picbasic +12c508a+soft_stack error

[omid_juve]

Hi
I wrote a simple basic program for my PIC (12c508A) in a Picbasic Pro compiler
but I have this error after compilation "unable to fit variable soft_stack"
please see my program below and tell me the errors as I am beginner in PIC programming .
Thanks


lowinput VAR GPIO.0
overtemprature VAR GPIO.1
overload VAR GPIO.2
outputpin VAR GPIO.3
green VAR GPIO.4
red VAR GPIO.5
count1 var byte
i var byte

OUTPUT outputpin
input lowinput
input overtemprature
input overload
OUTPUT green
OUTPUT red

while i=0
if (lowinput=0) and (overload=0) and (overtemprature=0) then

outputpin =0
High green
Pause 500
Low green
Pause 500

ENDIF


if (lowinput=0) and (overload=1) and (overtemprature=0) then

while count1<2
gosub overloadmode
count1=count1+1
wend
outputpin =1

while count1<7
gosub overloadmode
count1=count1+1
wend

outputpin =0
count1=0
ENDIF

if (lowinput=1) and (overload=0) and (overtemprature=0) then

while count1<5
gosub lowinputmode
count1=count1+1
wend
outputpin =1
count1=0
ENDIF
if (lowinput=0) and (overload=0) and (overtemprature=1) then

while count1<5
gosub overtempraturemode
count1=count1+1
wend
outputpin =1
count1=0
ENDIF
if (lowinput=1) and (overload=1) and (overtemprature=0) then

while count1<5
gosub overtempraturemode
count1=count1+1
wend
outputpin =1
count1=0


endif


wend


overloadmode:
High green
low red
Pause 500

Low green
high red
Pause 500
return

lowinputmode:
High green
low red
Pause 500

Low green
high red
Pause 500
return

overtempraturemode:

High green
high red
Pause 500

Low green
low red
Pause 500

return

end

[mister_e]

you may need to move on a bigger RAM PIC such as PIC12C509. 12C508 have 25 bytes of RAM, while 12C509 have 41 + twice the program space size.

The compiler use between 20 & 22 bytes of RAM itself.. so... no real space for more variables

[Darrel Taylor]

A 12F might be a good choice. Cheaper too.

However, what you have should fit, with a little modification.


The complex IF statements are creating 3 (T)emp vars, which are word sized. (even though it's only using 1 bit per WORD)
So it's using 6 bytes that it doesn't really need too.

If you change all the IF statements to something like this for the first one, I believe it should fit.

    ;if (lowinput=0) and (overload=1) and (overtemprature=0) then
    if (lowinput=0) THEN
        if (overload=1) THEN
            if (overtemprature=0) THEN
                while count1<2
                gosub overloadmode
                count1=count1+1
                wend
                outputpin =1
                
                while count1<7
                gosub overloadmode
                count1=count1+1
                wend
                
                outputpin =0
                count1=0
            ENDIF
        ENDIF
    ENDIF

hth,

[mister_e]

i was thinking of a select case and GPIO mask. mmm...

edit: Seems to work

lowinput VAR GPIO.0
overtemprature VAR GPIO.1
overload VAR GPIO.2
outputpin VAR GPIO.3
green VAR GPIO.4
red VAR GPIO.5
count1 var byte
Mask var byte


OUTPUT outputpin
input lowinput
input overtemprature
input overload
OUTPUT green
OUTPUT red

Start:
        mask=GPIO & 7
        select case mask
                case %000 ;if (lowinput=0) and (overload=0) and (overtemprature=0) then
                    outputpin =0
                    High green
                    Pause 500
                    Low green
                    Pause 500

                case %100 ;if (lowinput=0) and (overload=1) and (overtemprature=0) then
                    while count1<2
                        gosub overloadmode
                        count1=count1+1
                        wend
                    outputpin =1
                    
                    while count1<7
                        gosub overloadmode
                        count1=count1+1
                        wend
                    
