Keithley 427 User manual

INSTRUCTION MANUAL
MODEL 427
CURRENTAMPLIFIER
0 1975, KEITHLEY INSTRUMENTS, INC.
CLEVELAND, OHIO, U.S.A.
DOCUMENTNUMBER29103

CONTENTS MODEL 427
CONTENTS
1. GF,Ng3,&.L DESCRIPT~ON-------------------------------.-------- 1
2. OPERATION-------------------------------------------------- 4
3. APPLICATIONS----------------------------------------------- 9
4. ACCESSORIES------------------------------------------------ 10
5. ClRC"IT DESCRIPTION---------------------------------------- 12
6. REp~Cp‘&LE PARTS------------------------------------------ 17
7. CALIBRATION------------------------------------------------ 26
SCNEMATICS---------------------------------------------------- 33
1074

MODEL 427 ILLUSTRATIONS
ILLUSTRATIONS
Figure No. Title Page
1 Front panel. __-___---___-_____-___________________ 1
2 Front pane1 Controls. ________--______-____________ 3
3 Rear Panel Controls. - 3
4 shunt Method Measurement, -----------------__------ 4
5 Effect of Shunt capacitance. ---------------------- 4
6 Compensation for Shunt Capacitance. --------------- 5
6b Extended Frequency Response. ---------------------- 5
7 Fee&a& Method. _--__-_-__________-_______________ 6
8 Input Voltage N&se. -------------___-------------- 6
9 Bandwidrh of Feedback System. --------------------- 7
9b Effect of filter on Noise spectrum. -------------- 7
9, Effect of Input Capacitance on Noise. ------------- 7
10 Frequency Compensafian. ---------_____--_-__------- 7
11 Plot of Noise-Improvement Contours. __- _____ -__-__ 9
12 Block Diagram of a High-Speed Current Amplifier, -- 12
13 Filter circuit. ------_______--_-__-_______________ 13
14 power s”pp~y Regulator. --------------------------- 13
15 current suppression. ------------__---------------- 13
lb COmpOnent Layout - PC-291. ------------------------ 13
17 component Layout - pc-290, -----------------___---- 14
18 component Layout - pc-2?39. ---------_------___-____ 15
19 COmpOnenf Layout - P&292. ---------------__------ 15
20 chassis - Top vie”. ------------___--------------- 16
21 Chassis Assembly - Exploded “few. ----------------- 19
22 Bottom CO”er Assembly. -------------------------- 19
23 Measurement of Input Voltage Drop. ---------------- 27
24 Measurement of Rise Time. -------_--_____-___-_____ 28
25 Measurement of Filter Rise Time. ------------------ 29
1074 iii

