C.P. No. 310 RCY,c,; ! mFi)@qyq.._ C.P. No.

310
(18.199)
A.R.C. TechnicalReport

MINISTRY OF SUPPLY

AERONAUTICAL RESEARCH COUNCIL

CURRENT PAPERS

Vibration and Flutter

Flight Testing
BY

M. 0. W. Wolfe

LONDON: HER MAJESTY’S STATIONERY OFFICE

1956

THREE SHILLINGS NET

 C.P. No. 310

U.D.C. No. 534.134.001.4 + 533.6.013.422.0m.4

Technical Note No. Structures 170

ROYALALRCRAFT
ZZXABLI~T
Vibmtion and ETutter l?l..ight Testing

by

1,i. 0. W. Wolfe

A mvkrr is made of the techniqws that have been developed or are

being developed in the United Kilngdonfor flight vibration testing and
flight flutter testing aircraft. This includes a description of the
instnznentation used for recotiing the Vibratioml response, and a
comparative assessment of the vaious methods used for exciting the
aircraft in i'lipht flutter tests.
 LIST OP COKrExTS

1 Introduction 3
2 Instrumentation for Vibration measurementin flight 3
7.1 The general measurementproblem 3
2.2 Instrumentation 4
2.3 Amplifiers
2.4 Recorders z
3 Plight Vibration testing 5
4 Typical pmgrsxme for flight vibration tests 6
4-1 Limitations of control jerking
4.2 Special problems of forced vibration ;
5 Plight flutter testing 7
Goneral techniques 7
::: Xethods of excitation 7
5-2.1 General
5.2.2 Single inertia exciter ;
5.2,3 Phased inertia excitors 9
5.2.4 Eleotrod~namic excitation of control cirouits IO
5.2.5 Stick jerking IO
5.2.6 Rocket excitation 10
5.3 General remrks on teohniqw 11
55.54 Associated theoretical u;ork II
conolusions I.2

References 12

LIST OF ILTJJS~RATI~E
Firm
Analysis of phase Difference b&won Translational and
Rotational Gcq0nents of Ging ;;otion 1
Double Integrating Multi-Channel Ar;rplifiers and Reoorders 2
Performance Curves for Type IT.I-6 &@.fiers 3
Waveforms obtained from Stick Jerk Test on a Jet Pighterdimraft 4
Single Inertia Exciter Installed in the Nose of a keteor Aimraft 5
Electrcdynamic Bxsitor Installed in a Pure Servo Tab Lancaster
Aircraft for the Excitation of the Rudder Tab Control Circuit 6
Thrust-Time Curve for Rocket Impulse Unit 7
Gomparison of Amplitude Response curves Z&asured in a Plight
Flutter Test 1~5thPredicted Values 8

-2-
 1 Introduction
The expansion of our knorrledge of the flutter of aircraft has not
kept pace with the rapid changes in aircraft design and the considerable
increases in speed that have followed the introduction of the gas turbine
engine. In particul,arthe extent of our knowledge of the aerodynamic
forces involved in flutter is at present inadequate for the satisfactory
prediction of the flutter clz?racteristics of an aircraft at high air
speeds from data provided by pre-flight eqerimental and theoretical
investigation. In view of our l.M.ted rosourses for research on flutter
and the continuous advances in aircraft pctiormance this situation is not
likely to change materially for some considerable time. Consequently,
despite experiments and calculations carried out at all stages of a pmto-
type development prior to flight, the possibility of flutter being encoun-
tered, particularly flutter involving the control surfaces, cannot be luled
out. 'i'he increases in airwaft speed have not only enhanoeathe likelihood
of flutter occurr:ng but have also lea to an increase in the level of aero-
dynamic excitation due to bul'1etinC ati :Iach number effeots. koreover, a
condition of low damping due to a near approach to a critical f'luttor
coditioa can considerably augment the response of the stzuoture to aem-
ilymd.0 forces.

