 6-40
presentation time stamp (PTS) 3-44
pressure tap 7-118
preventive maintenance 8-1, 8-11, 8-32, 9-76
primary gamma rays 1-17
primary service area 2-5

prime center 8-65
prime focus, single parabolic reflector 7-93
prime mover 8-79
principal mode 4-11
probability density 1-82
product quality assurance 9-60
program clock reference base 3-44
program clock reference extension 3-44
program-specific information (PSI) 3-122
proof of performance 11-90
propagation 6-41
propagation constant 8-63, 8-64
propagation modes 1-19, 4-11
protected contour 2-82
protective zone 5-106
protocol analysis 11-79
protocol analyzer 9-31
pseudo random sequence 7-80
pulse amplitude modulation 1-79
pulse code modulation 1-80
pulse duration modulation 1-80
pulse frequency modulation 1-80
pulse length modulation 1-80
pulse modulation 1-79
pulse position modulation 1-80
pulse repetition frequency 9-56
pulse shaping 3-103
pulse stretcher 9-56
pulse time modulation 1-80
pulse width modulation 1-80, 2-9
pulse-width triggering 9-43
punctured convolutional code 3-124
purity 11-32
pylon antenna 5-69
Q
quadrature amplitude modulation (QAM) 1-74, 3-124
quadrature detector 7-35
quadrature distortion 7-29
quadrature phase-shift keying 2-60, 3-124, 7-149
quadriphase-shift keying 1-85
quality 9-62
quality assurance 9-61
quality audit 9-62
quantization 1-80, 9-38, 11-1
quantization error 1-80
quantization noise 1-82
quarter-wave monopol 7-92
quartz crystal 1-45, 6-55
quartz filter 6-46
quasi-parallel sound system 7-35
quasi-peak detector 10-8
R
radial ground 5-103
radial scale 8-67
radiating near-field region 5-14
radiation 7-86
radiation resistance 5-11, 7-86
radiative cooling 1-40
radio channel 6-27
Radio Data System 2-82
radio frequency band 1-12
radio horizon 2-38
radio receiver 6-7
radio refractivity 1-40
rain attenuation 7-147
random-noise signal 10-25
raster area 11-9
raster coordinates 11-8
raster spatial frequency 11-28
rating region table 3-117
ratio detector 6-63
ratio measurements 10-5
ray propagation 1-19
ray theory 1-30
Rayleigh channel 3-99
Rayleigh distribution 6-28
Rayleigh’s law 1-39
RDS 2-82
reactance curve 8-65
reactive near-field region 5-14
real-time analyzer 10-16, 10-24
receiver antenna system 7-85
receiver gain 6-8
receiver loop acquisition sequencing 7-63
reciprocal double exponential waveform 5-117
record length 9-39
Reed Solomon encoding/decoding 3-50
Reed-Solomon decoder 7-63
Reed-Solomon error-correction 7-73
Reed-Solomon outer coding 3-125
reference phase 4-21
reference tower 5-23

