• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 82 subscribers
  • Views 548 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

piccolo ADC issues

Max Sonnaillon
Intellectual 600 points
Other Parts Discussed in Thread: TMS320F28062

Hello,

I am working in a switching power supply, which is a noisy environment, controlled by a Piccolo 28035. The ADC has a very strange behavior when I change the ACQPS. Bigger ACQPS values don't reduce the noise in the conversions but they introduce DC effects or they even increase the noise level. I found some posts where other users mention this but nobody has responded.

I try with my switching converter turned off, my ADC measurements are very clean (+/- 1 bit of noise) but when I change the ACQPS value I see effects in the DC value of the measurement. Once in a while ONE of the channels (not all of them) goes to a strange state where is has a little offset error that stays until I reset the DSP (restarting the firmware does not eliminate it).

Is there any recommendation on the "optimum" ACQPS value? is there any way of avoiding this problem? (I am already discarding the first acquisition that seems to be a little bit off sometimes)

Thanks,

Max

  • 0
    Intellectual 280 points
    Max,
    Did you ever get clarification on this issue? I'm using a TMS320F28062 and having an issue that sounds similar to what you've described (now several years ago). I have been discarding the fist sample as required by the silicon errata. However, what I see occasionally is apparent corruption of the second sample (i.e. the first sample after he discarded one). The corruption appears to occur in the form of a change in the noise distribution. When things are working fine, I get a nice gaussian-looking distribution with a very tight standard deviation. But when things aren't working properly, the noise becomes bi-modal. That is, I now get a distribution with two gaussian bells! What makes tracking this down frustrating is that I cannot willfully make the problem come or go. It appears to come and go with re-compiles of my code. But once it's there (or not there), the problem stays (or stays away) until some random future re-compile. I have had numerous suspects for the cause... hoping your experience may lead me to think of something I haven't already. (I've been doing countless experiments to nail this down.)

    Sincerely,
    Bob W.