Chapter 7.

Digital Transmission Systems

The preceding chapters have presented the fundamental characteristics of the individual building blocks of an
optical fiber transmission link. These include the optical fiber transmission medium, the optical source, the
photodetector and its associated receiver, and the connectors used to join individual fiber cables to each other and
to the source and detector. Now we shall examine how these individual parts can be put together to form a
complete optical fiber transmission link. In particular, we shall study digital links in this chapter,
We consider the simplest case of a point-to-point link. The complete link design will include examining the
components that are available for a particular application and seeing how these components relate to the system
performance criteria (such as dispersion and bit error rate). For a given set of components and a given set of
system requirements, we then carry out a power budget analysis to determine whether the fiber optic link meets
the attenuation requirements or if repeaters are needed to boost the power level. The final step is to perform a
system rise-time analysis to verify that the overall system performance requirements are met.
Then we consider line-coding schemes that are suitable for digital data transmission over optical fibers. These
coding schemes are used to introduce randomness and redundancy into the digital information stream to ensure
efficient timing recovery and to facilitate error monitoring at the receiver.
To increase the end-to-end fidelity of an optical transmission line, forward error correction (FEC) can be used if
the bit-error rate is limited by optical noise and dispersion.

Point to point Links

The simplest transmission link is a point-to-point line that has a transmitter on one end and a receiver on the
other. This type of link places the least demand on optical fiber technology and thus sets the basis for examining
more complex system architectures.

The design of an optical link involves many interrelated variables among the fiber, source, and photodetector
operating characteristics, so that the actual link design and analysis may require several iterations before they are
completed satisfactorily.
The key system requirements needed in analyzing a link are:
1. The desired (or possible) transmission distance
2. The data rate or channel bandwidth
3. The bit error rate (BER)
To fulfill these requirements the designer has a choice of the following components and their associated
characteristics:
1. Multimode or single-mode optical fiber
a. Core size
b. Core refractive-index profile
c. Bandwidth or dispersion
d. Attenuation
e. Numerical aperture or mode-field diameter
2. LED or laser diode Optical source
a. Emission wavelength

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 1/17
 b. Spectral line width
c. Output power
d. Effective radiating area
e. Emission pattern
f. Number of emitting modes
3. pin or avalanche photodiode
a. Responsivity
b. Operating wavelength
c. Speed
d. Sensitivity
Two analyses are usually carried out to ensure that the desired system performance can be met:
1. Link Power Budget
2. System Rise-Time Budget Analyses
In the link power budget analysis one first determines the power margin between the optical transmitter output
and the minimum receiver sensitivity needed to establish a specified BER. This margin can then be allocated to
connector, splice, and fiber losses, plus any additional margins required for expected component degradation or
temperature effects. If the choice of components did not allow the desired transmission distance to be achieved,
the components might have to be changed or repeaters might have to be incorporated into the link.
Once the link power budget has been established, the designer can perform a system rise-time analysis to ensure
that the desired overall system performance has been met.
System considerations
In carrying out a link power budget, we first decide at which wavelength to transmit and then choose components
operating in this region. If the distance over which the data are to be transmitted is not too far, we may decide to
Operate in the 800- to 900-nm region. On the other hand, if the transmission distance is relatively long, we may
want to take advantage of the lower attenuation and dispersion that occurs at wavelengths around 1300 or 1550
nm.
After deciding a wavelength, next we interrelate the system performances of the three major optical link building
blocks; that is, the receiver, transmitter, and optical fiber. The procedure we shall follow here is first to select the
photodetector. We then choose an optical source and see how far data can be transmitted over a particular fiber
before an amplifier is needed in the line to boost up the power level of the optical signal.
In choosing a particular photodetector, we mainly need to determine the minimum optical power that must fall
on the photodetector to satisfy the bit-error rate (BER) requirement at the specified data rate. The system
parameters involved in deciding between the use of an LED and a laser diode are signal dispersion, data rate,
transmission distance, and cost. For the optical fiber we have a choice between single-mode and multi-mode
fiber, either of which could have a step- or a graded-index core. This choice depends on the type of light source
used and on the amount of dispersion that can be tolerated.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 2/17
 Link Power Budget
An optical power loss mode for a point-to-point link is shown below

