Simplified Flutter Prevention Criteria for Personal Type

Aircraft

Robert Rosenbaum1 and A. A. Vollmecke2

Civil Aeronautics Administration, Washington, D.C.

This report3 is intended to serve as a guide to the small plane designer in the presentation of design
criteria for the prevention of such aeroelastic phenomena as flutter, aileron reversal and wing divergence.
It should also serve as a guide to recommended and acceptable practice for the design of non-structural,
mass balance weights and attachments. The criteria developed in this report include: wing torsional
rigidity; aileron, elevator and rudder mass balance; reversible tab and balance weight attachment
criteria.

Nomenclature
𝑐 = chord
𝐹 = torsional flexibility factor
𝐾 = product of inertia
𝐼 = mass moment of inertia
𝐾/ = dynamic balance coefficient
𝐼
𝛾 = balance parameter
𝜆 = balance parameter
𝑉𝑓 = flutter speed parameter
𝑉𝑑 = design dive speed (IAS) of the airplane

I. Introduction

T HE simplified criteria appearing in CAM 04 were developed at a time when rational methods of flutter analysis
were not available. Because of the lack of available methods of analysis various attempts were made to set up
empirical formulae which, if complied with, would reasonably assure freedom from flutter. The sources of material
for these studies were threefold:
1) A statistical study of the geometric, inertia and elastic properties of those airplanes which had experienced
flutter in flight, and the methods used to eliminate the flutter.
2) Limited wind tunnel tests conducted with semi-rigid models. These models were solid models of high
rigidity so that effectively the model was non-deformable. The motion of the models was controlled by
attaching springs at the root and at the control surface to simulate wing bending, torsion and control surface
rotation.
3) Analytic studies based on the two dimensional study of a representative section of an airfoil.

For the most part these studies indicated that for a conventional airfoil in which the center of gravity of the
airfoil section is not too far back, that wing flutter could be prevented by designing for a certain degree of wing
torsional rigidity and by control surface dynamic balance, whereas empennage flutter could be prevented by
providing a degree of control surface dynamic balance. The limitations were based on the design dive speed of the
airplane and within certain ranges were functions of the ratio of control surface natural frequency to fixed surface
frequency.
Satisfactory rational analytic methods have been available for a number of years which would permit an engineer
to carry through computations to determine the flutter stability of a specific design. In view of the fact that flutter is
an aeroelastic phenomenon which is caused by a combination of aerodynamic, inertia and elastic effects, any criteria
1
Chief, Dynamics and Loads Section
2
Chief, Airframe and Equipment Engineering Branch
3
This 1955 report is a continuation of the series of reports which previously appeared as Aircraft Airworthiness
Reports and Engineering section Reports. Adapted to digital format by Jamie Halford, BSMET Undergraduate,
University of North Carolina at Charlotte.
1
Airframe and Equipment Engineering Report No. 45
which does not consider all three effects is bound to have severe limitations. That this is so, is evidenced by the fact
that in almost all cases where rational analyses have been carried thru for specific designs it has been found that the
balance requirements specified by the simple criteria have been too severe. In some special cases the criteria in
CAM 04 appear to have been unconservative, i.e. flutter has been encountered in some airplanes which complied
with these criteria. In spite of the fact that the old flutter prevention criteria for the most part yield over-conservative
results most small aircraft companies in the personal plane field prefer to comply with these criteria rather than
perform complex flutter analyses. In order to aid the small. manufacturer the CAA in October 1948 issued Airframe
and Equipment Engineering Report No. 43, entitled, “Outline of An Acceptable Method of Vibration and Flutter
Analysis for a Conventional Airplane”. The purpose of that report was to present to the inexperienced flutter analyst
an acceptable, three dimensional method of analysis by presenting in detail a step-by-step tabular technique of
analysis. Although a number of aircraft companies are using the methods outlined in the report, others are of the
opinion that this method entails too much time and expense and are therefore seeking other means of complying
with those regulations which require them to show freedom from flutter.
Although a rational flutter analysis is to be preferred to the use of the simplified criteria contained herein (since
in most cases a better design may be achieved by reducing or eliminating the need for nonstructural balance
weights), the application of these criteria to conventional aircraft of the personal plane type is believed to be
adequate to insure freedom from flutter.
The criteria contained in the present report have been developed after an exhaustive study of the American and
British literature as well as independent investigations. For the moat part the criteria contained, in this report are
new, however, some have been taken with little or no modification from other sources.
It should be noted that the empennage criteria developed in this report, have been developed on the basis of a
single representative (conservative) value of the empennage mass moment of inertia about the bending axes. The
value was chosen as a result of a study of the mass parameters of a number of airplanes of the personal plane type.
Therefore, for larger .03 aircraft than those usually classified as personal planes the criteria may not be applicable.
The wing criteria on the other hand should be applicable to all conventional .03 airplanes which do not have large
mass concentrations on the wings.
The criteria developed in this report are of a preliminary nature, and although considered to represent current
thinking on acceptable and recommended practices regarding flutter prevention measures for personal type
airplanes, these criteria should not be construed as required procedure to meet the flutter prevention requirements of
the Civil Air Regulations.

A. Definitions
Flutter: Flutter is the unstable self-excited oscillation of an airfoil and its associated structure, caused by a
combination of aerodynamic, inertia and elastic effects in such manner as to extract energy from the airstream. The
amplitude of oscillation, (at the critical flutter speed) following an initial disturbance will be maintained. At a higher
speed these amplitudes will increase.
Divergence: Divergence is the static instability of an airfoil in torsion which occurs when the torsional rigidity of
the structure is exceeded by aerodynamic twisting moments. If the elastic axis of a wing is aft of the aerodynamic
center then the torsional moment about the elastic axis due to the lift at the aerodynamic center tends to increase the
angle of attack, which further increases the lift and therefore further increases the torsional moment. For speeds
below some critical speed (the divergence speed), the additional increments of twist and moment become smaller so
that at each speed below the divergent speed an equilibrium position is finally attained (i.e. the process of moment
increasing angle and thereby increasing moment etc. is convergent); above this critical speed the process is non-
convergent.
Control Surface Reversal: This is the reversal in direction of the net normal force induced by the deflected
control surface, due to aerodynamic moments twisting the elastic “fixed” surface. This phenomenon can best be
illustrated by considering the case of aileron reversal. Normally the lift over the wing with down aileron is increased
by the aileron deflection, while the lift over the wing with up aileron is decreased by the aileron deflection, thus a
rolling moment results from an aileron deflection. However, since the center of pressure for the lift due to the
deflected aileron is usually aft of the elastic axis, deflecting the aileron downward tends to reduce the wing angle of
attack thus reducing the increment of lift. For the wing with up aileron the torsional moment due to up aileron tends
to increase the wing angle of attack. It can thus be seen that the rolling moment for an elastic wing is less than for a
rigid wing. Since the wing torsional rigidity is constant while the twisting moment due to aileron deflection
increases with the square of the velocity it is obvious that at some critical speed the rolling moment due to aileron
deflection will be zero. Above this speed the rolling moment will be opposite to that normally expected at speeds
below this critical speed. The critical speed so defined is the aileron reversal speed.
2
Airframe and Equipment Engineering Report No. 45
 II. Summary of Criteria

A. Wing Torsional Stiffness

The wing torsional flexibility factor 𝐹 defined below should be equal to or less than

200
(1)
𝑉𝑑 2

Where:

𝐹 = ∫ 𝜃𝑖 𝐶𝑖 2 𝑑𝑠 (2)

𝜃𝑖 = Wing twist at station 𝑖, per unit torsional moment applied at a wing station outboard of the end of the
aileron, (radians/ft-lb)
𝐶𝑖 = Wing chord length at station 𝑖, (ft)
𝑑𝑠 = Increment of span (ft)
𝑉𝑑 = Design dive speed (IAS) of the airplane

Integration to extend over the aileron span only. The value of the above integral can be obtained either by
dividing the wing into a finite number of spanwise increments 𝛥𝑠 over the aileron span and summing the values of
𝜃𝑖𝐶𝑖2𝛥𝑠 or by plotting the variation of 𝜃𝑖𝐶𝑖2𝛥𝑠 over the aileron span and determining the area under the resulting
curve.
In order to determine the wing flexibility factor 𝐹 , a pure torsional couple should be applied near the wing tip
(outboard of the end of the aileron span) and the resulting angular deflection at selected intervals along the span
measured. The test can best be performed by applying simultaneously equal and opposite torques on each side of the
airplane and measuring the torsional deflection with respect to the airplane centerline. The twist in radians per unit
torsional moment in ft-lb should then be determined, If the aileron portion of the wing is divided into four spanwise
elements and the deflection determined at the midpoint of each element the flexibility factor 𝐹 can be determined by
completing a table similar to Table 1 below. Figure 1 illustrates a typical setup for the determination of the
parameters 𝐶 and 𝛥𝑠.

Aileron Span

C11
C C2 C3 C4
C
C 2 C 3 4

∆S1 ∆S ∆S ∆S
∆S1 ∆S22 3
∆S3 4
∆S4

Figure 1. Typical setup for measuring torsional deflection.

3
Airframe and Equipment Engineering Report No. 45
 Table 1. Determination of Wing Flexibility Factor.
(1) (2) (3) (4) (5) (6)
STATION Δs c c2 θ θc2Δs
ft ft ft2 rad
ft-lb
1
2
3
4
F = ∑ column (6)

B. Aileron Balance Criterion

The dynamic balance coefficient 𝐾/𝐼 should not be greater than the value obtained from Fig. 2 wherein 𝐾/𝐼 is
referred to the wing fundamental bending node line and the aileron hinge line. If no knowledge exists of the location
of the bending node line the axis parallel to the fuselage center line at the juncture of the wing and fuselage can be
used.
Wherein:
𝐾 = product of inertia
𝐼 = mass moment of inertia of aileron about its hinge line

C. Free Play of Ailerons

The total free play at the aileron edge of each aileron, when the other aileron is clamped to the wing should not
exceed 2.5 percent of the aileron chord aft of the hinge line at the station where the free play is measured.

