Re: [gpsd-users] Help with extracting time information from GPS to NTPD

[Two Sank]

On 11/08/2011 11:21 AM, Jordan wrote:
> I just purchased a Haicom HI-206-USB GPS device looking to extract time 
> information from GPSD and pass it to NTPD to get an accurate
> system clock. I'm using Ubuntu 10.04 LTS.
> 
> I believe the GPS is working because I can run xgps and see the 
> satellites/time info/location/etc.
> 
> However I can't quite figure out how to get ntpd to pull the information from 
> gpsd. I've modified my /etc/ntp.conf to:
> 
> Code:
> 
> server 127.127.28.0 minpoll 4 maxpoll 4
> fudge 127.127.28.0 refid GPS
> 
> restrict 127.0.0.1
> restrict 10.1.1.0 mask 255.255.255.0 nomodify notrap
> 
> where 127.127.28.0 is supposed to be the IP to access the GPS information. I 
> can ping that IP successfully. However when I try to
> access the server using ntpq -p, it says 'No association ID's returned'.
> 
> If I stop NTPD and run 'ntpdate 127.127.28.0' it says there were no suitable 
> synchronization servers found. 
> 
> NTPD works fine if I point it to a public time server. 
> 
> Any help is greatly appreciated as I've run out of ideas.
> 
> Thanks

Do readers of this forum not think that there should be a standard response to 
people hoping to use GPSD with NTPD.
It comes up fairly frequently. It seems to follow the same route i.e.

someone tries to help make the link between the daemons.
the requestor manages to get the two daemons associated (sometimes with no 
other time server)
it fails to synchronise their PC clock
then someone points out that USB linked gps are not appropriate because the 
time at best is accurate only to one
    second with lots of jitter, and that a real serial line with pps is 
necessary
then someone points out that a single source is not what ntpd is looking for
then someone suggests Chrony as an alternative
then the user discovers some erratic behaviour in terms of reporting time from 
their gps
etc. etc.

The problem is that I don't know enough to give the best advice but the 
question does cause quite a lot of traffic and
the tests wastes a lot of the requestors time.

I know that searching the archives could turn up a lot of this stuff, but 
perhaps a standard summary might help.

Just trying to be helpful..
Ray