Interrupt problem with 12F675

[toddintr]

Hi,

First off: I am new to CCS! I am using PCWH 3.210 on Win XP. I am trying to program a PIC12F675 through a PICkit 1 board.

I have the following code:

[RKnapp]

Todd,

I'm just starting on my morning coffee so this is sort of a dumb question, but I can't quite see how you'd "lock out" the state change resulting in an interrupt if a person sits on the button.

Pushing a button to us seems fast, but not for these chips. I guess I'm wondering what happens if a button press slams you into the ISR, you perform some action, but then, the button remains pressed for hundreds of milliseconds before one can remove the finger. Does one keep re-entering the ISR, or is the interrupt somehow masked until you let go?

(In POLLING a button, it is easy to construct software that forces one to release the button before the system will listen to the next button press. But the problem is more intricate when interrupts are involved.)

(I am having a problem on some connected 8722s in which an interrupt is serviced a few times just fine, but then stops working. I am apparently not servicing it correctly, so I'm interested in all posts with interrupt problems.)

One of my problems is that I don't know what the little patch of assembly does in your init() routine.

Not much help, I'm afraid. But I urge you to think carefully about the "sit on the line" problem. There may well be a difference between the HiTech compiler and CCS in the "overhead" code for interrupt servicing. HiTech, for example, may enter the ISR and immediately mask the B0 interrupt until one exits the ISR. You appear comfortable in assembly, so you could check out the differences in entry/exit code.

Good luck,

Robert

[PCM programmer]

You're using #INT_EXT, but according to other things in your code,
you want to use GPIO change interrupts. Those are two different
types of interrupts. Look near the end of the 12F675.H file, and
you'll see a list of all the interrupt constants that can be used with
that PIC. You likely want to use INT_RA or one of the pin specific
versions of that interrupt. Look at the list. Change your setup
code, and the isr.

Also, you're doing a lot things that are not needed in CCS. All
the interrupt detection (ie., checking flags, etc.) is automatically
done for you by the compiler. You don't need that stuff in your isr.
Also, if you're using interrupt-on-change, you need to read the GPIO
port in the isr, to clear the "change" condition. See section 3.2.2 of
the 12F675 data sheet.

The compiler also will take care of reading the OSCCAL value and
updating the register. You don't need that ASM code in there.
Basically, if you used CCS standard i/o mode, you could reduce the
total line count in your program to at 25% or less, of your current
program size.

[toddintr]

PCM programmer,

THANKS! Per your comments, I changed #int_ext to #int_ra, which immediately solved the problem and enabled the pushbutton. The program now works as expected.

I realize that my code uses the CCS compiler in the least efficient manner (directly twiddling bits and registers). However, I haven't come across documentation that explains the effective methodology for using CCS C. The CCS C Reference Manual has hints in it, but not the big picture.

I know there are the examples that come w/ the code, plus there is this forum, but a central tutorial that explains the CCS way of doing things would be most welcome. (Otherwise, the true strength of the CCS compiler may not be understood. Just some feedback from a customer.)

Regards,

Todd

PS: You are doing an EXCELLENT job. Everyone should be so lucky to have their question answered by PCM programmer!

[toddintr]

Robert,

The assembly code retrieves the calibration value from the last position in memory (where it is stored when leaving the factory) and uses it to set the chip calibration register.

Thanks for your response,

Todd