the reading. Pick up from AC fields can also be detected by
observing the electrometer output on an oscilloscope. Line
frequency signals on the output are an indication that elec-
trostatic interference is present.
Means of minimizing electrostatic interference include:
Shielding. Possibilities include: a shielded room, a shielded
booth, shielding the sensitive circuit, and using shielded
cable. The shield should always be connected to a solid
connector that is connected to signal low. If circuit low is
floated above ground, observe safety precautions when
touching the shield. Meshed screen or loosely braided
cable could be inadequate for high impedances, or in
strong fields. The Keithley Model 6104 Test Shield can
provide shielding under many circumstances. Note, how-
ever, that shielding can increase capacitance in the measur-
ing circuit, possibly slowing down response time.
Reduction of electrostatic fields. Moving power lines or
other sources away from the experiment reduces the
amount of electrostatic interference seen in the measure-
ment.
2.14.3 Thermal EMFs
Thermal EMFs are small electric potentials generated by dif-
ferences in temperature at the junction of two dissimilar
metals. Low thermal connections should be used whenever
thermal EMFs are known to be a problem. Crimped or cad-
mium soldered copper to copper connections are methods
that can be used to minimize these effects.
2.14.4 RFI
Radio Frequency Interference (RFI) is a general term frequent-
ly used to describe electromagnetic interference over a wide
range of frequencies across the spectrum. RF1 can be especial-
ly troublesome at low signal levels, but it may also affect
higher level measurements in extreme cases.
RF1 can be caused by steady-state sources such as TV or radio
broadcast signals, or it can result from impulse sources, as in
the case of arcing in high voltage environments. In either
case, the effect on instrument performance can be consider-
able, if enough of the unwanted signal is present. The effects
of RF1 can often be seen as an unusually large offset, or, in the
case of impulse sources, sudden, erratic variations in the
displayed reading.
RF1 can be minimized by taking one or more of several
precautions when operating the Model 617 in such en-
vironments. The most obvious method is to keep the instru-
ment and experiment as far away from the RF1 source as
possible. Shielding the instrument, experiment, and test leads
will often reduce RF1 to an acceptable level. In extreme cases,
a specially constructed screen room may be necessary to suffi-
ciently attenuate the troublesome signal.
If all else fails, external filtering of the input signal path may
be required. In some &es, a simple one-pole filter may be
sufficient. In more difficult situations, multiple-pole notch or
band-stop filters, tuned to the offending frequency range,
may be required. Keep in mind, however, that such filtering
may have other detrimental effects (such as increased
response time) on the measurement.
2.14.5 Leakage Resistance Effects
At normal resistance levels, the effects of leakage resistance
are seldom seen because any leakage resistance present is
generally much higher than the resistance levels encountered
in the circuit under test. At the high resistance levels of many
Model 617 measurements, however, leakage resistance can
have a detrimental effect on the measurement. Such leakage
resistance can occur in the-circuit under test (on PC boards,
for example), in the connecting cable, or even at the elec-
trometer input itelf, especially if the input connector is not
kept clean.
To see how leakage resistance can affect measurement ac-
curacy, let us review the equivalent circuit in Figure 2-27. Es
and Rs are the source voltage and source resistance respec-
tively. The leakage resistance is represented by RL, while the
voltage, as seen by the electrometer, is VM.
Rs and RL form a voltage divider that attenuates the input
signal in accordance with the formula:
ML
viq =-
Rs + RL
Thus, if RL has a value of lCOGQ and Rs is lOGQ, the actual
voltage measured by the electrometer with a 1OV source
would be:
10 x 1WGSl
VM =
1OGQ + 1WGQ
vp”j = 9.09v
Thus, we see that the effects of leakage resistance can be
substantial, resulting in an error of almost 10% in this case.
2-32
To Next Page
Loading...
Need help?
General
