CCP1, CCP2 interrupts

[Steve]

Hello,

I am using the Both CCP modules in the capture mode to measure frequency on a PIC18F2410. I Have CCP1 set for Timer1 and CCP2 set for Timer3.

SETUP_CCP1(CCP_CAPTURE_RE);
SETUP_CCP2(CCP_CAPTURE_RE|CCP_USE_TIMER3);

I have CCP2 set for the RC1 pin

#FUSES HS,NOWDT,CCP2C1

When I see a pulse on the input pin I go to the interrupt routine and grab the time measurement. Then reset the timer to zero for the next measurement.

#INT_CCP2
void ccp2_isr()
{
time2 = CCP_2;
SET_TIMER3(0);
flips2=rollover2;
rollover2 = 0;
}


#INT_CCP1
void ccp1_isr()
{
time1 = CCP_1;
SET_TIMER1(0);
flips1=rollover1;
rollover1=0;
}

The rollover is coming from the Timer interrupt routine

#INT_TIMER1
void timer1_isr()
{
rollover1++;
}

#INT_TIMER3
void timer3_isr()
{
rollover2++;

}

My problem stems from only having one input. I place a square wave at one of the inputs and although the correct channel reads the correct frequency, the other interrupt routine is still being called. I know this because on a break, there is a value in time2, which is not set anywhere else. I know there are no other pulses coming in because I have grounded the input. If I comment out the ENABLE_INTERRUPTS(INT_CCP2); time2 will stay at a value of 0x00. Will an interrupt on any CCP module fire both interrupts? any help would be appreciated. I am basically trying to add two frequencies and with only one input the total frequency is messed up.

Thank you,

Steve

[PCM programmer]

Can you clarify what you're saying in your last paragraph ?
You keep referring to "one input". Do you mean one input signal ?
Or do you mean one input pin on the PIC ? Just re-phrase the whole
paragraph.

[Steve]

Sorry about that.

The object of the code will be to add two frequencies. Each frequency will be put through some circuitry so that the micro pin will have levels that it can handle. I started out with just the CCP1 module to see if I could get the micro to correctly read the frequency; which it did. I then tried to copy that using the CCP2 module of the micro. Once again it worked (I had shut down the CCP1 module while testing this half). The next step was to have both modules active.

With a frequency injected into the CCP1 module, the frequency is read correctly but the CCP2 interrupt is firing multiple times. I can tell this because the interrupt routine is the only place time2 is set with a value. Also, if I comment out the enable interrupt for CPP2, the time2 value stays at 0x00. Now I thought maybe the frequency was getting to the CPP2 input pin so a grounded it. I still had the multiple interrupts. If I switched the inputs around so that the frequency was at the CCP2 input and the CCP1 input was grounded, I get a good frequency calculation in time two and multiple interrupts on CCP1.

So I guess my question is do the CCP modules work in conjuction; that is if one is interrupted do both flags get set and both interrupt routines need servicing?

[PCM programmer]

Can you post a small, but complete test program that shows the
problem ? I want to see the lines of code for the CCP setup.
Make sure the program is complete. It should be "drop in" compilable
into an MPLAB project. ie., with all necessary #include, #use, #fuses
statements.

Also post the version of your PCH compiler. This will be a number such
as 3.191 or 3.239, etc. It's at the top of the .LST file which is in your
project folder.

[Steve]

Sorry for the time delay. This is a side project that I had not been able to get back to for the weekend. Here is the code, I believe this will compile but I did not actually compile it. I have multiple files and I did a cut and paste job.

#include <18f2410.h>
#include <math.h>


#FUSES HS,NOWDT,CCP2C1

#use delay (clock=4000000)

#define BASE 0.000001 // time base for a 4Mhz crystal
static int1 flag;
static int1 output;
static int rollover1, flips1;
static int rollover2, flips2;
static int1 measure;
static long time1, time2;
static float freq1, freq2, total_freq;
static float remainder;



static int ready;
static float useconds;
static long time;


#INT_CCP2
void ccp2_isr()
{
time2 = CCP_2;
SET_TIMER3(0);
flips2=rollover2;
rollover2 = 0;
}


#INT_CCP1
void ccp1_isr()
{
time1 = CCP_1;
SET_TIMER1(0);
flips1=rollover1;
rollover1=0;
}



