NATIONAL ADVISORY COMMITTEE

,
FOR AERONAUTICS

TECHNICAL NOTE

No. 1670

APKYCIATION AND PREDICTION

OF FLYING QUALITIES
I
By William H. Ph+ps ..

Langley Aeronautical Laborato

Langley Field, Va.

Washington
August 1948
 - -

NACA TN No. 1670

AERl3XATION AND PREDICTION
OFFLXlXGQUALITIE3
By William H. Phillips
August 1948

Page 14, line 16 of first paragraph: (x/c negative) should be changed

to (x/c positive).
Page 14, third line from bottom: The equation % = -0.05 should be
changed to E = 0.05.
page 38: In the two equations, Cn should be changed to CN.
Page 70: The symbol C, and its definJtion should be added aft&r C
as follows: mo

% normal-force coefficient
(
norm31 force
SS >
Figure 8: In the expression under the curve "Neutral stick-free
stability for static margin of - .O~C,~ the value -.Ogc should be
changed to .05c.
 TABLE OF COl!iTERTS

Page
SUMMA.RY ........................... 1
IlYTRODUCTIOl? ...................... 1
LOBGITUDIXAL STABILITY ABD CONTROL
CHARACTERISTICS IB STRAIGHT
FLIGHT .......................... 2
~C-csIE~IGmmEr ........ 2
Requirements e&Definitions ............... 2
Methods of Obtaining Static Longitudinal Stability ....
Dynmic Longitudinal Stability .............. %
l!zce!EcTsmPRo~o~oHmpawwoHsTABIIsTy .... 7
single4zngineAirp1anes .................
MultiengineAlrplanes .................. ;
Jet--PropeUedAirplanes ................. 10
coI!?rRoL cHAR4cmcs IH sfJT&imv .......... 10
-10x OF -PO~FROM~GEC~TS ...... 16
Stick4Mxed Neutral Point- ................ 16
Stick4Zree Neutral Point . ................ 16

Ei3TxzsoFc~LBILITyo~ TRIMmD8TABILITJT . . . . . . 17
Effects of Compressibility onVarious Airplane
Components . . . :. . . . . . . . . . . . . . . . . . 17
Examples of Effects of Ccnnpressibility ..........
Reasons for Compressibility Effects ............ 3
Div&ecoveryFlaps ................... 19
EEVECTSOFSTRlXTURAL MID COKI'ROI&URF~ DISTORTIOlV OR
L0IWxmuD~STABILITY.................. 19
LolQxEruD~TRlMc~mTO~mFLAps . . . . . . 20
Requirement . . . . . . . . . . . . . . . . . . . . . . . 20
Reason for Trim Change with Flap end Power Condition . . 20
LAJYDmmT- CVICS . . . . . . . ...*. 21
Requirement for Landing Characteristics . . . . . . . . . 21
Requirements for Take-Off Characteristics . . . . . . . . 21
M Discussion of Ground Effect . . . . . . . . . . . . . . . 21
I
LOWGITUDIlVAL STABILITY AND CONTROL
CHARACTERISTICS IN ACCELERATED
l d
FLIGHT . . . . . . . . . . . . . . . . . . . . . . . . . 22
Relations between Longitudinal Stability in Straight
and in Accelerated Flight . . . . . . . . . . . . . . . 22

i
 Calculation of Stick Forces in Accelerated Flight ; . . . . 25
Effects of pitching the whole airplane to a higher
angleofattack................... 25
Effects of curvature of the flight path . . . . . . . . 26

Discussion of Factors Influencing Stick Forces in

Accelerated Flight . . . . . . . . . . . . . . . . . . . 26
Distinction between Turns end Pull-Ups . . . . . . . . . . 26
Requirements for Elevator Control in Accelerated Flight . . 27
Examples of Stick Force in Accelerated Flight on Different
~pesofAirpl.anes................... 28
Means of Obtaining Satisfactory Elevator Con*01 Forces
insteadybuvere . . . . . . . . . . . . . . . . . . 28
Stick Forces in Rapid Pull--Ups . . . . . . . . . . . . . . 29
DISCUSSION OF TYPES OF CONTROL-
SURFACE BALAXCE.................. 30

Importance of ControlSurface Balance ........... 30

PlainControlSurface ................... 30
Balance Characteristics .................. 30
Overhanging or inset--hinge balance. .......... 30
Unshielded horn balance ........ ; ....... 31
Balhcingtab ..................... 31
Beveled-trailing1-idge balance . ; ........... 31
Sealedinternalbalance ................ 31
Othertypes of corrtroleflurface balance. ........ 31
Ccmparison of Various Balmcin@; Devices .......... 32
DIRECTIONAL STABILITY AND CONTROL
CHARACTERISTICS . . . . . . . . . . . . . . . . . 33

DlRFCTIOEKLTRJNC BICS . . . . . . . . . . . . . .
l 33
Req,uiremente . . . . . . . . . . . . . . . . . . . . . . . 33
Directional Trim Characteristics for Sizgle4ngine
Airplanes . . . . . . . . . . . . . . . . . . . . . . . 33

CHARAcTERIsTIcsm-s~m . . . . . . . . l . . . . 34
Requirements . . . . . . . . . . . . . . . . . . . . . . . 34
Directional stability and control characteristics
in sideslips . . . . . . . . . . . . . . . . . . . . 34
Pitching moment due to sideslip . . . . . . . . . . . . 35
Side-force characteristics . . . . . . . . . . . . . . . 35
Discussion of.Equilibrium of an Airplane in a Steady .
Sideslip........................
Typical Deficiencies in Sideslip Characteristics . . . . . ;z
Contributions of Various Airplane Components to the L -
Directional Stability . . . . . . . . . . . . . . . . . 37

ii
 Page 6
Directional stability of the fuselage . . . . . . . . . . 37 ‘-
Propelleryawingmoments . . . . . . . . . . . . . . . . 38 0,

W~ya~momente.................... 38
Yawing moments from the vertical tail . . . . . . . . . . 38
Design Considerations for Prevention of Rudder Lock . . . . 39
DihedralEffect . . . . . . . . . . . . . . . . . . . . . . 40
Requirements . . . . . . . . . . . . . . . . . . . . . . 40
Definition of effective dihedral . . . . . . . . . . . . 40
Influence of wing location, power, and sweepback on
effective dihedral . . . . . . . . . . . . . . . . . . 41
Measurement of effective dihedral. in fli@ . . . . . . . 42 ,
AILERON CONTROL CHARACTERISTICS . . . . 42

. Requirements for Satisfactory Aileron Control ....... 42

Typical Aileron Control Cheracteristics .......... 44
Calculation of Rolling Effectiveness ............ 45
Amount of Aileron Balance Required for Satisfactory
Cheracteristics ..................... 46
ktes on Aileron Balance, Wise ailerons, and spoilers ... 48
Adverse AileronYaw .................... 49
Requirement for Limits of Yaw due to Ailerons ....... 50 L
Rolling Maneuvers in Accelerated Flight .......... 51 -

STALLIBG CHARACTERISTICS........... 51
Requirements for Satisfactory Stalling Characteristics . . . 51
D~SCI.WS~OII 0f Typical Characteristics
stdin@s . . . . . . . 52
Influence of Various Design Factors on Stalling
Cheracteristics . . . . . . . . . . . . . . . . . . . . . 52
Fli@zt Conditions LeadinS to Inadvertent Stallin@; . . . . . 54
GroundLooping....................... 55

COETROL -FREE STABILITY OR SHORT-

PERIOD OSCILLATIONS . . . . . . . . . . . . . . 55

Requirements for Longitudinal Motion . . . . . . . . . . . . 55

Influence of Design Factors on Short4Feriod Longitudinal
oscillations . . . . . . . . . . . . . . 56
Requirements for iaieiai. &&x. . . . . 56
Influence of Deei& Factors on La&&i. ~&i&t-,~o~. 56
Relation between Rudder, Aileron, and Elevator Short-GeGild'
Oscillations . . . . . . . . . . . . . . . . . . . . . . 57

WIRD -TURIiEL TESTS AND CALCULATIOB

PROCEDURES FOR DETERMIIATIOE OF
FLYIHG QUALITIES . . . . . . . . . . . . . . . . . 58
JXTRODTJCTIOM.......................... 58
...
III
 SIMCL&TIONOFpawERCO~~OmS..;....- ......... 58
..
Criterions of Similitude ................. 58
Variation of Thrust in Flight ............... 59
Calculation of the Vmiation of Thrust Coefficient with
Lift Coefficient for aSpecific Airplane. .......
Selection of Model Propeller Blade Angle ......... 2:
Preparation of Operating Chart8 .............. 62
Simtiationof.PropellesIdlingCotition ......... 63

~TUKNEZTEsTSFaR~INGAKDTAm4FFCHARAC!TmISTICS.. 63
Wind&Funnel Tests Employing a Ground Board . . . . . . . . 63
Simulation of Power for Take--Off Condition . . . . .
Wind-Tunnel Test Procedure for Talrs-Off cond~t~o;l . . . . . 2
Computation of Ground-Reaction &%mmt~ . . . . . . . . . . 64
Tricycle landing gear . . . . . . . . . . . . . . . .: 64
Conventional landing gear . . . . . . . . . . . . . . . 65

~IONOF~POINTS . . . . . . . . . . . . . . . 65
Stick4FixedNeutral Point. . . . . . . . . . . . . . . . . 65
Stick-Free Neutral Point . . . . . . . . . . . . . . . . . 68
J CONCLUDING REMARKS CONCERNING
SELECTION OF AIRPLABE
CONFIGURATION TO SATISFY T.HE
I FLYING QUALITIES REQUIREMENTS . ..a 68

APPENDIX-SYME3OLS..................... 70

REFERENCES........................ 74

TABLE I.. ......................... 77

TABLE II ........................... 78

FIGURES............................ 79

iv
 NKI!IoNALADvIsoRYc~ FOR AER0NADr1cs

TECHNICAL NCTE NO. 1670

APPEEXIATIONAND E3KDICTION

By William H. Phillips

SUMMARY

The material given in this paper summar izes scme of the results of
recent research that will aid the designers of an airplane in selecting
. or modifying a configuration to provide satisfactory stability and
control characteristics. The requirements of the National Advisory
“Committee for Aeronautics for satisfactory flying qualities, which
Specify the important stability and control characteristics of an
airplane from the pilot's standpoint, ere used as the main topics of
the paper. A discussion is given of the reasons for the requirements,
d of the factors involved in obtaining satisfactory flying qualities, and
-- of the methods used in predicting the stability end control charact-
istics of 811 airplane. This material is based on lecture notes for a
training course for research workers engaged in airplane stability and 9
r control investigations. .

INTRODUCTION

In recent years, extensive flight, wind-tunnel, and theoretical

investigations of the stability and control characteristics of airplanes
have led to an improved understanding of this subject and to better
correlation between the results of these three research methods. The
present paper suunnar izes the more important aspects of this field of
research and presents information that will aid the designers of an
airplane in selecting or modifying a configuration to provide satis-
factory stability and control characteristics. The material given in
this paper 3s b.ased on lecture notes for a course, first given in 1942,
that was intended to train research workers engaged in airplane stability
and control investigationf3.
The flying qualities of an airplane sre defined as the stability
and control characteristics that have an important bearing on the safety
of flight and on the pilots ' impressions of the ease of flying an
airplane in steady flight and in maneuvers. Most of the available
tiwledge of flying quelities.has been obtained from flight tests made
by the NACA since 1939 on approximately 60 airplanes of sll types. In
these tests, recording instruments were used to obtain quantitative
2 NKA m NO. 1670

i
measurements of control movements, control forces, and airplane motions
while the pilots performed certain specified maneuvers. The results of -.

many of these tests have been published as N&CA Wartime Reports. -w

Reference 1 is a typical exam@e of this type of report. From the fund

of information accumulated in these tests, it has been possible to
prepare a set of requirements for satisfactory handling qualities, in
terms of quantities that may be measured in flight or predicted from
wind-tunnel tests and theoretical analyses. When an airplane meets
these requirements, the airplane is fairly certain to be safe to
fly snd to have desirable qualities from the pilot's standpoint.
Different sets of specifications for satief'actory handling charac-
teristics have been prepared by various agencies as a result of the
work done by the NACA. The requirements for satisfactory flying
qualities stated in this paper do not form a complete set and are .
not taken directly from any of the previously published specifications,
but include the more important requirementa that should, in general,
be met by all types of airplanes. For more complete flying-qualities
specifications, references 2, 3, and 4 should be consulted.
The original lectures on wind-tunnel procedure and control-surface
hinge+noment characteristics were prepared by Mr. I. G. Recant and
Mr. T. A. Toll, respectively, and the correeponding sections of the
present paper were based upon the material prepered by these two members
of the Langley Aeronautical Laboratory staff.
A list of synibole ie included as an appendix.

LONGITUDINAL STABILITY AN D
CONTROL CHARACTERISTICS I N
STRAIGHT FLIGHT
STABIlXTYCHARACTERISTICS INSTRAIGBT EIIGRT
Requirements and Definitions

An airplane is required to be statically longitudinally stable with

stick fixed or free in flight conditions in which it is likely to be
flown for long periods of time, and inthelanding4pproach andlanding
conditions. The mean- of thie requirement is explained in the-following I
sections. First, the concept of trim and the concepts of static and .-.
dynemic stability are considered.
An airplane is trimmed longitudinally in steady flight with etlck - -
fixed when it ie in equilibrium, that is, when the resultant force on
the airplane is zero and the pitching moment is zero. An airplane ie
trimmed in steady flight with- etick free when, in addition to the above
conditions, the stick force is zero. The methods of obtaining trim are
 NACA em NO. 1670 3

to adjust the pitching moment to zero by means of the elevators and to

adjust the stick force to zero by either a trim tab, an adjustable
stabilizer, an auxiliary airfoil near the tail, or an adjustable spring
in the control system. Of these devices, the trim tab is by far the
most common.
In order to determine whether an a-lane is stable, it first must
be trimmed. Stability is-related to the behavior of an a3rplan.e after
it is disturbed slightly from the trinrmed condition. Stability is
referred to as stick-fixed or stick-free stability, depending upon
whether the control is held f&xed in its trim position after the
disturbance or is left free. The behavior of an a-lane after such a
disturbance may consist of a divergence, a convergence, or an increasing
or decreasing oscillation. The definition of static longitudinal
stability is expressed in terms of this behavior as follows: anairplane
is statically longitudinally stable if, when disturbed slightly from a
trimmed condition (by changing angle of attack or speed), it will _
initially tend to return to its trimmed condition. An a-lane is
statically unstable if, when it is disturbed slightly frcanthe trFmnred
condition, it performs a divergence. The dynsmic longitudinal stability
may be defined as follows: an ajrplane is dynamically longitudfnally
stable if, after a disturbance, it performs a decreasing oscillation.
An airplane is dynamical&y unstable if, after a disturbance, it performEl
an oscillation of increasing amplitude.

Y
Methods of Obtaining Static Longitudinal Stability
An airplene will be statically longitudinally stable if, when the
. angle of attack is increased, the pitching moment acting on the airplane
becomes negative, tending to return the airplane to its original angle
of attack (duda negative). If this condition is fuDXl.led, the
airplane wiIl. also tend to return to its trim speed if the speed is
changed. For example, if the speed is seater than the trim speed,
corresponding to a lower eqQe of attack than that required for trim,
the adrplane will tend to pitch up to the trim angle of attack. As
a result, it will go into a climb and the speed will decrease and tend
to approach the trim speed.
An amroa-lte theory of static longitudinal stability is given
in order to show the effects of pAmary design features on the stability.
lil the folJowing anelysis, it is assumed that drag forces and propeller
effects may be neglected. The theory derived under these assumptions
applies approximately to the condition of gliding flight at low angles
of attack. The theory given herein is not sufficiently complete for
design purposes because the methods for determining the effects of the
fuselage and idling propellers are not discussed. The methods presented
in references 5 and 6 may be used to calculate the longitudinal stability
of an airplsne in the gliding condition for design purposes.
4 NACA TN NO. 1670

Any combination-of aerodynamic bodies that have linear v.c$ations.

of lift and pitching moment with an@;le of attack (such as a wing and
fuselage) may be shown to have an aerodynamic center. The aerodynamAc
center is defined as the point about which the pitching moment, remains
constant if the angle of attack is varied at a given airspeed. This
constant moment is indicated by the symbol Mo.
The moments and vertical forces acting on the airplane are indicated
in figure 1. The pitching moment about the center of-gravity is
M= Lx'+qyL$ (1)
By definition
L = CLc$3

dCL
CL =%

M = c,qsc
Making these substitutions gives i

M = ~SX’
dCL + ~qsc - % t
(2)

--
but

The following equation may therefore be derived:

cm - k(l -g) +.-%]($gT(IT5yz

+2g¶Sx~+C~¶SC (3)
--
= @c
This equation may be used to determine the tail incidence required for
trim (Cm = 0) at a given angle of attack for the simplified airplane
under consideration.. The degree of static longitudinal stability may
now be obtained from the preceding expression by differentiating with
respect to. a. The value of d&/da is:

+SLdcL (4)
da c
 NACA Tm NO. 1670 5

From equation (4) a value msy be found for xr, the distance from the
.-
-w
aerodynamic center to the center of gravity, such that -dCm = 0.
da
The concept of neutral point may now be introduced
because the neutral
point is defined as the centeMf-gravity location at which -%?l = 0 when
da
,the airplaneistrimmed (Cm= 0). When the center of gravity is ahead
of the neutral point, d&/da is negative and the airplane is statically
stable. When the center of gravity is behind the neutral point, the value
of dCm/da is positive and the airplane is statically unstable.
The preceding equations for determining the neutral point with
stick fixed can also be used to determine the neutral point with stick
free by using a value for the slope of the lift curve of the tail
corresponding to that obtained with the elevator free. If the elevator
tends to float with the relative wind (thatis, to float up when the
angle of attack is increased positively), the lift effectiveness Of the
tail will be reduced and the stick-free neutral point will be farther
forward than the stick-fixed neutral point. If the elevator tends to
.
float against the relative wind (that is, to float down when the angle
-*’ of attack is increased positively as it may with certain types of
aeromc balance), the lift effectiveas of the tail w-ill be increased
Y and the stick-free neutral point will be behind the stick-fixed neutral
point.
The stability of an airplane is expressed in terms of various design
parameters in formula (4). It is more convenient to trensformthis
formula so that the center-ofsavity position is expressed in terms of
its distance from the neutral point rather than from the aerodynamic
center of the wing--fuselags combination. Solving equation (4) for the
distance between the center of gravity and the aerodynamic center of the
wing-fuselage combination yields

(5)

At the neutral point, '2 = 0; hence, the distance between the

aerodynamic center of the wing-fuselage combination~and the neutral
point is L. -.
6 NACA TN NO. 1670

As mqy be seen from figure 1, the distance between the center of

gravity and the neutral point is obtained by subtracting equation (5)
from equation (6). This procedure gives the result

dcm
x -
C
-=-%
da (7)
=-- *Cm
*CL
Formula (7) shows that the degree of stability is determined solely by
the distance between the center ofgravity and the neutral point. The
distance between the center of gravity and the neutral point, expressed --
in percent of the mean aerodynamic chord, is fYequentl,y called the static
margin. If, in the design of the airplane, the centelr-ofvavity location
if3 COIISider8d t0 b8 VfXCiabl8, any degree Of Stability may b8 obtained
by suitable location of the center of gravity, and the tail msy then be
designed stiply from consideration of its ability to provide trim. On
the other hand, if the center of gravity is fix8d by other deeIgn
considerations, stability must be obtained by prwiding a sufficiently
rearward location of the neutral point; Formula (6) shows what design
features of the airplane may be changed to provide more resrwar d location
of the neutral point. These possibiliti8s include increasing the tail f
area, tail length, and tail aspect ratio.
Under the smlified a8sumptions of the preceding analysis, the
pitching+noment coefficient varies i.inesrly with angle of attack and,
as a result, the neutral-point location is independ8nt.of angle of attack.
These assumptions no longer hold in power-on flight or in flightnear
the St& where the drag is increasing or where appreciable flow
separation may have set in. In these cas88, the Variation of pitching
moment with angle of attack mey be nonlinear and neutral-point-location
will be a function of-e of attack.

IDynamIc Longitudinal Stability----

The position of the center ofgravity with respect to the neutral
point determines the static longitudinal stability but not- the dynamic
stability. Certain general relations exist, however, between the dynamic
stability and the position of the center of gravity with respect to the
neutral point. These relations sre s u?mDerized in figure 2, which shows
the behavior of an airplane following a disturbance, with stick fixed
and free, with various cente~f--@avlty locations. This method of
presentation is taken from a British report of lmted availability by
S. B. Gates, which gives a more complete discussion of these relations.
The period ofthe phugoid, or long-period, oscillation referred to In
figIn? 2 iS SO great that the dam@ng of this 0scKUation ha8 no
NACA m NO. 1670 7

correlation with the handling characteristics from the pilot's standpoint.

(See reference 7.) The occurrence of an unstable or poorly damped short-
period oscillation.with the elevator free is, however, very objectionable
and dangerous because of the rapidity with which large accelerations may
build up. (S ee reference 8.)

