N A S A TECHNICAL NOTE

-
h
G;

w 4e -v
LOAN COPY: RETURN T6--
AFWL (DOUL)
KIRTLAND AFB,N.M.
m
-
--
-
=
w n
m
g

=E

, --

' I MEASUREMENTS A N D ANALYSIS OF

~I SOLID-PROPELLANT-ROCKET VIBRATIONS

1
OBTAINED DURING A CAPTIVE FLIGHT

11
. by James A . Schoenster

1
Langley Research Center
1 1 Hdmpton, Va. 23365

$ N A T I O N A L AERONAUTICS A N D SPACE A D M I N I S T R A T I O N WASHINGTON, D. C. OCTOBER 1971

:4:&
%
 TECH LIBRARY KAFB, NM

1I 11I I1111
1l1IlII$I lHlIlI1
OL334LB
l
1. Report No.
NASA TN D-6517
4. Title and Subtitle
I 2. Government Accession No. 3. Recipient's Catalog No.

5. Report Date
October 1971
MEASUREMENTS AND ANALYSIS O F SOLID-PROPELLANT-
6. Performing Organization Code
ROCKET VIBRATIONS OBTAINED DURING A CAPTIVE FLIGH:
7. Author(s) 8. Performing Organization Report No.
J a m e s A. Schoenster L-7837
10. Work Unit No.
9. Performing Organization Name and Address 114-08- 13-03
NASA Langley R e s e a r c h Center 11. Contract o r Grant No.
Hampton, Va. 23365
13. Type o f Report and Period Covered
2. Sponsoring Agency Name and Address Technical Note
National Aeronautics and Space Administration 14. Sponsoring Agency Code
Washington, D.C. 20546
~~

5. Supplementary Notes

6. Abstract

As p a r t of an overall p r o g r a m t o investigate the vibrational environment of s m a l l solid-

propellant rockets, a captive test f i r i n g of a Nike rocket with a boilerplate payload was con
ducted i n a unique test stand which, while restraining the total vehicle displacement, allowed
substantial freedom f o r the test vehicle to vibrate.
Measurements w e r e obtained of the vibratory accelerations, p r e s s u r e s i n the rocket
case, rocket-case s t r a i n s , and total vehicle displacement and data are presented. A dis
cussion of the analytical procedures used t o interpret random phenomena is presented.

7. Key Words (Suggested by Author(s)) 18. Distribution Statement

Solid-propellant-rocket vibrations Unclassified - Unlimited

Random vibration analysis

9. Security Classif. (of this report) 20. Security Classif. (of this page) 21. NO. o f Pages 22. Price*
Unclassified Unclassified 39 $3.00
- .
 MEASUREMENTS AND ANALYSIS O F SOLID-PROPELLANT-ROCKET
VIBRATIONS OBTAINED DURING A CAPTIVE FLIGHT

By J a m e s A. Schoenster
Langley Research Center

SUMMARY

A s part of an overall program to investigate the vibrational environment of small

solid-propellant rockets, a captive t e s t firing of a Nike rocket was conducted in a unique
test stand. The stand was designed to allow the rocket to vibrate freely while the overall
displacement of the vehicle due to static thrust w a s restrained. Measurements were made
to determine the vibrational behavior of the vehicle, which included a boilerplate payload,
and the effects of the t e s t stand on the vehicle. These measurements included vibratory
accelerations along the length of the vehicle, pressures in the rocket case, rocket-case
strains, and total vehicle displacement. Data could be obtained only during a limited time
period because of the short burn time of the rocket and, therefore, a discussion of the ana
lytical procedures used by the working engineer to interpret these random phenomena is
presented.
The test results demonstrated the ability of the test stand to meet all dynamic
requirements; however, problems were encountered in measuring the oscillatory pres
s u r e s generated during burning of the rocket due to the low level of these oscillations.
The power-spectral-density vibration levels measured to 2000 H z were considerably
lower than those usually specified as flight environmental levels for solid-propellant
rockets. High responses were experienced at frequencies above 10 000 H z .

INTRODUCTION

A major problem i n insuring the success of a space mission is to determine the

dynamic loads which may affect the payload-launch-vehicle system and then to specify
tests which will demonstrate the ability of the system to perform satisfactorily under
those loads. Undertesting may not show potential defects which could compromise the
success of the mission, whereas overtesting could lead to structural overdesign. The
problems associated with overdesign, which may cost both time and money, a r e several;
among these a r e weight restrictions on the payload which limit the objectives of the mis
sion, the need for a larger launch vehicle to perform the mission, and additional design
effort to eliminate dynamic-response problems associated with individual components
f o r each payload.
In spite of the problems of overdesign, vibration environment specifications on
solid-propellant rockets usually a r e very conservative as a result of the lack of flight
measurements (ref. 1). F. R. Wagner, in his studies on solid-propellant-rocket load
definition, discussed the vibrational environment and concluded that the problems asso
ciated with overdesign are frequently encountered (in 1969 at the Solid Rocket Structural
Integrity Information Center, University of Utah). Therefore, because of the uncertain
t i e s involved in defining test specifications, a strong need exists f o r adequate analytical
and experimental methods f o r predicting flight vibration frequencies , amplitudes, and
damping values.
In a Langley Research Center program measurements a r e being obtained of the
flight vibrations on both a payload and its solid-propellant launch vehicle to aid in devel
oping techniques f o r predicting vibrational environment. These flight measurements will
be used for evaluating data obtained from a captive firing of the same payload-launch
vehicle combination in a special test stand to determine whether vibration measurements
from a captive firing can be used to establish test specifications for flight hardware. A
preliminary step in the program prior to either the captive firing or flight of the actual
payload-launch-vehicle system w a s the firing of a combination of the boilerplate payload
and the launch vehicle in the test stand with the following three objectives:
1. Obtain vibration data from the launch-vehicle-boilerplate-payload combination
2. Verify the capability of the instrumentation to obtain the necessary data
3. Verify the ability of the restraint system to allow the vehicle to vibrate freely
while withstanding the rocket firing loads
It is the purpose of this report to present some results from this preliminary ground
firing. The primary measurements obtained during this firing were restraining force on
the rocket, vibratory accelerations at various locations on the vehicle, pressures in the
solid-propellant rocket case, and acoustic levels in the test cell. Because much of the
vibration data obtained w a s random in nature, the data evaluation methods are discussed
in detail.

