Hi What are we trying to do here?
What is the input frequency? What is the phase noise of the OCXO? What is the phase noise requirement on the output? Are there other system based requirements on the output? (ADEV maybe ….) ===== Symmetry (by it’s self) going from 50/50 to 45/55 (or whatever) has zero impact on the phase noise of the output. If your logic implementation that *uses* the clock depends on 50/50 duty cycle, that is a very different “specific to your design” constraint. Bob > On Jun 19, 2020, at 3:06 PM, Gilles Clement <[email protected]> wrote: > > Current PICDIV concept generates exactly 50% output duty cycles, so only even > dividing factors are possible. > If we accept a slightly unbalanced duty cycles (ex: one more instruction on > the down side loop than for the up side loop), one could generalize to odd > factors. > (An idea suggested by Didier Juges) > What would be the down sides for a Clock signal ? > Gilles. > >> Le 18 juin 2020 à 13:58, Gilles Clement <[email protected]> a écrit : >> >> Hi >> I need to divide the output of an OCXO by a factor D=81 for testing >> purposes. So with minimum added phase noise. >> PICDIV-like approches would not work (D needs to be divisible by 8 or at >> least be even) >> I went through the archives and it seems that an Injection Locked Frequency >> Divider with resynchronization flip-flop could be a simple and acceptable >> solution. >> As described in the following Wenzel paper: Unusual Frequency >> Dividerswww.wenzel.com › uploads › dividers >> <https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwik49qGpIvqAhURahQKHTBVClAQFjABegQIARAB&url=http%3A%2F%2Fwww.wenzel.com%2Fwp-content%2Fuploads%2Fdividers.pdf&usg=AOvVaw2m-9lURROiSbG9XykiDNDU> >> Does this make sense? >> Gilles. > > _______________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to > http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com > and follow the instructions there. _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com and follow the instructions there.
