Odd ADC issue

Hi,

I can't completely follow what is going on but I think I get it. It does sound like you maybe see aliasing. The sampling process doesn't care if you sample at too low of a rate for the waveform at the input or skip samples coming out of a fast ADC, the impact can be similar. If the processes aren't completely synchronous then it might be hard to track down. If you were to downsample by filtering and decimating, then you can pull off skipping samples deliberately.

If the signals are truly square and you see a lot of samples at "in-between" values, that is a little suspicious.

I think if you set the sampling rates (and anti-aliasing filters) appropriate for the waveforms and always sample/process contiguous sample segments, you should be OK. As I said, I'm not completely sure I understand the assumptions.

If the scenarios (waveforms and schedule of how samples are drawn) are easy to define, mocking up the sampling system in Matlab/Octave or your tool of choice might be a good approach if you think there might be more to the problem.

Lance