SYMBOLS

B bandwidth, Hz

dB decibel, root-mean-square p r e s s u r e referenced to 2 X lom5 N/m2

f frequency, Hz

2
 function of time

acceleration of gravity, 9.807 m/sec2

statistical degrees of freedom

m maximum number of time-lag values

sample s i z e

probability-density function

autocorrelation function

record length, seconds

amount of time x(t) falls inside the range from x to x + A x, seconds

time, seconds

amplitude, g units, N/m2, dB, pm,, cm, o r N

change in frequency

At change in time

Ax change in amplitude

E sampling interval, seconds

mean value, g units or N/m2

root-mean-square value, g units or N/m2

time-lag value, seconds

power spectral density

chi- square probability distribution function

. .
Subscripts :

av average

meas measured

Pk peak

rms root mean square

APPARATUS AND TEST CONDITIONS

Description of T e s t Vehicle
The vehicle used in this firing consisted of a Nike type M88, solid-propellant rocket
and a boilerplate payload. The nominal rocket thrust w a s 213 000 N with a burning time
of 3.4 sec. In o r d e r to be representative of the flight condition, the boilerplate payload
w a s designed to bring the total weight of the vehicle to 705.2 kg with a center of gravity
2.05 m from the aft end of the Nike nozzle. The boilerplate payload was a steel cylinder
1.36 m long, 0.381 m in diameter, with a 0.3175-cm-thick wall. It had a 3.175-cm-thick
plate at the forward end and six 2.54- by 0.635-cm longitudinal stiffeners equally spaced
around its circumference. The boilerplate payload weighed approximately 83.4 kg.
Shown in figure 1 is the test vehicle mounted in the test stand. Figures 2 to 6 show var
ious details of the test vehicle.

Description of Test Stand

The test stand, shown in figure 1, is a built-up box-beam structure designed to
support a launch vehicle with cables to three torsion b a r s as shown schematically in fig
u r e 2. The launch vehicle is shown in the nozzle-up position used for firing in this test
stand. Three short cables, which a r e attached at the aft end of the Nike rocket case and
which run over pulleys, connected the vehicle to the torsion b a r s . The torsion b a r s
a r e spaced at 120' intervals around the thrust axis and have a combined stiffness of
10.33 kN/cm. The suspension-system natural frequency was experimentally verified
by attaching a 705.2-kg load to the torsion bars. The resulting 6.1-Hz frequency agreed
with a calculated value based on the assumption of a single-degree-of-freedom system.
The damping of the suspension system under this load w a s experimentally determined to
be 4 percent of an equivalent critically damped viscous system.
In addition to providing a restraining force which reacts to the thrust of the rocket,
the test stand provides a preload which may be used to r e s t r i c t the initial displacement
of the rocket, as shown schematically in figure 2. Lateral restraint of the torsion bar is

4
provided by three bearings, two at the top and one at the bottom. A quarter sector at the
top of the torsion b a r is connected by a short cable over a pulley to a fitting at the aft end
of the rocket. A rod, connected to the same fitting as the cable, passes through a hole in
a test-stand brace beam and a block of honeycomb. A washer and two jam nuts on this
rod hold the block of honeycomb against the brace beam. A preload turnbuckle at the
base of the torsion b a r , when adjusted, twists the torsion b a r and thereby causes the rod
to pull the honeycomb against the brace beam. The force in the cables caused by the tor
sion b a r s is measured by a force link located at the cable attachment to the rocket. When
the rocket is fired, the vehicle does not unload the honeycomb until the preload condition
is exceeded, at which time the rod is f r e e to slide through the hole in the brace beam and
release the load on the honeycomb. A spring-loaded spacer, shown i n figure 2, is origi
nally inserted between the honeycomb and the brace beam. After ignition, and after the
honeycomb has been unloaded, the spacer is extracted by the springs, which allows the
vehicle to oscillate past the initial r e s t point during the vibration s e t up by the initial
impact of the rocket firing. This permits the use of higher preloads which further limit
the initial acceleration and total displacement of the vehicle. At the end of the thrusting
period, the vehicle moves upward under the influence of the torsion-bar force and the
energy in the system is absorbed by the honeycomb. A photograph of the honeycomb-
spacer installation is shown in figure 3.

Description of Instrumentation
Six types of instruments used during the ground firing test a r e the force link, dis
placement potentiometer, p r e s s u r e transducer, crystal vibrometer, s t r a i n gage, and
microphone. Listed in table I a r e the linear frequency range, the linear amplitude range,
and the calibrate factors for each of the instruments. All data obtained from these instru
ments were recorded on magnetic tape. A brief description of the installation and loca
tion of these transducers is presented in the following paragraphs.
Force links.- Force links were placed between the cable and Nike attach fitting
(fig. 4) to measure the preload and overall force on the torsion b a r s . Each link consists
of a strain-gaged square s t e e l rod, 2.032 cm a c r o s s the flats and approximately 3.30 cm
long.
Potentiometer.- In o r d e r to measure the mean position of the vehicle during firing,
a linear-displacement potentiometer w a s attached between the t e s t stand and a bracket
bonded to the Nike case approximately 1.90 m from the forward thrust face of the Nike.
(See fig. 5.)
P r e s s u r e transducers .- Three pressure transducers were installed in the Nike
rocket case. Transducers 2 and 3 were flush mounted i n the forward thrust dome of the
Nike approximately 180 apart. Transducer 1 (fig. 4) was mounted in the convergent sec
tion of the nozzle. The signal from each transducer w a s divided into two parts. The part
5

of the signal r e f e r r e d to as the dc pressure measurement represents the total pressure

