[time-nuts] Happy Leap Hour Day !

Tom Van Baak tvb at leapsecond.com
Sun Mar 11 23:35:41 EDT 2007


> > might take one 2 AM transition. Or it might work in
> > another hour when 11-Mar-2007 UTC is over. Can
> > you try it again a few times and tell me when it first
> > comes up with the correct time?
> 
> Tom, I did another forced sync and now it has the correct time.

Cool. Yeah, it sounds like it has the bug. I predict that if
this receiver does not get a signal the night before, or is
reset on the transition Sunday between 2 AM local time
and 24h UTC (early evening, local time) it will be off by an
hour.

I suspect the person/company who wrote the firmware for
the RC clock didn't think about this case, or didn't test for
it, or didn't code general enough so that this case wasn't
special in the first place.

If you or anyone on the list wants a challenge, try reading
the spec yourself and pseudo code an algorithm that will
*correctly* convert WWVB subcode to local date & time,
accounting for USA timezones and DST. It's much harder
to get right than you might think. Extra credit: see if it's
simple enough to run on a 32 kHz 4-bit CPU...

/tvb




More information about the time-nuts mailing list