BRUEL & KJAER 4320 User guide

Triaxial
Accelerometer Set Type 4320
The Tr i axial Accelerometer is virtually three
accelerometers mounted in one housing. It
indicates the vibration in three mutually per-
pendicular directions . Principal application
areas are, balancing of machinery , vibration
testing , multidirectional shock testing and
others where the direction of vibration is not
known or where measurements in more than
one
direction are of interest.
BRiJEL &KJJER
Na!rum, Denmark . eye 80 05 00 . ,;;:' BRUKJA, Copenhagen . Telex: 5316
BB 4320

Accelerometer Set
Type 4320
April1967

Contents
0. Introduction 3
1. Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2. Factory Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3. Mounting Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4. Measuring Systems ................................. 28
5. Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
6. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
7. Accessories ........................................ 46
8. Conversion Charts, etc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
9. Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
'•

f,
0. Introduction
The accelerometer is an electromechanical transducer which produces at its
output terminals an e.m.f., proportional to the acceleration to which the trans-
ducer is subjected. The output signal can be electronically processed and
read on a
meter or some other suitable indicating device.
The Triaxial
Accelerometer is virtually three accelerometers mounted in one
housing. It
indicates the vibration in three mutually perpendicular directions
Principal application areas are, balancing of machinery, vibration testing, multi-
directional shock testing and others where ·the direction of vibration is not
known
or where measurements in more than one direction are of interest.
The
accelerometer is small, light and of rugged construction, suitable for
most vibration measurements, both in the laboratory and in field environments.
Versatility in use has been a prime goal and special
care has been taken to
ensure a
wide frequency range, high sensitivity and good temperature charac-
teristics. Furthermore, the accelerometer has low transverse sensitivity,
negligible mounting error and low sensitivity to severe environmental, con-
ditions, such as humidity, high temperature,
corrosive atmospheres and
magnetic fields. The acoustical sensitivity and influence of cable whip are
also minimized.
For absolute measurements it is necessary to know the sensitivity of the
vibration transducer employed. Each accelerometer is supplied with individual
calibration data and frequency response curves, all taken as part of the
production test procedure. The aim has been to supply the maximum amount
of information on the calibration sheet to ensure a completely predictable
performance of the accelerometer.
The Triaxial Accelerometer has type number 4340 when on its own in the
production and calibration stages. When it is sold it is accompanied by
certain
accessories such as connecting cable, mounting studs etc., and is
available only as
Accelerometer Set Type 4320. See Specifications at the end
of this book.
3

1. Description
Construction.
The Accelerometer Type 4340 is of the piezoelectric compression type with
a construction indicated schematically in Fig. 1.1 .
Mass
Base
Output
terminals
Fig. 1.1. Schematic drawing of a piezoelectric accelerometer.
Each transducing element consists
of two piezoelectric*) discs on which is
resting a heavy mass. The mass is preloaded by a
stiff spring and the whole
assembly is mounted in a metal housing with a thick base. When the accelero-
meter is subjected to vibration, the mass will exert a variable force on the
piezoefectric discs . This force is exactly proportional to the acceleration of
the mass. Due to the piezoelectric effect a variable potential will be developed
across
the two discs , which is proportional to the force and therefore to the
acceleration of the mass. For frequencies much lower than the resonance fre-
quency
of the mass and the stiffness of the whole accelerometer system the
acceleration
of the mass will be virtually the same as the acceleration of the
whole
transducer , and the potential produced will therefore be proportional to
the acceleration to which the transducer is subjected. This potential can be
picked' up from the
output terminals of the accelerometer and used for
determination of the vibration amplitude, waveform and frequency. ,
Although the accelerometer has been designed especially with multidirectipnal
measurement in mind , it is also a versatile general purpose accelerometer
which can successfully meet the requirements of most vibration laboratories.
*) Lead Zirconium Titanate.
4

