[ntp:questions] Re: PPS ATOM clock set to falseticker

David L. Mills mills at udel.edu
Fri Aug 15 03:38:44 UTC 2003


Udo,

Please see Slide 9 of the architecture briefing at the NTP project page
linked from www.ntp.org. You will not like the answer. The Byzantine
selection algorithm insists on a majority clique, which means the clique
conataining truechimers must have more members than the clique
containing falsetickers. The correctness intervals of your GPS reciever
and the PPS have no points in common. The algorithm concludes one of
them at least must be a falseticker. This is what happens when you shave
the microseconds unless you are really careful.

The problem is you have to shave down the apparent offset of the GPS
receiver so its interval overlaps the PPS interval. Sneakiest way to do
that is using the calibrate command described on the reference clock
drivers page.

Dave



More information about the questions mailing list