PIC Problems 101

[wrocko]

I have a nice working circuit that I am utilizing to control a set of H-Bridges for control of a robot. Yet, the oddest thing happens. When I energize the system, the PIC does not begin its program right away every time. Many times i'll see a brief (1-2s) delay, other times it will never come on. Yet again, other times it begins right away. This circuit used to work right away, so I swapped out the PIC, and the same results occur. It doesn't seem like it could be my code since the results are almost random. I'll post what I am running though. Any ideas?


Thanks

#include "C:\Documents and Settings\Admin Bin\My Documents\ECE497\Code\motorcode.h"

// GLOBALS
int16 RRotations=0, LRotations=0, RClicks=0, LClicks=0;
char command = '.', P1=0, P2=0;


// PIN REDEFINITIONS
#define LBRAKE PIN_D0
#define LDIR PIN_D1
#define RBRAKE PIN_D3
#define RDIR PIN_D2
// Both CCP set to PWM with Duty from 0 to 500
#define MAXDUTY 480; // True max is 500
// Constants to normalize motors, not used yet
#define LMC 1.0
#define RMC 1.0

// ENCODER 1, not implemented
#int_RB
RB_isr() {

}

// ENCODER 2, not implemented
#int_EXT
EXT_isr() {

}

// EASE OF USE MOTOR FUNCTIONS
/*
DIR = 0 : Forward
DIR = 1 : Reverse
*/

inline void LeftMotor(int duty, short dir, short brake)
{
int dutycycle;

if(dir == 1)output_low(LDIR);
else output_high(LDIR);

if(brake == 1)output_high(LBRAKE);
else output_low(LBRAKE);

if(duty>100)duty=100;

dutycycle = (duty/100)*MAXDUTY;
set_pwm2_duty(dutycycle);
}

inline void RightMotor(int duty, short dir, short brake)
{
int dutycycle;

if(dir == 1)output_high(RDIR);
else output_low(RDIR);

if(brake == 1)output_high(RBRAKE);
else output_low(RBRAKE);

if(duty>100)duty=100;

dutycycle = (duty/100)*MAXDUTY;
set_pwm1_duty(dutycycle);
}

void main() {

setup_adc_ports(NO_ANALOGS);
setup_adc(ADC_CLOCK_DIV_2);
setup_spi(FALSE);
setup_psp(PSP_DISABLED);
setup_counters(RTCC_INTERNAL,RTCC_DIV_2);
setup_timer_1(T1_INTERNAL|T1_DIV_BY_1);
setup_timer_2(T2_DIV_BY_16,124,1);
setup_ccp1(CCP_PWM);
setup_ccp2(CCP_PWM);
enable_interrupts(INT_RB);
enable_interrupts(INT_EXT);
enable_interrupts(global);

// PORT SETUP
set_tris_a(0x00);
set_tris_b(0x00);
set_tris_c(0x00);
set_tris_d(0x00);
set_tris_e(0x00);

for(;;){ // START MAIN LOOP

LeftMotor(100, 1, 0);
RightMotor(100, 1, 0);
delay_ms(50);
LeftMotor(100, 0, 1);
RightMotor(100, 0, 1);
delay_ms(50);

} // END MAIN LOOP
}
_________________
- - - - -
-- Wrocko --
- - - - -

[PCM programmer]

1. What PIC are you using ?

2. What version of the compiler ?

3. What do you have connected to the MCLR pin ?

4. Post the first part of your program, that shows all
the pre-processor statements. (Including #fuses,
#use statements, etc.)

[Neutone]

Remove unused interupt enables that cause the processor to wake from sleep on interupt but no other function.

[fpgeh]

Check your circuit and power supply. Motors draw more current than most typical CMOS circuits. Is your supply voltage correct and noise free? Do you have solid grounding? Also, do you have the correct free-wheeling diodes to prevent inductive spikes at the H-bridge? (often they are built into the H-bridge). Large inductive spikes could easily damage a PIC.

[wrocko]

#include <16F877.h>
#use delay(clock=40000000)
#fuses XT,NOWDT
#use rs232(baud=9600,parity=N,xmit=PIN_C6,rcv=PIN_C7)

The chip is a PIC16F877, and i have a 1k resistor going to 5V on the MCLR pin. Compiler is V3.048

Its not the motors causing the problem, i removed them and just put logic probes on the I/O. It still reacts the same. I'll look into changing the loop type, but I have always used for(;;) with success in the past.

As to the interrupts, i'll give it a shot. Yet, if that is the problem wouldn't that happen when I do have code for them too? (I will need them for the motor encoders).

Thanks for the ideas.
_________________
- - - - -
-- Wrocko --
- - - - -

[PCM programmer]

#include <16F877.h>
#use delay(clock=40000000)
#fuses XT,NOWDT
#use rs232(baud=9600,parity=N,xmit=PIN_C6,rcv=PIN_C7)


1. The #use delay is set for 40 MHz. This should be changed to 4 MHz.
Example:

#use delay(clock=4000000)


2. The #fuses should likely have NOLVP added to them. (Unless you're
using a programmer that supports LVP mode, and 99% do not).
Example:

#fuses XT, NOWDT, NOLVP


Also, I normally use the Power-up Timer and Brownout Reset.
So this might be better:

#fuses XT, NOWDT, PUT, BROWNOUT, NOLVP

[ajt]

Also, if the crystal is > 4MHz (not sure as both 4MHz and higher are mentioned) the HS fuse should be used not XT. Possibly, the oscillator is not starting up correclty.
_________________
Al Testani

[wrocko]

Thanks for all the great ideas. The clock being set to the wrong delay was not the culprit, but was definetly something that needed to be fixed. I tried each idea seperately to see which would solve the problem.

It turned out to be the interrupts, once i added a single line of code to each, the program fired up right away every time.

Thanks again.
_________________
- - - - -
-- Wrocko --
- - - - -