[ntp:questions] PPS problem
snews at lordynet.org
Sun Apr 6 11:47:06 UTC 2014
Simchaev Arthur-BAS071 wrote:
> Hi David
> I have the following system :
> uname -a
> Linux am335x-evm 3.2.0-4.08 #5 Wed Mar 12 15:07:28 IST 2014 armv7l GNU/Linux
> And pps is existed in the kernel , when I run ppstest /dev/pps0 , I see the pps clock grown . Also ntpd shows that the server is PPS "o".
> But I am confuse why I still don't see PPSFREQ,PPSTIME,PPSSIGNAL status . How I can be sure that my clock is synchronized to the PPS ?
I'm not at all familiar with recent ntpd on Linux and have never
run an armv71 system and my oldest GPS is a Garmin GPS18xLVC.
On my systems cron jobs run /root/scripts/ntp/ntp-stat.sh
at 6-10 minute intervals.
ntpq -c lpeers:
| Wed Apr 2 17:36:00 GMT 2014
| remote refid st t poll reach delay offset jitter
| -GPS_NMEA(2) .GPSb. 7 l 64 377 0.000 -1.940 8.828
| oPPS(2) .PPSb. 0 l 16 377 0.000 0.000 0.004
ntpq -c kerninfo:
| associd=0 status=011d leap_none, sync_pps, 1 event, kern,
| pll offset: -0.000317
| pll frequency: -35.8115
| maximum error: 0.406
| estimated error: 3e-06
| kernel status: pll ppsfreq ppstime ppssignal nano
| status 0x2107 (PLL,PPSFREQ,PPSTIME,PPSSIGNAL,NANO),
After I restart ntpd offset is sometimes >> 10ms when PPS kicks
but then fairly quickly reduces to a few us.
System load/local temperature changes give offset blips of >> 30us
once a day from daily logging, with a larger blip at weekend from
the weekly cron jobs.
You can also check your wiring and circuitry used for the pps
signal. I've tried input to printer port lpt0 which was no
different to serial/DCD but you may have gpio which could be
better if you have support compiled in.
More information about the questions