The divergence that occurs with the center of gravity behind the
neutral point is not violent, but is generally a slow, easily controlled
motion. Although this type of instability is not dang8rous, it is
objectionable to the pilot on a long flight because small corrections
must be made continually to hold a given flight speed. It is also
undesirable because of illogical control-force variations end stick
movements that sre required in changing the flight Sp88dS. For these
reasons, this type of instability is considered unacceptable for
satisfactory handling qualities. (This difficulty will be discussed
more fully in connection with control cheracteristics.)

EFE'ECTSOF PROPEiLiXROPEREI'IO~ AND

Single4Engine Airplanes

The following discussion applies primarily to propeller4riven

aircraft, though some of the effects of power on jet-propelled aircraft
are quite simllsr to those on propeller-driven aircraft.
The application of power introduces the follawing effects which
change the pitching moments acting on the a3rplane:
(1) Moment of propeller axial force about center of pavity
(2) Moment of propeller normsl force about center of gravity
(3) Increased angle of downwash
(4) Increased dynamic pressure at the tail
(5) Change in pitching moment of wing due to action of slipstream
These effects will cause a change in longitudinal trFm of the
airplane if the power is suddenly applied at a given speed. Since the
longitudinal stability depends on the variation of pitching moment with
angle of attack, the factors just listed will effect the stability if
they very in magnitude with the angle of attack. In steady flight,
the propeller thrust coefficient varies a+, as a result, all the
related propeller effects very with speed. The variation of propeller
thrwt coefficient with lift coeffici8nt in steady flight is ordinarily
similar to that shown in figure 3.
 NMA TN No. 1670 =
i
The moments of the direct propeller forces msy be estimated from
theoretical considerations or from experimental data given in various
papers. A theoretical treawnt-.of the propeller forces is given in .-
reference 9. Because the thrust coefficient-increases with lift coeffl-
cient, the moment coefficient caused by the axial force will increase
with angle of attack. If-the thrust line passes below the center of
gravity, this effect will be destabilizing. The normal forces act on
the propeller in a way simXLar to the force that would act on a smsJl
wing at the same location as the propeller. For a propeller located-
ahead of the center of gravity the propeller normal, force will therefore
give an appreciable destabilizing effect-
The effects oFthe dowmU h and increased aynamio pressure in
the slipstream on the pitching moments contributed by the wing and
horizontal ta2l surface are difficult to estimate from theoretical
considerations. For this reason, tests of powered models sxe normally
used to predict the stability characteristics of an airplane in the power-
on condition. Some general statements as to the effects of power on the
momente contributed by the wing and tail may, however, be made.
The increment in dynamic pressure in the slipstream caused by
propeller operation increases linearly with thrust coefficient. If the
tail is required to carry a down load for trfm (as for example, to .
offset the wing pitching nt0men-t with flaps down), the positive pitchlng-
moment coefficient given by the tail located in the slipstream will
increase as the angle of attack of the -lane increases, and a I
destabilizing effect will result. In extreme cases, the tail my
actual&y decrease the static longitudinal stability in pow8r-on flight.
Because of the increased normal force on the propeller with
application of power, the slipstream is deflected doummr d and thereby
causes an increased downwash over the tail. Also, with power on, the
slipstream increases the lif"t of the section-of the w&g that it covers.
The downwash in the slipstream, therefore, generally increases with
angle of attack more rapidJy than the downwash outside the slipstream.

As a result, the factor l- 2 that occurs inthe formula for the

stability contributed by the horizontal tail is reduced and the stability
of the airplane with power on is decreased.
If the tail is cmrying a down load end comes itio the high--velocity
region of the slipstream as the angle of attack incr8as8s, the positive
pitching+mme nt coefficient contributed by the tail wUl increase with
angle of attack.and a destabilizing effect will result. For this reason,
the horizontal tail surfaces of some airplanes have b88n located near 1 -
the top of the vertical tail in order to avoid enter- the slipstream at-
high angles of attack. . s
Though the effects of power on the longitudinal stability of single-
engine airplanes cannot be predicted in a completely rational manner,
 NACA TN No. 1670 9

=1 attempts have been made to devfse semiempilriosl methods that will yield .
fairly accurate results. The method given in reference 10 msy be used
-. for design purposes.

Multiengine Airplanes
The effects of power on the longitudinal stability of tti+ngine
or multiangtie air-planes sre similar to those on single-engine airplanes,
but certain.additional effects that depend on the mode of rotation of
propellers are introduced. If the propellers rotate in opposite direc-
tions, changes in downwash over the horizcntal tail will be introduced
by the slipstream rotation. This effect is most marked in the case of
twin-engine airplanes, because in mO6t oases the span Of the horizontal
tail does not extend far beyond the center lines of the two propellers.
The downwash betind the tibosrd portions of the propeller disks will
have a predominant effect on the angle of attack of the tail.
Experiments have shown that in the flamp condition of flight the
rotation of the slipstream behind the propeller continues in the same
dtiection after the slipstream has passed over the wing. If the
propellers rotate in opposite directions with the blades moving up in
. the center, the slipstream rotation will cause an increment of upwash
at the tail that will increase in strength as the speed is decreased
because of the resulting increase in torque coefficient. This upwash
J at the tail will cause a negative pitch3nga nt increment that
increases with increasing angle of attack; therefore a stabilizing effect
will result. Conversely, if the propellers rotate in opposite dlreotions
with the blades moving down in the center, an additional downwash at
the tail will be produced resulting in a destablizing effect. Figure 4
illustrates these conclusions.
-rants have shown that tith flaps down the direction of
slipstream rotation is reversed after the slipstresm has passed over
the wing. (See reference Il.) As a result, the effects on stability
discussed for the flap-up condition may be reversed in a flap-down
condition of flight. Ii~scune oases, inwhiohtests showthatthe
stabil.Pty of a m-engine airplane may be different with flaps up or
down, the mode of propeller rotation may b8 charged to utilize these
stability effects; for example, if the stability is satfsfactory with
flaps up but deficient with flaps down, the stability with flaps down
might possibly be improved by using propellers that rotate down In the
center.
In general, the mode of rotation cannot b8 readily changed because,
for r8asonB of servicing and maintenance, it is'desirable to employ
engines that rotate in the same direction.
10 NACA m NO. 1670

Jet&?t?ope~ed ALrpJan8s
on a j8i+pI?Ope~8d airplane in whlchthe jet is expelled from the
rear of the fuselage, the influence of the jet on the flow about the
a-lane will probably have a negligible effect on stability. Application
of the jet power will, however, introduce the mom8nt of the direct jet
thrust about the center of gravity. The moment coefficient caused by
this force varies wfth speed in a manner similar to that caused by the
propeller axial force, and its effects on stabiiity are the aam8* A more
serious effect on stability may occur if the jet exit Is unsymmetrical.
In this case, the jet may adhere to one side of the nozzle in some
flight -conditions and not in others. As a result, the direction of the
jet thrust may change In sn unpredictable msnner and cause large pitohing-
moment changes. For this reason, it is advisable to use a symmetrical
nozzle which is not located directly alongside other parts of the
airplane.
In Order t0 avoid damage t0 the StlllOtUT8, the jet iS always located
in such a way that it-does not 3mpinge directly on some psrt of the
airplane. Jets mounted on the wing, which pass below the tail, may,
however, cause considerable change in the dowrxa sh at the tail even
though they do not blow directly on it, because of the inflow of air
Into the mixing zone behind the jet. The destabilizing effect of this
downwash is similar to that of a propeller slipstream. The magnitude
of this effect may be estimated from data given in reference 12.
The flow into the inlets of a turbojet engine also causes a
destabilizing effect which may be estjmated from the change in direction
and the mass flow of the air entering the inlet.

CONTRCJLI
CHARclcTERIsTICS m STEADY FLIGEI!

In steady flighqthe elevatm must be used to offset any pitching

moment caused by the stability of the airplane or, in other words, by
the variation of pitching moment with angle of attack. ZEYhe airplane
is stable (dCm/da negative), more up elevator.(correspondin.g to a more
rearward position of stick) must be applied to hold the a-lane at a
higher angle of attack. Because steady flight at a higher angle of
attack corresponds to a lower flying speed, a stable airplane will
require a resrwsrd motion of the control stick to trim at a lower flight
speed and vice versa. Such a condition leads to a logical type of
control, that is, in order to reduce the speed, the pilot normally
noses the airplane up by pulling back on the stick. This stick position
mey then be maintained to hold the airplane In trim at a lower flight
speed. On the other hand, if the airplane is unstable, the pilot, in
order to fly at a lower speed, must first pull the stick back to nose
the airplane up and then move it forward ahead of its original position
to hold the atiplane in trim at the lower speed and prevent the speed
from continuing to decrease.
 .
NACA m NO. 1670 ll
-

The stability of the adrplane with stick free is similarly related

to the vsriation of control force with speed. IT the a-lane is stable
with stick fr88, a pull force will be required to trim at a lower speed.
Thus, for a stable airplane, if the speed were reduced and the stick
then releaSed, the stick would move forward end pitch the airplane down,
and its speed would therefore increaSe to the original trim speed. A
logical tspe of control results if the airplane has stick-free stability
because in order to reduce the speed, for example, the pilot must first
pull on the stick to p$tch the adrplsne up. He may then maintain this
control force to hold the airplane in trim at the lower speed.
The stick-fixed stability of an airplane is apparent to the pilot
through its influence on the variation of elevator angle with Sp88d or
with angle of attack. I.n steady flight the elevator is used to make the
pitching moment zero. The variation of elevator angle with speed may be
derived by u8e of this fact. The following relations ere obtained from
figure 5. The pitohlng moment due to 818vatOr a~@8 iS

(8)

This formula neglects the small pitching mment of the tail about its
qmrter-chord point. The p1tching-momen-t coefficient is

acL, 2Ecl
c, = -6e ag (9)
@c

In order to make Cm = 0, the pitching-moment coefficient due to the

elevator must be equal and opposite to the pitching-mame nt coefficient
due to the engle of attack. From formula (7) this cpantity is

(10)

Hence

or
12 - NMA ti ~b. 1670

or the elevator angle Be is directly proportional to the lift coeffi-

cient CL and to the distance between the center of gravity and the
neutral point. Because Ins-beady flight
.
CL=- (13)
&s
then

(14)

---
J .
or the elevator angle vsries inversely as the square of the speed.
Typical e-plea of the variation of elevator angle with speed for
stable and unstable airplanes are shown in figure 6. In general, curves
of the type predicted by formula (14) are measured in gliding flight
but considerable vsriations from this type of curve msy be obtained in
power+n flight, because of the effects of power mentioned previously
and also because of effects of sideslip that will be considered later.
The stick-Eree stability of sn airplane In flight is apparent to
the pilot through its influence on the variation of control force with
speed. The control-force variation with speed depends not- only on the
elevator-angle variation with speed but&o on the hinge-mome nt &aracte*
1st10s of the elevator. Some consideration of the hinge+uome nt cheracter-
istios of typical COnt3?01 surfaces will th8refore be required in order to
derive an ewession for the stick-fOrC8 variation with speed. A control
surfaoe‘that consists of a plain flap with no aeromc balance usually
has hinge moments that vary linearly with angle of attack orwith
deflection at- an&es below the stall. In practice, some type of aero-
dynamic balance on the surfaces is usually employed. In sdane cases, the
hinge-mome nt cheracteristics of anaerodynamically balanced surface are -
nonlinear. In order that the control characteristics of t.he airplane ?
shallbe normal, however, linear hing8+nome nt characteristics are very
desirable and an effort-is usually made to avoid nonlinear characteristics.
For this reason, it will be assumed in the following discussion that the L -
elevator hinge moment varies linearly tith angle of attack of the tail-.
 NACA TN No. 1670 13

and with elevator deflection. This statement may be expressed mathe-

matically as follows:

(15)

Einge moment may b8 expressed in terms of a dimensionless coefficient

similar to lift and m0ment coefficients. The hi% 84UO.IWII-t 008ffiOi8nt ch
is defined by the relation

H
Ch = (16)
qb,o,2

Formula (13~ then b8 8V8SSed 88 follows:

H= + "ec%, (17)

The term ChO has been added to take care of any initial hing8-mment
coefficient that may eldst when q.f and Se are zero. The trim tab
maybe usedtovary C$,.
The variation with speed of elevator hinge Jllament may be obtained
by substituting in formula (17) the eqressione fOr the values of QT
and 6, already derived. The expression for 6, (formula (14)) has
been modified by adding 8eo, the initial elevator d8fl8CtiOn when CL
is zero. This substitution gives

In steady flight CL = 2. The stick fOrC8 is directly prOportiOIUl. to

the elevator binge mament: F = KE. Making these substitutions and
8 implifyin@; giV8S

. .
14 NACA m NO. 1670

where Ch ' is the sum of the constant terms:

0

0’ = %Fs + Geochfj8 + cb (20)

Formula (19) my be used to show the effect of various design

features on the variation of stick force with speed. If the assumption
is made that the ratio qT/q does not vary appreciably with speed
(a condition usually true In gliding flight), the first two terms of
formula (19) sre seen to be independent of speed. The third term, which
depends on the trim-tab setting or stabilizer settdng, adds to the
constant force a force that varies as the square of the speed. These
conditions are shown graphically in figure 7. The slope of the CLWVB
of stick force agaIn& speed for a given trim speed is seen to be stable
when the sum of the first two terms gives a pull force. If chfje Is
assumed to be negative, factors contributing to stability are, first, a
cent-er-of-gravity location ahead of stick-fixed neutmtl point, and second,
a positive value of C The C&El8 Of a pOSitiV8 ValU8 Of C% is of
%c' 8
no practical interest because, as till be shown later, this condition
results in unstable short--period oscillations of the a-lane with stick
free. If the airplane is stable with the stick fixad (x/c negative)
increaBing c*, negatively will increase the slope of the cm8 of
stick force against-speed.
The relative importance of the t8rms (2% and C may be shown
e 43
by substituting the following typical values for th8 first two terms in
formula (19):

! = 40 POUIdS pa SqUSLT8fOOt K = 1.25

-*CL = 0.10 J?8r de@?ee

da

1 --=dE 04. - = 0.05 per- degree

da
ase
-X = -0.05
C
1 -
F = 0.2
I

The first two terms of formula (19) are 2CCC bece 2 and 62.~~ebec82.
%
 RACA TN No. 1670 15

For this particulsr value of static margin, therefore, a given change

in c has about three times as much effect on the sum of these two
hclT
terma, and hence on the stability chsracteristics,as a similar change
in (2%.
8
One type of diagram that illustrates graphically the relative effects
of c a* 'hse on the static stability, and that is also useful
%
in the design of an elevator, is shown in figure 8. This diagram is a
plot of c against Ch . On this plot is a line representing
%
combinaTtions of c and eC% which make the Bum of the first two
%c 8
terms of equation (lg), and hence the stick-free stability, equal to
zero. This line in drawn for the case of a static margin of 0.03 just
considered, and also for the case of the center of gravity at the stick-
fixed neutral point (static margin equal to zero). When the static
margin is equal to zero, variations of C have no effect on the
%e
stick-force variation with speed. In this diwam, each combination
of GQ) and c may represent the hinge-mome nt characteristics of
8 %c
an elevator with some type of aerodynamic balance. It is possible to
pick combinations of (2% that will give stability. A
8 and %c
range of types of aerodynamic balance which will give stability may
therefore be selected. Other lines, representing Buch quantities as
various degrees of stick-force variation with speed or acceleration,
trim changes due to flaps and -power, and boundaries between stable and
unstable short-period oscillations, may be drawn on a plot of this type.
The hinge-mome nt parameters which give the most desirable characteristics
for a given application may then be determined.
The relation between the control characteristics of the airplane
and the locations of the stick-fixed snd stick-free neutral points may
be sumarized on a diagram similar to that previously given for the
stability characteristics. This chart is shown as figure 9.
When a control surface is free to float it will assume a deflection
such that the hinge moment is zero. If the surface Is initially trimmed
at zero deflection, the floating angle is related to the angle of attack
by the fomula

s (a
.

It was previously mentioned that the Btick-free stability would be

.
1
increased if the elevator tended to float against the relative wind
and that a positive value of dCh/da would contribute to the stick-free
stability. Formula (21) indicates that a surface with a positive value
16 NACA TN No. 1670

i
of C& will float against the relative wind. The two methods of
considering the problem of stick-free stabill* ere therefore in agreement.

WON OF IKEZEAL POINTS FROMFLIGDT TESTS

Data for the determination of neutral points from flight tests sre
obtained by measuring the elevator angle and stick force required to
trim the airplane at-various speeds. The tests are made at two or more
centeMf*avity positions.

Stick-Fixed Neutral Point

The stick-fixed neutral point is determined from the variation of
the elevator angle with speed. Q-picsl flight data showing elevator angls
plotted against speed for various centerr>f+ravity positions are shown
in figure 10(a). The stick-fixed neutral point at any given speed msy
be determined by finding the cente~f~avity position at which the
elevator angle for trim remaim constant as the speed is changed slightly.
Because of the difficulty of reading the slopes of the curves plotted .
in figure 10(a) with equal accuracy at all speeds, it is desirable
to plot first-the elevator angles against lift coefficient as shown in
figure 10(b). Inasmuch as in this case these curves are not straight r
lines, the slopes of these curves are determined at the liftcoefficient
at which it is desired to find the neutral point. These slope8 me
then plotted against the center+f+$ravity position as shown in figure 10(c).
d6e
The stick-fixed.neutral point is the point at which slope r equals zero;
L
in this case, 36.5 percent mean aerodynamic chord (M.A.C.).

Stick4Vee Neutral Point

The stick-free neutral point is determined from the variation of stick
force with speed. Typical flight data showing stick force plotted against
speed for various centeLt%Jfwavity positions are &own in figure U(a).
From these.curves, a plot of F/q against lift coefficient is made as
shown in figure XL(b). The slopes of these curves are determined at the
lift coefficient at which it is desired tcfind the neutral pain-ix These
slopes are then plotted against the centemf-gravity position, as she-.
in figure 11(c). The stick-free neutral point is found as the center-of- .
dE !I
mavity position for which theslope 4 equals zero; in this case, at
@Ia m .
28.0 percent mean aerodynamic chord.

This method is strictly correct only at-the lift coefficient at which

the airplane is trFmmed, but the error involved at other lift coefficients
is generally within the accuracy of the flight data.
 NACA TN No. 1670 17

Another method to determine the stick-free neutral point in flight

is to trim the airplane, stick free, at various speeds and record the
trim.tab angle a~ a function of speed. The test is repeated at vsriou~
centeMf-gravity positions and the stick-free neutral point is determined
as the centerdf-avity position where the variation of trim-tab angle
with lift coefficient is zero. The procedure used is similar to that
described for finding the stick-fixed neutral point, from the measured
variation of elevator angle with speed.

EFFECTS OF COMPRESSIB~ ON TRIM MID STABILITY

Effects of Ccsrpressibility on Various Airplane Components

Ilarge changes in the aerodynsmic forces and moments exerted on a

wing do not occur until the wing critical Mach number is exceeded. At
the critical Mach number, a shock wave is formed. In order to define
the critical Mach nmber, a locus of points on the body where the
velocity of flow is a maximum must be determined. When the component
of velocity normal to this locus reaches the local speed of sound, the
critical Mach number ie reached, For Cwo-dimensional and exially
symmetrical flow, or other flows in which the locus of points where
the velocity is a msximum is perpendicular to the fLreMtream flow, the
critical Mach number is the speed at which the local velocity equals
the local speed of sound. At a &ch number apln?oximately l/l0 greater
than the critical Mach number, separation of flow occurs behind the
shock wave,and the lift and the moment acting on the wing sre g2eatl.y
changed. Generally the lift at a given angle of attack is reduced and
the pitching moment acting on the wing becomes more positive. The
critical Mach number of a wing depends principal&y on its thic3mees
and somewhat on its airfoil section. The critical Mach numbers of
various airfoil sections are given in reference 13.
The forces acting on the tail are influenced by ccm~essibility
effects in the same wey as the forcea on the wing. At Mach naberB l/10
or more above the critical Mach number of the tail section, the
effectiveness of a control surface such a8 the elevator may be expected
to be greatly reduced.
CompreBsibility effects on the fuselage may cause considerable drag
increases but they usually do not seriously affect the stability.

. Examples of Effects of CompresBibility

Typical effects of COmpreBSibility on the trim and Btability chsrac-
teristics of a straight-ing airplane designed primarily for flight at
I
I

subcritical speeds, a8 typified by fighter airplanes of World War II,

are a8 follows:
18 NACA TR NO. 1670

b’
(1) Large nosing-down tendency at high speed that may require pull
force on the stick exceeding the strength of the pilot .i
(2) Large increase in stability which requties unduly large elevatoi
movement and forces to produce a given change In lift coefficient or
acoeleration
An example of the variation with speed of the stick force required
for steady flight in a fighter a-lane of this type is shown in figure 12.
The stick forces required to pull out of the dive with various accelerations
are also shown. Although most airplanes experience a diving tendency
due to compressibility effects, som aIrplanes have shown a nosing-up
tendency.

