BO 0 1 7 9 - 1 2

FIXTURES FOR B & K VIBRATION EXCITERS

Revised October 1987

 CONTENTS

1. INTRODUCTION 1

2. DESIGN CONSIDERATIONS 2

2.1. GENERAL 2
2.2. MATERIALS AND FABRICATION 2
Some General Rules for Fixtures 4
2.3. PREDICTION OF RESONANT FREQUENCIES 4
Beams 5
Plates 5
Stretching Resonance 7

3. SAMPLE DESIGN 8

3.1 . DESIGN PARAMETERS 8

3.2. CONSTRUCTION DETAILS 11
3.3. EVALUATION 11
3.4. TROUBLE — SHOOTING 13

4. CHARTS AND TABLES 15

4.1. BOLT HOLE PATTERNS FOR B & K SHAKERS 15

4.2. VALUES OF B FOR BEAMS 16
4.3. VALUES OF A FOR PLATES 17
4.4. PLOT OF b/L VERSUS A 18
4.5. NOMOGRAPH FOR PLATE EQUATION 19

5. BIBLIOGRAPHY OF TEST FIXTURE LITERATURE 20

 1. INTRODUCTION

The design and construction of fixtures can be a very complicated undertaking, as a look
at some of the literature will show. Some firms and research organizations keep a full-
time staff of qualified engineers occupied w i t h nothing but designing fixtures. Indeed,
w h e n the test involves a large airborne radar unit w h i c h must be tested at several g in
the frequency range from 10 to 2 0 0 0 Hz, the design of the pertinent fixture is a job for
an expert.

However, we have found that many of our customers have neither the staff nor the expe­
rience for this kind of work, and quite often only need a simple general purpose fixture
to handle all their testing needs.

This Application Note is intended as a practical construction guide for those customers
w h o have purchased a B & K vibration exciter and are faced w i t h the problem of building
a vibration fixture for it. It is not meant to be a comprehensive study of fixture design,
w h i c h would indeed fill quite a large volume. A bibliography of fixture and related litera­
ture is given in the last section for further study.

1
 2. DESIGN CONSIDERATIONS

2 . 1 . GENERAL

The basic idea w i t h any fixture is that it should transmit uniformly the forces produced
by the vibration exciter to the test object. It is generally impossible to fix a test object di­
rectly to the shaker table itself, and the fixture acts as a transition piece between the
two.

In B & K vibration exciters, the force from the driver coil is transmitted via the skeleton
to the four mounting holes at each corner of the table. The object to be tested usually
has a number of finite points (feet, mounting brackets, screw holes) w h e r e the force is
to be applied, the idea being to duplicate as nearly as possible the force input experi­
enced in service. Between the table and the test object is the fixture. In order to get the
most out of the vibration exciter, this fixture should carry the force from the four m o u n ­
ting holes in the table to the mounting points on the test object w i t h a m i n i m u m of loss
and distortion. In other words the fixture should be as rigid as possible. This is condition
number one in fixture design.

The acceleration level possible w i t h the vibration exciter is inversely proportional to the
total mass it has to drive, therefore the fixture should be as light as possible. This is con­
dition number t w o . An ideal fixture, t h e n , would have infinite stiffness at all frequencies
and zero mass, and it is this paradox which is the basis of fixture design.

Although a fixturing problem may seem very complicated, in fact many of the design par­
ameters are determined before-hand. The weight, for instance, is limited by the weight
of the test object and the force available from the exciter. The shape is determined in
part by the fact that it has to fit the exciter table at its bottom surface and the test object
at one of its other surfaces. Finally, the test conditions impose restrictions on the fix­
ture. A typical test specification calls for testing to 2 0 0 0 Hz w i t h no fixture resonances
below 1 0 0 0 Hz and an allowable variation in vibration levels between 1 and 2 kHz of not
more than 2 ; 1 between any pair of points on the table.

2.2. MATERIALS A N D FABRICATION

Of the materials generally considered for fixture construction, steel, a l u m i n u m and

magnesium are the most common. For very small items, beryllium is sometimes used,
but its very high strength and light weight are counteracted by its cost and difficulty in
machining.

The ratio of Young's Modulus to density (E/p) is a controlling factor for the natural fre­
quency of a construction, and since this ratio is roughly the same for steel, aluminum
and magnesium, the choice of material will not greatly affect the natural frequency of
the fixture. It w i l l , however, in the case of steel, affect the weight, and in the case of
magnesium, the cost. Table 2.1 compares the E, p and E/p of the three metals. Tita­
nium, used in B & K accelerometers, is included for comparison.

