Gal Lager

Report So.
6445-T-29
INVESTIGATION OF A YAW DAMPER

FOR AIRCRAFT
by
Edward F. Gallaghiit Robert L. Jungklas

William D. Spiegel
MASSACHUSETTS INSTITUTE TECHNOLOGY

Cambridge 39. Mass.
92
INVESTIGATION OF A YAW DAMPER FOR AIRCRAFT
by
Edward F. Gallagher B.S., U.S. Naval Academy, 1941

Robert L. Jungklas B.S., U.S. Naval Academy, 1942
William D. Spiegel B.S., U.S. Naval Academy, 1941
SUBMITTED IN PARTIAL FULFILLMENT OF THE

REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
at
MASSACHUSETTS INSTITUTE OF TECHNOLOGY

1950
May 19, 1950
Prof. Joseph S. Newell

Secretary of the Faculty
Massachusetts Institute of Technology
Cambridge 39, Massachusetts
Dear Professor Newell:
In accordance with the regulations of the faculty, we hereby

submit a thesis entitled, "Investigation of a Yaw Damper for Air-
craft", in partial fulfillment of the requirements for the degree of
Master of Science (Aeronautical Engineering).

ACKNOWLEDGMENT
The authors wish to express their deep appreciation to Prof. Otto K. Koppen
and to Prof. Robert C. Seamans, Jr. for the interest, encouragement, and as-
sistance given by them in supervising the progress of this work.
The authors also wish to thank Prof. Dominic Amara and Mr. George Coury
of the Fire Control Instrument Laboratory for their technical advice and aid in
setting up experimental equipment.
Thanks are also extended to Mr. Charles Olson for his assistance in the
design of the wind tunnel model.
Valuable advice on wind tunnel testing and techniques was given by Mr.
Eugene E. Larabee of the Department of Aeronautical Engineering.
Thanks are given to Mr. L. E. Payne and his associates of Jackson &
Moreland for their part in the preparation of the text and illustrations.
The graduate work, for which this thesis is a partial requirement, was
performed while the authors were assigned to the Naval Training School,
Massachusetts Institute of Technology. This thesis was prepared under the
auspices of D. I. C. Project 6445, sponsored by the Office of Air Research
and the Armament Laboratory, Engineering Division, Air Materiel Command,
through Contract W33-038ac-13969, Project RR3.
IV
ABSTRACT
An investigation was made of a method for improving the lateral sta-

bility characteristics of an aircraft by means of a Yaw Damper. The Yaw
Damper was a simplified system to the extent that power amplification was
obtained as a result of a servo tab which positioned a rudder. The control
system consisted of a single degree of freedom gyroscope sensitive to yaw-
ing rates, a mechanical linkage to position a tab control surface, and a free
rudder whose position was a function of the tab deflection.
Model tests were conducted in the 4 1/2 by 6 foot wind tunnel of the
Guggenheim Aeronautical Laboratory located at the Massachusetts Insti-
tute of Technology to determine the damping effectiveness of the control
system.
The results of the investigation show that a Yaw Damper of this type
is very effective in damping large magnitude lateral disturbances. For
small amplitudes, the arrangement produces a control system stability
problem in that rudder and tab oscillations occur which are only slightly
damped. These oscillations are considered to be functions of the time lags
associated with unsteady aerodynamic forces and circulation of the air
stream as well as the effects of mechanical backlash and coulomb friction.
Further tests with the rudder fixed, which made the gyro and tab a
direct damping in yaw control, resulted in the elimination of the control
system instability problem for small amplitude lateral disturbances. It
was observed, however, that this stability was obtained at the expense of
the effectiveness of the original control system in damping in yaw.
The tests indicate that a small airplane of the Piper "Clipper" size could
be stabilized laterally by means of a gyroscopically controlled yaw damper to
increase the aircraft spiral stability at all airspeeds. At the same time, the
damping of the Dutch Roll oscillation would be also improved. High performance
aircraft using the same system would be able to replace the tab amplifier with
a servo system to provide an irreversible control system to produce the desired
response specifications.
VI
TABLE OF CONTENTS
Page
ABSTRACT v
Chapter 1 INTRODUCTION 1
Chapter 2 DEVELOPMENT OF THE YAW DAMPER 4
Chapter 3 DESCRIPTION OF THE YAW DAMPER 8
Chapter 4 PRELIMINARY DESIGN OF THE YAW DAMPER 11
Chapter 5 DESIGN OF THE WIND TUNNEL MODEL 17
Chapter 6 WIND TUNNEL TESTS 25
Chapter 7 DISCUSSION OF RESULTS 33
Chapter 8 CONCLUSIONS AND RECOMMENDATIONS 47
Appendix A CALCULATION OF THE ROOTS OF THE AIRPLANE

LATERAL STABILITY EQUATION, CONTROLS FIXED 50
Appendix B PRELIMINARY DESIGN STUDY FOR THE

VERTICAL TAIL OF THE WIND TUNNEL MODEL 55
Appendix C SIMULATION OF THE LATERAL AIRCRAFT

RESPONSE ON THE GENERAL PURPOSE SIMULATOR 71
Appendix D RESPONSE RECORDS 82
REFERENCES 100
vn
PILOT
CONTROLLED
RUDDER
FIG. lb. SCHEMATIC DIAGRAM

OF DIVIDED RUDDER SYSTEM
FIG. la SCHEMATIC DIAGRAM OF YAW DAMPER

Chapter 1
INTRODUCTION
This paper presents a theoretical investigation into a method of improv-

ing the lateral stability characteristics of an aircraft by means of a Yaw
Damper. It is a fact well known among aeronautical engineers that increased
damping in yaw is not only desirable but may even be a requirement in some
airplanes of the future.
A Yaw Damper can be designed so that an airplane that is basically spir-
ally unstable can bemade spirally stable to a high degree, and at the same
time can achieve a marked reduction in the amplitude and period of the Dutch
Roll oscillation. This improvement in lateral stability characteristics results
from the action of the Yaw Damper in effectively increasing the value of the
airplane stability coefficient, C . The result of increasing the value of C
is illustrated in Appendix A, Table DJ.
An aircraft would be considered to have high spiral stability if its spiral
mode would damp to one -half amplitude in one second. It would also possess
above average lateral stability characteristics if its Dutch Roll oscillation
would damp to one-half in one quarter of a cycle. Such an airplane would be
easy to fly and pleasant to ride in. In addition, an aircraft that could meet
these specifications for spiral stability and Dutch Roll, when properly trimmed,
could be flown "hands off even in gusty air. This would be ideal, not only
from the viewpoint of pilot and passenger comfort, but even more important,
from flight safety considerations. For instance, a pilot inexperienced in in-
strument flying who might find himself caught in poor visibility conditions
would have the comfort of knowing that the plane would not inadvertently wind
up in a graveyard spiral if flown "hands off. " In fact, such an aircraft could
be eased to the ground with wings level, simply by retarding the throttle and
maintaining flying speed — without touching the flight controls.
It can be shown that the lateral equations of motion for a conventional
type airplane, controls fixed, can be reduced to a quartic and that the four
roots will consist of a large roll subsidence, a small spiral divergence, and
.
a Dutch Roll oscillation. The spiral mode may be positive at and above cruis-
ing speeds, but even so the time required to damp to one -half amplitude would
be 52 seconds for a B-26C, Medium Bomber. This is practically marginal
stability, as shown in the calculations of Appendix A
Because the spiral instability is generally small and acts slowly, it is
usually corrected by the pilot without his knowledge, but the fact remains
that it has to be corrected by one means or another, and it remains a potential
source of trouble for the pilot. The Dutch Roll oscillation is present to some
degree in all aircraft, as it is a function of wing dihedral and vertical tail area.
It is desirable both from the viewpoint of pilot control and passenger comfort
to keep the amplitude and period of this oscillation as low as possible. From
military considerations it means a more stable gun or bombing platform.
In a flight analysis conducted by the Cornell Aeronautical Laboratory, a
large group of experienced pilots flew an F4U-5 Chance-Vought Corsair
equipped with a yaw damper. The vertical tail surfaces were so modified that
automatic damping in yaw was made proportional to a gain control knob setting
which the pilot could adjust in flight. It was determined by ail the pilots, with-
out exception, that the handling and flying characteristics of the airplane were
best for the maximum setting; that is, the optimum setting was that which gave
the maximum amount of damping.
In addition, Mr. Roland J. White, Aerodynamicist of the Boeing Airplane
Company, has stated' ' after considerable investigation and analysis of flight
test data on the XB-47 that some means must be found to improve the damping
on all future high-performance airplanes because of the adverse effects of
altitude and wing loading on yaw damping.
One method of improving the damping in yaw that is familiar to all aero-
nautical engineers is to increase the size and length of the vertical tail. How-
ever, there is a limit to which this can be carried, because the disturbances
of the airplane due to side gusts will likewise be greater. For smooth flying
qualities and optimum controllability and riding comfort, other means of in-
creasing the damping in yaw should be utilized. For example, a rate gyro
could be installed so as to automatically move a control surface in response
to a disturbance of the aircraft about the yaw axis.
The idea for using a rate gyro to increase the damping in yaw is not new.
It has been used in the XB-47, and in the F4U-5 mentioned above, and it is
known that there have also been other applications of this principle in the field
of automatic aircraft control. The important distinction between this Yaw

Damper and those previously designed is in its simplification.
This investigation is an attempt to simplify the Yaw Damper to the extent
where it will be readily available and usable, not only in the high performance
airplanes of tomorrow, but in the large and small aircraft of today as well.
Reference 4 describes a yaw damper as designed for use in the XB-47 airplane,
and concludes that "other methods should be investigated with the single aim of
simplifying the yaw damper." The unique simplification of the Yaw Damper de-
scribed in this paper is that although the system is fully automatic, and within
itself constitutes a closed -loop servomechanism, it operates without the aid of
an electronic servo amplifier or servo motor. The "muscle" of the system is
the output of a single -degree -of -freedom rate gyroscope, which is mechanically
linked to a servo tab — the output of the gyro and the deflection of the tab being
proportional to the angular velocity of yaw. A rudder tab with an aerodynamic
balance is the power amplifier which positions the rudder. The rudder itself
has an aerodynamic balance to reduce the hinge moments of the system. The
general arrangement of the yaw damper is shown in Fig. la.
The size and configuration of the aircraft in which such a simplified Yaw
Damper would be adequate to produce the required or desired increased damp-
ing in yaw is presently limited by the size of the gyroscopes available. A pre-
liminary investigation indicated that a small airplane of the Piper Clipper con-
figuration could be adequately stabilized with the gyroscope at hand for experi-
mental purposes. It is estimated that a larger airplane could be handled by a
gyro with a larger value of angular momentum (which is a function of the mo-
ment of inertia and the angular velocity of the gyro wheel), or a servo ampli-
fier could be inserted in the closed loop which would adequately increase the
gain of the Yaw Damper so that it would operate according to desired specifi-
cations.
Chapter 2
DEVELOPMENT OF THE YAW DAMPER
The calculations given in Appendix A indicate than an airplane with the

desired lateral stability characteristics (the spiral mode to damp to one -half
amplitude in one second, and the oscillatory mode to damp to one -half ampli-
tude in one-quarter of a cycle) would be realizable if C were increased
r
four times and if C\ alone were in-
were increased four times. If C
*3
n
r
creased sixteen times, the lateral stability characteristics would be even
better than that obtained by increasing Cn four times and C, four times;
r *p
but this is not a practical solution, as it would require a larger than normal
rudder surface.
Flight Condition 3 of Appendix A indicates that increasing the value of
C by a factor of four improves the spiral stability but does not bring it
r
within the specifications listed above. However, the Dutch Roll oscillation is
improved to such a degree that itwas decided to sacrifice some of this im-
provement in the interest of the spiral mode. This was done by increasing
the value of C, by a factor of four, as in Flight Condition 4, Appendix A. In-
asmuch as the value of C, is a function of wing dihedral and the Dutch Roll
oscillation is also a function of wing dihedral, it is observed that the Dutch

Roll oscUlation is slightly greater in Flight Condition 4 than it is in Flight
Condition 3, but is still within the specifications. It is seen that the spiral
stability is now within the desired specification.
One reason that more dihedral has not been used in airplane design is be-
cause of its adverse effect on the Dutch RoU oscillation. Dihedral improves
the spiral stability but, paradoxically, also increases the amplitude of the
Dutch Roll oscUlation, the latter effect generally determining the amount of
dihedral employed in a given design. Now, with the Yaw Damper to increase
the effective value of C and thus the damping, more dihedral can be
employed.
' "
The results of the investigation cited in reference 6 support the conclu-

sion that:
"The maximum tolerable effective dihedral at cruising speed

was indicated to be about 22 degrees. An effective dihedral
as high as 28. 4 degrees did not cause the airplane to exhibit
intolerable stability and control characteristics at landing-
approach speed, . even though the airplane was oscillatory
. .
[sic] unstable. The term 'tolerable' describes a condition

which would not be dangerous in normal fighter operation, but
which is not necessarily 'desirable' or 'pleasant.
"It was the opinion of the pilots that the optimum values of ef-
fective dihedral investigated were 6. 2 degrees (normal air-
plane) for the cruising condition and 14. 2 degrees for the
landing-approach condition. They thought more than normal
amounts of dihedral were desirable in the approach condition
because of the good response in roll to rudder control. It is
noteworthy that this is the direction of the variation of effec-
tive dihedral with lift coefficient for swept -back wings; that is,
increasing lift coefficient results in increasing effective dihedral.
As pointed out above and indicated in Appendix A, the Yaw Damper would
permit even greater values of effective dihedral. It is a fact known to aero-
nautical engineers that anything that improves the damping also improves the
controllability, particularly at the low speeds of landing and take-off.
No further mention will be made of the increase in C, ; it is assumed
that it can be accomplished by increasing the effective dihedral area or the di-
hedral angle, or by sweepback. The problem to be solved is how to increase
the effective value of C four times so that the desired specifications for
lateral stability will be met.

Neglecting the possibility of increasing the vertical tail area by four, or
doubling the length of the vertical tail, either of which would solve the immedi-
ate problem (but would also introduce additional difficulties resulting from high
weathercock stability and increased drag), it is observed that if artificial means
of increasing the damping are used, there are at least two designs available
which incorporate the use of a rate gyro. The rate gyro improves the lateral
stability by automatically moving a control surface in response to an aircraft
disturbance. The first is to connect the gyro to a tab which would move the
whole rudder surface. The second is to connect the gyro to a tab which would
move only part of the rudder surface. The arrangement of connecting the output
of the gyro directly to the rudder (i. e. , no tab) might be feasible if it would
not require too large a gyro to do the work. This increase in gyro size is
due to the loss in power amplification experienced when the tab is not used
to operate the rudder.
The first possibility was rejected because, if the damper were to operate
the whole rudder surface through the deflection of the tab, the rudder pedals
would move as the Yaw Damper operated unless a mechanical differential
functioned so as to keep the motion of the rudder surface from feeding back
to the rudder pedals. Such an arrangement is possible in airplanes equipped
with the control boost systems described in reference 4. The requirement of
having to design a mechanical differential was effectively bypassed by resort-
ing to the remaining practical solution; namely, connecting the gyro output to
a tab which would move only part of the rudder surface.
Inasmuch as C was to be increased only four times, it was decided to
divide the rudder in two, horizontally, and to connect the top half to the pilot's
rudder pedals so as to permit this section to be controlled in the conventional
manner. Figure lb illustrates this configuration. The bottom half is to be
controlled automatically by a rate gyro without the pilot's assistance or know-
ledge. The only information manifested to the pilot on the operation of the yaw
damper is that which is reflected in the improved flying characteristics and
lateral stability of the airplane.
The control surface that the pilot ordinarily would have subject to his
control is reduced by half; however, the controllability of the aircraft has not
been reduced in the same proportion. This is because the gyro -controlled
section aids the pilot, especially in the landing and take-off flight conditions.
However, it is estimated that the area of the vertical tail should be increased
by approximately 20 percent to give the pilot sufficient controllability to side-
slip 15 degrees — a value which is considered to be an adequate controlla-
bility criterion for a conventional aircraft. The reasons for this increase are
furthercommented upon in Appendix B.
Although the Yaw Damper is beneficial at the start of a turn, it does
act adversely during steady turning flight. Because the Yaw Damper acts to
bring the airplane out of a steady turn, the pilot will have to apply more than
normal rudder pedal force in order to make a smooth, steady turn. A method
for yaw compensation during steady turning flight may be desirable if the in-
creased pedal forces are considered objectionable from a "feel" viewpoint.
The actual increase in pedal force in pounds will not be high. It is mostly a
matter of familiarization and the acquisition of a new "feel" through flight ex-
perience. Reference 4 describes two methods of accomplishing yaw compen-
sation in steady turns — methods which apply to electro -hydraulic systems.
It is considered that yaw compensation, while desirable for a mechanical sys-
tem such as described herein, is beyond the scope of this thesis. However,
one simple expedient is mentioned — that of cutting out the adverse effect of
the Yaw Damper in steady turning flight by caging the gyro.
Chapter 3
DESCRIPTION OF THE YAW DAMPER
The Yaw Damper consists essentially of four parts: (1) a gyroscope; (2)
the linkage connecting the output motion of the gyroscope to the tab surface;
(3) the tab surface whose angular deflection is proportional to the precession
of the gyro; (4) the rudder surface whose angular deflection, (within defined
limits), is proportional to the tab deflection. It is planned to restrict the tab
deflection to a maximum of 15 degrees in order to keep the deflection within
the linear range. Figure la is a schematic representation of the Yaw Damper.
The gyroscope is a single -degree -of -freedom gyro in which the preces-
sion is resisted by an elastic restraint; this places the instrument in the cate-
gory of rate gyros. In this type of gyro an angular velocity input causes a
torque to be applied, as output, to the restraint. In the application under dis-
cussion, the elastic restraint is the air load, which creates a moment about
the tab hinge line proportional to the deflection of the tab. The gyroscope is
mounted in such a manner that its spin axis is directed parallel to the X-axis
of the aircraft; its input axis is parallel to the Z-, or Yaw, axis, and its out-
put axis is parallel to the Y-axis of the aircraft. Figure 2 is a schematic dia-
gram of a single -degree -of -freedom gyro, indicating its reference axes and
component parts.
The Yaw Damper system works as follows when the airplane is disturbed
by a side gust that induces an angular velocity about the vertical axis: The
gyro senses this angular velocity and, because of its restraint, produces an
output position angle. According to the gyroscopic law, * the gyro precesses
From page 8-3, ref . 7, "The condensed form of this law relates the applied
torque, M, the angular momentum of the rotor, H, and the absolute processional
— — — — H Xg M
angular velocity of the spin axis, W, byM = WXH, orW = -— . In non- —
vectorial form, this law may he written W = M/H, plus the rule that 'the spin
vector precesses toward the torque vector. If a torque is applied tending to
1
change the direction of the angular momentum vector, the vector will process
about an axis at right (concluded on following page)
en
O
0T
>
o
o
UJ
UJ
or
u.
i
U_
oi
UJ
UJ
or
UJ
oi
UJ
-J
o
z
(/)
CVJ
6
.
about its output axis, which is perpendicular to both the input axis and the spin
axis. The tendency is for the spin axis to precess toward alignment with the
input angular velocity vector. This moves the gyro pickoff link, which is at-
tached to the output axis. This motion in turn, is translated to the rudder tab
which, by deflecting, causes the rudder to deflect so as to oppose the motion
of the airplane. This, of course, reduces the input almost instantaneously,
only a small time lag occurring between the deflection of the tab and the re-
sponse of the rudder to the tab deflection. Figure 3 is a block diagram of the
complete Yaw Damper system.
Mgust „
AIRFRAME a
•
r "1
,v,
or I
MSt ML M6 SDF
RUDDER TAB LINKAGE GYRO
1
YAW DAMPER-
FIG. 3. YAW DAMPER FUNCTIONAL DIAGRAM
(continued) angles to Doth itself and the axis of the applied torque at a rate
proportional to the applied torque and inversely proportional to the angular
momentum. The law of the gyroscope is a reversible one, i.e., a torque input
results in an angular velocity output (precession) and an angular velocity in-
put (forced precession) results in a torque output being applied to whatever
restraint is provided in the ins trument *
10
Chapter 4
PRELIMINARY DESIGN OF THE YAW DAMPER
The airplane lateral stability criteria as developed in Appendix A, indi-

cated that an increase in the aircraft damping in yaw was desirable. To ac-
complish this, a method based on the output torque characteristics of a
single -degree -of -freedom rate gyroscope was selected.
In this design, the gyroscope is oriented to receive rate of yaw of the air-
craft as an input; the gyroscope torque output then acts to produce a rudder
position in the proper direction to oppose and dampen the yawing rate of the
airplane. The rudder is positioned by means of a tab which is operated di-
rectly by the gyroscope output, and is effectively a power amplifier between
the gyro and the rudder.
Without adequate information concerning the characteristics of such a
power system, was necessary to make a preliminary investigation into the
it
practicability of the design. A primary factor of interest was the friction

level in the system. For a reasonable friction level, the remaining torque
from the gyro had to be sufficient to position the rudder tab against the oppos-
ing aerodynamic hinge moment of the tab when operating in a moving air
stream.
In order to simplify the tests, the tab hinge moment characteristics were
assumed to be sufficiently linear to permit simulation by a spring restraint.
Since the rate gyro requires a spring restraint, this restriction on the tab
hinge moment parameter permitted the tab to perform this function directly.
Further restrictions on the maximum angular displacements of the tab

and rudder permitted the additional assumption that rudder motion would not
affect tab position during operation. With this assumption, the tab position
would be a function of the gyro output only, and a direct calibration could then
be made for gyro output moments. For a symmetrical system no angular re-
strictions are necessary up to the stall region.
In the preliminary tests, an estimate of aerodynamic tab hinge moments
was made in order to select the size of the gyroscope required in an actual
11
aircraft installation. From the law of the gyroscope, as developed in refer-
ence 7 and other sources, M . = H x W = HW cosAg. By proper selection
of the gyro-to-tab linkage ratio, the angular displacement of the gyro was
limited to 15 degrees, a value which would introduce a maximum output torque
reduction of only 4 percent if the cosine term were to be neglected.
For an estimated tab hinge moment of 1. 5 inch-pound, and an angular

rate of yaw of 1. 5 degrees/sec for full tab angular displacement, application
of the gyroscope law shows that a gyro H value of 64. 5 x 10
dyne-cm-sec
was required. The largest gyro available had a specification of 72 x 10
dyne-cm-sec and was selected on the premise that the friction level and un-
balance and uncertainty torques would not prevent the realization of the esti-
mated torque requirements. It was found that the physical size was also prac-
tical for possible installation in small aircraft of the Piper "Clipper" type for
later testing.
Since the preliminary setup was to be for the full-size system, efforts
were made to simulate closely the conditions of operation. A rudder trim tab
and a self -aligning bearing for a Navy SNJ type aircraft were obtained and
formed, with the gyro, the basic units in the mockup. No attempt was made
to reduce hinge moments in the tab; however, the self -aligning bearing was
cleaned and lubricated with a light oil, on the predication that such care would
be required on the actual installation for best operation of the system. The
previous assumption of non-interaction between rudder and tab positions en-
abled the mockup to be shortened by attaching the gyro linkage directly to the
tab, thus bypassing a free -link rudder support. This permitted the entire
system to be mounted on a board of convenient length for making turntable
tests. Figures 4, 5 and 6 are photographs of the mockup used in the tests.
From initial static tests, the friction torque level measured at the tab was
found to be 0'. 0394 x 10 dyne -cm. A calibration run using balance weights
was made at 26. 18 mr/sec for the gyro output and the tab input moments.
From this data, the friction torque was found to be 0. 039 x 10 dyne-cm, or
7 percent of the gyro output torque. With no spring restraint, the system was
sensitive to a minimum table rate of 0. 4091 milliradian/sec, indicating a
Coulomb Friction level of 0.029 x 10 dyne -cm. A slightly higher friction
level due to the rudder free -link support is to be expected, but the total is not
12
(
CO
H
CO
w
H
w
m
&
H
K
O
Q
w
<
H
CO
<
Q
<
H
o
<
<
«
w
o
o
»—
13
w
u
w
o
u
CO
o
O
O
O
CO
<
H
W
Q
14
FIG. 6. DETAIL SHOWING TRIM TAB, AUTOSYN
AND SPRING RESTRAINT
15
expected to exceed 15 percent. For this value of friction, the gyro selected
appeared to be of a reasonable size to carry out the design specification for
the tab positioning function.
The spring constant was calculated to be 0. 082 x 10 dyne -cm/degree for
the spring used in these tests. No attempt was made to reproduce an aero-
dynamic spring constant accurately, since a reasonable correlation of the
mockup data to wind tunnel test data could be accomplished by applying a con-
stant multiplying factor.
For a final check, the system was placed on a sinusoidal input table and
the frequency at which oscillations began was found to be 1. 6 cycles/sec.
Based on a normal design criteria of 0. 5 cycle/sec as an upper limit for air-
craft undamped natural frequencies, it is not anticipated that the yaw damper
unit natural frequency will create a stability problem due to vibration coupling.
In all the mockup tests a 60-cycle vibrator was exercised to simulate the
aircraft airborne vibrations. It thus tended to reduce all friction levels to a
minimum, particularly the Coulomb Friction effects.
On the basis of the test information obtained as described above, the feas-
ability of the system mockup was accepted and the construction of a wind tunnel
model incorporating this system was begun. The scale model was built to per-
mit wind tunnel testing of the Yaw Damper performance under simulated flight
conditions.
16
Chapter 5
DESIGN OF THE WIND TUNNEL MODEL
Theoretical analysis established the benefits that may be derived by in-

creasing the damping coefficient in yaw, C by a factor of four. Prelimi-
,
r
nary design studies indicated that this increase might be achieved through the
use of a rate gyro.
Under actual flight conditions the amplitude and phase relationships be-
tween the airplane motion and the rudder motion become functions of many
variables, such as the amount of inertia of the gyro linkage system (which
consists of the moment of inertia of the gyro about its output axis, the mo-
ment of inertia of the tab about its hinge line, and the inertia of the linkage
system), the moment of inertia of the rudder, the forcing frequencies of the
aircraft due to gusts, the aerodynamic configuration of the tab and rudder,
the degree of static and dynamic balance of the tab and rudder, size of the air
gap between the rudder and the vertical fin and between the tab and the
rudder, the slack or lost motion in the linkage system, and friction in the
system. It is conceivable that with all of these variables amplification,
rather than damping, could result under certain conditions of flight. In
an attempt to evaluate and select the variables that would produce an
optimum system, the investigation was carried into the wind tunnel.
Two gyros were available for this investigation. The first gyro G(l) had
an angular momentum of 72 x 10 dyne -centimeter -seconds. The second
gyro G(2) had an angular momentum of 0.88 x 10 dyne -centimeter -seconds.
By definition, angular momentum H is equal to the product I
SDin
wSDm •
For both gyros, Wg . equals 12,000 rpm. Preliminary design studies in-
dicated that the first gyro, G(l), was of sufficient size to produce the de-
sired increase in yaw damping in a light aircraft. Installation and test of
G(l) in a typical light airplane suggests an interesting extension to this
investigation.
17
It was decided to install G(2) in the wind tunnel model. Since the two
gyros were the controlling scale factor in deciding the model size, and since
the aerodynamic moments vary as the cube of the airplane dimensions, the
ratio of lengths between the prototype with G(l) installed, and the model with
G(2) installed became;
Scale Factor =
72 * 1q6 = ±™
6 *
.88 x 10
Assuming the fuselage length of the prototype to be 22 feet, the length of the
model became 5.06 feet. An overall length of 5.0 feet was actually used for
the model.
The wind tunnel model, mounted in the horizontal plane,was constrained
about all axes except the vertical axis. Since the only alteration in the air-
craft being considered in increasing the damping was in the vertical tail con-
figuration, themodel merely consisted of a fuselage and a vertical tail.
The fuselage was constructed of balsawood (Fig. 7), and proportioned to ac-
comodate the gyro, its associated linkage, and the angle -measuring instruments.
In other words, the fuselage was designed as a streamlined box to carry the
vertical tail and the necessary measuring instruments. Slab sides with
curved edges, and as many straight-line panels as possible, were used to de-
crease model cost and construction time. The proportions of the fuselage are
similar to those of the B-26. The fuselage was dowelled together with pine
wood inserts to facilitate the use of wood screws in joining the parts. Hard-
wood inserts were used at all of the pressure points to add sturdiness to the
model.
The fuselage was designed in four parts: the streamlined solid nose; the
underbody extending from the nose back to the tail; the forward top section
extending half way back from the nose to the tail; and the after top section (Fig. 8)
carrying the vertical tail, the gyro, and the associated linkage. This sub-
division was made so that access to the inside of the fuselage would be simple,
and so that installation and tests of the components could be carried out in
the laboratory prior to assembly. The gyro and linkage were secured to the
18
K
W
s
<
Q
<
fa
o
CO
H
CO
w
H
W
p
H
Q
55
K
O
fa
w
Q
O
s
w
<
u
CO
O
fa
19
I
CO
d
<
K
2
K
W
o
CO
o
H
U
W
CO
w
m
<
>
o
w
K
fa
o
i
W
CO
O
O
00
d
I—
fa
20
,
after top section containing the vertical tail, tab, and rudder. With this ar-
rangement the tests for friction, and gyro output versus angular rate, were
made comparatively easy.
The fuselage was supported by a two -inch vertical steel pipe fixed to the
floor of the tunnel. This support was mounted on an aluminum plate that was
bolted to the tunnel floor. Two tension cables helped to steady the support
during the actual tunnel tests. The support being hollow, a means was af-
forded for streaming the leads to the gyro and to the recording instruments.
The fuselage was supported at a point one -third of the way back from the
nose. This point was considered to be the center of gravity of the airplane.
The model was provided with a means for varying the moment of inertia,
I ., by varying the distance of a lead weight from the pivot point.

The vertical tail was made of balsawood having an NACA 0012 cross sec-
tion. Figure 9 shows the size and important dimensions of the tail. The
tail height, h -, was controlled by the selection of the aspect ratio, (AR) .
and the vertical tail area, S .. An (AR) . of 1.8 was chosen for two reasons,
the first being that the model differed from the normal geometric config-
uration in that no horizontal tail, which would act as an end plate and increase
the (AR)
ff
, was present on the model. This endplate effect increases the
(AR)
eff
by a factor of 1.55 (Ref. NACA TN 755). The second reason is that
there is a tendency in modern aircraft design to increase the (AR) ., within'
structural limits, thereby increasing the slope of the lift curve of the verti-
cal tail, and the tail effectiveness.
Since the wind tunnel tests are to be run with the idea of determining the
increase in yaw damping through the introduction of the gyro -controlled sur-
face, the design of the vertical tail was solid, simulating a stick-fixed flight
condition, except for the gyro -controlled portion. As has been previously
explained, the tail on an actual plane would consist of a vertical fin and two
rudders, one controlled by the pilot and the other controlled by the gyro -tab
combination. These two rudders would be of approximately the same size.
The pilot-controlled portion should be designed to allow the same pilot con-
trol of the aircraft as is required in present-day specifications. It was
21
1 1
1 1 1 1 1 1 1 '
1 1 1 1 1
INCHES
1 2 3 4 5 6 7 8 9 10 11 12
FIG 9 DETAILED DESIGN OF THE VERTICAL TAIL FOR THE WIND TUNNEL MODEL.
22
estimated that the gyro -controlled portion of the rudder would aid the pilot as
much as 20 percent in slow flight while correcting for yaw, such as in a land-
ing or take-off run. The gyro -controlled rudder, being more sensitive to rates
than the pilot, should greatly enhance the flight characteristics of the plane. In
steady turns the pilot would have to overcome the forces resulting from the con-
trolled -rudder output.
The vertical tail area, S t was selected to be 12 percent of the area of
,
the wing of the aircraft. Since the ratio for small planes rarely exceeds 10
percent, this increase in design area was an attempt to insure the pilot-
controlled section remaining as effective as is required by lateral control
specifications.
The gyro -controlled tab was made of pine because of the small size of the
tab and because of the magnitude of forces from the link and hinge that are
imposed on the tab. Several tabs were constructed and tests for system
stability were made for each tab. The tab was actually supported by ball
bearings fitted into aluminum end plates that were carried by the rudder.
See Fig. 12 for tab design and table of modifications.
The rudder was made of balsawood and hollowed out to accomodate the
linkage system. The actual construction of the rudder can best be understood
by studying Fig. 1 a.
The design configuration for the tab and the rudder resulted from a study
of systems that would produce small hinge moments per degree of rudder tab
effectiveness. See Appendix B. This led to an aerodynamically balanced rudder
and an under -balanced tab. The hinge positions, chord lengths and areas are
shown in Fig. 9. The leading edge of both the rudder and the tab were made
elliptical. It was felt that sharpening the nose might be necessary, but this
modification would be comparatively easy to accomplish in a wooden control
surface.
The rudder was statically balanced about its hinge line as
nearly as possible so that it would be insensitive to inertia effects as a result
of angular accelerations of the fuselage about the vertical, Z-axis. Means for
statically balancing the rudder were provided by inserting lead weights in the
rudder nose in a void provided for the weights. The linkage was constrained
in the XZ plane so that its inertia effects could also be neglected.
23
The angular position of the fuselage with respect to the center line of the
tunnel was measured with an autosyn, an instrument whose voltage output is a
linear function of angular displacement. The case of the autosyn was fixed
relative to the support; the rotor shaft moved with the fuselage as indicated in
Fig. 1 a. Rudder angle was measured by an autosyn that acted as the bottom
hinge for the rudder. Tab angle, being a function of the gyro gimbal position
about the output axis, was measured at the gyro. Voltages were picked off by
a three -channel recording instrument that photographed these angular meas-
urements on a permanent trace on the same time scale.
It was an important design consideration to keep the system as sym-
metrical as possible so that rudder motion would not effect tab deflection.
This was accomplished by running the linkage from the gyro along the fuselage
axis of symmetry, to the rudder hinge point. The installation of the gyro
linkage arrangement can be seen in Fig. 1 a. The gyro output axis was mounted
horizontally along the Y-axis of the airplane so that angular rates would result
in torque outputs directly proportional to this angular rate. The torque output
about the output axis of the gyro was transmitted back to the tab by the linkage
arrangement. A discontinuous rudder hinge was necessary so that the linkage
could be symmetrical.
The gyro with was balanced about the gyro output axis
the links attached
to insure that the forces on the tab were functions only of angular rate of yaw
and not of angular position of the output gimbal. The linkage was made ad-
justable so that the tab and gyro could be properly centered. Means were
provided whereby the ratio of tab deflection to gyro gimbal rotation could be
varied. A linkage ratio was chosen so that a 7.5-degree gyro gimbal angle
produced a maximum tab deflection of 15 degrees. In order to reduce weight
and friction in the system to the minimum, magnesium and aluminum links
and fittings were used with miniature ball bearings at all of the moving joints.
The air gap between rudder and vertical fin, and between rudder and tab
was made less than 0.025 inch by filling up the gap with masking tape until
the desirable clearance was obtained.
Control strings from either side of the tail to observers on either side of
the tunnel test section provided means for introducing yaw angles and for stop-
ping undesirable oscillations during the wind tunnel tests.
24
Chapter 6
WIND TUNNEL TESTS
In order to measure the effect of a control surface actuated by a rate

gyro in increasing the damping in yaw, model tests were conducted in the
small wind tunnel located in the Guggenheim Aeronautical Laboratory at the
Massachusetts Institute of Technology, Cambridge, Massachusetts. Figure
10 shows the layout and general dimensions of this wind tunnel.
The 4. 35:1 scale model was instrumented and mounted on a pedestal
which was located on the tunnel center line. Figure 11 illustrates the gen-
eral arrangement for the test runs. A preliminary test was made to observe
the air flow around the model. Small strings were attached to the tail sec-
tion to indicate any boundary layer separations. No separation was observed
at any speed for yaw angles up to 20 degrees, at which point the vertical tail
had reached the stalled condition.
It was observed that the tunnel air flow was not smooth and that the model
oscillated in response to the turbulence of the tunnel flow. This turbulence
or "noise" was a random effect and had high frequency harmonics. A more
detailed analysis of noise characteristics is beyond the scope of this thesis
but the effects on the model in the tunnel was qualitatively measured at air
speeds from 30 to 80 miles per hour at intervals of 10 miles per hour. Figure
12 is a plot of the amplitude and period of the over -all noise data and indi-
cates that these effects increase with air speed. Brush recordings of tunnel
noise are included in Appendix D.
The next step was to check the stability of the control surfaces. It was
immediately apparent that the rudder, with tab fixed, was unstable. For this
original configuration, shown in Fig. 9, C, was positive and C, negative.
a 6
The rudder system was made stable by increasing the tab chord dimension by
0.4 inch and this new tab configuration installed as indicated in Fig. 13. This
tab will hereafter be referred to as the original tab.
25
• «
S>
as
j ifif
• • n o u
eo 0*h a 3.
a to n
u ate - •
S a rt en
o *•* •
1
o >h UN •
1
>
«e
h
S< r> «o
"
"
• r c U
*JW ii
( . -u . . CI
a, -cm «h *m
© PC TJ M
O 00 3 • M « 3
W> » ICO 0> O
*»
^ *>*
1C•« M
mmSq
«0
« o
O OlOtO -
••H^ COW • JC-H
«o »^
3* 5
o.**
£ * • <-<
0= ©
• •
B
On n M
•
03 B
e*H « oo • u> «
lO O^HflN q •-« H
(--« «0Q m n
*»iX «H >i IH7
i
• *» la c eg
o o C o o o
0- offlfc. O C4K
\ /
/
/
/
1
1 l
i
t
-® ^1
> o £
«i v.
5
5€S I 3
>*
i H- J
>
I
i
> * u *~
6
i
\ **:
c t
Q O
VI \
li!
\
/- — J O "In
3« \
1
^ ti
-©- 1/
ii
1/
l> t
N
/ \ 7
-<i»
\ 1
/ \ '/ V / \
V \
4e 1
1 i
f °
w.e. -• —
.PS
-f0 1
S<o
° «
'
"a
/ \
^
\
k
^ i-
P.
^5>
- H -
t
t
3*
V^V
'
1'9- &
* »' ?' *i
26
UJ
o
o
_i
UJ
z
z
I-
o
z
or
o
o
Z>
o
5
o
z
<
g
I
z
UJ
S
z>
a:
27
SQN0D3S Nl Q0IU3d 3SI0N ' 3SI0N^
o h-
J r- O
X
< J
/
U_
•*• <
4 o
<£>
0)
<
\ UJ
\ cr
3 o
O 2 .
s o
m X ,
-J UJ
Q
\ \ ,
cr
UJ
Q-
NNE
SPE
cn
UJ
^ o:
l- —
UJ
</>
o S
_l
u.
<
o <*•
O _|
z
1- z w
UJ ^
.
o
UJ ITUD TUN
UJ
O (/> _i ,
rO
cr
< 20£
0-
V _i
<
\ z
Ul
Q
z Z
\N 3
O (-
(M
<
Q
\ O
or
\ UJ
\ Q.
\ C\J
\.
CVJ
S33«93a ni 3DNV"ingyni T3NNn±

ONIM 01 30Q lN3W30VldSlQ avinoNv wnwixvw , xvw
28
2-1 2-2 2-3
TAB CONFIGURATIONS
SCALE
-r
3.5"
RUDDER CONFIGURATIONS
FIG. 13. MODIFICATION OF VERTICAL TAIL
29
With this stable rudder arrangement, the rudder free and tab free system
was next investigated. This combination was found to be unstable also. The
thickness of the vertical tail was built up slightly to simulate the effect of
sharpening the rudder nose and a slight improvement was noted. The rudder
nose section was sharpened by shaving the balsa wood contours but no appre-
ciable improvement was obtained.
On the basis of the negligible improvements observed, the aerodynamic
balance of the rudder was reduced by removing the original nose section and
replacing it with an aluminum nose cap. The new configuration is shown as
number 2 in Fig. 13. At this point the rudder was only partially balanced
statically by inserting lead weights in the nose section. Due to the small
volume space within the nose cap and the reduced moment arm of the center
of gravity of the nose section, a satisfactory static balance could not be ob-
tained. Subsequent tests showed this configuration to be stable under both tab
fixed and tab free conditions and no flutter tendency under the tunnel noise
conditions was observed.
At the same time, a viscous damper was being installed to reduce the
sensitivity of the controller linkage system to the high frequency tunnel noise
excitation. This addition was found to improve the stability of the system and
the viscous damper was therefore retained as a portion of this controller ar-
rangement. Upon completion of these two alterations, this system configura-
tion was selected as the basic yaw damper for testing purposes.
All data was available as output voltages while the aerodynamic yaw re-
sponse and control surface positions were based on angular measurements.
It was therefore necessary to obtain the calibration data for the autosyn volt-
ages and the position angles. The permanent records were made on a Heiland
photographic recorder actuated directly by the autosyn voltages and the re-
sponse voltages were adjusted until all recording traces were within the trac-
ing range for the maximum angular values previously outlined. All three
traces were adjusted to the same reference center line for the calibrations
and the rudder trace was then offset slightly to permit ready identification of
the various traces. All data records have been included in Appendix D.
The moment of inertia of the model about the yaw axis was obtained from
oscillations with a known spring restraint attached to the tail section. This
2
was calculated to be 30. 8 x 10 gm-cm .
30
Upon completion of the instrumentation calibrations, the model was
tested for control effectiveness factors. A tunnel airspeed of 60 miles per
hour was selected for this data due to the velocity of 120 miles per hour which
it represented for the full scale model, a value near the cruising flight con-
dition. Data was obtained to determine static curves of <!* vs. 6
p , + vs. 6
T ,
and 6 vs. 6 In addition, a record was made for fixed tab angles and free
rudder and yaw. The respective control effectiveness curves are shown in
Fig. 14.
Having completed all the basic work, the actual model tests were then
begun. For comparison purposes, three tunnel speeds were selected for
these runs. These speeds, 40-60-80 mph compare with full scale aircraft
speeds of 80-120-160 mph. For light aircraft, these speeds will correspond
to the landing approach, normal cruise, and fast cruise flight conditions. It
is felt that this broad range of test speeds also provides a more comprehen-
sive evaluation of the yaw damper control system.
Test data was obtained for the model yaw response from an initial yaw
displacement of 15° for the following conditions:
1. Controls Fixed. No yaw damper operating.
2. Controls Free. The yaw damper is operating. Unless
otherwise noted, all controls free tests are made with
the gyro running.
3. Direct Control. This corresponds to controls free with
the damping rudder fixed at zero deflection. The tab
operates directly as the damping control.
The yaw damper system in condition 2 was found to oscillate in the zero
yaw angle position. A temporary mass balance was added
external rudder
but did not improve the oscillation tendency. A trailing edge extension was
added to the control tab to increase the tab hinge moment; this was not found
to improve the situation. For the present control configuration, the oscilla-
tion appeared to be a function of the control linkage and the time lags in the
system.
Further work to improve this tab and rudder control system was suspended
and an amended approach initiated in which the tab control surface was operated
directly by the gyro. For this approach, the rudder control surface was fixed
in the zero deflection angle. Yaw response data was taken for the several tab
configurations shown in Fig. 13. Basically, these tabs have two values of nose
balance with two extensions added to the original trailing edge.
31
This direct control system was found to perform satisfactorily at the zero
yaw angle position. As a final test, the tab system was operated in the balanced
configuration with the half -inch extension and without the viscous damper in
order to permit a comparison for the effect damper in the system.
of the
At this point all tunnel tests were concluded. The effects of changes in
the model moment of inertia and control system linkage ratio, while allowed
for in the design of the model and anticipated as a part of the tunnel work, were
not investigated due to the large volume of test data to be analyzed at this time.
15 15
10
/~~——
f =
5
°( ) 5 10 15 5 10 15
8R , RUDDER ANGLE IN DEGREES TAB ANGLE IN DEGREES
10
6.6*
20
5.5
5 15
o r FREE
o (v
<r
10
IT
«o
1
10 li 5
5 10 15 20
8T , TAB ANGLE IN DEGREES 8 Tl TAB ANGLE IN DEGREES
FIG. 14. CONTROL EFFECTIVENESS CURVES, SYSTEM 2-1-1.
32
Chapter 7
DISCUSSION OF RESULTS
In carrying out the investigation of a simplified yaw damping system for

an airplane, it soon became apparent that the theoretical approach could not
provide sufficient design information for an operating system. For this rea-
son, the theoretical design was reproduced in a scale model and the system
tested in the wind tunnel.
For testing purposes, the model was restricted to a fuselage and a verti-
cal tail arrangement with a single degree of freedom in yaw only. In consider-
ing the motion of the model in the tunnel, the equations below describe the
transient yaw response to a step input of yaw:
V + cn l
. i + cn 4, =0
qSb \ t\>
c. •
c
n
X—
<\, <|, . _ Controls Fixed
<\> + 1\> + -J- <\>
<\> + ¥— 4; + i- <\> = Yaw Damper Operating

C C
V + 2(DR)u) + w 2 =0
n i
<\,
n
where: DR = damping ratio u> = undamped natural angular velocity
C = -Jt— u =
n-\
\ ^
qSb
C
DR =
(2C)u>
*_ C„.
n
<\>
= ( —V
2
)C„n
r
n
[With negative sign of C and C . taken into account. ]
33
With the Yaw Damper operating, the equations of motion become more
complex than the simple second -order equation which describes the fixed-
control case. This is because of the introduction of the non-linear damping
term described in Appendix B. Since was desired to compare the two sys-
it
tems, the response equation is assumed to remain a second-order function,

and the primed coefficients reflect this effect.
In order to analyze the photographic data obtained on the Heiland re-
cordings, all Yaw Damper tests were standardized as indicated in Chapter 6.
The data has been correlated for three air speeds. This information is shown
in the plots of Figs. 16 to 27, with the operating condition and configuration as
indicated by Table I for each run. Specific configuration relationships may be
obtained by referring to Figs. 9 and 13.
All plots have been presented in a non-dimensional form. The undamped
natural period for the controls -fixed response at each airspeed was chosen as
the non-dimensionalization time factor for all data at that airspeed. In this
manner, the variations in undamped natural frequency due to changes in air-
speed have been removed. In addition, the yaw response ordinate has been
non-dimensionalized by taking the ratio of this angle to the initial yaw angle.
Figures 16, 17, and 18 show the comparative results for the three test
conditions. All the test airspeeds have been shown, but, except for minor
variations in the damping ratio, the general response data is similar for each
of these speeds. The remaining figures show the response results obtained
for the direct-control system with the tab extension dimension as the variable.
Any other individual comparisons between the several tab configurations are
readily obtained by overlaying the desired response curves. Figure 15 is a
non-dimensional plot of the response of a second-order system to a step in-
The response curves of Figs. 16 to 27
put as a function of damping ratio.
inclusive may be compared with a second -order system by overlaying a spe-
cific response curve on Fig. 15.
From these tests, the Yaw Damper installation was found to greatly in-
crease the effective damping of the model response from the initial yaw angle
of 15 degrees. The response plots show that the model returns to a zero yaw
angle position with no overshoot. However, at this position the control sys-
tem oscillated and did not possess the smooth performance characteristics
34
desired. Under similar conditions in an actual aircraft, this installation
would not contribute to pilot or passenger comfort and might eventually de-
stroy itself.
In order to improve the system characteristics, the vertical tail fin thick-
ness was increased and a small viscous damper added to the control linkage.
Both effects acted to smooth the system, but reduced the effective damping
sufficiently to cause a slight overshoot to occur.
The damper rudder was fixed at a zero deflection angle, and the tab then
operated directly as a control surface. This direct control with viscous
damper arrangement produced the smoothest operating system, but reduced
the effective damping still further and increased the magnitude of the over-
shoot.
Because of the large initial yaw angle, response rates were sufficiently
high to cause the tab to operate in the limit stops. In most cases, the yawing
rate after the first peak overshoot was small enough to operate the tab within
the 15-degree limit stops. For this reason, a measure of the direct control
system effectiveness is difficult to obtain.
The comparative effectiveness of the various control conditions and con-

figurations is shown in Table n. This table lists the time, T, in seconds
necessary to damp the fuselage yaw angle to ten percent of its initial ampli-
tude at each of the test airspeeds. A response time to damp to five percent
initial amplitude is normally used for comparison purposes. Because of the
small scale factors in the later portions of the response curves, the level has
been increased to ten percent. An approximate method of conversion to this
five percent basis is to estimate this level by an exponential decay from the
ten percent position.
A time ratio T/T-. . is arbitrarily selected as a figure of merit, where
Tfixed is the time for the controls -fixed response to damp to ten percent.
For evaluation, the smaller numbers correspond to the more effective system.
In order to remove the variations due to airspeed, a similar time ratio based
on TN (the undamped natural period of the controls -fixed response) is also
shown. Both ratios are included in the data of Table II.
Based on these figures of merit, it may be seen that the tab-rudder-gyro

system is the most effective yaw damping control. Because of the instability
35
that arises near the zero yaw angle, this system at present is not feasible for
an actual installation.
The direct tab control system is a simple and workable system. It was
found that the configuration 2-2 is the most effective of all of the tabs con-
sidered.
Based on real time considerations, the direct tab control effectiveness
appears to increase with airspeed. The tab rudder effectiveness decreases
with airspeed. For all tab configurations, the control effectiveness changed
through a range of only 30 percent.
The small viscous damper installed in the control system noticeably in-
creased the effectiveness of the control.
As a representative conventional B-26C has formed the basis

airplane, the
of the theoretical approach in the investigation of the yaw damper design herein
described. Comparative data for this aircraft and the wind tunnel model are
shown in Table in.
,DR = 2.0
08
DR =1.5
0.6
DR = 10
«£ °*
02
DR =0.7^
DR = 0.5-
O
=> -02
Q-
«= -0.4
06
0.5 10 1.5
NON-DIMENSIONAL TIME - t/TN
FIG.-15
NON-DIMENSIONAL AMPLITUDE RATIO ft/fi VS NON-DIMENSIONAL TIME RATIO t/T N
FOR VARIOUS VALUES OF DAMPING RATIO DR OF A SECOND ORDER SYSTEM WHEN
THE FORCING FUNCTION IS A STEP INPUT.
36
TABLE I
MODEL CONFIGURATIONS FOR NONDIMENSIONAL YAW RESPONSE CURVES
Number Control Speed Rudder Tab Viscous Damper
1 Fixed 80 2 1-1 In
2 Free 80 2 1-1 In
3 Direct 80 2 1-1 In
4 Fixed 60 2 1-1 In
5 Free 60 2 1-1 In
7 Fixed 40 2 1-1 In
8 Free 40 2 1-1 In
10 Direct 80 2 1-2 In
11 Direct 80 2 1-3 In
12 Direct 60 2 1-2 In
13 Direct 60 2 1-3 In
14 Direct 40 2 1-2 In
15 Direct 40 2 1-3 In
16 Direct 80 2 2-1 In
17 Direct 80 2 2-2 In
18 Direct 80 2 2-3 In
19 Direct 60 2 2-1 In
20 Direct 60 2 2-2 In
21 Direct 60 2 2-3 In
22 Direct 40 2 2-1 In
23 Direct 40 2 2-2 In
24 Direct 40 2 2-3 In
25 Direct 80 2 2-2 Out
A Fixed 80 1 1-1 Out

B Fixed 60 1 1-1 Out
C Fixed 40 1 1-1 Out
Dt Free 80 2 1-1 In
Et Free 60 2 1-1 In
Ft Free 40 2 1-1 In
G Direct 80 1-1 In
H Direct 60 1-1 In
I Direct 40 1-1 In
Jt* Free 50 1-1 In
K* Free 50 1-1 In
L* Free 40 1-1 In
Mt* Free 40 1-1 In
N Free 40 1-1 Out
Ot Free 40 1-1 Out
P Free 40 1-1 In
t Gyro Not Running

* Spoiler on Vertical Tail
37
Table n
CONTROL EFFECTIVENESS COMPARISON
Configuration
Viscous Damper in xunnel Airspeed in mph
Condition Rudder Tab 40 60 80
Controls Fixed 2 1-1

T 4.62 4.05 3.04
T /T fixed 1.0 1.0 1.0
T/T N 1.96 2.62 2.58
Controls Free 2 1-1
T 0.5 .76 .58
T / T fixed .108 .1875 .191
T/T N .212 .49 .492
Direct Control Fixed 1-1
T 1.5 1.16 .67
T/ T fixed .324 .286 .22
T/T N .635 .748 .568
T 1.66 .88 .9
T /T fixed .36 .217 .296
T/T N .704 .568 .763
T 1.62 1.16 .88
T / T fixed .351 .286 .29
T/T N .686 .749 .746
38
Table II (cont'd)
Configuration
Viscous Damper in Tunnel Airspeed in mph
Condition Rudder Tab 40 60 80
T 1.55 .98 .73
T/T fixed .335 .242 .24
T/T N .657 .633 .619
T 1.47 1.06 .91
T//T .318 .262 .299

fixed
T/T N .623 .684 .771
T 1.4 1.26 .92
T / T fixed .303 .311 .302
T/T N .594 .813 .78
No Viscous Damper
Installed
T 2.2 1.35 1.12
T / T fixed .476 .333 .368
T/T N .932 .871 .949
TN (sec) 2.36 1.55 1.18
where
T = Time to damp to 10 percent amplitude in seconds.
TN = Undamped natural period for controls fixed.
39
Table III
COMPARISON OF WIND TUNNEL MODEL AND B-26C DATA
MODEL B-26C
Speed in mph 60 265
Altitude in feet 10,000
u) in radians per
c second 4.05 3.95
n
DR 0.15 0.149
C -.0694 -.0735
C n .(
4>
—
b
) -0.123 -0.106
CL 0.3* 0.3
Corresponds to airplane of Piper "Clipper" type at 120 mph, from which

scale model was derived.
40
1.0
0.8
TUNNEL AIRSPEED = 80 MPH

0.6
Tn = 1.18 SEC
-£ 04 v.
0.2
2
^
3
. \
-0.2
•0.4
-0.6
N.
-0.8
-1.0
0.5 1.0 1.5
FIG.-16
1.0
0.8

0.6
TN = 1.55 SEC
<£ 04
02 /
<
EC
Ld
Q
Z)
-02
6
X
^4
-0.6
-0.8
-i.O
0.5 1.0 1.5
FIG. -17
NON-DIMENSIONAL PLOT OF WIND TUNNEL MODEL YAW RESPONSE TO A STEP INPUT.
41
1.0

6
Tn =£36 SEC
^ 04
1
8
1
9
02
-0.4
-0.6
7
-0.8
-10
0.5 1.0 1.5
NON-DIMENSIONAL TIME - t/Tu
FIG.- 18
10

TN = 1.18 SEC
04
11
02 /
O V 10 10 /
11
-0.2
^3
-04
^3
-06
-0.8
-1.0
0.5 1.0 15
NON-DIMENSIONAL TIME- t/T(M
FIG.- 19
42
1.0
0.8

0.6
TN = 1.55 SEC
0.4
Q2
3
\/ 2 /
13
U
o
3 -02
'
12 %
6
-0.4
-06
-08
-10,
Q5 10 15
FIG- 20
10
08

06
TN =2.36 SEC
^ 04
02
15 14
I
15 14
-0.2
-04
^9
-0.6
-Q8
-1.0
0.5 1.0 1.5
NON-DIMENSIONAL TIME - t/T|\j
FIG.- 21
43
1.0

U.6
TN = 1.18 SEC
0.4
0.2
o 18
1-
<
cr
17 18
UJ
Q 16 17
13 0.2
_l
Q. 16
5 4
<
06
o.e
0.5 1.0 1.5
NON- DIMENSIONAL TIME - t/TN
FIG. -22

06
TN = 1.55 SEC
O
02
o 19 21
20 21
1
02 20 "b 20
19-^
-0.4
-0.6
08
0.5 10 15
FIG- 23
44
J..U
0.8

0.6
Tn =2.36 SEC
o
22 24
23 24
0.2 ^1 /
/
23
22
0.4
0.6
0.8
0.5 1.0 15
FIG -24
10
Q8
06
TN = 1.18 SEC
O
04
0.2
17
-0.2
-0.4
25
-0.6
-0.8
-1.0
0.5 1.0 1.5
FIG.- 25
45
l.O
08

0.6
Tn = 1.55 SEC
o 20 26 \
1 o |i
J—
UJ
3 -yj.ir
I
\-
_l
0.
26
<
-0.6
-08
-10
0.5 LO L5
FIG- 26
1U
08

06
TN =2.36 SEC
O
04
02
g
i-
<
or
23
,
27^
04
06
08
10
<) 5 1 1 5
NON-DIMENSIONAL TIME- t/TN
FIG- 27
46
Chapter 8
CONCLUSIONS AND RECOMMENDATIONS
The object of this investigation was to develop a method for improving

the lateral stability characteristics of an aircraft by means of a Yaw Damper.
The results of an analytical study made in Appendix A, and a simulator
study reported in Appendix C, illustrate the marked improvement in the
lateral response of an aircraft as a result of increased damping in yaw. Not
only has the spiral stability of the aircraft been greatly increased, but the
Dutch Roll oscillation has been highly damped. On the basis of these two
studies, it is concluded that increased damping in yaw is desirable.
The results of the wind tunnel model tests indicate that increased damp-
ing in yaw can be obtained from a control surface moving in response to a
disturbing yaw rate input so as to oppose the motion initiated by the disturb-
ance. The amount of damping obtained is a function of the Yaw Damper con-
trol system configuration. Power amplification can be obtained by causing a
large surface to move as the result of the deflection of a smaller surface. In
the case of a single surface, the amount of damping obtained is a function of
the aerodynamic balance and the area of that surface. This means that the
amount of damping is directly related to the hinge moment and the aerody-
namic efficiency of the surface being moved.
An important conclusion based on the results of the wind tunnel test is
that one surface moving in response to the disturbing input can provide in-
creased damping, and this direct control configuration results in a simple,
smooth -working system. A single control surface system is simple to de-
sign and operates smoothly because the complexities associated with moving
a second surface as a function of the position of the first have been eliminated.
The tab-rudder control system gives the maximum damping but also pro-
duces an unsatisfactory oscillation at the zero yaw position in response to the
tunnel noise or turbulent air. Further investigation and study of the unsteady
forces associated with circulation lags (which result in a time delay between
47
the tab position and the rudder response, and between rudder position and
fuselage response) could possibly bring about the development of a smoothly
working, non -oscillating, tab-rudder system that would provide more damp-
ing than that obtainable from a direct control system.
It is concluded that the direct control tab configuration which gave the
most effective damping could, when linked to a gyro having an H value of
72 x 10 dyne -centimeter -seconds, adequately stabilize a light aircraft of the
Piper "Clipper" class so that it would be spirally stable at all airspeeds and
would have a highly damped Dutch Roll oscillation.
Higher performance aircraft could also use the same system, provided
the hinge moments were of the same order of magnitude. Low hinge moment
is a function of aerodynamic balance and surface design. If a low hinge
moment could not be obtained through design, a direct -control mechanical
system would require a gyro with a larger value of angular momentum, H.
One way to increase H is to increase the angular velocity of spin. If the size
of the gyro were bounded by dimensions or weight, a servo system could be
installed to provide an irreversible control with the desired amplification so
as to meet design specifications.
recommended that the possibility of making the Yaw Damper operate
It is
smoothly at the zero yaw angular position be further investigated. The follow-
ing steps are suggested:
1. The linkage ratio should be varied to determine its effect

on the smoothing of the system.
2. The effect of the viscous damper in the control linkage

system should be further examined.
3. Time lags in the surface control positions and the aero-

dynamic forces which form, should be investigated to
determine their effects on the system design.
4. The effect of friction and mechanical backlash in the con-

trol linkage should be studied as these effects will con-
tribute control lags that may be destabilizing the system.
For further wind tunnel investigations it is recommended that the tests

be made with smaller yawing rates and larger scale factors. Such a procedure
would permit a more accurate evaluation of the effectiveness of the various
control systems in damping in yaw.
48
Finally, on the basis of the results of the wind tunnel tests, it is recom-
mended that a direct -control system similar to the 2-2 tab configuration with
the 72 x 10 dyne -centimeter -seconds gyro be installed in an aircraft and that

the system be given a complete and thorough flight test.
49
APPENDIX A
CALCULATION OF THE ROOTS OF THE AIRPLANE LATERAL

STABILITY EQUATION, CONTROLS FIXED
Lateral equations for the airplane motion are given

in reference (1) These equations and notations will
.
be used with slight modifications. It can be shown

that for an aircraft flying at cruising speed and
above, the product of inertia, I xz , has but a small
influence on the lateral stability.
At speeds below cruising the terms involving this

product have a greater effect on the lateral stabi-
lity. However, if these terms are neglected, the re-
sults will be more conservative than if they are in-
cluded. The calculations of reference (2) indicate
that the inclination of the principal longitudinal
axis above the flight path causes a stabilizing shift
in the oscillatory-stability boundary. Therefore, for
reasons of simplicity the product of inertia I xz and
terms involving this product are neglected^ and the
following equations of motion result:
(C - 2d) 3 - 2dy + C L<p =
^ P
6 + _£
2
dy +
2
d - Jx d <p = (1)
%0 +
rCr
Jzd d\p + S
2
P dtp =
where
Jx " 2
L
—
b J
<>z
- ^ [i
L b-J
These equations may be written in the determinant

form as:
c -2 CL
v„ - 2)l
C Cl
l
P X - J \
"% 2 2
x =0 (2)
C
"rr -
U
S 2
J 7X
2
50
Expanding this determinant, a quartic in X is ob-
tained:
4 3 2
AX + BX + CX + DX + E =0 (3)
where
A = 1
B + C /J + C
lp/J x )
- -ft (C
yp nr z
C =
^XJ Z
<%% - VV * %.
Jz
+ _
Jx J,
D
^z Vl, VrJ
- ~ — Vlo " -^ ^V - C l Cn>
r
U^L C C C Cl >
<
lo n r - nQ X
4J XJ Z ^ "P r
The coefficients of the quartic in X given in eq. (3)

above are constants depending for their values on the
inertia terms, the airplane relative density para-
meter n, and the partial, or stability, derivatives.
The Values of these derivatives are listed in Table I,
The coefficients of the quartic in X are listed in
Table II.
The quartic stability equation, (3) , is readily solv-

ed by use of the approximate factorization formula:
E R D
(Xi + B)(x 2 +p-)d|,4 * + y) = o (4)
^5 ,
the roots being:
Xj = -B
R R2 D
^3 .*
2B D
2
4B 4 D 2 B
where R = BCD - D 2 - B 2 E = Routh's Discriminant.
51
From the roots of equation (4) , it is possible to cal-
culate the period, the time to damp to one-half ampli-
tude, the number of cycles to damp to one-half ampli-
tude, and the damping ratio, DR, by use of the fol-
lowing equations:
^1,2 = <*1,2
\ Sj4 = a + 3i
.
logp 2
T 1// 2 = — -Z — seconds
<x/t
_ —2nt— seconds/cycle
P =
(3/t
N l/2 =^- cycle

.11
DR =
The value of the roots Xj. 2 3 * the time to damp to

one-half amplitude T 1/2 , and the period of the Dutch
Roll oscillation, are tabulated in Table III for the
four flight conditions. The root \ ± indicates the ra-
pid roll subsidence. The root \ 2 indicates spiral sub-
sidence, which in most airplanes is generally a small
spiral divergence. This root gets more positive as the
speed is decreased. The roots \ 3 A represent the com-
plex Dutch Roll oscillation.
As a check on the approximate factorization method of

determining the roots of the quartic, the exact roots
have been calculated for flight condition (4) by syn-
thetic division methods and found to be:
\ ± = -4.97
\2 = -0.495
\B = -0.77 _
+ 1.441
r ,
The exact roots indicate that the plane is even more

stable than the approximate factorization methods in-
dicate, both from the spiral stability and the Dutch
Roll considerations. This is shown in Table III under
flight condition (4) .
52
TABLE A -I
AIRPLANE STABILITY DERIVATIVES
(D* (2) (3) (4)

Small Small Airplane Small Airplane
B-26 Airplane IAS = 120 mph IAS = 120 mph
Flight IAS = 256mph IAS = 120 mph Sea Level Sea Level
Condition 10,000 Ft Sea Level C n increased 4x C n & C 10 increased 4x
ip
r r
C 0.3 0.3 0.3 0.3

L
b(ft) 70.22 32 32 32
H 13.05 3.12** 3.12** 3.12**
t (sec) 2.02 0.567 0.567 0.567
J 0.0313 0.0546 0.0546 0.0546

X
J 0.0598 0.1048 0.1048 0.1048

z
c -0.720 -0.720 -0.720 -0.720

»
% -0.096 -0.096 -0.096 -0.384
C 0.0735 0.0735 0.0735 0.0735

*
0.055 0.055 0.055 0.055
r
C -0.106 -0.106 - 0.424 -0.424

n
r
C -0.505 -0.505 -0.505 -0.505

l
P
C -0.0168 -0.0168 -0.0168 -0.0168
n
P
* From ref. (3) ** From ref. (5)
53
TABLE A-H
AIRPLANE STABILITY COEFFICIENTS
Flight
(1) (2) (3) (4)
Condition
A 1 1 1 1
B 9.31 5.49 7.01 7.01
C 26.53 6.31 13.54 13.54
D 144.17 12.13 14.56 18.35
E 3.93 0.253 1.51 6.54

R 14520 281 1096 1084
TABLE A -m
AIRPLANE RESPONSE
(4)
Flight Approx. Exact
Condition (1) (2) (3) Roots Roots
RoU \
1
= -B -9.31 -5.49 -7.01 -7.01 -4.97
StabUity
T (sec) .15 .0717 .0562 .0562 .079
l/2
„
2
d
.-JL -.0272 -.0208 -.104 -.356 -.495
Spiral
D
StabUity
T (sec) 51.5 18.6 3.78 1.09 .795
l/2
X -,58±3.88i -.381 1.44i -.77+ 1.23i -.6± 1.51 .77±1.44i

3,4
T (sec) 2.42 1.035 .511 .647 .511
/2
Dutch P (sec) 3.27 2.48 2.91 2.36 2.48
RoU
Oscillation N (cycle) .738 .417 .176 .275 .206
l/2
DR, damping .149 .264 .625 .400 .534
ratio
54
Appendix B
PRELIMINARY DESIGN STUDY FOR THE VERTICAL TAIL OF THE

WIND TUNNEL MODEL
In designing the vertical tail for incorporation of the proposed yaw damper,
it was necessary to investigate the left effects of various combinations of geo-
metric airfoil surfaces so as to find the combination that would produce the
greatest lift per unit of control torque. The control torque was to be produced
directly from the gyro linkage system. The aerodynamic lift was to be ob-
tained through the use of a tab -rudder -fin combination. As has been previously
explained, the tab was to be controlled by the gyro, and act as an aerodynamic
amplifier in moving the rudder. The change in the airfoil camber produced by
the deflection of the tab and rudder was for the purpose of increasing the lift
in such a direction as to always oppose the angular yaw motion of the aircraft.
The design study was therefore reduced primarily to the selection of a
tab-rudder -vertical tail configuration that would produce the maximum lift per
unit of tab hinge moment.
Since the vertical tail design was to include both a pilot-controlled rudder
and a gyro -controlled rudder, it was decided to make the two rudder areas
approximately equal. In an attempt to give the pilot the same amount of con-
trollability that he now has in most conventional aircraft, (i. e. , rudder de-
flections capable of holding a 15° sideslip angle) the area of the vertical tail
was increased by a factor of about 20 percent. It is apparent that with this in-
crease of vertical tail area, the weather-vane tendencies of the plane have
been increased. Since the pilot -controlled rudder has not been increased in
the same proportion, it is admitted that the pilot could not hold the same
steady-state sideslip angle that he could hold with a tail of conventional de-
sign. The controllability of the airplane for this dual rudder configuration is
a subject for further analysis. The pilot -controlled section was fixed at
6
R = throughout the remainder of this investigation.
55
The tab hinge moment is the sum of four effects; that is,
3Ctt ^Cxr ^îi

ilrp JClm
C„ =
TIfti
i- 6
l
+ &p + — a + C„
n
^T 3 6
t d6
R da T(friction)
From Ref . 14 we find that, by using a balanced rudder and a partially

balanced tab, the tab hinge moment coefficient can be expected to be in the
order of magnitude of:
CH = 0.01/degree
where
3C W
tlrp
— = - 0. 009/ degree
3&
T
3C„
rim
i- = - 0. 002/degree
36
R
3C„
Xlrp
— = - 0.001/degree
da
For
6 degrees
R(max) - !5
6 degrees
T(max) = 15
5 de ^ rees
"(max) =
C„ = -0.17
T( max) (aero)
These hinge moment coefficients are based on the area and chord of the con-
trol tab.
The actual tab hinge moment for the wind tunnel model tab, based on the
dimensions shown in Fig. 9 and a q corresponding to 60 miles per hour,
becomes:
56
MT 002378 x (iL)
1
= C„
n ST
l
qC Tl = -0.17X •
x . 75 x5.25
(max)(aero) T 2 12
= 0. 043 in-lbs.
For the steady -state condition in which the aerodynamic moments on the
tab are balanced by the control moment from the gyro linkage system:
£MT = o = M T(aero) + M T(gyronnk)
The gyro output torque can be determined as follows:
M (gyro) " W (Z axis)

H (gyro wheel)
or
(gyro) ~ W (Z axis)
H (gyro wheel)
cos ( A g)
Assuming perpendicularity between M, W, and H, since the angle A is re-

stricted to plus or minus 7. 5 degrees because of the setting of the gyro stops,
we may write this equation in the form:
M (gyro) = W(Z axis) H (gyro wheel)
When the gyro is installed in the model, the gyro output torque and the
control moment at the tab are related by the following sensitivity:
M (gyro link) " M (gyro) X S

(link)
where
S " °' 482
(link)
The damping in yaw of an aircraft consists of three component parts; that
due to the wing, that due to the fuselage, and that due to the vertical tail. Ex-
pressing this in terms of the damping coefficient in yaw, we have:
C = C + C + C
r r(wing) r(fuselage) r(vertical tail)
where the yawing moment coefficient due to the vertical tail is simply
57
;
r Ldo; j b
(yT) b(w)
From the equation it can be seen that the damping can be increased by
lengthening the distance from the center of gravity of the aircraft to the center
of pressure of the vertical tail, by increasing the tail area, or by increasing
the slope of the lift curve of the vertical tail. The yaw damper changes the
lift of the vertical tail by changing the camber of the lift surface, thereby in-
creasing the effective angle of attack and the slope of the lift curve.
The increase in damping due to the yaw damper is proportional to the in-
crease in the damping force, that is:
dC L n dC L
AM (damp) a
*« '
qS (VT) X (VT)
^ a eff L do- ]
(VT) (VT)
Since the tail of the wind-tunnel model is designed so that only one -half
of the tail is controlled by the yaw damper, it becomes apparent that the damp-
ing will be made up of two parts:
M,(damp) . v = [C + Cn 1 qSb
n
r (gyrocont) r(pilot cont)
If we are to increase the damping coefficient four times, the merits of

which were discussed in Chapter 2, it means that the damping force must be
increased four times. Since the areas of the two controlled rudders are the
same, and since the damping force on the pilot -controlled rudder remains the
same as for the fixed -control case, it follows that the damping force on the
gyro -controlled rudder must be increased seven times, i.e.
Cn = 7C n
r(gyro cont) r (pilot cont)
Now, if the gyro were infinitely sensitive to angular rate of yaw, and
powerful enough to overcome the friction inherent in the linkage system, it
58
would cause the tab to swing to its designed stops as the fuselage sensed an
infinitesimal yaw rate. The tab would deflect the rudder, and the fin-rudder-
tab combination would produce an effective angle of attack that is a function
of angular rate of yaw. If this effective angle of attack is seven times as
great as the angle of attack experienced by the fixed control surface due to
'
the same angular rate, (i. e. ,
VT we have effectively increased our damp-
)
VA
ing coefficient by a factor of four. It can be shown that the damping coefficient
will be a non -linear function having its greatest values for small angular rates
of yaw and rapidly decreasing values for increasing angular rates. As will be
seen later, the aerodynamic lift coefficients are also non-linear, especially
for large angular deflections of the control surfaces. The fact that the damp-
ing is non -linearmakes the system more difficult to analyze, but it does not
detract in any way from the merit of the damping system. This analysis indi-
cates that the damper will be most efficient at low angular rates of yaw and
that the damper efficiency reduces to zero as the angular rate increases with-
out limits. (See Fig. B-l below.)
r, YAWING VELOCITY IN RAD/SEC
FIG. B-l. QUALITATIVE CURVES OF EFFECTIVE ANGLE OF ATTACK

AS A FUNCTION OF YAW RATE.
59
In the gyro -controlled portion of the vertical tail it was desired to deter-
mine the minimum control force for the maximum lift of a fin-rudder -tab
combination. It has been found that the optimum design in obtaining a given
amount of lift results from keeping the control surface within the linear range
of operation. The lift effectiveness after stall is always less than the values
obtained just before stall. The maximum control surface deflection of the
rudder was the first thing to be determined in the design problem.
Figure B-2 from reference (14), Fig. 3, shows the maximum rudder de-
flection as a function of c p/ c yrp and a.
/
Since a =
MVT)
—
VA
'-
— r
, we may de-
sign for an «/
max\ of two degrees which corresponds approximately to an
angular rate of 1 radian per second for V. = 60 mph. Entering Fig. B-2, we
see that a fifteen -degree rudder deflection is well within the stall limit for a
chord ratio of Cp/cyrp of 0. 33 and an angle of attack, a, of 2 degrees. This
a will always be positive for the purposes of using Fig. B-2, since the rudder
will always act in such a manner as to move into the relative wind. Fifteen
degrees maximum rudder deflection is considered a conservative estimate
since Refs. (9), (10), and (11) indicate that in most cases rudder stall is de-
layed when the tab moves in a direction opposite to that of the rudder. It is
to be noted that the tab deflection limits and the rudder deflection limits de-
pend upon the aspect ratio of the vertical tail, the plan form, and the scale
size. No attempt was made to correct for these variables. Figure B-2 is
based on information obtained from an NACA 0009airfoiL; an NACA 0012 air-

foil was used on the wind-tunnel model tail, no correction being made for the
difference.
The amount of tab deflection necessary to produce fifteen degrees of rud-
der deflection (or the most effective rudder angle) under the dynamic conditions
of flight was so difficult to estimate that the tab link was made adjustable so
that the maximum angular throw of the tab could be controlled. It is to be
noted also that the relative angular deflection of tab and rudder is a function
of aerodynamic balance on the two control surfaces.
It can be shown that aerodynamic balance on the control surfaces can
greatly reduce the control hinge moment. The more closely the controls are
balanced, the more apt they are to become unstable. For example, if the
60
rudder is overbalanced to the extent that C. is a positive number, the aero-
dynamic forces no longer act as a spring restraint, but act in the same
direction as the control moment. This would result in rudder lock, and a non-
stable control system. Figure B-3 reproduced from Fig. 144-fo of Ref. (15),
is indicative of the merits derived from using various values of aerodynamic
balance, i.e. Cj/c Cj/c selected to equal 0.3, and c p/ c
,
R . If
R is
vr *°
equal 0. 3, values of C = -0. 006 and C = -0. 0045 can be taken di-
h h
Rt6 R„,
a
R
rectly from the figure.
Figure B-4, reproduced from Fig. 106, Ref. (10), shows that by choosing
aerodynamic parameters of CrVcp equal to 0. 3 and Cp/ c vT etl ua l to 0. 26 for a
blunt -nosed overhang with a 0. 001 -inch gap, we might expect a small negative
value for C, , and a lift coefficient of 1. for an angle of attack of 5 degrees
and a 6p = 15°. It is to be noted that this curve is plotted for an NACA 66-009
airfoil with infinite aspect ratio. No attempt was made to reduce this data to a
finite aspect ratio for the NACA 0012 airfoil. Figure B-4 shows that for rud-
der angles greater than ten degrees, the lift effectiveness of the airfoil falls
off rapidly and the hinge moment rises rapidly. In an attempt to delay this
stalled condition, the nose of the rudder and tab of the model was made ellip-
tical. The combination of the balsawood rudder and pine tab was designed so
as to lend itself to easy modification if wind tunnel tests proved it desirable to
sharpen the nose.
Figure B-5 from Fig. 143 -b of Ref. (15) is an indication of the airfoil lift
coefficient C, and the lift effectiveness parameter a < as a function of bal-

6
Vr
ance. Here, again, the blunt -nosed configuration appears to be the most ef-
ficient.
The optimum must be determined from lift

size and balance of the tab
considerations and tab hinge moments, as well as from rudder hinge moments,
since the gyro linkage output must produce the desired tab hinge moments. In
a sense, the gyro output forces can be compared to the control tab forces of
Fig.B-6 from Fig. 148 of Ref. (15). From this figure it can be seen that the
optimum size of the balancing tab is about 20 percent of c R The lift effec- .
tiveness of the resulting combination is about two -thirds that of the rudder with-
out the tab.
61
.
References (11), (12), and (13) indicate that, for a constant tab chord,
using a large span tab at a small deflection angle is superior to using a small
span tab at a large deflection angle. The tab on the wind tunnel model was de-
signed for full span (i.e. , controlled tab and controlled rudder had the same
height)
Reference (15) indicates that for a given amount of overhang, the blunt-
nosed control surface gives the greatest amount of balance. When the nose of
any balanced control surface unports itself because of the deflection of the
surface, regardless of the nose shape, large increases of hinge moment and
decreases in lift result. Based on these considerations, a balanced rudder
with a blunt elliptical nose was used. The unporting angle of 25 degrees was
much greater than the designed 15 -degree maximum rudder deflection. The
tab nose was made similar to the rudder nose.
It can be seen that the tail design resulted from a study of various com-
binations of tab and rudder configurations as found in NACA reports. Repre-
sentative curves showing some of the most important parameter ratios as
functions of the lift efficien cy have been included in this report. The final
model configuration chosen as a result of this study has the following geometric
dimension ratios:
C c
R = 0.30
b
p
= 0.27
C C
VT R
C
= 0.20 —^_
c
c
b = 0.31
R T
The actual tail design can be seen in Fig. 9.
62
Flap chordJairfoil chord cf /c
,
.20 .40 .60 .80 1.00

1 1 1 1 1 1 1 1 ' 1
*0 = -12°
-
-10°
-
-e^
H
- -6°
ST
-
.0
-£
<J
-10
- /
//
0°
a - 1 1
-15- 1
-
V "C
\
6°
-20
8°
-
o
3: 10°
-
12°
-25--
14°
-
Figure 3.- Approximate maxima allowable flap deflections

for linear limits of airfoil characteristics
at various angles of attack. Data for HAOA
0009 airfoil with infinite aspect ratio and
at an effective Reynolds number of 3,410,000.
FIG.B-2
63
FIG.B-3
64
0.30 c,
FIG. B-4
65
FIG. B-5
66
FIG. B-6
67
CI TMhnlwl lot* to. TM n*.i
si cc
A "^f
s ?
*v $
sj
? Tt
jt *tlr
-it! 3
35
5
^^
fea
!-•'
!
^
8
lf
•
li^P
%x%
^ jirs
£
^
?*
*8
!
\
.>B"i
-iSi
ill -/
-yZ
i a \%
FE
^• ;
\ III t 1 Ga
b
[
v .fiy ./ vi^
V
S -t j
VSc- -!
-
-$-
it
%
\
T-\t
\t\
V
?
A S
\\
-
X—f Jl
r
&->
*£-*
t cQ
k$xv-
;lN'Tr
-8
fe
|Ni
6
1
!
>^v
It \
—J
l-^V
^âT- ,Z
7
V
/~7
2-
/
??
^ i
S§3 3S
^ 7T"
If
'
=**
%?7 ~ '
-t"
"+-
1^
I
CD
O
L. •
Si.
^1 —
-I?.
N- -^
7
~fc ni. t S "I
~-
v- T 32-
it
2
1
\£
S 8.
r
k.
-.5
zf
-F
a » B 5 5 5
X± it tl >S3
ii ti J-t O C o
t\- <\ Jw
* 2
e 4* i
XX Si-
44
1.
j-i
It
i£<S ^ ^
-^7 7 '
^
in
a / -
s^
1
yJT
^>'
/
zi
~
S.
i
^^
^-^^-^
"£
i
I
^z'---'' "
% *S S *S 8 8 8
Jl
<k
68
SYMBOLS
C, Airfoil section lift coefficient
C, Hinge moment coefficient of control surface
C, Flap section hinge moment coefficient (in this investigation

f C, and C. are synonymous).
C, Rudder section hinge moment coefficient

r
C, Rudder hinge moment coefficient
n
R
S
w Wing area
S Tab area
T
S Rudder area
R
q Dynamic pressure of free air
c Chord of airfoil section
c Chord of vertical tail*

VT
c
R Chord of rudder*
c Chord of tab*
T
c* Chord of flap*
a Angle of attack of finite span wing (angle between zero lift

line and relative wind).
tfff Effective angle of attack
a Angle of attack for infinite aspect ratio
a. Flap or rudder effectiveness
<\> Angle of yaw

r, <\> Angular rate of yaw
69
<\> Angular acceleration of yaw
6 Control surface deflection
6™ Tab deflection for model investigated
6pj Rudder deflection for model investigated
b Span of surface
A Aspect ratio
c. Chord of overhang
+
cl rifLi
a *-
3 a J
6
ac
c nh
L J
a da
3 a -U6
+
ch
6
[—]
L
6 a3
J
M Moment
H Angular momentum
V. Aircraft velocity
p Air density
S Linkage ratio
link
VT Vertical tail
s Control force, or stick force
t} Tail efficiency
ly Length from center of gravity of aircraft to center of

T pressure of vertical tail
A Angle between spin reference axis and gyro spin axis
NOTE:
* Measured from hinge axis to trailing edge of airfoil. + Subscripts out-
side of parentheses around the partial derivatives indicate the variables held
constant when the derivatives are taken. A straight line over a symbol means
the symbol is a vector quantity.
70
Appendix C
SIMULATION OF THE LATERAL AIRCRAFT RESPONSE

ON THE GENERAL PURPOSE SIMULATOR
In order to more clearly illustrate the effect on the aircraft lateral re-
sponse of increasing C and C, by a factor of four, a simulator study using
n *p
r
the equations of motion was made. The General Purpose Simulator of the
Instrumentation Laboratory, Massachusetts Institute of Technology, was used

for this study. The assistance of Mr. Samuel Giser and Mr. Frank Spada in
setting up the problem is gratefully acknowledged.
The aircraft lateral equations of motion when written in the dimensional
form appear as follows:
[Y -D]p -(JL) 9+ D* = YA (1)

P VA 6 r
LB + [LD - D 2 ]<p + [LD + -52. D 2 ]* = -L (2)

P p l
X
N p +
2
[N.D + J52_ D ]cp + [ND -I Z
^
D 2 ]4, = -N - N (3)
fl
p p t
°r °a
These equations may be rewritten:
D,P = KjD^ - K 2 6 r - K3 q> + K4 P (4)
D% = K 5 Dc|, + K6 D 2 (|, + K? 6a + Kg D 9 - Kg p (5)
D2 (|, = K 1Q 6 a + Kn 6r + K 12 p + K 13 D<p + K^D 2 ? + K^Dc^ (6)
71
For the B-26C airplane at 265 mph IAS, at a density altitude of 10,000
feet, the dimensional coefficients have the following values:
COEFFICIENT CONSTANT VALUE UNITS

- Kl 1
1
Y* K2 .0444 sec"
r
1
g/vA *3 .0709 sec"
1
Y K4 -.1772 sec"
P
1
Lr K5 .4320 sec"
.121
'xz^x *e
2
L *7 -6.901 sec"
'a
1
L K8 -3.961 sec"
P
2
L K9 10.757 sec"
e
% K 10 .2043 sec'
2
%
N
K ll
K 12
-4.110
3.881
sec"
sec'
2
2
e
1
n K 13 -.06852 sec"
p
'xz'-'z
K 14 .064
1
Nr K 15 -.4327 sec"
The method of simulation will not be discussed in this appendix. Suffice

to say that the equationswere simulated in such a manner that the constants
Kg and K 1& could be increased by a factor of four. In addition, constants Kg
72
and K 14 , which represent the products of inertia, could be set equal to zero.
The simulated response of the B-26C airplane for a rudder pulse input
and an aileron pulse input were photographed from the trace appearing on a
cathode ray oscilloscope. Response pictures were taken for Sideslip, p ;
Roll, 9 and Yaw,

, These photographs are described as they appear on the
<|» .
following pages.
Fig. C-l. Pulse Input = 1/4 sec
Note:
1. Same pulse applied to aileron and rudder.
2. Timing dots are 2 seconds apart.
73
Comments:
Oscillation is Dutch Roll.
DR = 0.15
Sideslip Response
Roll Response
Yaw Response
Fig. C-2 Responseof B-26C to a Rudder Pulse Input

Normal Configuration
74
Comments:
Neglecting Products of Inertia,
I has only a small effect.
xz
Sideslip Response
Roll Response
Yaw Response
Fig. C-3. Response of B-26C to a Rudder Pulse Input
Normal Configuration with *xz = ^
75
Comments:
Increasing C by a factor of
r
four has damped Dutch Roll

Oscillation
Sideslip Response
Roll Response
Note improvement in spiral mode.
Yaw Response
Fig. C-4. Response of B-26C to a Rudder Pulse Input with I =

xz
and Yaw Damping Increased by a Factor of Four
76
Comments:
Increasing C, by a factor of
P
four has increased Dutch Roll
Oscillation
Sideslip Response
Roll Response
Yaw Response
Fig. C-5. Response of B-26C to Rudder Pulse Input with I

xz =
and Dihedral Increased by a Factor of Four
77
Comments:
Increasing C and C, by a
factor of four has damped Dutch

Roll and improved Spiral Stability
Sideslip Response
Roll Response
Yaw Response
Fig. C-6. Response of B-26C to a Rudder Pulse Input with l^

z
=
Dihedral and Yaw Damping Increased by a Factor of Four
78
Comments:
Normal configuration DR = 0. 15
Sideslip Response
Roll Response
Yaw Response
Fig. C-7. Response of B-26C to Aileron Pulse Input with I

xz =
79
Comments:
Increasing C has damped
Dutch Roll. Spiral Stability

is improved.
Sideslip Response
Roll Response
Yaw Response
Fig. C-8. Response B-26C to Aileron Pulse Input with

of I
xz =
and Yaw Damping Increased by a Factor of Four
80
Comments:
Increasing C_ and C, by a
n x
r p
factor of four results in a
high stable system.
Sideslip Response
Roll Response
Yaw Response
Fig. C-9. Response of B-26C to Aileron Pulse Input with I =

xz
Dihedral and Yaw Damping Increased by a Factor of Four
81
APPENDIX D
RESPONSE RECORDS
Numbers on Individual Records Correspond to

Information in Table I, Chapter 7
82
i
.
S„= 15° R Sp=15°L f= 15°R S T =14°L 8 T =13.5°R
CALIBRATION FOR RUDDER NO. 1
....
WMflM M uwiwu
ii i w
iyuiyu
i;i;,s:i» iKWWWW)
=Z5°L S T = 14'L
^ = i5°R >
R =25°R ?>
R
S T = i3.5*R
CALIBRATION FOR RUDDER NO. 2
CALIBRATION DATA FOR PHOTOGRAPHIC RECORDS

EAS 80
" if.' '

il'Wnri,nlnl
""" «i. ii
nYiVliMHUIi'liMrK!
J Jl w»». mill|i L t
i
' V FimrHMiHriflm4fn^1>hll)mmrtM*UiwnVT
>
r*t!
r
r ii'i 'i'i hi 1
1
fWiVrfiii
i
. ^f 'i
i
EAS SO
—| 1 SEC t~ EAS 40
N lHWIHIWm
PPW
1
MMllMMIMlllMllMNMIiMMiriYlWlMilMMMiMlMM^mi'l'-l II 1 1 H I ,11 1 1! I
iuMiml'ni i
II lini Bflljljffl^ '
I'jiH'
1 n, riil[|J !
RESPONSE RECORD for FIXED CONTROLS

RUDDER 2 TABi-1 VISCOUS DAMPER IN
r
St
— &,
EAS GO
EAS 40
RESPONSE RECORDS for FREE CONTROL SYSTEM

RUDDER 2 TAB.1-1 VISCOUS DAMPER IN
GYRO ON
St
40mph
RESPONSE RECORD DIRECT CONTROL SYSTEM

for
RUDDER 2 TAB 1-1 VISCOUS DAMPER IN
EA5 80
^ i.ii n i. M iii i,
^ l|lii|iil i^ .Willi .m iiiiim, i
u^fjllj^ i
mmmmmm y li -m
'
U , i
,i
EAS SO
mMW i
'
MWA^ nM
I i. MM . J^mt^^ W'MV'Vi j ii, iW.«sWiV ih -,U sir .(ii: 11
.,.!"
i.i ii 'i'
«,w»«mwr.
"•'lililll
RESPONSE RECORD for DIRECT CONTROL SYSTEM

RUDDER 2 TAB. i-2 VISCOUS DAMPER IN
m EAS 80
r
h
!
!
;
.
.
;
j
,
i
:
:
'
.
:
!
:
!
.
j
i
|
,
. :
%/•'':' :'{' ;i
rjo' '':,;. $<% }^l :
ff
;
"||ip
EAS 40
1>>
*v.
i ,
\ |l!|ll|l||l"
r
7^',
iTrrrmTTTmi
VWWWJWWWV^^ wmUuuW^^ In .mi ,'i ,VliV
;
1
i
1
!:,liM.ilii'l.i''ii'J '
Mrr*^ Jill.,

EA5 80
A\ \JV/i/WW*W.V,' ,V
/^ ^
\y \y
16
EAS GO
Mh»^w^W¥WtViwvAW<w»\ft*w»viMAM«^^
EAS 40
1
J
S u ffiffl _^_ —
llS ^rZZS
v^
J
!
;
',
i
,
..,,i,..M,„zz:
22

RUDDER 2 TAB. 2-1 VISCOUS DAMPER IN
80mph
w^wwi^Mw^^ Y
;-';'
"-"^fl^flYftlflWfifiir"""'-
17
60mph
M pro rwi v n wii

ft ft i^fi* ftft 'rt vniVflii'fi
Y4 f^ ft ^ ft vii vifl vi^
riHifiiir-iHiffliM-riiTriWifrinil'
|j
u—-'
— | i SEC
;-'77"Y; i
;; ,
rTT; :y;;p;v::7
l
l
l !
l,
^p.

RUDDER 2 TAB. 2-2 VISCOUS DAMPER IN
EAS 80
r
p. ^7^ "friT-' iiM wii M i iiii iiii >i
EAS GO
EAS 40
W>"ilW i
C i. vri W |ti ;|
| Ww
,
24

RUDDER 2 TAB- 2-3 VISCOUS DAMPER IN
80mph
•v
25
COmph
H 1 SEC h-
26
40mph
;i." l
i l
i'i,iw,i,-ifr,iv,i7,i<-,iiwv<-Mf-<»r,i-M>Mii,»Mv l
i-yfy 1 iM ii
|
w .W.WMM
i
27
RUDDER 2 TAB. 2-2 NO VISCOUS DAMPER
EA5 40
RESPONSE RECORD for FIXED CONTROL SYSTEM

RUDDER 1 TAB l-± NO VISCOUS DAMPER
EAS 80
— *| 1 SEC k-
EAS GO
EAS 40
!
:
RESPONSE RECORDS for FREE CONTROL SYSTEM

RUDDER 2 TAB. i-i VISCOUS DAMPER IN
GYRO OFF
60 mph
,.X,'..ll<ll/lMlHM«ll?l»
" fill I'l I HI TO I
Wl'l'lll Hl'l'l I'l in HI IN n I ,;,!„':' Ull.'IVIVfiri'l.l'l I' 'I' 'I' r: i', 'ft T..1MV1 ,'!':'!'. ',<!!',; r
/~ "^ *-/
I>| I I'l
60 mph
c\.
— —XJlî-' -
'
•
'
,.,
;.
'
,;:
ittkiW «mfl«^îfe l
':
r,r
:
1
,^'
,
^y^rn'-^-^^^^'^^^^ m .-
r -'"^îi
v.wa-M'mw/mwihmnummum'w.'mvi
40 mph
/^"N
lAMtatfjWiVW '. '. jf:'.\::.
\^_ _ _.yiivr
I I
'
l II
'
l lll l lllllll 'II Il

'
11 '
" i
"
l
''
l
,
r: ,
l
" l
"i"i" |
l
,
;
i "i ';r'
,
i
' ;,l
'' r
'- '"
"
ji'iUii >î"
-" '"'
'',^'i i,,, '.
'jjSfi
> -
P^l WTi ^r
,.'.,-^^TT*"m;:
ii
ji
-
i
I'll'
"
.y
I

RUDDER 1 TAB i-i VISCOUS DAMPER IN
50mph
'
.
'
iWiVWVW lllHMH
ll
KW.W.Y. '
.'iV,',
."..','. ' .'.
'
.' '
'.' .'.' '
.
'
.
' '
.V. .v.
'
.
'
v.
'
.
'
.
'
mî ii .wumw iwiK W mMrx
L L i
50mph
'Jl^ f.YM
' " '
,
' '
. .
^' "
;
LL^y^/^ :
H 1 SEC h K
RESPONSE RECORD for FREE CONTROL SYSTEM

RUDDER i TAB. 1-1 V/SCOUS DAMPER IN
SPOILER ON
80 mph
WttwWl
*" "MB 1
",
J" 1
,
!
—I 1 SEC h- M
40 mph
Aa
x^
w fll*ffi^
»
^
Pi

itaw» y faww ^flVfJlfflffl
' ^^^w 4 wk»
l > vWjti
^m
|
)
l||(^
mmm TOiipwap^^^'ii ii ii
;
'
' '
H asec h
wflîVig^^^ f
wffli
1
:
.".^'.Viwfif. '

RUDDER i TAB. 1-1 40mph
-
—
NOISE CALIBRATION 2" YAW
—
3° YAW
MARK 5WM
'-OTm
30 mph 5 MM/SEC 60 mph 5 MM/SEC
-jfc»Af. .
4-0 mph 5 MM/SEC 70 mph 5 MM/SEC
50 mph 5 MM/SEC 80 mph 5 MM/SEC
CO mph 125 MM/SEC
80 mph 125 MM/SEC

sm
WIND TUNNEL NOISE DATA
MODEL RESPONSE
" "
REFERENCES
1. Perkins, Courtland D. , and Hage, Robert F. , Airplane Performance ,
Stability, and Control , John Wiley and Sons, Inc. (1949).
2. Sternfield, Leonard, "Effect of Product of Inertia on Lateral Stability,

NACA TN No. 1193, March 1947.
3. Stone, D. G. , and Muse, T. C. , Memorandum Report of the Air

Materiel Command, USAF, "Power-off and Power-on Tests of the .2375
Scale Model of the Douglas XA-26 Airplane in the N. A. C. A. 19 -Foot
Pressure Tunnel, Part II, Stability Characteristics."
4. White, Roland J. , "Investigation of Lateral Dynamic Stability in the

XB-47 Airplane," Institute of the Aeronautical Sciences Preprint No. 241.
5. McKinney, Marion O. , Jr. , and Maggin, Bernard, "Experimental Veri-

fication of the Rudder -Free Theory for an Airplane Model Equipped
Stability
with Rudders Having Negative Floating Tendency and Negligible Friction,
NACA ARR No. L4J05a, November 1944, (Declassified).
6. Liddell, Charles J. , Jr. ; Van Dyke, Rudolph D. , Jr. ; and Heinle,

Donovan R. , "A Flight Determination of the Tolerable Range of Effective
Dihedral on a Conventional Fighter Airplane," NACA TN No. 1936,
August 1949.
7. Weems, William R, "An Introduction to the Study of Gyroscopic Instru-

ments, " M.I. T. Instrument Section, Dept. of Aeronautical Engineering.
8. Street, WiUiam Ames, Milton B. Jr. "Pressure Distribution

G. and , ,
Investigation of a N.A.C. A. 0009 Airfoil With a 50 - Percent - Chord

Plain Flap and Three Tabs," T.N. No. 734, NACA, 1939.
100
" "
9. Ames, Milton B. , Jr. , and Sears, Richard I. , "Pressure Distribution

Investigation of an N. A. C. A. 0009 Airfoil with a 30 - Percent - Chord
10. Ames, Milton B. , Jr. , and Sears, Richard I. , "Pressure Distribution

N.A.C. A. 0009 Airfoil with an 80 - Percent
Investigation of an - Chord
11. Harris, Thomas A. , "Reduction of Hinge Moments of Airplane Control

Surfaces by Tabs, " Rep. No. 528, NACA, 1935.
12. Anderson, Raymond F. , "Determination of the Characteristics of

Tapered Wings, " Rep. No. 572, NACA, 1936.
13. Silverstein, Abe, and Katzoff, S. , "Design Charts for Predicting Down-
wash Angles and Wake Characteristics behind Plain and Flapped Wings,
Rep. No. 648, NACA, 1939.
14. Ames, Milton B. , Jr. , and Sears, Richard I. , "Determination of

Control -Surface Characteristics from NACA Plain Flap and Tab Data,"
T.N. No. 796, NACA, 1941.
15. Sears, Richard I. "Wind Tunnel Data on the Aerodynamic Character-

,
istics of Airplane Control Surfaces," Wartime Report (Originally issued

December 1943), NACA, 1943.
16. Crandall, Stewart M. , and Murray, Harry E. , "Analysis of Available

Data on the Effects of Tabs on Control Surface Hinge Moments,
T.N. No. 1049, NACA, 1946.
101
Thesis 15490
G14 Gallagher
Investigation of a yaw
damper for aircraft.
Thesis 15490
Gl 4 Gallagher
Investigation of a yaw
damper for aircraft.
thesGU 1
yaw damper for aircra
Investigation of a
3 2768 002 01002 7

DUDLEY KNOX LIBRARY

Gal Lager

Uploaded by

Copyright:

Available Formats

Gal Lager

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

Gal Lager

Uploaded by

Copyright:

Available Formats

Report So.

INVESTIGATION OF A YAW DAMPER

Edward F. Gallaghiit Robert L. Jungklas

MASSACHUSETTS INSTITUTE TECHNOLOGY

INVESTIGATION OF A YAW DAMPER FOR AIRCRAFT

Edward F. Gallagher B.S., U.S. Naval Academy, 1941

SUBMITTED IN PARTIAL FULFILLMENT OF THE

MASSACHUSETTS INSTITUTE OF TECHNOLOGY

Prof. Joseph S. Newell

Dear Professor Newell:

In accordance with the regulations of the faculty, we hereby

craft", in partial fulfillment of the requirements for the degree of

Master of Science (Aeronautical Engineering).

An investigation was made of a method for improving the lateral sta-

be stabilized laterally by means of a gyroscopically controlled yaw damper to

Chapter 2 DEVELOPMENT OF THE YAW DAMPER 4

Chapter 3 DESCRIPTION OF THE YAW DAMPER 8

Chapter 4 PRELIMINARY DESIGN OF THE YAW DAMPER 11

Chapter 5 DESIGN OF THE WIND TUNNEL MODEL 17

Chapter 6 WIND TUNNEL TESTS 25

Chapter 7 DISCUSSION OF RESULTS 33

Chapter 8 CONCLUSIONS AND RECOMMENDATIONS 47

Appendix A CALCULATION OF THE ROOTS OF THE AIRPLANE

Appendix B PRELIMINARY DESIGN STUDY FOR THE

Appendix C SIMULATION OF THE LATERAL AIRCRAFT

Appendix D RESPONSE RECORDS 82

FIG. lb. SCHEMATIC DIAGRAM

FIG. la SCHEMATIC DIAGRAM OF YAW DAMPER

This paper presents a theoretical investigation into a method of improv-

of automatic aircraft control. The important distinction between this Yaw

DEVELOPMENT OF THE YAW DAMPER

The calculations given in Appendix A indicate than an airplane with the

oscillation is also a function of wing dihedral, it is observed that the Dutch

The results of the investigation cited in reference 6 support the conclu-

"The maximum tolerable effective dihedral at cruising speed

[sic] unstable. The term 'tolerable' describes a condition

lateral stability will be met.

DESCRIPTION OF THE YAW DAMPER

FIG. 3. YAW DAMPER FUNCTIONAL DIAGRAM

PRELIMINARY DESIGN OF THE YAW DAMPER

The airplane lateral stability criteria as developed in Appendix A, indi-

practicability of the design. A primary factor of interest was the friction

Further restrictions on the maximum angular displacements of the tab

For an estimated tab hinge moment of 1. 5 inch-pound, and an angular

DESIGN OF THE WIND TUNNEL MODEL

Theoretical analysis established the benefits that may be derived by in-

I ., by varying the distance of a lead weight from the pivot point.

there is a tendency in modern aircraft design to increase the (AR) ., within'

WIND TUNNEL TESTS

In order to measure the effect of a control surface actuated by a rate

S33«93a ni 3DNV"ingyni T3NNn±

FIG. 13. MODIFICATION OF VERTICAL TAIL

8R , RUDDER ANGLE IN DEGREES TAB ANGLE IN DEGREES

FIG. 14. CONTROL EFFECTIVENESS CURVES, SYSTEM 2-1-1.

In carrying out the investigation of a simplified yaw damping system for

<\> + ¥— 4; + i- <\> = Yaw Damper Operating

where: DR = damping ratio u> = undamped natural angular velocity

tems, the response equation is assumed to remain a second-order function,

The comparative effectiveness of the various control conditions and con-

initial amplitude is normally used for comparison purposes. Because of the