[ntp:questions] ntp drops gps/pps - lets time drift , solution: use 'true' in ntp.conf

M Z mcrunch2 at yahoo.com
Tue Aug 2 00:49:22 UTC 2011


Thanks for all the suggestions. I finally saw an ntp.conf configuration command 
option named 'true' which forces ntp to use the source no matter what. This 
works great with my offset and jitter now < 0.010 for default tsc and also hpet 
clocks. Also, for the Globalsat MR350P, I don't usually have to use a 
time1 offset (I have 4 globalsats).
The ntp.conf below says "don't use the nmea stream for timing but always use the 
pps signal for timing". Since this
is using gpsd for the time calculations, the nmea time is folded into the pps 
signal at .28.1 and it is more stable to
make ntp ignore the not-too-accurate nmea time at .28.0.
So, the final ntp.conf for SuSE 11.2 looks like:

disable kernel
server 127.127.28.0  minpoll 6 noselect
fudge 127.127.28.0  time1 0.550 refid GPS stratum 3     #garmin
server 127.127.28.1  minpoll 4 prefer true
fudge 127.127.28.1  refid PPS

 *** Most of the original messages have been snipped for brevity *** 

 
>----- Original Message ----
>From: "questions-request at lists.ntp.org" <questions-request at lists.ntp.org>
>To: questions at lists.ntp.org
>Sent: Fri, June 10, 2011 7:00:00 AM
>Subject: questions Digest, Vol 80, Issue 11
>
>Today's Topics:
>
>  1. Re: ntp drops gps/pps - lets time drift
>      (E-Mail Sent to this address will be added to the BlackLists)
 > 2. Re: ntp drops gps/pps - lets time drift (David Lord)
>  3. Re: ntp drops gps/pps - lets time drift (Rob)
>  4. Re: ntp drops gps/pps - lets time drift (David Lord)
>  5. Re: ntp drops gps/pps - lets time drift (David Lord)
>  6. Re: ntp drops gps/pps - lets time drift (Rob)
>
>
>----------------------------------------------------------------------
>
>Message: 1
>Date: Thu, 09 Jun 2011 12:00:30 -0700
>From: E-Mail Sent to this address will be added to the BlackLists
>    <Null at BlackList.Anitech-Systems.invalid>
>To: questions at lists.ntp.org
>Subject: Re: [ntp:questions] ntp drops gps/pps - lets time drift
>Message-ID: <isr58f$92r$1 at dont-email.me>
>Content-Type: text/plain; charset=ISO-8859-1
>
>M Z wrote:
>> How can I force ntp to use the PPS signal or at least not drop it?
>
>prefer ?
>
>Message: 2
>Date: Thu, 09 Jun 2011 19:54:56 +0000
>From: David Lord <snews at lordynet.org>
>Try with prefer for the refclock rather than the PPS
>
>server 127.127.28.0  minpoll 4 prefer
>
>and don't use 127.127.0.1
>> server tick.usask.ca iburst    #stratum one canada server
>> driftfile /var/lib/ntp/drift/ntp.drift  #(has had values from -70 to 13)
>> restrict -6  ::1
>> restrict 127.0.0.1
>> restrict default  kod nomodify notrap nopeer noquery
>> restrict tick.usask.ca  mask 255.255.255.255 nomodify notrap noquery
>> 
>> after 12+ HOURS  looks good to me:
>> ra2:~> /usr/sbin/ntpq -p
>>      remote          refid      st t when poll reach  delay  offset  jitter
>> 
==============================================================================
>> xSHM(0)          .GPS.            3 l    7  16  377    0.000  108.591  32.540
>> *SHM(1)           .PPS.            0 l    6  16  377    0.000    0.007  0.001
>> xtick.usask.ca  .GPS.           1 u  41  64  377  92.184    6.839  0.316
>> 
>> So, then I run without external sources (like tick.usask.ca) and I run both 
>>without and with localclock, I get:
>> 
>> without 127.127.1.0 (SHM (GPS/PPS) only):
>> ra2:# ntpq -p
>>      remote          refid      st t when poll reach  delay  offset  jitter
>> 
==============================================================================
>> xSHM(0)          .GPS.            3 l  13  16  177    0.000  122.103  6.617
>> xSHM(1)          .PPS.            0 l  12  16  177    0.000  -0.113  0.153

>...SNIP...

>End of questions Digest, Vol 80, Issue 11
>*****************************************




More information about the questions mailing list