reflected wave 5-53
reflection coefficient 1-28
reflector dipole 5-21
refractivity gradients 1-40
registration error 11-36
relative velocity of propagation 4-6
relaxation oscillator 7-49
reliability 9-59, 9-61, 9-62
reliability growth 9-61
reliability prediction 9-61, 9-62
reliability-centered maintenance 9-77
residual distortion 10-43
resistance axis 8-65
resistance circle 8-65
resolution 9-45, 11-24
resolution bandwidth 9-56
retrace switch 7-51
retransmission consent 7-134
return loss 8-57, 9-56
return loss bridge 8-58
return path 7-129
reusable alkaline battery 8-82
revolving field alternator generator 8-79
RF combiner 4-17
RF emissions mask 11-90
RF interaction region 8-13
RF mask 11-90
RF processing DOCR 3-100
RF space current 3-88
RF system failure 8-7
RF troubleshooting 8-44, 9-52
RF watermark 3-99
RFI 5-133
RGB matrixing 7-43
Rice distribution 6-28
rigid coaxial cable 4-5
rime ice 5-83
ringing 10-51
rise time 10-49
rms 10-6
rolloff 4-18
root mean square 2-71, 10-6
root sum square 2-71
rounding 11-56
rubidium frequency standard 11-94
running status table (RST) 3-123
S
S/N 11-87
safety factor 5-35
salting 5-94
sample 9-61
sample loop 5-26
sampling interval 9-38
sampling lines 5-27
sampling principle 1-79
satellite 7-137
satellite antenna piers 5-108
satellite digital audio radio service 2-71, 2-82
satellite link 3-112
satellite receiving antenna 7-93
satellite relay system 7-137
satellite services 7-108
satellite transmission 3-113
satellite virtual channel table 3-118
saturation 3-12, 7-24
scan width 9-46
scanning 3-10
scan-synchronizing signal 7-47
scatter field 1-41
scatter propagation 1-41
scatter serration 5-60
scattering matrix 4-22
SDARS 2-71, 2-82
sealed lead acid battery 8-82
SECAM 3-11
SECAM III 3-22
SECAM IV 3-26
second detector 6-57
secondary effects of lightning 5-105
secondary gamma rays 1-15, 1-17
secondary service area 2-5
second-order loop 1-53
segment sync 7-62, 7-77
selective fading 6-41
selectivity 6-11, 6-40, 6-46
selenium thyrector 8-30
self-discharge 8-82
self-supporting tower grounding 5-101
sensitivity 6-8, 9-46, 9-56
sensitivity level 7-11
separate service information (SI) 3-123
separation 2-19, 10-13
sequence term 9-27

sequential sampling 9-61
service description table (SDT) 3-123
service multiplex and transport 3-43
set top box 3-115
shadow loss 1-36
shadow region 5-60
shape factor 9-56
shedding 5-84
short dipole 7-90
shot noise 6-9
shrinking raster measurement 11-26
sideband 9-56
sideband cancellation 3-60
sideband splatter 11-90
sideband suppression 9-56
Siemens star chart 11-37
signal crosstalk 2-19
signal multiplexing 1-67
signal processing 1-67
signal source 10-44
signal-to-noise ratio 1-76, 2-19
signature analysis 9-28
silver solder 5-99
sine-wave-square-wave IM test 10-41
single frequency network 3-97
Single Link DVI 7-162
single-frequency network (SFN) 3-126
single-phasing 8-31
single-shot bandwidth 9-39
single-sideband suppressed carrier AM 1-71
single-tone dynamic range 6-14
skew-symmetry 3-12
skin effect 4-1, 5-100
skywave propagation 2-7
slave synchronizer 3-109
slew rate 10-51
slew-rate limiting 10-37
slope-induced distortion 10-41
slotted line device 8-58
slotted-cylinder antenna 5-69
slugging 5-58
Smith chart 8-63
SMPTE 303M 11-40
SMPTE color bars 11-44
SMPTE IM distortion 10-38
SMPTE RP 166-1995 11-31
SMPTE RP 167-1995 11-31
SMPTE RP-219 11-46
soft gamma rays 1-17
soft X rays 1-17
software reliability 8-24, 9-79
soil conditions 5-90
soil resistance 5-88
soil resistivity 5-88
solar cells 7-138, 7-140
solar panel 7-140
solar wind 5-116
solid-state power amplifier 3-84, 7-141
source coding 3-42
source coding and compression 3-42
source-coding domain 3-44
spare parts 9-59
special-purpose electrode 5-89
spectral color reproduction 11-21
spectral components 11-13
spectral lines 1-13
spectral purity 9-49
spectral regrowth 11-88, 11-90
spectral spreading 11-88
spectroradiometric measurement 11-19
spectrum 9-57
spectrum analyzer 9-45, 9-47, 10-23
spectrum coordination 2-38
spectrum ripple 3-101
spectrum-efficient transmission 1-79
spiral antenna 5-22
sporadic E ionization 1-42
spread spectrum 1-86
spurious harmonics 9-49
spurious resolution 11-27
spurious response 9-57
spurious sidebands 11-90
spurs 1-52
square-topped sampling 1-79
standard broadcasting 2-5
standard PAL 3-12
standardization 9-63
standby carrier generator 7-113
standby power 8-72, 8-79
standby power system 8-78
standing wave 5-53
standing wave ratio 4-8
state 9-27
state analyzer 9-25

