> I have  a  register pair,  ADC_H and ADC_L , which contain a 16 bit value
> representing an  ADC sample. I want to take successive samples and
> add them into a 32Bit accumulator. ACC_3, ACC_2, ACC_1, ACC_0.
> Speed is essential so code needs to be short. I am having trouble
> with the 16bit / 32 bit Add that I attempted to write.  It does funny things
> for example when I  add FFFF successively. I have written the rest of
> the loop  and the ADC control part  but just can't get the accumulate to
> work.
>
> Would someone please give me a clue on how to  do the add and how to handle
> the
> carrys. Or maybe send  a snippet of code that will help?
>
> Also I need to find the Maximum positive value and the Minimum negative
> value of the
> register pair ADC_H and ADC_L. The value in this pair represents a signed
> number
> from +32767 to -32768.  Again code speed is important.
>
Remember that you have to sign extend the 16 bit value to 32 bits, before
you add them to ACC.
Here is a piece of code adapted from an earlier post by John Payson that
should work:

; ACC(32) = ACC(32)+ADC(16) (Signed)
        clrf    Temp,f      ;Default to positive
        btfsc   ADC_1,7     ;Test sign bit
        decf    Temp,f      ;ADC is negative
        movf    ADC_0,w
        addwf   ACC_0,f
        movf    ADC_1,w
        btfsc   C
         incfsz ADC_1,w
         addwf  ACC_1,f
        movf    Temp,w      ;Get sign extension byte
        btfsc   C
         incfsz Temp,w
         addwf  ACC_2,f
        clrf    Temp2,f
        rlf     Temp2,w
        addwf   Temp,w
        addwf   ACC_3,f

Looks like about 17 cycles.

Niki