WT A* C

-va-S7Zw 7/,40

MEMORANDUM
REPORT A UFAIRS.tV
PUBLIC E

Ohio
WRGT I DEVELOPMET CENTE

I.'. -1

SUJBJECT-: A Method for Evaluating Aircraft Stability

Parameters from Flight Test Data

SERIAL NUMBER: WADC-TR-52-71

WRIGHT AI DEVElOPMETCNTE
Reproduced From
CLASSIFICATION: UnclassifiedAvailable Copy Best

DATE: June 1952

AMC Form I"U S AMC Form 142D,

.pon""
( J U 3i0) w h ich ,will b e wsed. OIC A11i- 2 5 1 " VM
 NOTICES

When Government drawings, "pcificationa, or otber da~a,%

-used for am purpose other than.in camectibn vith a defililtAelyrelated
G u proemvrent operation, the United. States Govbftment there-
by incurs no responsibility nor any obligation vhatsoever and the fact
that the Govement my have formulated- furnished, or-aifaMW wSW sup-"
plied the said drawings, specifications, or other data, is not to be re-
garded by implication or otherwise as in any manner licensing the holder
or any other person or corporation, or conveying any rights or permis-
Ssion to mnafacture, use, or sell any patented invention that may in any
way be related thereto.

The informtion furnished herwith is made available for study

upon the mnderstanding that the Government's proprietary interes i-,in
and relating thereto shall not be impaired. It is desired that the Office
of the Judge Advocate (WCJ), Wright Air Development Center, Wright-Patterson
AFB, Dayton, Chio, be promptly notified of any apparent conflict between the
Government's proprietary interests and those of others.

The U. S. Governmet is absolved from any litigation which

may ensue from the contractor's infringing on the foreign patent rights
which may be involved.
AF TECHNICAL REPORT No. WADC-TR-52-71 June 1952

A METHOD FOR EVALUATING ATICRAFT STABILITY

PARAMETERS FROM FLIGHT TEST DATA

Lloyd E. Schumacher

All Weather Branch

Flight Test Division
Project No. R27h-9

United States Air Force

Wright Air Development Center
Wright-Patterson Air Force Base, Ohio
 ABSTRACT

A least-squares method for evaluating aircraft stability

derivatives and transfer function constants from flight test fre-
quency response data is derived. The normalized least-squares
equations for rotating vectors are derived by minimizing the sum of
the squares of the residual vectors without regard to the phase
angles of these residuals.

This method was applied to longitudinal oscillation data ob-,

tained from flight tests of a USAF bomber-type B-25J aircraft. It
is shown that a linear dependency in the longitudinal variables of
motion must be resolved before any solution can be obtained. The
assumption of a linear relationship between the stability derivatives
for the forces or moments resulting from the rate of change of the
angle of attack and the pitching velocity eliminated this dependency
and gave acceptable results0 The lift caused by pitching velocity
must be included in order to obtain a reasonable value for the lift
resulting from elevator deflection0

Close agreement was obtained for the related values of the

transfer function constants from the data of several independent
measr emeents.

The application of vector algebra greatly simplifies the numerical

computation required to obtain the normalized equations0

PUBLICATION REVIUW

Manuscript Copy of this report has been reviewed and found

satisfactory for publication.

FOR THE COMMANDING GENERAL-

Eo M. GAVIN
Colonel, USAF
Chief, Flight Test Division

WADC-TR-52-71 iii
 TABLE OF CONTENTS

Page

Introduction 1

Symbols and Notations 3

Theory 5

Application of Method and Discussion of Results 11

Conclus ions 17

Appendix I 18
Derivation of Longitudinal Dynamic Response
Equations

Table 1 23
Calculation and Summation of Scalar Products for
the Normalized Equations

Table 2 30
Method of Solving Normalized Least-Squares Equations
With Four Unknowns

Table 3 31
Sample Solution of Normalized Equations for
Stability Derivatives

Table 4 32
Sample Solution for the Pitching Velocity Transfer
Function Constants

Table 5 33
Sample Solution for Angle of Attack Transfer
Function Constants

Table 6 34
Sample Solution for Normal Acceleration Transfer
Function Constants

Table 7 35
Transfer Function Constants Computed from Flight Data

Table 8 36
Transfer Function Constants from Flight Test Data

References 46

WADC-TR-52-71 iv
 LIST OF ILLUSTRATIONS

Figure Title Page

1 Right Side View, B-.25J 433892 37

2 Frequency Response of Pitching Velocity to 39
Sinusoidal Elevator Oscillations

3 Frequency Response of Normal Acceleration to 39

Sinusoidal Elevator Oscillations

4 Effect of Frequency on Moment Derivatives Obtained 40

from Flight Test Data

5 Effect of Frequency on Lift Derivatives Obtained 41

from Flight Test Data

6 Effect of Neglecting CLo• CL , and CL on the

Values obtained for Cýxiand tLd C

7a Effect of Neglecting CLy CL and on the 14

Residual Amplitudes L& JY

7b Effect of Neglecting CL•, C, and CL on the 14

Residual Phase Angles

SPitching Velocity Frequency Response Computed from 45

Transfer Function Constants and Compared with
Flight Data

WADC-TR-52-71 v
 INTRODUCTION

Considerable effort has been expended in recent years to improve flight

test methods of evaluating aircraft stability derivatives and transfer
functions. These improved methods are required to provide more accurate data
for the design of automatic control equipment and to give the aircraft de-
signer a better check on wind tunnel and estimated parameters. Satisfactory
aircraft handling qualities are also usually expressed in terms of some
stability parameter, thereby necessitating flight tests to determine compli-
ance with specifications.

Most flight-test methods of evaluating aircraft stability parameters are

primarily steady-state maneuvers. A standard approach is to maintain constant
as many variables of the aircraft motion as possible while measuring the
effect of the remaining variables. For example, the slope of the lift curve
is usually found by measuring the change in the angle of inclination of longi-
tudinal axis with changes in the air speed in straight and level flight 0 This
method is subject to various possible errors, however 0 Lift resulting from the
change in elevator position and lift caused by changes in the flow field re-
sulting from propeller "wash" and airspeed changes ate included in addition to
lift resulting from change in angle of attack0 Also, when the lift coefficient
is a function of the Mach number, this method cannot be applied0 The elevator
effectiveness is usually evaluated in static methods by measuring the elevator
deflection required to offset a known moment applied to the aircraft. This
moment change is obtained by shifting the center of gravity location. Again
some error is involved in this method, since some elevator deflection is re,
quired to overcome the pitching moment resulting from changes in the aircraft
angle of attack0 This change in the angle of attack is a by-product of the load
shift between the wing and the tail as the position of the center of gravity is
changed0 These methods have the further disadvantage of requiring extensive
flight-.test time and expert piloting technique,.

The method described in this report adopts the approach of activating all
of the variables of the aircraft motion simultaneously while retaining the oper,-
ating conditions as constant as possible. Values of the stability parameters
are then deduced in accordance with the theory of probability0 It is necessary
to use an equation describing the relationship between the response and the
stability parameters. This relationship can be expressed in the form of equa,-•
tions of motiQn or as transfer function equations0

The flight test evaluation of the stability parameters by this method

requires that each variable included in the equations be measured directly or
be derived from measured data. The evaluation of the transfer fanction con•,
stants in the equation for the rate of pitch resulting from elevator deflection
requires that the first, second, and third derivatives of the angle of pitch be
obtained, as well as the elevator angle and its first derivative. With the best
instrumentation techniques available to date, it has been found practicable to
measure only the first pitching derivative with an acceptable degree of accuracy0
The remaining derivatives must, therefore, be obtained from the measured rate
by analytical, numerical, or graphical means.

WADCTTR-52-71 1,
 Two possible techniques for exciting the variables of the aircraft motion
under constant flight conditions are the transient method and the sinusoidal
oscillation method. In the transient method the aircraft is disturbed from
a steady-state condition by a step or pulse type motion of the aircraft con-
trol surface. The oscillation method consists of obtaining the aircraft
motion by a sinusoidal variation of the control surface about a trim point
at various frequencies.

The transient method requires a minimum of flight time and has the fur-
ther advantage of being applicable to flight conditions other than straight
and level, such as climbs, descents, or turns. However, the evaluation of
the variables of motion and all required derivatives with sufficient accuracy
to permit solving for the stability parameters does not appear feasible in
the time plane. Ir the above example of the transfer function constants for
the rate of pitch, 6 can be measured directly. If the time history of the
rate of pitch is quite smooths the second derivative of the pitch rate can be
obtained with fair accuracy by standard numerical or graphical methods. How-
ever, if the third derivative is then obtained similarly from the second, the
inaccuracies are no longer tolerable. For those elastic aircraft having bend-
ing modes that appear on the record in addition to basic airframe response,
even the first derivative that is obtained by graphical or numerical methods is
of doubtful value.

