Hi all,

I have developed (like millions others) a software based Clock.

First I develped the time base, using 12MHz crystal and 89c51 in my "TEST BOARD".

Ater busting the error enough down to as low as 60 seconds loss in a year, I implemented the software in the actual PCB with 89c2051 micro keeping the time base untouched.

But it is gaining time horribly, about a second an hour .... :(

I have used the same cyrstal by removing the test board actually.

I think one of the following may be the culprit:

+ the crystal is same but not the stabilizing caps 33 uF.

+ the internal divider (divide by 12) of the micro may be errornous.

Any help is highly appreciated.

Babar/