[LinuxPPS] Re: nmea

[Udo van den Heuvel]

Paul,

Please post to the LinuxPPS list.

Paul wrote:
> I thought I would update you on my curious problem with the nmea and
> pps.
> 
> My ntp.conf had time1 0.0 (I think this is for adjusting for VERY small
> delays on the pps signal - although I guess you would need access to a
> primary standard atomic clock to adjust this) so I changed it to time1
> 1.0
> 
> Now the pool servers and I agreed and after I let it run overnight the
> ntpq -p gave both good agreement and low jitter from the GPS source.
> 
> I don't think the length of the pulse can be adjusted on this board,
> there is a serial input, but only for differential GPS corrections (no
> access to other parameters. Udo, I guess from your reply you use the
> Garmin which is rather easier to adjust (and to interface).

I use a GPS-18LVC.

> Obviously I had checked the satellite number output -and these are good
> - when I first made my interface board I was suprised to see the pps LED
> flashing with no antenna so I was aware of a potential problem.
> 
> I am sure I have the assert/clear the right way round the assert leads
> the clear by 25.6 mS.
> 
> The Rockwell (manufacturer) data sheet says that the nmea time is output
> for the next pps assert. So if it outputs 10:21:36 that time is valid
> about half a second later when the pps signal goes high. Is this the
> same as the Garmin and other GPS units?

The Garmin sends the first PPS edge and at the same time starts sending
the NMEA data for that moment in time.


Udo