Hi All,

I would like to hear about your opinion on the most important parameters
which would need to be "tinkered" on ntpd (v4.2.2) to discipline a high
quality ovenised crystal (instead of a standard PC crystal).

As the stability of the oscillator is increased of several order of
magnitude, the maxpoll can be increase quite dramatically. I am testing
with 2^14, which seems a good compromised between longer averaging
period (i.e. better frequency estimation) and fast update.

An ovenised Crystal may have an Allan Intercept down to few ppb's or
less (>0.001ppm). So the allan intercept parameters should be increased
(the FLL should kick-in with longer averaging period). I am testing with
a value of 5000, I am least sure about this than the others parameters.

Another parameter is the dispersion value (I am using 15e-9, but it
could be better estimated). 

Any other parameter would you suggest to change?

Mauro

-- 
Mauro Fiacco 

ip.access Ltd                      


_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to