Fig. 1.2. Physical dimensions of the triaxial accelerometer.
Physical Dimensions.
It is generally true that an accelerometer should be as light as possible in
order not to influence the vibration of the specimen on which it is mounted.
For measurements on heavy machinery etc. this causes no problem but
for
lighter structures, such as for example a thin metal plate, the transducer
weight is important.
The resonance frequency of a vibrating single degree of freedom system is
fo= _1_
1/ k -
2n V mo
where mo is the effective lumped
mass and k is the lumped stiffness restrain-
ing the mass.
It is easily seen that adding another mass, e.g. an accelerometer, will result
in a change in resonance frequency according to the formula
V
---
~- -- ~
fo - ml + mo
where f1is the new resonance frequency and m1is the added mass.
Thus any mass added
to a vibrating structure will influence the vibration to
some extent, but provided the added mass is small the influence is often
negligible . A 10% increase in mass will reduce the resonance frequency by
about
5%.
The sensitivity of a piezoelectric accelerometer is directly proportional to
the mass acting on the piezoelectric disc. This limits the extent to which the
accelerometer weight may be reduced without loosing too much of the
sensitivity. For many
applications it is also important that any strain developed
in the material on which the accelerometer is fixed is not transmitted to the
piezoelectric element as this would show up in the output signal, indicating
an acceleration which actually is not present. It is therefore necess:ary to
make the accelerometer base thick , in order to isolate the piezoelectric disc
from such strain. The requirements to high sensitivity, low strain sensitivity

and good high frequency response are conflicting with requirements to small
physical dimensions and weight, so that
for general purpQse accelerometers
some useful compromise must be found.
The Triaxial Accelerometer Type
4340 has a titanium base and a total weight
of
35 grammes (1 . 23 oz). Fig. 1.2 gives its physical dimensions.
Environmental Sensitivity.
Accelerometers are often used to measure vibration in the field or on
specimens subjected to severe environmental tests. It is therefore important
that
their sensitivity to environmental changes is as small as possible. The
factors that may influence accelerometer performance are primarily tem-
perature, humidity and rapidly varying ambient pressure (Sound). The
temperature effect is to reduce the voltage sensitivity
of the accelerometer
at higher temperature, but if the accelerometer has undergone a suitable
temperature cycling process in the production stages, the sensitivity will
revert to its normal value w~en the temperature is brought back to normal
again. Beyond a certain temperature (the Curie point) the piezoelectric element
is permanently damaged.
-148'F
-100
dB
6
-3
-4
-5
-6
Voltage Sensitivity
571'F
300 •c
Temperature
1'+622
Fig. 1.3. Typical performance characteristics for the accelerometers in the
temperature range -100 to +260°C.
The triaxial accelerometer Type. 4340 is designed to be used for temperatures
up to 260°C (500°F) without cooling. The sensitivity will be slightly reduced
at the
higher temperatures, but' the necessary heat cycling process has been
carried out, so that no permanent change will take place.
6

It should1 be noted that great care has been taken to use materials which will
withstand high temperatures. The thermal coefficient of expansion of in-
sulating ', material and metal parts are carefully matched in order to maintain
humidity sealing.
It is, however, very
difficult to guarantee a perfect seal for immersion in
water when the accelerometers have been exposed
to the high temperature
limit. Therefore accel.erometers which are to be used in liquids should not
be exposed to more than 150°C for long periods of time.
The accelerometers a're
brought up to 250°C and left to cool down to about
50°C several times during production and calibration until performance is
stable.
Tests have also been conducted in order to find the influence of low
temperatures on accelerometer performance. The voltage sensitivity increases
steadily down to some -100°C and then levels out, while the capa<J:ity under-
goes a gradual decrease with decreasing temperature. The charge sensitivity
is
practically constant.
Typical performance characteristics are given in Fig. 1.3 for the temperature
range -100
to +260°C (- 150 to +500° F).
Cooling.
The Triaxial
Accelerometer has provision for water cooling , so that it may be
used
for measurements on surfaces with temperatures up to 1000°C (1800° F)
without seriously affecting performance.
Threaded connection tubes are supplied with the accelerometer
for easy
connection to a water supply. See Fig. 1.4.
2 67031
Fig. 1.4. Arr~ngement for water cooling of the accelerometer.
The water should preferably be supplied via
silicon rubber tubes, ~uch as
the B & K Type AF 502, 3.8 mm (0.15") internal diameter. The amount of
water necessary may be in the order of 1 to 20 litres per hour (1 /4 to 5
gallons per hour}.
7