Reasons for'Compre8sibility Effects

In most cases the diving tendency experienced at high Mach numbers
may be accounted for by a reduction in downwa8h at the tail resulting
from separation of flow at the wing root and also from the need to pitch
the a-lane to a higher angle of attack in order to maintain the ssme
lift on the wing as the Msch number increases. The increased stability
of the airplane at high Mach number8 results from the same cause, that is,
the airplane mu8t be pitched to a higher angle of attack than normal to
obtain a given lift increment and when this lift is obtained itis not
accompanied by downwash at the tail because of separation of the flow
from the inboard portions of the ting. when these compressibility effects
are experienced in flight they are generally accompanied by severe
buffeting and shaking of-the airplane caused by the action of the wing
wake on the tail surfaces.
Compres8ibility effects may be postponed to higher Mach numbers by
providing thinner wings and otherwise providing for a cleaner design.
The terminal Mach numbers of future airplanes may however frequently
exceed the Mach nmbers at which compressibility effects occur, in epite
of any refinements in design. With thinner sectiona, however, the
adverse effects of compressibility on stability and control are likely
to be much less severe.
Another method for reducing the adverse effects of com~essibility
is the use of sweepback. On a Elweptback wing of high aspect ratio, the
critical Mach number of sections not too close to the root or tip is
postponed until the component of velocity normal to the leading edge
exceeds the critical Mach number of the airfoil in twc+dimenslonal flow. I
(See reference 14.) On a finite-span swept wing, however, this amount of .I
gain is not obtained because the root section tends to behave more like
an unswept wing. Thus, the use of sweepback cannot-be expected to I .
eliminate stability difficulties similar to tho8e encountered with
straigh+wlng designs. The use of a large amount of sweepback also
introduces many low-speed stability and control problems. (See reference 15.)
 NACA TN NO. 1670 19

Dive-Recovery Flaps
One device which has proved successful in providing recovery from
dives at high Mach numbers on straighMng airplane configurations
designed primarily for flight at subcritical speeds is lmawn as the
div-recovery flap which consists of a pair of smsll mrrvable flaps on
the lower surface of the wing, generally located at about 30 percent
of the chord. Such flaps should be located in front of the horizontal
tail because their ma3n effect is to change the span load distribution
of the wing so as to provide en increased downwaah at the tail. For
a fighter ah-plane such flaps would have about 2-foot span and &Inch
chord. Uhen deflected in the dive these flaps will cause the a-lane
to pull out with an acceleration of about 59. The acceleration obtained
may be adjusted by vsrying the flap deflection. A typical dive-recovery
flap installation is illustrated in figure 13.

lZE!EXCTSOF - AND corn- ACEDlECOECION

. Another cause of difficulty with longituddnal stability and control

characteristics that appears in flight at high speeds is distortion of
the covering on the control surface, twisting of the stabilizer, or
bending of the fuselage. The most serious effect generally results from
deflection of the covering of the control smace. Such effects gen.eraUy
arise frqm two causes. These axe first, abulging or suckinginof
the covering due to ~sitive or negative internal pressure, and second,
a change in the mean camber line of the control surface due to external
aerodynamic loads.
The effect of positive internal pressure may bulge the surface so
that its trailing-dge angle is weatly inoreased. This change in
contour msy result in-the surface beocming overbalanced and will cause
violent short-period oscillations to occur. On the other hand if'the
covering is sucked in by negative internal. pressure, the effective
trailingedge angle mag be reduced so that values of Cb become more
negative. Thischangeinhinge4nom9 nt characteristics may result in
a loss of stick-free stability which may cause unstable control-force
variations with acceleration in ddve pull-outs.
Bowing of the mean camber line of the control surface which
increases progressively with speed may occur if the fixed surface ahead
of it is set at the wrong angle. For example, if the stabilizer incidence
is too great, up elevator will have to be carried in f1Lgh-L at -high speed.
The down load on the elevator will cause a progressive increase in
curvature of the surface which gives an effect similar to deflecting a
trti tab on the surface farther up as the speed increases. As a result,
20 NACA TN No. 1670

rapidly increasing pull force willbe required to maintain trim. The

opposite effect will occur if the stabilizer is set at a negative angle,
requiring down elevator for trim. The effects sre iU.ustrated in
figure 14.
In order to determine whether unusual control characteristics in
high-speed flight are caused by compressibility or by distortion, tests
should be made .at low and high altitudes. In this way different Mach
numbers may be attained at the same dynsmic pressure. Compressibility
effects will always seein at a given Mach number;whereas distortion
effects will set in at a given dynamic pressure,
These stability characteristics cannot be pedicted from wind-tunnel
tests of a rigid model; however, tests ofa rigid model should give
characteristics of the basic airplane configuration when it is free from
distortion effects. The diBtOI?tiOn effects may be minimized by correctly
setting the stabilizer and by properly venting the elevator to avoid
large internal pressures. In some cases the distortion effects msy be
employed to advantage to provide increased stability if the rigid
airplane is deficient in stability. A more ccrmplete analysis of these
distortion effects is given in reference 16.
.

Requirement *

The specifications of various agencies for satisfactory flying

qualities differ somewhat in the limits specified for allowable trim
changes. In general, the requirement is that the change in stick force
due to changing the configurationofthe airplane by changing the flap
position or power condition should be less than 35 pounds at any speed
within the structural limits of the design.

Reasons for Trim Change with Flap and Power Condition

In general, changing the flap or power condition will cause a change
in angle of flow and in dynamic pressure at the tail. These effects
combined with the change in wing pitching+noment-chacteristics will
require a change in elevator angle to maintain trim. The change in the
angle of attack and elevator angle influence the elevator hinge-mament
coefficient in accordance with the values of C!b and C%. A change .

in dynamic pressure tames a change in elevator hinge moment even if the I

hinge;mament coefficient remains constant. Trim change may possibly be

minimized by using values of and C such that the effects of I
%.t hs I

angle of attack and elevator deflection tend to cancelone another.

The maximum trim change frequently occurs when full power is applied
sfter the airplane has been trimmed for a landing approach with flaps
 NACA TN No. 1670 21 .-

down and power off. This condition usually requires full nose-up tr&
L.
tab deflection. With application of power the velocity of flow over the
trim tab generally increases lIlore than the average change over the tail
and lsrge push forces may be required to wevent the airplane from
nosing up.
Onlarge airplanes, the value Cof must be made small to obtain
hs
light forces in maneuvers Over a reasonably large centeMf-gcavity rsrge.
Since large changes in angle of attack of the tail usually occur when
the flaps ere deflected, the value of Cb must also be smsll to avoid
large trim changes. In general, a large positive value of C&
(obtained with a horn4alanced elevator or a 'beveled-trailing-edge
elevator) has been found to lead to excessive trim changes.

Requirement for Ianding Characteristics

The flyi~ualities requirements state that the elevator control

should be sufficiently powerful'to hold the airplane off the ground until
three-point contact is made for a conventional landing gear and, for a
tricycle landing gear, should be sufficiently powerful. to hold the
airplane from actual contact with the ground until the minimum speed
required of the airplane is attained, The stick force required for this
maneuver should be less than 50 pounds pull.

Requirements for Take-Qff Chsracteristics

During the take-Off run it should be possible to maintain the attitude
of the airplane by means of the elevator at any value between the level
attitude and that corresponding to maximum lift under the following
conditions:
(1) For a tricycle landing gear, after 0.8 take-off speed has been
reached
(2) For a conventional landing gear, after 0.5 take-Off speed has
been reached

l.
DisCU3SiOn~of Ground Effect
L
. The foregoing requirements were established because the landing
condition is often the most critical with regard to elevator control.
This condition results from the fact that the ground reduces the downwash
22 NACA TN No. 1670

angles near the tail and makes the airplane more stable. The size of-the
elevator is usually determined by the oontrol requirments near the l a

(pound. A simplified ex@anation of the effect follows.

The airplans wing may be replaced by a vortex whose strength is
proportional to the lift, as shown in figure 15(a). The vortex produces
a vertioal velocity w in the region of the tail and the dowmash angle

f=- W
v

The effect of the ground can be simulated by a mirror image of-the

airplane and its vortex system, since such an image will satisfy the
condition that there can be no vertical velociQ through the ground.
This vortex system is shown in figure 15(b). The effect of the image
vortex is to produce an upward velocity wg in the region of the tail.
The downwash angle whe?.ithe airplane is near-the ground is then

w-w
g
E=
v

The dowmash is therefore reduced by the presence of the ground and more
umlevator angle is required to trim the airplane.

LONGITUDINAL STABILITY AND

CONTROL CHARACTERISTICS
IN ACCELERATED FLIGHT
Relations between Longitudinal Stability
in Straight and in Accelerated Flight

In the preceding sections the static stability of an airplane in

straight flight hss been discussed. The stability was related to the
variation of pitching moment with angle of attack. Changes in angle
of attack were brought about by changing the speed while keeping the
airplane in straight flight at 1 g normal acceleration. This condition
applies in ordinary climbing, cruising, or gliding flight. In maneuvers,
5
however, it is more common for the pilot to make sudden or rapid changes
in angle of attack which occur before the speed can change appreciably.
The result of such changes in angle of attack is to cause an accelerated . -
maneuver. In this case, the normal acceleration is more than 1 g and msy
approach the structural limits of the airplane, which for fighter
airplanes corresponds to about gg, and for transport or bamber types,
 NACA TN No. 1670 23

to about 38. During an accelerated maneuver of this kind, the'elevator

' is used to supply a pitching moment which balances the pitching moment
caused by the variation of angle of attack. In this respect longitndinsl
stability in maneuvers is similar to that in straight flight. An
additional pitching moment is introduced, however, because of the *
curvature of the flight path in an accelerated manewer. In order to
calculate the elevator movement and control forces required in
accelerated maneuvers, the effects of both sources of pitching moment
must be considered.
The effects of curvature of the flight path are discussed feat.
Consider the a&plane performing a pull-up fromstraightflightwhile
traveling at cons-tart speed as illustrated in figure 16. The change
in angle of attack of the tail caused by the curvature of the flight
path is given by the expression

The radius of curvature may be expressed in terms of the normal acceles

ation and the speed by means of the formula

n-E-Z 1 ar
g
e
Rg
(23)

The c-e in angle of attack of the tail caused by curvature is therefore

given by the expression

For some calculations this formula is more conveniently expressed in

terms of lift coefficient instead of norma& acceleration. Erom the
definition of lift coefficient

(25)

. . This formula may be solved for p to give

Wn
82 =- (26)
24 NACA TN No. 1670

4’
Substituting this value in formula (24) gives the following expression:
.I

(27)

where

The quantity cr is called the airplane relativedensity coefficient.

This factor frequently occum in dynamic-stability-calculations.
The change in elevator mgle required in accelerated flight, ltie
the change in angle of attack of the tail, comes from two sources. The
first part, designated AEel, is that required to pitch the whole
airplane to a higher angle of attack, and the second part, desig- .
nated AEe2, is thatrequired to _.offset the additional lift on the
tail that results frmthe curvature of the flight path. The quantity tie1
.
is derived by equating the pitching moment due to the change in elevator
angle to the pitching moment due to change in angle of attack. The
ex$ression for the elevator angle was derived previously and is given
in formula (12). The change in elevator angle is

(29)

An additional change in elevator angle is required to offset-the effect

of curvature of the flight path. This change in elevator angle is given
by the expression

.
1

.
.
 NACA TN NO. 1670 25

where

(31)

The mm of these two increments of elevator angle gives the total change
in elevator angle required in accelerated flight.
.

Calculation of Stick Forces inhcelerated Flight

The change in elevator hinge moment msy be calculated from-the general
forDIula

AH= mecse + hrTc QTbece2 (32)

( 4
It is convenient to consider eepsrately the changes in hinge mmsnt
caused by pitching the whole airplane to a higher s&le of attack and
the changes in hinge moment caused by the effects of curvature of the
flight path.
Effects of pitching the whole airplane to a higher angle of attack.-
The change in elevator angle necesssry to substitute in formula (32) was
given in formula (29). The chsztge in angle of attack at the tail is
derived as follows:

(33)

Substituting the preceding values for AEel snd &,I in equation (32)
1
m and simplifying gives the following expression for the change in elevate
. hinge moment:

. .
26 NMA TN NO. 1670

AH1 =

as,’
- ke+,e

where AE$ is the change in elevator

+

hinge moment-neglecting the effects

due to curvature of the flight path.
Effects of curvature of the flight path.- The change in elevator angle
necessary to substitute in formula (32) was given previously (formula (30))
and the change in angle of attack at the tail due to the curvature of the
flight path was also presented (formula (27)): The elevator angle used
is that required to offset the additional lift on the tail caused by the
curvature of the flight path. When these quantities are substituted in
formula (32) and the result-s4mplified, the followjng ewesalon is
obtained for the change in elevator hinge moment caused by curvature of
the flight path:

$ShBe + ch,T g$ F 2bece'(n - 1)

I

Discussion of Factors Influencing Stick Forces

in Accelerated Flight
Formulas (3.4) and (35) showthatthe hingemomentandhence the
stick force fn a pull--up varies directly tith the normal acceleration
and that the force per g normal acceleration is approximately independent
of speed. The psrt of the stick force per g caused by pitching the
airplane to a higher angle of attack is proportional to the wing loading
and to the span times chord squared of the elevator. The contribution
Of %, to this psrt of the force per g is proportional to x, the
distance between the center of gravity--and the stick-fixed neutral point
in straight flight. The part of the stick force per g caused by curvature
of the flight path is proportional to the air density, the tail length,
and the span times chord squared of the elevator. This pa;rt of-the
force per g, therefore, vsxies with altitude and approaches zero at high
altitude where the density becomes small. This part of the force per g
is independent of the cente%ofsavity position,

Distinction between Turns and PuU.-Ups

In a steady turn the angle of bank rapidly approaches 90' as the
acceleration increases. For example, in a 2g-tmn the angle of bank
is 600, and ina &g-turnthe angle ofbsnkis 76. tin the airplane
is banked the acceleration of gravity which caused a reading of 1 g on
NACA TN No. 1670 27

the accelerometer in level flight is no longer applied to the instrument.

A turn and a pull-up made at the same value of acceleration as determined
by an accelerometer till, therefore, differ because 1 g which was supplied
by gravity in the pull-up must be supplied by a shorter radius of
curvature in the turn. The change in the angle of attack at the tail
caused by curvature of the flight path will, therefore, be greater in
a steady turn than in a gradual pull-up at the same acceleration. The
expression for change in angJ.e of attack at the tail caused by curvature
of the flight path in a turn is as follows:

=-
a"T 2 ( n2n;" )

When this expression is used to calculate the force per g it is found

that the force per g in a turn does not vary linearly with the acceleration.
The departure from linearity causes a slight difference between the values
of force per g measured in turns and pull-ups. This difference, however,
is generally within the experimental accuracy of flight tests. Many
other factors may cause a nonlinear variation of stick force with
acceleration on an actual. airplane. For example, nonlinear stick-force
variation may be introduced by nonlinear hinge-moment characteristics of
the elevator or by gyroscopic moments from the propeller.

Requirements for Elevator Control in Accelerated Flight

The elevator effectiveness is specified by the requircement that
either the allowable load factor or the maximum lift coefficient can be
developed at every speed. Ordinarily this requirement is less critical
than the requirement for making a three-point landing. Possible exceptions
to this statement are as follows: light airplanes for which the effects
of curvature of the flight path are large, and flight at high Mach numbers
where, because of large increases in stability caused by compressibility
effects, excessive elevator deflection may be required for maneuvering.
The veriati'on of normal acceleration with elevator angle and with
control force should be approximately linear. The theory developed
previously indicates that this condition will be satisfied if the elevator
hinge moment and effectiveness characteristics vary linearly with
deflection.
The variat+on of the elevator control force with nominal acceleration
should be in the following range:
(1) For transports, heavy bombers, and so forth, less than 50 pounds
per g
(2) For dive bombers, torpedo planes, and so forth, less than 15 pounds
per 63
28 NACA TN No. 1670

(3) For pursuit types, sport planes, and other highly maneuverable
airplanes, less than 8 pounds per g
(4) For any airplane it shouldrequire a pull force of not less
than 30 pounds to obtain the allowable load factor
These requirements vary somewhat in the specifications of various
agencies, but the force limits are in the same rsnge. Another requirement-
sometimes made is that-the airplane should not, under any conditions, be
flown with the center of gravity far enough back to reduce the force
gradient to zero pounds per g. An additional requirement that the force
in rapid maneuvers should be sufficiently heavy compared to the force
in steady turns has been ahownto be necessexy by recent research.

Examples of-Stick Force in Accelerated Flight

on Different Types of Airplanes
The stick force per Q of an airplane at any center-of-gravity
position may be conveniently shown on a plot of the type shown in
figure 17(a). The &fFects ofbbanges in some of the parameters that
influence the force per g are illustrated in figure 17(b). In order
to illustrate the effect of airplane size on the stick-force character-
istics, the force per g thatwould be obtained at various center-of---- --
gravity positions on three airplanes of different types has been
calculated. The calculations were based on the assumption ofan
unbalanced elevator with hinge- &parameters Cb =4.003 per degree
and (2% = -0.007 per degree. The results of the calculations are shown
in figure 18. The desired range of-stick force is also shown in this
figure. The airplane characteristics that were assumed in calculating
these results are given in table I.
From these examples, the use of a plain unbalanced elevav on
the fighter or bomber airplane types is seen to give stick forces that
do not satisfy the requirements over a sufficiently large center-of-
gravity renge.

Means of Obtaining Satisfactory Elevator Control

Force% in Steady MEum.IVerB
As illustrated in figure 17(b), the variation of stick force per g
with centerdf-avity position may be decreased by reducing the value
of Chs and the value of the stick force per g may be changed by a
constant amount at any centercof-gravity position by changing the value
of C&. A constant increment of stick force per-g may also be added
 NACA TN No. 1670 29

-P
by use of a bobweight. A bobweight, therefore, has an effect on the
stick-force characteristics similes to that of a more positive value
.. of c Means for independently varying the values of Ch and C%,
klcc*
were discussed in connection with the balancing of control surfaces.
Figure 18 shows that an unbalanced elevator will provide satisfactory
stick forces on a light airplane, but that a lerge amount of aerodynamic
balance will be required on larger airplanes. The required reduction
in C!h as a function of airplane weight is shown roughly in figure lg.
Since small variations in (2% will occur because of differences in
contours of the elevators within production tolerances, the stick-force
chsracteristics of very large airplanes may be difficult to predict and
may vsry widely between different airplanea of the same type if a
conventional elevator is used. These difficulties may be avoided by
the use of a servotab or by 6ome type of booster mechanismwhich multiplies
the pilot's effort by a large factor.

Stick Forces in Rapid PuU4Jps

When an airplane is equipped with an elevator that does not have a
. l&e amount of aerodynsmic balance, the stick force required to produce
a given acceleration in a rapid pull--up till be much larger than the
stick force required to produce the same acceleration in a steady turn,
because the elevator deflection required in a rapid pull-up is much
L
larger. On the other hand, if the elevator is very closely balanced
sothat c hs is zero and all the force in a maneuver results from the
use of a bobweight or a positive value of % the stick force in a
rapid maneuver-will be no weater than that in a steady turn. Such
srruements have been tried in order to provide desirable stick forces
in steady turns over a large range of centemf+avity position. Flight
tests of such an srrangament have shown it to be undesirable, however,
because the pilots object to the light stick forces in rapid maneuvers.
With such a system the pilot mey be able to deflect the elevator quickly
a large amount with practicallyno stick force and then the stick force
caused by the action of the bobweight will build up as the acceleration
increases. In order to avoid this undesirable control feel, the use
of very closely balanced elevators should perhaps be avoided. This
restriction will necessarily limit the cente~f~avity renge over which
desirable stick forces can be obtained unless some additional mechanism
is employed which increases the stick forces for rapid deflections
without affecting the forces under steady conditions.

I
.
30 NACA TN No. 1670

DISCUSSION OF TYPES OF CONTROL-

SURFACE BALANCE
Importance of Control-Surface Balance

The discussion of stiok-force characteristics in steady flight and

in maneuvers indicated the close relation between the stick-free
longitudinal stability characteristics of an airplane and the hinge-moment
peremeters of the elevator. The same type of relation is shown to exist
in the case of the aileron and rudder controls. Not only the stability
itself, but also the magnitude of the control forces in various maneuvers
is directly dependent on the control-surface hing e-moment parameters.
As larger and faster airplanes are made, an Increased degree of balance
(corresponding to values of C& and C% closer to zero) must be
employed on all control surfaces in order to prevent control forces in
steady flight and in meneuvers from becoming excessive. Several common
types of aerodynamio balance for control surfaces will be considered.
First, the oheracteristics of a plain control surface, which consists of
a hinged flap with no aeromc balance, is discussed.
l

Plain Control Surface

The values of Cb and Ch as a function of flap chord for plain
(unbalanced) sealed flaps on an RAC!A0009 airfoil of infinite aspect ratio
are shown in figure 20. These data are taken from reference 17. The
effect of finite aspect ratio usually is to reduce somewhat the negative
values of both C& and Cb. Reliable values of these hingnnt
parameters for a finite aspect ratio can be calculated from the two-
dimensional parameters only when methods based on lifting-surface theory
ere used. Lifting-line-theory methods, such as are generally used in
prediction of lift-curve slope, have been proven inadequate. Lifting-
surface-theory equations, applicable to fullepan control surfaces on
wings of finite aspect ratio, are given in reference 18.