There are numerous ways to fabricate a fixture; among t h e m , machining the fixture out
of one solid piece of material, casting, bolting together several pieces to form an assem­
bly, building up an assembly by laminating strips of material together, and welding.

2
 Steel Aluminum Magnesium Titanium

2 10
Young's Modulus (E) N/m 20,7 x 1CH0 6,9 x 1010 4,14 x 1 0 10,7 x 1010
3
Density (p) kg/m 7840 2770 1800 4510
7
E/p N m/kg 2,65 x 107 2,49 x 107 2,3 x 107 2,38 x 10
800756

Table 2.1. Characteristics of fixture metals

Machining. Machining from solid stock usually yields a good fixture because there are
no joints to work loose or bolts w h i c h act as springs at high frequencies. It is also a sim­
ple, easy method of construction and is often used for small test items such as electrical
connectors or electronic components. However, for larger fixtures it would be expensive
and wasteful to machine from solid stock.

Casting. A cast fixture has the advantage of homogeneous, monolithic construction, as

well as very little machining work to be done. However, casting is only economically
feasible w h e n it involves making a number of units, because a master has to be made
up before-hand, usually of w o o d .

Bolting. Bolted fixtures are not recommended. Although it is a simple way to assemble a
fixture, it can produce a lot of trouble later o n , w i t h bolts working loose under vibration.
Furthermore a bolted connection is not as rigid as a welded connection and is usually
considered as a pin connection rather than a fixed connection w h e n designing.

Consider the t w o identical aluminum beams in F i g . 2 . 1 .

Fig.2.1. Comparison of two beams

The beam in Fig.2.1a is simply supported and that in Fig.2.1b is fixed at both ends. The
ratio of the frequencies of the t w o beams is 2,3. This is, in effect, the advantage to be
gained by welding instead of bolting.

Laminating. Something can be said for the method of laminating, whereby a fixture is
built up in layers from different materials. This technique has t w o advantages: 1) It can
be done by lab personnel w i t h fairly simple tools, and 2) by alternating layers of rubber
or plastic a highly damped fixture can be built. The disadvantage is that a considerable
amount of time is used in the construction, w h i c h makes it expensive.

Welding. For most fixtures for the B & K interchangeable head exciters, welding is proba­
bly the best construction method, both from the point of view of economy and that of per­
formance.

All seams should be fully welded on at least one side. Spot welds are no more effective
than bolts and should be avoided. After welding and before machining, let the fixture
stand overnight in an oven at 2 5 0 ° C . This w i l l relax some of the stresses induced by
welding. It will also reduce some of the warping.

3
 To compensate for this w a r p i n g , the top plate of the fixture should be made thicker t h a n
the design thickness, then machined down to the design thickness after annealing.

Wooden Fixtures. It should also be noted here that wooden fixtures have been used suc­
cessfully at low frequencies and that a reasonable fixture can be made of laminated
w o o d , w h i c h has a high damping coefficient.

2 . 2 . 1 . Some General Rules for Fixtures

1. The easiest and cheapest fixture for the B & K interchangeable head exciters is that
made of welded a l u m i n u m plate.

2. The mating surfaces of the fixture should be machined as flat as possible. If the sur­
face is a plate, the fabrication thickness should be greater than the design thickness
as some warping w i l l occur during fabrication which can later be machined off.

3. All seams should be full welded on at least one side. Avoid spot welds and bolts.

4. Attachment bolts from fixture to exciter table should be counter-sunk so that the
stressed length of the bolt is twice the diameter.

5. It is very difficult to eliminate resonances completely in the frequency range 1 0 0 0

— 2 0 0 0 Hz. What can be done, however, is to introduce some damping into the
structure in order to reduce the severity of the resonance. This may be done by bond­
ing a damping material onto the member in question or fitting rubber or wooden
blocks between resonating plates. Hollow fixtures may be pumped full of polyure-
thene foam.

6. Make a drilling template to match the threaded holes in the shaker table. This in­
sures proper alignment of the bolts.

7. Tighten all the bolts as hard as the limiting torque of the inserts will allow.

2 . 3 . P R E D I C T I O N OF R E S O N A N T FREQUENCIES

The load carrying ability of the material almost never enters into fixture design. This is
because, if the structure is designed so as to avoid resonances, it is over-designed by far
w i t h respect to static or dynamic loading. The resonant frequency then becomes the criti­
cal design factor.