at the transducer location. The part of the signal r e f e r r e d t o as the a c pressure mea
surement is the s a m e signal passed through a high-pass filter to remove the dc component
and amplified 28 times. The recovery time of the amplifier for this part of the signal due
to the initial pressure-pulse overload is 0.075 sec. In this paper, the two signals obtained
f r o m a single transducer a r e considered to be coming from independent transducers and
are referred to as the dc p r e s s u r e measurement and the a c p r e s s u r e measurement.
Vibrometers .- Six vibrometers (vibration transducers) were mounted on the vehicle.
Vibrometers 1 and 2 were mounted at the aft end of the Nike case (fig. 4) to measure lon
gitudinal and tangential motions, respectively. Vibrometers 3 and 4 were mounted at the
forward end of the Nike (fig. 6) f o r longitudinal and radial measurements, respectively.
Vibrometer 5 w a s also mounted longitudinally at the forward end of the Nike case on
an attachment bolt approximately 180 around the circumference from vibrometer 3.
Vibrometer 6 was on the forward end of the boilerplate payload (fig. 7) to detect longi
tudinal response.
The vibrometers were crystal accelerometers using charge amplifiers and were
mounted on thermal-electrical isolation studs. This vibration measuring system w a s
checked on a calibration shaker to 10 000 Hz and the amplitude response was down 1.2 dB
at that frequency.
Strain gages.- Three strain-gage bridge circuits were placed on the Nike to mea
s u r e longitudinal strains. They were located l rocket-case diameter (0.406 m) away
f r o m the forward thrust face and spaced at 120 intervals.
Microphones.- Two microphones were mounted external to the vehicle. Micro
phone 1 was mounted near the nozzle end of the vehicle. (See fig. 4.) Microphone 2 was
s e t near the midpoint of the boilerplate payload. (See fig. 6.)

DATA ANALYSIS TECHNIQUES

In general, the vibrations of the payload and solid-propellant launch vehicle a r e

random in nature and require statistical methods for evaluating the response. Procedures
for analyzing random signals from constant-parameter linear systems a r e available (for
example, ref. 2) and have been used in studying of response in structures (ref. 3). Four
main parameters a r e usually used to describe the basic properties of random data:
(1) mean-square value, (2) power spectral density, (3) probability density, and (4) auto-
correlation. A problem in using these parameters to evaluate the data from a rocket
firing is that conventional techniques available a r e based on the assumption of a constant-
parameter linear system with stationary inputs. Because the vehicle consumes a n amount
of fuel equal to approximately half its initial m a s s , there are definitely some time-varying

6
parameters. In addition, it is not known whether the input itself is a stationary process.
Therefore, in order to use well-defined random-analysis techniques f o r stationary data,
the concept of "self-stationarity" is applied. Self-stationarity assumes that the variation
i n the properties of the system will be l e s s than the anticipated statistical variations due
to the sampling theory.
The entire firing time f o r this rocket is nominally 3.4 sec; however, information of
interest is not available during the initial shock transient, which lasts about 0.1 sec, due
t o instrumentation limitations. Also, in the test stand a preload will cause the vehicle to
return to the stops on the stand before burnout of the rocket. Therefore, the data period
i n which the response of the vehicle is due only to the burning of the propellant with mini
mum restraints on the structure is from approximately 0.1 to 3 s e c of the actual burn
time. Analysis of the data was performed during this time period.

Information Obtained
Time history.- The data a r e presented on oscillograph records giving the time his
tories of each of the transducers. Inspection of these data provides estimates as to the
levels, frequencies, and general quality of the data. From these records, the exact time
interval f o r the statistical analysis was selected.
Power spectral density.- Power-spectral-density analysis yields the frequency com
position of the data. Power spectral density is a measure of the mean-square value of a
time history divided into relatively narrow bandwidths of frequency and has dimensions of
(Quantity)Z/Hz. The selection of the frequency bandwidths is discussed subsequently in
relation to confidence statements.
Probability-density function.- The probability-density function describes the proba
bility that the data will have values within selected bounds. To obtain a curve of proba
bility as a function of amplitude, a summation should be taken for an infinite time inter
val. A more practical method with real data, used herein, is the amplitude histogram in
which the data a r e divided into amplitude intervals and all the observations in this interval
a r e summed and plotted as a function of the amplitude interval.
The histogram yields information such as strong sinusoidal components, nonlineari
ties, clipping of data, and actual peak amplitudes.
Autocorrelation function.- The autocorrelation function describes the general depen
dence of the amplitudes of the signals at one time relative to the amplitudes at another
time. F o r a stationary process, the autocorrelation function is related to the power
spectral-density function by means of the Fourier transform. However, data presented
i n the form of autocorrelation plots point up information i n a different manner from that
of power-spectra plots. The autocorrelation function (1) shows sinusoidal components in
the data, (2) at zero time delay is the mean-square value of the signal, and (3) at long time

delays is the mean value squared. This function was used to obtain the mean-square
value f o r the various parameters. For vibratory motions, the mean value squared must
be zero, a fact that should be verified by the computed autocorrelation function. A non
zero value indicates that further analysis of the data is necessary.

