Pcsu1000 question

Instruments PC Oscilloscopes and Recorders
1

Hi all,

I haven’t worked with an oscilloscope for decades, but I have pretty much decided on the 1000. What I would like to do initially is look at grid noise using magnetic sensors that will put out in the mV range. I would like to get down as low as possible in a 200Khz BW and no more than 1MHz BW. Since 8 bits is getting marginal, I was hoping for some averaging gain from the unit. If I reduce the sample rate by a factor of 4, can I assume that the 1000 will average those 4 samples giving me an SNR boost of 6 dB (adding another effective bit)? I also need accurate freq domain from the analyzer.

If 8 bits is deficient for these low level signals, I would have to go to 12 bits at higher cost (Pico 4224).

Thanks much,
Larry

Edit: It would also be great if the software could gate/window the spectrum and IFFT back to the time domain maintaining reference to the input waveform - but I assume that this (currently) would be an offline user-supplied function.

2

To reduce the FFT noise level you can use the “Vector Average” option of the PCSU1000.
For more info please see also these threads:
viewtopic.php?f=10&t=7460
viewtopic.php?f=10&t=4728
viewtopic.php?f=10&t=1670

[quote]It would also be great if the software could gate/window the spectrum and IFFT back to the time domain maintaining reference to the input waveform - but I assume that this (currently) would be an offline user-supplied function.[/quote]You are right. There is no such inverse FFT function in the PCSU1000.

3

[quote=“VEL255”]To reduce the FFT noise level you can use the “Vector Average” option of the PCSU1000.
For more info please see also these threads:
viewtopic.php?f=10&t=7460
viewtopic.php?f=10&t=4728
viewtopic.php?f=10&t=1670

Thanks VEL255 - I checked out the threads, now maybe you can tell me if I understand (excuse my ignorance):
For my signal, think 60Hz, 10mV ptp, with switching noise (harmonics) on top of the 60 Hz sine.

  1. I assume the front end amps will auto range the signal for optimum A/D
  2. The O-display is limited to ± 50 mV so will there be a prob viewing the time signal?
  3. If I select a much slower sample rate, I get no benefit from time averaging/decimation.
  4. Benefit is only achieved in the freq domain by averaging O-snapshots in trigger mode.
  5. The trigger set will jump around on the waveform due to the noise on the 60 Hz
  6. This will broaden the fundamental at best (low noise) and decorrelate the snapshots at worst (high noise) in the F domain

Does that sum it up correctly?

Thanks much!

4

Yes, the Vector Average needs rather good triggering.
The fundamental wave should be high enough.
If there is a lot of noise and the fundamental wave is low, then there is also some triggering jitter.
This will increase the FFT spectrum noise level.