[time-nuts] Follow-up on Z3801A high EFC reading
Magnus Danielson
magnus at rubidium.dyndns.org
Sat Jul 16 21:28:58 UTC 2011
Hi John and Graham,
On 16/07/11 22:50, Graham / KE9H wrote:
> John:
>
> If it was turned off for more than a year, in my experience, the HP10811
> (or similar)
> oscillators at the heart of the Z3801A will take about three weeks to
> settle down,
> and rejoin the previous mature aging curve.
>
> I have built a few GPSDOs using HP10811 oscillators, that I bought off
> of eBay,
> and they all took three weeks to settle down.
>
> You can ask the experts on the list why. Something about temperature
> related stress
> relief and gas sublimation on the surface of the crystal.
I was just about to make this point. While I am not an expert in crystal
oscillators (I think I can name at least 4-5 people here with their
hands dirty from it) from what I have read up and from what I have seen
the type of re-trace properties you describe is typical.
It is to avoid this re-trace that instruments keep their oscillators
oven heated in stand-by mode. If you look for it, you will find the
effect well covered in literature. There are several mechanisms in play,
but polution depositing on the crystal surface is one of them that I've
heard of. Mechanical stress as the temperature change causes re-shaping
of all mechanical aspects and the forces it puts on the crystal blank
until realigned again is another.
So, that was in the back of my mind when I said that you should wait.
For a crystal of 10811 class weeks is to be expected before saying it is
bad or badly out of tune.
Anyway, good to hear about the progress. The EFC track-in cruve looks
like expected from this type of event.
Notice that you might need to use a curve-fit cancellation matching the
drift curve. ADEV is sensitive to drift and isn't particularly gifted in
illustrating the drift either.
I could dig up a few references for you if you badly need it.
Cheers,
Magnus
More information about the Time-nuts_lists.febo.com
mailing list