Re: NTP via tcp NMEA feed

[Nick Taylor]

Yo Gary
> Something wrong with the time in your NMEA feed:
>
> gpsd:WARN: WARNING: WKRO bug: leap second 18 inconsistent with 1046516735, 
> corrected to 1665831935 (2022-10-15T11:05:35.000Z)
>
> Got a the matching NMEA for us to see?
>
> Any idea what $GPHWBIAS is?
>
> gpsd:PROG: NMEA0183: GPHWBIAS is just after a cycle ender. ({ONLINE})
Maybe this was a red herring - I used different GPS device (Huawei ME936 
Modem with GPS) due to my setup here and I think it's flawed. We are 
lucky that gpsd is clever enough to fix things up.

New test using the normal SIM68 feed shows the following:

nmea-raw.log is read direct from /dev/ttyS5 with socat

gpsd-serial.log is output from std working gpsd on local machine (ntpshm 
feed fine)

gpsd-nmea.log if from remote server with nmea feed (ntpshm not working)

I also tried using socat to feed the raw /dev/ttyS5 as NMEA stream 
direct without involving gpsd on the source machine - same result with 
ntpshm not working

I assume this same problem will always show with NMEA feeds but probably 
no-one has bothered to use ntp on remote machine as it's never going to 
be accurate, however for us it would give reasonable timesync - plenty 
accurate enough for our usage...

Best

Nick

gpsd-nmea.log
Description: Text document

gpsd-serial.log
Description: Text document

nmea-raw.log
Description: Text document