[time-nuts] Crystal aging

Richard (Rick) Karlquist richard at karlquist.com
Thu Sep 3 15:46:43 UTC 2009

> Maybe! But for the sake of clarity: I have been asking how HP/Symmetricom
> manage to handle the regulation loop appearantly WITHOUT time constant
> adaption/switching and with a fixed time constant that seems to be much too
> high to start with for a cold OCXO.  
> Best regards
> Ulrich Bangert

One answer might be the use of a second integrator.  The HP5061 used
an analog second integrator, which was quite non trivial to implement
at the time.  The 5071 also uses a second integrator, which is of
course digital.  The oven control loop of the E1938A uses first and
second integrators (PID + I^2) to get the best of both worlds.
None of these use switched time constants.

Rick Karlquist N6RK

More information about the time-nuts mailing list