D. Elevator Balance
Each elevator should be dynamically balanced to preclude the parallel axis flutter (fuselage vertical bending-
symmetric elevator rotation) as well as perpendicular axis flutter (fuselage torsion – antisymmetric elevator
rotation). If, however, the antisymmetric elevator frequency is greater than 1.5 times the fuselage torsional
frequency the perpendicular axis criterion need not apply.

E. Parallel Axis Criterion

The balance parameter 𝛾 as obtained from figure 3 should not be exceeded. In Fig. 3 the balance parameter γ and
the flutter speed parameter 𝑉𝑓 are defined as:

𝑏𝑆𝛽
𝛾= 𝐼
(3)

𝑉
𝑉𝑓 = 𝑏𝑓𝑑 (4)
ℎ

Where:
𝑆𝛽 = Elevator Static Balance
𝐼 = Elevator mass moment of inertia about the hinge line (lb-ft2)
𝑏 = Semichord of the horizontal tail measured at the midspan station (ft)
𝑉𝑑 = Design dive speed of the airplane (mph)
𝑓ℎ = Fuselage vertical bending frequency (cpm)

4
Airframe and Equipment Engineering Report No. 45
F. Perpendicular Axis Criterion
For each elevator the balance parameter 𝜆 as obtained from Fig. 4 should not be exceeded. In Fig. 4 the balance
parameter 𝜆 and the flutter speed parameter 𝑉𝑓 are defined as:

𝜆 = 𝑏𝐾
𝑆𝐼
(5)

𝑉
𝑉𝑓 = 𝑏𝑓𝑑 (6)
𝛼

Where:
𝑆 = Semispan of horizontal tail (ft)
𝑏 = Semichord of horizontal tail at midspan station (ft)
𝐾 = Elevator product of inertia referred to stabilizer center line and elevator hinge line (lb-ft2)
𝐼 = Elevator mass moment of inertia about the elevator hinge (lb-ft2)
𝑓𝛼 = Fuselage torsional frequency (cpm)

G. Rudder Balance
The value of 𝛾 as obtained from Figure 3 and the value 𝜆 as obtained from Fig. 4 should not be exceeded; where
in Figs. 3 and 4,

𝑏𝑆𝛽
𝛾= 𝐼
(5)

𝜆 = 𝑏𝐾
𝑆𝐼
(6)

and:
𝑆 = Distance from fuselage torsion axis to tip of fin (ft)
𝑏 = Semi-chord of vertical tail measured at the seventy percent span position (ft)
𝐾 = Product of inertia of rudder referred to the fuselage torsion axis and the rudder hinge line (lb-ft2)
𝑓𝛼 = Fuselage torsional frequency (cpm)
𝑓ℎ = Fuselage side bending frequency (cpm)
𝑆𝛽 = Rudder static balance about hinge line (lb-ft)
𝐼 = Mass moment of inertia of the rudder about hinge line (lb-ft2)

H. Tab Criteria
All reversible tabs should be 100% statically mass balanced about the tab hinge line. Tabs are considered to be
irreversible and need not be mass balanced if they meet the following criteria:
1) For any position of the control surface and tab no appreciable deflection of the tab can be produced by
means of a moment applied directly to the tab, when the control surface is held in a fixed position and the
pilots tab controls are restrained.
2) The total free play at the tab trailing edge should be less than 2.5% of the tab chord aft of the hinge line, at
the station where the play is measured.
3) The tab natural frequency should be equal to or exceed the value given by the lower of the following two
criteria.

𝑉 𝑆
𝑓𝑡 = 48 𝐶𝑑𝑙 𝑆 𝑡 cpm (7)
𝑐

or

2000 cpm, 𝑉𝑑 ≤ 200 mph

𝑓𝑡 = { (8)
(10 × 𝑑𝑒𝑠𝑖𝑔𝑛 𝑑𝑖𝑣𝑒 𝑠𝑝𝑒𝑒𝑑) 𝑐𝑝𝑚, 𝑉𝑑 > 200 𝑚𝑝ℎ

5
Airframe and Equipment Engineering Report No. 45
 Thus for an airplane with a design dive speed less than 200 mph if Eq. (7) above gave a value in excess of 2000
cpm it would only be necessary to show a frequency of 2000 cpm for the frequency criterion.

Where:
𝑓𝑡 = lowest natural frequency of the tab as installed in the airplane (cpm) - either tab rotation about the
hinge line or tab torsion whichever is lower.
𝐶𝑙 = chord of moveable control surface aft of the hinge line, at the tab midspan position (ft)
𝑆𝑡 = Span of tab (ft)
𝑆𝑐 = Span of moveable control surface to which tab is attached (both sides of elevator, each aileron and
rudder) (ft)

Particular care should be taken in the detail design to minimize the possibility of fatigue failures which might
allow the tab to become free and flutter violently.

I. Balance Weight Attachment Criteria

Balance weights should be distributed along the span of the control surface so that the static unbalance of each
spanwise element is approximately uniform. However, where a single external concentrated balance weight is
attached to a control surface of high torsional rigidity the natural frequency of the balance weight attachment should
be at least 50 percent above the highest frequency of the fixed surface with which the control surface may couple in
a flutter mode. For example, the aileron balance weight frequency should be at least 50% above the wing
fundamental torsional frequency. The balance weight supporting structure should be designed for a limit load of 24g
normal to the plane of the surface and 12g in the other mutually perpendicular directions.
It should be noted that the dynamic balance coefficient 𝐾/𝐼 can be reduced by (1) reducing 𝐾, (2) increasing 𝐼
or (3) reducing 𝐾 and increasing 𝐼. Since an increase in 𝐼 results in a reduced control surface natural frequency with
possible adverse flutter effects, the primary purpose of ballast weights used to reduce 𝐾/𝐼 , should be to decrease the
product of inertia 𝐾 and not to increase the mass moment of inertia 𝐼.

6
Airframe and Equipment Engineering Report No. 45
 Fig. 2
Aileron Balance Criterion

4.0
Dynamic Balance Coefficient, K/I

f(x) = ‐0.016x + 4.8
3.0

2.0

1.0

0.0
0 50 100 150 200 250 300
VD = Design Dive Speed, MPH

7
Airframe and Equipment Engineering Report No. 45
 1.2 Figure 3
Elevator and Rudder Balance Criteria
Parallel Axes
(Fuselage Bending - Control Surface)
1.0
Rudder
143333 49716 5764.4 224.35 0.11
10 10 4 10 901902 104470 6452.9 167.6 0.11 0.19
898.65 929.74 367.29 68.371 5.1739 0.19
Mass Balance Parameter, (bSβ)/I

0.8

Elevator
803472x6 − 798159x5 + 325171x4 − 69578x3 + 8303.3x2 − 538.7x + 16.035

0.6

0.4

Units
b - ft
Sβ - lb-ft
0.2 I - lb-ft2
VD - mph
Fh - cpm

0.0
0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 0.24 0.26
Flutter Speed Factor, Vd / (bfh )
8
Airframe and Equipment Engineering Report No. 45
 2.4 Figure 4
Elevator and Rudder Balance Criteria
Perpendicular Axes
(Fuselage Torsion - Control Surface)
2.0

1.6
Mass Balance Parameter, bK/(SI)

483058 320420 79195 8666.2 356.03, 0.151

240861 355895 218345 71284 13095 1291.5 54.537, 0.151 0.2
5158.3 10242 8236.3 339.4 741.9 75.22 1.7003, 0.2

1.2

0.8 Units
b - ft
S - ft
K - lb-ft2
I - lb-ft2
VD - mph
0.4 f - cpm

0.0
0.12 0.16 0.20 0.24 0.28 0.32 0.36 0.40 0.44 0.48

Flutter Speed Factor, Vd /(bfα )

9
Airframe and Equipment Engineering Report No. 45
 III. Dynamic and Static Balance of Moveable Control Surfaces

A. Definitions
Static Balance: Complete static balance of a moveable control surface is obtained when the center of gravity of
the control surface lies on the hinge line i.e. the resultant moment of the mass of the surface about the hinge line is
zero. If the center of gravity of a surface lies aft of the hinge surface it is called statically unbalanced, whereas if the
center of gravity lies forward of the hinge line the surface is called statically over-balanced.
Dynamic Balance: A moveable surface is dynamically balanced with respect to a given axis if an angular
acceleration about that axis does not tend to cause the surface to rotate about its own hinge line. The dynamic
balance coefficient 𝐾/𝐼 is a measure of the dynamics balance condition of the moveable control surface, wherein 𝐾
is the product of inertia of the surface (including balance weights) about the hinge and oscillation axes and 𝐼 is the
mass moment of inertia of the control surface (including balance weights) about the hinge axis. Physically the
dynamic balance coefficient 𝐾/𝐼 may be interpreted to represent:

𝐸𝑥𝑐𝑖𝑡𝑖𝑛𝑔 𝑇𝑜𝑟𝑞𝑢𝑒
(9)
𝑅𝑒𝑠𝑖𝑠𝑡𝑖𝑛𝑔 𝑇𝑜𝑟𝑞𝑢𝑒

Mass Balance Computations

Assume the X axis coincident with the oscillation axis and the Y axis coincident with the control surface hinge
line. After the reference axes have been determined the surface should be divided into relatively small parts and the
weight of each part 𝑊 and the distance from its c.g. to each axis tabulated. See Fig. 5 and Table 2. Referring to Fig.
5 the static moment of the element 𝛥𝑊 is 𝛥𝑊𝑥, the moment of inertia about the hinge line is 𝛥𝑊𝑥2 and the product
of inertia is 𝛥𝑊𝑥𝑦. The static unbalance of the total surface 𝑆𝛽 is then ∑𝛥𝑊𝑥𝑦; the moment of inertia of the surface
is ∑𝛥𝑊𝑥2 and the product of inertia is 𝐾 = ∑𝛥𝑊𝑥𝑦.

+Y Note: + Y-axis is taken

on same side of X-axis
as is CP of maneuvering
CP of maneuvering load on surface.
load distribution

90°
ΔW
x

90°
+X
Note: +X-axis
-X Hinge axis
taken rearward

Axis of oscillation
-Y
Figure 5. Elements of mass balance computations.

