[time-nuts] Re: HP Z3801A project update

[Hal Murray]

eb at telight.com said:
> I can't figure out why the 01 Jan 2006 initial date allows it to apparently
> work. I've looked at rollover info online, and ballparked some numbers -
> nothing makes sense to me to explain it  numerically

It's simple after you see it.

You get a date/time from GPS.  The week slot has been truncated to 10 bits.  
So the correct date is date-from-GPS + N*1024_weeks.  The question is what is 
N?

If you know some magic must-be-after date, then go into a loop:
  while date < magic:
    date = date + 1024_weeks

One good source for the magic date is when the software was built.  With a 
little help from the compiler, you can automate that.

That will work for roughly 20 years after your magic date.  (1024 weeks is a 
little less than 20 years.)

You could also get the magic date from NVRAM.  You could, for example, update 
that slot every year.  Then it won't stop working after 20 years.





-- 
These are my opinions.  I hate spam.