[time-nuts] Influence of Cycle Wraps on TInt-Measurementswith53132A
Tom Van Baak
tvb at LeapSecond.com
Thu May 1 17:38:38 UTC 2014
> hi tom,
> thank you very much! that is quite interesting. i am happy to learn that
> there is nothing wrong with *my* counter! converting the non-linearity
> effect into a correction table is beyond my abilities, but simply
> knowing that this effect is inherent to the 53132a counter helps a lot.
One has to be very careful at this level of detail to make sure these are not external effects in the experimental setup. So I have not published my report yet. But your plots look similar enough to my initial plots that we may be seeing the same effect.
The SR620 counter has explicit interpolator calibration (and you can read/write the calibration bytes!) which I why I got the idea that maybe the 53132 would exhibit a similar pattern. If you read TIC literature, everyone that does interpolators faces the same issue of linearity, and most apply some level of correction, to gain extra performance.
So *if* what we're seeing is evidence of low-level 53132 non-linearity, it should be possible to retroactively apply the non-linearity as a correction.
> indeed, my plots look similar to yours. after only three hours of
> warming up i measured the TI of an HP 10811 against the 1 pps output of
> my fury. the 10 mhz output of the fury was used as the external timebase
> of the counter.
I didn't even worry abourt warm-up. The effect is immediate, consistent, permanent, yet quite small and within rms specifications.
>steps and drift removed (detail):
This simple phase residual plot already tells your there is strong periodicity. So you will then expect any ADEV, PSD, or auto-correlation plot to also expose the effect.
> but almost invisible in the "standard" ADEV plot:
This is correct. There is no reason to use standard ADEV anymore; it was necessary back in the slide rule and calculator era, but not today.
See also adev4 and adev5 under my www.leapsecond.com/tools/ directory. These should run under any OS.
More information about the Time-nuts_lists.febo.com