Did you use watchdog timer?
That seems to be issue with a dead loop. Check for cases that would enter an endless loop . Using watchdog would restart the chip, not that you should rely on the WD to do this, you still need to fix that if WD does kick in.
The minipro may reset the chip when inserted.
Check if the values pass to a funtion can be out of range.
Just to be quite sure:
Are you using a MiniProg or a MiniProg3? Connection made with supplied the ribbon-cable??
Are you able to debug? Setting Breakpoints??
What happens if you do not unplug your MiniProg(3)
Post your schematic, or at least the portion that shows power supply connections
and miniprog connections.
Of course there is always issue of coupling (crosstalk) to pins from some external source
where a input pin is configed as Hi Z and close to other fast signals not of interest. Or
an interface connector that is electrically marginal and sensitive to miniprog presence/absense
from mechanical movement. These are more general signal integrity issues.
Yes, a floating input may also be the problem.
Thanks to everybody for their answers and very sorry for the late reply.
Indeed H L, I'm pretty sure it was due to a dead loop. I fixed it and that issue doesn't occur anymore. That was easy! Thanks for pointing it to me.
To answer your questions, I did not use watchdog timer. Bob, I'm using a MiniProg3, connection is made with the supplied ribbon cable.
glad that your problem is fixed.
And nice to have people come back to tell us the result.
I replied a little too fast and was not completely accurate.
The program doesn't stop anymore (issue was the dead loop) but it restarts automatically as soon as I plug back the MiniProg3 (which doesn't bother me as I'm about the reprogram anyway).
I think this is normal.
That's what I thought too. Anyway, the main issue was fixed. :-)