[ANSS-netops] Be aware of possible early leap second implementation!

[Dietz, Lynn]

On January 22, 2015, NCSN staff noticed that data produced by one of our
Earthworm adsend systems had a -1s timestamp offset when compared with
other nearby data. The initial time shift had occurred around 20150121 2330
UTC. The TrueTime XL-AK GPS clock used as the time source for this system
had apparently applied the future leap second as soon as it was announced
in the GPS almanac (see article in URL).

http://gpsworld.com/leap-second-implementation-confuses-some-receivers/

Only 1 of the 8 TrueTime XL-AKs operating in our network exhibited the -1s
problem, and it has been replaced with a correctly-functioning unit. But we
do have a 2nd (spare) unit that syncs with a -1s time shift.  Power cycles
and re-syncing to GPS do not fix the problem. We assume it is a bug in the
firmware. However, another TrueTime XL-AK with identical firmware revs is
not exhibiting the problem, so we are a bit stymied.

Note that this is not the first time we've found
early-implementation-of-leap-second bugs in GPS receivers used as time
sources in our seismic network.

Be aware!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://geohazards.usgs.gov/pipermail/anss-netops/attachments/20150203/12b78f65/attachment.html>