Check video - shows the issue at 2:10
The spectrum goes crazy and after about 10 samples it goes back to normal.
Also happens in the QA40xplot software - so maybe some HW issue?
Happens twice in this video - once at the start and once at 1:13
It can due to a transient spike being picked up, spikes have energy across all frequencies.
Hi @kawal, yes, @MarkT has it right. You have 8 averages turned on, so when you have a big burst of energy, it takes 8 more acquisitions to clear. For those 8 averages after the burst, the one noisy acquisition dominates the math.
Note the reset button in the Acquisition section. That button will zero out the history. So, if you are changing cabling or have done something else to pollute the acq history, you can push the button and start fresh and then you don’t have to wait for the noisy acquisition to clear.
The issue is no spikes and no change in connection or device under test. Just happens.
Happens when some settings are changed in the SW like sampling rate or FFT depth
I’ve had similar experiences for the past year using QA403 and v1.204. It seems to involve the hardware getting out of sync with the settings in the software. I’ll be doing some measurements, get a high noise floor like you photographed, and disconnect all cables from the QA403 as a sanity check and see the harmonics still there. Only fix is close the software, unplug the USB from the PC, then unplug the USB-hub power.
The reason I’m inferring a HW issue is that I can normally audibly hear a small click when changing the attenuator from the first four choices to the second four choices (low to high full-scale input). However, when “malfunctioning” (with no cables connected to HW), I never hear the click, however the displayed dBV of the harmonic noise does increase/decrease.
Happens when some settings are changed in the SW like sampling rate or FFT depth
Yes, if you change FFT, for example, the current acquisition will stop, the averaging buffer will be cleared, and then acquisition will start again. This is because it doesn’t make much sense to average a bunch of mixed-length buffers together.
If you are experiencing a USB issue you could see the problem you descirbe. Can you set averaging to off and start acquiring with 16 or 32k buffer at 48ksps in loopback with a 0 dBV tone and tell me if you see any glitching? Just watch hundreds of frequency waveforms go by. Waveforms should be rock solid with noise wiggling around a bit. If you are seeing glitching there, then it’s likely a USB issue. Try a new cable, try another machine, etc. See if that changes anything.
2 Likes