It is also possible to reduce the temperature of the accelerometer base by
inserting a
cooling plate between the base and the measuring surface.
When a
mica washer surch as the one included in the Accelerometer S' et
4320 is used between the mounting surface and the cooling plate, experience
has shown
that temperatures up to 350-400°C may be measured on the
mounting surface with less than 250°C in the acca.lerometer base.
A stream
of air passing over the cooling plate consi. derably aids the cooling
process.
For high temperature measurements one should always employ Teflon cable,
B & K
part No. AO 0038. The cable delivered with the accelerometer , B & K
No. AO 0037 is based on Polyethylene and PVC which
will not withstand more
than 100°C. Also for low temperature measurements, from - 40 °C downwards ,
the
T~flon cable should be employed.
267 0 30
Fig. 1.5. Sealing of cable entry.
Vacuum Test.
The
accelerometer is sealed and tested under water in a evacuated jar for
leaks. This eliminates the risk
of inferior performance in moist atmospheres or
environmental test chambers where heavy condensation may take place .
When the accelerometers are used for measurements in liquids or in very
moist environments it is necessary to seal the cable entry, as shown in
Fig. 1.5.
A good se?lant
for the cable entry is for example Dow Corning Silastic
RTV 731 (room temperature vulcanizing silicon rubber) or General Electric
equivalent. These sealants show excellent performance for a wide tempera-
ture range, (- 100 to +.500° F).
Acoustical Sensitivity.
Effects due
to acoustical excitation are unavoidable with piezoelectric
accelerometers, but may pe reduced by careful design. The B & K
accelero -
meters are
of a rigid, mechanically isolated construction and pressure vari-
ations in the air will have little effect on the force exerted on the piezo-
8
· "'

electric element. One may generally assume that the vibrations induced in
the vibrating specimen will give rise to a much higher acceleration signal
than the
direct acoustical excitation of the accelerometer itself. In cases
where very low accelerations are to be measured in an intense acoustic treld
however, care must be exercised in order to obtain correct results. Typical
acoustic sensitivity for the accelerometers has been measured and found to
be less than 0.2 fA- VI fA- bar, i.e. less than 0.2 fA- V for 74 dB sound pressure level.
At 140 dB sound pressure level the output is less than 400 fA- V.
Magnetic Sensitivity.
The magnetic sensitivity has been found to be less than 2 fA- V/Gauss for the
least favorable orientation of the accelerometer in the magnetic field.
Polarity.
The polarity
of the accelerometer is such that an acceleration directed from
the mounting surface into the body of the accelerometer results in a positive
e.m.f. on the centre
conductor of the output terminals.
ShQck Performance.
In general the accelerometer will withstand shocks higher than 500 g and
the
output will be linear for shocks of this magnitude both in the positive and
in the negative directions.
Effect of Mounting Torque.
The
effect of mounting torque is less than 1 % change in sensitivity up to
5000 Hz for 6-60 kgcm (5 to 50 lb-in) mounting torque.
Long Term Stability.
A prime goal during design and manufacture has been to achieve maximum
stability under severe environmental conditions. However, when an accelero-
meter is used
for measurement under severe conditions of shock and .heat,
one shour'd not rely on this as a primary standard. It would be advisable to
recalibrate such an accelerometer at relatively short intervals of time.
A
calibration vs. time history is closely followed for a number of representative
units takeri from
production lots and they show less than 2 % change p~r
year.
Piezoelectric Materials.
The quality
of piezoelectric accelerometers depends largely upon the per-
formance of the material used for the sensing elements. Monocrystalline
materials such as qu&rtz and Rochelle salt, have been used in the past, but
are now superseded by polycrystalline,
artificially polarized ceramics like
barium titanate, lead zirconium titanate, lead metaniobate or similar materials.
9

