Naming IO Pins

bob_4610 · Joined: 30 Nov 2003 Posts: 3 Location: Beervelde (Belgium)

I have an application that uses several IO pins to drive various leds. The pins are scattered over ports A, B & C (PIC16F73) because I use special functions like I2C, ext. IRQ etc... so some pins are not available.

To ease addressing I tried to use following approach :

Define pin addresses as an array

const int leddrv[6] = {pin_a1, pin_a2, pin_b2, pin_b4, pin_c0, pin_c2};

with the idea of using :

output_high(leddrv[var]) or output_low(leddrv[var])
However, this only works if var is a constant.

But if I want to use a variable e.g. :

for(i=0;i<=6;++i) output_low(leddrv[i]);

I get a compiler error "Expression must evaluate to a constant"

Does anyone know if there is another way to do this?
Regards,
Bob

rwyoung · Joined: 12 Nov 2003 Posts: 563 Location: Lawrence, KS USA

This may not be the cleanest implementation but I believe it will work. It does compile (V3.184). Based on the suggestion found in the FAQ titled:

"How can I pass a variable to functions like OUTPUT_HIGH()?"

bob_4610 · Joined: 30 Nov 2003 Posts: 3 Location: Beervelde (Belgium)

hi Rob

Thanks for your reply. This is very similar to the way I did it right now. However since there are 13 leds in total, the switch/case statement is quite extensive. Also, each led has 3 states (on, off, blink). The blinking is achieved by setting a bit in an int16 (led_blink_state). When a bit is set, the corresponding led is toggled at a preset blinking frequency.

That blinking function is again quite long since I can't pass a variable to output_high(pin#) and output_low(pin#). So I have to write the blinking code 13 times.

That's why I was looking for a way to pass variables to these output functions, but I'm afraid I'll have to stick to the current solution.

Anyway, thanks for your suggestion.

Regards,
Bob

Mark · Joined: 07 Sep 2003 Posts: 2838 Location: Atlanta, GA

This is how I do it. You can use a similar approach and overlay the union on the ports. Mine are actually on 6 pins of one port and then multiplexed. If you look at the LCD thread that was just recently brought up, you will see this technique used in the lcd driver file that CCS supplies.

bob_4610 · Joined: 30 Nov 2003 Posts: 3 Location: Beervelde (Belgium)

Thanks for the suggestion, but it has the same disadvantage as what I'm using now, i.e. you need as many "switch/case" or "if's" as there are I/O pins involved since you must use constants.

What I had in mind was something like this, but I guess to achieve this I will have to write my own I/O driver as you suggested.

rwyoung · Joined: 12 Nov 2003 Posts: 563 Location: Lawrence, KS USA

Note about enums and switch/case:

It is somewhere on the old forum but I had a discussion with RJ Hamlett about how CCS was generating the switch/case statements. We kept disagreeing about how it was generated and it turned out that there was something about the use of enums as the case arguments versus the use of a hard coded integer or #define. I think the code was tighter when #define'd values were used as the case arguments.

Just for the heck of it you might try a quick test of your switch/case statements where you try enum'd values, #defines, and hard coded constants. Also the presense of a "default" case makes the code larger if I remember correctly.

Look at the LST file and see how the switch/case code is generated under different condtions.
_________________
Rob Young
The Screw-Up Fairy may just visit you but he has crashed on my couch for the last month!