Accuracy
In the mathematics of statistics, the parameters are computed f o r a n interval of
time approaching infinity. In reality, the test time involved is finite. Therefore, the cal
culations can only be estimates of the t r u e values. It is appropriate then to make some
statements about the accuracy of these estimates. Listed in table 11are the mathematical
equations defining each of the functions of interest.
Time history.- The limitations on the time-history playback do not involve averaging
and are restricted only by the amplitude and frequency-response characteristics of the
galvanometer of the recorder. The galvanometers were selected to obtain the full capa
bilities of the data acquisition systems.
Power spectral density.- The power spectral density is a function of both the f r e
quency bandwidth of analysis and the record length. A confidence statement about the data
is also dependent on these parameters. It was assumed that the estimate of the power
spectra would follow a x2 distribution curve (ref. 4) with the statistical degrees of
freedom

k = 2BT (1)
where

B bandwidth of analysis, Hz

T record length analyzed, s e c

The x2 distribution curve for the 90-percent confidence band is shown i n figure 8.
Once the statistical degrees of freedom have been determined for the analysis conducted,
statements giving a n estimate of the reliability can be made by using this curve. For
example, if the true average value of power spectra +x(f)av is known, figure 8 may be
used to determine the 90-percent confidence limits. For an estimate having 120 statisti
cal degrees of freedom, the measured power spectra +x(f)meas will fall between 0.8
and 1.22 of the true average value o r within 51 dB of +x(f)av. Alternately, if the power
spectra +x(f)meas has been measured with 120 statistical degrees of freedom, then
it can be stated with 90-percent confidence that the true average value is between
0 a82+x(f)me a s and 1.25iPx(f)meas. Additional information on this concept may be
found i n references 2 and 4. Since there is a limited test time for obtaining data from

8
the Nike firing reported herein, it is necessary to select bandwidths of analysis wide
enough to obtain reasonable confidence intervals and wide enough to use the assumption
of self-stationarity.
Amplitude histogram (probability density) .- The amplitude histogram does not
involve averaging; therefore, when using a digital approach i n the data reduction, the
parameter of most significance is the sampling rate. This rate is selected based on the
highest frequency t o be analyzed and will be specified. Although a "goodness-of-fit" test
could be applied to the histogram to compare it with a known distribution, the confidence
limits f o r such a short time period would be of such wide bounds that it was not consid
ered a meaningful evaluation for the test reported herein.
Autocorrelation function.- The estimates of this function were assumed to follow
the same distribution as power-spectral-density estimates. The method for determining
the number of statistical degrees of freedom w a s slightly different from that used for
determining power spectra and is related to the digital method of data analysis. The
statistical degrees of freedom k a r e expressed as
n
k = 2 -m

where

n sample size

m maximum number of time-lag values

The sample size is related to the record length by the following equation:

n=T
- (3)
E

where

T record length analyzed, s e c

E sampling interval of data, s e c

This value f o r k is identical with that evaluated for power spectra if a value of m is
selected such that

m = -1 (4)
B
where

B bandwidth of analysis, Hz
9
The confidence intervals for the autocorrelation function are then determined in the same
manner as for power spectra. The mean-square value is a particular value of the auto-
correlation function and, therefore, the accuracy is also determined i n the s a m e manner.

RESULTS AND DISCUSSION

The results from the firing i n the test stand are presented in figures 9 to 18. Listed
in table Ill a r e the data analysis estimate parameters used in this study. Oscillograph
records of measurements from the force links, displacement potentiometer, the forward
and aft end Nike dc p r e s s u r e transducers, the vibrometers, the forward and aft end Nike
a c pressure transducers, s t r a i n gages, and microphones a r e presented i n figure 9.
Results from analysis of data from vibrometers 5 and 6 covering burn time 0.1 to 0.9 s e c
a r e shown as the acceleration power spectral density in figure 10, the amplitude histo
gram in figure 11, and the autocorrelation function in figure 12. Results from analysis
of data from vibrometers 2 , 5, and 6 covering burn time 1.2 to 2.8 s e c a r e shown as the
acceleration power spectral density in figure 13, the amplitude histogram in figure 14,
and the autocorrelation function in figure 15. Analysis of data from the external micro
phones was conducted for approximately the entire burn time (i.e., 0.1 to 3.3 s e c ) , and
the results a r e shown as the acoustic power spectral density in figure 16, the amplitude
histogram in figure 17, and the autocorrelation function in figure 18.

Qualitative Evaluation of Data

The oscillograph records (fig. 9) indicate that the measured data from the three
f o r c e links a r e well within their operating range. Due to the instrumentation polarity,
the recording of the measurement of force link 3 is out of phase with force links 1and 2.
(See fig. 9.) The electrical noise level of the force links, particularly link 1, is extremely
high relative to oscillatory signal levels during the firing.
The displacement signal is within the anticipated range of the transducers and is
consistent with the phase of the measurement of force link 3. (See fig. 9.)
The signals from vibrometers 1, 3, and 4 severely overdrove both the conditioning
and recording equipment well beyond the anticipated levels, and therefore could not be
analyzed. (See fig. 9.) The data from vibrometers 2, 5, and 6 appeared to be within the
capabilities of the equipment and were analyzed.
The diaphragm of pressure transducer 1failed approximately 0.24 s e c after ignition;
thereby all vibratory and steady-state pressure data from this transducer were lost after
this time. The dc pressure measurement from transducer 3 was not recorded on the tape.
The dc pressure measurement from transducer 2 is satisfactory with the exception of
approximate burn time 1.92 to 2.61 sec. Although it is not known what actually caused

10
the problem with transducer 2 in this time period, it may be attributed either to the over
load in the power supply due to the failure of transducer 1 o r to a malfunction i n the
pressure-transducer power supply. The ac p r e s s u r e s as measured from transducers 2
and 3 show a similar problem during the same time period. The oscillograph records
f o r the ac pressure measurements do not indicate any pressure oscillations. However,
the sensitivity of these transducers may not have been high enough to detect the p r e s s u r e
fluctuations inside the Nike case.
The strains were well within the capability of the system to record the data and are
shown i n figure 9. Strain gage 3 failed at approximately 2.58 s e c and provided no data
after that time.
The microphone data are shown in figure 9. The recording equipment was not over-
driven; however, the records indicate that microphone 2 clearly has a n amplitude limit.