                    outputpin =0
                    count1=0

                case %001 ;if (lowinput=1) and (overload=0) and (overtemprature=0) then
                    while count1<5
                        gosub lowinputmode
                        count1=count1+1
                        wend
                    outputpin =1
                    count1=0
                
                case %010 ;if (lowinput=0) and (overload=0) and (overtemprature=1) then
                    while count1<5
                        gosub overtempraturemode
                        count1=count1+1
                        wend
                    outputpin =1
                    count1=0

                case %101 ;if (lowinput=1) and (overload=1) and (overtemprature=0) then
                    while count1<5
                        gosub overtempraturemode
                        count1=count1+1
                        wend
                    outputpin =1
                    count1=0
                end select

        goto Start


overloadmode:
        High green
        low red
        Pause 500

        Low green
        high red
        Pause 500
        return

lowinputmode:
        High green
        low red
        Pause 500
        
        Low green
        high red
        Pause 500
        return

overtempraturemode:
        
        High green
        high red
        Pause 500
        
        Low green
        low red
        Pause 500
        
        return

        end

[Archangel]

. . . .

The compiler use between 20 & 22 bytes of RAM itself.. so... no real space for more variables

Hello Mister_E
Could you please explain? This statement appears to me to mean PIC BASIC PRO is an intrepreted code, if not why would the compiler require any of the PIC's ram?
Thank You
JS

[mister_e]

Yeah but the library/compiler need to reserve some variable to work to not interfer with your code... euh.. that's my understanding

support for 12 bit core is limited, and the real explanation can be found at the bottom of the folowiing page
http://www.melabs.com/support/12-bit.htm

so why 20-22... i don't know, seems hungry to me.. but, you have to deal with it.

We already heard something like that for 10F and bring some weird/nasty/dangerous/clever alternate solutions, some who remind...
http://www.picbasic.co.uk/forum/showthread.php?t=4789&

HTH

[Darrel Taylor]

After converting all the IF statements like I suggested above, it did fit in RAM,
and it compiled to 313 WORDS.

mister_e,
Yours fit in RAM too, and compiled to 295 WORDS. (unexpected for Select Case)

You know what that means RIGHT?
Yup, this ain't over yet!

172 WORDS

@ DEVICE WDT_OFF
DEFINE NO_CLRWDT 1 ' Don't kick the Dog 

lowinput       VAR GPIO.0
overtemprature VAR GPIO.1
overload       VAR GPIO.2
outputpin      VAR GPIO.3
green          VAR GPIO.4
red            VAR GPIO.5
FlashCount     var byte
FlashPattern   var byte

TRISIO = %000111

Start:
    If (lowinput=0) then
        IF (overload=0) then             ; lowinput, overload, overtemprature
            if (overtemprature=0) then                      ; 000
                outputpin = 0
                FlashCount = 1 
                FlashPattern = %0001  ; ModeOK
                GOSUB FlashLEDs
            else                                            ; 001
                FlashCount = 5
                FlashPattern = %0011  ; overtempraturemode
                gosub FlashLEDs
                outputpin = 1
            endif
        else                                                ; 010
            if (overtemprature=0) then           
                FlashCount = 2
                FlashPattern = %1001  ; overloadmode
                gosub FlashLEDs
                outputpin =1
                
                FlashCount = 7
                gosub FlashLEDs
                outputpin =0
            else                                            ; 011
                ; --- Action undefined ---
            endif
        endif
    else
        if (overload=0) then                     
            if (overtemprature=0) then                      ; 100
                FlashCount = 5
                FlashPattern = %1001  ; lowinputmode
                gosub FlashLEDs
                outputpin =1
            else                                            ; 101
                ; --- Action undefined ---
            endif
        else                                                ; 110
            if (overtemprature=0) then           
                FlashCount = 5
                FlashPattern = %0011  ; overtempraturemode
                gosub FlashLEDs
                outputpin =1
            else                                            ; 111
                ; --- Action undefined ---
            endif
        endif
    endif
GOTO Start

FlashLEDs:
    REPEAT
         green = FlashPattern.0
         red   = FlashPattern.1
         Pause 500

         green = FlashPattern.2
         red   = FlashPattern.3
         Pause 500
         FlashCount = FlashCount - 1
    UNTIL (FlashCount = 0)
return

However, omid_juve,

There are 3 states that are not accounted for

lowinput=0 overload=1 overtemprature=1
lowinput=1 overload=0 overtemprature=0
lowinput=1 overload=1 overtemprature=1

It may be intentional. Maybe they aren't needed. But normally, unchecked states will cause havoc in your program.