state qualified trigger 9-43
static buildup 8-50
static dissipator 5-125
static electricity 5-116
station keeping 7-138, 7-142
statistical quality control 9-62
stereophonic broadcasting 6-66
stereophonic transmission 2-14
stop band 4-18
store-and-forward 7-157
stress-analysis method 9-63
studio to transmitter link 2-21, 3-100
stuffing table (ST) 3-124
subjective CRT measurements 11-26
subjective testing 2-72, 2-82, 11-79
subrefraction 1-40
subscriber drops 7-107
subscriber networks 7-106
subsplitting 7-125
successive approximation converter 9-39
superheterodyne receiver 6-5, 6-39
supertrunk 7-110, 7-116
surface heating 1-40
surface-wave attenuation factor 1-25
surface-wave propagation 2-7
surge impedance 8-60
sweeping-filter technique 10-23
swinging burst 3-19
symbol clock recovery 7-62
symbol jitter 7-65
symbol level slicing 3-103
symbol rate 3-57
symbol synchronizer 7-63
sync-clamp AGC 7-31
synchronous AM 2-19
synchronous counter 6-54
synchronous demodulation 7-30
synchronous demodulator 6-60
synchronous detection 3-17, 7-63
synchronous detector 7-27, 7-30, 7-64
synchronous digital hierarchy 7-126
synchrotron radiation 1-13, 1-17
synthesizer 1-51, 1-53
system 9-61
system bandwidth 5-1
system downtime 8-8
system time table 3-118
system VSWR 5-53
T
tangential-fire mode 5-50
tangential-firing panel 5-71
tap-off device 7-118
TE waves 4-11
tee network 5-15
telemetry 7-142
television lines 11-24
television receive only (TVRO) system 7-108
television receiver 7-11
television sound system 7-33
television transmitter 3-75
television tuner 7-12
temperature control 8-32
temperature shock testing 9-65
temperature-compensated crystal oscillator 1-47
terrestrial microwave relay 7-108
test charts 11-14
test instruments 11-2
test pattern 11-43
test to failure 9-61
thermal distortion 10-37
thermal fatigue 8-32
thermal noise 6-8, 6-9, 7-138
thermal radiation 1-13
third-order IM dynamic range 6-18
Three-Layered Bottom-Up Noise Weighting Model 11-84
threshold 6-22
threshold-of-visibility (TOV) 7-61
thyristor power control 8-51
tilt 10-51, 11-56
time and date table (TDT) 3-123
time division multiplexing 2-60
time domain 10-49
time gating 7-40
time hopping 1-86
time-division multiplexing 6-66
time-domain multiplexing 1-68
time-domain reflectometer 8-57
time-invariant 6-36
timing analyzer 9-25
TM waves 4-11
tone burst 10-52
top sampling 1-79

topology database 9-30
topping-charge 8-83
torroidal current transformer 5-26
total harmonic distortion 2-18, 10-6
tower ground system 5-87
tracking filter 7-64
tracking generator 9-47, 9-57
training sequence 3-48, 6-32, 7-73
transfer characteristic 10-33
transfer function 6-35
transient intermodulation 10-41
transient overvoltage 8-30
transistor detector 7-27
translator 3-110
transmission impairments 11-87
transmission line 4-1, 7-95, 8-25, 8-57
transmission line impedance 4-7
transmission standards 3-1
transmitter field-strength 7-11
transmitter-to-studio link 2-22
transponder 7-139
transport stream 3-105, 3-118
transverse electromagnetic mode 4-7
transverse line image 11-9
traveling-wave tube 7-141, 7-144
Travis discriminator 6-62
tree-and-branch architecture 7-107
Trellis decoder 7-62
trellis decoding 3-103
triboelectric effect 5-127
trigger modes 9-43
tristimulus values 11-19
troubleshooting 8-39, 9-9
true north 5-25
true reading RMS voltmeter 8-18
trunk amplifier 7-119
trunk system 7-107
tuner 7-62
tuning linearity 1-52
tuning procedures 8-19
tuning sensitivity 1-52
tuning speed 1-53
turnover temperature 1-46
turnstile 5-68
turnstile antenna 5-51
TV limiting resolution 11-26
TVRO 7-144
twin-tone intermodulation 10-40
two-tone dynamic range 6-18
typical lightning flash 5-118
U
Ufer ground system 5-97
UHF tetrode 3-84, 3-86
UHF transmitter 3-76
UHF tuner 7-17
ultraviolet band 1-10
underscan 11-32
unit sequential sampling 9-61
up-link 7-137
upper-sideband 2-5
upper-sideband signal 1-68
UPS inverter 8-80
useful storage bandwidth 9-39
utility company power failure 8-72
utility service drop 8-73
UV band 1-12
V
V antenna 7-90
valid data window 9-11
varactor diode 6-55, 7-19
variable transformer 5-58
variable-frequency oscillator 6-7
varicaps 7-19
vector field 5-65
vectorscope 11-43, 11-49
velocity of propagation 8-60, 8-64
Venn diagram 9-70
vertical blanking interval 3-35
vertical interval reference 3-35, 7-47
vertical polarization 5-12, 5-36
vertical resolution 11-12
vertical scan system 7-49
vertical shading 11-36
vertical sync pulse sag 7-32
vertically polarized 7-86
vestigial sideband transmission 3-10
vestigial-sideband AM 1-71
VHF transmitter 3-76
VHF tuner 7-14
video amplifier 7-36
video demodulator 7-27
Video Electronics Standards Association 7-163

