Hi A lot depends on just how long your “frequency” samples are. If the GPS is good to a nanosecond, you are at 1x10^-11 at 100 seconds. It’s a rare crystal based GPSDO that will hold closer than that frequency wise.
The FLL is non peaking so that’s going to help things a bit as well. Bob > On Jun 23, 2019, at 12:47 PM, Dana Whitlow <[email protected]> wrote: > > Correction from Dana: > > I meant "... without accumulation of phase error during normal times ...". > > Sorry about that. > > Dana > > > On Sun, Jun 23, 2019 at 11:45 AM Dana Whitlow <[email protected]> wrote: > >> Leo, >> >> Are you saying that you want to abandon phase lock altogether in favor of >> freq >> lock? Or just during the reacquisition following loss of and restoration >> of the >> reference? >> >> By me definition of pure freq lock, there will generally be some permanent >> (but varying) >> frequency error, so that phase error could accumulate without limit; >> clearly an undesirable >> thing in most applications. >> >> My interest lies in having a stable LO for receiving, without accumulating >> phase error (at least >> during times of missing reference). When the reference goes away, I'll >> accept some phase >> error accumulation. So for me, I think the best approach is phase lock >> under normal >> circumstances, but switch to freq lock during reacquisition of phase lock. >> >> Dana K8YUM >> >> On Sun, Jun 23, 2019 at 9:01 AM Leo Bodnar <[email protected]> wrote: >> >>> I have to draw your attention to practical aspects of why some designs >>> use FLL rather than PLL. >>> >>> Consider a GPS locked OCXO outputting GPS synced 10MHz signal. >>> >>> Properly designed control loop will not produce much (if any) difference >>> when the reference (GPS signal) is present. In the end, integral of zero >>> is zero. >>> >>> When reference (GPS lock) is lost the things are very similar too, >>> holdover is just flying blind in the rough direction you were facing last. >>> Accumulating frequency and phase offset on the way. >>> >>> However, when reference is restored the things are much different. >>> After regaining the reference (which in case of GPS signal has >>> unambiguous absolute time embedded into its phase) *proper* PLL loop will >>> try to correct for slipped phase at the highest slew rate. This can be >>> huge. If phase has drifted 1ms apart the loop will have to slew the phase >>> all the way until it gets those 10,000 cycles out of the way. This usually >>> looks ugly in frequency domain and is very disrupting if you are using the >>> device as frequency reference rather than an absolute time reference. >>> >>> Proper FLL loop will just gently (and reasonably quickly) get your >>> frequency back and forget about all the lost phase. Which is what a lot of >>> users want. >>> >>> Initially, I have used PLL mode on GPS clocks that I am making, but >>> switched over to FLL during the last few years. >>> >>> Cheers >>> Leo >>> _______________________________________________ >>> 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. _______________________________________________ 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.