Balance Characteristics

Overhanging or inset&hinge balance.- The overhangingbalance or

inset--hinge balance has beenthetype most commonly usedinthe past on .
actual airplanes. The hing~ment parameters for control surfaces I
having such balancea are affected by the overhang length and by the
balance nose shape in the manner illustrated in figure 21. These data .
are t&ken from reference 19, which also oontains a large amount of .

information on the various types of aerodynamic balance. Inoreasing the

bluntness of the balance nose reduces the hinge moments for small
defleotions, but it also tends to make the flow separate from the
 NACA TN No. 1670 31
lt
balance nose at smaller deflections than those at which separation occurs
on an elliptical- or sharpllose seotion. A control surface with a very
.. blunt-nose balance therefore usually must be restricted to a smaller
deflection range than a control surface with a more rounded nose shape.
Unshielded horn belance.- The effects of varying the size of an
unshielded horn balance ere shown for a typical case in figure 22.
These data ere taken from reference 20. The amount of balance is
expressed in terms of the erea moment of the horn about the hinge line.
Balencing tab.- The effect of a balancing tab is to reduce the
negative value of (3% without appreciably changing the value of C&.
The value of Cb is not chenged because the configuration of the
airfoil is not affected by the.tab except when the control-surface
deflection is varied. The tab effects the value of Cb by changing
the pressure distribution in the vicinity of the trailing edge of the
control surface when the surface is deflected. Thischangefora
belancing tab results in a small loss in control-surface effectiveness
as well as a reduction in the value of C!%. A tab with a ratio of
tab chord to flap chord of about 0.2 gives the least reduction in
control effectiveness for a given change in Ch6. Q-pical effects of
a balancing tab on the hinge+noment characteristics are illustrated
. in figure 23. The data shown in this figure are derived from reference 19.
Beveled-trailing+dRe balance.- The flow in the vicinity of the
trailing edge of an airfoil equipped with a beveled-trailing+dge control
surface, when the control surface is deflected, is like that caused by
a deflected tab. For this reason, the value of (3% is reduced by the
beveled trailing edge. The beveled trailing edge also reduces the
negative value (or increases the positive value) of Ch. A beveled
trailing edge on an unsealed control surface.msy give exaggerated effects
at small deflections and angles of attack, which result in overbalance
of the surface for a small deflection range. For this reason, control
surfaces equipped with a beveled trailing edge should be sealed. The
effects of trailing*dge angle on hinge nt characteristic6 sre
shown in figure 24. The data shown in this figure ere derived from
_ reference lg.
Sealed internal balsnce.- The characteristics of a sealed internal
balance ere scmewhat similar to those of an overhanging balance. The
1 ratio of the area of any leaks in the seal to the erea of the vents at
5 the hinge line must be small if the balance is to be effective. In
practice, some type of rubberized cloth seal is most satisfactory. The
e . effects of a sealed internal balance on the hinge-moment cheracteristics
ere showninfigure 25.
Other types of control+urface balance.- Other types of conizol-
surface balance that are sometimes used are as follows: shielded horn
32 NACA m NO. 1670

balance (paddle balance), Frise balance, piston b&Lance, end vsrious types
of double&inge control surfaces, such as those described in references 21
and 22. Other devices that may be used to reduce control forces include ..
spoilers (reference lg), all-movable control surfaces (reference 23),
servotaba, and spring tabs (reference 24).

Comparison of Various Balancing Devices

The preceding discussion of the various balancing devices has shown

that some balances affect Cb more than C%, whereas other balances
have a predominsnt effect on C%. In order to obtain desired control-
force and stability cheracteristics, it is convenient tobe able to
very Cb and Ch independently through the appropriate choice of
balance or of combinations ofbalances.
A comparison of the relative effects of the various balances on
the hinge-mome nt parameters is given in figure 26 where values of Cb
are plotted against values of Chg. A point indicated by a circle on
figure 26 represents the values of the hing e+noment parameters ofa
typical plain control surface* The various lines radiating from that
point indicate the manner inwhichthe hing e4omentpsrameterssre
changed by the addition of vsrious kinds of balances. The distance
along any of the lines from the point-for the plain control surface
to a point- for a balanced control surface depends on the amount of
balance used. Through the appropriate choice of aerodynamic balance
a large number of combinations of C& and (3% can be obtained. A
considerably greater number of combinations of these parameters can
be obtained by combining two or more types of balance as, for example,
a small amount of bevel with any of the overhang balances or with a
balancing tab. .The value of Cb may be made to increase positively
while the value of Cb increases negatively by combining an unbalancing
tab with an unshielded horn balance or with a beveled--trailing-edge
balance. A plot of 4% against C% such as figure 26 showing the
balance characteristics may be used in conjunction with a similar plot
such as figure 8 showing the required hing e+noment characteristics. By
comparison of the two sets of curves, a balance which will provide the
desired stick forces may be selected.
Any of the-types of--balance discussed in this section may be used
to reduce the value of C% to zero if used in sufficient-amount; The
choice of the type-of balance to use in a practical instaUation depends
largely on the effect of the balance on ch&racteristics'other %han the . .
hinge moments at small deflections. The advantages and disadvantages
of various types of balance are briefly discussed in table II.
 NACA TN No. 1670 33

DIRECTIONAL STABILITY ABD

C.

CONTROL CHARACTERISTICS

Requirements

For ell Q-pea of airplanes, the rudder should be sufficiently powerful

to povide equilibrium of yawing moments in flight with the wings level
at any speed and in any flight condition. Whenthe airplaneistrinrmsd
at maximum level-flight speed, the rudder force required at any speed
from the stall to the maxUum diving speed should be as small a&possible
and should not exceed 180 pounds. In addition, the rudder control should
be sufficiently powerful to maintain directional control during take-off
andlanding. For multiengine airplanes, the rudder control should be
sufficiently powerful to provide equilibrium of yawing moments at all
speeds above Xl.0 percent of the staU.ing speed with any one engine
inoperative (popeller at low pitch) and the other engines developing
full rated power.

Directional Trim Characteristics for Single43ngine Airplanes

Typical variations of rudder angle, rudder force, and sideslip with
speed in straight flight with the wings laterally level are shown for a
singl~n@;in.e airplane in figure 27. The reasons for the rudder
deflection and sideslip required at low speed with poweron are
illustrated in figure 28. At high angles of attack the propeller produces
a yawing moment and the propeller-fuselage combination produces a side
force. For the normal direction of propeller rotation (clockwise when
viewed from the rear) the yawing moment and side force ere to the left.
Right rudder deflection is required to offset propeller yawing moment
end also to offset the aileron yawing lnoment when the ailerons ere
deflected to balance the propeller torque. The vertical tail, therefore,
develops an additional side force to the left. In order to offset the
left side force on the fuselage and tail, the airplane must sideslip to
the left because with the wings level no side-force component due to
gravity exists. Additional rudder deflection to the right is required
to provide directional trim when the airplane sideslips to the left
because of the airplane's directional stability. Right rudder deflection
is also required to offset the effects of slipstream rotation. The
provision of directional trim at low speed with flaps down and rated
power generally is a critical condition for the rudder power. It is
- desirable to have sufficient rudder deflection beyond that required fcrr
. trim to offset the yawing moments due to aileron deflection and rolling
velocity in a roll.

The variation of rudder force with speed is caused by the effects of

power and by distortion effects on the rudder fabric at high speed. In
34 N~~cA TN NO. 1670
*-
the power-off condition en airplane with zero fin offset would be-expected _
to require no rudder deflection or rudder force for trim at any speed. ..
The right rudder force which is shown by figure 27 to be required for
trFm in the lowdpeed power-on condition results from the right rudder
deflection required. The left rudder forces required for trim at very
high speed would occur if the fin were offset with leading edge to the
left, for the same reason that the elevator force variation with speed
depends on stabilizer setting. On actual airplanes the fin is frequently
offset to the left in order to reduce the rudder deflection required
for trim at low speed. This practice appears inadvisable on
high-speed airplanes because of its adverse effect on the rudder trim
forces in dives that result from distortion of the rudder.
A possible method for considerably redu&ng the rudder deflection
for trim at-low speed withoutltntroducing undesirable effects athigh
speed is to offset the center of gravity ofthe a-lane to the right;
This method is effective for the following reasons:
(1) The aileron deflection required for trim at low speed and
therefore the aileron yawing momentare ther8by reduced
(2) If the thrust force exceeds the drag, the excess of thrust .
over dreg produces a yawing moment to the right about the center of
gravity which reduces the rudder deflection required for trim
(3) Because of the smaller side force on the vertical tail, less
sideslip is required for equilibrium and hence the rudder deflection
required to produce this sideslip is reduced
The control deflections required when the center of gravity is offset
vary inversely ~EI the square of the speed.and therefore become very
small at high speeds. Flight-tests have shown that on a typical single-
engine atiplane a lateral centmfsavity shift of 1.8 percent of the
wing span reduced the rudder deflection required for trim at mInImum
speed in the wave-offcondition by 10'.
.

Requirements

Directional stability and control-..characteristics in sideslips.-

Right rudder deflection should be required to hold left sideslip, and .
vice versa. The variation of rudder angle with sideslip should be CL
approximately lineer for angles of sideslip up to kl?O. The variation
of rudderfarce with sideelip should be such that right rudder force
should be required to hold a rudder deflection to the right of the . *
trim position,and vice versa. If this requirement is met, the airplane
will tend to return to zero sideslip when the rudder is released. For
RA.cA TN NO. 1670 35

multiengine airplanes the directionel stability with rudder free should

be such that straight flight can be maintained by sideslipping, at any
speed above 140 percent of the staUing speed, with maximum possible'
asymmetry of power caused by loss of one engine.
Pitching moment due to sideslip.- The variation of elevator angle
and elevator force with sideslip angle should be as small as possible.
Requirements of different agencies are somewhat different. Flight tests
have shown that the pitchinS moments in sideslips should not be sufficient
to produce undesirable chan@s in acceleration if the elevator is left
free. A tentative requirement is that the application of a rudder force
of 50 pounds should not produce a change in normal acceleration greater
then 0.2g.
Side--force characteristics.- The.veriation of side force with
sideslip should be such that left bank is required in left sideslips and
vice versa.
The lateral stability and control characteristics in steady sideslips
are considered in another section.

Discussion of Equilibrium of an Airplane in a Steady Sideslip

In a steady sideslip the airplane flies straight with constant
attitude end speed end must therefore be completely in equilibrium. In
order to maintain this condition the rudder is deflected until the yawing
moment is zero. The ailerons ere deflected to make the rolling moment
zero end the elevators ere deflected to make the pitchinS moment zero.
The airplane must bank so that the lateral component of gravity offsets
the aerodynemic side force on the fuse.lage caused by sideslip. The
relation between the angle of bank and the angle of sideslip may be
derived by referring to figure 29

W sin@ = CysS

or for smell angles of bank

36 NACA em NO. 1670

(37)

This relation shows that at low speeds or high lift coefficients, a large
amount of sideslip will be required in combination with a smell angle of
bank in a steady sideslip. At high speed the engle of sideslip corre-
sponding to a given amount ofbank is reduced. The formula also shows
that an airplane with a small amount of side area will have to sideslip
t-o lKt?g8 a1@88 for relatively small EUnOuntsOf bank in steady sideslips.
E anairplane is banked and an effort is made to raisethelowwing by
uSe of the rudder-alone, the flight-path of the airplane will continue
to curve toward the low wing until the sideslip is sufficient to develop
side force on the fuselage to offsetthe lateral component of gravity.
A large side-force coefficient is therefore desirable in order to
minimize course changes that occur when the airplane is displaced in
roll by gusty air.

Typical Deficiencies in Sideslip Characteristics

One type of difficulty frequently encountered, known as rudder "lock,“
is reelly a condition of rudder-free directional in&ability that occurs
at large angles of sideslip. This difficulty is ueually found to be
caused by the vertical tail stalling or emerging from the slipstream at
large angles of Sideslip. If an airplane is directionally stable with
rudder free, left rudder force will be required to hold the airplane in
a right--sideslip and vice versa. When a condition of rudder lock is
encountered the rudder floats to an angle greater than that required to
hold the airplane in a steady aideslip, and the pilot mu& exert-right
rudder force to return the rudder toward neutral when the airplane is in
a right sideslip and vice versa. This condition may b8 very dangerous
on a lerge airplane because the rudder force required to push the rudder
from its stops and startit turning toward neutral mey exceed the
strength of the pilot. .- _-
Directional in&ability at Small angles of sideslip is sometimes
encountered, especially in the flap-ug condition at high angles of attack.
It is sometimes caused by the vertical tail operating in the wake of-the
fuselage. This type of in&ability makes it very difficult-to hold the
airplane on the desired course, especial&v in meneuvers in which high
angles of attack are reached athigh speeds. Lack of directional
stability at smell angles of Sideslip may be dengerous in flight at
high speeds becauSe in accelerated rolling maneuvers, in which the
airplane is subjected to large yawing moments, angles of sideslip may
build up sufficiently to exceed the design load of the vertical tail.
 NACA TN NO. 1670 37
.
+
Negative dihedral effect may be encountered in flight at lar speed
with power on, especiaUy in the flaps4own condition, even though the
airplane may have positive dihedral effect in high-speed flight. The
causes of this condition are discussed in subsequent sections. negative
dihedral effect is undesirable, but is not oonsidered to be a dangerous
condition provided that the aileron control .is more than adequate to
hold up the leading wing in a sideslip with full rudder.

Contributions of Various Airplane Components

to the Directional Stability
Directional stability of the fuselage.- The variation of yawing
mame& with sideslip for a fuselage is difficult to predict because of
the irregular shape of the fuselage. The effect of the fuselage cannot
be neglected, however, because it usually contributes a large unstable
variation of yawing moment with sideslip. Theoretical attempts to
predict the directional instability of the fuselage have been based on
calculations of the yaa IIloments on ellipsoids in en ideal fluid. The
flow around an ellipsoid in an ideal fluid does not simulate the flow
around an actual fuselage and for this reason the theoretical calculations
ewerate the directional instability. These calculations do show that
the directional stability of the fuselage depends principally on its
dimensions as seen in the side view and does not depend to w large
extent on its thiclmess. Since yawing moments of fuselage shap88 ere
frequently presented in the form of yawing- nt coefficients based on
the fuselagevolume, cere should be taken to convert these results to
the basis of side dimensiona when th8y are applied to prediction of the
moments on a body with different CrOSS-S8CtiO~ shape. In order to
predict the directional stability of an actual fuselage, wind-tunnel-
test results for a similer fuselage shape exe preferred. Wind-tunnel
results ere frequently presented as the variation of aerodynamic forces
and moments with angle of yaw, rather than angle of sideslip. Angle
of yaw is defined as the angle of the longitudinal axis of the airplane
with respect to a fixed direction, whereas engle of sideslip is the
angle of the lo~itudinsl exis with respect to the direction of the
relative wind. For an airplane in straight flight or in a wind tunnel,
the angle of yaw is equal to the negative of the angle of sideslip, and
the two a?@88 msy be used interChang8ably. When any type of maneuver
involving turning is analysed, however, the two angles must be considered
separately. In the present paper the term "angle of sideslip" will
therefore be used in the text When the angle with respect to the relative
wind is being COZLSid8r8d. Some of the figures presenting wind-tunnel
data, however, er8 given in terms of angle of yaw in accordance with
usual wind-tunnel practice.
One of the factors contributing to the problem of rudder lock is
the fact that the LXDStabl8 yawing moment6 from the fuselage'and propeller
continue to increase When l.erge.angles of sideslip are reached, whereas
the stabilizing effect of the vertical tail may decrease when it stalls
or emerges from the slipstream. Figure 30 shows the variation of yawing
38 NEA TN NO. 1670

moment with angle of yaw for en isolated fuselage with circular cross
a8ction. The effect of smeJl fins added on the rearpert of the body
..
is also shown. The addition of fins makes the fuselage very stable at
large angles of sideslip though it does not effect the instability at
small angles of sideslip.
Propeller yawing moments.- A tractor propeller gives an unstable
variation of yawing moment with sideslip becau.se it behaves like a
vertical fin located 'ahead of the center of gravity. The instability
contributed by the propeller mey be accurately estimated from theoretical
calculations of the direct propeller forces, such as those given in
reference 9. The propeller also affects the flow conditions at the
vertical tail end so influences its contribution to the directional
stability.
Wing yawing mom8nts.- The variation of yawing moment with sideslip
for the wing is generally small. A wing with positive geoketric dihedral
will give a slightdestabilizingeffect because of the iIIflU8nCe of the
lift force on the yawing moments. The reason for the unstable variation
of yawing moment with sideslip is shown in figure 31. The lift vectors
are drawn perpendicular to the relative wind and perpendicular to the
surface of the wing. Yawing moments contributed by the induced drag
in a steady sideslip are small because the ailerons are used to balance
oh the rolling mcment and hence tend to equalize the lift on the two
sides of the wing. For conventional designs the contribution of the
isolated wing to the directional stabilit;y is very small, but it-may
become important in the case oftailless airplanes.
Yawing moments from the vertical tail.- The vertical tail is
designed to-overcome the unstable yawing moments contributed by the
propeller, wing, and fuselage. The yawing moments produced by the
vertical tail mey be estimated from the following formula:

In practice th8 quantities entering into this formula are difficult to

estimate accurately. The principal source of error is'the determination
of the area and effective aspect ratio of the vertitial tail. Inasmuch *

as tests have shown that the portion of the verticaltail located behind 1

the fuselage contributes very little to the directional stability, it-

appears desirable to base these quantities only on the portion of the M

V8X'tiCal tail located above the fLU3ele&58. The aspe& ratio of the .

vertical tail should be inCreas8d by a factor ranging in value from 1.2

to 1.5 to take into account-the end-plate effect of the horizontal t-ail.
 NACA TN No. 1670 39
.
.
The sidewash and dynamic pressure at the vertical tail must also be
estimated. The sidewash and d@amic pressure that exist in the propeller
l *
slipstream may be determined frmvsrious theoretical or eqerlmental
data. Interference effects from the wing and fuselage ala0 have a large
effect on the sidewash and WC pressure at the vertical tail. These
effects ere discussed in reference 25. Wind-tunnel tests have shown
that a favorable sidewash factor a6 as large 88 -0.4 may exist for
as
low-wing airplanes. On the other hand for himing a-lanes an
unfavorable sidewash factor of 0.6 has been measured. Tests of powered
models of actual airplanes have generally shown much smaller sidewash
effects: The average favorable sidewash for low+&ng models seems to
be approxJmately 4.1, to which the propeller sidemash should be added.
The dynemic pressure-at the tail may be assumed equal to that in the
propeller slipstreamfor airplanes with clean canopies, but for airplanes
with poorly shaped canopies the vertical tail erea in the wake of the
cenopy must be assumed to be relatively ineffective.

Design Considerations for Prevention of Rudder Lock

The yawing moments contributed by the fuselage, propeller, and
vertical tail may increase with sideslip somewhat as shown in figme 32.
The yawing moment given by the tail does not increase beyond about 15O
sideslfp because the tail reaches the stall angle and also emerges from
the slipstream, whereas the yawing moments given by the propeller and
fuselage continue to increase all the way to about a 45O angle of
sideslip. For this reason the airplane my become directionally unstable
at large angles of sideslip evenwith the rudder fixed. With rudder free
the directional stability will be further decreased because when the
vertical tail stalls the midder always has a large tendency to float with
the relative wind no matter what type of balance is used. (See reference 26.)
A large amount of directional stability must be added at large azl@;las of
sideslip so that the rudder deflection required to hold the airplane in
a steady sideslip will exceed the angle to which the rudder tends to float.
One method of making the fuselage stable at large mgles of sideslip was
pointed out previously in the discussion of fuselage yawing moments. This
method consisted of the addition of small sher-pedge fins along the reer
portion of the fuselage. These fins, lmown as dorsal or ventral fins,
have proved very successful in eliminating rudder lock on mazly actual
airplanes. Another method that has been proposed to prevent rudder lock
consists of 'placing vertical tail surfaces at the tips of the horizontal
tail. These surfaces tend to preserve the directional stability up.to
. larger angles of sideslip because they remain in the slipstream longer.
‘
Wind-tunnel tests showing the effect of dorsal fins and of end plates on
the horizontal tail on the yawing moments of a typicaPsme-engine
- . fighter airplane ere ah- in fi&ure 33. Note that the curves pass
through zero because of the use of contrarotating propellers. With
single rotation, the curves for power-on conditiom,ere displaced at
40 NACA TN NO. 1670

zero sideslip, and thus rudde-force reversal is caused at a still smaller

angle of sideslip in one direotion (normally in right sideslips).