It is therefore realised in Se Exited I.ingdom that elaborate and

accurate te&xiques of' Plight Vibration Testing;', and in certain oinxun-
states FE&t Flutter Testin?* arc of very great importance as means of
obtaining the essential data on vibration and flutter chractetistics
necessary for the clearance of prototype aircraft. At the present time,
in view of the practical difficulty of installing the necessary equipment
and conducting comprehensive Flight Flutter Tests on all prototypes, ow
policy is to ask for a oomprehensiw FE@ Flutter Testing ptv&ramme
only in those cases where mwginal stabilities have been predicted by
pre-flight investigations. In all other cases we are proposing, as a
mintium requxrement, to make the installation o? multi-channel vibration
measurementequ$ment and a prograze of flight vibration tests mandatory
for one prvt0ty-p of all new Service aircraft.
2 &strumentation for Vibration Measurement in Plight

2.1 The General Sensurement Problem

The essential informaLon required for the diagnosis of a vibratory
condition on an aircraft is the frequency and mode of the oscillation
involved. To resolve an aircraft mode from measurementsmade in flight it
is neccssarr to know the linear translation amplitude ratios between a
certain m&mm numbor of points on the structure and their phase relation-
ships for the froquonoics concerned, and corresponding data for the angular
displaoemehts between certain parallel planes in the structure including
th control surfaces and tabs. The modes of the main structure and lifting
surfaces can usually be resolved adequately by using linear translation
transducers deployed at appropriate points on the structure. For example,
in the cmse of a wing the pich~~s are deployed cEaord\tise in spanwise pairs,
each pair comprising one trnnsduccr on the main and one on the Ear spar.

* By Flight Vibratiou Testing is zxnt the me~~suro,nent of the vibration

frequencies of an aircraft and their ~ssociatod mo&s in flight >+iithout
deliberate artrfiioial excitation of the aircraft
** By Plight Flutter Testing is meant a technique of testing involving
artificial oxcitation of ths aircraft in flight 7‘5th the object of
obtaining the dsmping in its modes of vibration.

-3-
 This arrangement is usually found to be satisfactory for resolving
the rotational and vertical translational components of motion. An
important analytical difficulty arises, however, with this method of
mcnsuromsnt when a phase difference exists between the translational and
rotational co~onents of motion.
The difficulty may be illustrated by reference to the diagram in
Fig.l(a) which shoo?rs a chord section of a vring which is assumedto have out
of phase motion3 of rotation an3 vertical translation. These motions are
referred to a spanwise axis passing through 0, whioh for simplicity is
assumedto be equidistant from the points A and B at rhich transduoers
measuring vertical translation are placed. Yiteofcmingto the vector
diam of Fig.l(b), the smplitude of the trsnslational component of xing
motion is assumedto be 0 and the amplitude of rotational motion 8, and
the phase difference betwoen them a.. The rotational motion produces
additional translational motions c ta 0, ~hci~ E is the distaru3e of 0
from A and B. These components arc shonn as x and -x on the vector diagram.
The translational anplitudcs a and b neasurcd at A and B are therefore
the vector sums of C and x, and C and -x respectively, and have somephase
difference p depending on the values of A, 0 and a.
%here there is no noticeable phase difference, the usual method of
determining the components of motion from tho readings of linear translation
transducers disposed in this manner is to tske half the sum of a and b
as being the translational component, snd half their difference as being
the rotational component. Howver if the rotational component is small
compared to the translational component, it zi.11 be clear from a consideration
of the vector diagrw shor+nin Fig.l(c) that a very small phase difference
bet:=enthe measurements at a end b does not necessarily in@y a
correspondingly small phase &ifferenco betTeen the components of the mot%on.
In fact, on a conventional rsllti-channel record, particularly if it contains
random transient frequencies due to aerodynamic disturbsnoes, a large phase
difference between the components of nation my not only be dif;"icult to
resolve but may not actually be notioed.