10
Airframe and Equipment Engineering Report No. 45
 Table 2. Calculation of the moment and product of inertia.
Dist. From Moment - wx Dist. from K=wxy
Weight I=wx3
Item Part hinge CL oscillation
Description w x2 - +
No. No. x axis – y - +
lbs. inch-lbs. inch-lbs. lb.-in.3
inches in.
(1) (2) (3) (4) (5) (6) (7) (8) (9)
1
2
3
etc.

B. Product of Inertia with respect to Other Axes

Having determined the product of inertia with respect to one oscillation axis it may be desirable or necessary to
determine the product of inertia with respect to some other oscillation axis. If the product of inertia was originally
calculated for an oscillation axis which was perpendicular to the hinge axis then the product of inertia with respect to
inclined axes O-O and Y-Y can be determined from the perpendicular axes product of inertia (X-X and Y-Y) by use
of the following equation:

𝐾𝑜𝑦 = 𝐾𝑥𝑦 𝑠𝑖𝑛 𝜙 − 𝐼𝑦𝑦 𝑐𝑜𝑠 𝜙 (10)

O x cos Φ

X X
O
Φ

y
y sin Φ

C.G. of W

Y
Figure 6. Product of inertia with respect to an inclined axis.
11
Airframe and Equipment Engineering Report No. 45
 Where: 𝛷 is the angle between O-O and Y-Y in the quadrant where the center of gravity of the surface is
located.

If the product of inertia was originally calculated for one set of axes and it is desired to determine the product of
inertia for another set of axes parallel to the original set, then the new product of inertia 𝐾2 can be determined from
the equation:

𝐾2 = 𝐾1 + 𝑥𝑜 𝑦𝑊 + 𝑦𝑜 𝑥𝑊 + 𝑥𝑜 𝑦𝑜 𝑊 (11)

Where:
𝑊 = total weight in pounds of the moveable surface
𝐾1 = product of inertia with respect to axes X1-Y1
𝑥𝑜 = distance between X1 and X2 axes
𝑦𝑜 = distance between Y1 and Y2 axes
𝑥 = distance from C.G. of surface to X1 axis
𝑦 = distance from C.G. of surface to Y1 axis

Y1
Y2 C.G. of
x surface

xx0
y

X1 X1

Y1
yy0

X2 X2

Y2

Figure 7. Product of inertia with respect to a parallel axis.

12
Airframe and Equipment Engineering Report No. 45
 IV. Experimental Determination of Static Unbalance,
Moment of Inertia and Product of Inertia

A. Static Unbalance

The moveable control surface should be carefully supported at its hinge line on knife edges or in a jig with a
minimum of friction. The force necessary to balance the control surface, when applied to a given point, is then
measured by an accurate weighing scale. The net force times the distance between the hinge line and the point of
application of the force is equal to the static unbalance.

Weighing Scale

Distance between
point of application of
balancing force and hinge line

Figure 8. Static unbalance or torque

B. Moment of Inertia
The experimental determination of the mass moment of inertia consists of supporting the surface or tab at the
binge line with a minimum of friction in a jig in an attitude similar to that described above and maintaining it in this
attitude by means of one or two springs, as shown in Fig. 9. One spring is sufficient for control surfaces with large
static unbalances, while two are generally used for surfaces which are fairly well statically balanced. The natural
frequency of the surface (for small oscillations) under the restraining action of the springs is then measured by
means of a stop watch by determining the time necessary for a given number of cycles. In order to reduce
experimental errors to a minimum, the time for a large number of cycles (about 30) is measured.
The spring stiffnesses are dynamically determined by placing a weight W1 on spring 1 which will deflect it an
amount approximately equal to the average spring deflection during the moment of inertia test and then determining
the natural frequency of the spring with W1 attached by determining with a stop watch the time necessary for a given
number of cycles; a similar test is conducted for the determination of the spring stiffness of spring 2, using a weight
𝑊2 .

13
Airframe and Equipment Engineering Report No. 45
 Spring No. 1

C.G.
x

Spring No. 2

Figure 9. Experimental determination of mass moment of inertia.

The moment of inertia can then be calculated by substituting the test results in either equation 1 or 2 depending
on whether the control surface center of gravity is above or below the binge axis.
If control surface center of gravity is below hinge axis:
2 𝑊0 𝑥
𝐼 = 𝑓𝑑 2 (𝑊1 𝑓1 2 + 𝑊2 𝑓2 2 ) + 9.788 𝑓0 2
(12)
0

If control surface center of gravity is above hinge axis:

2 𝑊0 𝑥
𝐼 = 𝑓𝑑 2 (𝑊1 𝑓1 2 + 𝑊2 𝑓2 2 ) − 9.788 𝑓0 2
(13)
0

Where:
𝐼 = Moment of inertia of surface about hinge axis (pound-inches2)
𝑊0 = Weight of surface (pounds); 𝑊1 , 𝑊2 Spring calibration weights (pounds)
𝑥 = Distance of surface C.G. above or below hinge axis (inches)
𝑑 = Distance from hinge axis to springs (inches)
𝑓0 = Frequency of surface when restrained by springs (cps)
𝑓1 = Calibration frequency of spring 𝐾1 under weight 𝑊1 (cps)
𝑓2 = Calibration frequency of spring 𝐾2 under weight 𝑊2 (cps)

C. Product of Inertia
The product of inertia 𝐾𝑥𝑦 of a moveable control surface can be calculated from three experimentally determined
moments of inertia. If the control surface moments of inertia are obtained by oscillating about each of the axes X-X,
Y-Y and then about a third axis O-O lying in the XY plane and making an angle  with the X-X axis, then the
product of inertia 𝐾𝑥𝑦 , is obtained from:

𝐼𝑥𝑥 𝑐𝑜𝑠2 𝛼+𝐼𝑦𝑦 𝑠𝑖𝑛2 𝛼−𝐼𝑜𝑜

𝐾𝑥𝑦 = 2 𝑠𝑖𝑛 𝛼 𝑐𝑜𝑠 𝛼
(14)

14
Airframe and Equipment Engineering Report No. 45
 Y

X X

Figure 10. Determining product of inertia.

Since this method of determining the product of inertia involves small differences between large quantities a
small experimental error in the determination of the moments of inertia may result in large errors in the product of
inertia. It can be shown (ACIC No. 711 “The Determination of the Product of Inertia of Aircraft Control Surfaces”),
that the error can be reduced to acceptable levels by the proper choice of the angle 𝛼. The proper value of  can be
determined after having determined 𝐼𝑥𝑥 and 𝐼𝑦𝑦 ; this value is given by the relationship:

𝐼𝑥𝑥
𝛼 = 𝑡𝑎𝑛−1 (15)
√ 𝐼𝑦𝑦

Appendix I - Discussion of Empennage Flutter Criteria

Studies made by the Air Material Command the Civil Aeronautics Administration and many independent
investigators have shown that for the most part empennage flutter modes can be closely associated with control
surface unbalance and the appropriate fuselage natural frequency with which the control surface will couple. Thus,
in the case of elevator coupling, for the most part, the fuselage vertical bending mode enters into the motion of the
system whereas for the rudder either fuselage side bending or torsion will couple. Although it is fully realized that
any analysis based on this type of simplification would of necessity be only approximate, it should be noted that the
results obtained are usually highly conservative, since other modes which generally enter into the motion of the
complete system tend to damp the motion with a resultant higher flutter speed. Thus, the fuselage vertical bending
mode is generally damped by coupling with wing symmetric bending and stabilizer bending whereas fuselage side
bending motion is usually damped by coupling with fuselage torsion the antisymmetric bending of the stabilizer and
bending of the fin.

15
Airframe and Equipment Engineering Report No. 45
 Based on these considerations the Air Material Command prepared a report Army Air Forces Technical Report
No. 5107 entitled “Charts for Fuselage Bending vs. Control Surface Flutter”. It has been found that these charts are
applicable to larger aircraft than those considered in the personal plane field. Each chart in AAFTR 5107 shows the
𝑉
variation of 𝜇𝑟𝛼 2 𝑔𝛼 with 𝑏𝜔 for various values of 𝜇𝑥𝛽 . Unfortunately the limits of the values of the parameter 𝜇𝑟𝛼 2 𝑔𝛼
used, are such that for most airplanes in the personal plane field these curves cannot be read with any degree of
accuracy, without doubtful extrapolation. Furthermore, it was considered that for simplicity a single curve would be
more suitable in treating the relatively low performance personal plane field, than a family of curves.

A. Fuselage Bending - Control Surface Rotation*

The following assumptions were made in the determination of the fuselage bending control surface rotation
flutter criterion (Fig. 3):
1) 𝜔 = 𝜔𝛼
2) 𝜔𝛽 = 0
3) 𝑐 − 𝑒 = 0
4) 𝑎 = −3
5) 𝜇𝑟𝛼 2 𝑔𝛼 = 5 for elevator rotation vs. fuselage vertical bending
= 8 for rudder rotation vs. fuselage side bending

The above assumptions are believed to be rational and valid for most aircraft in the field under consideration.
Justification for each of the above assumptions is given below.

1) The flutter frequency 𝜔 is equal to the fuselage bending frequency 𝜔𝛼 . Experience has shown that because
of the relatively large inertia of the tail the aerodynamic and inertia coupling terms are comparatively small.
The flutter frequency is therefore very close to the fuselage bending frequency.
2) For conventional aircraft with no springs in the control system the natural frequency of the empennage
control surfaces is zero. For the most part conventional tail control systems are so rigged that elastic
deformation in the control system takes place only if the controls are locked in the cockpit. Since under
actual flight conditions the pilot restraint in the cockpit is small, the assumption of 𝜔𝛽 = 0 is considered to
be valid.
3) In the low performance field it has been found that most control surfaces are not aerodynamically balanced.
Since in general an increase in aerodynamic balance will tend to increase the critical flutter speed this
assumption will yield conservative results for aircraft with aerodynamically balanced surfaces and yield
correct results for those aircraft with no aerodynamic balance.
4) The flutter mode involving fuselage bending and control surface rotation is analogous to the wing torsion-
aileron rotation case with the effective fuselage bending axis corresponding to the wing elastic axis. This
axis of rotation is the effective point about which the airfoil section (stabilizer-elevator or fin-rudder)
rotates when the fuselage bends and is not the nodal line of the fuselage in bending. A study made by the
Air Material Command from vibration measurements of a large number of airplanes indicates that the
effective fuselage bending axis is located approximately 1.5 tail surface chord lengths ahead of the tail
surface mid-chord (i.e. 𝑎 = −3.0)
5) An examination of the values of the parameter 𝜇𝑟𝛼 2 𝑔𝛼 for the empennage of a number of small airplanes of
the .03 type indicates that this parameter is small varying approximately between 4 and 8 at low altitudes
(based on 𝑔 = 0.03). For the case of fuselage side bending it has been found that the effective increase in
mass moment of inertia of the fuselage due to wing yawing is approximately 75% of the empennage mass
moment of inertia. By assuming 𝜇𝑟𝛼 2 𝑔𝛼 constant, one curve of allowable mass balance parameter versus
flutter speed parameter can be calculated for each value of 𝑒 thus simplifying the problem. The values of
𝜇𝑟𝛼 2 𝑔𝛼 = 5 for fuselage vertical bending and 𝜇𝑟𝛼 2 𝑔𝛼 = 8.75 for fuselage side bending are believed to be
representative, conservative values for .03 airplanes.

