Section 6: Optimizing SMU measurements
Model 4200A-SCS Source-Measure Unit (SMU) User's Manual
6-10
4200A-SMU-900-01 Rev. A December 2020
Cable capacitance
Without guarding, the effects of cable capacitance would adversely affect the settling time when
sourcing current. The rise time of the source depends on the total shunt capacitance at its output. For
a high-impedance load, even a small amount of cable capacitance can result in long rise times. For
example, cable capacitance of 100 pF and a load resistance of 1 GΩ results in an R
C
time constant of
approximately 100 ms. Guarding drastically reduces cable capacitance, resulting in much faster rise
times. With FORCE and GUARD at virtually the same potential, the cable capacitance cannot charge,
and rise time is not affected (refer to
(on page 5-1)).
When sourcing voltage, the rise time due to cable capacitance is usually insignificant. Because the
voltage source is low impedance (<1 Ω ), the R
C
time constant of 10
−10
seconds 1 Ω × 100 pF is
negligible.
Test system performance
When making a semiconductor I-V measurement, there is always a compromise between speed and
noise. Even with a fixed set of measurement settings, changes to the system configuration (such as
cable length or adding a switch matrix) change the measurement results. The 4200A-SCS has
settings to allow optimal I-V measurements. The settings fast, normal, and quiet are fixed. You can
also set a custom value.
To achieve a low-noise measurement, the quiet setting is recommended. The compromise is that
measurement speed is slower in comparison to the fast and normal settings. To make a fast
measurement, the fast setting can be selected, though the noise will be higher. Typically, the normal
setting is used to balance the speed and low-noise requirements. To further fine-tune the
measurement, the custom setting can be used.
The fixed settings are tuned to the 4200A-SCS for standard cable lengths connected to the DUT. In
general, this should be sufficient to make good measurements. However, when extra-long cables or a
switch matrix are used in the system, these settings may not be adequate. A typical phenomenon is
the appearance of a glitch or offset error. The magnitude of the error increases if the fast setting is
used to make the measurement. This is caused by insufficient settling time for the system. With
added load or capacitance (cables or matrix relays), it takes longer for transient effects to settle. If you
only use the measurement parameters optimized for short cables, you may generate erroneous
measurements.
The best way to minimize this effect is to allow extra settling time. The normal or quiet settings should
improve the measurement result. You can also use custom settings to fine-tune the measurement
settings. It may be a trial and error process. Various combinations of parameters can be used to
achieve the best results. In general, longer cables or slower settling of switch relays require a larger
delay factor.