Large variations in electrical and mechanical properties are obtained by small
changes in composition, and most manufacturers are continuously investigating
and improving their compositions in order to make them more suitable for
their particular purposes.
The most important factors for general purpose accelerometer~ are Curie
point, sensitivity, temprerature stability , capacity, resistance and time stability.
Unfortunately some of these are conflicting, high sensitivity for example, has
ofte,n to be sacrified for good time or temperature stability.
In tl)e table below are given some properties of commonly used ceramics as
given by
their manufacturers. The Curie point is the temperature at which the
ceramic changes its crystal
structure and looses its polarization, the piezo-
electric constant indicates the sensitivity, and the dielectric constant indicates
the
capacity for a given shape and size. Trade names are used C\S the true
composition of the ceramic is usually
not disclosed by the manufacturer.
Manufacturer and Curie
point
Piezoelectric
Dielectric
Trade Name Constant Constant
oc
10 -
12
Coui.!Newt. (Rei. Perm.)
Clevite
Ceramic
"B " 115 149 1300
Clevite I
PZT4
325 I
285
1475
Quartz & Silice
p 1-60
351
I
400 1500
Plessey !
Casonic Grade 3 120
I
83
1000
I
Quartz
300
I 2 4
I
B & K (1966)
350
300 1500
I
Natural Resonance Frequencies.
The natural resonance frequency of an accelerometer is not fixed. It depends
not only upon the mass and stiffness of the accelerometer but also upon the
mass (and stiffness) of the
object on which it is mounted.
Some confusion persists as to what is the natural frequency and, different
defihitions are seen.
The situation is best illustrated with a drawing. Fig. 1.6a shows a schematic
drawing of an
accelerometer which consists of several masses and springs
connected
in series. These springs are representations of the contact stiffness
between the various parts of the accelerometer.
10

Seismic mass
Contact
stiffness
Piezoelectric disc
Contact· stiffness
Piezoelectric disc
Contact
stiffness
Mass of base
a)
Fig. 1.6.
M
K= Equivalent
stiffness
·b) 2.
66o 20
a. Schematic of accelerometer as a dynamic system.
b.
Simplified system.
This mechanical system can be further
simplified into that of Fig. 1.6b. Here
M is the seismic mass resting on the piezoelectric element, B is the mass of
the accelerometer base and housing. K is the equivalent stiffness of the
system between M and B. The natural frequency
of such a system is equal
to
fo= fMV1 +
~
where fM is the natural resonance frequency of the mass M upon the spring
of stiffness K.
Now two resonance frequencies are easily thought of:
1. The tree hanging resonance frequency, i.e. the resonance frequency obtain-
ed with the system freely suspended in air .
This resonance frequency is entirely dependent upon
the ratio of M and B,
and it is seen that making the base B very Iight the resonance frequency
may be very high. This resonance frequency is therefore
of little practical
value, in fact the higher it is compared with the mounted resonance fre-
quency defined below
the poorer is the mechanical construction of the
transducer . (A thin base may cause bending of the piezoelectric due to
strain from the mechanfcal part on which the vibration is measured).
11

8
B=base+180 grams
Fig. 1.7. Possible configurations for definition of natural resonance frequency.
2. The mounted resonance frequency with the accelerometer firmly fixed to
a
structure of infinite stiffness and mass.
This results in a value for B of infinity and the resonance frequency is
equal to fM which is the natural resonance frequency of the mass M on
the spring of stiffness K.
This resonance frequency is of great practical value as it is approximately
the one obtained when the
accelerometer is mounted on a structure which
is heavy compared with the accelerometer mass.
Unfortunately it is
difficult to measure the mounted resonance frequency as
defined above, since an infinitely heavy mass is hard
to bring into motion.
A compromise has therefore
to be made and there are standards which
define
the mounted resonance as that obtained with the accelerometer mount-
ed on a one inch
cube of steel (Bruel & Kjrer use a steel block of weight 180
gram). Such a block is easily set into controlled motion at frequencies up to
some 50 kHz.
This mounted resonance may be used
directly as a basis for vibration
measurements as long as the accelerometer seismic mass is much smaller
than the mass of the calibration mounting blocck. The natural frequency ob-
tained is
practically identical with the mounted resonance frequency on an
infinite mass.
However when the accelerometer mass is large the natural frequency obtained
will be too high, thus if B = 5 M we obtain
fo = fM v 1+02 = 1.1 fM
i.e. a value which is 10% high.
For the Bruel & Kjrer acccelerometers the ratio
of weights of the seismic mass
to the weight of the base is about 1 : 1 and therefore the free hanging
12

resonance will be about 1.5 times the mounted resonance and the resonance
obtained on the
180 grams block of steel is 2-4% higher than that which
would be obtained on an infinite mass.
A small
difference in resonance frequency is obtained by considering the
damping present in the accelerometer. Thus undamped natural frequency is
defined as
the frequency at which the difference between the motion of the
mass and base lags
the motion of the base by a phase of 90°, whereas
damped natural frequency is the frequency of decaying oscillations after an
initial excitation. The
difference is negligible for high Q piezoelectric accelero-
meters such as the ones manufactured by Bruel & Kjcer.
·The
undamped natural frequency given on the 8 & K accelerometer calibration
charts is the free hanging undamped natural frequency.
13

