On Sun, 14 Jan 2007, David Woolley wrote: > In article <[EMAIL PROTECTED]>, > Timo Felbinger <[EMAIL PROTECTED]> wrote: > > > Or just use tickadj, which can adjust the frequency in steps of 100ppm > > and thus can bring your clock into +/-50ppm of the correct frequency; > > ntpd is then well able to cope with the remaining error. > > The reason I didn't suggest this is that I don't believe that there isn't > any reasonable way in which a working crystal oscillator could be out > by 500ppm, which implies that either the oscillator isn't being > disciplined by the crystal, or that something else is causing the > frequency to be misinterpreted.
It seems that such lousy crystals do exist: I have several (on various types of PC mainboards) which need ticks from 9992 to 10003. Without tickadj, ntpd will have a very hard time to stabilize or will not converge at all. After setting the optimal "tick", there are no more problems. The oscillator frequency seems to be pretty stable over time, just different from what it is supposed to be (in particular, I never had to change an "optimal" tick value once I had determined it). Timo _______________________________________________ questions mailing list [email protected] https://lists.ntp.isc.org/mailman/listinfo/questions