TO ovemome this difficulty a non-seismic r&ation sensitive trens-

ducer has been developed in this country lxrti.cularl;r for control surface
and tab measurementswhere the detection and measurementof phase differences
in the motion are most important from the flutter aspect.
2.2 Instrumentation
A considerable amount of rvDrk has been done in this country in recent
years on the instrumentation problen.. There are three main aspects of this
problem namely, the development of suitable transducers, the development of
suitable amplifiers and finally the analysis problem, The latter is not
strictly an instromontation problem oniy, but it arises because of the
increasing difficulty of analysing conventional multi-chsnnol trace
records, and the desirability of rc?cording the information in a manner
which mould permit the data to be dealt r&h directly from the records by
computational mchines.
Several different types of transducers are in current use for
vibration measurement. The one which has undergone the greatest development
and which is found to be the most useful for flight vibration and flutter
measurementsis the Lsn Elec Indw&o.co ficcolerometerl. This is si.mi.1s.r
in prinoiple to the correspording American Liller instrument but is in
many respects a better transducer. In particular the ;&ller instnvnent
has been found to be unsuitable for high altitude applications because of
leakage of &urging fluid. The Lan Elec accele~mcter incorporates a
dif?erential pressure compensating device and does not suffer from this
defect.
-4-
 The main advsnt~es of the ~&~ts.nce accelerometer for this 'work are
that the instrument itself 5s quite small and can be used in the rjnge f%~rn
0 to 100 c.p.s. without integration, and from 2 to 100 c.p.s. vrith into-
gration. It can also be used satisfactorily in manoeuvres involving applied
g, cordiitions for which certain other types of instrument are unsuitable.

2.3 Am&Siers
Tm types of m-lifier are tending to becomestandard for flight
vibration testing, one a str-iightfor;mrd A.C. amplifier T!ith single
integration for use ~6th generator type transducers, and the other a
cdcr tme of snv$ifier rtith double integration ,"or use with inducts.n~e
accclermeter type transducers. In both cases or*: of the major problems
of devologment has been miniaturisation because of the -.nstallation problem
on small Tighter type eimraft, Examples of the lastest type of equipment
are shown in the photograph, Fig.2. 30th types of C*lificr are inter-
changeable. The integration characteristics for the carrier type amplifier
are sham in Fig.3*.

2.4 Recorders
The multi-channel mirror galvanometer type of recorder has been found
to be the most suitable for flight vibration testing. Here also the
question of miniaturisation is important, an3 miniature multi-channel
recorders are being developed in this country. The analysis problem is
intimately associated Tvith that of recorder design. The conventional
multi-channel trace record is, in some cases, no longer a s;ttisfaotory
methed of presentation for nnalysis prposes beoause of the complexity of
the vibrations encountered on high speed aircraft. For this reason
multi-channel play-back recorders of the photographic and magnetic tape
type are being developed to enable subsequent anLyses to be made by means
of electronic wave analysers.

A twelve chanrrel photogt%phic type, operating on the variable area

tract; principle has been in use at tho R.A.E. for some time in cotiunotion
with a Huirhead Pamedra'Uave 1nalyser. It vw employed, for example,
on the Comet I accident investigation for the reoording & analysis of
the complex high frequency panel vibration excited by jet ef'flux effects.

The main ndvanws of this technique aw that it enables the

amplitudes of component frequencies in a complex :iavc form to be determined
accurately, and for a parricular frequency, for emle a flutter frequenoy
involving motion of r&n surfaces aUl controls, it makes possible the
measurementof phase different, * hetxeen the motions of different parts of
the structure where the motions are complex.

3 Flight Vibration Testing

L!ost of the flight vibration testing carried out in this country is
done on prototype aircraft during the flight development phase. During
the early stages the object is to detect and measuz vibrations wilenthey
arise vrith a view to establishing the cause and providing the data to
effect a Cure in severe. oases.. At a later stage tests are necessary to
show that the aircraft meets specification requirements in respect of
permissible vibration levels.

