[time-nuts] Allan deviation plots etc

[Magnus Danielson]

John Green wrote:
> Ulrich Bangert wrote:
> my point has not been to judge who has to be interested in what. My point
> has only been to use the right terms for it. Let me give an example: John
> uses the term
> 
>> ...but I am more concerned with ... and stability over time
> 
> Well, stability over time is what exacly is displayed in a tau-sigma-diagram
> of an oscillator. Since only a few words before he is saying that he is NOT
> intersted into Allan Deviation plots, then he is perhaps interested into
> something else?

Ehm... the Allan Deviation plots show the phase noise aspect of 
oscillator behaviour in terms of frequency stability, as this aspect 
proved harder to characterize than others. For long term aspects, linear 
terms like drift and drift processes plays a larger role and thus will 
not be included in the Allan Deviation plots.

> Allan deviation plots are interesting and if I had the means to generate
> them, I certainly would. Since I don't, I find it easier to interest myself
> in ways to try to measure oscillator performance in  practical terms with
> the equipment at hand. What I would really like is to be able to produce a
> specification in the same form as the oscillator manufacturers. I am used to
> data in that form and while the Allan deviation gives somewhat the same
> data, I rarely see one that gives a clue to stability over a year or 10
> years. Most of the ones I see seem more concerned with seconds to hours and
> maybe weeks at most. Of course, my method won't give any clue as to how an
> oscillator will perform over years either.
> I hope this clears things up a bit.

Allan Deviation plots themselfs isn't particularly hard to produce (a 
fairly simple math exercise), but the raw-data is a bit hard to acheive 
to for quiet sources.

Cheers,
Magnus