2. Factory Calibration
Each individual accelerometer has before leaving the factory undergone a
very thorough ageing, testing and calibration procedure to ensure the user of
a high quality product. It is supplied with three individual calibration charts
giving a complete set of data for each measuring direction .
Typical calibration charts as supplied with the accelerometer is shown in
Fig. 2.1.
C•llb,• tio nCh•rtfor
Acc:elerom•terTyfM4340
Br c~! ! .~ ._'!!Mr
·=:
1S...oH1wHy U
5:£'~P= Ao~
lOS lOS lOS P' '
2~7 2(,8 2/,5 mV (I'
22,9 24,4 2'0',8 pC/ 11:
C.p.c:lty (lnclulll ns c •blt l 1(}6( lf/9 116o pi"
~o~::·;o: t: en•s-oJtMtr ~o 3,7 { <~- .,,
· rt~ff.i:~1t§~~~~~I:~:~~~i. , .
ou,/-3 - 6'7s;
1
,.
1
turt ... EF:
:~ . ::;;- m:.:•
Cable Capacity.
BoUol & iJ;;; O
.. . :::.".:' ,' ·•· ···· .
=~~ t .:';~ IO.S2"F
IIMfrictiC-II~ I H o""''c:oultl
H-l'lllty oS. IJe<l
....,, _, ___ •2&0'Cor800'r<
§~~;~~?.~~~;~~: :: ::"
26703.3
Fig. 2.1. Typical calibration charts.
The capacitance of the accelerometers are given including the capacitance
of the ordinary B & K connecting cable, so that when other cables are used,
the change in cable capacitance must be taken into account. as this deter-
mines the effective sensitivity of the accelerometer. To find the capacitance
of the accelerometer alone, deduct the cable capacitance given on the
calibration chart. The standard length of cable included is 1.2 m (4 ft).
The
capacity of the 1.2 m cable is around 105 pF (micromicrofarad) but the
individual value is given on the calibration chart.
Voltage Sensitivity.
The voltage sensitivity of the accelerometers with connecting cable is deter-
mined at a frequency of
50 Hz and at room temperature (appr. 20°C).
Fig. 2.2 shows the instrumentation employed.
14

' . Microphone
Amplifier
Type 2603
Frequency
Analyzer .
Type 2107
Unknown
Accelerometer
265726
Fig. 2.2. Set-up used for calibration of absolute sensitivity of the accelero-
meters.
The.accelerometers are calibrated at
50 Hz, the frequ~ncy being monitored by
a cr,ystal controlled
electronic counter, and the perioaic time of the vibration
is kept within 200,000 ± 100 ftSec. Thus the frequency accuracy is ± 0.05 %.
The shaker providing the mechanical excitation is mounted on a large con-
crete block which is supported by steel springs, giving the whole assembly
a natural resonance frequency of about 1 Hz. This ensures isolation from
background vibration.
The calibration of all units is achieved by comparison between a standard
accelerometer and the unknown in a back to back mounting as shown in
Fig. 2.2. The standard is frequently checked with a
microscope observing the
displacement directly, with an accuracy of ± 1 %.
The distortion in actual vibration units is kept below 2 % and max. influence
on peak to peak amp)itude below 0.5 %. ·
The output from the accelerometer is measured with a B &K Frequency
Aanalyzer set to 50 Hz and calibrated with a reference voltage of exactly the
same frequency as the signal to the shaker,
giving a total voltage measuring
accuracy of ± 0.15 %.
The total accuracy on the sensitivity calibration of the accelerometers is
thus better than ± 2 %.
15

