WRT making a clock (as in a wall clock) based on a crystal, don't do it
unless you have to.  The drift is not much, but it accumulates over days
and weeks.  But if you are powering it by AC you have a good solution
available to you.  The power grid needs all the power to phase together
perfectly so there are exactly the same number of zero-crossings every day.
Add a few parts to create an interrupt for each zero-crossing of the power,
and you will have a stable clock.  For extra credit add a battery and have
it continue to run without display while power is out.

Of course you could get really fancy and have it receive WWV or GPS radio
signals.  But if all you want to do is to create a digital watch (no AC),
you should look into trimming the crystal with a tuning capacitor.

Vance Gloster           Friendship is far more tragic than love.
vance@eni.net                       It lasts longer. -Oscar Wilde

-----Original Message-----
From: wouter van ooijen / floortje hanneman <wf@XS4ALL.NL>
To: PICLIST@MITVMA.MIT.EDU <PICLIST@MITVMA.MIT.EDU>
Date: Friday, December 18, 1998 9:13 AM
Subject: delay accuracy


>Two questions:
>
>1
>Has anyone measured the accurracy of a 4 or 10 MHz crystal attached to a
>16x84 with
>the recommended capacitors? Is it enough to make a clock?
>
>2
>I have tried to make an accurate delay routine using the tmr0 and
>prescaler.
>I used mplab to simulate the code but found a jitter of 1 in the amount of
>instructions between each interrupt. Is this normal?
>
>regards,
>Wouter.
>