The optical power received at the photodetector depends on the amount of light coupled into the fiber and the
losses occurring in the fiber and at the connectors and splices. The link loss budget is derived from the sequential
loss contributions of each element in the link. Each of these loss elements is expressed in decibels (dB) as

where Pin and Pout, are the Optical powers emanating into and out of the loss element, respectively.
In addition to the link loss contributors shown in Fig. , a link power margin is normally provided in the analysis
to allow for component aging, temperature fluctuations, and losses arising from components that might be added
at future dates. A link margin of 6 to 8 dB is generally used for systems that are not expected to have additional
components incorporated into the link in the future.
The link loss budget simply considers the total optical power loss PT that is allowed between the light source and
the photodetector, and allocates this loss to cable attenuation, connector loss, splice loss, and system margin.
Thus, if PS is the optical power emerging from the end of a fiber flylead attached to the light source or from a
source-coupled connector, and if PR is the receiver sensitivity, then

Here, lc is the connector loss, α is the fiber attenuation (dB/km), L is the transmission distance, and the system
margin is nominally taken as 6 dB. Here, we assume that the cable of length L has connectors only on the ends
and none in between. The splice loss is incorporated into the cable loss for simplicity.
Example:
To illustrate how a link loss budget is setup, let us carry out a specific design example.
Consider we shall design a link with data rate of 20 Mb/s and a bit-error rate of 10-9 (i.e. at most one error can
occur for every 109 bits sent). Let’s say we use silicon p-i-n photodiode as receiver operating at 850 nm. To meet
above conditions of the link the required receiver input signal is -42 dBm (42 dB below 1 mW). And consider
we select GaAlAs LED that can couple a 50-µW (-13 dBm) average optical power level into a fiber flylead with
a 50-µm core diameter. Assume that a 1-dB loss occurs when the fiber flylead is connected to the cable and
another 1-dB connector loss occurs at the cable-photodetector interface. Including a 6-dB system margin, find
the possible transmission distance for a cable with an attenuation of α dB/Km.
Link Budget Analysis:
Here, Source power, Ps = -13 dBm
Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 3/17
 Required minimum receiver power, PR = -42 dBm
Thus allowable power loss is, PT = Ps - PR = -13dBm + 42dBm = 29dB
Now if L is the allowed transmission distance of optical fiber, then
𝑃𝑡 = 2(1 𝑑𝐵) + 𝛼 𝐿 + 6 𝑑𝐵
29 𝑑𝐵 = 2 𝑑𝐵 + 3.5 𝑑𝐵/𝐾𝑚 𝐿 + 6 𝑑𝐵
21 𝑑𝐵 = 3.5 𝑑𝐵/𝐾𝑚 𝐿
𝐿 = 21/3.5 𝐾𝑚
𝐿 = 6 𝐾𝑚
Hence, a 6.0 Km transmission path is possible for given point-to-point
The link power budget can be represented graphically as follow

The vertical axis represents the optical power loss allowed between the transmitter and the receiver. The
horizontal axis gives the transmission distance. Here we show a silicon pin receiver with a sensitivity of -42 dBm
(at 20 Mb/s) and an LED with an output power of -13 dBm coupled into a fiber fiylead. We subtract a l-dB
connector loss at each end, which leaves a total margin of 27 dB. Subtracting a 6-dB system safety margin leaves
us with a tolerable loss of 21 dB that can be allocated to cable and splice loss. The slope of the line shown in Fig.
is the 3.5 dB/km cable (and splice, in this case) loss. This line starts at the -14 dBm point (which is the optical
power coupled into the cabled fiber) and ends at the -35 dBm level (the receiver sensitivity minus a l-dB connector
loss and a 6-dB system margin). The intersection point D then defines the maximum possible transmission path
length, which is 6 Km.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 4/17
 Rise-Time Budget
A rise-time budget analysis is a convenient method for determining the dispersion limitation of an optical fiber
link. This is particularly useful for digital systems. In this approach, the total rise time tsys of the link is the square
root of sum of square of the rise times from each contributor ti to the pulse rise-time degradation:

The four basic elements that contribute pulse rise time are:
 the transmitter rise time ttx,
 Rise time due to fiber
o the group-velocity dispersion (GVD) rise time tGVD of the fiber,
o the modal dispersion rise time tmod of the fiber, and
 the receiver rise time trx
Single-mode fibers do not experience modal dispersion, so in these fibers the rise time is related only to GVD.
Generally, the total transition-time degradation of a digital link should not exceed 70 percent of an NRZ (non-
return-to-zero) bit period or 35 percent of a bit period for RZ (return-to-zero) data, where one bit period is defined
as the reciprocal of the data rate
The rise times of transmitters and receivers are generally known to the designer.
The transmitter rise time is attributable primarily to the light source and its drive circuitry.
The receiver rise time results from the photodetector response and the 3-dB electrical bandwidth of the receiver
front end.
Receiver Rise Time
The response of the receiver front end can be modeled by a first-order lowpass filter having a step response

Where, Brx, is the 3-dB electrical bandwidth of the receiver and

u(t) is the unit step function which is l for t  0 and 0 for t < 0.
The rise time trx of the receiver is usually defined as the time interval between g(t) = 0.1 and g(t) = 0.9. This is
known as the 10- to 90-percem rise time. Thus, if Brx is given in megahertz, then the receiver front-end rise time
in nanoseconds is

Fiber Rise Time

In practice, an optical fiber link seldom consists of a uniform, continuous, jointless fiber. Instead, a transmission
link nominally is formed from several concatenated (tandemly joined) fibers that may have different dispersion
characteristics. In addition, multimode fibers experience modal distributions at fiber-to-fiber joints owing to
misaligned joints, different core index profiles in each fiber, and/or different degrees of mode mixing in individual
fibers.
group-velocity dispersion (GVD) rise time
Determining the fiber rise times resulting from GVD and modal dispersion then becomes more complex than for
the case of a single uniform fiber.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 5/17
The fiber rise time tGVD resulting from GVD over a length L can be approximated by

Where, σλ is the half-power spectral width of the source

D is the dispersion.
Since the dispersion value generally change from fiber section to section in a long link, an average value should
be used for D.
In case of single mode fiber group velocity dispersion rises mainly from material dispersion. Hence,

Where Dmat(λ) is the material dispersions.

modal dispersion rise time
The difficulty in predicting the bandwidth (and hence the modal dispersion rise time, t mod) of a series of
concatenated fibers arises from the observation that the total route bandwidth can be a function of the order in
which fibers are jointed.
For example, instead of randomly joining together arbitrary (but very similar) fibers, an improved total link
bandwidth can be obtained by selecting adjoining fibers with alternating over- and undercompensated refractive-
index profiles to provide some modal delay equalization. Although the ultimate concatenated fiber bandwidth
can be obtained by judiciously selecting adjoining fibers for optimum modal delay equalization, in practice this
is unwieldy and time-consuming, particularly since the initial fiber in the link appears to control the final link
characteristics.
A variety of empirical expressions for modal dispersion have been developed. From practical field experience it
has been found that the bandwidth BM in a link of length L can be expressed to a reasonable approximation by
the empirical relation

where the parameter q ranges between 0.5 and 1, and B0 is the bandwidth of a l-km length of cable. A value of q
= 0.5 indicates that a steady-state modal equilibrium has been reached, whereas q = 1 indicates little mode mixing.
Based on field experience, a reasonable estimate is q = 0.7.
We now need to find the relation between the fiber rise time and the 3dB bandwidth. For this we assume that the
optical power emerging from the fiber has a Gaussian temporal response described by

Where σ is the rms pulse width

Fourier transform of this function is

If t1/2 is the time required for the pulse to reach half of its maximum value, i.e. g(t1/2) = 0.5 g(0) then the value of
t1/2 is obtained from above expression of Gaussian temporal response as

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 6/17
 If we define the time tFWHM as the full width of the pulse at its half-maximum value, then

The 3-dB optical bandwidth B3dB is defined as the modulation frequency f3dB at which the received optical
power has fallen to 0.5 of the zero frequency value.
The relation between the full-width half-maximum rise time tFWHM and the 3-dB optical bandwidth is,

Using for 3-dB optical bandwidth of the fiber link and letting tFWHM be the rise time resulting from
modal dispersion, then