-5-
 Aircraft vibration may be classified into trvo broad categories:-
(1) forced vibration, e.g. mec'hanical vibration ad buZfeting in
which the main agents arc the excitation and the resonance
properties of the structure,
(2) vibration characterised by a poorly dampedmode, resulting in
high vibrational response under moderate excitation.
The vibrations under group (<) are those associated with a definite
sourGe o? exoitation and o? which the significant featuws are the magritudes
of the exciting forces and their frequency relationships to the natural
frequencies of the aircraft. The primary object of flight vibration testing
in these cases is to identify the source of the excitation ad the nature
of the aircraft's resaonse so t&t curative measures involving either
rduction or elimination of the source, or detuning to avoid resonance
may be undertaken.
The vibrations under group (2) :-ii11 in general be amreoiably
influenced by air speed because the darr?pingin a mode is usually critically
dependent or the air speed. In the extreme, a mode may becomenegatively
damped, aen flutter occurs as a self excited oscillation. The object of
the tests in relation to the possibility o.? these vibrations is to obtain
some indication of the possible existence of any poorly dacrpedmodes,
together dth some quantitative measurc Or the dsxping and its v,xiation
tith air speed. Porthis purpose a technique of Flight Flutter Testing
involving deliberate excitation of the airwaft is required.
In general it is not possible to Foresee xi.th certainty ab initio to
rrhat extent each of the foregoing effects is going to contribute to the
vibrations o? a particular prototype sircraft, aad to cater for all
eventudities ww.ld involve equipping all prototype aircrdt with exaiting
equipment for exhaustive fl;&t flutter testing, in addition to the measure-
merit equipment. It is felt therefore that the more practical approach is to
do flight vibration tests over the whole perfonx03ce range on on2 prototype
of each new type, attempting to assess the damping in significmt modes 9
simple methods where these are ap$icable, for e-l+? by control jerl&ng
and by flights in bumpy conditions, in all oases where margind stabilities
have not been predxted by theory. In other cases e:haustive flight flutter
tests are required.
4 Typical Flight PKJQXINX
--_--- for Flight Vibration
..L.-_------ Tests
At moderate speeds and Each numbers tests are madeby increasing the
aircraft speed by reasonably small increments (about 20 knots I.A.S.) end
recording at eaoh speed. In suitable cases, an6 where necessary, each oook-
pit control is jerked in turn, records being taken at each jerk. Ideally
the records should be examined after each test before proceeding to the next
higher air speed; however at low speeds,and particularly where no instabilities
arc suspected,this procedure would be unnecessarily restrictive and in
practice recordings of several air speeds are made in one fli&t. If no
severe vibration is encountered this procedure is continued throughout the
performance range of the aircrart; at high speeds, however, the records a??3
analysed after each 20 knot increment before proceeding to the next? and in
the transonio region where rap% ohmges in the aerodynanic oocfficxents
are likely to be enoowtercd, tests are made at small increments of Mach
number, for example increments 0,* 0.02i'i in th range from 0.8 to 0.9 and
incrunents 09 0.0-W in the range from 0.3 to I.1 analysing the recor6.s after
each increment.

-L-
 There is always a possibility that unforeseen severe vibrations may
arise during a flight covering several airspeed increments without
nocessarFly being observed by the pilot. On large aircraft monitoring
devices are pruvided for the flight observers to guard aga%nst this
occu1x1cnc0,and on single seater aircraft it is recommendedpractice to
provide a monitoring device for the pilot to provide him writh in-formation
on the motions of the control surfaces.
With mgard to bu?:'cting, the effects of increasing CL at a given
Nach number may be Important. This aspect is investigated by straight-
forwam measurementsat increments of I+, at a partioular Mach number.

4.1 Limitations of Control Jerking

Control jerking as a moansof applying d sudden force to an aircraft
to exckte transient responses with the object of determining the damping
in the modes excited has certain limitations. sin many oases lack of high
frequency response of the control system will not permit the application
of a sharp jerk to the control surface. This may impost?a lovi ujjper 1id.t
on the frequency of the modes that can be excited. The technique is
therefore used with some discretion because of tine possibility of signi-
ficant modesbeing overlooked. Its chief virtue, in cases where it can
usefully be employed, is its siri:Licity. Typical exa@es of stick jerk
records taken on a high speed airoraft from wiiich values of damping have
been deduced are shown in Fig.&.
4.2 Special Emblems
--- of Forced
-- Vibration
Certain particular problems 0,* looalised severe forced vibratFons
sometiws arise which require special treatment, axsmples are buffeting
caused by open bombbays, panel vibrations exoited by jet efflux effects
and vibrations otiginating from pmpeller disturbances. In the first
example it is usually necessary to augment the structure vibration
measurementsby measurementsof the fluctuating pressures in, and in the
vicinity of, the bombbny,in order to Gisgnose the mechanismof the
disturbance. In the second erwnpio, very severe panel vibrations leading
to fatigue failures have been known to occur ca~od by jet eff'lux effects.
These are characterised by high frequencies di "white" exoitation spectra.
?or this pmblem accelemtion transducers oL p the Wezo cloctric type and
very swsitivc fluctuating pressure tronsducors capable of coverring a
52 quency range A-am 0 to IO Ko/seo am l-qrirod. In the thin3 oxsmple,
where severe propeller or?ier aircraft v3ration occurs the pmblem is
usually one of determining the ;leohanism of the energy transfer; whether,
for exsm$e, the forces am being transmitted mechanically through the
engine mountin@, or :~hother they are being transmitted aerodYIXU.Ii.CallY
in the form of pressure waves emmating from the propeller blades. HCIX?
again it may be neoessnrY to em$oy fluotuatmg pressure transducers.

