[LinuxPPS] Need help with setting up LinuxPPS/NTP server
Hal V. Engel
hvengel at astound.net
Wed Oct 8 21:17:21 CEST 2008
On Wednesday 08 October 2008 10:57:11 Richard Abbott wrote:
> Hal,
>
> Thanks for the quick response.
>
> I did a search on my system and the ppsldisc and ppstest commands were not
> compiled with the rest of the sources. So to answer your first question, no
> I wasn't running ppsldisc.
>
> I found their sources in the Documentation/pps/ directories of both the
> ntp and kernel sources. With some finagling I was able to get both ppsldisc
> and ppstest to compile from the kernel sources area. After running
> ./ppsldisc /dev/ttyS0 here is a sample output from ppstest /dev/pps0
>
> trying PPS source "/dev/pps0"
> found PPS source "/dev/pps0"
> ok, found 1 source(s), now start fetching data...
> source 0 - assert 1223487531.497205940, sequence: 10262 - clear
> 0.000000000, sequence: 0 source 0 - assert 1223487532.497206563, sequence:
> 10263 - clear 0.000000000, sequence: 0 source 0 - assert
> 1223487533.497206496, sequence: 10264 - clear 0.000000000, sequence: 0
> source 0 - assert 1223487534.497211739, sequence: 10265 - clear
> 0.000000000, sequence: 0 source 0 - assert 1223487535.497206593, sequence:
> 10266 - clear 0.000000000, sequence: 0 source 0 - assert
> 1223487536.497211575, sequence: 10267 - clear 0.000000000, sequence: 0
> .......
This looks like ktimer output. Notice how there are no clear events. For a
real refclock there will be both assert and clear events and both will have
sequence numbers and time stamps that increment over time. This output always
has the clear event as sequence 0 and time stamp 0 which is not correct.
>
> Besides the sound device I do not see any modules related to timers
> loaded. This however does not mean that I didn't build it directly into the
> kernel by mistake. Do you know what lines in the .config file should be
> disabled?
I should have said ktimer. Sorry if this caused confusion. This is located
in the PPS section of the kernel config and is named CONFIG_PPS_CLIENT_KTIMER.
The ktimer module is used to generate fake pps events for testing without
needing a real refclock. If the ktimer module is loaded you will have two
pps devices listed if you do tree /sys/class/pps at least of you also have
ppsldisc running. If you have two pps entries located in /sys./class/pps then
you have the ktimer module loaded. There is also the possibility that you
might be using the ktimer generated events rather than the events from the
real pps source and this is why you need to confirm that it is not loaded.
>
> After getting ppsldisc running I restarted the ntp server. Here is the
> output I got from ntpq soon after restarting ntp
>
> > ntpq -p
>
> remote refid st t when poll reach delay offset
> jitter
> ===========================================================================
>=== *GPS_NMEA(0) .GPS. 0 l 8 16 377 0.000 483.336
> 85.501 ntp2.your.org 216.218.254.202 2 u 11 64 17 30.433
> 1243.75 0.371 time.nist.gov .ACTS. 1 u 5 64 7
> 41.591 1235.97 0.489 mirror 210.138.109.164 3 u 8 64
> 17 27.716 1240.88 1.032 LOCAL(0) .LOCL. 10 l 48
> 64 7 0.000 0.000 0.001
>
>
> Then after it finished syncing up with the remote timeservers it went back
> to how it looked before.
>
> > ntpq -p
>
> remote refid st t when poll reach delay offset
> jitter
> ===========================================================================
>=== xGPS_NMEA(0) .GPS. 0 l 8 16 377 0.000 637.591
> 739.522 +ntp2.your.org 216.218.254.202 2 u 46 64 377 30.562
> 1245.10 1.322 *time.nist.gov .ACTS. 1 u 41 64 177
> 41.275 1237.31 1.642 +mirror 210.138.109.164 3 u 39 64
> 377 27.608 1241.07 0.739 LOCAL(0) .LOCL. 10 l 14
> 64 377 0.000 0.000 0.001
>
>
> --
> Richard Abbott <rabbott at ticom-geo.com>
> TICOM Geomatics, Inc.
> 9130 Jollyville Rd, Ste 300
> Austin, TX 78759
> Office : (512)610-8340
> Fax : (512)345-9992
> Cell : (512)422-2352
>
>
> Hal V. Engel wrote:
> On Wednesday 08 October 2008 09:52:54 Richard Abbott wrote:
>
> Hello all. I am having a bit of trouble getting a system configured to
> pull time and PPS from my GPS unit. I think I am close but I'm not sure
> where the disconnect in my system or configuration is and am hoping that
> y'all will be able to help me.
>
> I have followed the instructions on these sites as best I could and am
> seeing (more or less) what they say I am supposed to be seeing.
>
> http://time.qnan.org/
> and
> http://wiki.enneenne.com/index.php/LinuxPPS_support
>
> I am using a Garmin 35LVC for my OEM GPS unit. I have patched and
> compiled the 2.6.26 kernel with the appropriate LinuxPPS patch and have
> patched and compiled ntp-4.2.4p2 with the NMEA patch. For the kernel I
> have built in the 8250 serial and PPS directly into the kernel instead
> of as modules. RHEL4 update 5 AS standard install is the rest of the
> platform.
>
> The ntpd server is communicating with the GPS device but I suspect that
> it is not receiving the PPS signals. When ntpq is run it shows very
> large offset and jitter values when compared to servers it is connected
> to over the internet for comparison.
>
> Here are the config files as they are running on my machine and output
> from logs. Please let me know if there is other info that you would like.
>
> After booting I have to manually create symlinks /dev/gps0 -->
> /dev/ttyS0 and /dev/ppsgps0 --> /dev/pps0
>
> > dmesg | grep 8250
>
> Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing enabled
> serial8250: ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
> serial8250: ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A
>
> > dmesg
>
> .......
> PPS event at 56274
> PPS event on source 0 at 1223477791.159246144
> PPS event at 56524
> PPS event on source 0 at 1223477792.159245962
> PPS event at 56774
> PPS event on source 0 at 1223477793.159244902
>
> > cu -l /dev/ttyS0 -s 4800
>
> .......
> $GPRMC,145903,A,3022.9421,N,09744.6923,W,000.0,340.9,081008,005.8,E*61
> $GPGGA,145903,3022.9421,N,09744.6923,W,1,10,1.9,275.3,M,-23.1,M,,*70
> $GPGSA,A,3,03,,07,13,16,19,20,23,25,27,32,,3.3,1.9,2.6*37
> $GPGSV,3,1,11,03,48,100,36,06,40,080,30,07,34,304,38,13,62,325,36*72
> $GPGSV,3,2,11,16,24,041,30,19,36,142,33,20,19,200,31,23,85,133,40*76
> $GPGSV,3,3,11,25,48,325,35,27,26,300,33,32,09,181,31,,,,*40
>
> > cat /sys/class/pps/pps0/assert
>
> 1223478098.159245740#827
>
> > cat /sys/class/pps/pps0/assert
>
> 1223478100.159250016#829
>
> > tail /var/log/messages
>
> Oct 8 15:02:57 localhost kernel: PPS event at 153024
> Oct 8 15:02:58 localhost kernel: PPS event at 153274
> Oct 8 15:02:59 localhost kernel: PPS event at 153524
> Oct 8 15:03:00 localhost kernel: PPS event at 153774
> Oct 8 15:03:01 localhost kernel: PPS event at 154024
>
> ntp.conf
> --------------------
> restrict default nomodify notrap noquery
> restrict 127.0.0.1
>
> # LinuxPPS: GPS + PPS
> server 127.127.20.0 prefer minpoll 4
> fudge 127.127.20.0 flag3 1 flag2 0 time1 0.0
>
> server 0.us.pool.ntp.org
> server 1.us.pool.ntp.org
> server 2.us.pool.ntp.org
>
> server 127.127.1.0 # local clock
> fudge 127.127.1.0 stratum 10
>
> driftfile /var/lib/ntp/drift
> broadcastdelay 0.008
>
> >ntpd -g
> >ntpq -p
>
> remote refid st t when poll
> reach delay offset jitter
> ===========================================================================
> === xGPS_NMEA(0) .GPS. 0 l 17 16 376 0.000
> -641.83 72.089
> +ntp2.your.org 64.202.112.65 3 u 48 64 377 29.920
> 8.623 0.722
> *rrcs-64-183-55- .GPS. 1 u 63 64 377 66.258
> -2.181 1.880
> +mercury.element 128.252.19.1 2 u 33 64 377 42.065
> 6.104 1.858
> LOCAL(0) .LOCL. 10 l 40 64 377 0.000
> 0.000 0.001
>
> > ntpdc -c kern
>
> pll offset: 0.005085 s
> pll frequency: 63.751 ppm
> maximum error: 0.239794 s
> estimated error: 0.000879 s
> status: 0001 pll
> pll time constant: 6
> precision: 1e-06 s
> frequency tolerance: 500 ppm
>
> From the ntpq -p you can see that the ntp daemon says the GPS is being
> designated a flasticker and not being used. I do not know what is
> causing this.
>
> Any help you could provide or resources you could point me to would be
> appreciated.
>
> Thanks
>
>
>
> First some basics. The first thing you need to do is make sure that the
> pps signals are actually working.
>
> You don't have anything about running ppsldisc. Are you running it?
>
> Do you have the timer kernal module loaded? If you do you should unload it
> to prevent confusion about your PPS signals.
>
> After unloading the timer kernel module and starting ppsldisc in a separate
> terminal session what results to you get when running ppstest against your
> pps device?
>
> Until the above is working correctly it makes no sense to try to get ntpd
> working.
>
> Hal
>
> _______________________________________________
> LinuxPPS mailing list
> LinuxPPS at ml.enneenne.com
> http://ml.enneenne.com/cgi-bin/mailman/listinfo/linuxpps
> Wiki: http://wiki.enneenne.com/index.php/LinuxPPS_support
More information about the LinuxPPS
mailing list