12f1822, serial RX under interrupt, echo program doesn't want to play

[Chris Barron]

I've been trying to get some code working on the 12F1822 which works fine on the 16F690

I can't explain why there should be a big difference in operation. I've gone through the datasheet trying to find differences between teh two chips and making any necessary adjustments.

All I am trying to do at this stage is prove correct operation of the UART, so I have written a simple echo routine.

;        __config _CONFIG1, _FOSC_INTOSC & _WDTE_ON & _MCLRE_OFF & _PWRTE_OFF & _BOREN_OFF & _CLKOUTEN_OFF
;        __config _CONFIG2, _PLLEN_OFF & _LVP_OFF

Include "modedefs.bas" 

DEFine OSC 8
define INTHAND myint

RX_Byte     var byte

goto init

asm
myint:
    btfss   PIR1,5
    goto    int_out
    bcf     PIR1,5
    movfw   RCREG
    movwf   _RX_Byte
    movwf   TXREG
    nop
    btfss   PIR1,TXIF
    goto    $-1
       
int_out:    
    retfie
endasm

init:

TRISA   =   %00000010   ;   set pins for serial data
ANSELA  =   0           ;   set all pins digital
PORTA   =   %00000000


pause 200

asm
    movfw   RCREG
    movfw   RCREG
    movfw   RCREG
endasm

; Power on
; Initialise variables


Restart: 

; Initialise SFR's

;EUSART
TXSTA   =   %10100100
RCSTA   =   %10010000
BAUDCON =   %00000000
SPBRG  =    12            '38400 baud
;SPBRGH  =   0

;INTERRUPTS
INTCON  =   %11000000
PIE1    =   %00100000   ;   interrrupt on data received

;OSCILLATOR
OSCCON  =   %01110011

;PORT A
TRISA   =   %00000010   ;   set pins for serial data
ANSELA  =   0           ;   set all pins digital
PORTA   =   255

;COMPARATOR
CM1CON0 =   0

main:

while 1
    if  (RCSTA.1 = 1) or (RCSTA.2 = 1) then 
        gosub   Clear_err
    endif
wend


Clear_err:
RCSTA = 0
@   movfw   RCREG
@   movfw   RCREG
@   movfw   RCREG
RCSTA   =   %10010000
return


stop

The chip seems to repeatedly reset itself.

Any help getting this going will be very much appreciated. I have tried loits of variations of the commands in case there is a datasheet error ( I recall from previous experience with the 12F1822 I found a datasheet error).

Chris

[Chris Barron]

I can't remember how i resolved a similar issue like this with this chip previously, but I know I've managed it.

I've just looked at the project again and the receiving unit never sends data without a request for it to be sent, so the code doesn't need to expect the arrival of random data. I've just tested it using HSERIN and HSEROUT and it looks like I'll be able to work with the native Pbasic commands instead.

I like having control of interrupts, usually, but this time it's not important to me

[Darrel Taylor]

When you're working with PBP, it's easy to forget that there are "BANKs" in the RAM because it's all handled for you.
But when you are writing at the ASM level, you have to remember about the BANKs. And on the F1 type parts, there are a lot of them. (32 to be exact).

In your case, RCREG is in BANK3.
And since you only have 1 variable in your program, it will obviously end up in BANK0. But when the program gets bigger, that may not be the case.
It should be able to find your variables no matter what bank they end up in.

Also, when you get an interrupt on an F1 part. You have no idea what bank it is in when it gets to the ISR.
It's up to you to change it to BANK0 (if that's the bank your ISR needs).

[Mike, K8LH]

When you're working with PBP, it's easy to forget that there are "BANKs" in the RAM because it's all handled for you.
But when you are writing at the ASM level, you have to remember about the BANKs. And on the F1 type parts, there are a lot of them. (32 to be exact).

In your case, RCREG is in BANK3.
And since you only have 1 variable in your program, it will obviously end up in BANK0. But when the program gets bigger, that may not be the case.
It should be able to find your variables no matter what bank they end up in.

Also, when you get an interrupt on an F1 part. You have no idea what bank it is in when it gets to the ISR.
It's up to you to change it to BANK0 (if that's the bank your ISR needs).