5 Flight Flutter Testing

5.1 General Techniques

A considerable amount of work has been done on the development of
flight flutter testing techniques in this country in the last four years.

The techniques employed may be classified broadly into the

following categories, the "continuously forced oscillation" technique and
the "decaying; oscillation" technique. The former is characterised by a
continuous sinusoidal cxcitction of the aircraft structure or oontro~s
by means of a mechanical or some other form of exciter. The exoi.tatio,l

-7-
frequency is slowly increased fron ZCIU over a pre-?iotcrnined range and the
vibrstfon response of the aircrdt is reootied by means of multi-channel
vibrat3on measuring instruments. This prcccss is repe&-d at suitable
increasing inorements of afr sped or 1&h nwnber, and the an&itude speed
responses in the significast mo&s at-e deduced from the r?eesurements; an
approach to a critical flutter oon%ition being idicated by an increase
in wplitude response with air speed.
In the "decaying oscillation" technique either a sudden force is
applied to the aizcra..ft, or a sinusoidal ftirce at a fixed frequency is
stienly mruoved, and the subsequent transient responses of the aircraft
stature are measured. The overall wings in the nodes concerned ere
then deduced from the time rates of decay of the ensuing transient
oscillations; the process being repeated. at increasing increments of air
speed. An approach to a critical flutter condition is indicated by the
[email protected] appmachingzero. Stick jerking is a crude form of this technique.
The decaying oscillation technique is the more infomzttive of the
two, because it not only yields information on the approach to flutter but
also gives the values of the relative dampings in the modes concerned over
the Whole of the air speed range.

Quantitative information on damping cannot easily be derived from

the amplit~e/frequency curves in the continuously forced oscillation
technique. Apart from consideration of flutter prediction a knowledge of
the damping in flight modes of vibr;rt&x is becom& j.mpor&nt for other
aspects of airfxaft vibration, for example in the buffeting problem nrd
in the fatigun proolem dere the influence of &m&g on the response of
the aircraft- to Lusts may be simficant.
In both techniques the analysis of the fli,Tht records becomes
diffiCult when buffeting occurs because the artificially induced oscil-
lations cannot easily be clistin&.shed fro.2 those arising from the
buffsting. Since buffeting occurs on most high speed aircraft at the upper
end of the Mach number and speed ranges the analysis problem is 3 very
relevant one.
It has been found that in buffeting conditionqthe degree of
experimental scatter of amditude values at the same fli&t conditions for
several repeat amplitude n&r speed curves using the continuous oscillation
technique h3s been rather less than the corresponding scatter of &m&g
values for the transient technique. This difference is believed to be
almost entirely due to the andytical difficulties referred to.
In comparing the relative merits 02 these techniques, an 5mportant
aspect of the continuous oscillation teclxxi.que should be mentioned. It is
a d.ifficulty which arises when the wplitude/air speed curves are derived
from a study of the recordings fromtransducers placed at different
positions on an aircraft. Monndly the transducer positions chosen are at
extremeties such as tail plane and wing tips, and in the control surfaces.
It has been found that in certain oases the amplitude/air speed curves
for different transducer positions do not dways exhibit the same trwds
with air speed, this being due presucably, to changes in the flight mode
of vibration, for example, movementof the zones of mininur;l aplittie.
This difficulty can give rise to sozlt?clegree of diguity in flutter testing,
and makes it necessary to recod the +x.-en& at severd transdwer positions
to minimise the danger of overlooldng an importiant o&se.
This particular difficulty does not arise in the &ecayi.ng oscillation
technique, because the rates of decay for a given mode must be the same at

-8-
different points in the structure for a particular flight condition.
Analytical di~ffioulties may arise, however, with this technique when tw
or more trazx3ients are ewited decaying at different rates.

