BUG REPORT: 4.114

59seconds · Joined: 30 Dec 2010 Posts: 4

Compiling for the 24fj256da210 (with or without debug set), will generate configuration words in the wrong locations (reversed order).

Way to reproduce:
Compile simple 'hello world' with various fuses set. Import this hex file into MPLAB and under the 'Configure' -> Configuration bits, you can clearly see that the fuses are set incorrectly.

A workaround is to edit the table of the device (editing of table only works when NODEBUG fuse set).
Config words should be reorganized in this order:
1->4
2->3
3->2
4->1

Not sure if this is fixed with latest 4.116. Table 27-1 in device datasheet shows correct config address locations.

EDIT: There could be more bugs, this is just the first one I found trying to get this part working. Currently I'm not even able to output anything on the UART, but I can toggle a pin.

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

Sorry for confusing the discussion. The below statement is wrong. The bug is in PCD, and it's fixed in V4.116.

I briefly checked the issue. Assuming the present PIC24FJ256DA210 datasheet is correct, then the bug is in MPLAB, not in CCS C. I use MPLAB V8.53.

I guess, you noticed that the addresses are reversed related to the configuration word numbering. Basically, 24FJ256DA210 uses the same configuration word layout as other PIC24 chips I'm using, e.g. PIC24FJ256GA110. I see, that the configuration is correctly set in the *.hex file. If I read the DA210 *.cof file to a GA110 configuration in MPLAB, the configuration words are assigned correctly. If I select DA210, they are wrong. May be a newer MPLAB version fixes the problem.

I'm not using the new PIC24 chip and don't feel a need to investigate further on this issue now. In any case, you should contact CCS.

59seconds · Joined: 30 Dec 2010 Posts: 4

Thanks for verifying it. Do you know if 4.116 also fixes the UART? It's broke to high-heaven in 4.114. (with the da210 chip anyways).

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

I'm using V4.112 for my active projects, but it misses correct support of 24fj256da210. I didn't yet notice UART issues with V4.114 or V4.116. Can you post an example code that produces the problem?

59seconds · Joined: 30 Dec 2010 Posts: 4

Well Gosh dog it... it's working now, but I promise something funny was going on last week. After I solved the config word problem, I started on the Uart problem, and over the weekend I told myself to check that 0x8400 being written to the U1MODE (The '4' is a non-location, or perhaps even reserved). I also noticed you have to use UART1A in the #use rs232(*) or else it'll not use Uart1.

I've got some more boards to bring up. I see if I can reproduce the error.

59seconds · Joined: 30 Dec 2010 Posts: 4

O.k., I figured out what is breaking the UART. Whenever a delay is used, it never gets past the delay.

FvM · Joined: 27 Aug 2008 Posts: 2337 Location: Germany

MPLAB SIM executes the code correctly, when compiled with V4.116. If it's not working in the real hardware, you should trace the execution with a debugger.