Quantitative Evaluation of Data

Differences in event times presented in this section were obtained from high-
resolution oscillograph records and a r e not observable in the small-scale plots in
figure 9.
Force link and displacement measurements.- The static preload measured by the
force links w a s as follows:
Force link 1 64.5 kN
F o r c e link 2 64.5 kN
Force link 3 64.6 kN
for a total indicated load of 193.6 kN. An initial load of 2.22 kN per link w a s placed on
the vehicle before zeroing the force links, bringing the total preload to 200.3 kN. The
maximum load as measured by the force links occurred 0.025 s e c after ignition and
reached an instantaneous value of 239 kN. This peak value is associated with an initial
high-frequency oscillation of approximately 130 Hz. The thrust buildup near the end of
the burning (approximately 2.5 s e c burn time) is about 204 kN. This value includes the
effects of the initial preload and the 227 kg loss in weight due to propellant burning.
The measurements of the force links indicate that the thrust of the rocket only
slightly exceeded the preload and if the system had not had the retractable s p a c e r s , the
rocket would have been restrained by the honeycomb stops for almost the entire firing.
Although allowances were made f o r variations from the nominal thrust of the Nike, it was
not anticipated that it would be this great. The displacement trace shows the vehicle
lifting off the stops at 0.019 s e c and that its initial peak value is 1.08 cm; however, the
second cycle of oscillation is 1.9 cm which was the maximum displacement of the vehicle.

11
 The frequency shown by the f o r c e links, when averaged over 0.5 sec, is 7 Hz which
is in agreement with the anticipated suspension-system frequency. Evaluation of the
amplitude decay at the 7-Hz frequency indicates that the damping of this mode is i n the
range of 3 to 4 percent of a n equivalent critically damped viscous system.
P r e s s u r e measurements.- The maximum p r e s s u r e of 7.65 MN/m2 at the forward
end of the Nike case is reached at 0.088 s e c , which indicates a slower thrust buildup than
anticipated. The low-frequency suspension oscillation started before the maximum p r e s
s u r e was reached. This may contribute to the second cycle of oscillation being greater
than the initial one. P r e s s u r e transducer 1 indicated a p r e s s u r e buildup at the aft end of
the Nike 0.008 s e c after that at the forward end. The average velocity of the p r e s s u r e
pulse between these two transducers is approximately 366 m/sec. The maximum p r e s
s u r e at the aft transducer was 7.30 MN/m2.
Whereas the oscillograph records did not show any p r e s s u r e oscillations, a power-
spectra analysis of the a c pressure measurements did indicate signals at frequencies
other than normal electrical power noise frequencies. If a normal distribution is assumed
f o r these pressures, their estimates f r o m the power-spectra analysis indicated that the
4-sigma peaks do not exceed 69 kN/m2.
Vibration measurements.- The responses of vibrometers 2, 5, and 6 (fig. 9), which
yielded usable data, show two time periods with significantly different vibration responses.
The measured vibratory response, in g units, from 0.1 to 1.2 s e c shows considerably
higher amplitudes than those measured during the time period 1.2 s e c to burnout. The
amplitude histograms f o r vibrometer 2 indicated s e v e r e clipping of the signal, and there
f o r e no additional analysis was conducted on these data during this time period. A power-
spectra analysis was made, by use of a filter bandwidth of 40 Hz, of the longitudinal vibra
tions obtained from vibrometers 5 and 6 during the time period 0.1 to 0.9 sec. The r e s u l t s
show a strong response in a frequency range from 10 000 to 10 400 H z . (See fig. 10.)
Vibrometer 5, which is at the attachment of the Nike to the payload, had a peak spectrum
of 0.006 g2/Hz both in frequency band from 10 000 to 10 400 Hz and at 1600 Hz. The over
all root-mean-square value for vibrometer 5 during this time was 5grmS, as determined
f r o m the autocorrelation function (fig. 12), with maximum random peaks of +14.6g and
-13.7g (fig. 11). The oscillations of the autocorrelation function also indicate a strong
narrow-band random signal at approximately 10 000 Hz. Vibrometer 6, which is on the
forward end of the payload, had a peak power spectrum of 0.038 g2/Hz. The spectrum at
all other frequencies between 0 and 20 000 Hz a r e at least 8 dB down from the peak. The
overall root-mean-square value for vibrometer 6 during this time was 8.2grmS, with
maximum random peaks of +38g and -29g. (See fig. 11.) The autocorrelation function
(fig. 12) also indicates some 10 000-Hz random components. The amplitude histograms
for vibrometers 5 and 6 (fig. 11) approximate a normal distribution. The peak values are
about 3 to 4.5 times the root-mean-square value with a zero mean.

12

I-

Results from analysis of the data in the time period 1.2 to 2.8 s e c for vibrometer 5
show a peak power spectrum of 0.0018 g2/Hz at an approximate frequency of 1900 Hz.
(See fig. 13.) The spectra are similar in magnitude and shape for the time periods ana
lyzed except near 10 000 Hz. (Compare figs. 10 and 13.) The same can be said for
vibrometer 6, with the peak spectrum being about 0.0036 g2/Hz at approximately 600 Hz.
Peaks i n the power spectrum at frequencies 600 Hz, 1000 Hz, 1300 Hz, and 1500 Hz can
be seen in the data for both time periods. Due to the lack of a significant response near
the 10 000-Hz frequency, the root-mean-square vibration levels for both vibrometers 5
and 6 a r e less for the later time period as a r e the peak accelerations. The vibration
levels a r e 2.5grmS and 4.0grms for vibrometers 5 and 6, respectively, with peak
values again being in the range of 3 to 4 times the root-mean-square value. The ampli
tude histogram shows the distribution to be similar to a normal distribution. (See fig. 14.)
Filtering the data obtained during the first time period (0.1 to 0.9 sec) so that the
response is flat to 5000 Hz and down approximately 22 dB at 10 000 Hz yields root-mean
square accelerations of 2.4grmS and 3.9grmS for vibrometers 5 and 6, respectively.
A comparison of these values with those in the second time period (1.2 to 2.8 sec) where
the levels were 2.5grms and 4.Ogrms, respectively, shows them to be within the allow
able range of k0.72 dB for root-mean-square values at the 90-percent confidence level
for an estimate with 646 statistical degrees of freedom. (See fig. 8.) It should be remem
bered the mean-square value is determined from the autocorrelation function and, in using
figure 8, a factor of 2 is necessary to determine confidence levels for the root-mean
square values. This agreement in the root-mean-square value is sufficient to demon
s t r a t e that the signals a r e weakly self-stationary (ref. 3) and, along with the agreement
in the shape of the power-spectra curves, give reasonable confidence that the concept of
self-stationarity for the parameters is a reasonable assumption for the bandwidths of
analysis and confidence limits used in this paper.
Vibrometer 2, which is oriented in a lateral direction at the nozzle end of the vehi
cle, shows a decaying low-frequency oscillation of approximately 0.9 Hz on which the high-
frequency oscillation is carried. The power spectrum of this signal indicates a relatively
flat spectrum of 0.0008 g2/Hz up to 10 000 Hz, with a peak in the spectrum below 100 Hz.
(See fig. 13.) The amplitude histogram shows a distribution which is similar in shape to
a random signal with a normal distribution and a clipped sinusoidal component. (See
fig. 14.) The root-mean-square acceleration determined from the autocorrelation func
tion is 4grms; however, it should be noted that this function did not return to zero in the
analysis time period, an indication that the mean acceleration squared is not zero. (See
fig. 15.) This is due to the decaying low-frequency oscillation having a long period of
oscillation compared to the analysis time and thus yielding a nonzero mean-acceleration
square value. The peak accelerations were +7.5g and -log, which a r e about 2 to 2 1 times
2
the root-mean-square acceleration.