The oscillation method requires extensive flight-test time and is essen-

tially restricted to straight and level-flight conditionas However, the re-,
maining derivatives of the components of motion can be obtained analytically ih
the frequency plane with ease and accuracy from the phase and amplitude of any
measured components. The measured component can be expressed as a rotating
vector iin thq~compl eo pe•ge by the relationship 0 - 101e3- The second deriva-
tive is 0 = i•- Q e Consequently, the derivative of a rotating vector
can be obtained by advancing the phase angle by 90 degrees and increasing the
amplitude by the factorW. Higher order derivatives can be evaluated similarly.
The oscillation method has a further advantage in that the recorded data can
be corrected for the error introdved by the dynamic response characteristics of
the instrumentation by applying phase and amplitude calibrations at each fre-
quency,

The aircraft responses obtained by the two methods above can be related
analytically by Fourier or Laplace Transformation theory. This provides a
means of taking advantage of the best features of each technique. The flight
test time can be minimized by recording measured responses in the time plane
and transforming these responses to the frequency plane to take advantage of
the simplified analyses procedures available. The transform theories assume
that the principle of superposition applies, that is, that all components of
the motion are additive and that the system can be represented by a set of
linear differential equations with constant coefficients. Therefore, any non-
linearities existing in the response characteristics can, possibly, lead to
prohibitive errors in the transformation from the time plane to the frequency
plane. The validity of the transformation can be checked by observing the
agreement in several frequency response characteristics obtained by analyzing
transients having appreciable differences in frequency content. A numerical
method for accurately obtaining the Fourier transforms of arbitrary time
histories is given in Reference lo

WADC-TR-52-71 2
 SYMBOLS AND NOTATION

J perturbation elevator deflection, radians

O perturbation angle of pitch, radians

6 pitching velocity, dG/dt, radians/second

2 2
pitching ac celeration,, d idt2 radians/second
6'

c perturbation angle of attack, radians

perturbation flight path angle, radians

angular frequency, radians/second

air density , slugs/feet3

m/iVS, second

0 phase angle

downwash angle at the tail, radians

dC/cc rate of change of downwash angle at the tail with angle of attack,

summation of scalar products

4 determinant of coefficients
2
CL lift coefficient , 2mg/f V S

CL rate of change of lift coefficient with angle of attack

49

CL rate of change of lift coefficient with elevator deflection

04
CL6 rate of change of lift coefficient with pitching velocity

C rate of change of lift coefficient with rate of change of angle of

ýx attack

CL.6 rate of change of lift coefficient with angle of attack acceleration

Cm moment coefficient. 21y/'P VSC

C rate of change of moment coefficient with angle of attack

C rate of change of moment coefficient with elevator deflection

Cm6 rate of change of moment coefficient with pitching velocity

WADC-TR-52=71 3
 C rate of change of moment coefficient with rate of change of
angle of attack

C rate of change of moment coefficient with acceleration of angle

of attack

A in-phase component (with subscripts c, 4 gi n)

B out-of-phase component (with subscripts acj, 69i. n)

A,B;,C,E transfer function constants (with subscripts 0, 1, 2)

D differential operator, d/dt

i imaginary unit, 7 (complex notation)

i unit vector along vertical axis, (vector notation)

V true velocity, feet/second

it tail length, feet

IyYaircraft moment of inertia about lateral axis, slug-feet 2

m aircraft mass, slugs

2
a aircraft vertical acceleration, feet/second

S wing area, square feet

C mean aerodynamic chord, feet

n normal acceleration, gts

g acceleration due to gravity, 32.2 feet/second 2

2 2
2Iy/•V SC, second
h

WADC-TR-52- 71
 1H EDRY

Having obtained an accurate expression of the motion of all variables

of interest, the problem consists of obtaining values for the associated
parameters of edch variable, namely; the stability derivatives or the
transfer function constants. For simplicity, the theory will be developed
for the stability derivatives in the lift equation assuming that lift will
result only from the incremental changes in the angle-of-attack and elevator
deflection. Equation 8, Appendix I, will then be written as
CL C CCLn=
CL+ (1)

This equation relates the lift forces to the vertical acceleration at any
instant of time, irrespective as to whether the motion is any arbitrary
transient or a steady-state sinusoidal oscillation. In the latter case, the
variables can be considered for any given frequency to be constant amplitude
vectors rotating in the complex plane with fixed phase relationships0 These
amplitudes and phase angles can be experimentally determined for a number of
different frequencies at a constant flight configuration0

If these data were obtained without error of any form and the above equa-,
tion described the aerodynamic and inertial forces completely, it would then
be possible to solve for the unknown derivatives directly. The variables
could be resolved into components along the real axis for as many frequencies
as there are unknowns. A simultaneous solution of these real equations would
yield the accurate values of the derivatives. Since neither of these assump-•'
tions is true, this approach leads to large errors in practice0 Random
errors may be present that are larger than the effect of some derivatives at
the particular time in the cycle selected for taking real parts, The effect of
a neglected stability derivative will also be large when its associated vari-
able has relatively large components along the real axis as compared to some
included variables. Standard curve fitting theory for scalars can be extended
to minimize these errors when the variables are rotating vectors.

The problem of obtaining values of the stability derivatives in the sims,

plified lift equation given above will now be considered when the phase angles
and amplitudes have been obtained from flight data for a number of frequencies0
Any values that are found will not satisfy the equations exactly at each fre-
quency point at which data are obtained0 This is expressed algebraically by
transposing all of the terms in equation 1 to the left•-hand side and retaining
the residual resulting from the inequality on the right-hand side.

C i + CLi + Cni ý vi; i = im 2ý 39 m (2)

where m = the total of frequency points

v = residual

Since all terms on the left are vectors, the residual, vi, will also be a
vector at each frequency0 It is now required to find values of the de-
rivatives that will make the amplitudes of these residuals small. The

WADC-TR-52-71 5
relative phase angles of these residuals that result from minimizing these
amplitudes are of no concern at this point. Later 9 it will be shown that
these phase angles can be used as an indication of the validity of the
equation being fitted.

It can be shown (Reference 2) that the most probable values of the un-
knowns will be such as to make the sum of the squares of the residuals a
minimum. The square of a residual vector is a scalar equal in magnitude toq
the square of the amplitude of the vector,
= +2 + V 2 + 0 0 0 V 2

+ CL i + CLni) 2 (3)
=+(c
ff(C CL) 1, 2I,3, o m

The condition that this function be a minimum requires that its partial
derivatives with respect to the unknowns be zero0

af =2(C Cc + CLi+ CLni )OCi+ 2(Cix2 +CLJ 2 +CL2) 2 (C)

+ 0 ° C~m•m + CL2(Cm = 0
or C i2+ C di - W-7CLnio. ic i 1, 2, 0 0 0 m. (5)

Similarly, when

af =0
4)cL (6)
C 1 ii+ CL2,no, i 1, 2, 3, 0 0

Equations 5 and 6 are the normalized equations for the least-squares solution
for two unknown coefficients of variable vectors, The indicated products of
the various pairs of vectors are defined as the scalar or dot products
(Reference 3). The scalar product of two vectors is numerically equal to the
product of their amplitudes times the cosine of the angle between the two
vectors0 This is equivalent to multiplying the vector of the variable being
normalized by the respective components of the other vectors along this
normalized vector.ý In this way each individual equation is normalized for
each unknown parameter at that unique angle at which the associated vector is
at its peak or maximum value0 It is apparent that each parameter will have
the maximum effect on the response when its vector is at its peak value0 It is,
therefore, logical that the parameter should be so selected that it yields a
minimum error at the point where it has maximum effect. Basically9 in the
equations given above, this amounts to resolving the vectors into real parts
with the angle-of-attack vector lying on the real axis when normalizing the
equation for CT and then resolving the vectors into real parts again with the
elevator.deflection vector lying along the real axis when normalizing the
equations for CL 0o

WAPC-TR-52-71 6
 Using the definition given for the scalar products the normalized equations
can be simplified.

Q&6 olIcos %

jrxý crj(cos 0& cos + sinl sin )

where indicates the absolute length of the vector
is the angle between thee and j vectors
• is the angle between cK and the real axis
is the angle between. and the real axis

Let 1 = Pkx cos • A4 = lo cos

Boc = dKj sin , Br = fIj sin Or-

Then a od½= 6Ajd +BB(

Similarly, the remaining scalar products can be found in terms of real and
imaginary components of the complex variables. Equations 5 and 6 become

2C + 2 + C A
AAiA + BCB() CL(+BAn + BnB n (9)
Li X
hiii
2 2(A:iL An + •B B (10)
A +Bo B) + (i2 +CBi 2 )z ECL( n(

The amplitudes of all the vectors are usually given in terms of a unit input, and
the phase angles are all referred to the input vector which, in this case, is the
elevator deflection° Therefore, the elevator deflection vector will have a unit
length lying along the real axis in every case, Then A 1 and B = 0 and
the normalized equations become i i
(•.+•)+ C AiAn +BBni) (11)
Lk iY nii i

Cl Zti + mCL = -KC A,ni i = 1, 2., 3, 6 o 0 MO (]-2)

The original vector equations can also be separated into real and imaginary
parts and these scalar equations normalized by the conventional least-squares
method. When all the real and imaginary equations are summed in the normalizing
process, the same equations as given above will result. This approach, however,
makes it difficult to visualize the basic validity and effectiveness of the
method,

The least-squares method for vectors can now be applied to obtain values
for all of the moment and lift derivatives included in the equations of
motion in Appendix I. It is first necessary to resolve the linear dependency
that exists between the variables of motion, as shown in the derivation,

WADC-TR-52-71 7
before normalizing the eq-ationso In the simplified theoretical evaluation
of the stability derivatives, it is found that the ci and 6 derivatives are
directly related by a constant equal to the rate of change in downwash angle
at the tail with change in aircraft angle of attack (Reference 4). This can
be expressed algebraically as

c KCM (13)

C X KcL 4 (1 4)
dE
where K = d, the change in downwash angle at the tail with change in
aircraft angle of attack.
Substitution of equations 15 and ]1
4 into equations 7 and 9 of Appendix
I will reduce the equations to three linearly independent variables.

