Error programming 10F222

[JoelMurphy]

holy crow.
i was planning on setting up a breadboard to do ICSP on the thing... i'll still give that a go, while i'm waiting for my shipment from meLabs!
i have used the 12F parts [8 pin] with the standard ZIF with great success, and just assumed... yeah, right.
it is still possible to be smarter than yesterday.

thankyou kind Charles for waking me from my stupor.

[JoelMurphy]

Hi,
Got that 10F ZIF adapter. I had also thrown together an ICSP setup on my breadboard. Both work! Iʻd rather use the ZIF, and free up the breadboard!

SO, after getting a good test running, I put together my first draft program and got ʻUnable to fit Variableʻ error. Nice. Turns out that PBP uses 22 bytes of RAM for storing data that it uses to run my code. Well, the 10F222 has 23 bytes of RAM... I got 1 byte variable to use in my PBP program! This is a rude awakening.
The essence of my project is to fade 3 LEDs. Not much heavy lifting going on here. I started incorporating the TMR0 in my software PWM loop. That worked barely. The basic structure of my PWM loop is:

FOR X = 0 TO 100
FOR Y = 0 TO 100
IF Y < X THEN
GPIO = 1
ELSE
GPIO = 0
ENDIF
NEXT Y
NEXT X

to fade the other way, you just mirror the above loop.
You see that Iʻm using 2 byte variables here...
So what Iʻve done is to commandeer the OSCCAL register. There is an issue with bit 0 of OSCCAL and that it effects GP2. I just increment OSCCAL by 2 when Iʻm using it. I got some usable LED fading happening and found a free variable!
Iʻm still not happy though. I really need one more byte var to get the effect that I want. Anybody have any ideas? Is it possible to tap into the RAM area that PBP uses? Maybe thereʻs a DEFINE somewhere that I can make do my bidding...?

Iʻll post the code soon. Itʻs on another machine
Joel

[JoelMurphy]

Here's the code that I'm working with.

@ Device PIC10F222, IOFSCS_8MHZ, MCPU_OFF, WDT_OFF, PROTECT_OFF, MCLR_OFF


OPTION_REG = %11010010 'GPWU_OFF, GPPU_OFF, TOCS_OFF, TOSC_HIGH>LOW, PSA_TMR0, PS_1:8

ADCON0 = 0
TRISIO = 0
GPIO = 0

Time var byte

pause 100

Main:
    OSCCAL = 2
    GOSUB FADE1A
    PAUSE 1000
    OSCCAL = 2
    GOSUB FADE2A
    PAUSE 1000
    OSCCAL = 2
    GOSUB FADE3A
    PAUSE 100
    
GOTO MAIN




Fade1A:
    FOR TIME = 0 TO 250
    IF TIME < OSCCAL THEN
        IF TIME < 80 THEN
            GPIO = 7
        ELSE
            GPIO = 6
        ENDIF
    ELSE
        GPIO = 7
    ENDIF
    NEXT TIME
    OSCCAL = OSCCAL + 2
    IF OSCCAL = 250 THEN
        OSCCAL = 2
        GOSUB FADE1B
        RETURN
    ENDIF
    GOTO FADE1A
Fade1B:
    FOR TIME = 0 TO 250
    IF TIME < OSCCAL THEN
        GPIO = 7
    ELSE
        IF TIME > 170 THEN
            GPIO = 7
        ELSE
            GPIO = 6
        ENDIF
    ENDIF
    NEXT TIME
    OSCCAL = OSCCAL + 2
    IF OSCCAL = 250 THEN        
        RETURN
    ENDIF
    GOTO FADE1B
 
'the other two subroutine blocks are the same as above 
'with GPIO = 5 and GPIO = 3

It looks like it might be possible to use POKE and PEEK to access the RAM area that is restricted... I'm going to play around abit, I could get away with 1 more byte var, but I would love to have 2 more!

[JoelMurphy]

Well, I think I got what I need.
I have been tweaking the options and have found that if I use MPASM I can get two more byte variables! So total including commandeering the OSCCAL register is 4 byte variables. I'm on my way.
Would love to know if anyone has any further ideas about how to squeeze more RAM out of PBP