Now total system rise time is

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 7/17
Example: Let us perform Link Power Budget analysis of the link we discussed in Link Power Budget analysis
section. We shall assume that the LED together with its drive circuit has a rise time of 15 ns (ttx). Taking a typical
LED spectral width of 40 nm (𝜎𝜆 ), we have a material-dispersion-related rise-time degradation of 21 ns (tmat,
tGVD) over the 6-km link. Consider receiver bandwidth (Brx) to be 25-MHz. If the fiber has bandwidth-distance
product (BL) of 400-MHz-km and the q parameter of fiber is 0.7, calculate the link rise time.
Rise time analysis:
Here, Transmitter rise time, ttx = 15 ns
material-dispersion-related rise-time degradation, tGVD = 21 ns
LED spectral width, 𝜎𝜆 = 40 nm
Link length, L = 6 Km
bandwidth-distance product, BL = B0 = 400-MHz-km
q parameter, q = 0.7
Now,
Receiver rise time, trx is
350 350
𝑡𝑟𝑥 = = = 14 𝑛𝑠
𝐵𝑟𝑥 25
Rise time due to modal dispersion, tmod is
440 𝐿𝑞 440 ∗ 60.7
𝑡𝑚𝑜𝑑 = = = 1.1 ∗ 3.5 = 3.9
𝐵0 400
Now system rise time is,
2 2 2 2 1/2
𝑡𝑠𝑦𝑠 = (𝑡𝑡𝑟 + 𝑡𝐺𝑉𝐷 + 𝑡𝑚𝑜𝑑 + 𝑡𝑟𝑥 )

𝑡𝑠𝑦𝑠 = [(15 𝑛𝑠)2 + (21 𝑛𝑠)2 + (3.9 𝑛𝑠)2 + (14 𝑛𝑠)2 ]1/2

𝑡𝑠𝑦𝑠 = [225 + 441 + 15.21 + 196]1/2

𝑡𝑠𝑦𝑠 = [877.21]1/2
𝑡𝑠𝑦𝑠 = 29.62 ≈ 30 𝑛𝑠
For 20 Mb/s NRZ data stream maximum allowable rise time degradation is 35-ns. So, rise-time degradation is
below maximum allowable value. Hence, the choice-of components was thus adequate to meet our system design
criteria.
1 10−9
For 20 Mb/s NRZ data stream bit period is 20∗106 𝑠 = = 50 𝑛𝑠.
200

For NRZ maximum allowed transition time is 70% of bit period = 35 ns.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 8/17
Line Coding
In designing an optical fiber link, an important consideration is the format the transmitted optical signal. This is
of importance because, in any practical digital optical fiber data link, the decision circuitry in the receiver must
be able to extract precise timing information from the incoming optical signal. The three main purposes of timing
are:
1. To allow the signal to be sampled by the receiver at the time the signal-to-noise ratio is a maximum
2. To maintain the proper pulse spacing and
3. To indicate the start and end of each timing interval
In addition, since errors resulting from channel noise and distortion mechanisms can occur in signal detection
process, it is desirable for the optical signal to have an inherent error-detecting capability. These features can be
incorporated into the data stream by restructuring (or encoding) the signal. This is done by introducing extra bits
into the raw data stream at the transmitter on a regular and logical basis and extracting them again at the receiver.
Signal encoding uses a set of rules for arranging the signal symbols in a particular pattern. This process is called
channel or line coding.
The discussion here is limited to binary codes, because they are the most widely used electrical codes and also
because they are the most advantageous codes for optical systems.
One of the principal functions of a line code is to introduce redundant bits into the data stream for the purpose of
minimizing errors resulting from the channel interference effects. Depending on the amount of redundancy
introduced, any degree of error-free transmission of digital data can be achieved provided that the data rate that
includes this redundancy is less than the channel capacity. This is a result of the well-known Shannon channel-
condition theory.
The basic types of two-level binary line codes that can be used for optical fiber transmission links are the non-
return-to-zero (NRZ) format, the return-to-zero (RZ) format. In NRZ codes a transmitted data bit occupies a full
bit period. For RZ formats the pulse width is less than a full bit period.