*
The notation used in this section is similar to that appearing in CAA Airframe and Equipment Engineering Report
No. 43
16
Airframe and Equipment Engineering Report No. 45
B. Derivation of Criterion:
The two degree, three dimensional flutter stability equations used in the development of the criteria are:

𝜔
𝛼 2
{𝐼𝛼 + 𝜋𝜌𝐴𝛼𝛼 − 𝐼𝛼 (1 + 𝑗𝑔𝛼 )( 𝜔 ) }𝛼 + {𝑃𝛼𝛽 + 𝜋𝜌𝐴𝛼𝛽 }𝛽 = 0

𝜔 2
𝛽
{𝑃𝛼𝛽 + 𝜋𝜌𝐴𝛽𝛼 }𝛼 + {𝐼𝛽 + 𝜋𝜌𝐴𝛽𝛽 − 𝐼𝛽 (1 + 𝑗𝑔𝛽 )( 𝜔 ) }𝛽 = 0 (16)

Where:

𝐼𝛼 = mass moment of inertia of the entire empennage about the effective fuselage bending axis
𝐼𝛽 = mass moment of inertia of control surface about its hinge line (both sides of elevator for fuselage
vertical bending flutter and complete rudder for side bending flutter)
𝑃𝛼𝛽 = mass product of inertia about effective bending axis and hinge line = (𝑐 − 𝑎)𝑏𝑠𝛽 + 𝐼𝛽
𝐴𝑖𝑗 = Aerodynamic terms of the form below

𝐴𝛼𝛼 = [𝑀𝛼 − (12 + 𝑎)(𝐿𝛼 + 𝑀ℎ ) + (12 + 𝑎)2 𝐿ℎ ] ∫0ℓ 𝑏4 𝑑𝑥 (17)

Setting the determinant of the coefficients of Eq. (16) equal to zero and making the appropriate substitutions for
the assumptions the following equation is obtained:

𝜋𝜌𝐴𝛼𝛼 − 𝐼𝛼 𝑔𝛼 𝑗 𝑃𝛼𝛽 + 𝜋𝜌𝐴𝛼𝛽

𝐼𝛽 + 𝜋𝜌𝐴𝛽𝛽 | = 0 (18)
| 𝑃𝛼𝛽 + 𝜋𝜌𝐴𝛽𝛼

If 𝜋𝜌𝐴𝑖𝑗 = 𝜋𝜌𝑏4 𝑆𝐴′𝑖𝑗 where 𝐴′𝑖𝑗 is the aerodynamic portion of 𝐴𝑖𝑗 i.e. 𝐴′𝛼𝛼 = [𝑀𝛼 − (12+𝑎)(𝐿𝛼 + 𝑀𝑛 ) +
2 4 4
(2+𝑎) 𝐿𝑛 ] then dividing thru Eq. (18) by 𝜋𝜌𝑏 𝑆 and substituting 𝐼𝛼 𝑔𝛼 ⁄𝜋𝜌𝑏 𝑆 = 5 the following equation is
1

obtained:

𝐴′𝛼𝛼 − 5𝑗 𝑃 + 𝐴′𝛼𝛽
=0 (19)
| 𝑃 + 𝐴′𝛽𝛼 𝐼 + 𝐴′𝛽𝛽 |

Where:
𝑃𝛼𝛽
𝑃 = /𝜋𝜌𝑏4 𝑆

𝐼𝛽
𝐼 = /𝜋𝜌𝑏4 𝑆

𝑆 = Total span of surface (ft)

𝑏 = Semi-chord (ft)
𝑆𝛽 = Total static mass unbalance of control surface about hinge (slug-ft)

Equation (19) when expanded can be expressed in the following form:

𝑃 2 + (𝐴′𝛼𝛽 + 𝐴′𝛽𝛼 )𝑃 + 𝐴′𝛽𝛼 𝐴′𝛼𝛽 − 𝐴′𝛼𝛼 𝐴′𝛽𝛽 + 5𝐴′𝛽𝛽 𝑗 + (5𝑗 − 𝐴′𝛼𝛼 )𝐼 = 0 (20)

For a fixed value of 𝑒 and 1/𝑘 Eq. (20) when expanded results in two real equations, in 𝑃 and 𝐼, one a quadratic
equation in 𝑃 and the other a linear equation in 𝑃 . From the linear equation a value of 𝐼 is obtained as a function of
𝑃 . When this value of 𝐼 is substituted into the quadratic equation of 𝑃 , an equation in 𝑃 is obtained which does not
contain 𝐼. The resulting quadratic in 𝑃 can be solved and from the roots of this equation the associated values of 𝐼
17
Airframe and Equipment Engineering Report No. 45
 𝑉
can be obtained. The ratio of 𝑃 /𝐼 = 1 + 3𝑏(𝑆𝛽 /𝐼𝛽 ) as a function of 𝑏𝜔
can then be used as the flutter prevention
𝑏𝑆𝛽 𝑉
criterion. One curve of 𝐼𝛽
vs. 𝑏𝜔
can be obtained for each value of 𝑒, where 𝑒𝑏 is the distance from the airfoil
midchord to the control surface leading edge. Solutions were obtained for 𝑒 = −0.1, 0.0, 0.1, and 0.2 and it was
𝑏𝑆𝛽 𝑉
found that the variation in allowable 𝐼𝛽
for any 𝑏𝜔 value was small. Figure 3 was then chosen as a reasonable curve
to represent the envelope of curves, thus simplifying the problem by setting up a single curve applicable to
conventional small aircraft.

C. Fuselage Torsion - Control Surface Rotation

An approach to this problem was used which in essence is similar to that for the fuselage bending-control surface
case. The case involving fuselage torsion is analogous to the wing bending-aileron case. If the horizontal and
vertical tail do not deflect elastically then for an angular deflection 𝜃 radians of the fuselage, an airfoil section
located 𝑥 feet from the torsion axis will have a linear (bending) deflection of magnitude 𝑥𝜃 . It should be noted that
in the three dimensional analysis integrals of the form:

𝑀 = ∫ 𝑚(𝑥)[𝑓 (𝑥)]2 𝑑𝑥 𝐴ℎℎ = ∫ 𝑏2 𝐿ℎ [𝑓 (𝑥)]2 𝑑𝑥

𝑆𝛽 = ∫ 𝑆𝛽 (𝑥) 𝑓 (𝑥) 𝑑𝑥 𝐴ℎ𝛽 = ∫ 𝑏3 𝐿𝛽 𝑓 (𝑥) 𝑑𝑥 (21)

appear in the equations. If 𝑓 (𝑥) = 𝑥/𝑆 where 𝑆 is the distance from the torsion axis to the tip of the fin then the
mass and geometric parameters may be considered to be ‘weighted” parameters. In a three dimensional analysis the
integration for the 𝑀 and 𝐴ℎℎ terms must be taken over the complete horizontal and vertical tail surfaces whereas
the other terms involve integration over the rudder span only.

Although data was available for the evaluation of 𝜇𝑟𝛼 2 𝑔𝛼 in the case of fuselage bending vs. control surface
rotation similar data was not available for the evaluation of 𝑀𝑔ℎ 𝑗 (which bears a similar relationship to the fuselage
torsion case). For the analysis then 𝑔ℎ was assumed to be zero and a curve obtained for 𝑆𝑏 𝐾𝐼 versus 𝑏𝜔 𝑉
. Since the
assumption of 𝑔ℎ = 0 is known to be highly conservative the resulting curve obtained from the above analysis was
raised by an amount which experience indicates is reasonable. Table 3 below gives a comparison of the 𝐾/𝐼
determined by the proposed criterion with the allowable 𝐾/𝐼 as given by CAM 04 and ANC 12, as well as the actual
𝐾/ of the rudder on the airplane in service. It should be noted that since 𝜔𝛽/ is less than one, the allowable 𝐾/
𝐼 𝜔ℎ 𝐼
as given in CAM 04 is limited to a maximum value of unity.

Table 3. Comparison of K/I.

f Actual K/I ANC
Airplane VD b New K/I CAM 04
cpm on Airplane 12
(1) All American 10A 183 860 1.083 3.16 3.6 1.0 0.90
(2) Bellanca 14-13 240 510 1.458 0.69 0.708 0.96 0.69
(3) Cessna 190 259 685 1.917 1.53 0.994 0.65 0.61
(4) Howard 18 250 250 1.4 0 0 0.79 0.64
(5) Luscombe 8A 176 870 1.583 4.8 1.225 1.000 0.92
(6) Navion 210 480 1.208 0.655 1.00 1.00 0.81
(7) Rawdon T-1 200 450 1.625 0.886 1.59 1.00 0.84
(8) Thorpe T-11 164 950 1.183 4.06 4.08 1.00 0.96

18
Airframe and Equipment Engineering Report No. 45
 Appendix II - Discussion of Wing and Tab Criteria

In the case of empennage flutter prevention criteria the problem could be treated analytically. This was due to the
simplification of the problem by a number of rational assumptions, which experience indicated to be valid. Thus,
because of the structural elements involved, the problem could be reduced to a two degree of freedom flutter system
with but one elastic restraint. However, in the case of the wing no such simplification is available. An adequate
analytic treatment of the problem requires a minimum three degree of freedom consideration (with three elastic
restraints). It is true that if the ailerons are completely statically and dynamically mass balanced the system can be
reduced to a two degree case. However, since most light aircraft do not have completely mass balanced control
surfaces, the problem must be treated as a three degree of freedom one.
Because of the large number of parameters involved the development of criteria based on an analytic approach is
not feasible. However, experience to date indicates that for a conventional wing, where there are no large mass
concentration located far aft of the elastic axis and for which the ailerons are adequately mass balanced the aileron
reversal phenomenon will probably be the most critical of the aeroelastic phenomena of flutter, divergence and
reversal. Since the critical reversal speed is a function of the geometry and torsional rigidity of the wing the problem
of flutter prevention for a conventional wing can be resolved by providing adequate torsional rigidity to preclude
aileron reversal and by a criterion for balance.