13
 Strain-gage measurements.- The outputs of the s t r a i n gages indicated a maximum
s t r a i n of 430 pm/m on gage 1, which includes a response at 55 Hz during ignition through
0.4 sec. The gages then indicated relatively constant s t r a i n of 210 pm/m on gage 1,
227 pm/m on gage 2, and 252 pm/m on gage 3 until 2.6 sec; at this time the s t r a i n begins
to diminish and reaches z e r o at 3.6 s e c on all gages.
Acoustic measurements .- Both the power-spectral-density plots and the
autocorrelation-function plots f o r the microphones indicated a response primarily at
160 Hz. (See figs. 16 and 18.) The overall sound p r e s s u r e level at microphone 1 was
162 dB whereas the overall level at microphone 2 was 155 dB. However, the amplitude
histogram f o r microphone 2 shows this transducer to be severely overdriven. (See
fig. 17.) It is probable that the primary frequency is associated with a room resonant
frequency since the wavelength f o r 160 Hz is approximately equal to the distance between
the vehicle and the side walls of the test cell. It a l s o may be observed that this acoustic
level did not result i n a noticeably unique response in the t e s t vehicle at 160 Hz. (See
figs. 10 and 13.)

CONCLUDING REMARKS

A firing of a Nike rocket with a boilerplate payload was conducted in a special test
stand as a n initial step in the investigation of the vibrational environment of solid-
propellant rockets.
Since much of the data obtained during this project is random in nature, a detailed
discussion of the analytical procedures used to reduce these random signals is included
in this paper.
The test stand, which was designed to simulate flight boundary conditions, withstood
the thrust load without failure. The average fundamental resonant frequency of the vehi
cle in the stand w a s approximately 7 Hz, well below any anticipated resonant response in
the vehicle. The damping of the stand at its fundamental frequency was low enough to
allow the vehicle to respond freely.
This firing had two results which had not been anticipated. First, from a viewpoint
of mechanical vibration, it was expected that the frequency range of interest would be up
to 5000 Hz. However, data were lost when, during the first 1.2 s e c of burning, a vehicle
response i n the frequency range of 10 000 Hz to 10 400 Hz caused several of the vibrom
e t e r measuring systems to saturate. During the same time period, the remaining vibrom
e t e r s indicated peak values as high as 38g. The significance of these high accelerations
a t this high frequency is difficult to evaluate. Environment specifications covering
mechanical vibration seldom exceed 2000 Hz. Acoustic t e s t s may cover this higher f r e
quency range; however, there was no indication that this response was due to an acoustic
input.

14
I

The second unanticipated result was the inability of the pressure transducers in the
Nike rocket to measure the oscillatory pressure components during burning. A gain fac
tor of 2.5 over the steady-state pressures was far too low to detect pressure oscillations.
Evaluation of the pressure data indicated that the peak values of pressure w e r e below
69 kN/m2 and could not be accurately detected by these transducers. Since these very
low pressure oscillations caused vibration levels which were easily detected, improve
ment in the pressure detecting system is required in order to correlate input with
response.
Analysis of data from two vibrometers indicated that the acceleration power spectra
in the frequency range below 5000 Hz varied little between burn time 0.1 to 0.9 s e c and
burn time 1.2 to 2.8 sec. This agreement in both frequency and amplitude, along with
agreement of the root-mean-square value of vibratory accelerations in these frequency
ranges, provides the verification of the original assumption of self -stationarity. The
power spectra indicated peak responses at approximately 600 Hz, 1000 Hz, 1300 Hz, and
1500 Hz for both the vibrometers. The maximum power spectrum for the vibrometer at
the Nike-payload attachment was 0.0018 g2/Hz and occurred at approximately 1900 Hz.
The maximum power spectrum for the vibrometer on the forward end of the payload was
about 0.0036 g2/Hz at approximately 600 Hz.

Langley Research Center,

National Aeronautics and Space Administration,
Hampton, Va., September 30, 1971.

REFERENCES

1. Neff, W. J.; and Montes de Oca, R. A.: Launch Environment Profiles for Sounding
Rockets and Spacecraft. NASA TN D-1916, 1964.
2. Bendat, Julius S.; and Piersol, Allan G.: Measurement and Analysis of Random Data.
John Wiley & Sons, Inc., c.1966.
3. Houbolt, John C.; Steiner, Roy; and Pratt, Kermit G.: Dynamic Response of Airplanes
to Atmospheric Turbulence Including Flight Data on Input and Response. NASA
TR R-199, 1964.
4. Blackman, R. B.; and Tukey, J. W.: The Measurement of Power Spectra. Dover Publ.,
Inc., 1959.

