[time-nuts] Rohde & Schwarz GPSDO

Martin Burnicki martin.burnicki at burnicki.net
Thu Oct 8 16:56:33 EDT 2015

Bob Camp wrote:
> Hi
> One thing you can expect - some sort of issue with the GPS rollover issue. 
> Can it be fixed on this or that unit? That’s going to be highly variable model to
> model. Not much of an issue for a frequency standard, but a pain for NTP.

No need to worry, there is no problem with GPS rollovers.

The built in GPS receiver has been developed by Meinberg, and I wrote
parts of the firmware which evaluates the satellite data.

Handling of the week number in these receivers is *not* by simply adding
another 1024 weeks if the truncated wn is below a certain limit.

Instead, the wn is extended to 16 bit internally and simply incremented
on each 10 bit rollover into the next epoch.

If the device's backup battery fails and thus the epoch number is lost
you can simply set the internal date to the current date, so the
extended week number is computed and the device outputs the correct date
and time.


More information about the time-nuts mailing list