Post by thread:Decoding grey code with a pic.

Hi, I need to get position info from a Stegman single turn absolute position encoder. Getting the info is only a matter of clocking the unit and shifting in the 24 bit grey code result. Does anybody know the fastest way of converting this to binary using a pic. A pic is my preferred solution as I would like to do some position comparision (ie. accept position info from another microcontroller and read the encoder until that position has been reached.) Other suggestions are welcome though. Thanks. Joe _________________________________________________________________ Joe McCauley Physics Department, Tel: 353-1-6082218 Trinity College, Fax: 353-1-6711759 Dublin 2, Email: spam_OUTpmcculeyTakeThisOuTmail.tcd.ie Ireland.

Hi Joe and all, At 09:58 14/08/96 +0100, you wrote: >I need to get position info from a Stegman single turn absolute position >encoder. Getting the info is only a matter of clocking the unit and >shifting in the 24 bit grey code result. > >Does anybody know the fastest way of converting this to binary using >a pic. One easy way to convert the Gray code to binary is to do the XOR of the current word with the shifthed_rigth version (lsb droped, 0 to msb) of this same word. Best regards: -- Roberto Deza Asensio | .....rdezaKILLspam@spam@popmail.cti.unav.es Universidad de Navarra | rdezaKILLspamcun.unav.es Centro de Proceso de Datos | .....rdaKILLspam.....cpd.unav.es

Hi Rick and all, At 07:16 14/08/96 -0500, you wrote: >> One easy way to convert the Gray code to binary is to do the XOR of the >> current word with the shifthed_rigth version (lsb droped, 0 to msb) of this >> same word. > >That's wonderful! :-) I can use it! > >You wouldn't happen to know an easy way to *generate* grey code, >would you? Yes, I know the related trick: To get the Grey code of a binary number, you do the XOR of the binary word with the shifted_left version (0 to lsb, msb droped) of this same word. I know this "old tricks" from a book of the year 74 related to RTL & DTL logic (Resistor/Diode Transistor Logic...), but they have currently use. Best regards: P.D. I post this reply also to the list, as perhaps it will be of use to someone. -- Roberto Deza Asensio | EraseMErdezaspam_OUTTakeThisOuTpopmail.cti.unav.es Universidad de Navarra | rdezaspam_OUTcun.unav.es Centro de Proceso de Datos | @spam@rdaKILLspamcpd.unav.es

Joe McCauley <KILLspamPICLISTKILLspamMITVMA.MIT.EDU> wrote: > Getting the info is only a matter of clocking the unit and shifting > in the 24 bit grey code result. > > Does anybody know the fastest way of converting this to binary using > a pic. Joe: See the answer to Question #5, in the "Other Processors/ Miscellaneous" section of the "Answers" area on my company's web page. -Andy Andrew Warren - RemoveMEfastfwdTakeThisOuTix.netcom.com Fast Forward Engineering, Vista, California http://www.geocities.com/SiliconValley/2499

I am sorry to be a kill joy but .... the xor ideas are not quite complete, or ambiguous. To generate Gray code *is* just a mater of using a shifted binary copy, the reverse isnt true. For example assume the Gray value is 7, thus the expected binary is 7 xor 3 (7 right shift) = 4. However a binary 4 xor 8 (4 shift left) is 0c not 07. Gray code should be generated by Gray = binary xor (binary * 2) But binary is extracted as follows:- For n the nth bit of n bits, B being the binary output bit, and G the Gray input bit 1) B(n) = G(n) 2) B(n-1) = G(n-1) xor B(n) 3) B(n-2) = G(n-2) xor B(n-1) 4) B(n-3) = G(n-3) xor B(n-2) ...... 5) B(0) = G(0) xor B(1) From this it is clear that the binary value depends on the previous higher binary bit. The good news is if the serial comes in most significant bit first you can do the conversion on the fly. For the serious chip by chip design of finite state machines transitions from states ordered according to Gray code result in minimal wiring, because the Hamming distance is always 1. What is user experience of the UK Forrest PICDE software like? Happy coding. Mike

