[ntp:questions] NMEA driver using common PPSAPI code
hart at ntp.org
Fri Oct 1 06:55:11 UTC 2010
0.3 seconds (or 0.7 seconds) might be the offset between the
end-of-sentence timestamp and the PPS event. The NMEA driver in ntpd
now allows fudging the EOL and PPS timestamps separately, so you might
be tripping over the change in interpretation of fudge time1 and fudge
time2 compared to your older configuration.
You might also try disabling kernel discipline and/or kernel PPS to
see how that changes the behavior, too. I don't have either luxury on
Windows, where I use the NMEA driver, so I may have flubbed something.
More information about the questions