[LinuxPPS] Offset problem with NTP and linuxpps

Jon Schewe jon.schewe at honeywell.com
Wed Jan 24 20:30:14 CET 2007


I've got NTP patched for NMEA/PPS and the kernel(2.6.18) patched for
PPS.  I've got /proc/pps/sources:
driver:~ # cat /proc/pps/sources
id	mode	echo	name			path
----	------	----	----------------	----------------
00	1133	no	pps_8250_0		/dev/ttyS0
01	1133	no	pps_8250_1		/dev/ttyS1
02	1133	no	pps_8250_2		/dev/ttyS2
03	1133	no	pps_8250_3		/dev/ttyS3

I've got a good sync according to NTP:
driver:~ # /usr/local/ntp/bin/ntpq -p
     remote           refid      st t when poll reach   delay   offset
jitter
==============================================================================
*GPS_NMEA(0)     .GPS.            0 l    8   16  377    0.000    3.905
2.678

I've got a serial board that brings the PPS signal in on the same serial
port as the NMEA messages.

Now I've noticed some issues with time on this machine compared with an
embedded device syncing off the same GPS, there's a 350-400ms difference
in time.  I wrote an application that toggles a ping on my parallel port
once a second at realtime priority to avoid scheduling issues.  I then
hooked a scope up to that pin and the PPS signal from the GPS and notice
that the period of the two signals is right at 1 second (the standard
deviation is in the nanosecond range).  However the leading edge of the
signals is off by 361ms with a standard deviation of 1ms.  Does anyone
have any idea where this delay is coming from and why NTP isn't syncing
the machine to the PPS signal exactly?

Thanks.


________________________________________________________________________
http://lug.htc.honeywell.com/people/jschewe [Honeywell Intranet Only]
*My views may not represent those of my employers




More information about the LinuxPPS mailing list