Code runs in debugger but not standalone

[hucker]

I have a project running on a 16F887 processor using MPLAB as my build/debug environment. In this environment the code is running fine however if I disconnect the ICD3 and try to run the programmed code the device does not start up...I have an LED that lights very early in the boot and it never flickers.

I have checked this FAQ

http://www.ccsinfo.com/faq.php?page=standalone

Since I build from the MPLab environment it is not clear what I need to do to run with regard to disabling the debugger for PCW. I do have the debugger disabled in PCW and have selected the release build option in MPLab.

here is my #FUSE code

[temtronic]

When using MPLAB , the 'build configuration' must be set to 'release' and not 'debug'then recompile then burn the PIC.

My install of MPLAB has 'release' permanently selected as I never use 'debug'.

Jay

[PCM programmer]

[hucker]

When I build for release I do comment that line out.

When I compile from MPLab, I am using the CCS compiler using the plugin. I have never enabled the debugger in CCS (e.g. no debugging tools are selected from the "Enable" button on the debug screen.

[Ttelmah]

The obvious comments/questions are:

What is on the MCLR line?. Needs something to pull it up when the ICD is not connected.
Look carefully at the ground and PSU connections. Remember the ICD will connect these to the chip.
Remember you have to compile the code without DEBUG & load this into the chip, in 'programmer' mode. The code put into the chip when ICD is enabled will not run without the ICD connection.
Be aware that MPLAB always defaults to compiling with DEBUG enabled, even if a debugger is not selected. You have to explicitly select 'release' for it to not do this.

[temtronic]

MPLAB defaults to 'debug' mode for the 'build configuration'. It has to be set to 'release' in order for the code to be correctly compiled to burn a PIC for Real world use.

Jay

[PCM programmer]

He said that in his first post. He said:

[Ttelmah]

I was slightly 'dubious', since he seems to think that not having a debugger disabled it. Not true.

[hucker]

I have been setting release mode and changing the defines around when I build. I believe this is not related to the programming environment at this time.

It does appear that this is an MCLR issue. When the board was redesigned (16F877 to 16F887) the mcrl pullup with diode protection was not included (not sure why). Thanks Ttelmah!

I notice that there are MCLR and NOMCLR fuse settings for the 887 while the 877 does not have these fuses. Can this compiler fuse setting be used to get the PIC to boot from code or do I need to deadbug the pullup/diode? (the MCLR pin is only used for programming in our system).

[hucker]

There are a lot of things going on here and everybody seems to have opinions, some of them "dubious", as to how programming and debugging work using MPLab. Here are a few points worth noting:

I am using MPLAB 8.92 and have it configured to use the CCS compiler using the MPLab integration tool from CCS. Under this configuration I have discovered the following:

1) There is a release/debug option on the MPLab compiler.
2) There is a code directive #device ICD=True (or commenting out this line since ICD=FALSE doesn't work)

I can use ANY combination of these 2 (4 options total Release/True,Release/False, Debug/True,Debug/False) and get a hex file to load and run standalone (e.g. ICD disconnnected)...AND I can use any combination of those and have the micro fail to start.

There is a third variable. As far as I can tell if you program your hex file from the Debugger Menu you will always get an install that won't boot without the ICD AND if you program a hex file using the Programmer Menu you will always get an install that will boot REGARDLESS of the above settings.

[Ttelmah]

A lot of this depends on which version of MPLAB is involved. There have been versions that did not work at all properly in this regard. Even one that wouldn't switch out of debug mode whatever you did....