The prediction of natural frequencies of a structure can be very complicated. However,

for the purpose of designing a fixture, only the first natural frequency is important. The
first natural frequency is the most damaging and the appearance of any resonances w i l l
often disqualify a fixture for use.

A further simplification is the assumption of a single degree of freedom. Probably every­

one reading this is familiar w i t h the concept of a simple mass m supported by a pure
spring of stiffness k, w h i c h w h e n deflected, resonates at a frequency

f - -^ - -1_ / J T Hz
By breaking the fixture down to a combination of simple elements and assuming a single
degree of freedom system, this formula can be used as the basis for almost all frequency
calculations in fixture design. The trick then comes in deciding what k should be.

4
2 . 3 . 1 . Beams

As an example, take the aluminum cantilever beam shown in Fig.2.2.

Fig.2.2. Cantilever beam

This is a fairly common type of configuration. For instance, resonant fatigue specimens
are typically bars of material clamped at one end in a fixture. The k value for a beam is

BEI
k = —~- N/m
3
L

where
2
E = elastic modulus N / m
4
I = second moment of area m
L = length m
B = a constant based on end conditions, given in Table 4 . 2 .

For a rectangular section,

3
bd d
I = —— nv and m = p bd L kg
where

b = width m
d = thickness of the beam m
3
p = density k g / m

Hence,

3 2
fiT /BEI / 12Ebd /Ed ~
3
d
co = /— = /—-pr - /— T = / — r = 5 x 10 -o- rad/s
3 3 4 2
ym J mL yi2pbdLL J pL L

which shows that for a rectangular section the resonant frequency is independent of
width. For the above example,

3
f = — * 5 x 10 x —«■ - 800 x — = 8 0 0 Hz
2 2
2TT L 0,1

2 . 3 . 2 . Plates

Because of the great number of edge conditions and shapes possible there has been a
great deal of work done on plate resonances. A very useful compilation of all this rese­
arch has been made by Arthur Leissa and published as a NASA bulletin (see bibliogra­
phy). However, anyone using this information is cautioned to read very carefully because
the various authors state their results in slightly different terms and it can be mislead­
ing.

5
 Most plate formulae are based on the work of W a r b u r t o n , w h o used the Rayleigh-Ritz
method of equating kinetic and potential energy and evolved a formula for resonant fre­
quency.

For a plate:

4 3
? 7T D o Eh
co = —A A , where D = —- 0
4 2
b ph 12 (1 -v )
where

h = thickness of the plate and

v - Poisson's ratio.
and A is a mode constant based upon edge conditions.

Fig. 2.3. Piate resonance parameters

Since the use of steel, aluminum or magnesium is assumed for fixtures, this can be sim­
plified to
A h -
f = 23,5 x 10 —n2
A, where h and b are in centimeters
b
(D
A
4 h
or f = 23,5 x 10 —?- X
L

f
minimum ^ Maximum <b, L)

(see Fig.2.3) and A may be found in Table 4 . 3 . This simplification holds true as long as
the following conditions are met:

1. The plate should have uniform thickness, no applied load and h less than 1 / 2 0 of b,
although this last is somewhat flexible.

2. Clamped edges are rigidly restrained. Simply supported edges allow bending at the
support. Free edges have little or no restraint.

6
3. E/p is approximately 25 x 1 0 N m / k g and v (Poisson's ratio) is 0 , 3 .

Note that the frequency thus obtained is the first mode frequency. For quick estimations
and comparison between alternative designs, a graph of b / L vs. A is given in Table 4 . 4
and a nomograph for computation of f is supplied in Table 4 . 5 . Note also the relation­
ships given in Fig.2.4, w h i c h w i l l simplify certain calculations.

6
 Fig.2.4. Frequency relationships of different plates

2 . 3 . 3 . Stretching Resonance

One question w h i c h usually occurs to the novice fixture designer is whether to take into
account the stretching of the fixture as well as the flexing modes of the plates. Gener­
ally speaking, the answer is no.

The stretching resonance frequency of any object is f = V / 4 L where V is the velocity of

sound and L is the length. Shape, width and thickness have no effect. V and fl_ for the
three fixture metals and titanium are given in Table 2 . 2 .

Steel Aluminum Magnesium Titanium

V m/s 5148 4990 4796 5990

fLm/s 1287 1248 1199 1497

80075?

