[time-nuts] frequency stabilty question

[Tom Van Baak]

> Hi
> 
> This brings in another subtle but significant issue.
> 
> We talk about the ADEV being done as the standard deviation of the frequency
> differences, but often that's not what's done. Even with zero dead time,
> there's another bit of magic in there. Drift is removed before the samples
> are used. 

Bob, look at the Allan deviation formula again. A constant phase
offset is effectively removed (because of the 1st difference); a
constant frequency offset is also removed (because of the 2nd
difference). But frequency drift is not removed with ADEV. Are
you thinking of Hadamard deviation instead? With HDEV linear
frequency drift is also removed (because of the 3rd difference).

> Oddly there are multiple approaches to drift removal. It comes as no
> surprise that the more aggressive the drift removal, the better looking the
> result. If you are looking at ADEV, it's always worth asking if (and how)
> the drift was removed. 

Agreed. A good example of this is the selection of an oscillator
for a GPSDO. Strictly speaking a drift-removed ADEV or HDEV
is a better metric to use when selecting an OCXO in this case.

/tvb