SPECIFICATIONS MODEL 427
SPECIFICATIONS
RANGE: 10’ to 10” volts/ampere in eight decade ranges.
(10-13 ampere resolution to 10.” ampere full output).
OUTPUT: *10 volts at up to 3 milliamperes.
OUTPUT RESISTANCE: Less than 10 ohms dc to 30 kHz.
OUTPUT ACCURACY: 12% of reading to the lo9 vattsl
ampere range, +4% of reading on the 1O’O and 10”
volts/ampere ranges exclusive of noise. drift and current
offset.
RISE TIME (10% to 90%): Adjustable in lx and 3.3x steps
from “Fast Rise Time” listed below to 330 rnsec.
NOISE VS. RISE TIME’:
I FASTRISETlML
STABILITY: Current offset doubles per 10°C above 25°C.
Voltage drift is less than 0.005% per “C and less than
0.005% ,,er da” of full outDut after l.hour warmur,.
,-
OFFSET CURRENT: Less than lQLz am&e at 25”C’and up
to 7001, relative humidity.
CURRENT SUPPRESSION: lo-lo ampere to 10-a ampere in
eight decade ranges with 0.1% resolution (lO.turn poten.
tiometer). Stability is +O.Zo/. of suppressed value per “C
bO.Z% per day.
INPUT VOLTAGE DROP: Less than 400 /.IV for fullaxle
output on the 10” to 10” volts/ampere ranges when
properly zeroed.
EFFECTIVE INPUT RESISTANCE: Less than 15 ohms on the
10” and 105 volts/ampere ranges. increasing to less than
4 megohms on the 10” volts/ampere range.
MAXIMUM INPUT OVERLOAD: Transient: 1000 volts on any
range for up to 3 seconds using a Keithley (or other 10
mA-limited) highaoltags supply. Continuous: 500 volts
on the 10” to 1O’voltsjampere ranges, decreasing to ‘ZOO
an the 10”. 70 on the lo5 and 20 volts on the 10’ volts,
ampere ranges.
OVERLOAD INDICATION: Lamp indicates pre-filter or post.
filter overload.
DYNAMIC RESERVE: 10 (20 dB).
CONNECTORS: Input: (Front) ENC. Output: (Front and
Rear) BNC.
POWER: 90.125 or 180.250volts (switch selected), 50.60 Hr.
5 watts.
DIMENSIONS; WEIGHT: Style M 3%” half.rack, overall bench
size4’~highx8%‘widex12L/a”deep(100x217 x310 mm).
Net weight, 7 Ibs. (3.0 kg).
i” 1074

MODEL 427 GENERAL DESCRIPTION
SECTION 1. GENERAL DESCRIPTION
l-1. GENERAL; The Model 427 Current Amplifier is a C. Built-in Current Suppression. Small changes in
high-speed, feedback-type amplifier with particular the signal level can be measured since large ambient
features useful for automated semiconductor testing, current levels can be easily suppressed.
mass spectrometry, and gas chromatography applications. d. Overload Indication. Accurate measurements are
assured since overloads are automatically indicated.
l-2. FEATURES. e. Variable Cain. The GAIN Switch is designated
a. Wide Dynamic Range. Selectable rise times permit in eight gain positions from lo4 to 1011 volts per
low-noise operation important when resolving small ampere - therefore gain adjustment is straight forward.
current levels. f. Variable Kise Time. Optimum response can be
b. High Speed. Typical resolution is 20 picoamperes
out of a 10sSZpere signal with a 100 microsecond selected for each gain setting since a separate RISE
TIME switch is provided on the front panel.
rise time.
0471 1

GENERAL DESCRIPTION MODEL 427
TABLE 1-l.
Front Panel Controls and Terminals
Functional Description Paragraph
PUSH Power Switch (S302)
GAIN Switch (5201)
RISE TIME (5101)
SUPPRESSION
MAX AMPERES Switch (S303)
FINE Control (R333)
POLARITY Switch (5304)
ZERO ADJ Control (R235)
INPUT Receptacle (5202)
OUTPUT Receptacle (5102)
~ OVERLOAD Indicator (DS302)
Controls power to instrument.
sets gain in Volts per ampere.
Sets rise time Fn milliseconds.
Sets maximum suppression.
Adjusts suppression.
Sets polarity of suppression.
Adjusts output zero.
Input source connection.
Output connectFon.
Indicates overload condition.
2-4, al
2-4, a2
2-4, a3
2-4, a4
2-4, a5
2-4, a6
2-4, a7
2-3, a
2-3, b
2-5, d
TABLE 1-2.
Rear Panel Controls and Terminals
Control or Terminal Functional Description Paragraph
Line Switch (5301)
PowerReceptacle (P305)
Fuse (F301)
Sets instrument for 117V or 23411.
Connection to line power.
Type 3AG Slow-Blow, 117V @ l/4 A (w-17)
234V @ l/S A (w-20)
2-4, b
2-3, c
0lJTPuTReceptacle (5103) Output connection. 2-3, b
WARNING
Using a Line Power Cord other than the one supplied
with your instrument may result in an electrical
shock hazard. If the Line Power cord is lost or
damaged, replace only with Keithley Part No. CO-7.
0878

