Fw: [time-nuts] GPS Leap Second
Tom Van Baak
tvb at leapsecond.com
Fri Aug 5 00:20:29 EDT 2005
> > Yeah, my 58503B SmartClock just picked it up too.
> > But as I suspected, based on what was reported in
> > 1998, it has the date wrong:
> > scpi > ptime:leap?
> > 1
> > scpi > ptime:leap:stat?
> > 1
> > scpi > ptime:leap:acc?
> > +13
> > scpi > ptime:leap:date?
> > +2005,+9,+30
> That looks similar to the alleged 8 bit week counter overflow Tom
> reported for the HP 59551A
> (http://www.leapsecond.com/notes/leapsec256.htm). Maybe the 58503B
> has a week counter that started at 0 in 2000 ??
Nope, the 2005-09-30 leap second isn't the 256
week bug. What I recall is that it shows up only
in HP SmartClock receivers only when the leap
second is announced more than 3 months ahead
of time. Can someone else confirm this? It was
reported in 1998 too when IERS also gave more
notice than usual. Doug, you remember this
happening in 1998, right?
I think it's a simple HP firmware bug where they
sort of took the leap second guidelines too literally:
"first preference is given to the opportunities at
the end of December and June, and second
preference to those at the end of March and
My wild guess is the firmware looks at the leap
second subframe fields and then incorrectly
schedules the leap second at the end of the
nearest quarter year.
If there's someone from Agilent or Symmetricom
on the list, please consider scanning a page or
two of the now ancient SmartClock firmware for
the sake of history so we can see what really
happened and learn from it.
More information about the time-nuts