View Single Post
  #371   Report Post  
Posted to sci.electronics.basics,sci.electronics.repair,alt.comp.hardware.pc-homebuilt
Gordon reeder
 
Posts: n/a
Default Why aren't computer clocks as accurate as cheap quartz watches?

"Never Mind" wrote in
news
On Wed, 26 Oct 2005 17:46:23 -0700, do_not_spam_me wrote:

I'm looking for a techical explanation.


Somewhere in the Far East there is a big factory making ****loads of
crystals. These crystals have a nominal spec, but are subject to
tolerances.

Somewhere in that factory these crystals are measured. Those closest to
the designated frequency are shipped to Seiko, the farthest to Dell.

Watches tend to spend a lot of time strapped to a temperature-regulated
heat source (a human). PCs usually are not.

As mentioned in previous (non-technical) replies, PCs are not often

used
as timepieces, so lower tolerance (cheaper) components will be used.




I'll add that frequency of opperation will depend on supply
voltage. So the difference between running on battery 2.5V
and running on AC 5.0 to 3.3V (depending on chipset) will cause
the clock to gain or loose time.

Parasitic capacitance also affects opperational frequency.
These are tightly controled in a watch, but are of little
concern in a PC.

Solution: Get a program called Dimension 4. It runs on startup,
syncs your PC clock to an internet time server, then exits. It
can work in other modes too.