CM+ C + Cm (6 + F = h6 (15)
a @r

Cc + Cr0+ CLO(Q + ) _=CLn (16)

The solution of these equations by least-squares will yield values of the

derivatives directly. These derivatives will act along the new vector,
(6 + j), and thereby include the effects of the i derivatives. Since theol
derivatives are relatively small the effects of inaccuracies in the assumed
values of K should be minor. Equations 15 and 16 will be written in matrix
notation for compactness before normalizing.

[ (+C+CLn = or h6 (17)
c

C•; J = morL

1
The normalized equations for m data points can be obtained directly by scalar
multiplication of each equation of motion by each element in the matrix of
variables in turn and summing the results.

[cc2 a< + i~c) ICI1)

zcc (~+ NO 90 CH(G + KdC) :F N~ + A j L4J

or
or

:T
I
WALn c-
10 1
j - L [hi5 (9 ++ j W_M
ýE CLn (9 ++

WADC-TR-52-71
 Expanding and simplifying the scalar products and letting 4- AC + iBC
1, the normalized equations become

(19)

1:-CL(A An + Vn)

E-CLAn

2-CL (AnA + BB 0 ) + K(AoA + BBn) J L

[ ~ RcA 4)1
or i- %
B,

L 2 hK(A Aý + kxBZ)J
&A =m

Similar solutions can be made for obtaining values for the transfer
function constants given in equations 12, 13 and 21. Appendix I. The
normalized equations for the constants in the equation for 6/1d will be de-
rived first. Rewriting equation 12, Appendix I, and transposing terms to
eliminate fractions gives
0 2
AO0 + A1D6 + A2D = BO0 + BDdo (20)

To obtain a unique solution for the constants, it is first necessary to reduce

the number of unknowns by one. Since A2 is almost entirely made up of known
quantities, it is the constant most convenient to eliminate. This can be done
by dividing through by A 2 . Rearranging and placing a bar over the remaining
constants to indicate division by A2 , the equation becomes

fS+ f1 D6 = D26 (21)

Substituting D - io and writing in matrix notation gives

-ig 011LBF0 (22)

~B1~
Ii
A0

The normalized equations for m data points can be written directly by the
method given previously.

WADC-TR-52-71 9
 2 1!
1. -
a-sJoiWQ
• iA w-40iW6 C-iuJ)
j-•,,(%) 2:Q
a(o)2
(-iL4) A1i
Ao (l

L~J 1o-
d .-

•,2Ao6 ,

Expanding the scalar products under the summation symbols and letting
J= A + iBd = 1, the normalized equations become