It has been found that the best technique ADpears to be a combination

of the two methods. A variable frequency exciter is used to locate the
amplitude peaks in the frequency range, and a decay record is then obtained
by outting off the excitation as rapidly as possible when tuned to the peak.
Both the amplitude response and the damping can then be watched as the air-
speed is increased. Even if buffeting is encountered, making the damping
estimates less reliable, some guidance can still be obtained from the ampli-
tude responses at different parts of the structure and in the Control
surfaces.
5.2 -Plethods of Exoitation
5.2.1 caeml
The tv,u main types of ewitation are: firstly those in which a
sinusoidal f'oroe of variable frequency is applied to the aircraft structure
or controls, and secondly those in &&ch an impulse, or series of iiZPul.ses
are applied. In this country the last type is only used when instdlbtion
difficulties and time preclude the use of the former type.

5.2.2 Single Inertia Extoiter

In this technique a single inertia exciter of the rotating out-of-
balance weight type is attached to the structure, usually either at the
nose or 531the mar fuselage of the almraft. A typical such installation
is shovJnin the nose of a Meteor aircraft in Fig.5. Where necessary,
provision is made for lateral as Trell as vertical excitation by providing
two eziters operating in the appropr%ate planes and incorporating
clutohing arrangenents so that either may be used singly. On single
seater aircraft it is usual to employ M automatic frequency sweeping
device for the exciter, and the pilot is provided. with a visual indication
of amplitude and an overriding control of the frequency to p+?rmit him t0
tune to each amplitude peak if neoossary On larger aircraft a manual
frequency control is nozmKLly used.

To enable the &c&ng oscillation technique to be employed, and to

provide the pilot with a rapid means of stopping the excitation, as a
safety precaution, two methods of rapid exciter cut-off have been developed.
Onemethod employs the principle of regenerative braking, where the
exciter is driven by an electric motor, and the other involves the use of
a specially designed inertia exciter in which the out-of-balance effecrf; may
be cancelled with great rapidity by mechanical means. In a particular
design the cancellation ten be effected in half of a cycle.
In partioular cases symmetric and antisymmetric bending modes of
wings and tail units have been excited satisfactorily with this technique.
Asymmetr5.omodes have been excited with exciters placed in the v&ng tip or
tail-plane tip.

5.2.3 Phased Inertia ExciteI‘

Certain flight modes of vibration may be difficult, or i&e&.
impossible to excite w:+h an exciter placed at a single point in the
stmoture, for exmplo a wing torsional mode. Consequently, a method of
driving two or more identical inertia exciters by means of phased D.C.

