[time-nuts] cheap 5V OCXO in 14DIP has about 1E-9 drift per day

Magnus Danielson magnus at rubidium.dyndns.org
Mon Apr 18 05:22:34 UTC 2011


On 04/18/2011 01:06 AM, Chris Albertson wrote:
> I asked the question because I might want to build one of there but I
> could not see how they coud be stable with just an RC filter.  So the
> answer is "they just are".
>
> OK, so they are stable but why would it run at 10+16=26 and not 10.01+15.99=29

As I look around, the non-linearity provides a sample/injection lock 
effect which creates the asynchronous mode. I haven't read the papers 
through, not the time, but that's the concept I gather from the quick 
browsing thru.

> And what happens if I want to build a 10.01 reference clock?  Could I
> do that with a regenerative filter?

The capture range would be narrow while the 10.00 mode would capture it 
as it would have a wider capture range.

As far as I have gathered the knowledge.

Cheers,
Magnus



More information about the time-nuts mailing list