PIC18F6722 getting reset intermittently

[cypher]

Hi All,

I have a a very weird and frustrating problem happening. I am using a PIC18F6722. The problem is that sometimes the PIC gets reset for some unknown reason.

The nature of the reset taking place is when I get an interrupt on RB4-RB7 pins.

So far the things I'm considering that could cause a hardware reset are:
1. Watchdog timer
2. Brownout

Here are my fuses:

[PCM programmer]

[cypher]

Ok thanks, I will try your suggestion....also the links are very helpful, I will try a few things soon

[cypher]

So I read thru all those previous posts and used the restart_cause() function to find out exactly why my reset it happening. And interestingly enough, its because of a RESET_INSTRUCTION.

So, first of all, the reset_cpu() function in the code should never be called, however, I have commented out everywhere in my code where I use the reset_cpu() just to make sure it never gets executed. But the problem still exists, the intermittent resets I am getting are due to RESET_INSTRUCTION.

What else could be causing a RESET_INSTRUCTION reset other than the reset_cpu() command being executed?

[PCM programmer]

Make a short test program that calls restart_cause(), displays the result
on a terminal window on your PC, and then sits in a while(1) loop forever.
If you are getting random resets, the terminal window will fill up with
additional lines of "Restart cause = xxxx" reports, besides the first one.
Because this will be such a minimal program, if you do get problems,
the cause will probably be in hardware.

[cypher]

Also, interestingly enough, I do not see this intermittent problem if I just keep the program running forever. The problem only occurs when I introduce a hardware interrupt on the RB4-RB7 pins (interrupt on change). Is there something related to this interrupt that could cause a RESET_INSTRUCTION?

Also, for the while(1) infinite loop, I should put the printf statement displaying the result of restart_cause() inside the while loop right? If I just sit in a while loop forever, the terminal window would not display if the printf is not inside the loop.

[PCM programmer]

No. Just do this:

- call restart_cause().
- printf the result.
- wait in a while(1) loop forever. Don't do anything in the loop.

Restart_cause() is usually only valid if you read it at the very beginning
of main(). There are several posts on this in the archives.

The purpose of this test is to see if you are getting random resets just
while sitting in a loop, doing nothing. If you get a reset, you will get
another printf of the restart_cause() result. If you don't get a reset,
you'll get nothing.

[cypher]

Ok so tried doing that and I haven't seen any resets while just sitting in a loop....I only read a NORMAL_POWER_UP once and thats it. I've been running it for more than 30 mins and haven't seen a reset yet, so I think it's definitely something in my code that causes intermittenty resets.

I have another thought, don't know if it makes any sense though. So, like I said I have interrupts firing continuously on the RB4-RB7 pins and during these interrupts I have some printf statements I make to the terminal (not inside the interrupt sr but interrupt sr calls these printf). Now could it be possible that these intermittent resets are being caused since the printf are not executed all the time (I mean if an interrupt is called before all printing to the screen is done)?

I also do disable my int_rb manually in my isr and enable & clear it only when I'm done doing the actions (printf etc)....

[PCM programmer]

Are you enabling Global interrupts inside the isr ? If so, that could
definitely cause the problem. Delete that code, if you're doing it.

[cypher]

Yes, I :

[PCM programmer]

[cypher]

Thanks a lot for your help, it seems to be fairly robust now. So, removing the enable global from the isr definitely did it!