DT Interupts PortA or PortB?

[ShaneMichael]

I need a little help, I'm using a 18F2550, and I'm counting how often a switch opens/closes on PortA.5 (Maybe 4, I don't have my circuit in front of me). I'm looping in my main code counting the changes, but I'd like to use DT's interupts, seems like a good fit. I'm using them already for comm and timing. It's working great. What I'm not finding in the posts I've read so far is when I create and enable "INT_INT" how do I control what port triggers the interupt? I'm already doing A/D conversion on PortA.0-3 And have things happening on PortB, so i don't want the changing values on the other pins to trigger the interupt, just a single pin if possible?

What I have works, but I think I may miss some state changes if the switch operates too fast.

Any idea's or suggestions would be great.
Thanks,
Shane

[Charles Linquis]

There is no hardware support for interrupts on PORTA. But that doesn't mean you can't use interrupts to do what you want.
A good way is to determine the shortest switch closure that you must detect (50mSec?). Then set up a timer that interrupts
at that rate. You mention that you are already using DT-INTs for timing, if you are using a timer for that, just "piggyback" on
that INT. If you aren't, just set up another timer (I like TMR3 for such things). The interrupt rate must be as fast as the shortest
switch closure you want to be able to detect. You ISR could be something like -

SwitchInt:  ; The ISR

IF PORTA.5 = 0 THEN
    IF SwitchHasBeenClosed = 0 then
            Pause 10;    Debounce
         IF PORTA.5 = 0 then   ; Still low?
            SwitchHasBeenClosed = 1
            SwitchHasBeenClosedFlag = 1
         ENDIF
     ENDIF
ELSE
    IF SwitchHasBeenClosedFlag = 1 then
         Pause 10    ; Debounce time
         IF PORTA.5 = 1 then  ; Still high?
             SwitchHasBeenClosedFlag = 0
             SwitchHasBeenClosedThenOpened = 1
         ENDIF
     ENDIF
 ENDIF
          
      

@ INT_RETURN

; You can clear the flags as appropriate such that you don't perform multiple operations with one switch push.

[ShaneMichael]

Charles, Thanks for the suggestion, that's working better. How do I make TMR3 faster? It seems to check the interupt at about 76 hz, I'd like to be around 350 hz? Is there a way to do it without messing up TMR1 that I'm already using? I'm figuring 350 hz because the switch will open/close at about 44hz, but it only stays closed for about 45 degrees or 1/8 of the time. So 44hz=22.7ms, 1/8 of 22.7ms = 0.00284 s, .00284 s = 352 hz. I'm using a 20Mhz XTAL, here are some of my other settings:

asm
          __CONFIG  _CONFIG1L, _PLLDIV_5_1L & _CPUDIV_OSC3_PLL4_1L & _USBDIV_2_1L
          __CONFIG  _CONFIG1H, _FOSC_HSPLL_HS_1H
          __CONFIG  _CONFIG2L ,_BOR_ON_2L & _BORV_2_2L
     
        
    endasm
    
    INCLUDE "DT_INTS-18.bas"     ; Base Interrupt System
    INCLUDE "ReEnterPBP-18.bas"     ; Include if using PBP interrupts
    INCLUDE "Elapsed_INT-18.bas"  ; Elapsed Timer Routines
    ASM
INT_LIST  macro    ; IntSource,        Label,  Type, ResetFlag?
            INT_Handler    TMR1_INT,  _ClockCount,   PBP,  yes
            INT_Handler    RX_INT,     _Check_Command, PBP, yes 
            INT_Handler    TMR3_INT,  _CheckWind,   PBP,  yes 
        endm
        INT_CREATE               ; Creates the interrupt processor
    ENDASM
   
@ INT_ENABLE    TMR1_INT     
@ INT_ENABLE    RX_INT
@ INT_ENABLE    TMR3_INT

For reference here is how I count number of times the switch closes, my version of what you posted, seems to work good.

CheckWind:
 If Wpin = %1 Then
    WindThsState = %1
 Else
    WindThsState = %0
 EndIF
 
 If WindThsState <> WindLstState Then 'Spinning 
    WindCnt = WindCnt +1
 ENDIF
  WindLstState = WindThsState
@ INT_RETURN

Let me know where I should modify, Thanks again for the help.
Shane

[HenrikOlsson]

Hello Shane,
I'm jumping in here with my couple of cents worth...
TMR3 trips an interrupt when it overflows. It's 16bits wide so it counts from 0 to 65535 and then overflows back to 0, tripping the interrupt. If you have the TMR3 prescaler set to 1:1 (which you have) the timer "ticks" at Fosc/4 or 5Mhz in the case of a 20MHz x-tal. 5MHz/65536=76.3Hz which is your interrupt frequency.

If you want it to interrupt at higher frequency all you need to do is preload TMR3 with a suitable value each time it interrupt. If you preload it with 51350 then you'll get an interrupt frequency of roughly 5Mhz/(65536-51350)=352Hz.

Remember that you have to load the timer each time it overflows, not just once. This is best done by stopping the timer and then ADD the preload value to the content of the timer registers, this way the interrupt latency is catered for.

Using TMR2 may be better since it interrupts when it matched PR2 and automatically resets so you don't need to preload in each interrupt.

With all that said, if you're already HAVE a timer interrupt (TMR1 in your case) why not use that. There's no reason you can't check your input AND keep time (if that's what you're doing) from within the same ISR. There's usually no need to have a separate timer for each and every task.

/Henrik.

[Charles Linquis]

The preload value for a timer is easily calculated with Mister E's PIC MultiCalc.

http://www.picbasic.co.uk/forum/atta...7&d=1225550328

[ShaneMichael]

Charles and Henrik;
Preloading TMR3 Worked like a charm, I made it go just a little faster than I need just for some room for error. With a preload of:

TMR3Preload       CON 51650

I put this at the end of my TMR3 interupt routine to preload:

T3CON.0 = %0   'Turn Timer Off To preload for frequency I need.
TMR3Now.Highbyte = TMR3H
TMR3Now.LowByte =  TMR3L
TMR3Now = TMR3Now + TMR3PreLoad
TMR3H = TMR3Now.HighByte
TMR3L = TMR3Now.LowByte
T3Con.0 = %1  'Turn Timer on

Seems to run at 355-360 hz. The error is probably in how I'm capturing the frequency, it's probably more accurate than that. That's what I need though, MISTER E's MultiCalc is handy for this, so thanks also to Mister E and everyone else. I didn't know I could preload the timer, I'm learning more all the time.

Thanks Again,
Shane