Yes, GPZDA is doing the trick. You know I tried other sentences before but a) 
probably wasn't patient enough and b) I have found the fudging time2 value is 
critical - Time2 is now 560 to get the NMEA refclock to stop being false 
ticker. Surely that shouldn't be a critical value? 

As a footnote, I have a number of GPS rx lying around from various gpsdo 
projects. I am tempted to try some of them out as well and see if the time2 is 
really touchy. Hmm, Actually, I have a thunderbolt and a GCRU (scpi gpsdo) I 
can hook in to my test ntp server too. The secret to planning projects is make 
em so big they will never get started ;) Anyway, there I go raving again..

Thanks Dave

>It's the only one that matters here.  refclock_nmea.c treats both as valid 
>timecodes, but the timing mode GGA output tells it the GPS Quality >Indicator 
>is 0, meaning "Fix not available or invalid."  That makes some sense in that 
>it's not attempting to provide position fixes in timing mode, >but the NMEA 
>driver can't distinguish that from "I've just started and have no clue if my 
>time code is synchronized to UTC."

>Perhaps instructing the driver to use a different NMEA sentence will cure it?


> Footnote:
> To further complicate matters, I wired up a tee cable and sent the NMEA to 
> another NTP server here running windows NTPD 4.2.7p238-o.
> NMEA works fine on that server in either mode.

GPGGA I kid you not, it's in clockstats...
>Which NMEA sentence is the Windows ntpd showing in ntpq clockvar (cv) output?

Cheers,
Dave Hart


_______________________________________________
questions mailing list
[email protected]
http://lists.ntp.org/listinfo/questions

Reply via email to