[ntp:questions] First attempt GPSD/PPS ->NTP time server

David J Taylor david-taylor at blueyonder.neither-this-bit.nor-this-bit.co.uk
Fri Jan 25 16:56:32 UTC 2008

Dennis Hilberg, Jr. wrote:
> I was actually referring to the time emitted in the NMEA data (GPS
> time), not the PPS signal.  Sorry, I should have specified that.  The
> PPS is accurate to a microsecond, but not the GPS time.  I happen to
> use a fudge factor of 0.190 to get the GPS time close to zero, but
> the PPS is of course very accurate and doesn't require any fudging.
> He's not using the GPS_NMEA (type 20) driver, so he has to deal with
> the PPS and GPS time seperately, hence SHM(0) and SHM(1).  The type
> 20 driver uses both the PPS (if the OS supports it) and GPS time, but
> the offset you see in 'ntpq -p' is the PPS offset.

Thanks for that clarification, Dennis.  I appreciate what you mean by the 
GPS time.


More information about the questions mailing list