Not correct running without the ICD3 attached

Jody · Joined: 08 Sep 2006 Posts: 182

Hello,
I have a strange problem..
I compile my code and when using debugging it works correct.
I switch to release mode and compile and program the chip 18F87j50 and that is going well.
Remove the ICD3 and the code is running.. USB is running I see that is is connected to the PC and when I send a command it goes berserk. Starting random things..

When I connect the ICD3 (no MPLAB or what ever running) the PIC is behaving like I want...
And sometimes it is also behaving itself without the debugger... but not always.

Anybody any clue??????

Regards,
Jody

kWoody_uk · Joined: 29 Jan 2015 Posts: 47 Location: United Kingdom

Could be anything: your circuit design/layout, power rails etc.

However.. sounds like it could be an uninitialised variable in your code. Very easy to miss.

Keith

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

One possibility is grounding. The ICD3 could provide an adequate ground
when you have it attached, and when it's removed, that good ground
goes away.

Check the ground connections on your board. Possibly a cold solder joint ?

Ttelmah · Joined: 11 Mar 2010 Posts: 19539

Other possibility. What is on the MCLR line?.

Could be floating.

Jody · Joined: 08 Sep 2006 Posts: 182

Okee I have checked the MCLR.. there is is a pull up to VDD 50k.
All the ground resoldered but were looking good.

What I did and seems to help: - Cleared the USB buffer after powering and not start using it right away.. so far it looks like it is working...
Maybe when I power up garbage is in the USB buffer?
I keep on testing... but thanks so far for the ideas!!!

Jody

Ttelmah · Joined: 11 Mar 2010 Posts: 19539

Most likely the time delay.

With the ICD attached, the processor does not start running your code as fast. It wakes in the debug executive code, and this talks to the debugger for a moment, before the processor is then seen to actually start.

temtronic · Posted: Wed Jun 08, 2016 5:18 am

Don't know about this but...
if the program still has 'use ICD...' in the code, maybe that 'overrides' the MPLAB build 'release mode' command ?
One of those 'whose in control' issues.

I got into the habit of using 'zero RAM' after some 'quirky' operations 2 decades ago. Maybe I'm just paranoid due to my age, but I like a 'clean slate'.

Jay