A. Wing Torsional Rigidity Criterion

The 𝐶𝑇𝑅 criterion given in CAM 04 requires that at certain specified distances from the wing tip the torsional
rigidity of the wing exceed a value which is a function only of the design dive speed of the airplane. This criterion
was considered to be adequate to preclude wing bending-torsion flutter as well as divergence and reversal. Since the
reversal speed is a function not only of the torsional rigidity and design dive speed this criterion was reviewed and a
new one developed which is a function of the dive speed, the torsional rigidity of the wing over the aileron portion
of the span, the wing chord and the aileron span. The criterion developed for the torsional rigidity, is in essence
similar to the criterion developed by the Air Material Command in TSFAL 2-4595-1-11 “A Simplified Criterion for
Wing Torsional Stiffness” dated June 1945. The basic difference in forms between the two criteria is that in the
Army criterion the wing rigidity and chord length is chosen at one station only, whereas in the criterion proposed
herein the variation of torsional rigidity and chord length over the aileron span of the wing is used. For conventional
wings both criteria should yield approximately the same results.
This criterion was checked on a number of light aircraft and it was found that in all cases calculated reversal
speed by the proposed method resulted in a slightly more conservative answer than that predicted by the Army
criterion.

B. Aileron Balance
Experience to date indicates that the aileron balance criterion in CAM 04 is conservative. In some cases recently
checked by analytic means, allowable values of 𝐾/𝐼 of approximately five times that permitted by the criteria were
obtained. However, since the wing flutter prevention criteria are based almost completely on empirical methods and
since the success of the torsional rigidity requirement as a flutter prevention method is dependent on a well-balanced
control surface, any major change in existing criteria is believed to be unwarranted. It should be noted that in a
recent check on several light aircraft the allowable value of aileron unbalance was much higher than that given by
any existing balance criteria, However, in every case checked, the wing torsional rigidity was higher than the
minimum permissible rigidity.

C. Tab Criterion
The tab criteria proposed herein are essentially the same as those in AMC 12. A recent study of tab frequency
criteria indicated that the ANC 12 criterion although very conservative was the most satisfactory, consistent criterion
available, However, the use of the second of the two frequency criteria as applied to small, low performance aircraft
has in the past yielded satisfactory results. It is therefore suggested that in any particular case the less conservative
of the two criteria (the one permitting the lowest frequency) be used.

19
Airframe and Equipment Engineering Report No. 45
 NCT

L1RMLAM AM~ EQUIPMENT VICUEEING

REPORT No, 4&5 D

o ITILIFIED FDTTER. PMEVENON CRITERIA D I
N FOR PERSONAL TYPE AIRCRAFT :
C~%I ~JUL231987

Prepared by:
Robert Rosenbaum
Chief, Dynamics Section

Approved by:

A, A. 4rollinecke
Chief, Airframe and
"Equipment Engineerin
..
4 Branch

appeared as Aircraft Airworthiness Reports and Engineering Seution

R•eports*

87 * 4

•WW M -M 'M -
 Sfl!PIFIED FIJJTTEf PFR,"E7,17TION CPITERIA
FOR PERSONAL TYFE AIRCPAFYT -

This report is intended to serve as a guide to the small plane designer

in the presentation of design criteria for the prevention ot such aero-
elastic phenomena as flutter, aileron reversal and wing divergence. It
should also serve as a guide to recommended and acceptable practice for
the design of non-structural, mass balance weights and attachments. The
criteria developed in this report include: wing torsional rifidity;
aileron, elevator and rudder mass balance; reversible tab and balance
weight attachment criteria,.

Introduc ti on

The simplified criteria appearing in CAM 04 were developed at a time when

rational methods of flutter analysis were not available. Because of the
lack of available methods of analysis various attempts were made to set
up empirical formulae which, if complied with, would reasonably assure
freedom from flutter. The sources of material for these stadies were three-
fold:

1. A statistical study of the geometric, inertia and elastic

properties of those airplanes which had experienced flutter'
in flight, and the methods used to eliminate the flutter.

2. Limited wind tunnel tests conducted with semi-rigid models.

These models were solid models of high rigidity so that
effectively the model was non-deformable. The motion of
the models was controlled by attaching springs at the root
and at the control surface to simulate wing bending, torsion
and control surface rotation.

3. Analytic studies based on the two dimensional study of a

representative section of an airfoil.

For the most part these studies indicated that for a conventional airfoil
in which the center of gravity of the airfoil section is not too far back,
that wing flutter could be prevented by designing for a certain degree of
wing torsional rigidity and by control surface dynamic balance, whereas
1% empennage flutter could be prevented by providing a degree of control sui-
face dynamic balance. The limitations were based an the design dive speed
of the airplane and witbIn certain ranges were functions of the ratio of
control surface naturai 'frequency to fixed surface frequency.

Satisfactory rational analytic methods have been available for a number Lý

of years which would permit an engineer to carry through computations to
determine the flutter stability of a specific design. In view of the
fact that flutter is an aeroelastic phenomenon which is caused by a corn-
bination of aerodynamic, inertia and elastic effects, any criteria which

"'777
"4 911Ali4
-does not consider all three effects is bound to have severe limita-
tions. That this is so, is evidenced by the fact that in almost all
cases where rational analyses have been carried thru for specific de-
signs it has been found that the balance requirements specified by
the simple criteria have been too severe. In some special cases the
criteria in CAL: 04 appear to have been unconservative, i.e. flutter
has been encountered in some airplanes which complied with these cri-
teria. In spite of the fact that the old flutter prevention criteria
for the most part yield over-conservative results most small aircraft
companies in the personal plane field prefer to comply with these cri-
teria rather than perform complex flutter analyses. In order to aid
the smnall. manufacturer the CAA in October 1948 issued Airfraime and
Equipment Engineering Report No. 43, entitled, "Outline of An Accept-
able Method of Vibration and Flutter Analysis for a Conventional Air-
plane". The purpose of that report was to present to the inexperienced
flutter analyst an acceptable, three dimensional method of analysis by
presenting in detail a step-by-step tabular technique of analysis. Al-
though a nrinber of aircraft companies are using the methods outlined
in the report, others are of the opinion that this method entails too
much time and expense and are therefore seeking other means of comply-
ing with those regulations which require them to show freedom from
flutter.

Although a rational flutter analysis is to be preferred to the use of

the simplified criteria contained herein (since in most cases a better
design may be achieved by reducin or eliminating the need for non-
structurp.l balance weights), the application of these criteria to con-
ventional aircraft of the personal plane type is believed to be adequate
to insure freedom from flutter.

The criterie contained in the present report have been developed after
an ezhaustive st,.dy of the American and British literature as well as
indepc:'.dent investigations. For the moat part the criteria contained,
in this report are new, however, some have been taken vith little or
no modification from other sources.

It should be noted that the empennage criteria developed in this report,

have been developed on the basis of a single representative (conservative)
value of the empennage mass moment of inertia about the bending axes.
The value was chosen as a result of a study of the mass parameters of
a number of aerplanes of the personal plane type. Therefore, for ]Arger
*03 aircraft than those usually classified as personal planes the cri- .,,
teria may not be applicable. The wing criteria on the other hand should
be applicable to all conventional .03 airplanes which do not have large
mass concentrations on the wings.
;~I £v l ~ I t v Co,""0,

II
reo
Ifd C 'Iel g
 -3-

The criteria developed in this report are of a preliminary nature,

and although considered to represent current thinking on acceptable
and recommended practices regarding flutter prevention measures for
personal type airplanes, these criteria should not be construed as
required procedure to meet thd flutter prevention requirements of
the Civil Air Regulations.

Definitions

"Flutter: Flutter is the unstable self-excited oscillation of an airfoil

and its associated structure, caused by a combination of aerodynamic,
inertia and elastic effects in such manner as to extract energy from
the airstream. The amplitude of oscillation, (at the critical flutter
speed) followring an initial disturbance will be maintained. At a higher
speed these amplitudes will increase.

Divergence: Divergence is the static instability of an airfoil in tor-

sion which occurs when the torsional rigidity of the structure is ex-
ceeded by aerodynamic twisting moments. If the elastic axis of a wing
is aft of the aerodynamic center then the torsional moment about the
elastic axis due to the lift at the aerodynamic center tends to increase
the angle of attack, which further increases the lift and therefore
further increases the torsional moment. For speeds below some critical
speed (the divergence speed), the additional increments of twist and
moment become smaller so that at each speed below the divergent speed
an equilibrium position is finally attained (i.e. the process of moment
increasing angle and thereby increasing moment etc. is convergent); above
this critical speed the process is non-convergent.

Coprcl Surface Reversal: This is the reversal in direction of the net

normal force induced by the deflected control surface, due to aerodynamic
moments twisting the elastic "fixed" surface. This phenomenon can best
be illustrated by considering the case of aileron reversal. Normally the
lift over the wing with down aileron is increased by the aileron deflec-
tion, while the lift over the wing with up aileron is decreased by the
aileron deflection, thus a rolling moment results from an aileron deflec-
tion. However, since the center of pressure for the lift due to the de-
flected aileron is usually aft of the elastic axis, deflecting the aileron
downward tends to reduce the wing angle of attack thus reducing the irnre-
ment of lift. For the wing with up aileron the torsional moment due to
up aileron tends to increase the wing angle of attack. It can thus be
seen that the rolling moment for an elastic wing is less than for a rigid
wing. Since the wing torsional rigidity is constant while the twisting
moment due to aileron deflection increases with the square of the velocity
it is obvious that at some critical speed the rolling moment due to aileron
deflection will be zero. Above this speed the rolling moment will be
opposite to that normally.expected at speeds below this critical speed.
The critical speed so defined is the aileron reversal speed.

