BRUEL & KJAER 2620 User guide

Accelerometer Preamplifier
Type 2620
\
R-•Vft
-n .. ao . > ·~-~
A two channel preamplifier for feeding accelero-
meter signals to vibration controlling and in-
dicating equipment . The design obviates calibra-
tion of the measuring system.
Ncerum, Denmark • ~ 80 05 00 • ~ BRUKJA, Copenhagen . Telex: 5316
88 2620

Accelerometer Preamplifier
Type 2620
SEPTEMBER 1963

Contents
l. General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1. Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.
Impedance Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.
Gain Adjustment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4.
Shaping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1. Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.2.
Input Circuits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.3.
Low-Pass Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.4.
Attenuators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.5.
Monitor Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.6.
Output Amplifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.
7. High-Pass Filter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.8.
Main Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.9.
Cross- Talk . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.10. Phase-Frequency Hesponse ........ ... .. .... . ................ 11
2.11. Power Supplie s ...... .... . . . . .. . . . .. . .. .. . ................. 11
2.12. Noise ...... ..... ....... . ....... .... .... .......... ......... . 11
3. Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1. Controls and Sockets ....... . ................................ 12
3.2.
Power-Line Voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3.
Accelerometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4.
Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.5.
Hum and No i se Heduction ............................ . .. .. .. 15
3.6.
Warning ... .... . .. ......... .. ...... . . ...................... 16
3.7. Warm-Up ........ . ......................................... 17
3.8.
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.9.
Summarized Proceedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Appendix A Sensitivity Corrections ... ... ....... ......... .......... .. .. 19
Appendix B
Noise , and its Heduction by the Use of Filters .. ........ . . . 24
4. Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

1. General
1.1. Purpose.
This instrument has two channels, one of them an interconnecting link
between an accelerometer and the Automatic Vibration Exciter Control
Type 1018/1038 or any suitable vibration meter, whilst the other channel
can be used for monitoring the signal from a second accelerometer.
It performs three functions, which are set out in sections 1.2 to 1.4: note that
the preamplifier itself does not provide velocity or displacement signals
since the latest philosophy is to incorporate the necessary integrators in the
main measuring instrument.
1.2. Impedance Transformation.
The preamplifier is placed as near as possible to the accelerometer in order
Lo isolate the transducer from the influence of the long cables often
necessary in practical arrangements . The accelerometer is then looking into
an extremely high impedance which will not normally load it, and at the
s ame time the low output impedance of the 2620 allows instruments and
comparatively long cables to be connected to it without affecting the
performance.
1.3. Gain Adjustment.
The Automatic Vibration Exciter Control is calibrated to give direct accelera-
tion readings with an accelerometer of sensitivity 10 mV/ g. Individual gain
settings on each of the two channels of the preamplifier enable accelero-
meters of any sensitivity between 10 mV /g and 100 mV /g to deliver a signal
to the measuring instruments which is accurately scaled to the standard
sensitivity value. By means of a switch, either of the two accelerometers in
use may be selected for the " main " measurement, whilst the signal from the
other accelerometer is monitored at a separate output.
1.4. Shaping.
High Frequency. Although vibration tests are seldom carried out at fre-
quencies above 10 kc/s and the mechanical resonance of accelerometers does
not usually appear until above this frequency, it is nevertheless undesirable
to allow the large gains associated with such a resonance. Harmonics of
the drive frequency can sometimes excite this, with the result that equip-
ment receiving the signal from the accelerometer may not behave correctly .
To eliminate these unwanted high frequency signals, the 2620 incorporates
low-pass filters cutting off above 10 kc/s.

db up from flat
response level
1- -
2 5 10 20 kC/S 50 100
162+3+
Fig. 1. The high frequency response of a typical accelerometer, showing tl1e
resonance peak.
Low Frequency. If the accelerometer is to sense low dynamic levels and
its output is to be integrated either once or twice, thereby providing velocity
or displacement readings respectively, the signal must be almost free of
low frequency noise because integrators emphasise low frequencies.
To enable the indicating instruments to give a true displacement value, three
alternative high-pass filters are built into the preamplifier , and the use of
these is discussed in section 2.7.