#INT_TIMER0
void timer0_isr()
{
output = !output;
if (output)
OUTPUT_HIGH(PIN_A0);
else
OUTPUT_LOW(PIN_A0);
SET_TIMER0(time);
}

#INT_TIMER1
void timer1_isr()
{
//zero_flow1=1;
rollover1++;
}

#INT_TIMER2
void timer2_isr()
{
ready++;
if (ready == 30)
{
measure = 1;
ready = 0;
}
}

#INT_TIMER3
void timer3_isr()
{
//zero_flow2=1;
rollover2++;

}



void main(void)
{
SET_TIMER0(0);
SETUP_TIMER_0(RTCC_DIV_1|RTCC_INTERNAL);

SET_TIMER1(0);
SETUP_TIMER_1(T1_DIV_BY_1|T1_INTERNAL);

SET_TIMER2(0);
SETUP_TIMER_2(T2_DIV_BY_16, 255, 16);

SET_TIMER3(0);
SETUP_TIMER_3(T3_DIV_BY_1|T3_INTERNAL);

SETUP_CCP1(CCP_CAPTURE_RE);
SETUP_CCP2(CCP_CAPTURE_RE|CCP_USE_TIMER3);

// turn on the pullup for the sensors
// OUTPUT_LOW(PIN_A4);
// OUTPUT_LOW(PIN_A5);

// turn on the 2,5 Volt Level for the sensors
OUTPUT_HIGH(PIN_A2);
OUTPUT_HIGH(PIN_A3);





// ENABLE_INTERRUPTS(INT_TIMER0);
ENABLE_INTERRUPTS(INT_TIMER1);
ENABLE_INTERRUPTS(INT_TIMER2);
ENABLE_INTERRUPTS(INT_TIMER3);
ENABLE_INTERRUPTS(INT_CCP1);
ENABLE_INTERRUPTS(INT_CCP2);
ENABLE_INTERRUPTS(GLOBAL);




while (1)
{
if (flag)
OUTPUT_LOW(PIN_A1);
else
OUTPUT_HIGH(PIN_A1);
if (measure)
{
//freq1 = Calc_Freq1();
freq1=1/(((65536 * flips1) + time1) * BASE);
flips1=0;
if (freq1 > 2000000)
freq1=0.0;

// freq2 = Calc_Freq2();
freq2=1/(((65536 * flips2) + time2) * BASE))
flips2=0;
if (freq2 > 2000000)
freq2=0.0;

total_freq = freq1 + freq2;

measure =0;
}//end of if measure

}// end of while
}// end of main

[PCM programmer]

What is your compiler version ?

[Steve]

I am sorry, I have been doing this posting on a different computer than the development is on. The version is: 3.218. Thanks for the help, as I have still not yet been able to progress past this problem.

[PCM programmer]

I don't have that version, so I can't look at the .LST file for it.

The implication of your post is that the CCP1 isr is being called.
But if you ground the CCP1 input pin, you still get the problem.
So some internal process is setting the CCPIF bit. The question is,
how is this happening ?

The data sheet offers a few suggestions:

[Steve]

Hello again,

THe only Setup calls are those listed in the first few line of main. As you suggest, I do not change prescalers anywhere on CCP1 or CCP2 or Timer1 or Timer 3, so I do not think that is the case. I do not have a TRIS staement for the CCP2 input, so I think that is what I will try first. If I set all the inputs/outputs of the port maybe it will point me in a direction. The tris statement seems to be a confusing thing, depending on who I talk with. Some say it is out of date and the compiler should set the port pins as needed and some use the Tris statement in every project they have.

[PCM programmer]

The compiler does set the TRIS in the CCS functions. (PCH vs. 3.239)

[Steve]

After many frustrating hours, a friend has located an errata sheet. I had no idea about such a beast. In it, it states that a breakpoint can cause unwanted interrupts. Well it seems that this is the case in my scenario. If I free run the instrument, it appears to be working correctly, when I attempt to actually read the internal variables by breaking at a certain point, interrupts fire and cause values to be written in certain variables which makes no sense. I guess I will have to take into account these strange occurances when verifying my code. Anyway thanks to all for the help.

[ckielstra]