Wierd port C behavior

Guest · Guest

Something funny is happening with my portc on a 16f876.

Using output_c(0xff) all the pins go high.
When I use output_high(PIN_C0) C0 goes high as expected.
When I use output_high(PIN_C0) followed by output_high(C1) only c1 goes high.

Any ideas? Nothing else exists in the program, just those lines mentioned above in main. I tried turning off ccp1, ccp2, and timer1 but result is the same.

Guest · Posted: Fri Jan 30, 2004 4:35 am

Hello

I had this problem on a PIC12F675. I disabled the comparator module completely, and everything worked fine. I don't remember if the 876 has comparators. I it hasn't, there is maybe another hardware module at this pin.

mfg
Felix

Guest · Guest

Felix,

Good guess but that isn't the problem. The analog stuff on a 876 is on portA. The only thing on c0 and c1 is timer1 and ccp2 respectively, of which I have already tried manually disabling.

I have a feeling port c may need some special sort of initalizing when not using the TRIS command. I did some quick and dirty assembly as suggested by someone..

Ttelmah · Guest

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

I tested the following program with PCM vs. 3.184, and it
works OK. Both pins C0 and C1 are set high, as observed
with an oscilloscope.

#include "16F877.h"
#fuses XT, NOWDT, NOPROTECT, BROWNOUT, PUT, NOLVP
#use delay(clock = 4000000)

main()
{

output_high(PIN_C0);

output_high(PIN_C1);

while(1);
}

Guest · Posted: Fri Jan 30, 2004 6:58 pm

Thanks for letting me know about the read-modify-write issue.

Although, I don't think timing was my issue as observed by the following code that didn't work:

output_high(PIN_C1);
delay_ms(1000);
output_high(PIN_C0);

Only C0 remained lit (I had small leds directly connected to each pin). I rememberd the mention of current draw in the microchip website, so I added limiting resistors in series with each led (even though I know the leds weren't drawing more then the rated current per pin), and the above code worked even without the delay (@20mhz).

So I guess drawing too much current (or too fast) from any port pin, even if the others are stable, will lead to issues. Well, you learn something new every day.

Charlie U · Posted: Fri Jan 30, 2004 9:59 pm

Now that we know "the rest of the story" it is fairly easy to surmise that, indeed you _are_ seeing a "read-modify-write" problem. The "small" LEDs that you had on the port pins, without current limiting resistors, were limiting the output voltage to somewhere between 2 and 3 volts, because they are diodes. The inputs to port C are Schmitt triggers and require 0.8 x VDD to be high, which is 4 VDC in a 5 volt system. Therefor, even though you have set the outputs to a high state, the output drivers can not pull the outputs above the diode operating voltage of 3 volts max (estimated), which is below the input threshold of 4 volts, the pins will be read as a 0 during the read modify write action of the bit set, and driven back as a 0.