Assembly debug

BobbyV · 2011-02-16 14:00

Hello All,

I'm new to assembly and I am stumped on a problem I'm having! I'm trying to write a bare-bones spi routine for and ADC (AD7323) a I believe the problem is with the program not looping. What has me really confused is that when I debug the program with PSAD and just let it run without breakpoints, the routine seems to work just fine (I checked the chip's data output, and the clock signal going into the chip with an oscilloscope and I see consistent, regular pin activity). However, when I run it without PSAD, I only see a very quick pulse on scope and then no pin activity which has led me to think that the program ran through the code but did not loop.

I'm guessing there's some bone-headed thing I'm missing here that more experienced eyes will catch!
Thanks for the help,

Here's my assembly code in an object that I writing:

''
''      ASSEMBLY SHIFT IN PROTOTYPE
''               FOR AD7323
''                VER. 1
''           Robert Valtierra
''             FEB. 13 2011
''
VAR
        long     Cog

PUB START(SAMPLE)

  Cog := cognew(@entry, @SAMPLE) + 1

PUB STOP

  if Cog
    cogstop(Cog~ -  1)


DAT

                        org     0
entry

                        mov     Addr, par               ' Copy "Sample" address to Addr
                        muxnz   dira,DAT_CLK            ' Configure DAT_CLK and CS as outputs (1)
                        muxnz   dira,CS 
                        muxnz   outa,CS          
                        muxnz   outa,DAT_Clk            ' Set DAT_Clk and CS initially high

SHIFT_IN                mov     count,N_Bits            ' Set counter to number of sample bits
                        muxz    outa,CS
                        
SHIFT_LOOP              test    D_in,ina         wc     ' Test if data pin is high or low, write carry if high
                        rcl     TEMP1,#1                ' Shift in carry flag one to the left
                        muxz    outa,DAT_CLK            ' Cycle clock low 
                        muxnz   outa,DAT_CLK            ' Cycle clock high 
                        djnz    count,#SHIFT_LOOP       ' Decrement count, loop if not zero                
                        muxnz   outa,CS                 ' De-trigger Chip Select
                        shl     TEMP1, STP1             '
                        mov     TEMP2, TEMP1
                        and     TEMP2, CLRB             ' Throw out first four bits and multiply by 16 using shift lest  
                        shr     TEMP1, STP2             ' Convert to "little-endian"
                        shl     TEMP2, STP2
                        or      TEMP1, TEMP2
                        wrword  TEMP1, Addr             ' write TEMP value to address of global var "Sample" 
                        mov     TEMP1, #0               ' clear temp variables
                        mov     TEMP2, #0
                        jmp     #SHIFT_IN               ' loop back to SHIFT_IN
'                        
' VARIABLES, PARAMETERS
'                  
STP1                    long    $0000_0004
STP2                    long    $0000_0008
CLRB                    long    $0000_00FF
DAT_CLK                 long    $0020_0000   ' Pin 21 assigned to clock pin
CS                      long    $0002_0000   ' Pin 17 chip select 
D_in                    long    $0010_0000   ' Pin 20 Data in
TEMP1                   long    0            ' Temporary shift in variable
TEMP2                   long    0            ' Temporary shift in variable 
N_Bits                  long    16           ' Number of sample bits
count                   long    0            ' Shift in loop counter
Addr                    res     1            

                        fit

JonnyMac · 2011-02-16 14:03

Are you passing "sample" as an address? If yes, then you don't need the @ in front of "sample" in your cognew line.

Dave Hein · 2011-02-16 14:32

Yes, your PASM program is writing the result to the temporary stack variable SAMPLE. Depending on what your Spin program does after calling start you may be trashing an area of the stack that is being used for another variable. Also, you are always writing to the same location, so there is no way for the Spin program to know when a new sample is available.

The other problem is that you are using MUXZ and MUXNZ, where you probably really want to use ANDN and OR. MUXZ and MUXNZ depend on the zerio flag.

Edit: I thought about the MUXZ and MUXNZ a bit more, and that would work if the zero flag is guaranteed to be zero at the beginning. I don't know if that's a valid assumption or not.

kuroneko · 2011-02-16 15:47

Dave Hein wrote: »

Edit: I thought about the MUXZ and MUXNZ a bit more, and that would work if the zero flag is guaranteed to be zero at the beginning. I don't know if that's a valid assumption or not.

As valid as you make it, I nearly had my head bitten off for asking the initial flag state. Both flags are cleared initially. But I agree, andn & Co would be a better approach.

BobbyV · 2011-02-16 16:54

Got it working, Thanks for the help.

The Sample address was an issue, but not the cause of the lack of pin activity (Thanks for catching that before I had to deal with the frustration of of having a working chip but no data!) Turns out that it was indeed an issue of setting the pin addresses properly.

Thanks again!

Dragan2020 · 2020-12-14 00:07

Hi,
did u ever get your "bare bones routine " working?

I could use it, since I would like to use the same IC...the AD7323.

Best regards.

Assembly debug

Comments