23117

3--
~ 3ý
~ ~ ~ ~ ~ ~ ~33 V A
3 w
"33 3 - r. .

r r~.'
 •umaryof Criteria

Wing Torsional Stiffness

The wing torsional flexibil.ty factor F defined below should be equal

to or less than 2U0U

Where: F - LC:iL2ds

~i =-Wing twist at station i, per unit torsional moment

applied at a wing station outboard of the end of the
aileron. (radians/ft - Ib)

=Wing chord length at station i5, (ft)

=Ci

ds = Increment of span (ft)

Vd = Design dive speed (IAS) of the airplane

Integration to extend over the aileron span only. The value of the
above integral can be obtained either by dividing the wing into a
finite number of spanwise increments AS over the aileron span and
summing the values of 9iCi 2 &5 or by plotting the variation ot eiCi . 2
over the aileron span and determining the area under the resulting
curve*
In order to determine the wing flexibility factor F,# a pure torsional
couple should be applied near the wing tip (outboard of the end of the
aileron span) and the resulting angular deflection at selected inter-
vals along the span measured. The test can best be performed by
applying simultaneously equal and opposite torques on each side of the
airplane and measuring the torsional deflection with respect to the
airplane centerline. The twist in radians per unit torsional moment
in ft-lbs should then be determined, If the aileron portion of the
wing is divided into four spanwise elements and the deflection deter-
mined at the midpoint of each element the flexibility factor F can be
determined by completing a table similar to Table I below. Figure 1
illustrates a typical setup for the determination of the parameters C
andes

23177

WM-

-
~~~~~~~~~~~' w" oe U;.W* -- %~~
 Aileron Span

I I I
IIc 8CCNIl S
5

Fig. 1

TABLE I
S_ C1).(2) ( 4),() _•)(6)

STATION AS C C2 e OC2S

ft ft 2 ft-b
_ft

F = column (6)
Aileron Dalance Criterion

The dynamic balance co-efficient K4 should not be greater than the value
obtained from figure 2 wherein Ki is referred to the wing fundamental
bending node line and the aileron hinge line. If no knowledge e2d sts
of the location of the bending node line the axis parallel to the fuse-
lage center line at the juncture of the wing and fuselage can be used*
J

.3177

• U-3, 1f"
• . ... 3i 3 . W ."3 "3 : .•?3U . ~ 3 -~
 Wherein: K = product of inertia

-mass
aI moment of inertia of aileron about its

hinge line

Free Play of Ailerons

The total free play at the aileron edge of each aileron, when the other
aileron is cla:nped to the wing should not exceed 2.5 percent of the
aileron chcrd aft of the hinge line at the station where the free play
is measured.

Elevator Balance

Each elevator should be dynamically balanced to preclude the parallel

axds flutter (fuselage vertical bending-symmetriL elevator rotation)
as well as perpendicular ayds flutter (fuselage torsion - antisymmetric
elevator rotation). If, however, the antisymmetric elevator frequency
is greater than 1.5 times the fuselage torsional frequency the perpen-
dicular axis criterion need not apply.

Parallel Axis Criterion

The balance parameter Y as obtained from Figure 3 should not be ex-

ceeded. In Figure 3 the balance parameter r and the flutter speed
parameter Vf are defined as:

Vf Vd

'Where: S Elevator Static Palance about hinge line (ft - lbs)

I = Elevator mass moment of inertia about the hinge line

2 )
(lb - ft

b A Semichord of the horizontal tail measured at the mid-

span station (f t)

Vd = Design dive speed of the airplane (mpb)

th = Yaselage vertical bending frequency (cpm)

Perpendicular Axis Criterion

For each elevator tte balance parameter X as obtained from Fi.gre 4

should not beo exceeded. In ire 4h tbe balance parameter X and the
flutter speed parameter Vf are defined as:

.23177
•

l
U-+ 'W ... 14 -I US•N' . U I..

IS
 - 7-

S=bX

Vf Vd

¥Where: S = Semispan of horizontal tv.il (ft)

b a Semichord of horizontal tail at midepan station (ft)

K - Elevator product of inertia referred to stabilizer

center line and elevator hinge line (lb - ft?)

I -Elevator mass moment of inertia abodt the elevator

2
hinge (lb - ft )

f• Fuselage torsional frequency (cpm)

Rhdder Bala-rc e

The value of K as obtained from Figure 3 and the value A as obtained

from Figure 4 should not be exceeded; where in Figures 3 and 4#, •
•I•, ) bK and:

S = Distance from fuselage torsion axis to tip of fin (ft)

b = Semichord of vertical tail measured at the seventy

percent span position (ft)

K U Product of inertia of rudder referred to the fuselage

2
torsion axis and the rudder hinge line (lb - ft )

Vc= Fuselage torsional frequency (cpm)

fh a Fuselage side bending frequency (cpm)

= Rudder static balance about hinge line (lb - ft)

I a Mass moment of inertia of the rudder about hinge line

(lb - ft 2 )

Tab Criteria

All reversible tabs should be 100% statically mass balanced about the
tab hinge line. Tabs are considered to be irreversible and need not
be mass balanced if they meet the following criteria:

I. For any position of the control surface and tab no

appreciable deflection of the tab can be produced
by means of a moment applied directly to the tab,
when the control surface is held in a fixed position
and the pilots tab controls are restrained.
*t3A7 p
' " ...1 .. 3 " g i '.* i ..
S• ,••.•.'•. • . e •% f% 'dr-,•~~ • .
 S•, ,,,8-

2. The total free play at the tab trailing edge should be

less than 2.5% of the tab chord aft of the hinge line,
at the station where the play is measured.

3. The tab natural frequency should be equal to or exceed

the valup given by the lower of the following two cri-
teria

(a) -163 Vd St cp.

or

(b) ft= 2000 opm for airplanes having a design dive speed of
less than 200 mph. For airplanes with a design dive
speed greater than 200 mph the frequency in cpm should
exceed the value given by 10 times the design dive
speed in miles per hour.

Thus for an airplane with a design dive speed less than 200 mph if (a)
above gave a value in excess of 2000 cpm it would only 'be necessary to
show a frequency of 2000 cpm for the frequency criterion.

Where: ft lowest natural frequency of the tab as installed

in the airplane (cpm) - either tab rotation about
the hinge line or tab torsion whichever is lower.

C1 - chord of moveable control surface aft of the hinge

line, at the tab midspan position (ft)

St a Span of tab (ft)

c- Span of moveable control surface to which tab is

attached (both sides of elevator, each aileron
and rudder) (ft)

Particular care should be taken in the detaiil design to minimize the

possibility of fatigue failures which might allow the tab to become
free and flutter violently.

Balance Weight Attachment Criteria

Balance weights should be distributed along the span of the control sur-
face so that the static aniAlance of each spanwise element is appraxi-
mrately uniform. Howev'er, where a single external concentrated balance
weight is attached to a control surface of high torsional rigidity the
natural frequency of the balance weight attachment should be at least

* 23177

"OP U UUV 4. ' S1 rA1A

V '-W a 0 , -OFU W~ ~ I
N
 5O percent above the highest frequency of the fixed surface with Which
the control surface may couple in a flutter mode. For example the
aileron balance weignt frequency should be at least 50% above the wing
fundamental torsional frequency. The balance weight supporting struc-
ture should be designed for a limit load of 24g normal to the plane of
the surface and 12g in the other mutually perpendicular directions.

It should be noted that the dynamic balance coefficient W/ can be re-

d•eed by (1) reducing K, (2) increasing I or (3) reducing K and in-
creasing I. Since an increase in I results in a reduced control sur-
face natural frequency with possible adverse flutter effects, the primary
purpose of ballast weights used to reduce K4, should be to decrease the
product of inertia K and not to increase the mass moment of inertia I.

.4,

i.J

2277
21..

4, 1O W. A
 T77---,.

I ._.._.._..._._.._.._ a___7_

jj1 t
77~ 177-l'
~--------- -~--r------ - IW

CC 4-- c.:: C
w 77- =7 --7
~
I jCI I ,

* -.-- ____________ -
7--.-.

u m 77=

o~~ 1_ W~~

Auj
/_____

114 / .~W

____ __ _ __ _-.--
~4'l

L L . 7
~za
4Ei-E ~
i /I__ - -- --
_

7-

77h

0 0 0

111'1N3I31.3-300 30NV-IVG OIV'JVNkA

29
3 S U 1
to*:-:4 '4 P-~
 tIt

- - . ~2~2Lj
L4 c CDCL0

- Ii

S.-I

-ILL

'00

N;ILL

a W

LL w

Ix 4
~c1

- 0 00
9~~~~~'~~
.5 @C S S-
I V-SL
OF,

4
* -T 4 -w--r -- r v. 5

*55 ,..,-.5 5---.. S-.. .H*5

* . -. . . . -F-.5,. LL. .~. . - .-.

.? ~ ~ ~ ~ -- ~ ' ~ ~ J. ~ -*5
S-, "l *S4I5 Sj
~~u .. .. -.- I . - .. ."L .~
 -4

. . . . 04

-L. ... 3Et-.

5--. w +,

4.K-~-tt

ILL
2k~ C%
-.---.. c-o - . --

--- 5-- - --.

L- I,,I

- .¾
***j

0.

4- -~ - I
44j
1-4

L.-- -ILL

4 1--i--4-
J* It -- 4-

~IOR~ It
4Z,

---------- 7-
 Dynemic and Static Palanee of Moveable
Control Surfaces

Definitions

Static Balance: Complete static balance of a moveable control surface

is obtained when the center of gravity of the control surface lies on
the hinge line i.e. the resultant moment of the mass of the surface
about the hinge line is Fero. If the center of gravity of a crface
I ies aft of' trhe hinge surface it i. called staticaly unbetLaced, where-
as if the center of gravity lies forzard of the hinge line the surface is
c~aled statically ovr-balanced.

Pamc Balance: A moveable surface ini dynamically balanced with respect

-o a given axis if an angular acceleration about that axis does not tend
to cause the surface to rotate about its own hinge line. The dynamic
balance coefficient K/I is a measure of the dynamics balance condition
of the moveable control surface, wiherein K is ths product of inertia of
the surface (including balance weights) about the hinge and oscillation
axes and I is the mass moment of inertia of the control surface (including
balance weights) about the hinge axis. Physically the dynamic balance
coefficient KI/I may be interpreted to represent:

Exciting Torque
Resisting Torque

Mass Balance Computations

Assume tIw X axis coincident with the oscillation axis and the I axis
coincidenT with the control surface hinge line. After the reference axes
have been determined the surface should be divided into relatively small
parts and the weight of each part VI and the distance from its c.g. to
each axis tabulated. See Figure 5 and Table II. Referring to Figure 5
the static moment of the element &W is &Wx, the moment of inertia
2
about the hinge line is hJx and the product of inertia is aWxy * The
static unbalance of the total surface $ is then r.AWx; the moment of
inertia of the surface is ZWx2 and the product of inertia is K -- /x=

N U

N, 1
 014 SAME 31015I OF )k^11

^SO COP OF MNVtEVZCQ

LOAD OStrP.AmuTtoh

NOTE' +x"AXISI
TAVON QCARWAQD

AxtS OF OSCILLATr low Fig. 5

TABLE ii

~i~tflecuitloi ~Dist. fTom -M miwzDist, trovil I KtrrY

lbs. inches tneb-lbs. inch-W3~. IbAns.1 In. -

() (2) (3) (4) () (6) ) () ()

etc.

Product of Inertia with respect to Other Axecs

Having deternIned the product of inertia with respect to one oscillation

axis it may be deial rnecessary to determine the product of inertia
with respect to some other oscillation ruis, If the product of inertia
was originally calculated for an oscillation axi~s which was perpendicular
to the binge axis then the product of inertia with respect to inclined
 amow 0.0O and Y-I can be determi~ned from the perpendiciular waxe product
o0± inertia (.X.4 and Y...Y) by u,.,e of Lie follcming equation:
Koy mKxy sin~ - I7 coso

0 - 0

F1 g., 6

whee i is the angle between O..4 and Y-I in the quadrant where the
center of gravity of the surface is locatede
If the product of inertia was originally calculated for one set of axes
andl it is desired to detezmine the product of inertia for another set
p ~of axes parallel to the original set., ther the new prodact oXLinertia
ý2can be determined from the equation:

2Ki4X07 W i
a2 7!tW x0yW

There: W ntotal weight in poundsi of the moveable surf ace

11 2 product of incr~.a with. respect to axes

xo distance between X1 and 12 axes

23171.

~ ~ w ~ ~' ~w ~ v S *W"~ ~.i-~ w ~ WIN

 yo = distance between Y1 and Y2 axes

T x distance from C.G. of surface to 71 axis

7 a distance from C.G. of surface to 11axis

rt
COG. of
Ssuface
i xo

To

X2 Z2

•2 Fig*

Erpe~rimntal Determination of Static Unbalance, Moment of

Inertia and Product of Inertia

(a) Static Unbalance

Te moveable control surface should be carefully supported at its

hinge line on knife edges or in a jig with a minimum of friction.
The force necessary to balance the control surface, when applied
to a given point, is then measured by an accu-Late weighing scale.
The net force times the. distance between the hinge line and the
point of application of the force ie equal to the rtatic unbalance

SJ#

23177

-U W -- W 1 SP -W, W
6-W
 Weighing Scale

Distance
of appiction T lFig. 8
balancing force and hinge line

(b) Moment of Inertia

The experimental determination of the mass moment of inertia consists

of supporting the surface or tab at the binge line with a minirmum of
friction ir. a jig ia an attitude similar to that described above and
maintaining it in this attitude by means of one or two springsL, as
show~n in Figure IX. One spring is sufficient for control surfaces
with large static unbalanraes., while two are generally used for surfaces
which are fairly well statically balanced. The natural frequency of
the surface (for small1 oscillation~s) under the restraining action of
the springs is i~hen measured by means of a stop watch by determining
the time necessary for a given numnber of cycles, In order to reduce
experimental errors to a minimum, the Waae for a large nu~mber of cycles
(about 30) in measured4
The spr4ing stiffnemses are dynamically determined by placing a weight
Won 3pring: 1 which will deflect it an amount approximately equ.al to
* the average spring deflection during the moment of inertia test andi
then~ determinring the natural frequency of the spring with W.1 attached
by deter-mining with a atop watch the time necessary for a given number
of cycles; a similar test is condu~cted for the determination of the
spring stiffniess of spring 2s, using a weight W2 .

23117

~~~~ ISV~U ~Z
 Noig
I

The moment of inertia can then be calculated by substituting the test

results in either equation 1 or 29 depending on whether the control
surface center of gravity is above or below the binge axis,

rf control surface center of gravity is-below hinge axis:

d2=Wf3 W2 f 2 2 ) X9.788 Wocx B I

If control surface center of gravity is above hinge axis:

d2 ( 1f ~f 2 - 9o788 WoX (2)

Where: I a Moment of inertia of surface about hinge axis (pound-

inches 2 )

Wa x Weight of surface (pounds); W 1. W2 Spring calibration

weights (pounds)
M D1istance of surface C.0. above or below hinge axis
(inches)
d Z Distance from hinge axis to springs (inches)
fo a Frequency of surface when restrained by springs (c.p~s.)

f* Calibration frequency of spring K1 under weight W1 (c.p.S.)

f2MCalibration frequency of spring X2 under weight W2 (copes*)

23171

U U'U S 'U U V 'U~iU' U-

34v~~\~"- A
 *-\-.19-.

(C) Product of Inertia

The product of inertia Kqxof a moveable control surface can be calculated

from three experimentally determined moments of inertia. If the control
surface moments of inertia are obtained by oscillating about each of the
axes X-Y-, Y-Y and then about a third axis 0-0 lying in the XY plane and
making an angle o4 with the X-%X axis, then the product of inertia Kn, is
obtained from:

Cos aC. + rsý 1Mdýr

0
Fig. 10

Since this method of determining the product of inertia involves small

differences between large quantities a small experimental error in the
determination of the moments of inertia may result in large errors in
the product of inertia. It can be shown (ACIC No. 711 "The Determination
of the Product of Inertia of Aircraft Control Surfaces")p that the error
can be reduced to acceptable levels by the proper choice of the angle 0( *
The proper value of OL can be determined after having determined I,,and
T •j this value is given by the relationship:

23•77

W~~~A *s w AIL-
 - .

"-20-

Appendix I - Discussion of' Emnna Flutter Criteria

Studies made by the Air Material Coamand the Civil Aeronautics Admin-
istration and many independent investigators have shown that for the
most part empennage flutter modes can be closely associated with con-
trol surface unbalance and the appropriate fuselage natural frequency
with which the control surface will couple. Thus, in the case of
elevator coupling, for the most part, the fuselage vertical bending
mode enters into the motion of the system whereas for the rudder either
fuselage side bending or torsion will couple. Althouguh it is fully
realized that any analysis based on this type of simplification would
of necessity be only approxinate, it should be noted that the results
obtained are usually highly conservative, since other modes
generally enter into the motion of the complete system which
tend to damp I
the motion with a resultant higher flutter speed. Thius, the fuselage
vertical bending mode is generally damped by coupling with wing sym-
metric bending and stabilizer bending whereas fuselage side bending
motion is usually damped by coupling with fuselage torsion the anti-
ymnetric bending of the stabilizer and bending of the fin.

Based on these considerations the Air Material Command prepared a re-

port Army Air Forces Technical Report No. 5107 entitled "Charts for
Fuselage Bending vs Control Surface Flutter". It has been found that
these charts are applicable to larger aircraft than those considered
in the personal plane field. Each chart in AAFTR 5107 shows the
variation of -44 with Vo for various values of &A Unfortu-'
nately the limits bf the values of the parameter A.°used, are
such that for most airplanes in the personal plane field these curves
cannot be read with any degree of accuracy, without doubtful extrapo-'
lation. Furthermore, it was considered that for simplicity a single
curve would be more suitable in treating the relatively low perform-
ance personal plane field, than a family of curves.

Fuselage Bending - Control Surface Rotation,

The following assumptions were made in the determination of the fuse-

lage bending control surface rotation flutter criterion (Figure 3):

(1) Wz
-
(2) W0-

(3) C-es 0
"(14) 0-v -3
(5) gr, " for elevator rotation vs fuselage
vertical bending

for rudder rotation v 9 fuselage side

bending
*The notation used in this section is similar to that appearing in
CAA Airframe and Equipment Engineering Report No. 43
3u. , . - t- .. -. .;..3.,7...
The above assumptiorns are _,elie'ied to te rational and valid for miost
aircraft in tVle field under considerationi. Justification for each of
the above assumptionis is givei. below.

(1) The flubter frequeiicy 4')

s equal to the fusela.ge
bendin .tf. tquency 4~ . xporience has shown -,hat
'because of the relatively large inertia of the tail
t>k, ahrodynariic and inertia coupling, terms are ccm-
paratively small, The flutter frequency is there-
fore very close to the ft7,elage bending frequency.

