Re: [gpsd-users] Leap second has arrived but GPS not updated?

[Tomalak Geret'kal]

On 02/07/2012 01:47, Alexander Carver wrote:
> On 7/1/2012 16:34, Tomalak Geret'kal wrote:
> > On 01/07/2012 20:31, Alexander Carver wrote:
> > > On 7/1/2012 12:19, Eric S. Raymond wrote:
> > > > Tomalak Geret'kal <address@hidden>:
> > > > > > Now I'm even more confused.  As of about an hour ago, 
> > > > > > the GPS UTC
> > > > > > offset has gone *backward* from 16 to 15 again.  I'm 
> > > > > > not sure what
> > > > > > happened or why.  At about 0700 UTC it started 
> > > > > > showing an offset
> > > > > > of 16 seconds. Somewhere after 1300 it stepped 
> > > > > > backwards.  I still
> > > > > > have satellites in view, six are locked for a fix 
> > > > > > plus the one
> > > > > > SBAS is good.
> > > >
> > > > > Sounds to me like your receiver has a "default" of 15 
> > > > > in its NV RAM,
> > > > > plus a bug preventing it from properly extracting the 
> > > > > UTC Correction
> > > > > parameter from almanac.
> > > >
> > > > I agree with this diagnosis.
> > >
> > > I'm forced to agree with it but I don't want to. :)  I'm 
> > > exhausting
> > > all other possible avenues before declaring the receiver 
> > > badly
> > > programmed.
> > >
> > > The sad part is that I have four more of these identical 
> > > units.  I
> > > suspect there are many in the field considering they 
> > > were used in
> > > remote telemetry devices.  I'll be disappointed if they 
> > > can't be fixed
> > > but I doubt GlobalSat even supports them anymore.
> > The good news is you got them free.
>
> True, tough to beat free.  So it seems that it defaults to 
> a hardcoded value if it loses the UTC information from the 
> sats.  Once it picks up on the info it will use it.  It 
> has once again started reporting 16 seconds for the 
> offset.  But, I suspect if I cover up the antenna and have 
> it lose all signals it will revert to 15 seconds.
Ideally it wouldn't do that until you cleared the NV RAM and 
performed a cold start, but we've already established that 
it appears to be buggy software so I guess all bets are off.

Tom