[time-nuts] Measuring short term stability minus linear drift

[Magnus Danielson]

On 08/10/11 06:32, Rick Karlquist wrote:
> I want to measure the short term stability of a source
> with substantial linear drift.  I would like some measure
> of stability along the lines of Allan deviation, but I
> only want to measure the "noise" and ignore the "drift".
> AFAIK, ADEV treats linear drift like a form of noise.
> Has this problem been solved before?

Yes.

> Any ideas?

Oh yes.

Record your time-stamps (say using TimeLab). Use HDEV plot as a first 
approximation, as it will remove first degree linear drift.

If this drift does not meet your needs (for high taus higher degrees of 
drift leaks through even HDEV), then on a copy of the raw data use a 
high degree curve-matching (again TimeLab can do this for you) and 
remove the matched trend. Use ADEV on the resulting value. Use ADEV and 
MADEV to separate between WPM and FPM levels.

Look at the Allan Variance article on Wikipedia, and you will find more 
about it there, but also references to online manuals such as NIST 
Special Publication 1067 for instance.

Cheers,
Magnus