David Woolley wrote: > Heiko Gerstung wrote: > >> In the NTPv4 draft you will find a (similar) definition: Root dispersion >> indicates the maximum error, that does not necessarily mean that this is >> the current error. > > And maximum error means theoretical worst case, not that this value has > actually been reached at any time.
I think a basic question is: If the epoche (i.e. the correct time in this case) is set by some means (IRIG in this case) and time is continuously disciplined via a good PPS signal, should the returned reference timestamp continue be updated, and the root dispersion kept low in this case? If for some reason the absolute time on the server is not correct then an increasing root dispersion would not help anyway ... On the other hand, if the operators have the chance to keep the epoche synchronized via IRIG, why don't they just leave the IRIG signal connected? Martin -- Martin Burnicki Meinberg Funkuhren Bad Pyrmont Germany _______________________________________________ questions mailing list [email protected] https://lists.ntp.org/mailman/listinfo/questions