-Y-
electric motors suitable for use on onliasy aircraft D.C. electrical
suppplics has been developed, and at present marks satisfactorily in the
laboratory. A flight version is at present being installed on a Xeteor
aircratt.
5.2.4 Electmdynomic Excitation of Control Circuits
A method of applying a sinusoidal force to a control circuit by
means of a moving coil electrical vibrator has been developed and used
in flight. This method has shown promising results, althou& it is
probably only applicable to sircrxft having either pure servo t&b, or
spring tab controls. The advantages of this system are that both the
force amplitude and the frequency can be controlled independently in
flight, and the force may be removed instantaneously. A further
advantage is that the vibrator.may be used as a regenemtive brake
thereby increasing the damping in the control circuit. The vibrator
unit as installed in the rudder tab circuit on a Lancaster aircraft
fitted with pure servo tab controls is shown in the phwtograph, Fig.6.
5.2.5 Stick Jerking
This method of exciting the aircraft control surfaces is still used
on sane aircraft rrith manual or po;-rer assisted controls, but only when
better methods are not pxnctiuble. It is nonnall~ only successful in
exciting the fundamental frequencies of the aircraft structure. It is of
little value on aircraft fitted tith Mly powered controls. An exemple of
records of transients obtained from stick jerks on the elevator controls of
a fighter aircraft at increasing air speeds is shown in Fig.4, a fall in
damping tith air speed is clearly indicated. in this case.
5.2.6 Rocket Excitation
The shortcomings of the stick jerk technique combined with the
diffiCulty of installing equipment -F‘or0ontLnuous ewltation on mall very
high speed aircraft, particularly supersonic aircraft with thin wing
sections, has led to the develqx~nt of a rocket excitation technique. In
this technique small rockets are attached to appropriate p&s of the
siromft structure on?. used to provide 3 thrust of an tirrpulsivc nature to
excite transient osoilintions of thfi aircraft. One type of rocket unit,
known as a Va&eril is oyifndticd in shape, of I$ inches diameter and 45/8
inches overall leng;h, its thrust is 200 lb and duration 50 milliseconds.
A thrust time curve for this unit is shxn in Fig.?. The duration can be
varied by modifying the ex@osive charge. Standard times forthis unit
3re 75, 50 3.~3 25 milliseconds. A smaller unit is at present under
development. This unit is 5 inches diameter, snd of lengths 6, 4 and 2$
inches depending on the corresponding durationtimss of 50, 25 and jZ.5
milliseconds. Its mean thrust is 200 lb, end for ease of installation in
thin section wings, tail Rlanes or fins, the thrust axis maybe arranged
to be normal to the longitudinal axis of the unit.
The units are fired electrically and one great advantage of this
technip is that several units deployed at different parts of the aircraft
structure can be fired either simultaneously, or in a pre-arranged time
sequence appropriate to tie frequency and the modebeing investigated.
In this way some degree of selectivity may be achieved end unwanted modes
suppressed.
This technique has been used with success for flutter investigations
of pmtotypc a5.mraft.

- 10 -
 On one aircraft a flight flutter test x~3 made on a mode involving
tail plane antismxstric yawing motion at 8 c ;LS. Rockets were fitted at
each tail plane tip, firing fore ard aff respectively. In this particular
case the tail plane ms mounted approximately half way up the fin and scn?e
degree of fin torsion ins involved in the motion. Satisfactory de.&ne/air
speed curves were obtained up to the speed at which compressibility
buffeting amplitudes became large in relation to the mplitudes emited
by t‘ne rockets.

In oases swh as the foregoing the excitation can be maze more

effective by firing a second pair of rockets mounted in the ssme positions
as the first pair but arrangd to fire ~.nthe Opposite dirWtio%3 NE"ter a
time interval equal to half a period of the appropriate oscillation. For
this purpose the duration of the rockets must not be &re,ater than half a
period of the oscillation.

5.3 General Remarks on Technique

One of the difficulties of ez$loying the mechanical inertia type of

exciter is that these units are very inefficient at low frequencies.
Wthods involving the oscillation of a control surfaooe tab overcome this
difficulty since they exploit conpnrativeljj large aerodyn?mi.c forces
arising f'rw small applied hinge moments; however tnis method is limited
in scope. An extension of the technique is to enroloy an auxiliary control
surface installed solely for tne purpose of excitation. This has never
been attempted in t&s country, but a unit has been designed an.3 is being
manufactured with a view to examining tne possibilities of the technique.
Clecrly, such a surface must be irrever;;ible, rind the backlash in the
system should be as srxll as possible to avoid any possibility of flutter
occurring involving the auxiliary control surface itself. 1'1 a particular
case this aspect iwould have to be thoroughly explored.

Another method of pmduoing co,pamtively large sinusoidal forces

at low frequencies Ta&i& is at present being investigated at the R.A.E.
is the use of a rotating rocket nozzle. This involves the use of liquid
pmpellants which ere burnt m a combustion chamber, the products of
combustion being expanded through a nozzle which is rotated by means of
n speed controlled electric motor. In this way the full thrust of the
rocket msy be applied. to produce sinusoidal forces dove to zero frequency.

5.4 Associated Theoretical "ior&

It is the policy in this country to carry out comprehensive

theoretical cdculations, and in SWe asses windtunnelmo&elexpe&nents,
to deten%ne the flutter ohzxcteristics of an aircraft before embarking
on flight flutter tests. In particular, advantage is taken of the
rapidity with ~uhich amplitude/air speed curves can be deduced for modes
imrolving several degrees of free?.om on electronic analogue computers.