Dihedral Effect
Requirements .-The dihedral effect as indicated by the variation of
aileron angle with sideslip in steady sideslips should be such that up
aileron is required on the leading wing. The variation of aileron angle
with sideslip should be approximately linear. The variation of aileron
force with sideslip angle should be such that the stick will tend to
return toward its trim position at zero sideslip when it--is released.
This requirement is equivalent to stating that the dihedral effect shall
be positive with stick fixed or stick free.
The max3mum allowable dyhedral effect is specified indirectly by the
following requirements:
(1)Whenthe airplane is displaced laterally and the controls are
released, the resulting oscillation should damp to one-half amplitude in
less than 2 cycles
(2) The rolling velocity in a roll made with rudder f.ixed should
never decrease to zero as a result of the sideslip produced in the roll
The foregoing re@rements for the maximum allowable dihedral effect .
are rather lenient and a more severe requkement should possibly be
provided. Some a-lanes with large dihedral effect and low directional
stability have paved obJectionable because of the violence of the rolling
motion caused by small movements of the rudder in hlgtipeed flight.
Further research is required before a definite requirement can be formulated
to cover this condition.
Definition of effective dihedral.- The geometric dihedral angle is
defined as the angle, as seen in the front view, between the wing panels
of an airplane and the spanwise axis of the adrplane. The effective
dihedral anglemay differ from-the geometric dihedral angle because of
the interference effects of the fuselage and propeller slipstream. The
effective dihedral of an a&&ane is defined as the number of degrees
of geometric dihedral that would be required on an isolated wing of the
same plan form to give the same vat?iation of rolling-moment coefficient
with sideslip. The.effective dihedral is taken on the basis that it is
constant fromthe root to the tip of the wi?. Thus, a wing'with tips
upturned at a 45O angle might have about 10 effective dihedral. .
,
The variation of rolling moment with sideslip per degree dihedral.
for wings of various plan forms and aspect ratios has been determined
theoretically and may be obtained from various papers, such as reference 27. . *.
For an aspect ratio of 6, lo of effective dihedral corresponds to a
 NACA TN No. 1670 41
,-
.
a

acz
vsJm3 of the variation of rolling -Munent coefficient with sideslip
aB'
-.
angle, of 0.0002 per degree.
Influence of wing location, power, and sweepback on effective dihedral.-
Ordinarily a high+ing errangement has about 3O more effective dihedral
thangeometric dihedral. A lowa errangement has about 3O less effective
dihedralthangeometric dihedral.
The effective dihedral on a tractor-type airplane frequently decreases
' with the application of power. This condition is most marked in the
climbing condition with flaps down at low speeds because in this condition
the ratio of dynamic pressure in the slipstream to free+tream dynamic
pressure is highest. The reason for the decrease in effective dihedral
with power is illustrated in figure 34. The decrease in dihedral effect
is caused by the additional lift developed by the trailing wing when the
slipstream, which is deflected in the sideslip, covers a larger area of
that wing. The lift results in a rolling moment tending to raise the
trailing wing. Because of the increase in the thrust coefficient as the
speed is decreased, the effective dihedral in power-on conditions of
flight becomes progressively more negative (unstable) as the lift coeffi-
cient increases.
A wing with sweepback is found experimentally to have a positive
dihedral effect that increases in proportion to the lift coefficient. This
effect mey be used to offset the decrease in dihedral effect due to power.
A typical example of the variation of effective dihedral with lift coeffi-
cient for as airplane in the power--on condition is given in figure 35.
The beneficial effect of a relatively small smount of sweepback in avoiding
negative dihedral effect at high lift coefficients is shown. With flaps
down sweepforward or sweepback of the hinge line of the flaps rather than
the quartemhord line of the wing sections is the important factor in
determining the dihedral effect. The difficulties encountered with large
positive dihedral effect in h-peed flight have been mentioned previously.
It is therefore very desirable to reduce as much as possible any increase
of dihedral effect with increasing speed. Eqerience has shownthat
negative dihedral effect at low epeeds is less serious than excessive
positive dihedral effect at high speeds. Though sweepback is beneficial
in offsetting the decrease in dihedral effect due to Bower, sweepback of
a wing even in small amounts is usually detrimental to its ataU.ing
characteristics.
The use of a large amount of sweepback (that is, 30° or more) on
jet-propelled aircraft for the improvement of performance at transonic
end supersonic speeds generally produces very large positive dihedral
effect at high lift coefficients. The increase in dihedral effect with
lift coefficient and with sweepback may be estimated qualitatively by
I

I celculatingtheliftontheleft andrightwings onthe ase~ptionthat

the component of velocity norm&L to the leading edge is responsible for
the lift of the wing. The predictions based onthis theory ere in fair
42 , NACA TN No. 1670

agreement with experiment so.long as the flow on the wing remains.unstalled.

With large sweep angles, however, flow separation may stert~at relatively
low angles of attack, and the dihedral effect obtained under these .-
conditions increases with lift-coefficient less rapidly than predicted by
the theory. Tests of sweptback wings with sharp leading edges have shown
that the dihedral effect changes from pos'3,tive'to negative values at
moderate lift coefficients, as a result of st&linn of -the leading wing. '
Quantitative data on the dihedral effect end other aerodynamic-cheracter- v * ,
istics of swept wings may be found in reference 28 andmany other papers.
High dihedral effect-at high lift coefficients or low flight speeds is
not so objectionable as it would betat high speeds, and acceptable flight' I T,
characteristics may be obtained Provided? '&&the directional stability
is also fairly large and the aileron effectiveness is normal. .
Measurement of effective dihedral in flight.- From the veriation of G.-T J .-'.
aileron angle with sideslip measured in steady sideslips~the variation I
of rolling mamentwith sideslip or the dihedral effect may be determined, * .
provided that the variation of rollingament coefficient with aileron,,' -
deflection is known, by means of the formula
..
&2
-z--w aBa X2
(39) - -
as a 38, 1 .

The variation of rollingmoment coefficient with aileron angle may--- r

be obtained from flight measurements of the rolling velocity by means
of the formula

The damping in roll C2 may be obtained for wings of various plan forms
-0
from theoretical calc~tions. The value of C2 is between 0.4 and 0.6
P
for unswept wings of normal aspect ratios.

AILERON CONTROL CHARACTERISTICS

Requirements for Satisfactory Aileron Control - .

Early research on lateral-control devices was concerned mainly with I

improvement of the lateral control of the airplane beyond the s-tell. .
Attempts were made on the basis of this work to set uP.requirements for
satisfactory aileron control characteristics. One Proposed criterion .
stated that the ratio of rolling-moment coefficient to lift coefficient l

should exceed a certain value. This criterion would in effect require

an airplane to have a rolling velocity thatvaried inversely as the
 NACA m NO. 1670 43

airspeed. Measurements of'flying qualities.of numerous ad@anes have

shown that a,criterion of this type,does not conform to the pilots'
opinions of satisfactory rolling performance. With conventional ailerons
the rolling velocity obtained with a-given-aileron deflection increases
%n~opori;ioptothe speed. Thb reason for thl.s Increase is that the
ailerons'introduce an effective twist InLo the wing that &uses the
a&plane to rol& essenti&ly on a geometric helix. In R steady roll'with
-, .' a given aXLeron deflection, therefore, the drplane always rolls through
, the same~engle ofLUnk in traveling a given ddstence no matter what the
airspeed.'
\
r \
. The conc&tthatthe 8iGplphn describes ahel3xwhenitrolls has
, ! led to the practice of specw the rate of roll in terms of the helix
generated by the wing tip. The tangent of this helix angle is given by
.f . . -. the eweseion Eb/2V as shown in figure 36. In practice pb/2V is of
. the order of 0.1 or less so that it is sufficiently accurate to consider
,
the tangent equal. to the angle expressed in radians. For this reason,
pb/2V is generallycalledthehelixangle.
\ . Flight tests of numerous aIr@anes have shown that'pilots demand a
. higher rolling velocity as the speed is increase?d and they also require
that a small ajrplane should be.able to roll faster than a larger airplane.
a These observations lead to the conclusion that the rolling ability of any
I airplane will be considered satisfactory by pilots if the value of pb/2V
is greater than a certain amount. Tests have shown that the rolling
. ability of an airplane is considered satisfactory when the value of pb/2V
exceeds 0.07 radian. (S 88 reference 29.) This criterion is consistent
with logical design of the a&?pti, became geometr1caU.y sdmiler wing-
aileron arrangements of different sizes with a given aileron deflection
_. will hsve the same helix angle independent of size or airspeed. IJY a
given rolling velocity were required to satisfy the pilots, the aileron
proportions would have to increase rapidly with the size of the airplane.

r With an aileron control system in which the ailerons are directly

linked to the control stick, the pilot is generally unable to obtain
full deflection of the ailerons above some definite speed because the
stick force required becomes too large. For nonmilitsry airplanes the
requdrements state that full aileron deflection should be obtainable
with 3&pound stick force or 80-pound wheel force up to 0.8 times the
maxjmum level--flight speed. ConibaterperiencewithmilitargadrpJanes
has emphasized the tipcrtance of rolling ability in both normal flight
and hi-peed dives. The present Armv and Navy requdrements, therefore,
specify that lsrge values of pb/2V IX rolling velocity should be
avaiJ.ab1e.u~ to the maximum diving speeds of fighter--type &planes with
the stick force not exceeding 30 pounds. The Army and Navy requirements
also specify a value of pb/2V considerably greater than 0.07 for low-
speed or c-sing flight In order to provide for r0lJA.g ability greater
”
. than that desired simply on the basis of satisfactory handling
characteristics.
44 NACA TN No. 1670
..
In addition to the previously stated requirements for aileron
effectiveness and stick force, the following requirements must be satisfied: .-
(1) The aileron force and rolling velocity should very approximately
linearly with aileron deflection and the stick force should be sufficient
to return the control to neutral when the stick is released
(2) The rolling acceleration should always be in the correct direction
and should reach a meximum value no more than 0.2 second after the ailerons
are deflected; this requirement-has alweys been met by conventional ailerons
but certain types of spoiler ailerons have proved unsatisfactory because of
excessive lag or initial reversal in their action

Typical Aileron Control Characteristics

If the ailerons ere suddenly deflected an airplane ordinarily reaches
its steady rolling velocity very rapidly. For this reason only the steady
rolling velocity is considered in the requirements for aileron effectiveness.
Z? the rudder is held fixed during the roll, the rolling velocity may
decrease after it reaches the maximum because of thesideslip developed
during the roll. Any sideslip in conjunction with the dihedral effect of
the airplane introduces a rolling moment opposite to that given by the
ailerons. E-the rudder is used to maintain zero sideslip, the rolling
velocity may continue to increase during the roll because of the rolling
moment due t0 y&Wing Velocity. Typical time histories of rudder-fixed
rolls are given in figure 37. Although innormal flight the rudder is
coordinated with the ailerons to avoid excessive sideslipping, tests
for aileron control characteristics are usually made with the rudder
fixed in order to obtain a maneuver that can be reproduced.
The variation of aileron effectiveness with speed is ordinarily
similar to th8t shown in figure 38. This diagramshowsthatwitha
rigid wing a constantvalue of pb/2V should be obtained at sll speeds
with full aileron deflection. In practice, however, the ailerons cause
the wing to twistin such a wey as to reduce the rolling velocity, until
at some very high speed; aown as the aileron reversal speed, the wing
twist completely offsets the effect of aileron deflection end the
ailerons fail to produce rolling velocity. The aileron reversal speed
should, of course, be well above the maximum diving speed ofen airplane.
A method for estimating the aileron reversal speed is given in reference 9.
Figure 38 also shows that some loss in aileron effectiveness may be
expected neer the s-tell because.of reduction inthe r.olling moments given
by the ailerons and because of the increased sideslip reached in rolls
at low speed. With a given stick force the pilot can fully deflect the .
ailerons up to some definite speed but at higher speeds ths aileron
deflection is reduced because- of the high stick forces, hence the
value of pb/2V is reduced. This reduction is illustrated in figure 38.
With a given aileron configuration and conventional types of aileron
balance, the aileron performance at low speed msy be improved at the
 RACA m NO. 1670 45

expense of high-speed characteristics by increasing the aileron travel

while keeping the s&m~ stick travel. Conversely, the aileron effectiveness
r. at high speeds may be improved at the expense of lowepeed rolling ability
by decreasing the aileron travel while keeping the eeme stick travel.
These effects are shown in figure 39. With increased aileron travel, the
value of pb/2'V for fuU. aileron deflection is increased but the speed
above which the pilot is unable to obtain full aileron deflection is
reduced because of the reduced mechanical advantage of the stick over the
ailerons.

Calculation of Rolling Effectiveness

The value of pb/ZV attainable with a given aileron deflection and
with given wing and aileron dimensions can be calculated accurately enough
for design purposes. The rolling velocity may be estimated within
about ri;lO percent for conventional types of ailerons in unstalled flight.
The calculation is based on the assumption that in a steady roll the
rolling moment due to the ailerons is equal to the damping moment in
roll

La=$=c2p (41)
( $@b
> .

. The damping moment is caused by the increased angle of attack on the

downgoing wing end the reduced angle of attack on the upgoing wing.
,Fornmla (41) shows that this moment is proportional to the helix angle,
the dynamic pressure, and the product of area end span of the wing. If
formula (41) is expressed in coefficient form, the following result is
obtained:

c2, = c2 22
( 1
P2v

The dsmping+nome nt coefficient C2 is a function only of the wing

P
plan form: Its value has been calculated theoretically and may be found
in reference 27 as a function of wing aspect and taper ratios. The
value of pb/ZV may be readily calculated if the aileron rollingmment
coefficient is known. This cgzantity may be determined from wind-tunnel
tests or msy be determined with equal accuracy framthe aileron dimensions
by the following procedure. The aileron rolling- nt coefficient may
. be expressed tithe form
s

C2,=6a -C28
0fT
where the coefficient Czg is equal to -3% and the value of 7 is the
asa
46 NACA TR NO. 1670
l

ratio of the variation of section lift coefficient with aileron deflection

to the variation of section 1zS.fL-tcoefficient with angle of attack. Notice
that the symbol T is equivalent to the symbol k used in reference 27. .I

Thevalue of 7% represents the rolling+nom.ent coefficient that would be

given by a wing if the spaawiee pert that includes the ailerons were twisted
1 radian. men this quantity is multiplied by T the rolling-moment
coeffI.cient is reduced to correspond to 1 radian of aileron deflection.
The value of %
7 may be found in reference 27 as a function ofthe wing -
aspect and taper ratios and of the spawise locationof the aileron. The
value of T may be obtained from section data but more accurate calculations
may be made by computing from values of pb/2V measured in flight-a value
of T for ailerons of a type similar to those under consideration. A
somewhat more exact procedure for calculating the value of pb/zV is given
in reference 9.

Amount of Aileron Balance Required for

Satisfactory Characteristics

The foLiowing example illustrates the degree of aerodynamic balance

required for ailerons on airplanes of various sizes.~~Conaider a fighter .
type &plane with the dimensions shown in figure 40. The value of pb/2V
reached with dLL5’ aileron deflection may be calculated as foXLows: For
20-percentcchord plain ailerons, assume that T = 0.4. From reference 27:

c2P= 0.46
-=
% 03,
7
From formula (43)

= $(0.3)(0.4) = 0.0314
‘Ia .

From formula (42)

*

EL 0.0314 = 0.068 radian I

2v 0.46
.
The stick forces ere calculated by assuming that plain ailerons with no
aerodynamic balance are used. The foUowlngtypicalvalues are assumed
for the hinge- ntparameters:
 NACA TN No. 1670 47

Assume 9 inches stick travel is required to deflect each aileron lgo.

The force required per aileron is then determined from the aileron hinge
moment as follows:

FAXs=E46a (44)

F = 0.35 H
The hinge moment is given by the equation

H= Ck + AEa Ck gbaca2
>

where L&C is the change in angle of attack at the aileron caused by the
rolling velocity. This change in angle of attack at the ting tip is
equal to the value of pb/2V. The change in angle of attack at any
point on the aileron mey be calculated by multiplying pb/2V by the
ratio 2b' where b' is the distance from the longitudinal &a to
b'
this'point on the aileron end b is the wing span. More complete
analyses, such as that given in reference IL, have shown that a point
near the inboard end of the aileron should be used to give the best
average measure of the angle-of-attack change.
In the present example

Aiz =--pb 2b'

!2Vb

= (0.068) 2

= 0.048 radian or 2.8'

.
I
where b* is the distance from the longitudinal axis to a point on
aileron 0.7 foot from the inboard end. lttmm equation (45), the hinge
moment on the-fully downward deflected aileron is
48 NACA TN NO. 1670

H = b-2.8)(-0.003) + 15(-0.007~ ~~(6.7)(1)~ = -0.00077v2 foot-pounds

where V is in feet per second.
The variation with airspeed of stick force to deflect two ailerons
is therefore as shown in figure 41. With plain ailerons, full deflection
cannot be reached with 30 -pounds stick force above 1% miles per hour.
Above this speed, the deflection and hence the value-of pb/2V vary
inversely as the square of the speed.
In order to meet the present Army or Navy requirements for aileron
control at high speed, the ailerons on en airplane of this size would have
to be aerodynamically balanced to reduce the hinge moments to about l/3 of
those for a plain aileron, even with the kl5O deflection range that was
assumed. The aileron deflection range would, however, have to be
increased to ~El9.5~ to meet-the low-speed requirement of a value of pb/2V
of 0.09. The mechanical advantage of the control stick would therefore
be reduced end the hinge moment for full deflection increased and a still
closer degree of balance would be required for satisfactory high-speed
cheracteristica.
Consider next a large bomber of 24%foot span, assumed to have a
wing-aileron arrangement geometrically similar to that of the fighter
airplane discussed previously. If a stick-type control is assumed, the
mechanical advantage of the stick over the ailerons will remain the same.
If plain, unbalanced ailerons are again assumed, the only quantity in
the equations that changes is the product baca2. This quantity is

or 63 which equals 2l6. By use of a wheel-

type control, the pilot's mechanical advantage may be increased about
60 percent, so that the forces wouldbe multiplied by -216 or 135. The
1.60
order of magnitude of the wheel forces is indicated in figure 42.

A very close degree of balance of the ailerons (approx. C% = -0.00014

and C&,, = 0.00000, for example) would-be required to reduce the wheel
forces to acceptable limits. In practice, this degree of balance is
unattainable because minor differences in the contours of the ailerons,
within production tolerances, can cause veriations in C% and C&
of kO.0005. Some type of servo or booster control is therefore required
for adequate control of an airplane of this size, or even for one of
considerably smaller size. The ailerons should be aerodynamically
balanced as far as possible, while a definite force gradient-is still
maintained, in order to reduce the power requirements for the booster.

Notes on Aileron Balance, F'rise Ailerons, and Spoilers

The example given previously showed that the change in angle of
attack at the aileron during the roll was about l/5 the chenge in aileron
 NACA TN No. 1670 49
_

deflection. A given change in the value of -C& will therefore have

only l/5 as much effect on aileron forces as the same chenge in the value
of C&. The aileron control-feel characteristics sre not markedly
affected by the ratio of the values of C!b and (2% althoughwhen C&
is positive, the control force required to suddenly deflect the ailerons
will be lighter than the final force reached in a steady roll; whereas
when the value of (2% is negative, the opposite will be true. All the
types of control--surface aerodynemic balance discussed previously have
been successfully applied to ailerons.
Certain additional means'of providing aerodynamic belance for aileron8
have been frequently used. These methods depend upon balancing the system
consisting of the two ailerons and their connecting linkage rather than
balancing each aileron individually. In the case of one frequently used
type of aileron balance, called the Frise aileron, the upgoing aileron
is overbalanced and therefore helps to deflect the downgoing aileron.
In using this arrangement the control system must be very rigid so that
the upgoing aileron will not deflect to excessively large angles and
cause the system to overbalance at high speeds. A differential linkage
is frequently employed in conjunction with Frise-type ailerons as well
as with other types of ailerons. With this arrangement the upgoing
aileron deflects through a larger range than the downgoing aileron. If
both the ailerons have an upfloating tendency, (trailing edge tending to
go up) the differential linkage will result in reduced stick forces.
The use of spoiler-type ailerons has been proposed to permit
increasing the span of the landing flaps, thereby decreasing tak-ff
and lending speed without sacrificing aileron performance. The hinge
moments of spoiler-tne ailerons may be erratic unless cere is taken to
use a desiep. that develops very small hinge moments. One successful
spoiler srrangement incorporated a thin circular-arc spoiler which
develops smell hinge moments, in conjunction with a small conventional
aileron to provide the necessary control forces. The spoiler should be
located far back on the chord in order to avoid undesirable lag in its
action.

Adverse Aileron Yaw

Use of the ailerons to produce a rolling moment also introduces a
yawing moment for two reason8. When the aileron8 are first deflected
the induced drag on the side of the downgoing aileron is increased and
that on the side of the upgoing aileron is decreased. The difference
, in induced drag causes a yawing moment. When the airplane starts to
roll the lift vectors on the downgoing wing are inclined forward and
those on the upgoing wing are inclined backwerd. A yawing moment is
therefore introduced called the yawing moment due to rolling which is
in the same direction as the yawing moment due to the ailerons. These
two yawing moments tend to swing the nose of the airplane to the right
in a left roll and vice versa. The change in heading is in the opposite
50 NACA TN NO. 1670

direction from that desired and this effecthas therefore been called
adverse aileron yaw. An additional yawing moment due to the profile-drag
differences on the left and right wings when the ailerons are deflected
must also be added to-the induced yawing moment and the yawing moment due
to rolling mentioned previously, but this profile-drag difference is
relatively small for conventional ailerons. With spoiler-type ailerons
the profile-drag differences msy introduce an appreciable favorable yawing
moment; Even when spoiler ailerons sre used, however, at high lift
coefficients this favorable moment is generaUy smaller than the sum of
the adverse yawing moments due to induced-drag differences and due to
rolling.
The adverse aileron yawing moment-in a roll may be calculated by
adding to the yawing moments measured in a wind tunnel. the yawing moment
due to rolling. The yawing mcment due to rolling may be determined as a
function of wing plan form by methods.from reference 27' and other papers.
If wind-tunnel data are not available, the induced aileron yawing moment
may be found-from theoretical calculations in referenoe 30. An approxi-
mate formula for the adverse aileron yawing- nt coefficient is as
follows:

c, = -cL &
82v

This formula, which is accurate within ck5 percent for ordinary wing plan
forms, gives approximately the sum of the yawing mnts due to induced
drag and due to rolling. The adverse aileron yawing moment is directly
proportional to lift .coefficient. ._ .

