Re: [avr-chat] Strange problem with ATMega32A-PU

[David A. Lyons]

On Apr 30, 2011, at 1:56 AM, Royce Pereira wrote:
> Well I discovered a solder bridge from a MCU pin to ground. ... After 
> removing the bridge, the board works with any chip (32/L/A).

Which pin?  How is that pin configured in software?  What else in your project 
is connected to that pin?

It's probably too late to collect this data, but:  What was the resistance of 
the solder bridge?  Was Vcc (at the MCU) lower than expected when it was 
misbehaving?

I imagine:  (1) the bridged pin was an output, and when driven high it acted as 
a partial short across your power supply, or (2) the bridged pin was an input, 
and the software looped forever if an expected input transition never happened. 
 Small differences in the MCU's "Electrical Characteristics" could make all the 
difference here.  Without knowing more details of your project, I can only 
guess.

> I finally disabled the Timer2 (my system tick) interrupt, & put a while(1) ; 
> after the LCD welcome message ... Now I get the welcome message.

What does your project do with the bridged pin normally, that it doesn't do 
when you disable Timer2 and stop after the welcome message?

> Probably that was hanging my MCU ? Strange, the 'L' version chip did not hang.

I don't think we've established that the MCU was hung, only that your project 
didn't behave as expected.  The software could be stuck in some way, but still 
executing instructions correctly.

If no debugging hardware is available, I'd suggest narrowing down the problem 
by continuing the experiments you started:  If disabling Timer2 and stopping at 
a certain point is OK, stop at later and later points, and enable/disable 
various pieces of the work done by Timer2 until you find something specific.

Cheers,

--Dave