The calibration vs. time history is closely followed for a number of re-
presentative units taken from production and they show less than 2 %
change per year.
Note that the calibration is valid for the accelerometer with its individual
connecting
cables ,. ,and remember to take any ch~-nge of cable capacitance
into account when different cables are employed.
Also the input capacity
of the preamplifier must be taken into' account. The
influence
of this capacity in the calibration set-up is eliminated with a
substitution measurement. In
practice the capaci ty will be less than 10 pF
using a normal B & K amplifier and plug such as B & K Cathode Follower
Type 2615 , Source Follower 2616 or 4 pF in the case of preamplifier 2623 .
Note" also the units employed in the calibration. The sensitivity is given in
mV /g1= mV RMs/ g RMs = mV P••k/ g pe•k· If, for example, the sensitivity is required in
mV RMs/ g P••k the sensitivity given should be multiplied by 0. 707.
Charge Sensitivity.
This sensitivity is calculated from the voltage sensitivity and the equivalent
capacitance
of the accelerometer with normal connecting cable.
Charge sensitivity is expressed in pico-coulomb/g and is independent upon
the capacitive loading on the accelerometer. It is determined by multiplying
the voltage sensitivity with the total capacity used in the calibration, i.e.
capacity
of transducer and cable.
S chorge = S voltage (C . + Cc)
where c. = accelerometer capacity and Cc= Cable capacity.
Since the voltage sensitivity is given as mV/g the charge sensitivity will be
in pico-Coulomb/g
or pico-Coulomb peakl g pe•k·
,.
Capacity.
The capacity of the accelerometer is given for each direction Including the
cable capacity as specified on the calibration sheet. Accelerometer capacity
comes into the question when the low frequency cut-off
of the measuring
system is computed, as
it determines the effect of loading on the accelero-
meter.
The capacity is measured at 1000 Hz in a capacitance
bridge , comparing
with a standard capacity equal to
the nominal capacity of the accelerometers.
Accuracy ±
0.5 %.
Transverse Sensitivity.
Transverse sensitivity i~ the maximum sensitivity to a transverse acceleration
expressed in percent
of the reference sensitivity in the intended measuring
direction.
16
'

The transverse sensitivity of the accelerometers is primarily due to irregular-
ities in the ceramics, and limitations in the mechanical coupling between
ceramic and metal parts. Careful mechanical machining helps to minimize
transverse sensitivity.
Transverse sensitivity can be regarded as a result of the particular axis of the
accelerometer.making a small angle with the direction
of the maximum sensi-
tivity as shown in Fig.
2.3.
Main axis
sensitivity
100%
\
\
\
\ Minimum
1 sensitivity
in this
direction
I
I
I
I
I
I
I
I
I
I
I
Maximum
sensitivity
Fig. 2.3. Graphical illustration of transverse sensitivity.
Imagining the accelerometer placed in a rectangular coordinate system, as
shown,
the vector representing the maximum sensitivity can be resolved into
two components: the main axis sensitivity which is the one called voltage or
charge sensitivity on the calibration chart, and the transverse sensitivity
given on the calibration chart as a percentage
of the main axis sensitivity.
The voltage obtained from an accelerometer is a
product of the acceleration
vector and the component
of the sensitivity in the same direction, so that
in a certain transverse direction, at right angles to both the main axis and
the max. transverse sensitivity axis the accelerometer output will
be minimum.
In order to maintain a low transverse sensitivity the user should always
mount the accelerometer on a flat, clean surface and keep the accelerometer
mounting surface free from burrs and scratches. Also he should avoid very
large shocks, such as might be caused by dropping the accelerometer on
the floor, and large temperature shocks.
17

If thd accelerometer is properly handled and mounted, the transverse
sensitivity will normally be below 4% up to 5000 Hz. Above 5000 Hz it
is very difficult to establish useful test facilities for transverse measurements.
The type of mounting of the accelerometer will, however, undoubtedly set
a
limit to the good performance of the accelerometer in the transverse
direction at a frequency somewhat lower than that for operation in the main
direction .
The transverse sensitivity for each axis of the accelerometer is specified as
being less than
4% of the sensitivity. However, the individual maximum
transverse sensitivity is given on the
calibration chart.
In the factory calibration the transverse sensitivity is measured at approx-
imately 30 Hz on a rotating table whose motion sideways is more than 100
times larger than the motion in the vertical direction. See Fig. 2.4.
Fig. 2.4. Part of the arrangement for obtaining transverse sensitivity.
The accelerometer is first mounted with its main axis parallel to
the direction
of motion of the table and the deflection on the meter is adjusted to a
reference
mark (100 %). Then the accelerometer is mounted with its main
axis at
right angles to the motion and the table is rotated slowly. The meter
is set to indicate full scale deflection for 10% transverse sensitivity and the
maximum value is noted.
18