(2) For conventional aircraft winth nio springs in the con-'

trol system the natural frequency of the empennage
ccntrol surfaces is zero* For 'the mosit part con-
ventional tail. control systems are so rio-ed that
ela~stic deformation in the control system takes place
only if the controls are locked in the cockpit. Since
under actual flirzht condiLicns the pilot restraint in
the cockpit is sn~all., the assumption of Wk, = 0 is con-
sidered to be val-id,
(3) In the low performance field it has b~een found that
most control surfaces are not aerodynamically balanced,
Since in general an increase in aerodynamic balance will
tend to increase the critical flutter speed this assump-
tion will. yield conservative results for ai.rcraft Ydth
*aerodynarically balanced rurfaces and yield correct re-
sults for those aircraft with no aerodynamic balance.
(.)The flutter mode involving fuselage bending and control
surface rotation is anmlogoiis to the wing torsion-aileron
rotation case 'with the effective fuselage bendirC axis
corresponding to the wing, elastic axis,, This aoxis of
rotation is the effective point about which the airfoil.
section (stabilizer-31evator or fin-rudder) rotates when
the fuselage bends and is not the nodal line of the fuse.-
lage in bending. A study wiade by the Air M~ater'ial Command
from vibration measurements of a large number of airplanes
indicates that the effective fuselage bending amxis is
located approximatel~y 1.!5 tail s~Aface chord lengths ahead
of the tail sr..rface add-chord (ie. a =-3.0)

(5)Atn exaninmtion of the values of the parameter Arex for

the enrperxnage of a number of small airplanes of the .,03
type indi~cates that this parameter is small varying approx-
imately between 4~ and 8 at lcow altitudes (based on g = .03)e
For the case of ftuselage side bending it has been found that
- the effective increase in mass moment of inertia of t~he

W S._FVU N''
- Q ~
2%fuselage due to wingg yawing ia approximnately 75%
of the eznpennge mass moment of inertia. By as-
Guminlg constant, one curve of allcwable
mass b~alance pArameter versus flutter speed pa-
rameter can be calculated for each valiue of Q
thus simlif ying the problem. The values of
ACK= 5 for fuselage vertical beniding and
a'3 = 6.75 for fuselage side bending are be-
lieved to be representative, conservative velues
for .03 &irplanes.
Derivation of Criterion-.

The two degree,, three dimensional flutter stability equations xused

in the development of the criteria are:

Wheres mass moment of inertia of the entire empennage about the

effective fuselage bend~ing axis
M maiss moment of inertia of control surface about its binge
lina (both sides of elevator for fuselage vertical bending
flutter and complete rudder for side bending flutter)
P=mass product of inertia abouit effective bending axis and
hinge line n(- Lb5+1

A,.nAerodynamic terms of the form

Ak M) (L Let
Setting the determi~nant of the coefficients of equation (1) equal to
zero "rx making the appropriate substitutions for the assumptions the
following equation is ebtained:
(2)

I1)f 1IA+ IIA4

+
+
 -23-.

I 1rp Ar1 . 4ShA

6rb w e is the aerodynamic portion of
k,~4.e.'
- NV. CMc-•("i(.L-N) + (+a.)'L•then dividing thru
equjtion (2) by rfs'.5 and substituting T /•,ý $ = 5 the
following equation is obtained:

S0 (3)

+/ P + AkT+A
Whe re z P u P/f hS

S Total span of surface (ft)

b - Semi-chord (f t)
S Total static mass unbalance of control surface
F about hinge (Slug-ft)
Equation (3) when expanded can be expressed in the following forn:

P+(A + gP+A + S"A)3r

(4i)
For a fixed value of e and 14&equation (4) when expanded results in
two real equations, in P and I, one a quadratic equation in P and the
other a linear equation in P. From the linear equation a value of I
.is obtained as a function of P. •hen this value of I is substituted
into the quadratic equation of P, an equation in P is obtained which
does not contain 1. The resulting quadratic in P can be solved and
from the roots of this equation the associated values of I can be ob-
tained. The ratio of PA I 3b as a function of Vito can
then be used as the flutter prevention criterion. One curve of
vs V/w can be obtained for each value of e, where eb is the distance
from the airfoil nidchord to the control surface leading edge. Solu-
tions were obtained for e *a,0,.,
-. and .2 and it was found that the
variation in allowable b4/,for any V/bow value was small. Figure 3
was then chosen as a reasonable curve to represent the envelope of
curves, thus simplifying the problem by setting up a single curve ap-
plicable to conventional small aircraft.

,('1

•+'I • +' +I~... 7I • . ..+1 I I" ""11 I :I ++-+l:;ZJ.17 .. . ..U ... q•+
 FUselage Torsion - Control Surface Rotation

An approach to this problem was used which in essence is similar to

t•at for the fuselage bending-control surface case. The case in-
volving fuselage torsion is analogous to the wing bending-aileron
case. If thie horizontal and vertical tai] do not deflect elastically
then for an angular deflection 9 radians of the fuselage, an airfoil
section located X feet from the torsion axis will have a linear
(bending) deflection of magnitude XG e It should be noted that in
the three dimensional analysis integrals of the form.

w W'dc Ahh bLLf~ov UW34

Appear in the equations. If 3 where S is the distance from the

vTiO
torsion axis to the tip of the fin then the mass and geometric pa-
rameters may be considered to be 'weighted" parameters. In a three
dimensional analysis the integration for the M and Amterms must be
taken over the complete horizontal and vertical tail surfaces whereas
the other terms involve integration over the rudder span only.

Although data was available for the evaluation of A(. t in the case
of fuselage bending vs control surface rotation similar-data was not
available for the evaluation of M,&(which bears a similar relation-
ship to the fuselage torsion case).,. or the analysis then ?hwas as-
sumed to be Zero and a curve obtained for 1-1versus '/bW , Since the
assumption of 0%:ais known to be highly conservative the resultirz
curve obtained from the above analysis was raised by an amount which
experience indicates is reasonable. Table III below gives a compar-
ison of the Y/• determined by the proposed criterion with the allow-
able /I as given by CAM Oh and ANG 12, as well as the actual VI of
the rudder on the airplane in service. It should be noted that since
""_is less than one, the allowable K/I as given in CAL' Oh is limited
to a maxi mim value of unity,

23177

- - - - - - - - -
 N -25-

. IActual
4/I on
SNew
Air- CAPU. ANC
Airplane f KVI plane 0h 12
VD cpm b

(1) All American 1OA 183 £60 1.083 3.16 3.6 1.0 .90
(2) Bellanca 14-13 2J40 510 1.458 .69 .708 .96 .69

(3) Cessna 190 259 685 1.917 1.53 .994 .65 .61

(4) Howard 18 250 250 1.40 G 0 .79 .64

(5) Luscombe 8A 176 870 1.583 4.8 1.225 1.000 .92

(6) Navion 210 480 1,208 ,655 1.00 1.00 81

(7) Rawdon T-1 200 450 1.625 .886 1.59 1.00 .84

(8) Thorpe T-11 164 950 1.183 4.06 4.08 1.OO .96

Appendix II fiscussion
D- of Wirin sand Tab Criteria

In the case of empennage flutter prevention criteria the problem could

be treated analytically. This was due to the simplification of the
problem by a nwmber of rational assumptions, which experience indicated
to be valid. Thus, because of the structural elements involved, the
problem could be reduced to a two degree of freedom flutter system with
but one elastic restraint. However, in the case of the wing no such
simplification is available. An adequate analytic treatment of the
problem requires a minimum three degree of freedom consideration (with
three elastic restraints), It is true that if the ailerons are completely
statically and dynamically mass balanced the system can be reduced to a
tro degree case. However, since most light aircraft do not have completely
mass balanced control surfaces, the problem =st be treated as a three
degree of freedom one.

Because of the large number of parameters involved the development of cri-

teria based on an analytic approach is rot feasible. However, experience
to date indicates that for a conventional wing, where there are no large
mass concentration located far aft of the elastic axis and for which the
ailerons are adequately mass balanced the a- leron reversal phenomenon will
probably be the most critical of the aeroelastic phenomena of flutter,
divergence and reversal. Since the critical reversal speed is a function
of the geometry and torsional rigidity of the wing the problem of flutter
preventi.on for a conventional wing can be resolved by providing adequate
torsional rigidi~.y to preclude aileron reversal apd by a criterion for
• (
\aileron balance.

23177

w---~ q a V 9 3 3 3 - 3 3 - U
ý%q R
- ~ ~'A
 * .-26-.

Wing Torsional -igiLty Criterion M

wThe C Tot criterion -iven in CLY 04 requires that at certain specified

distances from Lw wing tip the torsional rigidity of t win -he
exceed
a value whIich is a function o.-ly of the &dsigrdive speed of the air-
plane. This criterion was considered to be adequate to preclude vying
bending-torsion flutter as well as divergence and reversal. Since the
reversal speed is a function not only of the torsional rigidity and
design dive speed th-is criterion was reviewed~and a new one developed
which is a function of the dive speed, the torsional ri'gidity of te 771.
wing over the aileron porticn of Lhe span, the wing chord and the
aileron span. The criterion developed for the torsional rigidity, is
in essence si.mlar to the criterion develcped by the Acr Materi~al Corn-
mand in 'TSFAL 2-4595-1-11 "A Simplified Criterion for WinC Torsional
Stiffness" dated June 1945. The basic difference in forms betveen
the two criteria is that in the Army criterion the wing rigidity and
chcrd 2ingth is chosen at one station only, whereas in the criterion
proposed herein the variation of torsional rigidity and chord lenth
over the dileron span of the wing is used. For conventional wings
both criteria should yield approximately the same results.

This criterion was checked on a r=mber of light aircraft and it was

"found that in all cases calculated reversal speed by the proposed i
method resulted in a slightly more conservative answer than that
predicted by the Army criterion,.

Aileron Balance

Experience to date indicates that the aileron balance criterion in

CAM Oh is conser-iative. In some cases recently checked by analytic t","U
means, allowable values of K/1 of approximately five times that per-
mitted by the criteria were obtained. However, since the wing flutter
*. prevention criteria are based almost completely on empirical methods
"ardsince the success of the torsional rigidity requirement as a
flutter prevention method is dependent on a well balanced control sur-
face, any major change in existing criteri-a is believed to be unwar-
ranted. It should be noted that in a recent check on several light
aircraft the aillwable value of aileron unbalance was much higher
than that given by any existirg balance criteria, However, in every
case checked, the wing torsional rigidity was higher than the minimum
permissable rigidity.

Tab Criterion
The tab criteria proposed herein are essentially the same as those
in AMC 12. A recent study of tab frequency criteria indicated that AL
"the ANC 12 criterion although very conservative was the most satis-
"factory, consistent criterion available, However, the use of the
second of the two frequency criteria as applied to small, low per-
formance aircraft has in the past yielded satisfactory results. It
4 is therefore ruggested that in any particular cdse the less conserv-
ative of the two criteria (the one permitting the lowest freqaency)
be used.

4P nW

l2%.

" : • -. ',.- ,,% ,/•• • 1t % -• ,, . • ',

% - -. : *%
q '%. *' ,