NRZ Codes
A number of different NRZ codes are widely used, and their bandwidths serve as references for all other code
groups. The simplest NRZ code is NRZ-level (or NRZ-L), shown in Fig. For a serial data stream an on-off (or
unipolar) signal represents a 1 by a pulse of current or light filling an entire bit period, whereas for a 0 no pulse
is transmitted. These codes are simple to generate and decode, but they possess no inherent error-monitoring or
correcting capabilities and they have no self-clocking (timing) features.

The minimum bandwidth is needed with NRZ coding, but the average power input to the receiver is dependent
on the data pattern. For example, the high level of received power occurring in a long string of consecutive 1 bits
can result in a baseline wander effect, as shown in Fig. This effect results from the accumulation of pulse tails
that arise from the low-frequency characteristics of the ac-coupling filter in the receiver. If the receiver recovery
to the original threshold is slow after the long string of 1 bits has ended, an error may occur if the next 1 bit has
a low amplitude.
Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 9/17
In addition a long string of NRZ ones or zeros contains no timing information, since there are no level transitions.
Thus, unless the timing clocks in the system are extremely stable, a long string of N identical bits could be
misinterpreted as either N - 1 or N + 1 bits. However, the use of highly stable Clocks increases system costs and
requires a long system startup time to achieve synchronization.
Two common techniques for restricting the longest time interval in which no level transitions occur are the use
of block codes and scrambling.
Scrambling produces a random data pattern by the modulo-2 addition of a known bit sequence with the data
stream. At the receiver to recover same known bit sequence is again modulo-2 added to the received data, and
the original bit sequence is recovered. Although the randomness of scrambled NRZ data ensures an adequate
amount of timing information, the penalty for its use is an increase in the complexity of the NRZ encoding and
decoding circuitry.

RZ Coding
If an adequate bandwidth margin exists, each data bit can be encoded as two optical line code bits. This is the
basis of RZ codes. In these codes, a signal level transition occurs during either some or all of the bit periods to
provide timing information. A variety of RZ code types exists, some of which are shown in the figure below.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 10/17
The base line data are shown in a) in above figure. In the unipolar RZ data a 1 bit is represented by a half-period
of optical pulse that can occur in either the first or second half of the bit period. A 0 is represented by no signal
during the bit period. A disadvantage of the unipolar RZ format is that long string of 0 bit can cause loss of timing
synchronization.
A common data format not having this limitation is the biphase or optical Manchester code as shown in the figure
d) above. Note that this is a unipolar code, which is in contrast to the conventional bipolar Manchester code used
in wire line. The optical Manchester signal is obtained by direct modulo-2 addition of the baseband (NRZ-L)
signal and the clock signal. In this code, there is a transition at the center of each bit interval.
A negative - going transition indicates a 1 bit, whereas a positive-going transition means a 0 bit was sent. The
Manchester code is simple to generate and decode. Since it is an RZ-type code, it requires twice the bandwidth
of an NRZ code. In addition, it has no inherent error-detecting or correcting capability.

Block codes
An efficient category of redundant binary codes is the mBnB block code class. In this class of codes, blocks of
m binary bits are converted to longer blocks of n > m binary bits. These new blocks are then transmitted in NRZ
or RZ format. As a result of the additional redundant bits, the increase in bandwidth using this scheme is given
by the ratio n /m. At the expense of this increased bandwidth, the mBnB block codes provide adequate timing
and error-monitoring information, and they do not have baseline wander problem since long strings of ones and
zeros are eliminated.
For example in 3B4B code 3 input bits are converted into 4 output bits. So there are 8 possible input sequences
and 16 possible output sequences. Among available 16 sequences most appropriate 8 sequences are only used at
output each corresponding to one input sequence.
Consider a simple 1B2B block code in which a 0 may be transmitted as 01, and a 1 as 10. This encoding format
is shown in Figure below and is commonly referred to as biphase or Manchester encoding.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 11/17
With this code there are never more than two consecutive identical symbols, and that two symbols must be
transmitted for one information bit, giving 50% redundancy. Thus twice the transmission bandwidth is required
for the 1B2B code which restricts its use to systems where pulse dispersion is not a limiting factor.
The most suitable codes for high data rates are the 3B4B, 5B6B, and 6B8B codes. If simplicity of the encoder
and decoder circuits is the main criterion, the 3B4B is the most convenient code. The 5B6B code is the most
advantageous if bandwidth reduction is the major concern.