Requirement for Limits of Yaw due to Ailerons

Since undesirable heading ohanges occur in maneuvers because of the
effects of aileron yaw if the directional stability of an airplane is too
small, a requirement in the handling-qualities specifications has been
provided to set an upper limit on the sideslip reached in rolls. This
requirement states that the change in sideslip occurring in a rudder-
fixed roll made.with full aileron deflection at 1.2 times the stalling
speed should not exceed 20°. It is important that tu degree of'stability
should be obtained at small sideslip angles in order to limit inadvertent
sideslipping which causes heading changes inmaneuvers involving small
aileron deflections such as those used in flying through rough air. Also,
it is important to avoid large smounts of sideslip in hig'lwpeed flight,
as discussed in the following section. Thus-in a roll with 5 percent of
full aileron deflection, the sideslip should not exceed lo. With
conventional tspes of ailerons the designer can do little to reduce the
adverse aileron yawing moment. The rudder-fixed direction& stability of
the airplane must therefore be made sufficiently great tomeet the above
requirement. In flight tests, this requirement-can be checked more
 RACA TN No. 1670 51

conveniently by rolling out of a 45O banked turn, so that excessive angles

of bank sre not reached before the maximum sideslip is attained.

Rolling Maneuvers in Accelerated Flight

When an airplane is rolled out of a pull-out or out of an accelerated
turn, the values of pb/2v, lift coefficient, and airspeed may sll be
relatively large. The aileronyawinglame nt coefficient will therefore
be large, as shown by formula (46). The amount of sidealip developed
in a rudder-fixed roll at high speed in this type of maneuver may therefore
equal the amount of sideslip developed in a roll from straight flight
nesr the stalling speed. Reference 31 indicates that because of the high
speed, the loads imposed on the vertical tail msy be exceptionally large.
The provision of adequate directional stability, especially at amall
angles of sideslip, in order to prevent excessive sideslipping in rolls
at high speed is therefore important from structural considerations as
well as from the standpoint of providing desirable flying qualities.

STALLING CHARACTERISTI.CS
Requirements,for Satisfactory Stalling Chsracteristics
Conventional airplanes are unable to fly if the flow on the wing
. has completely stalled. In setting up the requirements for satisfactory
stalling chsracteristics the fact that normal control chsracteristics
cannot be maintained beyond the stall has been considered. The purpose
of the requirements is, therefore, to prevent inadvertent entry into a
atsUed condition of flight and to assure recovery from a stalled
condition if the pilot stalls the airplane intentionally.
The required chsracteristics are as follows: First, the approach to
a complete stall should be unmistakable to the pilot. Any of the following
cheracteristica sre considered to constitute satisfactory stall warning:
(1) Msrked buffeting or shaking of the airplane or control system
(2) Msrked rearward motion of the control stick or increase in pull
force required to stall the airplane
(3) Sufficiently slow developnt of instability
(4) A mechanical werning device may be used, in the event that inherent
stsll warning is not present
Second, it should be possible to effect a prompt recovery from a complete
ateLl2 by normal use of the controls. Finally, a'desirable characteristic,
although not required, is that the rate of roll of the airplane after
the stall should be low.
52 BACA TN No. 1670

Discussion of Typical Stalling Characteristics

Flight tests have been made by the RUA to determine the stalling
chsracteristice of many different airplanes. In these tests measurements
were made of the control motions, acoelerations along each of the three
-3, angular velocities about each of the three axes, angle of sideslip,
and airspeed. In some cases the progression of the stsll on the wing
has been visualized by means of tufts. Many different types of stall
behavior have been obsemed. In some cases a violent roll without any
form of warning occurs at the stall. In a fighteYctype airplane the rate
of roll has in some cases exceeded 90' per second. In other cases violent
oscillatory motion occurs in which the airplane ro118, pitches, end yaws
through a fairly large amplitude in an erratic fashion. This type of
stall is not so dangerous as the first-mentioned type but is, nevertheless,
COnSider8d .unsatiefactorg if the violent motion occurs without warming.
In some other cases, violent bUffetin@; of the airplane occurs several
miles an hour above the mjn-lrmrm speed and. full up elevator may be applied
without causing the airplane to roll. This tme of stall behavior is
considered satisfactory. Another type of motion at the stall consists
of-a gradually increasing oscillation in roll and pitch that, if allowed
to continue, msy eventually cause the airplane to roll on its back. This
type of stall is comidered satisfactory if the pilot has time to apply
corrective action before.the smplitude of the motion becomes excessive.
The stalling characteristics may be markedly different in different
conditions of power and flap setting. They may be also effected to a large
extent by minor chenges in configuration, such as change in cowl-flap
position. A stsll made from a high+peed.,turn is frequently more violent
than a stall made from straight flight because of the increased aerc+
dynamic moments acting on the stalled airplane.

Influence of Various Design Factors

on Stalling Characteristics
The stalling characteristics of an airplane cannot be accurately
predicted by my available methods. The uncertainty in the prediction
of stalling chmacteristics is due partly to the large nmiber of variables
which may influence these characteristics and partly to the lack of 821
adequate theoretical treatment of phenomena involving flow seperation.
A few general statements with regard to the present knowledge of stalling
characteristics will be given in the following paragraphs. In any
individual design, hoWeVer, other factors than those considered may have
a large effect on the stsU.ing characteristics. As~offull-scale
Wind-kIILLI81 StUdi of stalling cheracteristi0s is given in reference 32.
The progression of the stall on the Wing is usually comidered to be
of primary importance in determining the stalling chsracteristics. If
the stall starts first at the tip and progresses inboard, the type of
stall characterized by a violent roll without warning is likely to result.
 NASA TN No. 1670 53

A violent roll is caused because the region of stalled flow is at a large

distance from the airplane center line and, therefore, exerts a large
-- rolling mom8nt. As soon as the airplane starts to roll, the angle of
attack on the downgoing wing is increased farther beyond the stalling
angle while that on the upgoing wing is decreased. As a result the
downgoingwingis completely stalledwhilethe upgoingwingremains
unstalled. The large rolling moment produced by this asymmetric-flow
condition msy be accompanied by a large yawing moment which will tend
to sausethe airplane to enter a spin. Stall warning is likely to be
absent because the stalled flow does not strike the tail of the airplane.
Aileron control m~ly also be lost because of the stalling of the flow over
the ailerons. Initial stalling of the wing tips is likely to be caused
by a high degree of taper or by the use of sweepback. .In the case of
a tapered wing, the induced velocity at the wing caused by the trailing
vortices increases the effective engle of attack of sections at the tip
and decreases the effective angle of attack of sections at the root.
The tips therefore stall first unless the tip airfoil sections are
designed to have a higher stalling angle than those at the root.
Sweepback has a similsr effect in promoting tip stalling. The flow
field about the wing creates an induced velocity and also an induced
camber at the tip which tends to promote tip stalling. In addition,
the boundary lsyer tends to flow towsrd the tip, which helps to prevent
separation at the inboard sections.
. A stall Qhich starts at the wing root and progresses symmetrically
toward the tips is usually considered beneficial. This type of stall
may provide warning in the form of buffeting because fluctuations in the
flow occur at the tail over a region approximately twice as wide as the
region of reduced dynamic pressure in the wake. Furthermore, the large
loss of lift at the center portion of the wing may result in a decrease
in downwash at the tail. A large nosing-down moment will result and a
marked increase in upelevator deflection or a pull force on the stick
will be required to maintain trim. The smallmomentsrmof the stalled
srea contributes to a low initial rate of roll and the aileron control
may be maintained.
Initisl stalling of the wing root is promoted by use of a wing of
rectangular plan form or by sweepforwsrd. The induced velocities and
bo-undary-leyer effects sre then opposite from those of the tapered and
sweptback wings.
Some factors which may be overlooked in conn8ction with stalling
characteristics are as follows:
(1) On a large airplane a stall at the wing root may be unsatisfacto-ry
because of excessively violent buffeting of the tail.
.
(2) The wake from a wing stalled at the root mey b-et the vertical
tail. As a result rudder control may be lost and the airplane msy become
directionally unstable. This instability in combination with the high
54 NACA TN No. 1670

effective dihedral of a stalled wing may result in a violent directional

divergence and roll.
(3) "Stability' of the stall pattern is important, In other words,
several degrees change in angle of attack should.be required for the
stall to progress from the root to the tip. lYonlyasmalLchang8in
ar43le of attack is required to cause the whole wing to stall, then as
soon as the airplsne starts to roll the increased angle of attack of the
downgoing wing will. cause this wing to stall and a violent roll will
result. If stability of the stsll pattern is attained by means of "wash-
out" of the wing tips, a loss in maximum lift CO8ffiCi8R-t Will IleCeSS&XJ?ily
result because not all portions of the wing will reach their maximum
lift at 'the s&e time. Stability of the stall pattern may, however, be
provided by Use of slots on the outer portions of the wing. These slots
increase the maxjm~m lift coefficient at-these stations. This procedure
will not result in any loss of me23mum lift coefficient.
(4) If the wing stalls first at the hailing edge opthe wing root,
the apead of the stall to the leading edge rather than outboard on the
Wing is beneficial. This characteristic causes a large loss in lift
as the angle of attack is increased which will cause the airplane to
pitch down rather than to roll.
It is possible for some airplanes to have good stslling characteristics
even though the tip sections stall first. These desirable characterletics
are usually obtained by the Use of an a&foil section at-the tip which .
has a curPe. Withthiety-pe of lift curve the
airfoil maintains its lift beyond the stall and as a result large rolling
mom8nts are not applied to the airplane. Thin highly cambered sections
with f3mall 18adiDgedge radii generally have lift curves of this type.

Flight-Conditions Leading to
&advertent stalling
The handling characteristics of an airplane at speeds above the stall
may have .a decided effect on the danger of inadvertent stalling. A large
pitching moment dU8 to sideslip is undesirable because the pilot has very
little ability to judge the amount of sides7lp existing in flight at low
speed, and because changes in sideslip such as those occurring in a roll
out .of a turn in the landing approach may result in pitching moments
sufficient to stalLl the airplane. Longitudinal instability in the landing-
approach condition also increases the danger of inadvertent stalling
because the airplane will tend to at& by itself u~iiess the pilot applies , -
increasing push forces to the stick. Directional instability may reeult
in inadvertent large sideslip angles while rolling into or outof turns.
The maximum lift coefficient-may be considerably reduced at these large m .
sideslip angles, and the airspeed meter may give false indications, so
that the airplane may stall at indicated speeds at which it would normally
remain unstalled.
 EACA TN No. 1670 55
I
.

The formation of ice on the leading edge of the wing or on the

retaining strips of deicer boots may have a serious adverse effect on the
-w
stalling chsracteristics of an airplane end may also greatly reduce the
maximum lift coefficient.

Ground Looping
Ground looping end 8talLing chsracteristics sre closely related.
Ground looping difficul.ties have generally been caused by large yawing
and rolling tendencies caused by an unsyzmn8trical stall on the wing of
an airplane while it is in the three-point attitude. The groundangle
of an airplan with a conventional landing gear should be-approximately 20
less then the stalling angle in order to avoid this difficulty. The use
of a tricycle landing gear usually eliminates this trouble.

CONTROL -FREE STABILITY OR SHORT-

PERIOD OSCILLATIOES
Requirements for Longitudinal Motion

Jf en airplane which has static longitudinal stability is disturbed

from a trimmed condition end then allowed to fly for a long period with
the controls either fixed in the trim position or free, it will normally
perform a motion consisting of two types of oscillations. A short
oscillation, which generally dsmpe out within 1 or 2 seconds, occurs
immediately after the disturbance. A long-period oscillation then
occurs which consists of a gradual increase and decrease of speed about
the trim speed with a corresponding variation in the altitude of the
airplane. This long-period oscillation, called the phugoid oscillation,
has a period given approximately by the formula: period in seconds
8qlMlS one-qusrter times the v8lOCity in miles per hour. The period is,
therefore, of the order of a minute for high-speed airplanes in cruising
flight. Because the period is so long the pilot has no difficulty in
controlling the oscillation and causing it to damp out. Tests have
shown that the damping of the phugoid oscillation has no correlation
with the pilotte opinion of the handling qualities and, therefore, no
r8qUiremeRtS ere specified for its damping. In many actual airpkIl88,
this oscillation is unstable.
If the controls are held fixed following a disturbance, the short-
period oscillation always damps out so rapidly that it is difficult to
detect. With the controls free the short-period oscillation generally
damps out very rapidly, but in som8 ca688 the pitching motion of the
airplane may be coupled with the osciXLa-tions--of the elevator to cause a
viOl9nt unstable oscillation. The period of this oscillation varies
inversely as the speed and is generally about 1 qecond in hi-peed
flight. If the oscillation does not damp out, it may cause large
56 NMA TN NO. 1670

accelerations approaching the structural strength of the airplane after

1 or 2 cycles, Such an oscillation cannot be tolerated and the
requirement is therefore made that this oscillation should damp out so
that the motion of the elevator and the airplane has complete-ly
disappeared in 1888 than 1 cycle.

Influence of Design Factors on Short-Period

Lo,ngitudinal Oscillations
Reference 33 ah&S that theoretically an airplane with a positive
value of (2% of the elevator is likely to experience unstable, short-
period 1ongitudineJ oadillations. An aLcplan9 having a positiv9 value
of C% will be statically unstable with stick free Unless the value
Of chs iS SUffiCi8nt~ pOSitiV9. ti a positive valU8 Of c& iS Used .
in combination with a positive value of C% to provide Stick-free static
stability, unstable short-period oscillations are likely to result. For
this reason a fairly accurate rule to follow in comection with the
design of aerodynamic balance for the elevator is that C% should
always be negative. The tendency for short+period longitudinal oscillations
to become unstable is greater at high altitude and with a bobweight in the
control system. Theoretical analysis end flight tests have shown that a
continuous shortcperiod oscillation msy exist under these conditions
~nl.988 the value of ..C% is sufficiently negative.

Requirements for Lateral Motion

When an airplane is disturbed lateraUy from a trimmed condition and
the controls are left free for a long perfod or held fixed An their trimmed
positions, the airplane will generally perform a ahorkperiod oscillation
and will eventually go into a spiral dive. The divergence into the spiral
dive, known as spiral instability, is very slow and, like the phugoid
oscillation, has no correlation with the pilot's opinion of the handling
cheracteristics. -j?or this reason there 93~8no requir&8nte for spiral
stability. Almost sll actual airplanes are spirally unstable. Two types
of lateral oscillation which are difficult to distinguish from each other .
m&y occur. Li?heaeare known as Dutch roil and snaking. The requirement
is made that the88 oscillations should damp to one-half-amplitude
-. in less
than 2 cycles.

.
Influence of Design Factors ,
on Lateral OsciUations
Dutch roll oscillations mey OCCUTwith the controls either ftxed or
free. The period of this type of oscillation on conventional airplanes . .
varies inversely as the speed and generally varies from approximately
 NACA TN No. 1670 57

6 seconds nesr the stalling speed to about 2 seconds near the maximum speed.
This oscillation is a combined yawing and rolling oscillation that is
generally well damped for normal values of directional stability and
dihedral. With normal values of directional stability an effective
dihedral of approximately 15’ would be required to cause instability of
the Dutch roll oscillations. On airplanes with a large amount of weight
in the fuselage, the inclination of the fuseleg to the flight path has
an important effect on the stability of the oscilJations. A positive
angle of attack of the fuselage has a stabilizing effect. (See
reference 34.) The tendency for this oscillation is inCr8aEted on
airplanes with high wing loading flying at high altitude and the
requirement for damping of the osclltion may set an upper limit on the
allowable dihedral angle for heavily loaded airplanes intended to fly
at very high altitude.
The type of oscillation called snaking is a constant-amplitude
motion that can oocur only with the rudder free. It is caused by the
use of a rudder that tends to float against the relative wind in
conjunction with friction in the rudder control system. Iftheairplane
is disturbed from a trimmed condition the rudder will tend to float
in a direction to oppose any sideslip that is introduced. The friction
in the rudder control system will then hold the rudder as the airplane
swings back through the trimmed position. The rudder, therefore, tends
to feed energy into the OSCillatiOR and a constant amplitude OsCillatiOn
is built up. This sequence Of 8VeRtS 16 illustrated in figure 43. The
period of the oscillation varies inversely as the speed, and the amplitude
is proportional to the friction in the rudder system. A theoretical
analysis of this type of oscillation is given in reference 35. Because
the motion of the airplane in this type of oscillation is very similar
to that in a Dutch roll, it is difficult to distinguish the two types
of motion. In some cases the pilot msy hold the rudder pedals fixed
but the flexibility in the rudder control system will sJlow the rudder
to move slightly and maintain 89 oacilL3tion of constant amplitude.
Nearly all cases of small amplitude yawing oscillations which have
been reported on numerous airplanes have been cases of snaking rather
than Dutch roll. A good rule to u8e in connection with the design or
rudderbalanoe isthatthe value of C& should always be negative so
as to avoid the possibility of snaking oscillations. Theoretically, a
small positive value of Cb may b8 Used without causing oscillations
provided (2% has a sufficiently large negative value.

Relation between Rudder, Aileron, and Elevator

* Short-Period Oscillations
The rudder snaking oscillation discussed previously is the most
. frequent type of &or-&period oscilJation caused by motion of a control
surface. Shor-G-period longitudinal oscillations with the elevator free
. are less likely to occur, and the range of hlnge-mcXment parameters that
58 NACA TN NO. 1670
.
l

can be used is less restricted by the requirements for stability of the

oscillations* Short-period aileron oscillations can also occur but these
oscillations are more difficult to obtain than those of the elevator. It .-
has been shown th8oreticalJ.y that unstable oscillations of the ailerons
can occur oxi& When Ck and Ch have appreciable positive values.
Short+eriod oscillations of the ailerons have been observed in cases
for which the controls were overbalanced for smalll;teflections because
Of llOIiLiIl8E12 hinge-mmne nt characteristics. OVerbn'LaTlr.8 Of either the
elevator or rudder controls at small deflections would be even more likely
to cause Sho+periOd oscillations of these controls, in addition to
probably causing static instability with controls free. The short-period
osci~ations discussed herein are quit9 distinct from fhXbh3r in that
they do not involve much deformation of the airplane structure. Usually
the oscillations caused by flutter'have much shorter periods than the
oscillation discussed in this section.

WIND -TUNNEL TESTS AND CALCULATION

PROCRDURES FOR DETERMINATIOR
OF FLYING QUALITIES

lXCRODUCTION

For many years wind-tunnel tests were ordinarily made ofmodels

Without-propellers. Som&izlaes 9mpiriCal methods were used t0 allow
for the effects of power on stability, such as, for example, a criterion
that required that the slope of the curve of pitching mom8nt against
lift coefficient should'lie within certain specified limits. Such a
procedure was shown to be unsatiefastory when quantitative flight4est
data became available. T8StS of poWered models are now ordinarily made
and it has been shown thatthe stability of an airplane may be correctly
predicted from these tests. The procedures for makiT@ such t8StS aY?9
discussed in reference 36.

SZtMUlXTION OF POWERCOIKDITIOHS
Criterions of Similitud8

Since the 8ff8CtS of power result from the~action of the propeller

forces and slipstream effects of the airplane, the68 factor8 must be
simulated as closely as possible in the mod81 tests. If the slipstream
velocities are correctly reproduced in relation to thefree-streem
velocities, the forces of the propeller will else be reproduced, since
they are equal to the changes in momentum of the air in the slipstr9sm.
The slipstream c.onsists.of a mass of. air towhich is imparted an increase
of axial velocity, a rotational velocity, and a vertical or lateral
V8lOCity. Propeller theory indicates that the &xial velocity is a
 RACA TN No. 1670 59
.
.
function of the torque coefficient, end the vertical velocity is a
fUnCtiO?I of the nm-force coefficient. Because the relation between
the thrust coefficient and the torque coefficient is a function of the
prOpell8r efficiency, a popener on the model would have to have the
same efficiency as that on the amlan8 in order to simulate correctly
all the pItYOpeu9r 8ff8CtS. Generally, the efficiency of the model
propeller i8 Somewhat 1eSS than that Of the RSlplane propeller, b8CaU68
of its Smaller scale. Therefore, exact skulation of both the thrust
and torque coefficients may not be possible in longitudinal.~tability
t8StS. However, the thrust coefficient is the most important parameter
and should be 8XaCtu repI?tiUCed. The v8rticaLforce coefficient
msy generally b8 reproduced with SUffiCient accuracy by using a propeller
geometrically similsr to th8 fti-sCal8 propeller.

Variation of Thrust in Flight

The definition of popeller efficiency is given by the following
equation:

TV
c rl =-
55op

Hence, the thrust is given as a fun&ion of speed by the equation

c

Ordinarily with constant--speed propellers, the horsepower remains

approximately constant, and the propeller efficiency does not vary greatly
thrOUghOUt the speed range. The thrust, therefore, varies approximately
inversely as the speed.
In Order to test a pOWer8d model, t'he variation of thru& coefficient
with lift coefficient must be known. The thrust coefficient based on wing
area is usually employed in order that it should be directly GOmpeJ?abl8with
the drag coefficient. mom the preceding formula, the thrust coefficient
based on wing erea May be obtained as follows:

Tc' = 55w
9 P-w
60 MACATN No. 1670
.
.
The speed mey be elcpresged in tern of the lift coefficient by the formula:

(49)

Hence, the equation for the thrust--coefficient becomes

(50)

This formula shows that the thrmt Coefficient increases approximately

a6 the three-halves power of the lift coefflcisnt. The effects of power
on stability are umally greateet where the thrmtcoefficient and henoe
the axial velocity of the slipstream is greatest. Formula (50) ind.lcates
that theBe effects will be most marked at high lift coefficiente or low
speeds. The affects will also be greater at sea level than at high
altitudee.

