Circuit Analizer proposal

I have both PCSU1000 and PCGU1000 and I like Circuit Analizer feature of this setup.

But as I see on Bode Plotter, I can measure Vrms (and Phase) correctly in narrow 0…-25dB range. Of course, I can turn on ‘Automatic Voltage Scale’, but it doesn’t help if I measure milivolts, or have a noisy circuit.

But if we assume that circuit has linear response, we can filter out all frequencies except desired, and calculate it’s amplitude and phase much better.

To measure V and phase of desired frequency, I propose to use a part of ‘simple’ Fourier transform:

  1. apply one of ‘FFT window’ to input data, for exapmle, Hamming window-
    for(i=0, i<4096,++i) data[i] = (1.0-cos(i/2048.0pi));

  2. integrate data with sin(ik) and cos(ik) (where k is calculated from frequency and time/sample rate). As a result you get ‘x’ and ‘y’ amplitude values.

  3. Calculate phaseCH1 = atan2(x,y); VrmsCH1 = sqrt(xx + yy);

  4. If you need to show phase on plot, do the same calculations with second channel and get final Phase = phaseCH1 - phaseCH2 (rounded to 0…360 region)

I think this approach can give us extra 20…30dB of signal/noise resolution (0…-50dB range).

This approach is not valid for nonlinear circuits (such as frequency multipliers), so I think it can be an option in menu.

Nice that you like the Circuit Analyzer (Bode plotter) feature of the PCSU/PCGU1000.
Yes, you are right, using the FFT we may select only the frequency of interest.
As you mentioned, this is usable only if the output signal of the devise under test is sine wave.
The tracking of the FFT analyzer to the generator output frequency may be a problem…
This idea has to be considered anyhow (as an option to the future release maybe)…
Thank you for your comment.