On 3/15/06, markp@cannontech.com <markp@cannontech.com> wrote:
>
>
> >>I needed to convert a 0to5 volt input to 16 discrete alarm
> >>temperatures. Dividing a 0-4096 12bit A/D count by 256 yields
> >>a perfect 0-15 result.  To do this, I simply grab the high
> >>byte from the conversion.  I then use a 16-entry lookup table
> >>to convert to literal temperature values.
> >>
> >>This is the analog-read subroutine:
> >>
> >>      ; Read alarm temperature 1
> >>      movlw b'00001001' ; Configure A/D, RA1/AN1.
> >>      movwf adcon0      ;
> >>      call  D_20US      ; Delay 20 microseconds.
> >>      bsf   adcon0,2    ; Start conversion.
> >>      btfsc adcon0,2    ; Test go/done bit.
> >>      goto  $-1
> >>      movf  adresh,w    ; Conv complete, get high A/D result.
> >>      movwf ALARM1_RAW  ; Write data to register.
> >>
> >>It works except for one thing, I never get a zero for the high
> >>byte. See results *:
> >>It works except for one thing, I never get a zero for the high byte.
> >>See results *:
> >>
> >>Input           12 bit      Expected    Actual
> >>Volts           Count       High Byte   High Byte
> >>0.00            0           0           1 *
> >>0.31            256         1           1
> >>0.63            512         2           2
> >>0.94            768         3           3
> >>1.25            1024        4           4
> >>1.56            1280        5           5
> >>1.88            1536        6           6
> >>2.19            1792        7           7
> >>2.50            2048        8           8
> >>2.81            2304        9           9
> >>3.13            2560        10          10
> >>3.44            2816        11          11
> >>3.75            3072        12          12
> >>4.06            3328        13          13
> >>4.38            3584        14          14
> >>4.69            3840        15          15
> >>
> >>
> >>Any ideas on what would cause the high byte to always have a minimum
> value of one?
>
> >What device are you using?  The only 12bit integrated ADCs in the
> >16bit core that I know of are in the old 16C77x parts.  Have you
> >correctly configured ADCON1?
> >
> >Regards
> >
> >Mike
> >
>
> You're right, I'm using a 16C773.  ADCON1 is configured correctly.  Three
> other analog inputs work perfectly, however, this one is just short of
> perfect due to the value being 1 when it should be 0.

Maybe you have a wrong carry flag somewhere ?
BTW, why did you use such a complicated algorithm while probably was
much easy with the internal reference and one comparator? Probably
those aren't available in PIC16C773 ?

greetings,
Vasile

-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist