[ntp:questions] How to debug PPS?

David Lord snews at lordynet.org
Fri Apr 23 09:24:12 UTC 2010


Steve Kostecke wrote:
> On 2010-04-22, David Lord <snews at lordynet.org> wrote:
> 
>> Chaofu Chen wrote:
>>
>>> Hi all. I want to build a NTP server but PPS doen't work for the time
>>> being. Status bits from "ntptime" and "ntpdc -c kern" include PLL and
>>> NANO, sometimes PPSTIME and PPSFREQ, but PPSSIGNAL never appears.
>> If you have a PPS signal on serial DCD, I'd suggest adding pps driver
>> "server 127.127.22.(n)" which requires /dev/pps(n)
> 
> The NMEA driver (127.127.20.X) will use the PPS signal if it is present
> on the DCD pin.
> 
> http://doc.ntp.org/4.2.6/drivers/driver20.html provides set-up details.
> 
> You don't need the PPS driver (127.127.22.X).


With NetBSD-5, ntpd 4.2.4p6-o, gps18x-LVC, last year I seemed
to get lower offset and jitter from PPS driver that was from
just couple of days with each config. Having just rmc sentence
seemed to make a difference.

                RMC-only    Default
NMEA only       0.07 ms     0.12ms
+PPS type 22    0.002ms    0.004ms

Since then I've kept separate PPS driver
server 127.127.20.2 mode 1 minpoll 6 maxpoll 8 prefer
fudge  127.127.20.2 time1 0.100 refid GPSb
server 127.127.22.2 minpoll 4 maxpoll 4
fudge  127.127.22.2 flag2 1 flag3 1 refid PPSb

past 4 days from peer_summary
127.127.22.2
mean(ms)  -0.000 0.000 0.000 0.000
rms(ms)    0.002 0.006 0.007 0.006
max(ms)    0.019 0.040 0.036 0.033

127.127.20.0
mean(ms)   0.002 0.003 0.003 0.003
rms(ms)    0.002 0.006 0.007 0.006
max(ms)    0.018 0.038 0.033 0.031

I'd guess large day/night temperature change is causing
higher rms/maximum offsets than normal.


David




More information about the questions mailing list