video gain 7-36
video information 11-7
video processing 11-1
video spectrum 11-15
video system distortion mechanisms 11-50
videoconferencing 7-166
viewing environment 11-31
virtual channel table 3-117
visible light band 1-10
visual 3-75
visual amplifier 3-79
Viterbi algorithm 6-32
VITS 11-49
voltage standing wave ratio 5-53, 8-25
voltage-controlled oscillator 1-53, 6-53, 6-54, 10-23
voltage-limit battery charger 8-83
von Kries model 11-22
VSB carrier recovery 7-62
VSWR 5-53, 8-25, 8-44, 8-60
VSWR overload 8-49
W
wave optics 1-30
waveform monitor 11-43
waveform subtraction 9-57
waveguide 4-1, 4-10
wavelength 5-1, 7-85
web caching 7-154
weighted quasi peak 2-82
weighting filter 10-15
white balance 11-22
white noise 2-70, 2-80, 10-26, 11-92
windload 5-68
window function 10-27
windstorm effect 5-122
windsway 5-40, 5-73
Y
Yagi-Uda antenna 5-21
Yagi-Uda array 7-92
yaw 7-143
Z
zone plate 11-69
zone plate chart 11-14
zone plate pattern 11-69
zone plate signal 11-73
About the Editors

Jerry C. Whitaker is Vice President of Standards Development at the Advanced Television Systems
Committee (ATSC), Washington, D.C. He was previously President of Technical Press, a consulting company
based in the San Jose (CA) area. Mr. Whitaker has been involved in various aspects of the electronics
industry for over 25 years, with specialization in communications. Current book titles include the following:
• Editor-in-chief, Standard Handbook of Video and Television Engineering, 4th ed., McGraw-Hill, 2002.
• Editor-in-chief, Standard Handbook of Audio and Radio Engineering, 2nd ed., McGraw-Hill, 2001
• DTV Handbook, 3rd ed., McGraw-Hill, 2001
• Editor, Television Receivers, McGraw-Hill, 2001
• Editor, Audio and Video Professional’s Field Manual, McGraw-Hill, 2001
• Editor, Master Handbook of Video Production, McGraw-Hill, 2002
• Editor, Master Handbook of Audio Production, McGraw-Hill 2002
• Editor, Audio/Video Protocol Handbook, McGraw-Hill 2001
Mr. Whitaker has lectured extensively on the topic of electronic systems design, installation, and
maintenance. He is the former editorial director and associate publisher of Broadcast Engineering and Video
Systems magazines, and a former radio station chief engineer and television news producer.
Mr. Whitaker is a Fellow of the Society of Broadcast Engineers and an SBE-certified professional
broadcast engineer. He is also a fellow of the Society of Motion Picture and Television Engineers.
K. Blair Benson (deceased) was an engineering consultant and one of the world’s most renowned television
engineers. Beginning his career as an electrical engineer with General Electric, he joined the Columbia
Broadcasting System Television Network as a senior project engineer. From 1961 through 1966 he was
responsible for the engineering design and installation of the CBS Television Network New York Broadcast
Center, a project that introduced many new techniques and equipment designs to broadcasting. He advanced
to become vice president of technical development of the CBS Electronics Video Recording Division. He
later worked for Goldmark Communications Corporation as vice president of engineering and technical
operations.
A senior member of the institute of Electrical and Electronics Engineers and a fellow of the Society of
Motion Picture and Television Engineers, he served on numerous engineering committees for both societies
and for various terms as SMPTE Governor, television affairs vice president, and editorial vice president. He
wrote more than 40 scientific and technical papers on various aspects of television technology. In addition, he
was editor of four McGraw-Hill handbooks: the original edition of this Television Engineering Handbook, the
Audio Engineering Handbook, the Television and Audio Handbook for Engineers and Technicians, and
HDTV: Advanced Television for the 1990s.