If we knew what was supposed to happen in those states, I think mister_e ..., uh I mean, I could probably reduce it further.

[omid_juve]

it a program for 24v to 12v converter that has 3 input 1.over temperature 2.low input 3.overload and we have 3 output that 2 of them is just indicator and one (outputpin)
is the output for controlling the output voltage according to the situation of inputs.
and just the modes that you saw in the program can occur .
thanks for all of your help. my problem is solved .

[mister_e]

You know what that means RIGHT?
Yup, this ain't over yet!

yup.. sounds like a challenge... 100% PBP??? Maybe doable... scratch, scratch...mmm...

[Darrel Taylor]

12C508A's bad enough,
you want 100% PBP??? too?

Did you know it doesn't even have an overflow flag on timer0?
I didn't, Oi vey!

I had so much fun this time last year with what ended up as the 12-byte cylon scanner,
perhaps I'm a bit over zealous.

The OP is satisfied, and probably wouldn't use it anyways,
So even if the program size was only 83 words, like I have it now,
It'll never be used.

We need a real Optimization contest.
They're just too fun.
<br>

[mister_e]

yup, no eeprom either... so nothing interesting. Think it's like a 16F84A, 16C54... but in 8 pin package

There's always place for an optimization contest, next time, no rules (pic, language etc etc )... agreed?

[skimask]

yup, no eeprom either... so nothing interesting. Think it's like a 16F84A, 16C54... but in 8 pin package
There's always place for an optimization contest, next time, no rules (pic, language etc etc )... agreed?

Make it tough...use a 10F200...no SOT->DIP adapters allowed, only hand soldering, dead bug style. In fact, no soldering, hand wrap wires around the legs....
and no power supply, hand crank a motor to charge up a cap...

[Bruce]

Any chance we can see that 12-byte cylon scanner?

[Darrel Taylor]

Sure,

The 12-byte version is toward the end on page 2.
http://www.sfcompiler.co.uk/forum/viewtopic.php?t=72

Edit: Hey, you were there to.
<br>

[Bruce]

Yipes I totally forgot that one. That was way interesting and the end result was really creative.

Nice work DT..;o}

Tim B and I got into one of these a long time back, and here's one he came up with.

Can we beat this without EEPROM?

Standard cyclone thingie moving the LED back & forth with ~1S delay periods.

  	list      p=16F628A
	#include "P16F628A.inc"
	errorlevel  -302        ; suppress message 302 from list file
	__CONFIG _CP_OFF & _BODEN_ON & _MCLRE_ON & _WDT_ON & _PWRTE_ON & _LVP_OFF & _INTOSC_OSC_NOCLKOUT

	movlw	0x01
	movwf	PORTB	     ; Set Portb,0
	bsf	STATUS,RP0   ; Bank 1
	clrf	TRISB	     ; RB all outputs
	bcf	OPTION_REG,0 ; 1:64 prescaler to WDT
	bcf	STATUS,RP0   ; Bank 0

Left
	sleep		    ; go to sleep for 64 * 18mS ~1 second
	rlf	PORTB,F     ; rotate bit across portb from lsb to msb
	btfss	PORTB,7     ; jump to Right after Portb,7 = 1
	goto	Left
	
Right
	sleep		    ; go to sleep for 64 * 18mS ~1 second
	rrf	PORTB,F     ; rotate bit across Portb from msb to lsb
	btfss	PORTB,0     ; jump back to left after Portb,0 = 1
	goto	Right       ; loop until Portb,0 = 1
	goto	Left        ; now rotate back to the left.
	
