SUMMARY:    You CAN overclock a PIC but can never be certain of the
results.
Many people report success at 50% plus speed increases.  YMMV.


DETAIL    There's been lots on this here before.
Yes, you can always overclock a PIC with SOME degree of success.
You will never know how high the limit is for a given IC without
exhaustive testing.
SOME operations may fail before others.

If you have eg a 4MHz version and you can buy an eg 10 MHz version
(eg 16F84) then it is entirely possible (even probable) that you can
run your 4MHz part at 10 MHz or even better. Presumably some key
parameters are tested by Microchip when the 4/10 MHz parts cut is
made. Common opinion is that they may well arbitrarily label SOME
parts as 4MHz if the yield of full spec parts is too high :-) (a nice
problem to have). Many people also report running various PICs at up
to around twice their rated speed.

It seems likely that parts like the 16F84 MAY fail eg EEROM writes
while other features work OK. Also, the parts are spec'd across a
temperature and voltage range. As speed increases you will almost
certainly find that operation at the lower voltage becomes less
certain.

You can also operate parts below their design voltage spec and MAY
have no problems - I have seen 4 MHz 16F84s run at 3 volts at 4 MHz.
Not recommended.



RM




-----Original Message-----
From: Sam Laur <slaur@SEKUNDA.PP.UTU.FI>
To: PICLIST@MITVMA.MIT.EDU <PICLIST@MITVMA.MIT.EDU>
Date: Monday, July 05, 1999 7:44 AM
Subject: Re: OverClocking a PIC


>> Does anyone know if you can "overclock" a PIC?  (I remember in the
olden
>
>Yup, like I've written before to this list, I've clocked some
16F84-04
>individuals to 18.8 MHz (at 18.9 they failed, ha!) and some old
12C509 (non-A)
>to 30 MHz, where I ran out of range on my RF generator. Now I'd have
some
>new PICs to test (16F877, 16C622A, 16CE625, 16C505) too, but haven't
got
>much spare time. But I suspect I would run out of generator range on
just
>about all of them, so I need a better generator first.
>
>Nice note on the 12C509 - when running at 30 MHz, toggling an output
on/off
>and running a delay loop the other time, it still took only about 3
mA. And
>ran cool.
>