[ntp:questions] What is the NTP recovery time from 16s step in GPSserver?

Terje Mathisen "terje.mathisen at tmsw.no" at ntp.org
Wed Oct 31 09:09:23 UTC 2012

Rob wrote:
> The main point is that the local clock is not supposed to be broken.
> It worked OK years ago when I implemented it, and I wonder what has
> changed now so that it no longer works.  Maybe a change in the Trimble
> firmware or a change in gpsd (although I do not see that in the git
> browser).
> When this has been cleared up and fixed, gpsd should not provide time
> until the receiver reports the correct values.

My guess would be that the Trible has some form of almanac backup 
battery, but that this has worn out over e number of years, so it no 
longer works.

I do know that when I made my first Oncore, based on the TAPR board, I 
decided to _not_ put in a backup battery, in order to avoid problems 
when it eventually failed.

A quick&dirty hack would be to either:

a) Don't provide any timestamps until the Trimble reports all-up.

b) (Very dirty!) Remember (in a file) the TAI-UTC offset from the last 
run, and start with this value until you get an update from the GPS. If 
said update is different from the stored value, update the file.

With this approach you will be OK unless you get a leap second while the 
GPS is without power...

- <Terje.Mathisen at tmsw.no>
"almost all programming can be viewed as an exercise in caching"

More information about the questions mailing list