MODEL 427 GENERAL DESCRIPTION
GAIN
SWITC
S20
,----SUPPRESSION-,
FINE POLARITY
ADJUST SWITCH
INPUT
5202
. P
ZERO ADJ OVERIDAD Power OUTFW
R235 DS302 S302 5102
FIG"RF 2. want Panel Controls.
I
FIGURE 3. Rear Panel Controls.
0471

OPERATION
SECTION 2. OPERATION
MODEL 427
2-1.
MEASUREMENT CONSIDERATIONS.
a. Current-Detection Devices. The DleasUrement of
‘small electrical c~rrent8 has been the basis for a
number of instrumental methods used by the analyst.
Ion chambers, high-impedance electrodes, many forms
of ch=“metog=aphic detectors, phototubes and multipli-
ers are coaronly-used t=ansduce=a which eequire the
measurement of small currents. Devices used for this
measurement a=e often called electraaeters.
b.. In any measure-
ment, if e”“=ce noise greatly exceeds that added by
the inst=“mentatian, optimization of instrumenteti””
is unimportant. When source noise approaches theoret-
ical minimum, optimization of instrumentation charac-
teristics becomes imperative. TO determine the cate-
gory into which this meas”=ement falls. the researcher
needs t” be familiar with the characteristics which
impoee theoretical and practical limitations on his
me*surement . Most researchers a=e familiar with the
theoretical limitations present in voltage measueements
The noise inceeases with 8”“=ce realstance, and the
familiar equation for the mean-square noise voltage is
q = 4kTRAf Eq. 1
whe=e k is the Boltzma”” Constant, T ia the absolute
temperature of the s”“=ce resistance R, and Af is the
noise bendwidth( 3 times the 3 dB bandwidth for a
single RC rolloff.) In the case of cureent measure-
ments it Is more appropriate to consider the noise
current generated by the ~l”“=ce and load resistances.
The mean squaee noise c”==ent generated by a resistor
is given by Eq. 2.
FIGURE 4.
In the shunt method c”=re”t is measured by
the voltage drop ac=“ee a resistor.
Prom this equation it is irmnediately apparent that the
m%QB”rement of emall current =equi=es large values of
R, i.e., high impedance levels. Howwee, thfe gives
difficulties for meas”=ements requiring wide bandwidths
because of the RC time constant associated with a
high-megohm resistor and even a few picofarads of cir-
cuit capacitance.
Figure 4 shows a c”==ent s”“=ce
generating a voltage across a parallel RC. The fre-
quency response of this current measurement is limited
by the RC time constant.
Figure 5 shows this response
end the -3 dB p”int “CC”=B at a frequency
Lor* noise and high speed, therefore, a=e contradictory
requirements. TO optimize a current-measuring system,
techniques must be used which obtain high speed “sing
high-impedance devices.
C. Hiah Speed Methods.
1. High epeed can, af c”“=se, be obtained in .
shunt-type meae”=eme”t by “sing a low value for the
shunt resist”=. As pointed ““t above, such a srmll
resistor value int=ad”ces excessive noise into the
meQ8”rement.
2. A second method to achieve bandwidth is to
keep R large, to accept the frequency roll-off
starting at F”, and t” change the frequency eesponae
of the voltage amplifier a8 ehown in Figure 6a. The
combined effects of the RC time c”“sta”c folloved
by this amplifier is shown in Pigure 6b and it is
seen that the frequency response of the c”==ent
measurement has been extended to Pl. The frequency
at which the amplifier gain sta=‘ts to increase must
be exactly equal t” the frequency F” determined by
the RC time c”nstant in order for this approach t”
result in a flat frequency respanse. Therefore,
FO
LOG FREQUENCY
FIGURE 5. The frequency respanse of the shunt method
1s limited by omnipresent ahunt capacitance.
I 0471

MODEL 427 OPERATION
this oethad is ueeful only far application* where
the shunt capacitance C is constant. Aaide from
thin drawback this is * 1eSitimste approach which
is being wed in low-noise, high-speed current-
meesuring applicatians. In addition to current noise
in the *hunt and in the amplifier input stage, B
maJor source of noise in this system *ri*** from
the voltage-noise generator ssaociated with thb in-
put atage (reflected a* current noise in the shunt
resistor) caused by the high-frequency peeking in
the following stages of amplification. More will be
said about this in the discussion on noise behavior.
3. A third method used for speeding up * current
measurement asas guarding techniques to eliminate
the effects of capacit*nces. Unfortunately only
certain type* of capacitance*, such ** cable cap=-
itances, can be conveniently eliminated in this
manner. To eli,r,inate the effect of parasitic c*p*c-
itences associated with the *ource itself became*
very cumbersome and m*y not be feasible in many in-
stsnces. The major *ourc** of noise in this *“*tern
*re identical to those mentioned in the second
system.
4. A fourth circuit configuration combines the
capability of low-noise and high-speed performance
with tolerance for varying input C and eliminate*
need for separate guard by making the ground plane
*n effective guard. This is the current-feedback
technique. This technique gives * typical improve-
ment of 3 over shunt technique*. Again, the major
sources of noise are identical to those mentioned
in the second system.
d. Noise in Current Measurements. Noise forms *
b*aic limitation in *nv hinh-speed current-measurinn
system. The shunt *y&m give; the simplest curren;
measurement but does not give low-noise performance.
A properly designed feedback *y*tem gives superior
noise - bandwidth performance. Noise in these two
systems will be discussed next.
1. Noise Behavior of the Shunt System. High
speed end low noise *r* contradictory requirements
in any current meesurement because *orw capacitance
is always present. The theoretical performance
limftetion of the shunt *yetem c*n be calculated **
FO F> LOG FREWENCY
FIGURE 6. ny tailoring the frequency response of the
amplifier (Pig. 6a) the frequency response
of the shunt method c*n be extended.
The rms thermal noise current (in) generated by *
resistance R is given by
Eq. 4
The equivalent noise bandwidth (.f) of * parallel SC
combination is Af = 1/(4RC) snd the eignal hand-
width (3 dB bandwidth) F, = 1/(2nRC). For practical
purposes peak-to-peak noise is taken 88 5 times the
ml* value. The peak-to-peak noise current can now
be written a*
i
UPP = 2 x 10-9 F, F Eq. 5
In practice, e typical value for shunt cape.cit*nce
is 100 picofarads. With this value the following
rule-of-thumb is obtained. The lowest ratio of
detectable current divided by signal bandwidth using
*hunt-techniques is 2-10-14 ampere/Hertz for B peak-
to-peak signal-to-noise ratio equal to 1. A coroll-
ary far this rule-of-thumb expresses the noise cur-
rent in term* of obtainable risetime (lo-SO% rise-
time tr = 2.2 RC). The lowest product of detectable
current and risetime using shunt technique* is 7 x
lo-l5 ampere seconds. In this derivation it has been
assumed that the voltage amplifier does not contri-
bute noise to the measurement.
2. Noise Behavior of the Feedback System. There
are three *ource* of noise in the feedback system
that have to be looked at closely. The firat two,
input-stage shot noise and current noise from the
mea*urinS resistor, are rather straight-forward. The
third, voltage noise from the input device of the
amplifier, cau*e* *ome peculiar difficulties in the
measurement. Any resistor connected to the input
injects white current noise (Eq. 4). In the circuit
of Figure 7 the only resistor that is connected to
the input is the feedback resistor R. As in the
shunt system, R must be made large for lowest noi*e.
Beceuse this noise is white, the total contribution
can be calculated by equ.,ting Af to the equivalent
noise bandwidth of the system. The second *ource of
noise is the current noise from the amplifier input.
This component is essentially the shot noise asaoci-
ared with the gate leakage current (io) of the input
device. Its rms value equals . . .
FO F, LOG FREQUENCY
FIGURE 6b. Extended frequency response.
0471