15

 TABLE I. - OPERATING RANGE AND CALIBRATIONS

FOR THE DATA MEASURING SYSTEMS

-. .-

Linear operating range Calibrate signal

-
Instrument
Amplitude Frequency, Amplitude Frequency,
~~ __ -_
HZ
-~
Hz
Force link 1 80 kN 0 t o 500 64.5 kN dc
Force link 2 80 kN 0 t o 500 64.5 kN dc
Force link 3 80 kN 0 t o 500 64.6 kN dc
Displacement potentiometer 15 cm 0 t o 100 8.62 cm dc
a c pressure measurement 1 rt2.07 MN/m2 40 t o 3000 rt2.07 MN/m2 rtdc
a c pressure measurement 2 rt2.07 MN/m2 40 to 3000 *2.07 MN/m2 rtdc
ac pressure measurement 3 *2.07 MN/m2 40 to 3000 rt2.07 MN/m2 *dc
dc pressure measurement 1 10.18 MN/m2 0 to 5000 10.18 MN/m2 dc
dc pressure measurement 2 10.63 MN/m2 0 to 5000 10.63 MN/m2 dc
Vibrometer 1 rt14g 3 to 5000 *log 1000
Vibrometer 2 rt 14g 3 to 5000 *log 1000
Vibrometer 3 rt14g 3 to 5000 *log 1000
Vibrometer 4 rt14g 3 to 5000 *log 1000
Vibrometer 5 rt14g 3 to 5000 k log 1000
Vibrometer 6 rt42g 3 to 5000 k30g 1000
Strain gage 1 1000 pm/m D to 1000 600 pm/m dc
Strain gage 2 1000 pm/m D to 1000 600 pm/m dc
Strain gage 3 1000 pm/m 3 t o 1000 600 pm/m dc
Microphone 1 180 dB 10 000 150 dB 1000
Microphone 2 150 dB 10 000 140 dB 1000
~.

Note: The vibrometer measuring system wa vithin 1.2 dB at 10 000 Hz.

16

TABLE It.- PROPERTIES O F EXPERIMENTAL DATA

Time history :
x = f(t)

Power spectral density:

Probability density:

Autocorrelation:

Mean-square value :

Mean value:

17

 TABLE ID.-DATA ANALYSIS ESTIMATE PARAMETERS

Galvanometer response ........................ Flat, 0 to 10 000 Hz

t o 0.9 s e c (At = 0.8 sec)
Data analyzed during burn times .............. to 2.8 s e c (At = 1.6 sec)
Power spectra:
Frequency range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 to 20 000 Hz
Bandwidth, B .................................. 40Hz
Degrees of freedom, k, f o r -

Burn time 0.1 to 0.9 s e c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Burn time 1.2 to 2.8 s e c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

Amplitude histogram :

Burn time, s e c . . . . . . . . . . ............. 0.1 to 0.9 1.2 to 2.8

Sampling interval, E, psec . . . .............. 25 100

Autocorrelation :
Burn time, s e c . . . . . . . . . . . . . . . . . . . . . . . 0.1 to 0.9 1.2 to 2.8
Sampling interval, E , p s e c . . . . . . . . . . . . . . . .. 25 100

Maximum number of time-lag values, m . . . . . . . .. 99 99

Samplesize, n ...................... 32 000 16 000

Degrees of freedom, k . . . . . . . . . . . . . . . . .. 646 323

Bandwidth, B, Hz . . . . . . . . . . . . . . . . . . . . . 40 1 101

18

 --
L-70-8341.1
Figure 1.- Rocket test stand and payload.

19

 Pulley -

Force l i n k
- Bearing

Sector

Bearing

\\,
-Test-stand b r a c e beam
-S p r i n g - l o a d e d s p a c e r
-Honeycomb stop

-T o r s i on b a r

< Bearing

- Preload turnbuckle

Figure 2.- Schematic view of torsion-bar support system.

20
 L-70-8348.1
Figure 3.- Honeycomb stop and spring-loaded spacers.
 L-70-8347.1
Figure 4.- Mi: end of test vehicle in test stand.
 ,

L- 70-8346.1
to Figure 5.- Displacement potentiometer.
w
 L-70-8343.1
Figure 6.- Vibrometers at forward end of Nike and
microphone near midpoint of payload.
 L-70-8345.1
Figure 7.- Vibroineter on forward end of payload.
 1000 - I
I
I
I

I
+ I
I

I
500 - I I
I I
I I 9 0-per c e n t c o n f i d e n c e
1 I band
I I
I \
- I

+e
200 I
I l
k0.36 dB f o r power s p e c t r a
and c o r r e l a t i o n

- I \
100
\
I \
I
I \
I \
W
0
50 - I \
I \
I
I \
I \
\
4
m
I
U I \
.d
U
m
20 - I \
.d
U I \
I \
\
I \
I \
10 - I \
I \
1 \
I \
\
I \
5- I \
/ \
\
I \

I \
I \
f \
\
I \
2L \

I I I I 1 I - I
0 3

Figure 8.- ~2 distribution curve for estimates of power spectra and correlation.

26

- . .. ..
 Force l i n k 1
64.5m

0 Li Force l i n k 2

64.6

Displacement

r-.
rl
a

m
.rl
cl

S t r a i n gage 1

ti
625 L
-XI .
c I
II

E
zG
S t r a i n gage 2
*rl
m

&I
CI
cn 608
0 S t r a i n gage 3
.,w,.,...

' 6 40 """' ' " ' "" './

I..I I I I I I 1 1 1 1 1 I 1 1 1 1 I I I I I I I I ~ I II I I I I I I I I I I 1 1
0 1.0 2 .o 3.0 4 .O
Time, sec
Figure 9.- Oscillograph records of data from Nike firing in test stand.

27

 Vibrometer 1

Vibrometer 2

I
+10

I
/I
Vibrometer 4

0
-10

L
Vibrometer 5

Vibrometer 6

+30
0
-30
I

0 1.0 2.0 3.0 4.u

T i m e , sec

Figure 9.- Continued.

