[LinuxPPS] PC Clock Offset

[Andrew Hills]

Hi all,

I'm pretty sure I'm making a newbie mistake here, but it seems that my 
PC clock is, on the order of hundreds of milliseconds, ahead of the GPS 
clock, a Motorola Oncore M12+--usually about 370ms. I've attached the 
quick-and-dirty C program I used to interpret what I read from 
/dev/ttyS0 (note: it's really quick and REALLY dirty; don't trust its 
reliability). Short summary: for every @@Ha message it receives, it 
prints the GPS reported date/time and then the system date/time and the 
difference between the two (in nanoseconds). Playing with the #defines 
will control which portions of the messages are displayed, including a 
very verbose @@Bb interpretation. (I based it entirely on the M12+ 
documentation.)

Other information: I'm using NTP's Oncore driver to update the clock. It 
does show continuous reach to GPS_ONCORE in ntpq's peers when the 
antenna is in the right spot.

So, is this offset something done intentionally by NTP, an issue with my 
own program, a natural serial port/communications delay, or something 
else entirely? Even 100ms sounds much too high to me, especially because 
the "DELAY" setting for NTP's Oncore driver is on the order of nanoseconds.

--Andrew Hills
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: oncore_satellite_status.c
Url: http://ml.enneenne.com/pipermail/linuxpps/attachments/20090611/e46dd9cc/attachment.txt