> It's interesting that you can buy 0.1ppm (commercial temp range)
> TCXO (not OCXO, so only a few mA @ 3.3V) oscillators for $35 or so:
>
> http://www.conwin.com/datasheets/tx/tx350.pdf
>
> That should be a few seconds per year, but the initial calibration
> is only +/-1ppm (about 30 seconds per year).
>
> or maybe one of the new < 150mW atomic clocks for about +16dB$

A GPS synchronised watch would be easy enough [tm] but overkill.
Almost practical would be a watch that synced to a pulse that occurred
say on minute boundaries whih it read when it was nearby. Could eg be
a low cost IR LED sender that plugs into a USB port and the watch sees
it optically.
If the computer is linked to a time server (as many are ) the watch
could be maintained at well under 1s absolute error at all times.
Buying a TI Chronos when they are on special at $US25 would probably
be as cheap :-).
BlueTooth would probably be cheaper than a small scale custom
implementation but raw transition based on 'internet time' will
probably allow greater (and usually unneeded) accuracy than Bluetooth.

I had desynced my PC from the time server a while ago as the co
conspirators had between them not managed to cope with NZ's change in
the date of applying daylight saving and at one stage insisted on
"correcting" the time to an invalid value.

A friend's PC was always some hours out. Desyncing from the time
server allowed them to set the time correctly. Odds are, the fact that
they had lived in Chile and bought the laptop involved there was
probably relevant, even though the various time zone settings had been
NZ adjusted. Deep and mysterious can be the ways of the Gates.



     Russell
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.