Really? I'm surprised! The BoostC compiler I use (sorry, still no PBP to play with here) apparently keeps track of the variables and registers used in my assembly language instructions and automatically inserts the 'movlb x' instructions for me. Now you have me wondering if that feature is unique to BoostC or if other compilers have it as well? Anyway, please notice how I had to comment out all of my bank select instructions after I discovered they were redundant;

  /*                                                                  *
   *  (C) 2012, K8LH - low level One-Wire R/W (Carry) Bit Driver      *
   *  (hard coded for "owpin" defines and 8-MHz clock)  (30 words)    *
   *                                                                  */
   void owrw()                  // read/write Carry bit
   { intcon.GIE = 0;            // interrupts off
   //asm movlb  owport          // bank 0 (inserted by compiler)
     asm bcf    owport,owpin    // clear owpin output latch
   //asm movlb  owtris          // bank 1 (inserted by compiler)
     asm bcf    owtris,owpin    // owpin low (falling edge)
     asm movf   _pcl,F          // pseudo bra $+1 (2 cycles)
     asm movf   _pcl,F          //   "
     asm movf   _pcl,F          //   "
     asm skpnc                  // if a '1' (C=1)
     asm bsf    owtris,owpin    // set owpin to hi-z '1'
     asm movlw  18/3            //
   dly1:
     asm decfsz _wreg,W         // DelayCy(14*usecs-10)
     asm bra    dly1            //
   //asm movlb  owport          // bank 0 (inserted by compiler)
     asm btfss  owport,owpin    // exactly 14-us from falling edge
     asm clrc                   // clear Carry if 'owpin' was low
     asm movlw  84/3            // balance of 60-usec "r/w slot"
   dly2:
     asm decfsz _wreg,W         // DelayCy(47*usecs-10)
     asm bra    dly2            //
   //asm clrf   _wreg           // wreg already 0 (see above)
     asm rlf    _wreg,W         // save C (wreg = 0x00 or 0x01)
     asm xorwf  _crc,F          // xor b0 bits, result in 'crc'
     asm rrf    _crc,F          // xor result 0 (C=0)?
     asm skpnc                  // yes, skip, else
     asm iorlw  0x18            // wreg = x8+x5+x4+1 poly and b0 bit
     asm rrf    _wreg,W         // restore C (wreg = 0x8C or 0x00)
     asm xorwf  _crc,F          // apply the polynomial, or not
   //asm movlb  owtris          // bank 1 (inserted by compiler)
     asm bsf    owtris,owpin    // set owpin to hi-z '1' @ 61-usecs
     intcon.GIE = 1;            // interrupts on
   }                            //

  /*                                                                  *
   *  (C) 2012, K8LH - One-Wire R/W Byte Function       (11 words)    *
   *                                                                  */
   char owrw(u08 pbyte)         // read/write byte
   { u08 i = 0;                 // bit counter
     do                         //
     { asm rrf _pbyte,W         // put output bit b0 in Carry
       owrw();                  // write then read bit
       asm rrf _pbyte,F         // save input bit from Carry
       i++;                     //
     } while(i.3 == 0);         //
     return pbyte;              //
   }                            //

[Chris Barron]

Thanks Darrel. Can't tell yoiu how close i came to putting in banksels, but then tried something else.

Job done,

Much appreciated, many thanks

Chris

[Dick Ivers]

Hi Chris,
Could you post your corrected code? Myself and others also use 16F1 parts, so this is an important issue.

Thanks,
Dick

[Chris Barron]

Hello Dick

I tested the same code as above, changing the interrupt handler to the following

asm
myint:
    banksel PIR1
    btfss   PIR1,5
    goto    int_out
    bcf     PIR1,5
    banksel  RCREG
    movfw   RCREG
    banksel _RX_BYTE
    movwf   _RX_Byte
    banksel TXREG
    movwf   TXREG
    nop
    btfss   PIR1,TXIF
    goto    $-1
       
int_out:    
    retfie
endasm

If in doubt, add another banksel

I'm not using the code now but I did test it upon receipt of Darrel's input to check it out.
I have used this part as a battery cell monitor and I can recall scratching my head over the same issue as then.
Hopefully I'm twice bitten, third time shy

[Mike, K8LH]

Are you missing a bank select in that code snippet?

asm
myint:
    banksel PIR1            '                    |B0
    btfss   PIR1,RCIF       '                    |B0
    goto    int_out         '                    |B0
    bcf     PIR1,RCIF       '                    |B0
    banksel RCREG           '                    |B3
    movfw   RCREG           '                    |B3
    banksel _RX_BYTE        '                    |B?
    movwf   _RX_Byte        '                    |B?
    banksel TXREG           '                    |B3
    movwf   TXREG           '                    |B3
    nop                     '                    |B3
    btfss   PIR1,TXIF       '                    |B0 <-- *****
    goto    $-1             '                    |B0
       
int_out:    
    retfie
endasm

If _RX_Byte is in bank 0, then maybe this would work?

asm
myint:
    banksel PIR1            '                    |B0
    btfss   PIR1,RCIF       '                    |B0
    retfie                  '                    |B0
    bcf     PIR1,RCIF       '                    |B0
    banksel RCREG           '                    |B3
    movfw   RCREG           '                    |B3
    movwf   TXREG           '                    |B3
    banksel _RX_BYTE        '                    |B0
    movwf   _RX_Byte        '                    |B0
    btfss   PIR1,TXIF       '                    |B0
    goto    $-1             '                    |B0
    retfie                  '                    |B0
endasm

[Chris Barron]

Whatever.

Darrel's answer covered it.