1
$\begingroup$

I'm sampling 8 bioelectric signals with a embedded board which use an 8-channels ADC (AD7175-8). My sampling rate for every single channel is about 6250 Hz. When move the analysis in the frequency domain, what I obtain is the power spectral density as follow

enter image description here

instead of an (almost) flat one as I expect. It seems that there is a square-modulated interfering signal. I want to understand the nature of this signal and for this reason I'd like to know if you can tell me what kind of signal generate this particular power density spectrum.

| improve this question | |
$\endgroup$
  • $\begingroup$ If you hadn’t said it was a modulation I would have guessed that zeros were interleaved between your samples $\endgroup$ – user28715 Aug 27 '18 at 21:09
  • $\begingroup$ Looking at the result, I am suspicious. Can you post/point to the raw data and your code? So you have 8 channels? Are they independent or scanned through a multiplexer? Can you post the circuit diagram? The diagram looks like your actual sampling might be ~3kz. With reflection about 1.5khz, there are other things that look like that though. $\endgroup$ – rrogers Aug 28 '18 at 19:41
  • $\begingroup$ @rrogers The ADC has the inputs multiplexed: it scan all the channels at about 50kHz. The circuit is very simple I have only a passive RC low-pass filter stage before the input of ADC. Actually I'm working to analyze the raw data (I'm not the developer of the firmware/GUI). What should I find in the raw data linked to the spectrum above? $\endgroup$ – thoraz Aug 31 '18 at 10:49
  • $\begingroup$ To directly address your last question: I would do a cross-correlation channel to channel. Probably not a real mathematical one unless something showed up but a simple graph to look for "leakage" between adjacent channels. "simple RC" is not so simple in a multiplexed system; capacitor component choice is a "gotcha" sometimes together with the Thevenin equivalent impedance of your input. Been there, done that; so personally, I would check that out first. Of course, experience prejudices approaches. I would still like to see the raw data to check it against the chip specs. $\endgroup$ – rrogers Aug 31 '18 at 14:28
  • $\begingroup$ Also, just to be thorough, what internal filtering are you set to? In addition, it really looks like you are sampling at 3kHz???? What is the input signal? If "grounded" then is it grounded: internally (as in the data sheet part), at the pin, before the RC or before the device before the RC? $\endgroup$ – rrogers Aug 31 '18 at 14:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.