Calculation of the Variation of Thrust Coefficient with

Lift Coefficient for a Specific AIrplane
For most imestigationf3 of specific models in a wind tunnel, the
manufacturer will furnish a chart showing the variation of thrust coeffi-
cient with lift coefficient for several constant-power conditions.
When such information is not supplied, however, this variation may be
calculated by the following method, The me of a conetant-speed propel&er
is asswned. Constant engine power is assumed becauee, in calculating the
stability of an airplane, it is dealred to detmmine the forces and
moments that result when the trfm speed or angle of attack is changed
and the throttle setting is maintained constant.
The following factors are known: eng?.ne brake homepower, propeller
speed, propeller diameter, airplez&e weight, and wing area. The procedure
may be outlined a8 follows:
(1) For several values of lift coefficient compute the speed
from the relation

For the first

to be zero.
appoxWation,
V=
r 23 CO8 8
@L
the angle of climb 8 may be asmmed
(51)
 II&IA TN No. 1670 61
.
.
(2) Compute the advance ratio for level flight for
-.
each value of lift coefficient.
(3) Csl.culate the power coefficient, Cp = 2.
A5
(4) From popeller chsrts applicable to the propeller under
consideration, determine Gf, B, and IJ for each of the values
of V and cp. These charts are frequently presented in the
0 %a t
formshown infigure 44. Examples of these charts msybe found in
reference 37.

(5) Compute the thrust coefficient based on wing srea

(6)The engle of CCL now be computed from the equilibrium

relation which applies in a steady ~1% or dive. Thfs formula may
be derived by considering the forces acting on the a-irplane as shown
in figure 45. Equating the forces in the direction of flight gives
the formula

T -D = W sin 8
W =- L
CO8 8
T -D =Lten8
hence

Tc* - CD
tme=--E--

The drag coefficient for use in calculation may be estimated or

measured on the model with the propeller removed.
62 NACA TN NO. 1670

(7) To correct the data for the engle of climb, recompute

and obtain new values of CT, p, and 9 for the corrected

values of -.V
nD
(8) The thrust coefficient may be corrected more simply by
use of the equation r

Tc’
T,’ = - LF
CO8 e

(g).The torque coefficient msy be obtained..from the formula

Selection of Model Propeller 3lade Angle ,

In the full-scale airplane the propeller blade angle changes with

flight velocity for constantipeed operation. It is desirable to select
a blade angle for the model propeller which will simulate as closely as
possible the efficiency and normal-force cheracteristics of the actual
airplane propeller. The model propeller msy be calibrated by making
measurements at verious propeller speeds with the model held at 0' angle
of attack. The drag of the model with propeller removed at the same angle
of attack C!% is also obtained. The thrust-coefficients may be computed
from the formula

T,' = CD -c%

and the torque coefficient may be obtained from the measurements of the
power .input to the model motor. From plots of torque coefficient against
thrust coefficient for each qf the blade angles tested, the blade angle
which most closely simulates the full-scale propeller may be selected.

Prepsration of Operating Charts

The procedure of the previous section has resulted in two charts:
the variation of thrust coefficient with lift coefficient for the airplane
.NACATN No. 1670 63

and the variation of thrust coefficient with rotational speed for the
model propeller at the selected blade angle. These chsrte may be combined
to give the variation of propeller rotational speed with lift coefficient.
In order to determine the variation of propeller rotational speed with
angle of attack, the variation of lift coefficient with angle of attack
must be determined with the correct variation of thrust coefficient and
also with the correct stabilizer setting variation to keep the model in
trim. A sufficiently accurate curve may be obtained from the tests with
two stabilizer settings. The results of these tests may be applied as
shown in figure 46. At given propeller rotational speeds the engle of
attack is selected to give the correct lift coefficient for a given
power condition for the two stabilizer settings used. A chart showing
the variation of propeller rotational speed with angle of attack must
be prepared for each power condition and flap condition to be tested.
The curve of lift coefficient agamt angle of attack for trimmed
conditionsmustbe used inprepsringthis chart.

Simulation of Propeller-Idling Condition

A windmilling propeller on a wind-tunnelmodelwill. usually give a
fairly accurate representation of an idling propeller on the actual
airplane provided there is no undue amount of friction in the model
propeller drive. In order to obtain the maximum accuracy in simulating
a propeller with engine idling, test data for the variation of engine
torque with speed on the actual airplane must be used.

Wind-Tunnel Tests Employing a Ground Board

Tests to determine elevator control near the ground are usually made
by instslling a ground board in the tunnel with just sufficient clearance
between it and the model lending gear to permit a reasonable variation in
angle of attack. The tests are made with the model in the landing
configuration, that is, flaps down, landing gear down, propeller windmilling,
and stabilizer set to the value used on the airplane for this condition.
The model is run through the angle-of-attack range with a series of elevator
settings. The pitching moment is plotted against s&l.e of attack for each
elevator setting. A cross plot is then made of elevator deflection
for trim against angle of attack. Because of scsle effect, the model angle
of stall and maximum lift coefficient will be lower than those of the
airplane. Consequently, the model usually stalls before it reaches the
angle of attack corresponding to the three-point attitude. The curve of
elevator angle against angle of attack must, therefore, be extrapolated
to this point in order to determine the elevator deflection required.
64 NACA TN NO. 1670

Simulation of Power for Take-Off Condition

The variation of thrust with speed and thrust coefficient with speed
have been discussed previously. On-the ground, as inthe air, the thrust
coefficient is determined by the velocity. In the air there is a definite
relation between velocity and the lift coefficient and therefore between
the thrust coefficient and the lift coefficient,--On the ground there is
no relation between the thrust coefficient and the lift coefficient-, The
airplane may be moving with a given velocity at almost ally liftcoefficient.
In wind-tunnel tests the model propeller operating conditions may be
determined by procedures similar to those given for the normal flight
renge . The calibrations must-extend to very large values of thrust coeffi-
cient since these values are.encountered at the airplane velocities below
take-off speed. It will probably be necessary to reduce the tunnel speed
considerably in order to obtain the required values of the thrust
coefficient.

Wind-Tunnel Test Procedure for Take-Off Condition

The take-off condition requires large control mments from the
elevator because of the ground-reaotion moments. The requirement amounts
to specifying that the elevator give sufYicient-aerodymunic moment to
counteract the ground-reaction moments. It is desirable to refer sJ2 the
mcments to the center of gravity, since the airplane in take4ffis
-.
accelerating. A summation of mcments about any other point would require
that-the inertia effects be considered.
The model is tested in the presence of a ground board at 0' angle
of attack with the thrust coefficient varied through a suitable range.
For a tricycle landing gear the maximum up-elevator deflection and the
most forward center-of-gravity location are used, and for conventional
landing gear the msximm dmlevator deflection and most rearward
centeMf-gravity location are used. Curves of aerodynamic pitching
moment available and moment required to balance ground.-reaction effect
can then be plotted against the thrust coefficient or velocity.
At 0.8 take-off speed the summation of the two should be positive for
the tricycle-landing-gear case; and at 0.5 t&e-off speed, it-should be
negative for the comentioaal-landing-gear csse.

Computation of Ground4eaction Moments

Tricycle landing gem.?.- In figure 47(a), the forces acting on an
airplane tith a tricycle landing gear during the tske-off run ere shown.
The ground-reaction moment is given by the formula
 BACA TN No. 1670 65
.
Mg = -Frh - Rd
-. R =w-L
Fr = (w - L)f

% = -(W - L)fh - (W - L)d

Mg =-Wfh+Lfh-Wd+Ld
Mg = afh + CLqSfh -Wd + CLqSd (55)
The corresponding moment coefficient is given by the formula

=- M
% MC
4ifh CL@fh Wd CLqSd
me-+-
Gm@;=qGc+ @c @c @c

(56)

From the wind-tunnel measurements, the speed at which the aerodynsmic

moment is sufficient to balance the ground reaction may be determined.
Conventional landing gear.- In figure 47(b) the forces acting on
an airplane with a conventional landing gear during the take-off run
sre shown. The equation for the ground-reaction moment coefficient
may be derived in the am manner as before.

(57)

Stick-Fixed Neutral Point

The stick-fixed neutral point may be determined from the measured

variation of pitching-mame nt coefficient with lift coefficient determined
with two or mare stabilizer (or elevator) settings. One way to determine
the neutral point would be to recompute the pitching moments about several
centeMf+ravity positions from the wind-tunnel bslance readings. With
 .
66 NACA TN No. 1670

sufficiently extensive calculations, the neutral points

couldbe found as
the cente+of-gravity locations for which Cm = 0 and da dcm = 0. .-
.
A simpler pocedure, given in reference 38, w%U now be described.
Assume that the wind-tunnel results sre presented as lift and pitching
moment about some particular point (1) on the model. As shown by figure 48,
the moment about another point (n) is given by

Mn = LXn + Ml (58)

Converting to coefficient form: -.

(59)

(60)

also

d%l
-= 5 d%
dCL 0 + q (61)

If point (n) is the neutral point, Gmn = 0 and dc%

-= o. lience,from
dCL
equation (60)

cml =-- xn
m (62)
CL c
and from equation (61)

(63)

In order to find the stick-fixed

neutral point, a point on the curves .
Cm1 d%ul .
of -c against CL where -=-mustbefound. The distance
ml CL dCL
between the neutral point and point (l), the pitching-mome nt reference .

point, is then equal to either - cml

- or ---.ac,,
CL dcL
 NACA 'TN No. 1670 67
.

A graphical method based on the above relationship that may be used

I-
to determine the stick-fIxed neutral point is shown in figure 49. At a
lift coefficient of 1.0, the following relation exists: .

=-- %l
c

= -0.05
Hence, the neutral point IS at 0.25 + 0.05 = 0.30 or 30 percent Rlesn
aerodynamic chord. At a lift coefficient of 0.6, the following relation
exists :

=- 3l
c

= -0.10
Hence, the neutral point 16 at 0.25 + 0.10 = 0.35 or 35 percent
aerodynamic chord at CL = 0.6.
At other lift coefficients, the results obtained from the tests
at two stabilizer settings must be interpolated or extrapolated. For
example, at a lift coefficient of 0.3, the values of -ac,
dCL =q
obtained from the measured results of figure 49 may be plotted as shown
in figure W(a).
The neutral point is found from the relation

-s-z
ac, c, -0204
.
dcL CL
. Hence, the neutral point is at 0.25 + 0.204 = 0.4% or 45.4 percent mean
. aerodpamic chord.

- Another graphical construction, known as the methcd of tangents, 18

* illustrated in figure 50(b) for the same data that were plotted in
figure 49. At a lift ooefficient of 0.3, the.neutral point is given by
the slope of the line from the origin to the intersection of tangents
to the pitching-moment curves at CL = 0.3. This slope is
68 NACA TN NO. 1670
.
cm
-c--z Rl -0.204
CL c

Hence, the neutral point is at 0.25 + 0.204 = 0.49, which agrees with the
value obtained by the previous method.
The pitching- nt curves presented in these examples ere idealized.
In Factice, experimental scatter of the data will make exact determination
of the slopes of the curves difficult. In order to reduce errors in
detemiaing the neutral points, it is desirable to obtain data for three
stabilizer settings with rather lerge increments of deflection.

Stick-Free Neutral Point

The stick-free neutral points msy be determined from wind-tunnel tests
in which the pitching moments and elevator hinge moments sre measured with
at least two stabilizer settings and two elevator settings, and the pitching
moments are also measured with tail eff. A graphical procedure similar to
that for the stick-fixed neutral points msy be used. This procedure is
described in reference 39. Alternatively, the model may be tested with
a free, mass-balanced elevator and the S&M procedure as was used for
calculating the stick-fixed neutral-point may be employed.

COHCLUDIBG REMARKS COHCERBIIG SELECTION

OF AIRPLANE CONFIGURATION TO SATIBFY
THE FLYING -QUALITIES REQUIREMEFTS

The various design factors which may be employed to obtain satisfactory

handling quslities have been discussed in connection with the various
requirements. Many of these design conditions are of a conflicting nature
so that compranises in the design will generally have to be made in order
to meet all the requirements as closely as possible. A few typical
examples of the conflicting requirements are given as illustrations.
The use of a slightly sweptbackwingto improve the dihedral effect in
lmdpeed climbing flight msy cause unsatisfactory stalling characteristics.
The use of a closely bslanced elevator to provide desirable atick4orce
gradients in steady maneuvers over a large oent~f-gravity range mey
result in undesirably light control forces in rapid maneuvers. The use
of a positive value of Ch on the rudder to improve the directional ,a

stability with rudder free will probably result in unsatisfactory snaking .

oscillations. Offsetting the fin to provide sufficient directional control

for trim at low speeds with power on may cause undesirably large variations
in rudder force with speed in high-speed dives. Increasing the chord of
 NACA TM No. 1670
I

any control surface to provide additional control power will make the
problem of balancing the control surface to obtain sufficiently light
stick forces more difficult. Many other similar examples may be found
by studying the handling+$mlities requiremnts in detail.
In spite of the conflicting nature of mmy of the design requiremmts,
several airplanes have been built which meet almost all the handling-
qualities requirements without appeciably sacrificing performance
characteristics. Desirable handling qualities in these cases have been
attained by considering the stability end control characteristics in the
early stages of the design and arranging such basic design factors as
the horizontal and vertical tail ereas and locations, wing plan form,
and centeMf-gravity location in such a way that the handling-quelities
requirements may be more easily satisfied.
The ability of an airplene to meet many of the handling~ualities
requirements mey be estimated quite accurately sim@.y from the dimensions
of the airplane. Methods of making these estimations have not been
discussed in detail in the present paper but they may be found in the
various NASA papers given as references. Some factors which cannot be
accurately estimated frcm the airplane dimensions at the present time
are the effects of power on longitudinal and directional stability.
Wind-tunnel tests of a complete model ere desirable in estimating these
effects. Themethods of calculatingthe flying qualities of an airplane
from wind-tunnel tests are described more fully in references 40 and 41.
In order to meke a complete evaluation of the handling qualities of a
proposed airplane, the effects of compressibility should be determined
by means of tests of a complete model in a hi-peed tunnel, end the '
hinge moments of the control surfaces should be measured by means of
tests of large-scale or full4ize models.

Langley Memorial Aeronautical Laboratory

Rational Advisory Committee for Aeronautics
Langley Field, Va., April l-2, 1948
 NACA TN No. 1670

.
APPENDIX
.-
SYMBOIS

&r radial acceleration, ft/sec2

b span of wing, unless subscript is used to indicate otherwise
b' distance from longitudinal axis to a point-on aileron

CD drag coefficient (D/qS)

drag coefficient of -lane with propeller removed
%
ch elevator hinge- ntcoefficient (H/Sbece2)
elevator hinge+noment coefficient-when ark = O" and 6 = 0'
%
%s variation of control-surface hinge-mcone nt coefficient with
angle of attack

chs variation of controlHurface hdng- ntcoefficient with

defLection

CL Uft-coefficient (L/qS)

% rolling-moment coefficient (L/qSb)

dsmping4noment coefficient in roll

c2P

variation of rollingament coefficient with aileron

deflection

cm pitching-moment coefficient (M/qSc)

.
pitchlng4lome n-t-coefficient at zero lift mo/qsc) .

Y~ng-mame nt coefficient m/q- 1

propeller power coefficient (P/t&D5)
propel&er thrugt coefficient (~/pn~D~)

.
 L NACA em NO. 1670 71
.

CY side-force coefficient (y/qs)

-c
c wing mean aerodynamic chord, with subscripts indicates
roohmwquere chord of indicated. surface
D drag, or propeller diameter

d horizontal distance between center of gravity and wheel hub

F stick force

Fr friction force
f coefficient of friction
acceleration due to gravity (32.2 ft/sec2)
hinge moment
h' vertical distance between center of gravity and ground
whenairplane is on-the ground

iT . incidenoe of tail
K ratio between elevator stick force end elevator hinge moment
-*
L lift, or rolling moment
.
Lp dempingmomentinroll I
2 tail length measured from the center of gravity to quarter-
chord point of tail
M pitcHng mDment

MO
pitching moment at zero lift
m mass of atip7

nr yawing moment

n $ropeller speed, rps, or normal acceleration in g

P shaft horsepower

. P rol3ng velocity
.
Qc propeller torque disk-load- coeffeiclent
72 -. NACA TN No. 1670

Q dynamic pressure y *
( >
I-

R ground reaction, or radius of curvature of flight path

S wing srea
T propeller thrust

% propel&r thrust disk-loading coefficient (T/pV2#)

Tc' propeller thrust coefficient based onwdng area

V true airspeed
W weight of adrplane
W vertical velocity of flow at-tail

X8 stick movement
X distance from center of gravity to neutral podnt
X1 distance from center of gravity to aerodynamic center of
wtng-fuselage combination

X0 distance from aerodynamic center of wing--fuselage

ccztiblaationto neutral point
side force
angle of attack
propeller blade angl-e, or angle of sideslip
changeinaquentity
control surface deflection
elevator deflection required for trim when CL = 0
downwash angle c

propeller efficiency
angle of clMb
airplane relative-density coefficient (m/@Z)
NACA m NO. 1670 73

P air density
&L
-j- =- as control+mface effectiveness factor
%
aa

Q sidew-ash en&e
angle of bank, or trailing+dge angle of airfoil
Subscripts:
a aileroti
b balance
8 elevator
f fl&P
Q . due to presence of ground
LF level flight
n pointn
t tab
T tail

W wing
 74 NACA TN No. 1670

REFERENCES

1. Phillips, W. H., Williams, W. C., and Hoover, H. H.: Measuremnts of

the Flying Quslities of a Curtiss SB2C-1 Urplane (No. 00014).
NACA MR, March 14, 1944.
my 2. Gilruth, R. R.: Requirements for Satisfactory Flying Quellties of
Airpl&ls. NACA Rep. No. 755, 1943.
3. Anon.: Stability and Control Characteristics of Airplanes. AAF Speci-
fication No. ~-1815-~, April 7,'1945.
4. Anon.: Specification for Stability and Control Characteristics of
Airplanes. Sl&UgA, Bur. Aero., April 7, 1945.
+ 5. Gilruth,,R. R., and White, M. D.: Analysis and Prediction of
Longitudinal Stability of Airplanes. NAf!A Rep. No. 711, 1941.
4 6. White, Maurice D.: Estimation of Stick4!'lssd Neutral Points of
Airplanes. NACA CB No. LsOl, 1945.
.=
7. Soul-e', Hartley A.: Flight Measurements of the Dy9ami.c Longitudinal
Stability ofseveral Airplanes and a Correlation of the Measurements
with Pilots' Observations of Hsndling Characteristics. - MACARep.
MO. 578, 1936.
8. Phillips, William H.: A Flight Investigation of Short-Period
Longitudinal Oscillations of an Airplane with Free Elevator.
NACA ARR, Msy 1942.
1)c 9. Ribner, Herbert 8.: Notes on the Propeller and Slipstream in Relation
to Stability.' NACAARRNo.L&Il&l~.
..!:- 10. Goett, Hexry J.', and Del-any, Noel K.: Effect of Tilt of the Propeller
Axis on the Longitudinel4tability Characteristics of Single4ngim
AirplEUles. EACH Rep. No. 774, 1944.
ll. Sweberg, Herold H.: The Effect of Propeller Operation on the Air Flow
in the Region of the Tail Plane for a Twi&Eugine Tract-or Monoplane.
NACAARR,Aug.19&2. ,
12. Ribner, Herbert S.: Field of Flow about a Jet and Effect of Jets on
Stability of JetcPropelled Airplanes. NACA ACR No. L&X3, 1946.
13. Baalet, Max. A.: Critical Mach Numbers of Various Airfoil Sections.
NACA ACR No. 4~18, 1944.
14. Jones, Robert T.: Thin Oblique Airfoils at Supersonic Speed. NACA Rep.
No. 851, 1946.
 NACA TN No. 1670 75
.
15. Sod, Hartley A. : Influence of Large Amounts of wing Sweep on
Stability and Control Problems of Aircraft. NACA TN No. 1088,
1946.

