Timer issue

[Ioannis]

I am not getting the time I expect from the following Interrupt Driven timer.

Controller is 16F887,internal OSC at 8MHz, expected interrupt every 25usec and I get every 101usec.

If I change the reload value to 63, the expected interrupt is 100usec and I get 146usec.

A check on the IntOSC with Clock out on RA6 prooves that the Clock is correct @ 2MHz (8MHz/4).

I tested also with the Timer2 w/o pre-post scalers with exactly the same results.

Here is the test code that is making me nervous:

DEFINE OSC 8

OSCCON = %01110001 'SET SYSTEM CLOCK SWITCH BIT

@Line1 = _DEBUG_OFF & _LVP_OFF & _FCMEN_OFF & _IESO_OFF & _BOR_ON
@Line2 = _CP_OFF & _MCLRE_OFF & _PWRTE_ON & _WDT_OFF & _INTRC_OSC_NOCLKOUT

@ __CONFIG _CONFIG1, Line1 & Line2

@ __CONFIG _CONFIG2, _WRT_HALF & _BOR40V

PORTA=0:PORTB=0:PORTC=0:PORTD=0:PORTE=0

TRISA=%00101101
TRISB=%00000000
TRISC=%10000000
TRISD=%00000011
TRISE=%00000000

OPTION_REG=00001111

WPUB = %00111111

ADCON0 = %11000001
ADCON1 = %00000000	'Set PORTA/E analog/digital, right justify result
ANSEL  = %00101101  'lower port A as Analog input
ANSELH = %00000000  'all others Digital inputs

DEFINE ADC_BITS 10  'Number of bits in Adcin result
DEFINE ADC_CLOCK 3  'ADC clock source (rc = 3)
DEFINE ADC_SAMPLEUS 50
                     
pwm1            var portc.2

INCLUDE "DT_INTS-14.bas"     ' Base Interrupt System
INCLUDE "ReEnterPBP.bas"     ' Include if using PBP interrupts

'**************************************************************************
'****
'****                            Initialization
'****
'**************************************************************************

't2con = %0000100    ;Fosc/4, no prescaler, no postscaler

't1con = %00110001   'Enable Timer1 (16bit) /8 prescaller, int. Fosc/4 clock
                    '@ 20MHz times out every 104,86ms
':::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
ASM
INT_LIST  macro    ; IntSource,        Label,       Type, ResetFlag?
			INT_Handler   TMR0_INT,  _UpdateTimer,   PBP,  yes
;			INT_Handler   AD_INT,    _ADC_Sub,       PBP,  yes
	endm
	INT_CREATE               ; Creates the interrupt processor
ENDASM

':::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::       

@ INT_ENABLE  TMR0_INT      ; enable Timer 1 interrupts
;@ INT_ENABLE  AD_INT        ; enable ADC Interrupt

CLEAR

loop1:
pause 100
goto loop1

UpdateTimer:                        'This interrupt should happen every 25usec :(
    tmr0=213
    toggle portc.2
@ INT_RETURN

I miss something here but ... what?

Ioannis

[mackrackit]

What happens if you make it an ASM interrupt instead of a PBP int.?

[rsocor01]

Ioannis,

Try changing the PAUSE 100 to PAUSE 1. I read somewhere that using Pause 1 instead of a bigger number is better for the interrupts. Give it a try and let me know.

Robert

[Ioannis]

Thanks Dave.

Hmm, almost half.

PBP int., 207 reload, 101 usec
ASM int., 207 reload, 41 usec

But according to the calculations, 8MHz clock/4=2MHz or 0,5usec clock to TMR0. So a 207 reload should give about 24usec.

Instead we get 41 or 101 usec, according to selected Interrupt.

And this leads me to ask, why/when to use ASM instead of PBP as long as I use PBP in the ISR?

@Robert. This is true for ON INTERRUPTS of the Compiler. Here I use DT_INTS14, so this is of no importance. Thanks anyway.

Ioannis

[Ioannis]

OK, after some tweaking, and using ASM as TYPE (Don't know why yet), is getting correct results up to a reload value of around 180.

The changes to get close are:

@ INT_DISABLE TMR0_INT
    TMR0=TMR0+180
    toggle pwm1
@ INT_ENABLE TMR0_INT

By adding the value,we have more precise control of the reload value.

But not close to my target of 25usec

Ioannis

[mackrackit]

Only guessing here, but I think Darrel ment the ASM type when there is ASM code in the ISR. I have noticed better accuracy with timers just by changing the type though.

Now that I read this page again about the underscore, I may be doing something wrong. http://darreltaylor.com/DT_INTS-14/asm_ints.html

Hmmm. Always something interesting..

[Ioannis]

There are some shadows on the DT-INTs routines as there is no Manual. I am willing to help writing a manual document whenever Darrel can provide more insides.

But for the moment the ASM idea seem to get closer to the target.

Ioannis