    End

[mister_e]

seems to be hardly optimized to me

food for thought... let's see...

[Darrel Taylor]

Tim's compiled to 15 words.

Here's 13 ... &nbsp; Note: LED's go to VDD

    list      p=16F628A
    #include "P16F628A.inc"
    errorlevel  -302        ; suppress message 302 from list file
    __CONFIG _CP_OFF & _BODEN_ON & _MCLRE_ON & _WDT_ON & _PWRTE_ON & _LVP_OFF & _INTOSC_OSC_NOCLKOUT

    clrf    PORTB         ; PORTB all 0's
    bsf	    STATUS,RP0    ; Bank 1
    bcf	    OPTION_REG, 0 ; 1:64 prescaler to WDT
    bcf     TRISB, 0      ; set RB0 to output
    bsf     STATUS, C     ; set carry flag for rotation
Loop
    btfss   TMR2, 3       ; switch Left/Right every 8 times
    rlf	    TRISB,F       ; rotate Left across TRISB
    btfsc   TMR2, 3
    rrf     TRISB,F       ; rotate Right across TRISB
    btfsc   STATUS, C     ; No sleep if output rotated into carry
    sleep                 ; nite nite
    incf    TMR2          ; Timer2 used for it's 00 value on reset.
    goto    Loop          ; Repeat

EDIT: Rats!
While it works in MPSIM, it's not going to do it on the real thing.
Time for a breadboard.

<br>

[Darrel Taylor]

OK, ignore the last one. That's what I get for using a simulator.

This is 14 WORDS, 1 less than Tim's.
At least it beats him

    list      p=16F88
    #include "P16F88.inc"
    errorlevel  -302        ; suppress message 302 from list file
    __CONFIG  _CONFIG1, _XT_OSC & _CP_OFF & _BODEN_ON & _MCLR_ON & _WDT_ON & _PWRTE_ON & _LVP_OFF

    clrf    PORTB         ; PORTB all 0's
    bsf     STATUS,RP0    ; Bank 1
    bcf     OPTION_REG, 0 ; 1:64 prescaler to WDT
    bcf     TRISB, 0      ; RB0 output
Loop
    sleep                 ; nite nite

    btfss   SPBRG,0       ; SPBRG used for it's 00 reset value   
    rlf     TRISB,F       ; rotate Left across TRISB
    btfsc   SPBRG,0
    rrf     TRISB,F       ; rotate Right across TRISB
    
    btfss   TRISB, 7      ; At LED7? - go the other way
    incf    SPBRG,F
    btfss   TRISB, 0      ; At LED0? - go the other way
    incf    SPBRG,F 
    
    goto    Loop          ; Repeat

LED's go to VDD.
Tested, working on 16F88, should be the same (except configs) on 16F628A.
<br>

[mister_e]

yup.. hence why i never trust sim...

[Bruce]

Nice work DT.

Here's one at 13 words.

  list      p=16F628A
  #include "P16F628A.inc"
  errorlevel  -302
  __CONFIG _CP_OFF & _BODEN_ON & _MCLRE_OFF & _WDT_ON & _PWRTE_ON & _LVP_OFF & _INTOSC_OSC_NOCLKOUT

   clrf  PORTB        ; clear port
   bsf   STATUS,RP0   ; Bank 1
   bcf   TRISB,0      ; RB0 = output
   bcf   OPTION_REG,0 ; 1:64 prescaler to WDT

Left
   sleep             ; go to sleep for 64 * 18mS ~1 second
   rlf   TRISB,F     ; rotate bit across portb from lsb to msb
   btfsc TRISB,7    
   goto  Left
	
Right
   sleep          ; go to sleep for 64 * 18mS ~1 second
   rrf   TRISB,F  ; rotate bit across Portb from msb to lsb
   btfsc TRISB,0
   goto  Right
   goto  Left     ; now rotate back to the left.
	
   End

[Darrel Taylor]

Oh Yeah! Very nice Bruce.

Now that's going to be hard to beat.
If you don't hear from me for a month.
You'll know what happened.