For an aluminum fixture to have a stretching resonance below 2 0 0 0 Hz it would have to

be at least 6 2 4 mm long, w h i c h eliminates the problem for most fixtures. The usual
place w h e r e stretching resonances occur is in slip tables but these are a subject in t h e m ­
selves and won't be covered here.

7
 3. SAMPLE DESIGN

As an illustration of the foregoing principles w e will design a general purpose fixture

w h i c h w i l l fit a Type 4 8 1 3 or 4 8 1 7 Exciter Head. The requirements are:

a. It must have a working surface 4 0 0 mm square.

b. It must have no resonances below 1 0 0 0 Hz.

3 . 1 . DESIGN PARAMETERS

As in most cases, the dimensions of the fixture are pretty much predetermined. The top
surface w i l l measure 4 0 0 mm x 4 0 0 m m . The bottom surface (see Table 4.1) will mea­
sure 1 0 0 mm x 1 0 0 m m . If we simply mounted a plate on the shaker table it w o u l d have
to be quite thick and therefore heavy and would obstruct the movement of the table on
the downstroke. If we make the fixture very high, any off-centre weight in the test object
w i l l cause rocking. So we choose a height of 1 0 0 mm as reasonable and set the top
plate atop a 1 0 0 mm cube.

Since the forces in the drive coil are delivered to the mounting points on the table w e
add four diagonal braces such that the forces are led from the mounting points to the far
corners of the table. The layout of the fixture at this point looks like Fig.3.1 .

Fig. 3.1. Initial layout of fixture

The top plate area bounded by the diagonals can be approximated by a right isosceles t r i ­
angle w i t h the equal edges clamped and the remaining edge free (see Fig.3.1). The
clamped edges have a length of 2 8 , 3 cm. To have a resonant frequency above 1 0 0 0 Hz,
the required plate thickness would be given by eq. (1) using A = 1,31 from Table
4.3.

2 2
1000x b =
1000 x 28,3 =
h = 2. A * 2,60 cm
4 4
23,5 x 10 x X 23,5 x 10 x 1,31

This plate would weigh 1 1,52 kg by itself. In order to reduce this, w e can add another

8
brace to bisect the angle of the plate. This would allow us to reduce the thickness to
6,5 mm but we will actually make it 12 mm thick. There are t w o reasons for this; 1) If
all the plates are designed for exactly 1 0 0 0 Hz we could get a beating effect w h e n we
near that frequency and may even get resonances around 8 0 0 — 9 0 0 Hz due to the
composite effect of all the plates. It is therefore a good idea to spread the resonant fre­
quencies a little. 2) The top plate is the best place to put any " f a t " in because it is here
that the test object is mounted and there w i l l be extra stress induced at the mounting
points.

Fig. 3.2. General Purpose Fixture for 4813 or 4817

9
 The fixture now looks like that shown in Fig.3.2. The resonant frequency of the top plate
can be calculated using eq. (1) and the relation given in Fig.2.4.

4
12
f 4 x 23,5 x 10 x '—- x 1,31 = 1845 Hz
2
28,3

The side braces are shown in Fig.3.3. For brace (§) the calculations are as follows:

b
= 16
T '
From Table 4.3

2
X = 0,39+ 0,96 ( — ) + 0,36 (—) - 4,37

X = 2,09

For a 6 mm thick plate the resonant frequency of brace (§) is given by:

A 0,6
4
f = 23,5 x 10 x —^ * 2,09 = 1310 Hz

Fig. 3.3. Side Braces

The plate, of course, could have been 4,5 mm thick but 6 mm was a more convenient
size. Fig. 3.4 shows the diagonal brace (A) . The calculations follow:

b
— - 2,09
L

2
X = 0,39 + 0 , 9 6 ( — ) + 0,36 \—\ = 11,45

X = 3,38

10
 A
4
0,6
f = 23,5 x 10 x —~ 2 x 3,38 = 1091 Hz
20,9

Fig. 3.4. Diagonal Braces

3 . 2 . C O N S T R U C T I O N DETAILS

As it is not necessary to have a full 1 0 0 mm cube for a core, a 6 0 mm hole was milled
out after the main parts were welded together. This serves t w o purposes:

1 . It removes some of the weight, and

2. It provides a convenient mounting point for the accelerometer.

The hole is milled all the way through the core to the top plate and a 1 0 - 3 2 UNF
threaded hole is provided in the centre. The accelerometer is mounted upside down at
this point, thus putting the control as close as possible to the test object without being
in the way. 4 channels, each 3 mm square in section, are milled into the base so that
the accelerometer cable can be led out in whatever direction is convenient. It is a good
idea to write the sensitivity on a piece of tape and stick it on the fixture. This saves
looking for the calibration chart w h e n you set up.

