time-nuts@febo.com said: > That’s kind of why I’m going down the road of multiple samples - to see if > there’s anything to it. I would hack up some way to grab a clump (say 10) of samples and print them out where you can capture them on a PC and analyze them. I'd start by looking with the old Mark 1 eyeballs, then write hack code to filter out the good stuff so I can see the bad/interesting cases. If you have a scope, it might be interesting to trigger on PPS and look at the ADC trigger and input voltage. If you have a digital scope with the remember forever option, that would catch the delayed interrupt bug that Jim Harman reported. ------- The simple way to discard occasional bogus samples is to loop until you get 2 identical samples. -- These are my opinions. I hate spam.

I always try to calculate things like the standard deviation and peak-to-peak to get some idea if the measurement is valid. A DSO with infinite persistence or envelope mode is great for tracking this sort of thing down during development. Only toy DSOs will lack both. On Wed, 05 Oct 2016 22:24:09 -0700, you wrote: >time-nuts@febo.com said: >> That’s kind of why I’m going down the road of multiple samples - to see if >> there’s anything to it. > >I would hack up some way to grab a clump (say 10) of samples and print them >out where you can capture them on a PC and analyze them. > >I'd start by looking with the old Mark 1 eyeballs, then write hack code to >filter out the good stuff so I can see the bad/interesting cases. > >If you have a scope, it might be interesting to trigger on PPS and look at >the ADC trigger and input voltage. If you have a digital scope with the >remember forever option, that would catch the delayed interrupt bug that Jim >Harman reported.