[time-nuts] End Of The World

Martin Burnicki martin.burnicki at burnicki.net
Thu Jul 2 09:36:08 UTC 2015


Bob Camp wrote:
> Well it sounds like there are (so far) no major reports of extreme behavior. There still seem to be
> *lots* of reports of systems that do not handle things gracefully. I wonder if the data above is
> specific to a chip set (which it probably is) or to how Glonass  handles leap seconds (yikes !!!).

As said in my previous email, it was a NVD8C-CSM v3.1 module 
specifically set to GLONASS-only mode.

I'm assuming the problem is specific to GLONASS.

The GPS satellites sends their data frames in 6 second intervals aligned 
with GPS time which is linear, with constant offset to TAI, so the data 
frames are not affected by a leap second.

It's been quite a while that I looked at some GLONASS specs, but if I 
remember correctly the data frames sent by GLONASS satellites are 
aligned with UTC. So if a leap second is inserted or deleted there *is* 
some interruption in the sequence of the data frames.

So it's much more difficult for receivers to keep tracking GLONASS 
satellites than GPS satellites over a leap second, and I'm not 
specifically surprised that GLONASS receiver at least loose sync when a 
leap second occurs.

Martin




More information about the Time-nuts_lists.febo.com mailing list