OPERATION MODEL 427
T;; = J-zTp-
where e is the electronic charge. The contribution
of this noise generator is also white. N*t only do
these two noise sources generate white current noise,
the noise in a given bandwidth is also independent
of the input capecitence C. The mejor source of
noise in e feedback current meesurement is the noise
contribution aseocisted with the voltage noise of
the input amplifier. The voltage noise ten be rep-
resented by a VOltage noise generator (0,) et the
emplifier input es shown in Figure 8. This wise
generator itself is assumed to be white. However,
its total noise contribution to the current-measuring
system is not white. Inspection of Figure 6 will
reveal that et low frequencies P large em*u*t of feed-
beck ie applied around the voltage noise source {en).
However, the SC combination ettenuetes the high-
frequency components of V,,t so that no feedback is
present et high frequencies. Thus, the noise con-
tribution to the output voltage V,,t from the valt-
age noise source a* is no longer independent of
frequency. The noise is “colored” and increases in
intensity for ell frequencies higher than F,. The
resulting noise spectrum is shown in Figure 9b. The
tote1 system noise is related to the are* under
this curve. Because the logarithm of frequency is
plotted on the horizontal axis, the eree under the
curve et higher frequencies represents e signifi-
cantly larger amount of noise then e similar eree
*t low frequencies. For comparison, Figure 9a show
the frequency response of the current measuring
system. Figure 9e ia identical to Figure 6b. It is
interesting et this point to compare this noise
spectrum with the frequency response of the voltage
amplifier in Figure 4 es shown in Figure 6a. A volt-
age noise eouec.e et the input of the amplifier would
generate a noise spectrum according to the amplifier
frequency response as shown in Figure 6a. The noise
spectrum of such e system, then, is identical to the
noise spectrum of the feedback system as given in
Figure 9h. This illustrates the well-known fact
that signal-to-noise performance of a measurement
cen**t be improved by feedback techniques. At the
high-frequency end the voltege noise is limited by
the frequency FA which is the high-frequency roll-
off point of the operational amplifier. It should
FIGURE 7. Beslc circuit configuration for the feed-
back method.
he noted that even though the useful bandwidth of
the system extends only t* Fl, there era noise com-
ponents of higher frequency present. To obtain
best widebend-noise performance, these high-frequency
noise components have to be removed. This ca* be
achieved by adding a low-pass filter section follow-
ing the feedback input stage. If the band-pass of
thin low-pass filter is made adjuetable this filter
can nerve the dual purpose of removing high-frequency
noise end of limiting the signal bandwidth of the
system.
2-2. THEORY OF OPERATION.
8. Current Feedback Technique. The basic circuit
configuration used in the current-feedback technique
is shown in Figure 7. In this configuration the
current-measuring resistor R is placed in the feedback
loop of e* inverting emplifier with a gain of A*. The
frequency response obtained with this circuit is iden-
tical to thet s+nvn in Figure 6b. F* agein is the
frequency associated with the RC time constant:
F, = SE Eq. 6
The frequency response of the syetem is extended t* a
frequency fl where
F , = AoF, Eq. 7
Note that the frequency rerponse is automatically flat
without heving to match break points. However, the
total bandwidth of the system (Fl) is still limited
by the value of the ahunt capacitance C across the
input. This improved frequency response of the feed-
back technique avoids the use of low values for R
which could generate exceesive current noise.
.
b. Refinements of the Feedback System. A major
difficulty of the feedback system ariees from shunt
capacitence esaociated with the high-megohm resiaeor R
in the feedheck path. If the shunt cepacitence acroes
the resistor is CFr then the bandwidth (FF) of the
system is determined by the time COnstent RCF:
FIGURE 8. The voltage noise associeted with the am-
plifier input device is en important eourc~
of noise in the high-speed feedback syatew
6 0471