ZWB;
m
0 ~

,V- Ai
6
£-A
2
0

2
Ae
B0
E.B§

;rh9
_W
2 A;
-4
2
62
~~~
i-A6

0
62
L1k6II(4
So

0
ZVAo

[0 0
,, _ o262
Similarly, the normalized equations for the constants in the angle of
attack and normal acceleration transfer functions are found to be

m 0 Z( BMS -- Ax C(25)
0 2 k t.2 I Bl 6 Ig) B
:9uj%, £ 4Ocx 0 A1 0

m 0 JAn ;£wBn t'Ir _d0E

0 W2 :2 -L'un2n I-&A 0 E (26)

ý-An X n X n2 0 X4 2 An 10
:E jB :1E-?An 0 X w2n 2 E _10B A!

.E=JBn
2n2

0
An

WADC-TR-52-71 i0
 APPLICATION OF METHOD AND DISCUSSION OF RESULTS

The least-squares method for vectors was applied to actual flight test
longitudinal response data to obtain values for the stability parameters. The
test aircraft was a twin-engine light bomber-type B-25J and is pictured in
Figure 1, Appendix I. A complete set of frequency response characteristics
for the air speed, altitude, and center of gravity range of this airplane has
been measured in flight. These measured responses are intended to provide a
store of flight-test data for the development of data reduction methods and
to investigate the possibility of extrapolating the results of limited test
programs. The test methods and instrumentation are described in Reference 5.

The response data, used to demonstrate the application of the methods of

this report, were obtained at an altitude of 10,000 feetý 155 mph indicated air
speed, and with the center of gravity located at 27 per cent of the mean aero.=
dynamic chord. The low air speed was selected to provide a large range of
values of the reduced. frequency, K, where
WsC
K r= (27)

W is the frequency of oscillation in radians per second.

C is the wing chord9 feet

V is the free stream velocity, feet/second

Unsteady flow effects are, theoretically, a function of the reduced frequency.

The results obtained from this example will give some indication of the possibili-,
ty of evaluating these unsteady flow effects from flight-test data.

The measured frequency response characteristics of the aircraft pitching

velocity and normal acceleration to sinusoidal elevator oscillations are given
in Figures 2 and 3 respectively0 Some of the scatter in the data is caused by
the small variations in air speed, altitude, and weight that occurred during the
course of the oscillation tests. The parameters and responses obtained at each
frequency are listed in Table 1, Columns 1, 2, S, 99 15, 19ý, 27, and Lj2-o

Table 1 also presents a procedure for obtaining all of the scalar products
and sums of products required for the normalized least-squares equations for
the lift and moment derivatives and the pitching velocity, normal acceleration,
and angle-of-attack transfer function constants. When only the stability
derivatives are required, this table can be redued from 5) columns to 40.

A convenient short-•cut method of solving simultaneous linear equations

on a desk calculator is outlined in Table 2 and is self- explanatory. The
theory and development of this method are explained in References 6 and 7.
Basically.9 iý consists of combining the steps that are normally required in
reducing the equations, one variable at a time, into a complete operation for
each element. The advantage, when using a desk calculator, is that the results

WADC-TR•-52=71 11
of intermediate steps are retained in the machine, thereby eliminating writing
down the results of each step.

The solution of the normalized equations of motion for the stability

derivatives is outlined in Table 53 The augmented matrix, in this case con,-
sists of the matrix of coefficients together with the right-hand columns for,
the lift equation and the moment equation° The matrix of the coefficients
will be the same for the lift and moment equations as shown in the derivatitn
(Appendix I). The auxiliary matrix for the coefficients needs only to be cal.-
culated once to obtain both the lift and the moment derivatives. The second-
right hand column in the auxiliary matrix is found by the same steps as the
first.

By arranging the variables in the equation of motion in the order of

descending magnitude before normalizing, it becomes possible to check the
effect of neglecting the smaller derivatives on the values obtained for the
remaining derivatives. This can be done by assuming that the values of the
derivatives to be neglected are zero and completing the remaining solution as
before. This is shown in Table 5 where CLZ is assumed equal. to zero and values
for the remaining lift derivatives are found. Then CL6 and. CL are both dropped
to obtain a solution when only C is left. This method will not work if inter-
mediate derivatives are assumed zro and the remaining derivatives to the right
are not zero0

The values of the derivatives listed in Table 4 agree closely to values

obtained by other methods where comparable data are available0 The slope of
the lift curveg CTX9 was found to be 5.lll compared to the value of 5.25 from
wind tunnel results (Reference 8). The value of the pitching moment deriva-
tive, CW, equal to -.553, agrees well with the wind tunnel value of .. o525
(Reference 5).

Theoretically, the lift and moment resulting from elevator deflection are
related as follows (Reference )4)-

CM it X CL/C = 2o57 CL
where C - the mean aerodynamic chord = 9.69 feet for the B-25J

it = the effective tail length = 25.9 feet for the B-25J 0

It can be seen that the results in Table 3 where Cm =lo_1..l and CL( 0O556
are in good agreement with this equation0 The same relationship should hold
for the pitching velocity derivatives, however, the agreement is not as good
here for Cm,4 = -0o270 and CL, = Ooltlo

As pointed out in the previous section, the effect of the rate of change
of the angle of attack is included by an assumed linear relationship with the
rate of pitch derivatives. The derivativeP CLY is so small that it can be
neglected without appreciable effects on the ýilues for the remaining lift

WADC-TR-52-71 12
derivatives. This was verified by actually carrying out the solution and ob-
taining values of 5.061, 00526, and 0o194 for C C-L, and CL. respectivel-y0

When the lift caused by pitching velocity is not included, the value
found for CL is considerably reduced (0.287), although the relative effect
on C is miuSh less (5.206). The value for the slope of the lift curve when
the angle of attack is the only variable included , is 5.176.

One of the reasons that the value found for 0C is only slightly affected
by neglecting the smaller lift derivatives is that~fhe sums of the products
in the normalized equations are predominately composed of the responses in the
low frequency range, At these frequencies, the angle of attack effect is large.
whereas the amount of lift from the elevator deflection and the rate of pitch
are approximately equal and 1•0 degrees out of phase, thereby having little
net effect. At higher frequencies, these latter derivatives become increasingly
important as their associated variables are relatively large and their relative
phase angle changes,

. At the high frequencies, the effects of other higher order derivatives

such as the lift and moment resulting from the angle of attack acceleration,
Ci and C W should become appreciable. To examine these effects the
normalized equations were obtained from the results of a small range of frequency
and solved as in the example in Table 3. The sums of the products from six ad-
jacent frequency points such as 1 through 6, 2 through 7, etc., were taken from
Table 1. The values of the derivatives found in this way are plotted versus the
median frequency of each group in Figures 4 and 50 The derivatives frund in
Table 3 are plotted as dashed lines over the entire frequency range to provide a
comparison with the overall averages,

The angle of attack stability derivatives show consistent trends which can
be attributed to the angle of attack acceleration derivatives . The variables
a and c are vectors along the same line but 190 degrees out of phase. The re=
sponse caused by the & stability derivatives will either increase or decrease
the effective value of the •c stability derivatives depending on the relative
signs. The remaining parameters are surprisingly constant throughout the
frequency range, considering that the theory of unsteady flow indicates varia-
tions of these derivatives with frequency0 . The scatter in the low frequency
range is believed to be caused by errors in the data resulting from slight tur-