[skimask]

Oh Yeah! Very nice Bruce.

Now that's going to be hard to beat.
If you don't hear from me for a month.
You'll know what happened.

I just checked on my '628A on my breadboard...about a ba-zillion times...
I set up code/hardware to do nothing but RESET, and put Port B onto a bank of LEDs, Vdd on the other side of the LED.
EACH time, Port B came up ZERO...no LEDs lit up, until I set a TRIS bit for output.
So, my submission...
Coming in at 12 bytes...

  list      p=16F628A
  #include "P16F628A.inc"
  errorlevel  -302
  __CONFIG _CP_OFF & _BODEN_ON & _MCLRE_OFF & _WDT_ON & _PWRTE_ON & _LVP_OFF & _INTOSC_OSC_NOCLKOUT

;   clrf  PORTB        ; clear port - not required on my breadboard, might be different on others
   bsf   STATUS,RP0   ; Bank 1
   bcf   TRISB,0      ; RB0 = output
   bcf   OPTION_REG,0 ; 1:64 prescaler to WDT

Left
   sleep             ; go to sleep for 64 * 18mS ~1 second
   rlf   TRISB,F     ; rotate bit across portb from lsb to msb
   btfsc TRISB,7    
   goto  Left
	
Right
   sleep          ; go to sleep for 64 * 18mS ~1 second
   rrf   TRISB,F  ; rotate bit across Portb from msb to lsb
   btfsc TRISB,0
   goto  Right
   goto  Left     ; now rotate back to the left.
	
   End

Somebody else check this out and see what happens.
As I stated, maybe it's my breadboard, maybe it's my 628A (I've only got 1 left), maybe it's the rise time on my PSU.

EDIT: Datasheet says that the port registers come up in an unknown state. Apparently, in this case, I now know my state...

[Bruce]

It might work on one series, but I wouldn't count on it as being reliable across the board.
I have seen various PICs come up with totally random values in port latches at POR, so
you really do need to clear portb latches first if it's to be reliable.

[mister_e]

i'm thinking of something using interrupt.. not sure if it could reduce things to 8-10 words... mmm..

[Bruce]

I had considered that approach too, but more than half of your code space (or more) would
be eaten up with interrupt configuration setup.

Making it "reliable" with less than 13 words is going to be a real task.

[mister_e]

yup, but where is it stated that it has to be reliable, where it is stated that we need to use the same PIC and hardware ?

i still believe it's doable... but with some hair lost

[skimask]

It might work on one series, but I wouldn't count on it as being reliable across the board.
I have seen various PICs come up with totally random values in port latches at POR, so
you really do need to clear portb latches first if it's to be reliable.

I don't doubt that at all...
I just checked the CARRY bit, datasheet says unknown on POR, mine come's up cleared every time (at least every time I've check anyways). That shoots that idea down...

[Darrel Taylor]

yup, but where is it stated that it has to be reliable?

I think that's a given. It has to work when you turn on the power.
As for the rest of the rules, well, I'm assuming No EEPROM is the only one. Edit: with a 16F628A

i still believe it's doable... but with some hair lost

I'm starting to believe NO.
But I haven't given up yet.

I've got 3 completely different versions at 13 words now.
1 using external circuitry to TMR0
1 using external circuitry to a comparator
and another that simply shaves a word off my 14 word program. (wish I'd done that to begin with)

It seems that no matter what you do to save a word, you end up having to add something to make it work.

I'm sure getting some "Quality Time" in with the datasheets though. All kinds of stuff I didn't know before.
<br>

[mister_e]

yup all good points.. the only thing i've never dare to mess with is the stack overflow behaviour and ORG... 8 bit stack seems nice... but not sure how bad/good it behave...mmm

interesting how a led chaser can be a food for thought

[skimask]

yup all good points.. the only thing i've never dare to mess with is the stack overflow behaviour and ORG... 8 bit stack seems nice... but not sure how bad/good it behave...mmm
interesting how a led chaser can be a food for thought

Something tells me that somebody is about to break the law...
The law of diminishing returns!

Does this 'thing' have to have 8 LEDs? Or can it have 16 LEDs?
I'm thinking something along the lines of 2 LEDs per pin, back-to-back, with the common for all running to another pin...don't know what I'm thinking yet...

[mister_e]

Yeah maybe breaking the law, but as i said, i didn't test it...it just looks good in my mind... 8 deep level stack, 8 leds... mmm

[skimask]

Yeah maybe breaking the law, but as i said, i didn't test it...it just looks good in my mind... 8 deep level stack, 8 leds... mmm

Great...now you got me thinking...and I was about ready to sleep...
Flip a bit...recursive call's without return's, slide a bit over...wash/lather/rinse/repeat/STACK OVERFLOW...flip a bit...repeat...

[Darrel Taylor]

Ladies and Gentlemen!

Please direct your attention to the Center Ring ....

12 WORDs

 	list      p=16F628A
	#include "P16F628A.inc"
	errorlevel  -302        ; suppress message 302 from list file
	__CONFIG _CP_OFF & _BODEN_ON & _MCLRE_ON & _WDT_ON & _PWRTE_ON & _LVP_OFF & _INTOSC_OSC_NOCLKOUT

Init
    clrf    PORTB         ; PORTB all 0's
    rlf     STATUS, F     ; sets Bank1 and clears carry
    bcf     OPTION_REG, 2 ; 1:8 prescaler to WDT
    comf    SPBRG,F       ; Make SPBRG all 1's
Loop
    btfsc   STATUS,C      ; no sleep if rotated into carry
    sleep                 

    rlf     SPBRG,F       ; rotate SPBRG left msb to carry
    rrf     PR2, F        ; rotate carry into PR2, lsb to carry
    movf    SPBRG, W      ; put SPBRG into W reg
    andwf   PR2, W        ; and it with PR2
    movwf   TRISB         ; result goes to TRISB

    goto    Loop          ; Repeat

Bwwaaa Haaa Haaa!

Added:
<script language="javascript" src="http://www.pbpgroup.com/files/CylonRotate.js"></script><br>

[mister_e]

Woohoo, now if we see this code example somewhere, we will know where it come from PR2, SPBRG... euh in a light 'chaser' ,deuh, doh... hein???

Nice one.... smell of the burned brain-cells, but nice ;o}

[Bruce]

Well, that's going to be tough to beat. Nice work Darrel.

Interesting side note: Using your rlf STATUS,F approach reduced my 13 word version to
only 12 words..;o}

  	list      p=16F628A
	#include <P16F628A.inc>
	errorlevel  -302    ; suppress message 302 from list file
	__CONFIG _CP_OFF & _BODEN_ON & _MCLRE_OFF & _WDT_ON & _PWRTE_ON & _LVP_OFF & _INTOSC_OSC_NOCLKOUT

	clrf  PORTB	   ; clear port
	rlf   STATUS,F     ; Bank1, clear carry for 1st rotate
	bcf   OPTION_REG,2 ; 1:8 prescaler to WDT

Left
	sleep              ; go to sleep for ~1 second
	rlf	TRISB,F    ; rotate 0 from carry into lsb, then left across trisb
	btfsc	TRISB,7    ; jump to Right once trisb,7 = 0
	goto	Left
	
Right
	sleep              ; go to sleep for ~1 second
	rrf	TRISB,F    ; rotate bit right across trisb
	btfsc	TRISB,0    ; jump to left once trisb,0 = 0
	
	goto	Right       ; loop until trisb,0 = 0
	goto	Left        ; now rotate back to the left

        end

[Darrel Taylor]

OMG! How HORRIBLE!

That means yours doesn't pause at each end like mine.
Which makes your 12, better than my 12.



Actually, it's awesome,
I'm crying because now I have to spend the next week trying to beat it.
<br>

[Bruce]

Oh I wouldn't call it better since your version is WAY more creative, and your nifty rotate STATUS thing was the only reason mine was reduced.

Man that was some clever & creative problem solving, and that's really what it's all about...;o}