2. Description
2.1. Block Diagram.
Fig. 2 shows how each of the two accelerometer channels has its own input
cathode follower, but shares a common output stage. The high frequency
cut-off and sensitivity adjustments are separate for each channel , but only
the main output, which may be switched to either channel, benefits from
the low frequency cut.
The functions of each block will now be considered in turn:
Accelerometer
A
Monitor
Output
Fig. 2. Block diagram of the Accelerometer Preamplifier.
2.2. Input Circuits.
These are designed to present the highest possible impedance to the accelero -
meter and also to eliminate the effects of input cable capacitance. An input
impedance of more than 600 M.Q in parallel with 10 pF has been achieved
by careful choice of circuitry. The printed circuits used here are mad e
in an unconventional manner, out of a special base material printed with
contact gold. Silicone-based protective coatings prevent the flow of minutel y
small but troublesome cross-currents even under extremely humid con -
ditions.
The impedance of a crystal accelerometer can be represented by a capaci -
tance Ca pF, but normally it is not true to say that if a charge q appear s
on the faces of the piezoelectric crystal, there is a potential difference
~. at the end of the accelerometer cable. The capacity of the cable, which
is normally composed of a main cable plus a lead associated with the
accelerometer , reduces the sensitivity. The effect of the accelerometer 's
associated lead is easily allowed for, but the different loads of various main
cables and input circuits mean that it is usual practice to calibrate ever y
set up independently in order to take account of an uncertain sensitivit y.

One of the prime objects of the 2620 is to obviate this calibration and to this
end the effective capacitance of the main cable and preamplifier input has
been very much reduced.
This is achieved as shown in Fig. 12 by using twin-screened cable and con-
necting the inner screen to the input tube cathode so that its potential
follows that of the central conductor. This means that there is almost no
charge build up or decay in the grid c~.rcuits. The leads from the input
sockets to the tube grids are similarly constructed and the 10 pF input
capacitance mentioned earlier is an effective value, which rises to 25 pF
if the inner screen is disconnected from the cathode.
Accelerometer
Cable
I
I
I
I
I
I
I
==::=cc
I
I
I
I
I
I
I
Input Cicuit
162~29
Fig. 3. Equivalent circuit of a piezoelectric accelerometer connected to one
of the preamplifier's inputs.
As indicated in Fig. 3, the total capacitance in the input circuits, CT, is made
up from three capacities,
CA = accelerometer capacitance, including associated lead,
Cc = overall effective capacitance of cable,
and c, = input capacitance of preamplifier itself.
i.e. CT = CA + Cc + Ca.
CA is usually large compared with Cc and Ca for the 2620 and provided an
accuracy of ± 1 db is adequate, no sensitivity corrections are necessary.*)
The frequency response of the input circuits should also be considered. At
high frequencies the cathode follower performance falls off very slightly
as explained in section 2.8, but the only significant attenuation occurs at
very low frequencies where the accelerometer impedance becomes large
compared with the input impedance. The loading effect of the preamplifier
and cable will be
*) For very accurate work or when CA is small, or Cc becomes significant (e.g. if the inner
screen is disconnected or the cable is long), sensitivity deviations should be accounted for
as explained in Appendix A.
6

which for most practical cases reduces to
jwR1CA
1 + jwR1CA'
where w is the angular frequency.
The corner frequency (- 3 db point) at which the accelerometer / preamplifier
combination begins to exhibit a linear frequency characteristic is therefore
given by
1
fc = 2.nR,c ~ ·
Two common values for CA are 400 pF and 1ooo pF, so knowing this and
the appropriate R1, the low frequency response of the input circuits can be
calculated: Figure 4. Either accelerometer feeding into a 600 M.Q input
impedance gives a response which is perfectly flat down to 4 c/s, i.e. to
below the usual frequencies for measurement.
Attenuation
db
0
I-- I-
I--
I--
1--
2
I-- 1-- -1000 pF
Accelerometer =:
--
400pF Accelerometer ::
4
a ~
~~
b
6 All
asymptotic
. to lines of slope 20 dbfdecade I=~
80.1
0.2 0.5
2 5
10 20 C/S 50 100
162~1/f
Fig. 4.
a) The low frequency response of the 2620 input circuits connected to
accelerometers.
b) For comparison , the response for the same accelerometers when using
common types of preamplifiers having an input impedance of 200 M.Q.
2.3. Low Pass Filters.
The characteristics for these filters, one in each a channel, are shown in
Fig. 5.
They are passive L-C circuits which in no way influence the low
frequency response of the instrument, though in order to achieve the correct
shaping they introduce a little attenuation which is, however, corrected for
in the output amplifier. To eliminate pick-up of the power-line frequency
and its harmonics, certain inductances are split.
7

