It doesn't sound like you have any brownout protection.  During a
brownout, the PIC is running but RAM is scrambled: this means the PC
can contain a random address, and execution can jump into the middle
of your EEPROM write code.

On Wed, 4 Oct 2006, Bob Axtell wrote:

> I'm getting some strange results with an application that powers on/off
> a low-end PIC repeatedly (about once per second): the EEPROM
> is getting scrambled randomly.
>
> The application reads something out of the EEPROM and sends data
> out. All signal and power lines are clamped with low-voltage varistors
> and the power is applied from another PIC several feet away. Yet before
> the night is over, the EEPROM is scrambled.  Yes, there IS code to
> write to EEPROM in the data space, but it is never actuated, in fact has
> been prevented from writing to the EEPROM..
>
> It seems that the device is powering up and running out of the wrong data
> space. The device is a PIC12F629 in SOIC8. The code simulates perfectly
> (never jumps to wrong area in MPLAB 7.41) and  otherwise works perfectly.
> Uses config: internal RC, WDT ON, Pwrup Timer ON, Brwn OFF, Mclr EXT
> (but has 22K pullup).
>
> Anybody ever seen anything like this before? I am very perplexed...
>
> --Bob
> -- 
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
John W. Temples, III
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist