I've posted this little routine several times. ÀIs it the one you ask for?

;
; Encoder read
        clrf    auxint     ;  Clears auxiliar input var.
        btfsc   encx       ; Gets bits 'encx' and 'ency' (encoder bits)
        bsf     auxint,0   ; in auxint (bits 0 and 1)
        btfsc   ency
        bsf     auxint,1
;
        movf    auxint,W   ;  Xor's actual and previous encoder
        xorwf   encod,F    ; states. Results in encod.
;
        rrf     auxint,F   ;  Save's actual encoder state (and rotates it)
        rlf     encod,F    ; in encod, for the next time.
        rrf     auxint,F
        rlf     encod,F
;
; Right now, bits 2 and 3 of encod keep the encoder's rotate direction:
        btfsc   encod,2    ; Rotates in one direction
        incf    posenc,F
        btfsc   encod,3    ; ...or in the other one.
        decf    posenc,F
;

This routine runs in a fixed-time interrupt, with many other things. Don't
care too much about the interrupt's frequency. You can detect overspeed when
both direction bits are set.

The last time I posted this little routine, I put with it a little
explanation... but my PC dead last month, and with it all my e-mails. If
anyone of you have the previous message, please, repost it. Thanks.

Bye.

-----Original Message-----
From: Walter Banks <walter@BYTECRAFT.COM>
To: <PICLIST@MITVMA.MIT.EDU>
Date: martes, 08 de diciembre de 1998 22:23
Subject: Re: Quadrature Encoder


>> a couple of months ago there was a thread about Encoding
>> a A-B Decoder,Quadrature Encoder, with the PIC.
>> There was one idea with just shifting the the input bits and making
>> a compare to decide what action to be done next.
>> So my questions who posted this, and can this person pls
>> repost it again ?
>
>
>I posted one that compared current state with previous state
>some months ago either here or on comp.embeded. Essentially it
>was a finite state machine that ran asynchronously to the
>incoming quadrate data. The main advantages of it (and Bob Blick's
>earlier post as well) is that it requires no signal conditioning for
>debouncing the incoming lines and is not execution time critical.
>
>The decoder software self corrects incoming noise. The
>application sees the position indication as toggling between two
>adjacent positions with no accumulated error.
>
>
>Walter Banks
>