bulence, although variation in the speed of the aircraft during the oscillation
cycle may have contributed also0

The normalized lift equations for these limited frequency ranges were also
solved neglecting first the CLo and C derivatives and then also the CL deriva-
tive. The results are plotted in Fig~e 6o The importance of including these
latter derivatives in the solution at high frequencies is clearly shown in this
graph0 Values found for the slope of the lift curve are more than double the
theoretically expected size in the high frequency range when these derivatives
are neglected,

The effectiveness of the least-squares process in fitting the equations of

WADC0TR-52,7V1 13
motion to the original data can be examined by plotting the phase angles and
amplitudes of the residual vectors for all frequencies. If the equations of
motion include all of the forces that act on the aircraft, the residuals should
consist of small vectors with random phase angles caused by the normal dis=
tribution of errors in measurement. If consistent trends do appear in the plots
of the phase angles of the residuals versus frequency, it should be possible to
find another stability parameter for the equation of motion. This parameter
should be nearly in phase with the residuals. This effect is demonstrated tn
Figures 7 a and 7 b, where the phase and amplitudes of the residuals are plotted
for several lift equations. The lift derivatives are obtained from Table 5o

The top graphs in Figures 7 a and 7 b are the amplitudes and phase angles
of the residuals, respectively, when only the lift from angle of attack is con-
sidered. The amplitudes are quite large and the phase angles are mostly in the
vicinity of 200 degrees. Now, since CT, acts along the 0 or 180 degree vector
line, it would appear that a value coul1 be obtained for this parameter. The
effect of including CL in the equation of lift is shown by the residuals in the
middle graphs, Figures 7 a and 7 b. The fit has been improved, as shown by the
decrease in the sum of squares of the amplitudes0 However, the phase angles
still show a fairly consistent trend. The inclusion of the CLo and CW deriva-
tives reduces the amplitudes of the residuals still further lwhe the scatter
in the phase angles becomes considerably more random0

It is possible that further analysis will reval other derivatives that

will improve the fit of the lift or moment equations to the data. Care would
have to be exercised to avoid introducing linear dependencies in the variables
of motion in order that valid results would be obtained0

The effects of some consistent errors in the measured flight-test data

can be pointed out at this time0 Calibration errors in air speed, altitude,
temperature, weight, and the moment of inertia will be consistent for all fre-
quency points. These items appear primarily on the right-hand side of the
normalized equations of motion, Since the equations are linear these errors
will affect all of the stability derivatives by the same percentage. The true
air speed is also involved in the computation of the angle of attack. so,
additional error will result if this item is not accurate. Control surface de-
flections are difficult to measure accurately because of twist and distortions
in the surface caused by the air loads0 Since only the amplitude is usually
in error, the effect on the accuracy of the derivatives is minor0 The surface
effectiveness will be in error inversely as the error in measurement while the
other parameters will not be affected0

The transfer function constants were also computed from flight test data
using average values of the flight conditions and flight test values of the
stability derivatives. These computations are shown in Table 7. The computed
values of the constants are compared in Table 8 with those obtained directly
from the normalized equations0 The agreement is good for most constants al-
though there is appreciable difference in several cases such as Bo0 The effect
of these differences was examined by plotting the transfer functions for the
pitching velocity in Figure 8. The results are in very close agreement throughout

WADC-TR-52-71 114
the frequency range. The actual flight test results are plotted in Figure 9
also to show the OfitN obtained by the two methods. The transfer function
equations fair the flight test data very closely in all respects except for
the amplitude ratio in the lo frequency range.

The individual data points could be weighted, prior to n6rmalizing, by a

factor proportional to their relative accuracy of measurement. It can be
seen that this would improve the accuracy of the parameters that are obtained
in the solution since the effect of the less accurate data in the normalized
equations would be decreased. The errors in the data obtained from flight
test are predominantly of an incremental nature resulting from reading in.=
accuracies. In the normalized equations for the stability derivatives, the
weighting is automatically included by summing the data directly as measured.

In the normalized equations for the transfer function constants the results
of high frequency oscillations comprise the largest part of the coefficients as
a result of the high powers of the frequency that are involved0 Since the re-ý
sponses in this frequency range are small these results are probably less ac-,
curate and therefore should have less weight. That this might be the case is
indicated by the comparison of the plots of the transfer function equation
using constants from the least-squares solution and of the actual flight data
(Figure 8).

The effect of increasing the weight of low frequency data points was in-,
vestigated to determine whether the "fit" in this region could be improved.
This was done quite simply by including only the first 17 data points in the
normalized equations. These results are also included in Table S. In most
cases these values are in close agreement with those obtained from the previous
solution0 The frequency response characteristics for the pitching velocity from
these constants corresponded so closely to those already shown in Figare 8 that
it was not possible to plot the results as a separate curve0

The consistent results obtained throughout this investigation give a strong

indication of the validity and accuracy of the aerodynamic theory, flight test
measurements, and data analysis methods 0 For example, the constants, AO and A
are common to each of the three transfer function equations investigated0 The
data for the rate of pitch and the normal acceleration were obtained simultan-
eously but by independent measurements. Yet the agreement in the values of these
constants obtained from these two sources is quite good as also is the <•alne
found by substituting stability derivatives and flight conditions into the theo-
retical expression for these constants (Table S).

The p.ots of the derivatives in Figures 4 and 5 are other illustrations of

these consistent results. Since the data selected for each set of derivat'i•es
covered only a very small frequency range the equations for the six data points
in each group are nearly the same in magnitude0 In other words, the equations
are nearly parallel and therefore the results are highly sensitive to errors0
The values plotted in these figures show little scatter, however. The magni-
tude of the normal acceleration residuals in the lower graph, Figure 7 a, com-=
pared to the size of the measured data in Figure 3 is another example. Note
that the scale in Figure 7 a is five times as large as Figure 3.

WADC-TR-52-71 15
 The method of least squares for vectors also provides a possibility of
solving for the aircraft longitudinal moment of inertia. The transfer function
A2 is directly proportional to the moment of inertia. All the remaining con-,
stants are divided by A, in the solution of the normalized equations. ohereo-
fore, the values found or those constants which are not a function of Iy before
division by b will be inversely proportional to the moment of inertia. These
include B B andi E0 By keeping all the other flight conditions
0 0. ' 0
constant and 6btaining data at two values of the moment of inertia differing
by a known increment, it is possible to arrive at an explicit solution for the
total moment of inertia for each of these constants0

WADC-TR=52-71 16
 CONCLUSI ONS

The method of least-squares for vectors provides a method for evaluating

aircraft longitudinal stability derivatives from frequency response data pro-
vided the linear dependency existing between the variables of motion is elimlnatec.o
The assumption of a linear relationship between the angle of attack rate and
pitching rate lift and moment derivatives eliminated this. dependency an4 gave
acceptable results. This method of least-'squares can also be used to find values
from flight test frequency response data for the constants in the aircraft tran•,,
fer function equations.

The lift caused by pitching velocity must be included for conventional air-
craft in order to obtain reasonable values for the lift resulting from elevator,
deflection.

The use of vector algebra greatly simplifies the numerical computation re-
quired to obtain the normalized equations.

WADC TR- 52-71 17

 APPENDIX I

DERIVATION OF LONGITUDINAL DYNAMIC RESPONSE EQUATIONS

The motion of an aircraft in flight is prescribed by aerodynamic, pro-

pulsive, gravitational, and inertial forces. The equations of motion of an
aircraft for small perturbations from a steady-state flight condition have
been derived in considerable detail in Reference 9. For the dynamic motion
considered in this report, the incremental changes in forward velocity are
considered negligible. The system of axes are fixed in the body during the
motion with the longitudinal or X axis being parallel to the flight path
during level flight. Nose-up moments and angular displacements, velocities,
and accelerations are positive, Normal acceleration and force components are
positive downwards. Elevator displacements and hinge moments are positive
when the trailing edge is moved downwards.

The equations for the pitching and normal acceleration degrees of free-,
dom can now be derived in the form that was found most convenient for analyzing
flight-test response data,

Fz- maz (2)

where M is the sum of the aerodynamic moments about the lateral axis, pound
feet.

axis,, slug-feet2
is the aircraft moment of inertia about the lateral
is the aircraft pitching acceleration, radian/second 2

Fz is the sum of the aerodynamic forces along the vertical axis, pounds

m is the aircraft mass., slugs

2
az is the aircraft vertical acceleration, feet/second
T he important incremental variables of aircraft motion that are capable
of producing aerodynamic moments about the lateral axis are the angle of attack
(a:), the elevator deflection (J"), the pitching velocity (6), and the rate of
change of angle of attack (i). Usually, only the angle-of-attack and the
elevator deflection are considered important in producing normal accelerations.
In the higher frequencies of the range being considered, the lift resulting
from pitching velocity also becomes an appreciable part of the total lift and
is, therefore, included in the development of the equations of motion. The
lift resulting from cc has small effect on the aircraft motion but, since its
inclusion simplifies the analysis of the flight-test data, it will be retained
here.

Equations I and 2 can now be written in the form of moment and lift
coefficient equations,

WADC-TR-52-71 19
 c1~~fv~sSC
+ f~ cs + Cm 1 V2 SCQ + C 2SO*
vN (3)
d 0
CIft& 0C
SOCKJ
+LL f2. S
jVS C 02M
If2,VtS = -mgn (4)

where Cm or CL.,I i 019 d 00 or'c are the respective moment or lift co-
efficienrs or, more generally, stability derivatives.

fis the air density, slugs/feet 3

V is the true velocity, feet/second

S is the wing area, square feet

C is the mean aerodynamic chord (mac)9 feet

n is the normal acceleration, g~s

feet/second2
g is the acceleration due to gravity9 32.2

Transposing the common terms to the right hand side and letting

h = /fV2 SC2 second2 (5)

CL - 2mg/IPV 2 S
the equations become

C -M•+ C d,+ CL 0 + h~ 0"C~ (7)

o+C +C +C Or hO
Cýý+ CL +L6eC S

If the elevator deflection 4, is considered as an input forcing function',

a simultaneous algebraic solution of these equations of motion will define the
response of any variable of the aircraft motion0 This solution will. be ex-
pressed as the ratio of two polynomials in D. the differential operator, d/dt,
and corresponds to the transfer function equation as given in servomechanism
theory0

The normal acceleration, a , is related to the rate of change in the

angle between the horizontal and the tangent to the flight path. t o This
is a result of the fact that velocity is a vector and an acceleration occurs
when the direction of the velocity is changed.
Vo
az = ng = V* = V(& GO)or n = V - q) (9)
99

The equatioma can naw be expressed in operational form.

(C + C0D) c + (C - hD)O = -m/ (10)

(, +cD + 27b)o + (CL - 2) CL d

WADC=TR--52-71 19
where'r= m/, VS. second.

Solution of the operational equations by use of determinants yields the transfer

functions

9 Bo+ BID (12)

d A0 + AID + A2 D'

+ CID SCO (13)

+ AID + A2D' SA0

where A0 CICmo- Cm CLo + 271C (l4)

Al• ClaCm + 2nt 6 hC - CCLO + 2rC (15)

A2 = -hC lb - 2Th (16)

B0 CL C - C (17)

C CC C cc - 27t (19)

0
C0= C MCL C C -2TC (19)

c hCLj (20)

The transfer function for the normal acceleration is obtained by combining

equations 12 and 13o

n V Dx 2
Z-. - (T- -)• E0 + EI D + EýD
2 (21)
21
A0 + AID + A2 D Sg
where E0 = -VB0 /g = V(CmCL - CLC )/g (22)

0
V(C 0 - Bl)/g = V(C CL C
C L C Cmo + Cj Cm S-
)/g (23)

E2-= VC/g - VhCL /g (24)

The amplitude ratio and phase angle lag of any variable in response to a
sinusoidal oscillation of the elevator can be obtained from the transfer
function equations. These will be derived for the pitching rate only, as the
procedure is the same for the remaining variables. The elevator angle is
expressed as a function of time

r = Vl0 sin W t (25)

where fJlis the peak positive value of the elevator deflection in radians.

WADC-TR-52-71 20
 W is the angular velocity of the oscillation in radians per second.

Then 9 B
B+BID o2 • sin Wt (26)
A0 + A1 D + A2D

For sinusoidal motion the differential operatorg D can be replaced by iw,

where i is the imaginary unit,, V-7
Now 6 B + iWB 1 s
1) + iWA1
(A 0 - A2

The complex numbers in the numerator and denominator can be expressed as

vectors in the complex plane having an amplitude and a phase angle, that is,

A + iB M reio

wherer B (29)
0= tan-IB

Therefore T1302 +W 2121 e ol

1F A+ W A 1 • lsin
"/ wt

01 + 02) (29)
) 2+ j 222 sin (Wt-
S 7(B0 A 2

where 01 tan- 1 El
-1jB0
02 - tan _
A0 - A~kF
____

The condition that must be satisfied in order that a unique set of values
will exist for the stability derivatives in equations 7 and S is that the
coefficients of the derivatives must be linearly independent. This is expressed
mathematically by stating that the determinant of the coefficients of n equa-
tions with n unknowns shall be non-zero, The determinant of coefficients that
must be considered in this case is the same for the moment and lift equations
and is

I J2
0'c3 (50)

This determinant can be shown to be equal to zero because of the relationship

WADC- TR-52-,71 21
given by equations 9 and 9. Solving equation 9 for c gives
%ni~
Lc ýJi4 1 - 9it 4x (31)
CL L CL. 0 cJ
S UL ~ 0
CLb

3 9i + KK i
K1ni + Kdi + KK

where-whre
K,
1 = --CL -;ec; etc. i11,~
- 19 29 39~h 4-
CL

Substitution of equation 31 in A gives

(K (K2Sl)4 1 (K 3 1 )J 1 1II o
• =(Kln2"f2'42d[2
(Kln )?s'Q% + (K2'•)29•(•
(K2 J 5 )j3• +(3 (K363)V36
's2)d2G2T
2V2 23 ++():)"2K1# :•(2 (32)
3 3
(Kin~hg• (K2 J 2 )J h(
2 2 (K3 h02)( K?
(KMAXM`49. (K2c()sdi#I* (K36h4V~L'1 (KL*h)d4O!k~h
The three determinants to the right are equal to zero since each has two
columns that are in constant proportion. In the same manner the remaining
determinant can also be shown to be equal to zero by the substitution of
i-O for ni in equation 320

WADC-TR-52-71 22
 G\A
CU Uf)\0 0 E--'6 Cr\O 1,K\-t \ Zt\ 0 W M~C0
KN K
oa 0 0 0 . 0 0 0 0 0 0 .0 0 0 0 0 * 0 0 0
ON 0 - : ONK\O Y NC,-IL-NW Z NC-Kýý
L~ : w0
'CDrIr- HH CO C\j uN(jN mCUiNcuNC\jj j

~ Lg . 0O H\CQ CQ 000 O\

-a a0 0 0 .0 . . . 0 0 0 0 0 0 0

t- H t-HH~b r-i -4bo r , \10\ 1 0Ký

CY\0Lrv'o 0 wt-- - r-4H0 cuU U-Z0 CU\O.ZNOOW0

U-t-ON 0ON C\W --- \4 -: K\C\JC\Jr-4,-4 Lý I I I 0 0

(m
CU - ~ L\-ft
LC') (, O (. 0N0,\oC'\ oOa\ t-O Nt
G\ N\
0-10(jroK\L- '0 (-C~ fo 0NCC~(C~H040 V-- 8-O-1OLr

o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0000 0 0 0

E- E_ rH0C--.C H'CCCCUUHHHri

(Ii ONlK\01 OLfNý

Y 0 C- K-iU\ C~--ON0 H C\ KC'N H L(*ýO\

co \0 0 l0\0 0 10 0- \0 0z 0~ 0~ 0 0 0 0 0

H H- H- H HA 0
H

o t-- N, \0 C\J --4 C\Jt-\ýO ,0-4-1 --H- - C - )\rIHK- (