Attenuation
db
0
20
60
800.1 0.2 0.5
t- -
2 5 10 20 kC/S 50 100
f62it3~
Fig. 5. Typical amplitude-frequency response for the low-pass filters. The
broken line indicates the overall response of the preamplifier and an accelero-
meter with the characteristics shown in Fig. 1.
2.4. Attenuators.
Each of the two attenuators consists of a two-position coarse adjustment
switch and also a fine control so that the overall gains of the two channels
are separately and continuously variable over a range of 10 to 1, or 20 db.
2.5. Monitor Output.
Whichever signal is not selected for the main output appears here scaled
to a constant 5 mV/g provided the transducer sensitivities lie within the
prescribed range and the attenuators are correctly set.
The maximum output impedance is 10 kQ, but this value decreases if the
attenuator wiper is moved towards ground.
2.6. Output Amplifier.
This has a fairly high input impedance (400 kQ) which will not load the
attenuators. It is a single stage amplifier and the gain, tied down by
negative feedback, is set to give the correct level (10 mV/g) at the main
output.
2.7. High-Pass Filter.
As already pointed out in section 1.4, when reading small displacement
signals at low frequency, the acceleration signal sensed by the transducer
is minute, and so the noise voltages must be kept to an absolute minimum.
As explained in the appendix it is extremely important to set the high-pass
filters to as high a frequ e ncy setting as circumstances will allow. In this
way the maximum amount d noise is rejected.

"
..
•
The high - pass filter is an active circuit in which there are three different
pairs of condensers that can be switched in according to the cut-off fre-
quency required.
Attenuation
dB
0
5
10 "5 CIS"
15 =="10CIS"
0.2db
50 100 200 500 1000 CIS
Frequency
16
m'
Fig. 6. Typical amplitude-frequency response for the high-pass filter.
2.8. Main Output.
Assuming that the accelerometer sensitivity lies between 10 mV/g and
100 mV/g, and that the controls are correctly adjusted , the signal scaling at
this point is 10 mV / g.
The main output comes from a cathode follower with a low output im -
pedance equivalent to less than 400 .Q in series with 8 ftF.
The instruments fed by this preamplifier are likely to have an input im -
pedance of about 80 kQ so that at low frequencies the loading will be
negligible. At high frequencies the capacitance of the output cable is the
dominant factor, not only on account of a simple attenuation effect, but
also because the load line of the output tube becomes distorted when the
load is capacitive. This means that the total cable capacitance that can
be used depends on frequency, voltage level and acceptable distortion .
Fig. 7. indicates maximum values of capacitance for less than 3% distortion,
and it will be seen that for most practical applications a very long cable
can be used: 100 pF per metre is quite a high value.
Referring to section 2.2 and to Fig. 12 it will be clear that the capacitance
between the two screens of the input cables has a similar loading effect on
the input cathode followers , ( see Fig. 8), but this effect does not impose
serious limitations in practice.
9

Output
Output Volts R.M .S.
20
10
1\
~\
1\
~1\1
1\ '
\ 1\ 2 nF
1\
\
~rF
5nF
15
1\
l Set for 10 mt
9
J
'T
10
100
3 4 5 B 1k 3 4 5 B
10 kC/s
263371
Fig. 7. Ch art indi cating th e capaci tiv e lo ading on the main output whicl1
results
in 3% overa ll distortion. The output is assumed to be connected to an
Automatic Vibration Exciter Contro l (input impedance 80 kQ). Distortion is
not materially affected by the external capacitance betwe en the preamplifier
input feeder and ground, provided thi s capacity does not exceed 2ooo pF ,
which covers all practical cases.
Input
In put Volt s R.M .S
10 2 3 < 5 8 llO 2 3 < 5 8 1000 2 3 < 5 8 10kc,1
5
263372
Fig. 8. Typical response curves valid for various le ngth of input cable. (The
co rresponding ef fective capacities be tween th e screens are given in the table).
The response curves sh own are actua /111 (( limit " curves causing less than 3%
overall distortion and a reduction in sensitivity of o.3 dB. Measurements were
made with a (( pick-up " capacitance of 500 pF.
Note:
If it is desired to use sti ll longer cables the reduction in sensitivity due
to added cable capacitance can be calculated from the formula given in
Ap pe ndi x A, p. 22.
10
,

