[time-nuts] OCXO phase noise

Magnus Danielson magnus at rubidium.dyndns.org
Mon Feb 13 09:32:13 UTC 2012


On 02/13/2012 10:16 AM, John Miles wrote:
>> That is exactly my thinking, but I was looking for a second opinion :)
>> We have anyway asked Abracon for pn data on the 40MHz version to be
>> sure.
>
> Most likely, the PN improvement will be somewhat better than the expected 12
> dB close to the carrier, and somewhat worse than expected (if you see any
> improvement at all) farther out.
>
> The corner frequency that separates "close in" and "farther out" will be
> device-specific,  but it'll probably happen between 100 Hz and 2 kHz.

A more proper analysis would separate the effects of oscillator Q, 
resonator loop noise and the buffer noise and then scale them accordingly.

Assuming the same Q (which isn't correct) the oscillator core 1/f^2 or 
1/f^3 flips up on. The buffer noise then masks on top of that.

Cheers,
Magnus



More information about the time-nuts mailing list