Basically, with the Kalman method, you end up in tuning your mathematical model coefficients to follow exactly the OCXO; then you can do whatever you want: drive it to have a fast lock-in, drive it for extremely low drift, apply a forward correction to extend the holdover...
On Sun, Apr 8, 2012 at 7:27 AM, Charles P. Steinmetz < [email protected]> wrote: > Ulrich wrote: > > The Smartclock in difference seems to be able to adapt regulation >> parameters to its "measurements" of ocxo stability and long term drift. >> > > I do not know any details about Smartclock, but I believe one of the > things it does is to adjust the oscillator disciplining parameters > (primarily the loop time constant, I suspect, but perhaps more) to allow > 3801s to reacquire lock quickly after a holdover event (or at startup), > then switch to parameters that reduce the deviation once it is firmly > locked. You have to do that manually with a Tbolt if you want both quick > lock and best stability once locked. > > It may also do as you suggest -- use heuristic methods to tweak the "low > deviation" discipline parameters -- but I do not know if that is the case > or, if so, if it is better than Kalman filtering. > > Best regards, > > Charles > > > > > > > > _______________________________________________ > time-nuts mailing list -- [email protected] > To unsubscribe, go to > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there. > _______________________________________________ time-nuts mailing list -- [email protected] To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts and follow the instructions there.
