[time-nuts] Z3805A cooling requirements?

Magnus Danielson magnus at rubidium.dyndns.org
Thu Dec 20 22:24:27 UTC 2012


On 12/20/2012 01:36 PM, Bob Camp wrote:
> Hi
> Did I miss part of the data?
> The plot I got shows about 9 minutes. Temperature step stress takes a *long* time to settle out.

I too felt the data was short. For relative mediocre oscillators I at 
least measure an hour. For quality rocks, that's just a "get a feel" 

Looking at the frequency plot, you can almost see the frequency going 
down, where as for most part the frequency is rising, and the ADEV plot 
shows a tau linear rise as you would expect form such a linear drift. 
However, as you would measure longer the OCXO oven loop would hunt in 
(unless you run outside of the control range) and frequency would return 
towards normal. The ADEV plot would settle back. ADEV isn't a good 
measure for such transient behaviours, which really isn't noise but 
systematic effects.


More information about the time-nuts mailing list