Mike Halliday <spamBeGonePICLISTspamBeGoneMITVMA.MIT.EDU> wrote: > Gray code should be generated by Gray = binary xor (binary * 2) Mike: This is ALMOST correct. Gray code is generated by: Gray = binary xor (binary / 2) That is, you must shift the input number RIGHT one position before XORing it with the original number. {Quote hidden}> But binary is extracted as follows:- > > For n the nth bit of n bits, B being the binary output bit, and G > the Gray input bit > > 1) B(n) = G(n) > 2) B(n-1) = G(n-1) xor B(n) > 3) B(n-2) = G(n-2) xor B(n-1) > 4) B(n-3) = G(n-3) xor B(n-2) > ...... > 5) B(0) = G(0) xor B(1) Another way of looking at it, which may be easier to implement, is: 1. Reading from left to right, accept all the "0"'s and the first "1". 2. INVERT = TRUE. 3. Read the next bit. If INVERT is true, invert that bit. Otherwise, accept it as-is. 4. If the bit you just read was a "0", loop back to Step 3. 5. Otherwise, invert the state of the INVERT flag, then loop back to Step 3. -Andy Andrew Warren - TakeThisOuTfastfwdEraseMEspam_OUTix.netcom.com Fast Forward Engineering, Vista, California http://www.geocities.com/SiliconValley/2499

I just wrote, on the subject of decoding Gray code: > 1. Reading from left to right, accept all the "0"'s and the > first "1". > 2. INVERT = TRUE. > 3. Read the next bit. If INVERT is true, invert that bit. > Otherwise, accept it as-is. > 4. If the bit you just read was a "0", loop back to Step 3. > 5. Otherwise, invert the state of the INVERT flag, then loop > back to Step 3. I wasn't thinking as clearly as I should have been... Here's a slightly-more-generalized version of the above: 1. INVERT = FALSE. 2. Read the next (reading left-to-right) bit. If INVERT is true, invert that bit. Otherwise, accept it as-is. 3. If the bit you just read was a "1", invert the state of the INVERT flag. 4. Loop back to Step 2. Off the top of my head, an 8-bit Gray-code decoder could look like this: FLAGS EQU [any register] COUNT EQU [any other register] GRAY EQU [yet another register] #DEFINE INVERT FLAGS,0 ;TO MAKE THE CODE EFFICIENT, THIS ;FLAG MUST BE DEFINED AT ;BIT-POSITION 0. ; ENTER WITH THE GRAY-CODED NUMBER IN "GRAY". EXITS WITH THE ; DECODED NUMBER IN "GRAY". DECODE: BCF INVERT ;START WITH "INVERT" FALSE. MOVLW 8 ;SETUP TO DECODE 8 BITS. MOVWF COUNT ; LOOP: RLF GRAY,W ;ROTATE THE NEXT BIT FROM BIT- RLF GRAY ;POSITION 7 TO BIT-POSITION 0, ;LEAVING A COPY OF IT IN THE ;CARRY. MOVLW 00000001B ;PREPARE TO INVERT THE BIT, IF ;NECESSARY. IF "INVERT" WAS ;DEFINED AT BIT-POSITION 0, THIS ;INSTRUCTION ALSO PREPARES US FOR ;INVERTING THE "INVERT" FLAG, ;SHOULD THAT PROVE NECESSARY. BTFSC INVERT ;IF WE SHOULDN'T INVERT THE BIT ;WE JUST READ, SKIP AHEAD. XORWF GRAY ;OTHERWISE, INVERT THE BIT. SKPNC ;IF THE BIT WE JUST READ WAS A ;"0", SKIP AHEAD WITHOUT INVERTING ;THE STATE OF THE "INVERT" FLAG. XORWF FLAGS ;OTHERWISE, INVERT THE STATE OF ;THE "INVERT" FLAG. DECFSZ COUNT ;HAVE WE DONE ALL 8 BITS? GOTO LOOP ;IF NOT, LOOP BACK. ; We're done... The decoded value is in "GRAY". Twelve instructions, 82 cycles. -Andy Andrew Warren - RemoveMEfastfwdTakeThisOuTix.netcom.com Fast Forward Engineering, Vista, California http://www.geocities.com/SiliconValley/2499