The top plate was originally 2 0 mm thick, then after welding and soaking overnight at
2 5 0 ° C was milled down to 12 m m . The bottom surface was ground off to obtain a
smooth mating surface w i t h the shaker table.

Mounting holes can be made in the top surface according to need. At B & K w e made a
series of holes around the edge at 4 0 mm centre-to-centre. Into each hole is fitted an
M 1 0 helicoil. The total weight of the fixture was found to be 7,7 kg.

3 . 3 . EVALUATION

When all the welding and machining is finished, the fixture should be evaluated before
use, both as a check on the design and construction and as a way of knowing the condi­
tions at the table surface during testing.

11
 Fig. 3.5. Setup for fixture evaluation

Fig. 3.5 shows a typical evaluation setup. The fixture is mounted on the Exciter w i t h a
control accelerometer mounted just under the centre of the top plate. The signal from
this accelerometer is fed to the Input of the Exciter Control via the Conditioning Ampli­
2
fier Type 2 6 2 6 . Here the vibration level is set ( 1 0 m / s is a good level for evaluation)
and monitored. The acceleration level is held constant at the centre by a voltage con­
trolled amplifier in the Exciter Control.

A second accelerometer, preferably as small as possible so as not t o affect the resonant

frequencies, is mounted at various places on the fixture. The signal from this
accelerometer is conditioned in a second 2 6 2 6 preamplifier. The frequency of the Exciter
Control is now scanned in search of resonances at various points on the fixture. The fre­
quency can be adjusted manually on the Exciter Control t o locate the resonance within
0 , 0 0 2 Hz.
r

It is often useful if a "signature" can be made for different parts of the fixture. This is a
plot of the output of the second accelerometer versus frequency and is made by feeding
the output of the second 2 6 2 6 to a level recorder. Such a signature is shown in Fig.3.6
for the fixture designed earlier in this chapter.

The control accelerometer is mounted on the centre of the top plate, while the second ac­
celerometer is mounted on the mid-span. In this case the level is the same as the con­
trol level throughout most of the frequency range. Just before the first resonance the le­
vel creeps up a little so that it is about 3 d B (40%) higher at 1 0 0 0 Hz. The first peak is at
1 1 1 6 Hz. The calculated first peak should be at 1091 Hz, thus there is an error of only
2,2%. A second peak appears at 1 5 5 8 Hz, which does not comply with either of the next
two calculated resonances, but this is to be expected after the first resonance because
of interaction between resonances and anti-resonances.

What is important in evaluating a fixture is to find out what the actual levels are ac­
ross the whole table so that if the control is set for a certain level one can be reasonably
sure that no over-testing will occur. A vibration contour map of the fixture can also be
made at the first resonance to see which parts of the fixture have the highest accelera­
tion levels at this frequency. Then, if desired these could be damped or changed.

12
 Fig. 3.6. Fixture signature

3.4. T R O U B L E - S H O O T I N G

If it happens that the fixture fails its evaluation for some reason, it is still possible to re­
medy it. If a resonance appears at too low a frequency for example, that frequency can
be changed by bonding an extra brace on w i t h epoxy.

If the resonance build-up occurs in such a way that the level is too high near the upper li­
mits of the range even though there is no resonance w i t h i n the limits, this can be remed­
ied w i t h damping. Say, for example, that the build-up shown in Fig.3.6 should be
2
brought down to 2 5 % over 1 0 m / s instead of 4 0 % . We could then try to damp the re­
sonance d o w n .

We can be reasonably sure that the peak shown corresponds to the resonance frequency
of the diagonal brace. Therefore w e could either add some damping material to this
brace by coating it w i t h a damping paint or paste, or we could fit wooden blocks be­
t w e e n all the braces and epoxy them in place.

It is sometimes difficult to ascertain exactly where a particular resonance is coming from

because of transmission throughout the entire structure. If this is the case, a thorough
investigation can be made by impedance methods.

Fig.3.7 shows a setup for measuring the inertance of the different parts of the fixture. A
small hole is drilled at the point of application and tapped w i t h a 1 0 - 3 2 UNF thread. A
Mini-Shaker Type 4 8 1 0 and Impedance Head Type 8 0 0 1 are mounted at this point. The
Impedance Head measures force and acceleration simultaneously at the same point. The
force signal is used as a feed back signal to hold the force constant. The acceleration sig-
nal is fed to a level recorder via a second conditioning Amplifier Type 2 6 2 6 . A plot of ac­
celeration as a function of frequency is thus obtained from which the resonances can be
identified.