Standard Handbook of Broadcast Engineering

Uploaded by

Copyright:

Available Formats

Standard Handbook of Broadcast Engineering

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Standard Handbook of Broadcast Engineering

Uploaded by

Copyright:

Available Formats

ii

Standard Handbook of Broadcast Engineering

Copyright © 2005 by The McGraw-Hill Companies, Inc.

Christy and Tara

You never really leave the broadcast business…

Other books in the McGraw-Hill Video/Audio Series

Section 1: Frequency Bands, Propagation, and Modulation 1-1

Section 2: Radio Transmission Systems 2-1

Section 3: Television Transmission Systems 3-1

Section 4: RF Interconnection Devices and Systems 4-1

Section 5: Transmitting Antennas and Systems 5-1

Section 6: Radio Receivers 6-1

Section 7: Television Receivers and Cable/Satellite Distribution Systems 7-1

Section 8: RF System Maintenance 8-1

Section 9: Test Equipment 9-1

Section 10: Audio Test and Measurement 10-1

Section 11: Video Signal Measurement and Analysis 11-1

Edward W. Allen D. E. Manners

About the Book

Low Frequency (LF): 30 to 300 kHz

Medium Frequency (MF): 300 kHz to 3 MHz

• Fixed and mobile commercial communications

High Frequency (HF): 3 to 30 MHz

Very High Frequency (VHF): 30 to 300 MHz

Ultrahigh Frequency (UHF): 300 MHz to 3 GHz

• Fixed and mobile communications

Superhigh Frequency (SHF): 3 to 30 GHz

Chapter 1.1: The Electromagnetic Spectrum 1-9

Modulation Index 1-74

Reference Documents for this Section

Frequency Bands, Propagation, and

Low Frequency (LF): 30 to 300 kHz

Medium Frequency (MF): 300 kHz to 3 MHz

1-2 Section One

• Fixed and mobile commercial communications

High Frequency (HF): 3 to 30 MHz

Very High Frequency (VHF): 30 to 300 MHz

Ultrahigh Frequency (UHF): 300 MHz to 3 GHz

Frequency Bands, Propagation, and Modulation 1-3

• Fixed and mobile communications

Superhigh Frequency (SHF): 3 to 30 GHz

Chapter 1.1: The Electromagnetic Spectrum 1-9

1-4 Section One

Chapter 1.2: Propagation 1-19

Chapter 1.3: Frequency Sources and References 1-45

Chapter 1.4: Modulation Systems and Characteristics 1-67

Frequency Bands, Propagation, and Modulation 1-5

Modulation Index 1-74

Reference Documents for this Section

1-6 Section One

Frequency Bands, Propagation, and Modulation 1-7

Norgard, John: “Electromagnetic Spectrum,” The Electronics Handbook, Jerry C. Whitaker

1-8 Section One

1-10 Frequency Bands, Propagation, and Modulation

1.1.2 Spectral Sub-Regions

1.1.2a Optical Spectrum

Visible Light Band

The Electromagnetic Spectrum 1-11

1-12 Frequency Bands, Propagation, and Modulation

2. Some references use 4, 5, or 6 nm as the upper edge of the UV band.