It is of considerable value in a fli&t test to have some fore-

knowledge of the probable shape of t he amplitude/air speed, or damping/air
speed curves particularly when con5ition.s of low damping are being
approached. In fact the theoretical calculations and the flight tests
e3?enow regarded as [email protected] aspects of the Lnvestigation as a whole
with "feed back" of information from both sides. In particular, the flight
flutter tests my be expected to pmvide some check on the valiiiity of
the aerodyrwniic dcriva!,~ves used in the celoulations end the numbor of
important degrees of f~cdom involved in a particular flight mode.

- II -
 The extent to which good correlation may be achieved between the
results of flight tests and calculations is exemplified in the CWS
shown in Fig.& These curves sho~tho cd.culated amplitude/air speed.
GWXCSfor an aircraft on which a series-of flight flutter tests w.s carried
out for a nmde of vibration involving primarily elevator rotation and tail-
plane bending, cornoaredwith the curve obtained from the tests. In this
case the continuous oscillation technique was en@oyed, ad single inertia
exciters wers installed in the nose of the aircraft for vertical exoitation,
and in the fin for lateral excitation.
5.5 Conclusions
It is considered that, in cases where the equipment can be installed
in the aircraft, the continuous exdtation method employed to obtain values
of dan&ng end amplitude response with air sped fbr the &or amplitude
peaks is the best technique of those at present aveilable. Phased.inertia
exoitera provide a further refinement to this teohnique. In other oases
the rocket excitation technique provides the best alternatS.ve.

-No. Author Title, etc.

I H.K.R. Neubert. A Variable Inductance Acceleration Transducer
R.A.C. Tech. Note No. INSTN.lj5
2 E.T. de la Perelle An Integrating Amplifier for Instrumentation
R.A.E. Teoh. Note No. INsTN.14-4.

- 12 -

4't.Z078.C.P.310.K3 - Prrnted m Smat Srrta;n.

 (b> I a c b

I
(cl
FIG. I(a-c) ANALYSIS OF PHASE DIFFERENCE
BETWEEN TRANSLATIONAL AND ROTATIONAL
COMPONENTS OF WING MOTION.
 FREQUENCY 6 p. S)

FIG. 3. PERFORMANCE CURVES FOR TYPE IT. I -6

AMPLIFIERS.
FlG.4. FROM STICK JERK TEST ON A JET FIGHTER AIRCRAFT

FIG.5. SINGLE I ERTIA EXCITER I STALLED IN THE NOSE QF A METEOR AIRCRAFT

2oc

15C

100

50

I .
4 0 15 0 6 o-07 0 '8 0

TIME (SECONDS)

FIG.7. THRUST - TIME CURVE FOR ROCKET IMPULSE UNIT.

 NOTE:- THE SAFETY MARGINS REFER TO AIRSPEEDS FOR CRITICAL
FLUTTER CONDITIONS AND ARE RELATED TO DIFFERENT
AMOUNTS OF ELEVATOR MASS BALANCE.

PREDICTED FOR 23%

SAFETY MARGIN. T-1
/
/
1’
/I

/AS MEASURED

PREDkTED FOR 43%

SAFETY MARGIN

AIRSPEED -

FIG. 8. COMPARISON OF AMPLITUDE RESPONSE

CURVES MEASURED IN A FLIGHT FLUTTER TEST
WITH PREDICTED VALUES.
 C.P. No. 3 IO
(18.199)
A.R.C.Technlcal Report

Crown copyrrght rewved

Pubbshed by
HER MAJPSTY’S STATIONERY Orrrcr
To be purchased from
York House, Kmgsway, London w c 2
.+zx Oxford Street, London w I
P 0 Box 569, London S.BI
‘3.~ Castle Street, Edmburgh 2
‘09 St Mary Street, Card~fi
39 Kmg Street, Manchester 2
Tower Lane, Bristol r
z Edmund Street, Blmungham 3
80 ChIchester Street, Belfast
or through any bookseller

S.O.Code No. 23-9010-10

C.P. No. 310