13
 Fig.3.7. Setup for measurement of inertance

14
 4. CHARTS A N D TABLES

4 . 1 . BOLT PATTERNS A N D INSERT SIZES FOR B & K SHAKERS

Square Tables

Table Dimensions
Depth of Hole Thread Size Insert Depth
Type A B

4811 44,7 35,9 7,5 M4 5,6

4812 66,5 53,9 7,5 M5 3,9
4813 100,0 81,0 15,9 M6 4,7
4816 66,5 53,9 6,8 M5 3,9
4817 93,5 81,0 12,7 M6 4,7
4818 139,7 121,0 12,7 M8 7,1
Mode Study Diameter
Head

4814 19,0 7,5 M5 3,9

800760

Table 4. 1.

15
4.2. VALUES OF B FOR BEAMS

Type of Load B

CANTILEVER
12
Uniform load

Point load at end 3

24 ( L - a)
3 4
Distributed load at end 4a a
3 L +
" L2 L3

2
Load at any point
F (-1)
Point load at end 24 ( w L + P)
plus uniform load 3 wL + 8P

FIXED ENDS
384
Uniform load

Point load 192

2 4
3 ^ +a2) L
Off-centre load
48^ a23

Uniform load plus point 384 (P + wL)

load 2P + wL

800761

Table 4.2.
4 . 3 . V A L U E S OF A FOR PLATES

2 2
X X * x*

0,14 0,14 0,37

0,14+ 0,47 f — j +0,14 ( — j 0,75 0,87

5,33 5,33 2,31

0,14+ ( — j +5,34 I —\ 6,48 2,55

14 21
5,33 + 3,55 (—) + 5,33 ( — ) > 3,77

0,39 + 0,96 (—) +0,36(—] 1,71 1,31

I ^ — _ ^ — - ^ 1 1
800762
* Indicates that b / L = 1

Table 4.3.

17
LU
>

-Q
LL
O
O
m

18
z
o
<

a
LU
UJ
H
<

Q_
cc
O
LL
X
a.
<

cc
O
O
z

19
 5. BIBLIOGRAPHY OF TEST FIXTURE LITERATURE

1. "Vibration and Shock Test Fixture Design", Tustin Institute of Technology, Santa Bar­
bara, California, USA, 1 9 7 1 .

2. "The Vibration of Rectangular Plates", G.B. Warburton, Proc, of Inst. of M.E.,

1953.

3. "Lectures on Shock and V i b r a t i o n " , Robert L. Stallard, Avco Corporation, W i l m i n g ­

t o n , Mass., USA, given March-June 1 9 6 7 at Northeastern University.

4. "Vibration of Plates", A . W . Leissa, NASA Publication SP-1 6 0 .

5. Handbook of Shock and Vibration, edited by Harris and Crede, M c G r a w - H i l l .

6. Mechanical Design and Systems Handbook, Harold A. Rothbart, Editor, McGraw-

Hill.

7. "Estimating Natural Frequencies", H.W. Schmitt, Machine Design magazine, Dec.

9, 1 9 6 5 .

8. "Speed of Sound Limits Size of Vibrating Structure", Wayne Tustin, Test Engineer­
ing and Management, Sept. 1 9 6 6 .

9. "Built-in D a m p i n g " , R.P. Thorn, Machine Design, Nov. 25 and Dec. 9, 1 9 6 5 .

10. "Impedance Simulation Vibration Test Fixtures for Spacecraft Tests", Terry D. Schar-
t o n . Bolt Beranek and Newman Inc. Van Nuys, California, USA*

1 1. "Vibration of Sandwich Panels in a V a c u u m " , Clemans A. Powell, Jr. and David G.

Stephens, NASA Langley Research Center, in Shock and Vibration Bulletin, No. 3 8 .

12. "Formulas for Natural Frequency and Mode Shape", Robert D. Blevins. Van Nos-
trand Reinhold Company.

There is also a wealth of information available from:

Shock and Vibration Information Center

Code 6 0 2 0 ,
Naval Research Laboratory
Washington, D.C. 2 0 3 9 0 USA

Quite often an article in their publication, "Shock and Vibration B u l l e t i n " , will have infor­
mation on the fixturing for a test even though the test itself is the main theme of the arti­
cle.

20