[time-nuts] Another arrival time histogram: SiRF III

Magnus Danielson magnus at rubidium.dyndns.org
Tue Aug 2 12:00:10 EDT 2016


On 08/01/2016 05:49 AM, Hal Murray wrote:
> It was setup to only send GPRMC sentences.
>   http://users.megapathdsl.net/~hmurray/time-nuts/RMC-hist.png
> Here is what it looks like over time:
>   http://users.megapathdsl.net/~hmurray/time-nuts/RMC-offset.png
> There are interesting glitches at 3 hours and 23 hours.  Anybody have any
> ideas about what would cause that?

This most likely is related to scheduling inside the RTOS of the GPS 
receiver. Due to the frequency difference between GPS second and 
internal clock processing that we can expect, the scheduling needs 
re-alignment. There is an overlap between "late" and "early" which 
indicate that a dynamic scheduling, so depending on when things get 
ready it gets early or late. Variations within each time-slot is then 
due to processing variations.


More information about the time-nuts mailing list