Table of Contents
Advertisement
2.8.1 Source Resistance
Increasing capacitance beyond this level may increase noise
and induce
instrument
instability.
The noise gain of the
As shown in Table 2-2, a minimum value of source resistance
measurement circuit can be found from:
is recommended for each range. The reason for this can be
Equation 2.
understood by examining Figure 2-3. Considering effects on
low frequency noise and drift, CS and CFB can momentarily
Output en = Input en x(1
+
be ignored.
where ZF =
Input amplifier noise and drift appearing at the output can be
calculated as follows:
Equation 1.
Output
enOiSe
=
IflpUt
enOjSe
x(1+
-@)
Thus it is clear than as long as RS>>RFB,
Output enoNe fz
Input enoNe. When RFB = RS, Output enojse = 2 x input
enoise,
and ZS =
RFB
\j
RS
42a
fRSCSP+
1 j
The same applies for eos.
Clearly ss f -0
equation (2) reduces to equation (I).
The frequency range in interest is O.lHz to IOHz which is the
, , cFB
noise bandwidth
of the AID convener.
The value of CF is
CS
I I
22OpF for nA ranges and O.OlpF otherwise.
II
-
RFB
In general, as CS becomes larger, the noise gain becomes
RS
Slrm
larger.
An
application
where
CS may be greater than
WVV
ANALOG
10,OOOpF is leakage measurement of capacitors. In this case
T--
OUTPUT
Input en must include the effects of the voltage source (ES)
lOon
used to bias the capacitor. The Keithley Model 230 is recom-
(1 Es
mended for this application.
.
When measuring leakage currents on capacitors larger than
6
TO A/D
10,OOOpF. stability and noise performance can be maintained
CONVERTER
by adding a resistor in series with the capacitor under test.
v
The value of this resistor shoulde be around 1 MB. For large
capacitor
values (z lrFj,
the value of the series limiting
Figure 2-3. Simplified
Model for Input Signal
resistor can be made lower in order to improve settling times:
Conditioning
however, values below 1OkD are not generally recommended.
Model 485 will typically
show insignificant
degradation
in
displayed performance with the noise gain of 2 resulting from
allowing RS = RFB. Typical amplifier input enOjSe is about
5@ p-p in a bandwidth
of 0.1.IOHz. Amplifier
ECS can be
nulled
with
front
panel ZERO adjustment,
but available
resolution
limits
this
adjustment
to
about
5V.
The
temperature
coefficient
of
ECS is
< 2OpV/"C.
These
numbers can be used with
Equation 111 to determine
ex-
pected displayed
noise/drift
given any source resistance.
Remember that 1 displayed count
= lOO$/ except on 2nA
range where 1 displayed count = lO@V. Note also that the
values given in Table 2-2 for minimum source resistance also
represent the value of RFB on that range.
2.8.2 Source Capacitance
The Model 485 is designed to accomodate up to 10,OOOpF in-
put capacitance
ICS). This limit will preclude problems in
most test setups and allow extremely long input cable lengths
without
inducing instabiliw
or oscillations.
This resistor is not critical in terms of tolerance or stability.
Any carbon composition
resistor will prove adequate.
A second-order advantage to using this limiting resistor is sd-
ded protection to Model 485 in the event of capacitor failure.
2.8.3 Leakage Resistance
The effect of leakage currents should be considered when
making small current
measurements
with high impedance
sources. Leakage current and its effects can be minimized by
using high resistance insulation in the test circuits and guerd-
ing. Since the Model 485 is a feedback picoammeter,
it is ef-
fectively guarded at the input. The effect of leakage paths on
the measurement can be further minimized by using a gusrd-
ed text fixture as shown in Figure 2-4. In the configuration
shown, the current through
the component
under test will
not be shunted significantly
by either leakage resistance path.
If a high voltage supply is used to make high resistance
leakage measurements,
it is suggested that a series current
2-7
Advertisement
Table of Contents