•
2.9. Cross- Talk.
The cross-talk between the two channels is somewhat dependent on fre-
quency, but is always more than 60 db down at frequencies above 10 c/s.
From 5 c/s to 10 c/s it is about 50 db down.
2.10. Phase-Frequency Response.
The accelerometer preamplifier has been designed with a view to achieving
certain attenuation characteristics and consequently exhibits phase advance
( dependent on filter setting) at low frequencies and phase lag at high fre-
quencies. It is not intended to be used for phase measurements.
There are no difficulties when making sweep sine tests in conjunction with
the Automatic Vibration Exciter Control, but if the accelerometers are
sensing complex signals (such as in shock tests), the various harmonics will
experience different phase changes. An R.M.S. indicating instrument measur-
ing the distorted signal will give the correct values, but an oscilloscope .
"average" reading or peak reading voltmeter will give a false impression.
2.11. Power Supplies.
These have been very carefully designed to avoid injecting low frequency
disturbances into the amplifier.
Most of the tubes are heated by stabilised DC in order to ensure a minimum
of power line hum pick-up and low frequency drift.
The H.T. voltage is accurately regulated by a circuit containing two vacuum
tubes and one gas-filled reference tube. A certain change in power line
voltage results in an H.T. level alteration of only about one twentieth of
this variation.
2.12. Noise.
Some practical hints for reducing noise are given in section 3.5.
Since it is too difficult to specify the noise voltage density at all frequencies,
for all working conditions and a direct wideband measurement of noise is
not particularly helpful, the Specification (Part 4.2) quotes the actual noise
level appearing on the associated measuring instrument in the worst case.
Much of the troublesome noise is picked up in the accelerometer cables.
Unfortunately any noise voltages generated between the central conductor
and the inner screen (Fig . 12) are amplified several times by the first tube,
and so under adverse conditions a higher effective cable capacity may have
to be accepted and the inner screen disconnected. As explained in 2.2,
individual calibration or correction rather than straightforward attenuator
settings must be relied upon for correct scaling in this case.
The appendix shows the importance of reducing low frequency noise. Some
of this, originating in the tubes, has been cut to a minimum in the 2620, but
also a substantial part comes from the accelerometers themselves. Accelero-
meter noise is largely caused by air passing the sensitive element, and can
be reduced by the use of wind shields. The higher the capacity of the
accelerometer, the more easily can the noise currents leak away to ground.
11

Channel
A
3. Operation
Charnel
8 .?6/!205
Fig. 9. Front view of tile Accelerometer Preamplifier.
3.1. Controls and Sockets.
INPUT A and INPUT B are Lemo OM 2 X o.9 sockets* ). Either one or two
accelerometers can be fed to these points via twin-screened cable
and a suitable plug. In each socket is a · pin protected by a piece
of Teflon: when the 2620 is live these must not be shorted to
ground · since they are at the potential of the input tube cathodes,
about 60 v.
Channel A pk.m Vl pk.g and RANGE mV/g control the attenuation in channel A
so that accelerometers with various sensitivities give rise to a
signal with a standard scaling in peak mV per peak g.
Channel B pk.mVlpk.g and RANGE mV /g control the attenuation in channel B
quite independently of channel A.
LOW FREQUENCY LIMIT switch is used to rounte either channel A or
channel B through to the main output and at the same time to
select Lhe desired low frequency limit for the main output signal.
*) The co rresponding plugs - B & K number JP 0201 - are coded F. OM2 X o.9 by their
make rs , Lcm o AG , Ro ute de La us a nn e. l\1orges, Switzerl a nd.
12
,.
•

•
0 0
Voltage
Selector
-and
Cartridge
Fuse
0
~...___ Monitor
o Output
~~-----------n~
2620
26220 6
Fig. 10. View of the rear of the instrument.
MAIN OUTPUT is a cannon XLR 3-14 socket delivering whichever signal ha s
been selected ( see above ).
MONITOR OUTPUT is a Cannon XLR 3-14 socket at which app e ar s a s ignal
from the channel not selected for the main output. The high -
pass filter is not in this circuit so the low frequency limit is
governed by the input circuit.
I
L _________ _
Preamplifier AVEC
Main Output Acceleration
Gen . Input
~- ~Connection to
screen optional,
depending
on
'------- _ grounds
Preamplifier Monitor
Monitor Output Instrument
Sockets (not plugs). viewed externally .
(AVEC= Automatic Vibration Exciter Control) f62Hf
Fig. 11. Connections to the sockets. (Viewed externally.)
3.2. Power-line Voltage.
The preamplifier may be operated at any of the mains frequencies 50,
60,
or 400 c/s, but the voltage selector must be correctly set to the supply
voltage. If it is not, unscrew and withdraw the cartridge fuse and then , with
the aid of a coin, turn the selector until the white bar is in the desired
position. Replace the fuse.
3.3. Accelerometers.
Fig. 12 shows how the accelerometers should normally be connected via
13

