int_RDA and 18F6680

Markdem · Joined: 24 Jun 2005 Posts: 206

Hi Everyone,

Lost a bit of hair over this one.
Below is a very simple bit of code,

Ttelmah · Joined: 11 Mar 2010 Posts: 19433

Critical thing. INT_RDA, _must_, repeat _must_ read the serial data. If it doesn't, the PIC _will_ be hung when three characters are received.

However E0, should go high, but the E1 cycling would stop.
However though, the E0 pin defaults to waking up setup for PSP bus operation, so you need:

setup_psp(PSP_DISABLED);
setup_external_memory(EXTMEM_DISABLE);

at the start of main, or there is no guarantee, that you can toggle E0....

Best Wishes

Markdem · Joined: 24 Jun 2005 Posts: 206

Hi Ttelmah,

Thats what I thought about reading the RX buffer. I just wanted to make the test as simple as I could. In saying that, however, there is nothing that could send any data to the pic. This happens even if I disconnect the MAX232, so I can't see the buffer overflowing.

I will give the two lines of code a try tomorrow when I get to work. Might point me into the right direction.

Thanks

Ttelmah · Joined: 11 Mar 2010 Posts: 19433

If you disconnect the MAX232, you need to pull the RX line high at the PIC, and then pulse it low for a moment to test. Otherwise, if it floats low, then the buffer will overflow, while if is stays high, the interrupt will never trigger.
Obvious comment at this point, have you actually verified that you are receiving/sending data. Simple getc/putc code, rather than interrupt driven.

Best Wishes

Markdem · Joined: 24 Jun 2005 Posts: 206

Hi Ttelmah,

Yes, if I have a simple kb_kit - getc, it will work fine. I can then print the data back to the PC. I am 99.99% sure that the hardware is OK.

I have now given this a test. I tied the RX pin high. PIC will still lock up! If I do the same thing with enable_interrupts(int_rda) commented out, the PIC runs fine.

I am still waiting on a new 18F6680. I can't see how it could be doing this if it is faulty, but I am running out of options.

Thanks

Markdem · Joined: 24 Jun 2005 Posts: 206

Right, I have now lost ALL my hair over this.

New PIC18F6680.
Disabled int_rda = no problems
Enable int_rda = lockup.

I have found somthing new with this one. If I put my finger on the MCLR pin, the PIC will come back alive!
I have tried to put it high, but it does not help. I also tried to pull it low, but that did not help.

Why would the MCLR pin have any effect on int_rda?
Anyone with any ideas? I would put a diffent PIC into the project, but the other CAN pics have different pinouts.

Thanks

Ttelmah · Joined: 11 Mar 2010 Posts: 19433

It shouldn't.
_But_ MCLR, is (almost?) unique on the PIC, on having no protection if it goes above the supply rail. If it does, even fractionally/momentarily, it can trigger all sorts of odd things.
When you 'pull it high', what sort of resistance is involved?.
What is the electrical environment like close to the PIC?.

Best Wishes

Markdem · Joined: 24 Jun 2005 Posts: 206

HI Ttelmah,

When I pull up I used a 4k7 resistor.
There is nothing around my bench that could be affecting it in any way. VDD is coming from a bench power supply.

I have now thrown in the towel. chucked in a PIC18F4680 with some modifications to the board (was not a hard as I thought). Nothing got changed around MCLR, just some output pins for the LCD.
Works like a champ even when I enable int_rda.
I would love to know what the hell was going on with the 6680, but I just don't have the time to keep playing with it.

I hope this helps anybody else that come across something like this.

Thanks for all the help guys.

Mark