Another useful tool of the SBench 6-Professional software is data interpolation. This powerful feature can be used to dramatically improve parameter measurements in situations where the sampling rate of a digitizer is limiting measurement precision. Based on the SinX/x (or sometimes called SinX) algorithm the SBench 6 interpolation function inserts samples in a predictive manner between the actual acquired data points. When done correctly, this produces a waveform with a higher effective sampling rate and a closer approximation to the analogue waveform that’s being analysed.
The benefit of interpolation becomes apparent as the input signal frequency begins to approach the Nyquist limit (half the sampling rate). Consider figure 3. Here a 20 MHz sine wave with 1.82 V amplitude is acquired by a digitizer sampling at 125 MS/s and with 16-bit resolution. The yellow trace on the left shows the signal without interpolation and with the sample points connected by linear lines. The blue trace on the right is the same signal but with interpolation activated. Both traces have been expanded vertically so you can examine the top of the sine wave.
Note that the 20 MHz signal is still well below the Nyquist limit of 62.5 MHz. However, the yellow trace (without interpolation) already has insufficient sampling rate to precisely measure the signals amplitude. The problem is that at 125 MS/s each cycle of the sine wave only contains around six sample points. Activating the interpolation function effectively improves the sampling rate. In this example an interpolation factor of 10 was selected, creating a trace with 60 samples per cycle. Parameter measurements for the Peak to Peak amplitude (lower left corner) made for both traces show the significant improvement in amplitude accuracy that’s possible.
Providing the frequency content of the input signal doesn’t exceed the Nyquist limit, interpolation can be also be used to improve other parameter measurements (such as rise and fall times, as well as cyclic measurements like frequency and period).