H W,0W u(N A --70 \\DrcK-%K-ýH WNUK'n f\ H
ONH t--\0 hH
Fl C-~--Lr\ CUjKN\to --I \'O\'0 Lr\ \0 R(J H- '-q0
r.zt- \.0('\ H H 0 nf 0
_ON bU.H C--ON (Y\t- H LfýH C\N'Q \.0 H H\0 K 4 O\ r-4 L(f\
ONNGr 60 w) t>-\0
Oi ONOONG\ n _t :I4 W K

r- o
0 0 ' n 0 0 0- 0- H 01 0 \0

0 \ 0e 1- ~ _:a fi f \(at- t-- w w G\ a\ al aNG\

C. 0 0 0 0 00 0 0 0 0 0 0 00 0

C)I00 0 0 0c 0, 0 0 0. 0 0 0 0 0 0 0. 0 0 0
CUj O"CkJK' -IUH K V'i'C
\D ,D C\U.4tL(LNnC>-H G\
.-Zo-Tc-i
CK~\
NS t--WO ON,0 H CU CUj :c
L(N'\.O
0
4 týb
"N'.- -Z Lf\ UNý nf'n' LV..\
\10
H a a I H H iH HH HH HH HH H H

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 04 04 0 04 0
N CH H ONw ri
U\ -: -I K'\CU CUHHH AHH

WADCO-TR-~52~-71 23
 UN 1r- - ~j 0 00T\J N o o-\ C jt . 4 Or\ \ 0 L C\,
III Ifi l Ir-ir'4r41 I hII I I

1-4\ a\tia\Cjr- .

0 H 0 0 0 0A 0 0- 0- G\ 0- 0 0- 0 0- 0 CU 0) 0- 0~ 0

0 0 0 00 0 0 0 0 0 0 0 0 0 0 0 0 0 o 0 o

I ~ ~ r- H Hr4 WHH

(R-0S 8\oN wL%0r. B - )

K1\ L0 r-1 0H SNE-L)(- H

Hi W\ t<, C~ \00 ~ I oU 0UsýTw 0m ~ 0 0 C00o 0 0

I~~ 4r- ID 04r1r1rqrIrAr

iC'r\
L(,\fN IC'g 0S ~ O ' O - 0

H l - N - \K aj -- -60 ci cm LN--%-
"D
0 0 0%
0 o. 0 0 0 0 0 0 0 0

k- -r-i' LrN-iCM
Oi mcmty- O i m H M IiIII
HH

50-NO C
10 0 K-\z(R.Zj0 O 0P I'C',I
K\60 CMKO K-1 10 CM CM
r'- 0"-C
H~ C~ \
C~j--I ON % 10,N1
1-0- 1E wN m~zNN C\1MI I- I,-, WONC-0-
H cm
HH0CN~'4> V.0 t'l M -tm o\
r 0
0 . . 0 0 0 0 0 0 0 0 0. .

0j 150_q w o 0C w0 C_1 HH C'0U\D LC M 0 CN N\

0 !ri [~--\10 -cML\ w '\E-- - 0 0 H f'NIScr
t C-4

0 0m\0 0 0r \ 0 0 011 0r 0 0 0 0' 0 0- 0\ 0: 0 0 0

4." 0 H CMK
tc~LC"'\ E-O ONC\O H CM K-\U0 -~ OOHC
Pý ý N -4 -1 r1 r4 ' H, N aM
CMC

WADC-TR--52-71 24
 N &WMO Z 0 1,0 -4 W K Ot H --I w 0 H

UN ~~~ C\ ~CA)-w% 0 \
Cy\ r4j K- 0 "
\0 0- 0 0r 0~ 05 N 0 011 0 0N o
r--IM Cm 01 cQ
W~~ IP7t
El I HH~
1A O,-:1-:fIo \Z 6 'O oc~-t-o
r -0C-0-*-r-O -1 0
0
H

(\J~~~ 3- 0 0 0 0
N~
H~goC0Pco\
0 0 0
0
30 . . 0a 0 0 0 0 0 0

CCM
-o w 0 K-\S C rr-0 H tc
_4 U-, N, C3

CM~~Hfl CM c~O- 0N4c

o 0 o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
H- \ r.-zt
wN -1\D I-- (3\ H U\O N r-4 ON UN CM I. 0 0" 0

r- H CO0K' \0p0 E \O o
rA ID\
W
LIN \K 0 NN N
0o~ HUHP PC\\WCC\o~ K-

0- 60 ý% s0 f-
+3 CM Lf\' C 0H MH 0 '0 H-
oo . 0 0 0 0 0 0 0 0 0 0 0 0 0

w 10OO ýrUNHLr%
1\ O\ N0 NfCt,&
UN
CMCH~ N 18 -'OC OLr{'ý 0
CM
Cr) 0- 0o "0 0 0 0 0 0~ 0 0N N
0\ 0 0 0 0 Ht0

u 0 0 0 0 0 1

Hf H H CMCMCMCM N'\ LC'

\ HC' " H
U 0"0
N H 0"
O
n 0N

0O\Lf0 -
4
0 t- 0w\ ~-cy\t 0WLfH
H wcQ t- L% KI
v-I
0 ~ t-v
w
0CQ tK- cý
0
0\U\
0 0 0
A ~tow0
00 0 0
HNK,CMj
H2 H
0
H
0 0
0 0\tt-I0--'0 0 0
0 0 0 010

H opCM. t~~L 0 0'~ 0 0\ C tc

H 000 .\4 ý t a 0(ý0H C
r- i(X
ri HHHHHHHHHP,\Cc
-:IUN MN\0'6
b M

W~~~nG- IT-5-712
 Q~ - Cm U\ 0 \0 0\o\ E- Lf sC W E,-o0 0
ROto0 LrAU O ~
cO0C\ El 0I 0r

CMJ0
3 -0'
0

r-4
0 0 0
mT0rIr
o
0

HCJC4r-,
0

\ ..4
0 0 0 0 0 0
o
LfV.O'
0 0
\OI--
0

t- ON G\ HCcm K%
0
(M
o-_:
0 0
Lrl%
n
0

0 1 HH- r4 0

ft'J\TG nr*C0 r--

HCQC-CQ0 cqo00HU'.N H
cl0 P 0- 0, 0ý 0m 0, 0ma- 0 0 0- 0 0 0 0z 00 0 0

00 0 0
fN 0 (,0 - ) Y0 r- 0 t- 0 o , w H 0 0 0( E- 0
C)I , r4w r.HHc\J w nC\Jo - H 0

00\ f~j a~<jHtN K0N t- \0LN ONWN0HHr-qOaH L

.y\ \D0 K\ K\\J
o44O' N W - ~ 4t
KO's. --
OH \0 0
IC~L'\t-- H
0003 0 0 0 0 0 01 0 0 0 0 0 0 0 0 0 0

r403 IrI 0 r mKN \-: III In - If' \ 1

J0 \J 0 Cij

00

00c 0 0' 0 00 S 0 0 0 00 0j H 0~ 0 0 0 . 0t

W -4 Lc,%r-i
Hy Y , Lfr\C~j0\rNLCC
LPA %LC\ EC
- <I\ 0f Gt\I-L Y

0 0 0 0 0 0 0 0 0 0 0 0 o 0 o . 0 0 0 . 0 0

I rI f-l -I -I H r H H H- H H H H Hi

El CN\ cun -b :tO\0--bO boPC\

"o N b
C~ ~ \HHM-EHHHCHH0G HHHH
HHHHHHHHH-4 K0\t
00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 00

rI fo\ W, INK \ O 0
H00 \
bCbjt- fkC-- G-\\04 br- - K ON\
CiM~CN-Ob H ~OK- MfNýLrN~\
L- o Cis-\U-\U-\ LC o \HNt
LN L HNL'

n3UN r t-- 0N 0\ rO4(\j O f.\ (H\\0 '6

0 0 0 0 0 0 000
0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nm-Hr- t--NMI r-4O r-Or-\O
r--b HC\J4 t

40 H~r~-4 ONO \ \j-ztw

' H 04 0
(\jIV, 010
-b 0 H r
Ci\
H r- HH
H r HH i-i H H-i H r-4HHH riiH

WA0T-27 26 .. 0 o*
 'o Lu- WNl- r ir*- CUlN M~\~k
'o ~ ~ r t O~-- l r-r-CI
+-4ON 1O~~ - N 0 tlOON (\ E- 'Klf\MCuHHH4C- C4O
N 0i 0~ 0 0 -0 P 0 0 0 0 01

O OL,\\O
ONu U\0 C'-C\ t--CQ LCrHH Hr-A r-I r-HH 0N1
+ W~uH c l H - N0 (\L' l HOA
-C- ONt
oND0 'OO4uuNt-4Cu00000 P CO

Nu CUjCu Cu H rir-

- C' 0
rUNOC- - L(,.N)cu QLC\NQfO iWE- UN f
ONUH u 4cu W'
w"400 NNCJý (-0-& r-- ~\0'- U' C
-1'0\ K.5OCuiN\-., N t--K-N\ -- 0 HONONC),Lr C' -LC- ý
H CUjCU Cuj\0\0 LCA
-- I--\0 H.C (MONw\0"LS-\

0 r-I C\J WN\H 15? W)COIN

t--0E-- 'WOHN\N3 UPC\
I- I-INN \ 0rl T

0 0m 07, 0\ r 0 0 0- H 0r- 0 0.:1 0 0 (I 0- D 0 10 0 0N 0\

HC 0 o \0 0 UN"0 ir04
0t- 0 Cu HII\ 0.t=uO
a 0
IH\D[- '\Lf\-OONUý\ Lr\
I- K" Df\t-(\HK
~ C-0 CUcýGu ULi'N\0 O 0-4j N
IH 0H~~
H ýUw1- Cuj Lf'\ N 41A 0\

~
f O-:To j ON KN4~t-~ \JD~-
K%~~tCuj
\0 '- ON r-4 O0\ - uN C7\ -NHLCu0CoLrA aHj K\
OH -ztQuwtbo -4
o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

H- H- H- H H Cuj Cu Cu Cu
Cuj CuCu NN C %N.J
I l\ n

0\O0 (3 ýj \Dc 5 \D 0
-o6O0 \I Lc~--0 --.
400 L--
U0,0CjC\
-6"0O ~ 0uL\ NH W \O~H'NN'
rl %
O\UouuONoO Oj Nq\0KCl ONO c-H~Z H-
0 0 0 0 a 0 0 0 0 0 0 . 0 0 0 . 0 0 0
KN Cuj 9hNV-K,0-1 KNL' CU ý bO El-O b 160
KN cQ 0SNI\ 0

rl61
W \O IN WCuM" 0
H H E-Cu \fCK\L\O Q

-4 H8ý.,Ou~H-'OCW0t- W\ N\
0 i
Cu-7C H
Cu ~ 0 0 0 0 .0 0 0 0
!4 ri 0
0 0 0 0 0 0 0
H
C\JCuJCuQuK% n C1 CC ' O uHl0--1 N
Cuj \0 $~c ~ONW f\C.--! \0 \0 k>-- '-0
X ýil r% L- c-\ r~- 4

P14z H Hr-H HHHHH ICUUNCN

WADC-TR-52-71 27
 00D o a',\H LVC'C~jc-. t- CUO tN
m~~~~c '~0 to~~O~~
0 ~~ * a 0 0 0 0 0 0 a 0 0 0 0 0 0 0 0 0
3- HHr4HHHA -IH
HHHH r-IO-

rA t-- C'oO0 Rý j0 Lf\ H 0'\ LCý -- C- R' La'0 G\

.4iCr-4oOr
L(-N \ý -7t H uý- KVQ 6
i !j a' SQo r H tq-\Z 0 C)w C~ Lf\rt--. (\J LN or\ LfN
0 . ~ 0 0
a tr0rC 0

0 H

El- w t !ia\(j -LIW a"-o O-

Q) w%~r<,
CUjKN~ K-\ CUjCj H4 H

O 00 HU Cj rt WKI H N\t- l N KI CU

o . 0 0 0 . 0 0 0 0 0 0 0 0
bO V- Lf,%tqC\ r I H CJHUCXJ
NH r-4 H HH r-4 -I

"-4.
U~j~'\
--r<-
a',
c0
Hi rt~-.z6
t- a' 6- Lr''.aý-I
_: w ' 0 CUjH Hsr.- ON
0\ 0x~~'
U'N\\ n-'(bvUNr ic'\
r-a-.r--- t0
LCI
C--K\%
w WKNNtC- -H4.\0 r4 H H -ZI H '60 00
0i 0e -0 0M 0*I 0 0 r0 -Z 0 0 0 00 0~ 0~ 0- 0- 0- 0 0- 0

r- H H

K\'~ 'Wu _zt 0t--\0o 0 'CtC\Hj H H 0 00 CU\

0 0 0 00 0 0 0 0 0 0o

I
-- YIp T--
H I
- ~ - II 0r
I I 1 ti IC

w NH KN Ke\Q _:t.0 iC\ % r t-- cl0-w '\O H Lfl

H U - ()N \ '. -Lf\ "0 cv\,OK0
HC\K LCIN4.4wo0-4
Ur\a

0o . 0 0 0 0 0 0 0 0 0 0

Q~ :~t0E-
C~0 tC-- r- 4
ý - 4IC -:
.H
NW
\0Hri\0
~ -
c)-\
,

00 V'N 0 0 0 o 0 0 0 0 0 0 0 0 0 0 0 0
0 0 O0
(% K-% 0 \O.Uc
--I C\J \H lI I 12 1 i

WADC-TR-52-71 2
 0 \ 0r
0 0 0 0 0 0 0<, 0~ 0i H
0i 0 0 0 \ 0 0 0 , 0

B e - N4 - s~OO
o PC,PC~N
YO KIN
L K ND ) - \ O O-\Cj w t--
N4---z~~' r-

0 0 0 0 0 0 0 00 00 000 0 0 0 0 0 0

.4.4 H w r1H
0 t-- KH I '0 6 - ýOt' 4 r ,8

+r C\ N10C~ 0U U GN E-.l- LI --j [-U-~ N ~j Cod

MoW oojcO
Kr-W w w y 6 \_' 0\0Cjt vt
U T+N-- CJ \0 wtofr
C~" NtI~.r) \Ir K K\Db LC--H
rl\ N o 00\DLr
0 ~ NUN 0-0-~--;-;0;;
W ~ CT 0 0(' 0 0 0- 0- 0- 0 0 0 0 00 0

r-4~ 4OH
r-A N~ Co.,~ H~N Hr-4 NO

\0 ~
\0ON- 11 y\D\)
1 L\ ~' o0 : ý~

ICij t-- K's CM t-V~K-060 t-4.OCf4'\O 0c~CoOlc

C\0fo 0j 0 0 _0: 0 -0 0 0 0l 0\ 16 00 0(' 0- 0 H 0 0 0 "

CMoHH a oJ H o I . o oo o.

CM)

.40r- cr .4'.OU ON
xIUN z\O-
C-K WNr\q
00W\ \
0k H -4 '\ 0 0t
H+ 00 7'u~ KN tC- W\boO \JK"HCMj rlfN\-LfN'~C\Jt -o -It -K
Q.Jm. ON-

m 0 0r, 0c 0 0- 0- 0:: 0~ 0~ 0- 0 1 0 0 0 0 0 0 0 0 0
VZO (\j r-4 o~j311 MO re CMAH ON

.4 O\ aN RH0 tQ
O OU\ \
06 .4ý r- -
LI& +~ ONý VNC--OtCM !l,-) N C-- 0 CAvON--I
ON\D
I s 0--0 0c 0ý 0- G\0
0N 0 0f 01 0t- 0" 0 t- 0 0 0j
\ 0 P

.4 > O ý --
N'- zt-
Nr o-j4() E--OO~C t-b 0~ - -

0K C~ "'\O t` w1 rt-\O'NHU \D . ['

$,'r ~-O
0\OK O\YOO to-\rA
-4
+ 9\0 ON S.O r-OtLNU.NCC
w I
+ 0, 0 OA0 0 0 q0 '0O0 0 0O GO 600r 0f, 0 ~ N 0 0 0
CMC 4 r-A HHN - Ný oH I ~- N

oI\0NU, -f -.IN cNi

o.4c% 0N0 -

0+,
0 HMC.Lf fV.O0 - CiOON -A0 \0HCM LvO\_0 tr--bO
ýe(\4 0 0H C

Hr-4
HHHrI H H H \ CMCMCN

WADC-TiR-52-71 29
 C'JCM
A Cuj
HICu m Cuj

4)~C
oc Cd C u
000
H~~~ C :HrIC~

H- CM , -4
_z

C~Cu

cuy Cu
Cu~
H- I ~

11" aU1 I I a 4

N N t.

H~u

r-i~c~ NC ~ f 0\

I- I- H

H~ ~~ d \U Hl
HOCO Cd~ CudC
0K

HNKNH( 0

0 H+ + + C -j Nc

coIC HC (\j N CM \ Cu

HC\J "1 X:-4 -4 0

+ + + + ~ -
H
H HH H HINCH
H CH
r,H-4 H- _q x H Cujc PI\C'0c
.H K\ ' - 0 ___4

WADC-TR-52-71 30
 TABLE 5
SAMPLE SOLUTION OF NORMALIZED EQUAITONS FOR STABILITY DERIVATIVES

Columns Summed in Table I to Obtain Augmented Matrix.

G (9 G
(See Theory, Equation 19)

Augmented Matrix:

20o72395 -2o10654 17.789218 107.27281 -13-30576

-2o18654 22 -43o35784 -4o65182 -18o33S20
*17 o89218 -4-555784 169.81986 91o512)2 5°79131
Auxiliary Matrix:
2072355 -o,10l65 o 6337 5o17634 -. 64206
-2°l0654 21 °78587 -1I9o67o .28699 -- 90393
170 S921 -hlo55912 75°16967 ol4125 -.o26959
Solutions from Auxiliary Matrix

C = o14i25
CL = 29699 - ol125 x (-1,90670) = o55631
CL 5.17654 - l1425 x .86357 - .55651 x (-o10165) 5ll094
Cm = -. 26959
Cm -°90383- (-,26959) x (-1.90670) = -1h41786
C -o642o6 (-,26959) x 96337 - (-1,4178>6) x (-.10165) -o55342

If C is assumed to = 0, then:

CL - °28699
w
Cý, 5o17634 - .29699 x (-.10165) 5o20551
If CL6 and CLa are assumed to = 0, then.

c 5ol7634

WADC-TR-52-71 31
 TABLE 4

SAMPLE SOLUTION FOR THE PITCHING VELOCITY TRANSFER FUNCTION CONSTANTS

Columns Summed in Table 1 to Obtain Augmented Matrix:

m 0 G -G -
0 -® -0 -0

(-@ 0
-O -@ 0
(Reference Theory, Equation 24)

Augmented Matrix:

22 0 100.54611 30.15384 295.41151

0 679.0712 295.4l151 -l00o 54611 -4690.1187
100.54611 295.o41151 960.15451 0 0
30o15384 -100.54611 0 88. 81077 960.15451

Auxiliary Matrix:

22 0 14.57027 1.37062 13.142779

0 679.'0712 .143502 -. 114806 -6.90666
100.54611 295.41151 372.12172 -. 25279 1.851475
30.15384 -100.54611 -94.07162 8.811409 4.-00456

Solutions from Auxiliary Matrix:

K0 = 4.00456

A = 1.85475 - 4.00456 x (-.25279) = 2.86706

S= -6.9o666 - 4.00456 x (-.14906) - 2.86706 x .43502 = -7.56097

1

0 = 13,42779 - 4.00456 x 1.37062 - 2.86706 x 4.57027 -5.16418

0

WADC-TR-52-71 32
 TABLE 5

SAMPLE SOLUTION FOR ANGLE OF ATTACK TRANSFER FUNlTION CONSTANTS

Columns Summed in Table 1 to Obtain Augmented Matrix-

S0 -0 -0 -
0 0 -G _0-
-6 -0 0
-0 0

(See Theory, Equation 25)

Augmented Matrix;

22 0 29-3W224 2ol0654 -72.9769o

0 679o0712 -72.97690 -29-345224 -411•54432
29-3L224 -72MS7690 753o66311 o o
2.1i0654 -29ý3.2214 0 20.723S5 73 o66311

Auxiliary Matrix:

22 0 1o33373 .09575 -353125S

0 679o0712 -. 10731 -04352o - . 6o6oL
290o5b224 -72-97690 26o7o006 -°22307 1o9S562
2.,lO654 -293L.224 - 595927 17o92545 4o 16667

Solutions from Auxiliary Matrix:

10 h4,16667

A, = 1.99562 - (-.22307) x 4-16667 2o915OS

1 = -60604 - (-°04320) x 4-16667.- (-o10731) x 2.91508 -. 11522

co = -3o31258 - ,09575 x 4°16667 - 1o33373 x 2o9150o -7o59947

WADC-TR-52-71 33
 TABLE 6

SAMPLE SOLUTION FOR NORKAL ACCELERATION TRANSFER FUNCTION CONSTANTS

Columns Summed in Table 1 to Obtain Augmented Matrix:

M. 0 -0 (2 -0 -

0-0 -0 0 -0

-0 -0 0 0 9 0
0 -0 0 5 -@ o
_@ 0 G -0 Q G
(Reference Theory, Equation 26)

Augmented Matrix:

22 0 -6°91452 -227.21809 -679o0712 955o47050

0 679o0712 227.21809 955°47050 0 4195o8905
-6.91 4 52 227.21809 107041754 0 -955o47050 4941.9625
-227.2109 955-47050 0 4941.9625 4195o 905 0
-679o0712 0 -955-47050 4195o8905 49370o050 -63366.258

Auxiliary Matrix:

22 0 -,531429 -10o32809 -30.96687 43o43047

0 679o0712 o33460 140702 0 6.17886
-6.91452 227.21809 991o97420 -. o39427 -1o17835 3.86936
-227°2109 955447050 -391.1120 I096o66347 -2.98951 4o99499
-697o0712 0 -1169o89579 -327To47851 16230o9349 -o79940

Solutions from Auxiliary Matrix:

E2 - o7994o

A, 2.,6051l

A0 = 3o95453

R1 = lo19013

9 609ohgo
0
WADC-TR-52-71 514
 TABLE 7

TRANSFER FUNCTION CONSTANTS COMPUTED

FROM FLIGHT DATA

Parameters Obtained From Flight Test Datag

2
have = .190 sec Cm -o555 C• 5o11

7-ave = 3.O0 sec C L 556

Vave - 265 ft/sec cmo= -270 sec CL; .l40 sec

g = 32.2 ft/sec2 Cma = -. 122 sec C °063 sec

Transfer Function Constants Calculated from Flight Test Parameters-

A2 = -hC .&2 hlu= -l1 0 91 sec

(CL Cm. Cm CL0 + 2iCm)/A2 x 4°25 4 See-2

10 -
a*Ca o
Lac~
( CmC + 2T-C . hCL CmoCLo + 2"'C,/A a 2.998 sec
2 (Cc -C C)/ , 2

07
0 (c-m -CICm)/A2 =-6o359sec

B1 = (CmC C c.C - 2 Cm)/A2 ,-:7816 sec

(Cm C CC - 2 rm )/A 2 - -7o752 sec-2

U0

2I hCL /A 2 -,•.092 ec-1

2
= V(CMYCIo C CmC) 2 a 52<334 sec
to

V(CmdCL• CLCmQ + CL C + C Cm )/A 2 0o527 sec- 1

ý2 VhCL /gA 2 - •0o755

WADC-TR-52-71 35
 TABLE 8

TRANSFER FUN TEON CONSTANTS FROM FLIGHT TEST DATA

2)
f1D)/(Ko + -1D + D)
(fro +

From Data Points From Data Points From Stability

1 through 22 1 through 17 Derivatives , Table 7

Ao 4o005 3o84O 4.234

A1 2°867 2.917 2o998

B-0 =5.164 -5.116 -6o359

B -7561 -7-777 =7.816

Ot, (C +UD)/(/o + 1KD +D2)

i h4167 4o174
0
Al 2o915 2.959
0 -7°599 -7-732 -7-752

01 -00113 =0.122 -0,092

n/- ( 2 2)
(to 1 D+ D)/(A0 +A1D )

A 3-954 4o189
0
A1 2.,605 2.99
2 06°9o5 51.211 52o354
S.lo90 0°L35 0.527

E2 -0-799 -o.655 -0o755

WADC=TR=52-71 36
WA -T 37
-5-7
 4 H

.44-

r++

tt

hill I
:m MM-1 f m. I
+ I MH
4 tmff-t: wt

44-f-111Tmý IF.

all$

g g fM Mis TT
T T
+ IT
®r
tMI, MIM
EM

I iH
4+

NOR

filml Iff 9i N.mmm

V41

41 .. .

M Mv fim: +
4 1
A ým
Ira H 44 4 4T T-:P:I+

#69
T; tl-+t
+1ý11t
1; + ......
4
w

P11 4ý

Rin

Ti
194

tT

H L T tt: t 1 Z±U1 ma. I EH

IM O-M!#4 4-'W't"UM, tt if
7 44t 4414- 4"

ig.
1+ 44414
4MM

21 ,Ml
w"t N111-T
IRIMIIIIZIP, I f, i ll, a
ifl
44

41

ILL L

WAD-C-TR-52-71 39
 I ~I

-';,- i - -~I" T. .. '

14 I.

.... .....-. 77.

K: ~ 7rVt
-~rtt i

A.:.... 1 . .. f ..
Ft - TJt'
4

-4~~~ Oil 7i :..*:.

i Ji
4L -7-H-j;

WALE-T-52A7 -!*A9
 T
:I.:
wt 41r,

.......
. . :z

Olt

t,
tW

i7
f

Itt
W-0 ýT,
7

JL.;-

4.1; ý4-ý ý4 .. . 'Imp :z.-.

-5 ' ...... .
..

Tr
4 44ý

4j-
jr

L
11

.114

4 1,J14
144-
7T=r

44, 'r

llf + 1H

ff,-, AR 4j;

fi% ttý 4r

M 4 Ill
4f- Mr
14W ill it-

MW

.rtfl
41
41
.14

ýWADC-TR-52-71 40
 i7j jji77 -,77177
-V
42~47~1VL7Ij
h
j44L~+
_71,
yfi7 47

H-V7 7 -4: -7i- t

7_ -j IL -i~I1
:- 7.
4-I- ± J-p~
11j774 - * i1 j.'
7 "tT

ifi77i};f;; .1.

7"-7T~
t7- t
T
I',~~~~7 -77i--.i~-~I
V77~ ii..Ii+2....J iw...
.
.I

t1 :JAt
-4
-'I.-7 ,

~ __ ______________L~LLLiL iiq

527 aJD TR
 F - -,7 -71 ,7 =1 4----

I4 L_ 4

K~7
I i -pCý~{

i _7 7-

I7 7

I~ Tf

Ft-_ ..777
7T7

r 7

- ~ ~ I"it
:1 ~ F 40

L- 7 11

--. -- -T 1 r7+

L ~~ -IC)9I ~ ~ LLA

'1:-

A -C T;- -j
 __ 2'.M

Il TH
I :f

Ptf aft. W I

44
I41
FITA

i4 -fl- -ff

44"4

~ 4M;
V;A4N
1. -

JAW TRo2711
 17 'F E., 1,1- li!j
TIM

TF
I I
it ....

Fýrý 177 L

.. 7 TI; il 1
11
4"t Mt T419-321 I
C, V:
t! '4qd; I 114, 1 +

7i
T

"1"
444 i 4'l 4 frj U 4,44,
r! . .....
t ji t
ff 4 ýTj
4+1
'it
!j 1, 411 1t
ti tIA- 'Hi
4

ý'Pl 4- 4"
jrý III t. 1
INCH M
t v
L6 t H- pi I
f

4 AA
90, Rký
# .
.. ..
i _,11T i 1, 4- _11Vt

+ ... T m :1

1 44-

#
+
I Ar M.-
ti I Mwýma gl
r tT

4f,,Wif: +
I +i+
+
7- + +

Tý 11`71- T-1

4 44 +

ýT
IT
1:1
Ur MU:
M m 441 7l; 4--,-rq

ý1,
ý4
41t
M V,

_ff"t_1mq#W-V 4 p T

Tf

Hý ý4ý4 ++ iý,,: -
if,; 4 T T_; W IM
t,
41h i

;;I- tiýýj
iI I; TM , 'N -4 -1
+ 4i 4-+
XX i

7r

TR-C-2-71 44
 1 4 -
1q, 11
I` IMM
1.'ý

INN Ti i'vT ýýI:F

T71
r
ciii IMPIT 31 14 +

RT TV
T,
7
11
IN 114, jj 14 tt I RJ'ilf tit i
Ij
..
til itk
j. i
. .11 1 , -11- 1 J,4
4

ttttttt
4

1 11ffl
p
417

4 4 Wtý I
if
TrI,
ýj4
AV 91
-4 Tflll
#fff

'w,

ý4: AM ýd
I- z
14V trRt
lN

itfill 1;i-T111
It ig: 1:1
-F,
i-I
r, HU T P-- :l4 I1 4114:1
"L4 4,
444

...........
PIT -- :1 :

I-iq
#i
I - ; :-, -ý,
T71 tR
TH 14
ý4 t V,
HIM
Tt TT
I - - - ir -, .11 4, 1
4t ft 1 ' - :
A f Ila
'17 1:4 1TO,4 J,ii- 41
Mq 4,1
Tlý 11ý ll W.
S M 44 --
-
r it 4-
Jý 5 WN Ti 1,111
4111 IT :4
_4
+ +
t+4
;TIT
14 -:T4ý - 4 iV
z, 41 Cý 41
;7ý

f", Ht +I 41
+4f i+

4ý t4t1i .

Mlý VT
7 F.'t :ý .. ;T4

,-H 11,11' ýIf It

L - - 1
ýr 'Ll tt
rý 4 -4
Ti
M;-

44 ir::I
77
4_1
, ; T - fH
+ 111t-ý -1
ELI -Itul-Tri ITT

WAX-TR-52--fl 45
 REFERENC ES

1. Schumacher, L, E., Methods for Analyzing Transient Flight Data

to Obtain Aircraft Freqiincy Response
USAF Flight Test Division MR No. 1ACRFT-2268
2. Scarborough, J. B.,9 "Numerical Mathematical Analysis", January
17, 1950, Edwards~, Ann Arbor, 1947

3. Phillips~, H. B., Vector Analysis, J, Wiley and Sons, Inc., New York

4. Perkins, C.D., Hage, R.R. Airplane Performance Stability and

Control, J. Wiley and Sons, Inc., New York

5. Wener, N. L., Oscillation Method of Stability Flight Testing,

USAF Flight Test Division MR No. MCEFT-2253

6. Milne, W. R., Numerical Calculus, Princeton University Press.,

Princeton, N. J., 1949

7, Scanlan., R. R. and 'Rosenbaum, R.I.

0 Introduction to the Study of
Aircraft Vibration and Flutter. Macmillan, New York

9. Campbell, G. F., Whitcomb.. B.W., and Brouhaus, W.O.,, Dynamic

Longitudinal Stability and Control Flight Tests of a B-25J Airplane,
TB-405-F-3, Cornell Aeronautical Laboratory, Buffalo~, New York, 1947

9. Durand, W. F. Aerodynamic Theory, Volume V, Durand Reprinting

Comm, C.I.T.9 Pasa.dena, California

WADCTR-52-71 46