28

d c pressure 1

d c pressure 2

ac pressure 2

+2.1.

- ac pressure 3

11 I I I I I , i I 1.1.1 I i I I I I J I I I I I 1.1 I I

1
0 1.0 2 .o 3.0 4.0

Time, sec

Figure 9.- Continued.

29

0 I Microphone 1

+a95
-895

.3
a

Microphone 2

!
 -.

Vibrometer 5

L)
0
a
v)

! I I .l -I I I I I I 1

3.6~10

I I I I I I I I
~- I I I ,
0 2 4 6 8 10 12 14 16 18 2 0 ~ 1 0 ~
Frequency, Hz

Figure 10.- Vibrometer acceleration power spectra for burn time 0.1 to 0.9 sec.

31

 Vibrometer 5

5grms

I-I
cdw
-13. 7gDk 1 +14.6g
Pk
E O
&I
0
z
0

Acceleration, g units

Vibrometer 6

+38g
Pk

Acceleration, g u n i t s

Figure 11.- Vibrometer acceleration histogram f o r burn time 0.1 to 0.9 sec.

32

 Vibrometer 5

! ! I
- -

Vibrometer 6

0 5 10 15 20

Time d e l a y , sec

Figure 12.- Vibrometer autocorrelation function f o r burn time 0.1 to 0.9 sec.

33

m
&
a

1
Vibrometer 6
&?
--.
CY
M
~ r!~
.44 ~ ~ ~
$ 3 . 6 ~ 1 0 --~ \,$by.!
u
c)

m
i %;dQ\

$4
a

8
P. 3.6~10-~ -

3.6xlO-

34

 4%ms

A-
I Vibrometer 2

0 +7.5g

0 ---. '

Acceleration, g units

Vibrometer 5

Acceleration, g units

-12.3g Pk Vibrometer 6

0
Acceleration, g units

Figure 14.- Vibrometer acceleration histogram for burn time 1.2 to 2.8 sec.

35

 Vibrometer 5

16
N Vibrometer 6
M
m

0
c)

3
-4 0 f4-iK-A" --iJ--u--=
-4
I I -I I 1 4
0 2 4 6 8 10~10-~

Time delay, sec

Figure 15.- Vibrometer autocorrelation function for burn time 1.2 to 2.8 sec.

36

 Microphone 1

Microphone 2

Overall l e v e l
155 dB

L I I I I I ~~
1
I
-. ..

I
. ..

I I I
3
0 2 4 6 8 10 12 14 16 18 20x10
Frequency, Hz

Figure 16.- Acoustic power spectra.

37

 2.68 (kN/m2)rm
h

O
Microphone 1

P r e s s u r e , kN/m

Microphone 2

P r e s s u r e , kN/m '

Figure 17.- Acoustic pressure histogram.

38

 .... .. ..

Microphone 1

I I I I

1.28

Microphone 2

-. 43 - 1 - I I

2 4 6 8

Delay time, s e c

Figure 18.- Acoustic pressure autocorrelation.

NASA-Langley, 1911 - 32 L-7837 39

[S)
N A T I O N A L A E R O N A U T I C S A N D SPACE A D M I S T R A T I O N
W A S H I N G T O N , D.C. 20546
- P O S T A G E A N D FEES P A I D
NATIONAL AERONAUTICS A N D
O F F I C I A L BUSINESS
P E N A L T Y F O R P R I V A T E USE 5300
FIRST CLASS MAIL SPACE A D M I N I S T R A T I O N
USMAIL

015 0 0 1 C 1 U 32 711015 S00903DS

D E P T OF THE A I R F O R C E

A F WEAPONS L A B I A F S C I

TFGct LI B R A R Y / % C n t /

4TTN: f L O U BQhlMANV C H I E F

KIRTLAND ,AFS NM 87117

If Undeliverable (Secrion 158

Posral Manual) Do Nor Rerurn

" T h e aeronautical and space activities of the United Stntes shall be

conducted so as t o contribute , . . t o t h e expansioiz of h u m a n knowl
edge of pheitoiiiena in t h e atmosphere and space. T h e Adniinistrrrtion
shall provide for t h e widest practicable and appropriate disseminatioiz
of i n f o r ~ i a t i o nconcerning i t s activities and the results thereof."
-NATIONALAERONAUTICS
AND SPACEACT OF 1758

NASA SCIENTIFIC AND TECHNICAL PUBLICATIONS

TECHNICAL REPORTS: Scientific and TECHNICAL TRANSLATIONS: Information
technical information considered important, published in a foreign language considered
complete, and a lasting contribution to existing to merit NASA distribution in English.
knowledge.
SPECIAL PUBLICATIONS: Information
-.
% Y TECHNICAL NOTES: Information less broad derived from or of value to NASA activities.
in scope but nevertheless of importance as a Publications include conference proceedings,
contribution to existing knowledge. monographs, data compilations, handbooks,
.. sourcebooks, and special bibliographies.
' . .".
, I '-' TECHNICAL MEMORANDUMS:
. -.
.I i ".Information receiving limited distribution TECHNOLOGY UTILIZATION
. :,T&e&'use o-f jheliminary data, security classifica-
.1? PUBLICATIONS: Information on technology
tioi, or $tier reason% used by NASA that may be of particular
c i.
.J,.
interest in commercial and other non-aerospace
CONTRAkTOR REPORTS: Scientific and applications. Publications include Tech Briefs,
technical informati& g<perated under a NASA Technology Utilization Reports and
contract 25 grant and consklered an important
Technology Surveys.
.
Contribution ,tp existing knowledge.
.+.. .1..

.? . .
:?'..
'

Details 6rf.fhe availability of f h e s e publications may be obtained from:

SCIENTIFIC AND TECHNICAL INFORMATION OFFICE

NATIONAL AERONAUTICS AND SPACE ADMINISTRATION

Washington, D.C. PO546