twin-screened cable. *) The cables should preferably be not more than 5 metres
long: Fig. 8 and section 2.8 explain the implications of having too muoh
capacity between the inner and outer screens.
0
Inner Scr!en
~ connected at
accelerometer end.
"'"" ., . Electrical
Insulation
~
Coble
Preamplifier
162SOf
Fig. 12. The input cables should be connected in this manner. Each twin-
screened cable is usually joined to the accelerometer's associated lead rather
than running directly to the pick-up.
It is usually advisable to retain the single screened lead supplied with each
accelerometer for the final connection to the pick-up itself. There are two
reasons for this, a) the twin screened cable is rather stiff and heavy for
some applications, and b) the accelerometer data normally refers to the
output from this lead.
Connect the accelerometer(s) lo Input A (and to Input B).
Select the two accelerometers to be used and determine for each its
sensitivity, a figure that is usually quoted for every individual accelerometer
by its manufacturer. Make sure thi ~ s value refers to the condition with the
associated lead which is normally in series with the twin-screened cable. If
an accuracy of 1 dB is adequate, the ef1fect of the twin- , screened cable and
input circuits on sensitivity is small enough to be neglected unless the
accelerometer capacity is below 400 pF. ** )
Switch the RANGE mV /g for each respective channel to the correct setting.
i.e. either 10-30 mV /g
or 30-100 m V/ g.
Set the sensitivity value on the corresponding scale knob which may be locked
in position by inserting an Allen key through the hole in the centre and
turning clockwise.
If the accelerometer sensitivity lies outside the range 10-100 mV/g, the
attenuator setting and vibration meter readings will have to be suitably
*) Note that the accelerometer should normally be isolated from ground.
**) Wherl the given sensitivity values do not correspond with the arrangement as used
(for instance if the accelerometer 's associated lead is discarded) , or if maximum
accuracy is wanted, consult Appendix A.
14

scaled up or down. Bear in mind that the accelerometer signal level should
be between o.1 mV and 20 volts peak, and if the signaVnoise ratio is to be
satisfactory, should represent more than o.1 g. Make sure that the accelero-
meter is shielded from draughts: wind noise can be a major problem.
3.4. Outputs.
The MONITOR OUTPUT may, if desired, be connected to an instrument such
us a vacuum tube voltmeter or oscilloscope which should have an inpl.lt
impedance of at least 1 M.Q if any quantitative measurements are to be made.
The MAIN OUTPUT is connected directly to the Acceleration Generator Input
on the back of the 1018 Vibration Exciter Control, or to any other suitable
vibration meter. See Fig. 11.
The length of cable which can be used depends upon:
The capacitance / unit length of the cable,
acceptable distortion.
voltage leveL
frequency ,
and terminating load.
Working up to 2ooo c/s at maximum voltage levels with 50 metres of 100 pF/
metre cable, the distortion will · be less than 3 %. For further information,
see Fig. 7.
Remember that the signal is still at a low level and hum and noise pick-up
must be avoided.
Choose which channel is to be fed to the main output and turn the LOW
FREQUENCY LIMIT swlitoh to select this channel with the highest lower
limiting frequency acceptable for the case in hand. The remaining channel is
automatically switched to the monitor output.
3.5. Hum and Noise Reduction.
Although "field" tests away from other electrical equipment and multitudinous
wiring are unlikely to raise problems in this respect, there are unfortunately
hum and noise fields, sometimes powerful, in almost every indoor test room.
Efforts must be made to cut down the effects of these, and if that is not
successful, to reduce the fields at source. The following remarks are included
as a guide to help those who are setting up a test system. Each problem
must be solved by the user employing trial-and-error methods and building
up experience.
Be careful to avoid earth loops, through which serious hum problems can
arise. Make the screen of the main output cable the only connection between
the chassis of the preamplifier and that of the main instrument. The P.A.
is usually grounded so if the monitoring instrument has one input terminal
grounded, the screen of the lead for the monitor signal should only be
earthed at one end. Some pick-up will in these circumstances appear on the
monitoring instrument, so if this is unacceptable the instrument must have
15

