Advertisement
05 | Keysight | Low Current Semiconductor Measurements Using the B2980A Series Ammeter - Technical Overview
3. Measurement speed, iltering/averaging and zero cancelation
functions improve measurement data quality
Null (Zero cancel) capability:
To remove the effects of any residual offset currents, you can have the instrument
automatically subtract out these currents by pressing the Null key.
Figure 3 shows a measurement example using the B2980A Series ammeter. In this
example the B2985A is in its 2 pA measurement range, the measurement speed is set
to "Quick", the iltering is set to a 100 point moving average and stray leakage currents
have been removed using the NULL function. Using the B2985A's Roll View the data is
displayed over a 100 second time span. With these settings it is easy to see that a stable
100 aA level measurement can be made.
Zero cancel Filtering
The Roll View graph shows the measurement is
stable at the 100 aA level within a 100 sec interval.
Quick measurement speed
Figure 3. Low current measurement example (B2985A).
4. Histogram View enables real-time data distribution analysis
In order to reduce or eliminate measurement noise it is helpful to be able to analyze
measurement data statistically in real-time. Once you know the characteristics of the
noise you are seeing you can decide if it is better to reduce the noise at its source or to
ilter out the noise from the data. Of course, if the noise is symmetrically distributed then
you can simply use the average of the histogram.
Figure 4 shows an example of this procedure using the B2980A's Histogram View. The
two histograms show the profound effect that measurement speed and data iltering can
have when making measurements in a noisy test environment.
Note: This measurement was made using a banana clip lead to triaxial cable with the input terminal
open and without shielding; therefore, there is a considerable amount of noise entering the ammeter.
Figure 4(a) shows Histogram View results using 0.1 power line cycle (PLC) measurement
speed and no iltering. The histogram distribution exhibits peaks at both ends and the
measured sigma is 0.17 pA. Figure 4(b) shows the same measurement performed using
Normal speed (which does PLC averaging) and a 100 point moving average ilter on the
measurement data. In this case the measured sigma is 0.018 pA, which is about 9.5 times
(= 0.17/0.018) smaller than the sigma shown in Figure 4(a).
(continued)
Measurement condition:
- Speed: Quick
- 2 pA range
- Filter: 100 moving average
- Null: zero cancel
- DUT: Open condition
Advertisement