16. Mathews, Charles W.: An Analytical Investigation of the Effects of

ElevatozGFabric Distortion on the Longitudinal Stability and Control
of an Airplane. WA ACR No. L4E30, 1944.
17. Sears, Richard I.: Wind-Tunnel Data on the Aeromc Characteristics
of Airplane Control Surfaces. NACA ACR No. 3LO8, 1943.

rk 18. Swanson, Robert S., and Crandall, Stewart M.: Lif'ting~urface-lllhwoq

Aspect-Ratio Corrections to the Lift and Hlngw-Momwnt Parameters
for Full-Span Elevators on Horizontal Tail Surfaces. NACA TN
No. ll75, 1947.
lg. Langley Research Department: &maary of Lateral-Control Research.
(Compiled by Thomas A. Toll.) NACA TN No. 1245, 1947.
20. Lowry, John G.: R&LID& of Hinge-Moment Data for Unshielded Horn4alancwd
Control Surfaces. NACA IIB Ho. 319, 1943.
21. Sears, Richard I., and Purser, Paul E.: Wind-Tunnel Investigation
of ControlSurface Characteristics. XIV - NACA 0009 Airfoil with
a 2%Percent+Chord Double Plain Flap. NACA ARR No. 329, 1943.
22. Spesrman, M. Leroy: Wind-Tunnel Investigation of ControlSurface
Characteristics. XXIII -A 0.2woilGhord Flap with Tab Having
a Chord Twice thw Flap Chord on an NACA Ooog Airfoil. NACA ARR
NO. L5G25, 1945.
23. Jonws, Robert T., and KLeclmwr, Hsrold F.: TheorgandPrelimlnsry
Flight Tests of an Allaovable Vertical Tail Surface. NACA ARR,
Jan. 1943.
24. Phillips, William H.: Application of Spring Tabs to Elevator Controls.
NACA Rep. No. 797, 19&b.
25. House, Rufus O., and Wallace, Arthur R.: Wind4unnel Investigation
of Effect of Interference on Lateraldtability Characteristics of
Four NACA 23012 Wings, an Elliptical and a Circular Fuselage, and
Vertical Fins. NACA Rep. No. 705, 1941.
26. Sews, Richard I., and Hoggsrd, H. Page, Jr.: Characteristics of
Plain and Balanced Elevators on a Tgpical Pursuit Fuselage at
Attitudes Simulating Normal-Flight and Spin Conditions. NACA ARR,
March 1942.
27. Pearson, Henry A., and Jones, Robert T.: Theoretical Stability and
Control Characteristics of Wings with Various Amounts of Taper snd
Twist. RACA Rep. no. 635, 1938.
76 NACA TN No. .1670
.
28. Letko, WllUaa, and Goodman, Alex: Preliminsry Wind-Tunnel Investigation 1
at Low Speed of Stability and Control Characteristics of Swept-Back
wings. NACA TN No. 1046, 1946. .-

29. Gilruth, R..R., and Turner, W. N.: Lateral Control Required for
Satisfactory Flying Qualities Based on Flight Tests of Numerous
Airplanes. NACA Rep. No. 715, 1941.
30. Pearson, H. A.: Theoretical Bpan Loading and Moments of Tapered Wings
Produced by Aileron Deflection. NACA TN No. 589, 1937.
9. Gilruth, Robert R.: Analysis of-Vertical-Tail Loads in Rolling Pull-Cut-
Maneuvers. NhCA CB No. L4ECL4, 1944.
32. Sweberg, Barold H., and Dingwldein, Richard C.: Gummary of Mwasurem.wnt0
In Langley Full-Scale Tunnel of Maximum Lift Coefficients and Btalling
Characteristics of Airplanes. NACA ACR No. ItjC24, 1945.
33. Greenberg, Harry, and Sternfield, Leon&l: A Theoretical Investigation
of Longitudinal Stability of Airplanes with Free Controls Including
Effect of Friction in Control System. NACA Rep. No. 791, 1944.
34. Sternfield, Leonard: Effect of Product of Inertia on Lateral Stability.
NACA TN No. 1193, 1947.
35. Greenberg, Harry, and Sternfield, Leonard: A Theorwtical Investigation
of the Lateral Oscillations of an Airplane with Free Rudder with
Special Reference to thw Effect of Friction. NACA Rep. No. 762, 1943.
36. Recant, fsidore G., and Swanson, Robert 9.1 Determination of the
Stability and Control Characteristics of Airplanes from Tests of
Powwrwd Models. NACA ARR, July 1942.
37. Biermann, David, and Her&an, Edwin P.: Test8 of Five Full-Scale
Propellers in the Presence of a Radial and a Liquid-Cooled Engine
Nacelle, Including Twsts of mo Spinners. NACA Rep. No. 642, 1938.
38. Schuldenfrei, Mervin: Some Notes on the Determination of the Stiokqixed
Neutral Point from Windbl Data. NACA RB No. 320, 1943.
39. Schuldenfrwi, Marvin: Some Notes on the Determination of the Stick-Free
Neutral Point from Windmel Data. NACA RB No. 4B21, 1944.
40. &y-ten, Gerald G.: Analysis of Wi~&a 1 Stability and Control Tests
in Terms of Flying Qualities of Full-Scale Airplanes. mCA Rep.
No. 825, 1945.
41. Goett, Harry J., Jackson, Roy P., and Belsley, Steven E.: Wind-e1
Procedure for Determination of Critical Stability and Control
Characteristics of Airplanes. NACA Rep. No. 78l, 194.4.
 NACA TN No. 1670 77
I

--
TABLE1

AlREumc-IcsAss~m CALGuLATm STICK FORCES

REQUIREDWMANEWXB GIVEHINFEUREl8

The folloting characteristics were assumed to be the same for

all the airplanestakenas examples:
ET& . . . . . . . . . . . . . . . . . . . ..*......... 0.2

%F
-,perdegree . . . . . . . . . . . . . . . . . . . . . . . . . 0.03
38,
t/c . . . . . . . . . . . . ..'................. 3

(4
SL
aa w
,perdegree . . . . . . . . . . . . . . . . . . . . . . . . 0.07

.- dG,/dX,, radian/ft . . . . . . . . . . . . . . . . . . . . . . . . 0.6

The following characteristics were assumed for the

individual a-trplanes:
Light airplane:
W/S,lb/sqft . . . . . . . . . . . . . . . . . . . . . . . . 6
b e,ft........... C................
ce,ft............................ l.;
Fighter airplane:
W/S,lb/sqft . . . . . . . . . . . . . . . . . . . . . . . . 45
b e,ft.........................,..
ce,ft............................ 1:;
Bother:
W/S,lb/sqft . . . . . . . . . . . . . . . . . . . . . . . . 45
c
b e,ft............................ 35
.
ce,ft............................ 3
.* r' i
 1

. : I -

. .

FQure l.- Moments and vertical forces acting on an airplane

in stead$ mgh;ht.
 Stick-free R.P. Stick-fixed N.P.

--tMvergence rrith stick free -Divergence rith stick fixed

Short-period 0ecillZZ Short-period OBCdii

(ueually roll damped, aluaysuelldamped. Lon&
sometimes unstable). period oBcillation (phu-
Lang-period oecillation goid) stable or unstable
(phugoid)stable or un- nith stick fixed.
stable with stIck free.

??igure2.- Chart describing the dynamic longitudinal stability of an airplane as a fUnction of

center-of-gravity position. The arrows indicate location of center of gravity witi respect
to neutral point.

.
.
 XACA TN No. 1670
.
.

-.

0 1.0 2.0
% .

Figure 3.- Typical variation of propeller thrust coefficient T,

with lift coefficient in steady flight. T, = T
PV2D2’
 Direction of fLh

Figure 4. - Effect of mode of propeller rotation on downwashat tail on a twin-engine airplane.

, ‘, :

$ 6,, positive

Figure 5.- Forces acting on airplane due to elevator deflection.

.
 *
I\
'
)
Stable

OC

+ Stelling speed

Unstable '-
E
65 I I I I
0 100 200 300 4000 WJ
Indicated airspeed, mph

F’igure 6.- Typical examples of the variation of elevator angle with speed for stable and
unstable airplanes.
 .-

I I

%
sVI Tail heavy
I I I I J I
0 100 200 300
Indicated airspeed, mph

FQure 7.- Variation of stick force with speed in steady flight, as calculated by formula (19).
Values below the staJGngspeed have no physical significance.
 ,005 Stable region
Positive values of Cb not used
hecau~e of unstable sh&-period
Neutral stick-free stability
for static margin of 0 7 oscillations wfth stick free

0SF O

Neutral stick-f’ree
-.oof static &-gin of
-.05c
Unstable re&on
v Unstable side of boundaries
indicated by cross-hatching

-.OlO
-.015 -.OlO -A05 0 ,005

Figure 8.- Boundary between stable and unstable values of %a and Qs for the example
e
given in the text.

. *
 ‘.
. I I
! I’ I

Stick-free N.P. Stick-fixed N.P.

I
Yore rearward p&t*
of control stick re- More f&ard position
quired for steadyflight of control stick required
at lower speed, and for steady flight at
vice versa lower speed,andvlceveraa
More pull. force
More push force
required for steady
required for steady
flight at lower flight at lower speed,
speed, and vice
and vice versa
versa

Figure 9.- Chart describing the control characteristics of an airplane as a function of

center-of-gravity position. The arrows indicate location of center of gravity with
respect to neutral point,
08 NACA TN No. 1670
.

.-
i+ \ C*g. Position, percent M-A-c

10 -

F3ight data
=
-.
0. .-_I

I I I I I
100 200
Indicated airapead, Eph
(a) Varlatian of elevator angle with Indicated &-speed.
20 t

I c.g. position, percent H.A.C. /

..

0 .L .8 1.2 1.6
Lift aoefficient, CL
(b) Veriatlon af elevator angle with lift coefficient.
20

N8Utl'd point

0 1.
20 30
.-
c.g. position, percentY.A.C.
.
(c) Variation of the d%
quantity do, with oentWsavitg

position. .

Figure 10. - Method for determining stick-fixed neutral point from

I flight data.
 NACA TN Ro. 1670

c.g. position, percent Y.A.C.

Flight data

I I , , \\ c
loo 200
Indicated alrapeed, mph
(a) Variation of stick force with indicated airspeed.

I-

0 .4 .8 1.2 1.6
Lift coefficient, CL
(b) Variation of the quantity B/q with lift coefficient.

4-

.
. c.g. poaitfon, percent M.A.C.

(c) Variation of the quantity w with centerof-avity

+
. position. -

Figure 11~ Method for determining stick-free neutral point from

flight data..
 loo-
g(etralght flight)

Indicated eirspeed at 25,000 ft altitude, mph

Figure 12.- Typical example of effects of compressibility on the variation of stick force with
speed in steady flight and in flight with constant values of normal acceleration.

c .’ :
 \
Section A - A

Figure 1%- Typical dive-recovery-flap installation.

 Stabilizer incldenct
positive-
/
\ /
0
----c/

No distortion

100 200 300 400 500

Indicated airupeed, mph

Mgure 14.- Effect of ‘stabilizer incidence on the variation of stick force with speed in straight
flight. The variations in stick-force characteristics result from distortion of the elevator
covering and from stabilizer twist, Angles and distortions greatly exaggerated on sketches.
 93
NACA TN Ho. 1670

v _

(a) Airplane away from ground.

Real airplane

Cirouna

Image airplane w
(b) Airplane near ground.

Mgure 15.- Effect of image vortex system on downwash at tail as

airplane approaches the ground.
”

t
.
94 NACA TN No. 1670
-
.

..

Figure 16.- Effect of curvature of- flight path on the angle of attack at
the tail during a pull-up.

.
NACATN No. 1670 95

(a) IXLustrative variation of force per g with

center-of-gravity position.

\ -Stick-fixed maneuver point

\
\
\ (c-y; pc~~fon where -d6e = 0
\ eG
\
Original \
\
More positive C
ight requ%%Fpull

FW
O-
Forward Back
c.g. position
(b) EPfect of various design variables on the
variation of force per g with center-
of-gravity position.

Figure 17.- Graphs showing stick forces required inmaneuvers.

 96 NACA mv NO. 1670

20
Stick-fixed neutral point
in straight flight
M I

9 rut- - i

0 1 I
-20 -10 0
Fomard e.g. position, percent M.A.C. Bagk

\\
(a> Light airplane. -- _-

c\
Sea lqve)
4wCoft~,
I
1. t

01 I I\ \ I
Forward -20 -lo ' 1
c.g. position, percent M.A.C. B;ck
(b) Fighter airplane.

0”
r
r)n L

40,oCC it 4
\ k
\ \
Sea level!
I

I
t

o! I I \ \. I
-20 -10 \
Forward c.g. position, percent M.A.C. ' B:ck
(c) Bomber.

Figure 18.- Stick-force characteristics in maneuvers for three types of

airplanes with unbalanced elevators; airplane characteristics given
in table II. ch = -0.003 per degree; chs = -0.007 per degree.
a
 -.008,

-.006 -

-.004- w
M
8
ka

j -.002 - Approximate
uncertainty
indLue0fC~ _

0 ,_ I I I ----- I J
0 10 20 N ------
40 50 J -
Airplane weight, thousandsof lb

Figure 19.- Approximate reduction Ln c$, required to meet elevator-control-

i3
force requirements, as a function of airplane weight,
98 MACATN No. 1670

-.004

-.008
\

\
-.012

a.016

0.020

-.W
0 .2 4 .6 .8 1.0

Flap chord/airfoil chord, af/c .

.

Figure 20, - Variation of flap section hinge-moment parameters with II

ratio of flap chord to airfoil chord. Plain flaps with sealed gaps on .

NACA 0009 airfoil of infinite aspect ratio. Data from reference 17.
NACATH NO. 1670 99

k-- =b = ‘.
-+-
Round nom

Round or
elliptical nose
0

Round nose
Elliptical norm

c3 0.008

Figure 21.- Typical effects of overhang balances on control-surface

hinge-moment parameters. Derived from data of reference 19.
100 NMA TN NO. 1670

Control area

bo
2
!z
a
-.004

i
k

2 -.008

T!$i7-
-.012 I I I I
0 l 05 .lO .15 .20
Area moment of horn
Area moment of control
Figure 22. - Typical effects of unshielded horn balances on control-
surface hinge-moment parameters. Derived from data of
reference 20.
 NACATN No. 1670 101

.004

.- 0 /

tm
-2 / Ao.10
l

k
a
p.
20.10 or 0.20
-.004
i
0”
2 -.008

-.012
0 1.2
I

-
Figure 23.- Typical effects of full-span balancing tabs on control-
.
surface hinge-moment parameters. Derived from data of
reference 19.
102 NACATN No. 1670
.
.

.-
,-Gap sealed

.004

0 -.

f%l .
%
b
a
m -.004
8 .,
k0

2 -.008

-.Ol2
0 10 20 30
B ’ deg
.

Figure 24. - Typical effects of. beveled-trailing-edge balances .

on control-surface hinge-moment parameters, Derived from .
data of reference 19.
 NACATN No. 1670

I-

-.0x?
0 .2 .4 .6
Gf

Figure 25.- Typical effects of sealed internal balances on control-

surface hinge-moment parameters.
 .008’

F Unshielded horn balance

Beveled-trailing-edge balance
balance
.004
Elliptical-nose
E3Zpt,ical-nose overhang balance
Round-nose overhang balance
bdan~e

0
Sealed internal lmlanae

c Plain control surface I

-.cQ4

%i $
s!
Figure 26.- Comparison of effects of various aerodynamic balances on hinge-moment 2.
parameters of typical control surface. P
3

‘. ’ .
. . .
 NAC!ATN No. 1670 105
-
i
2ho
-. a” Fin 0'
\ Fin 3’ left -- - -

lower1\ __
I
P
-L- ----- ----- --- - - E

Power on

Pouer

Pouer off

Indicated airspeed, mph

v Figure 27.- Typical variations of rudder angle, rudder force, and

.
sideslip angle with speed in straight flight with wings laterally level.
Single -engine tractor airplane.
 Thrust m propeller blades
Thrust on prcpaller
bladea

Farce on fuselags

hag on aileron0

Flow dire&Ion

OIV rudder deflection

Sidsslip = 0 Sideslip to left
Side force tc left Side form = 0
Airplane in equilibrium

figure 28.- Forces ahd moments acting on sfngle-engine tractor airplane in flight at
high angles of attack with wbgs laterally level. Propeller rotation clockwise when
viewed from the reax.

c
’ 1
.’ :
NACATIVNo. 1670

Figure 29.- Forces acting on an airplane in a steady sideslip.

 .

108 NACA TN NO. 1670

.4
r
0

.4

-1.2' I 1
0 30 60
Angle of yaw, deg

Figure 30. - Effect of..sm.alJ fins 011-tie yawing moments of a

fuselage with circular cross section.

.
.
 HACATN NO. 1670
-
.

-.

Figure 31.- Illustration of cause of unstable variation of yawing

moment with sideslip for a wing with dihedral.
110 NACA TN NO. 1670

..

Tail stalls or emerges

from slipstream r

Angle of yaw, deg Right

Figure 32.- Variation with sideslip of the yawing motients contributed

by the propeller, fuselage, and vertical tail for a single-engine
tractor airplane.

.
.
 NACATN No. 1670 111

-With dorsal fin

(no tip fins)
-k ---
-Ah
With tip
fins Approximate proportions
-3 of added areas
/

.
Original
model

Angle of yaw, deg

Figure 33.- Wind-tunriel measurements showing effect of a dorsal

fin and of vertical fins on the tip of the horizontal tail on the
directional stability characteristics of a single-engine fighter
airplane. Power-on condition (contrarotating propellers),
rudder free.
--

.
.
112 ITACATN NO. 1670
.
.

*-

-.

Large increase in lift on

trailing wing caused by
increased flap area in
slipstream
.

Rolling moment to left

‘Figure 34. - Illustration of unstable dihedral effect caused by power

( tractor -type airplane >.

.
 mcA TN NO. 1670 113
.
.

-.

Oo Sweepback

150 Sweepback- ---

Flaps up
10 I- /7-

f O/ ,<
Flaps down

,-
I
0 . .

Lift coefficient, CL

Figure 35.- Effect of a moderate amount of sweepback on the

variation of effective dihedral with Uft coefficient. Single-
engine tractor airplane; power-on condition.

.
 of the formula for helix angle.

. .
. : *’
 NACA TN No. 1670 115
I

*
20 -
%
r;! .

0
\

,
i2 Rolling Rolling

Ob d
-1 Yawing O "%- Yawing

I I I
0 2 4
Time, set Time, set
(a) I.40 miles per hour. (b) 200 miles per hour,

Figure 37. - Time histories oftypicalrudder-fixed aileronrolls in

a medium-bomber airplane.
. +
.

.
.
 r Reduction due to
sldeslip
,Wl aileron
rigid wing
deflection,

pull aileron deflection,

actual ah-plane (Yaqying
stick force)

Indicated airspeed-
i .
=:
P

Figure 38.- Typical variation of aileron effectiveness with speed.

.
.
,’ I
 I?ACATN No. 1670 117

Original aileron travel

-<Reduced aileron travel

. 30 lb stick force

0
Indicated aIrspeed-

Figure 39.- Effect of changing aileron travel while keeping same

stick travel on tie variation of aileron effectiveness with speed.
118 N.KA Tm NO. 1670

..

Taper ratio 0.5

Aspect ratio 6
Aileron root-mean-square chord 1 ft
Aileron travel &15o -icpggr
Stick travel +9 in.

FYigure 40. - Airplane dimensions used in example for calculation

of aileron control characteristics.
 XACATN NO. 1670
I

---
desired)
(llUXXiHlUlll

--

Full deflection
\
\
30 lb stick force
'\<
-- --4
a
0 I I 1
loo 200 300
Indicated airspeed, mph

Figure 41.- Variation of stick force and aileron deflection with

speed for airplane used as example.
120 NACATN No. 1670

10000

8000

2 4000
8
g
2000

0
Indicated airspeed, mph

Figure 42.~ Aileron wheel force for full aileron deflgctiog as a

function of speed for airplane with 240-foot span. Unbalanced
ailerons.

.
.
 ’ . ‘a

Figure 43.- Illustration of rudder and airplane motion during a snaking oscillation.
122 NACA TN No. 1670

at .7!%, dog

-.

-
.
.

l.O+-

.6-

57 l 4-

.2 - Blade angle
35 at .75R, deg
0
0 .2 .I .6 .8 1.0 1.2 1.4 1.6 1.8 2.0
.
V/nD

Figure 44.- Typical charts showing propeller characteristics.

Figure 45.- Forces acting on an a&plane in a steady climb.
 -.
124 NACATN No. 1670

.-

+ -5 Stabilizer incidence, it, dog

1.2-

.8- Lift coefficient for trimmed conditions

CY Stabilizer incidence, it, dog

deg
.A-

0 I 1 I 1
0 4 & 12 16
a, da

Figure 46.- Method of determining approximate variation of lift

coefficient for trimmed conditions with angle of attack for
preparation of propeller operating charts.
 NACATN NO. 1670 125

-.

(a) Tricycle landing gear.

.-

.
.

(b) Conventional landing gear.

Figure 47.- Calculation of ground reaction moments.

1.26 NACATN No. 1670

-.
Figure 48. - Diagram illustrating calculation of moments about
point (n) when forces and moments about point (1) are given. .
.
NACATN NO. 1670

Data for c.g. = 25 percent M.A.C.

.l Stabilizer setting, deg

l-

CT 0

-0 1

1 I I I I I I I i I ! I I
0 .2 .4 .6 .8 1.0 1.2
CL

Figure 49.- Wind-tunnel test data for determination of the stick-

fixed neutral point. Neutral points determined directly at
cL = 1.0 and 0.6.
128 NACA TN No. 1670
.

Along this line,

.1 CL - 0.3 “-

-. 2

-
-. 2 -0 1 0 .l .2
-.
Cm/CL
(a) Method of extrapolation of slopes.
.
.

.1 Stabilizer setting, dog

Lv
.
0 .-
0s

-. 1

I I I I I f I I I I 1 t I I
-. 4 -. 2 0 .2 .4 .6 .8 1.0
CL .
(b) Method of tangents.
..
.
Figure 50.- Graphical procedures for determination of stick-fixed
neutral point from wind-tunnel tests.