an input clear of ground, allowing a direct twin connection to the pre-
amplifier. The mains transformer in a power amplifier may have consider-
able capacitance between its windings, so if such an amplifier is used it is
best to ground it, making this the one and only earth point in the whole
system.
Automatic
Vibration
Exciter
Control
Accelerometer
Preamplifier
Control Signal (Main Output:A) '\
I(Monitor
:Output 8)
I
T
I
I
I
Accelerometers t
262191
Fig. 13. A vibration test system.
The outer screens of the accelerometer leads should not reach an earth at
bolh ends. Most acceleromete · rs have their casing as one connection, so they
must not be in electrical contact with the test object if it is a grounded metal
structure.
The case of the 2620 is not directly connected to its chassis, so it is often
possible to ground this case, thu s improving screening.
See that there is a minimum of wind noise from the accelerometer. An
efficient draught shield can be a great help even in an apparently still room .
Change to an accelerometer of larger capacity and charge sensitivity if
necessary and possible.
Provided a larger accelerometer cable effective capacitance is acceptable,
(sec tion 2.2), noise can be reduced by disconneting the inner screens of the
accelerometer cables. See Fig. 12 and also section 2.12. This will involve
setting corrections which are explained in Appendix A.
Extraneous noise can be particularly troublesome in test arrangements such
as that shown in Fig. 13 because the R.M.S. value of the control signal be-
comes larger than it should be, possibly with the result that the shaker table
stops vibrating. However, when the system is logically and tidily arranged,
bearing in mind the above points, there should be no difficulties.
3.6. Warning.
Particularly when the preamplifier is in the feedback loop of a system
controlling a vibration exciter, it is dangerous to unplug leads or switch
knobs .
16

3.7. Warm-Up.
When the vibration test system is correctly set up, connect the preamplifier
to the power line and switch on . Allow 5 minutes before taking any accurate
measurements.
3.8. Application.
The preamplifier has been designed for use in conjunction with the Auto-
matic Vibration Exciter Control Type 1019 or 1039 and reference should
be made to the appropriate instruction manual. However, some of the
advantages of the two-channel system will be briefly pointed out here.
Propeller Shaft
1 or 2 channel recorder,
Vibration
meter(s) or
Elect ron ic voltmeter(s)
161+17
Fig. 14. .4 typical field test arrangement.
The main channel will often be used in the feedback path when the vibration
exciter is being controlled in a closed loop system as in Figure 13. The
"main" accelerometer is sensing the behaviour of the shaker table, but the
second accelerometer can be used to monitor the vibrations of the test speci-
men itself. In this way much can be learnt, without the need for another
preamplifier, about the mechanical transfer function of the specimen and
about its various modes of vibration.
Motor
Type 4330/
Accelerometer
Accelerometer
Preamplifier
Fig. 15. Studying vibration insulation.
2622ff
17

Again, in measurements under actual operating conditions, the two channel
facility can be invaluable: vibration insulation tests are one example, and
the s tud y of propeller-shaft vibrations are another.
3.9. Summarized Procedure.
Sequence of Operations
I
References
Sections
Figures
Mount accelerometer ( s), electrically
isolated from
3.5
ground. Shield from wind.
2.12
Connect accelerometer(s) to input(s)
via twin
3.1 9
screened cable{s) in series with the accelerometers
12
associated leads.
2.2 8
Appx. A.
Set
accelerometer mV /g
with RANGE
mV /g and
3.1
pk.mV/pk.g for botlh inputs. Clamp dials.
3.3 9
2.4
Select
either Channel A or Channel B for MAIN 3.1 2
OUTPUT, using LOW FREQUENCY LIMIT switch . 3.4
Set LOW FREQUENCY
LIMIT
to
as high a fre-
3.1 6
quency is acceptable.
3.4 I
2.7
Appx. B.
Connect MONITOR OUTPUT
to
V.T.V.M., oscillo-
3.1 10
scope etc. if aes , ired. 5 m V ./ g. 3.4
11
2.5
Connect MAIN OUTPUT to ACCELERATION
GE- 3.1 10
NERA TOR INPUT on back of Automatic Vibration
3.4
11
Exciter Control.
2.8 7
Checking voltage selector panel, connect mains lead
3.2 10
and switch Power "On". Allow 5 minutes to warm
up.
Read vibration levels directly on
main measuring
instrument. Normally no calibration necessary.
If noise difficulties
arise, and when long or non-
2.2
standard
input
cables are used, take
special pre-
2.12
cautions.
3.5
18