How to make RTCC more accurate?

[Analyzer]

Hi,

I am a newbie about rtcc interrupt and i referenced ccs example program.This code (below) works but it is not accurate.Because 20M/(4*256*256) = 76,2939453125 and INTS_PER_SECOND is 76! So how can i make it the most accurate? How can i use it?

Thanks.

Analyzer.


#define INTS_PER_SECOND 76 // (20000000/(4*256*256))

byte seconds; // A running seconds counter
byte int_count; // Number of interrupts left before a second has elapsed


#int_rtcc // This function is called every time
void clock_isr() { // the RTCC (timer0) overflows (255->0).
// For this program this is apx 76 times
if(--int_count==0) { // per second.
++seconds;
int_count=INTS_PER_SECOND;
}

}


void main() {

byte start;
byte c;

int_count=INTS_PER_SECOND;
set_timer0(0);
setup_counters( RTCC_INTERNAL, RTCC_DIV_256 | RTCC_8_BIT);
enable_interrupts(INT_RTCC);
enable_interrupts(GLOBAL);

lcd_init();
lcd_clear();
c = 0;

do {
lcd_gotoxy(1,1);
lcd_putc((seconds-start)+48);

} while (TRUE);

}

[MGP]

Check out the algorithm here:

http://www.romanblack.com/one_sec.htm

Works great!

[Analyzer]

Thank you for your great help.But i can not understand asm code and i am too lazy to get theory and write a new code

Anywayz, if somebody has an instant ccs c code, please write it here

Analyzer.

[Ttelmah]

[Mark]

Can you use another timer? Do you have timer1 available? Do you have a CCP still free? If so, respond back and I will show you a way using a CCP int. Another option is to drive timer1 with an external clock. One of those nice frequency values that RJ was referring to.

[Analyzer]

Thanks for your help.It is a good idea to change xtall and get nicer values but i have many 4 mhz xtall.I need to use them.

My all timers are free now i only use that code above.I'm using 877 for this.I'm really curious on that "another way"

Analyzer

[Guest]

[Hans Wedemeyer]

I don;t have a PIC16 setup at the moment, but form memory this should give a 1 second tick.
hansw

#include <16F877.h>
#device *=16
#device adc=8
#use delay(clock=4000000)
#fuses NOWDT,HS, PUT, NOPROTECT, BROWNOUT, NOLVP, NOCPD, NOWRT, NODEBUG
#use rs232(baud=9600,parity=N,xmit=PIN_C6,rcv=PIN_C7,bits=9)

int32 Seconds;
int OneSecondFlag;
int Tick;

/////////////////////////////////////////////
// Using Timer 1 to interrupt every 100mS
// Timer 1 is a 16 bit timer.
// using a 4MHz xtal and Timer1 Prescale = 8
// Timer1 counter sees 8
// for 100mS preload 15536 or 65536-(0.1/(8/4000000))
#int_TIMER1
TIMER1_isr()
{
Set_Timer1(15536);

Tick++;
if( Tick >= 10 )
{
OneSecondFlag++;
Tick=0;
}
}
////////////////////////////////////////////
//
void main()
{
Tick=0;
Seconds=0;
OneSecondFlag=0;

setup_adc_ports(NO_ANALOGS);
setup_adc(ADC_OFF);
setup_psp(PSP_DISABLED);
setup_spi(FALSE);
setup_counters(RTCC_INTERNAL,RTCC_DIV_1);
setup_timer_1(T1_INTERNAL|T1_DIV_BY_8);
setup_timer_2(T2_DISABLED,0,1);
enable_interrupts(INT_TIMER1);
enable_interrupts(global);


while(1)
{
if( OneSecondFlag > 0 )
{
Seconds++;
printf("%lu\r\n",Seconds);
OneSecondFlag=0;
}
}

}

[Mark]

Lets say you are using a 20MHz crystal. The internal speed is 5MHz. This equates to timer1 incrementing every 0.2us with no prescaler assigned. Now we are going to use the CCP1 interrupt to keep track of our time. We will actually be counting 100ths of a second for this example. 100th of a second is 50,000 timer1 ticks (0.01/0.0000002=50000). This will be our CCP1 value.

[dyeatman]

All this to avoid buying a 90 cent crystal.... what a waste of time......sigh...

At some point you can cut your losses and save your self a LOT of time!

I had lot of 4MHZ stuff too. But I bit the bullet and started with about 20 of the half can 18.432MHZ oscillators for $1 apiece.

The benefits:
They are very stable and reliable in all environments.
They give me exact RS232 speeds and frequency divides
No Oscillator startup/loading hassles
Less component count
Uses about the same real estate.
Oscillator does not have to be as close a possible to the PIC.
Can be socketed and the frequency changed in seconds. (I always use sockets)
etc...etc...

I have used them in all my designs ever since. The 4MHZ xtals/capacitors and resonators are still sitting in the drawer and suppose they always will be....

something to think about...

[Analyzer]

Thank you for all your replies.I especially want to thank to code senders.I will try all codes and sim in proteus.I agree to buy new xtalls, i'm totally aware of all benefits but this will be a part of production and i have MANY 4 mhz xtalls and want to spend them.
I will test all and report here.

Thank you again.

[KerryW]

#define INTS_PER_SECOND 76 // (20000000/(4*256*256))

byte seconds; // A running seconds counter
byte int_count; // Number of interrupts left before a second has elapsed
int32 count;

#int_rtcc // This function is called every time
void clock_isr() { // the RTCC (timer0) overflows (255->0).
count = count + 65536;
if(count>5000000)
{
seconds++;
count = count - 5000000;
}

// For this program this is apx 76 times
//if(--int_count==0) { // per second.
//++seconds;
//int_count=INTS_PER_SECOND;
}

}


void main() {

byte start;
byte c;

int_count=INTS_PER_SECOND;
set_timer0(0);
setup_counters( RTCC_INTERNAL, RTCC_DIV_256 | RTCC_8_BIT);
enable_interrupts(INT_RTCC);
enable_interrupts(GLOBAL);

lcd_init();
lcd_clear();
c = 0;

do {
lcd_gotoxy(1,1);
lcd_putc((seconds-start)+48);

} while (TRUE);

}