Error correction
For high-speed broadband networks, the data-transmission reliability provided by the network may be lower than
the reliability requested by an application. In this case, the transport protocol of the network must compensate
for the difference in the bit-loss rate. The two basic schemes for improving the reliability are automatic repeat
request (ARQ) and forward error correction (FEC).
Automatic Repeat Request (ARQ)
The ARQ schemes have been used for many years and are widely implemented. As shown in Fig, the technique
uses a feedback channel between the receiver and the transmitter to request message retransmission in case errors
are detected at the receiver. Since each such retransmission adds at least one roundtrip time of latency, ARQ may
not be feasible for applications that require low latency. Among these applications are voice and video services
that involve human interaction, process control, and remote sensing in which data must arrive within a certain
time in order to be useful.

Forward Error Correction (FEC)

Forward error correction avoids the shortcomings of ARQ for high bandwidth optical networks that require low
delays. In FEC techniques, redundant information is transmitted along with the original information. If some of
the original data is lost or received in error, the redundant information is used to reconstruct it. Typically, the
amount of redundant information is small. So the FEC scheme does not use up much additional bandwidth and
thus remains efficient. Depending on the application, some considerations of FEC code properties include the
ability to accommodate self-synchronous scramblers (with characteristics polynomial l + x 43) used in SONET
the 4B5B line code used in FDDI, or the 8B10B line code used in Fibre Channel.
The most popular error-correction codes are cyclic codes. These are designated by the notation (n, m), where n
equals the number of original bits m plus the number of redundant bits. Some examples that have been used
Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 12/17
include a (224, 216) shortened Hamming code, a (I92. 190) Reed-Solomon code, a (255, 239) Reed-Solomon
cod, and (18880, 18865) and (2370, 2358) shortened Hamming codes.

Noise effects on system performance

In the analysis of Link power budget and Rise-time budget we assumed that the optical power falling on the
photodetector is a clearly defined function of time. In reality various interactions between spectral imperfections
in the propagating optical power and the dispersive waveguide give rise to variations in the optical power level
falling on the photodetector.
These variations create receiver output noises and hence give rise to optical power penalties, which are
particularly serious for high-speed links. The main penalties are due to modal noise, wavelength chirp, spectral
broadening induced by optical reflections back into the laser, and mode-partition noise. Modal noise is not present
in single-mode links; however, mode-partition noise, chirping, and reflection noise are critical in these systems.

Modal Noise

1. Mechanical disturbances along the link, such as vibrations, connectors, splices, microbends, and source
or detector coupling, can result in differential mode delay or modal and spatial filtering of the optical
power. This produces temporal fluctuations in the speckle pattern1 at the receiving end, thus creating
modal noise in the receiver.
2. Fluctuations in the frequency of an optical source can also give rise to intermodal delays. A coherent
source forms speckle patterns when its coherence time2 is greater than the intermodal dispersion time δT
within the fiber.
The bit error rate (BER) performance of a digital link with the addition of modal noise to an avalanche-photodiode
receiver system is shown in the figure below

1
A speckle pattern is an intensity pattern produced by the mutual interference of a set of wavefronts.
2
the coherence time is the time over which a propagating wave from laser be considered coherent, meaning that its phase is, on average,
predictable.
Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 13/17
The analysis is for 280 Mb/s at 1200 nm with a gaussian-shaped received pulse. The factor M’ in this figure is
related to the number of speckles falling on the photodetector. For a very large number of speckles (M’= 2910),
the error rate curve is very close to the case when there is no modal noise. As the number of speckles decreases,
the performance degrades. When M’= 50, one needs an additional 1.0 dB of received optical power to maintain
an error rate of 10-6. When M’ = 20, one must have 2.0 dB more power to achieve a 10‘6 BER than in the case
of no modal noise. This number becomes 4.9 dB when M’ = 4.
The problem of modal noise almost exclusively confined to the use of high quality (narrow linewidth, coherent,
stable) laser sources on multimode fibre. For multimode fiber, the light is carried by numerous modes in the core
of the fiber. If a coherent optical source is used, there will be interference between the modes. Although power
is coupling back and forth between the individual, modes, the net sum of all of the modes’ power is constant.
Modal noise can be reduced by:
1. Use LEDs (which are incoherent sources). This totally avoids modal noise.
2. Use a laser which has a large number of longitudinal modes (10 or more). This increases the graininess3
of the speckle pattern, thus reducing intensity fluctuations at mechanical disruptions in the link.
3. Use a fiber with a large numerical aperture, since it supports a large number of modes and hence gives a
greater number of speckles.
4. Use a single-mode fiber, since it only supports one mode and thus has no modal interference.
Mode-Partition Noise
Mode partition noise is associated with intensity fluctuations in the longitudinal modes of a laser diode, that is,
the side modes are not sufficiently suppressed.
This is the dominant noise in single-mode fibers. Intensity fluctuations can occur among the various modes in a
multimode laser even when the total optical output is constant as exhibited in Fig. This power distribution can
vary significantly both within a pulse and from pulse to pulse.

3
a texture of a surface or edge that is not smooth but is irregular and uneven
Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 14/17
The output pattern of a laser diode is highly directional, each of the longitudinal modes that is coupled into the
fiber has a different attenuation and time delay, because each is associated with a slightly different wavelength.
It causes significant variations in signal levels at the receiver in systems with high fiber dispersion.
The signal-to-noise ratio due to mode-partition noise is independent of signal power. Mode-partition noise
becomes more pronounced for higher bit rates.
To prevent the occurrence of a high system bit error rate due to power partitioning among insufficiently
suppressed side modes, one must select lasers carefully.
Chirping

A laser which oscillates in a single longitudinal mode under CW Operation may experience dynamic line
broadening when the injection current is directly modulated. This line broadening is a frequency “chirp”
associated with modulation-induced changes in the carrier density. Laser chirping can lead to significant
dispersion effects for intensity-modulated pulses when the laser emission wavelength is displaced from the zero-
dispersion wavelength of the fiber. This is particularly true in systems operating at 1550 nm, where fiber
dispersion is much greater than at 1300 nm.
One approach to minimize chirp is to increase the bias level of the laser so that the modulation current does not
drive it below threshold. However, this results in a lower extinction ratio (in this case, the ratio of on-state power
to off-state power), which leads to an extinction-ratio power penalty at the receiver because of a reduced signal-
to-background noise ratio.
The best approach to minimize chirp effects is to choose the laser emission wavelength close to the zero-
dispersion wavelength of the fiber.
The figure below shows the effects of chirping at a 5-Gb/s transmission rate in different single mode fiber links.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 15/17
 Reflection Noise
When light travels through a fiber link, some optical power gets reflected at refractive-index discontinuities such
as in splices, couplers, and filters, or at air-to-glass interfaces in connectors. The reflected signals can degrade
both transmitter and receiver performance. In high-speed systems, this reflected power causes optical feedback
which can induce laser instabilities. These instabilities show up as either intensity noise (output power
fluctuations), jitter (pulse distortion), or phase noise in the laser, and they can change its wavelength, linewidth,
and threshold current. Since they reduce the signal-to-noise ratio, these effects cause two types of power penalties
in receiver sensitivities.

The first as shown in the figure below is multiple reflection points set up an interferometric cavity that feeds back
into the laser cavity, thereby converting phase noise into intensity noise.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 16/17
A second effect created by multiple optical paths is the appearance of spurious signals arriving at the receiver
with variable delays, thereby causing intersymbol interference as shown in the figure below.

These effects are signal-dependent, so that increasing the transmitted or received optical power does not improve
the bit error rate performance. Thus one has to find ways to eliminate reflections.
Techniques for reducing optical feedback includes the following:
1. Prepare fiber end faces with a curved surface or an angle relative to the laser emitting facet. This directs
reflected light away from the fiber axis, so it does not reenter the waveguide.
2. Use index-matching oil or gel at air-to-glass interfaces.
3. Use connectors in which the end faces make physical contact (the so-called PC connectors).
4. Use optical isolators within the laser transmitter module.

Class Note, Optical Fiber Communication (Chapter 7), By: Krishna Prasad Phelu Pg. 17/17