Post by thread:Reading ternary inputs

This may be old hat. Anyone care to critique my thought processes. Ternary (3 state) inputs are handy for increasing the number of options that one can get from strapped option inputs. eg 3 binary inputs give 8 states (2^3). 3 ternary inputs give 27 states (3^3) I had a discussion today with a friend who said an input could be read as high, low or floating by using a single resistor to any of these inputs and perhaps a hi value pullup if the pin does not have an internal pullup. I suggested a small value capacitor was also needed at the pin. He commented ~~ that eg an input strapped to ground "could not be pulled up so its state is known". Hmm, comes clear as I type. He works often in the 8031 world to which I am not accustomed. They have I believe a "pseudo I/O" type of pin which may be legally driven against a hi/lo load and which is used as an input by simply setting it high and pulling it down against the weakish high drive. I may have detail wrong but I believe this is the case in essence. With a high current pin this is illegal (and dangerous to the cpu's health). So, my ternary reading then goes like this. 10k from pin to ternary input 1M from pin to eg high 100 pF from pin to ground. 10K and 100 pF form a 1uS delay (choose actual value to suit code). 1M and 100 pF form a long (100uS) delay. So - Set pin to output - Drive pin high - Set pin to input - Read pin ->Sa - Set pin to output - Drive pin high - Set pin to input - Read pin -> Sb 10k/100 pF need to have a delay safely less than time to set pin to read and read pin. Values will need to be changed for most processors. If Sa = high and Sb = low then input is O/C If Sa = low and Sb = low then input is low If Sa = high and Sb = high then input is high. Because: A floating pin will retain the last written state due to slow pin time constant from 1M. A pin held during 10K will be temporarily driven to the opposite state by a write but will return to the externally driven state after a few time constants The large value pullup returns the pin to an eg high if it left floating as an input. Not needed if user exercise proper discipline and leaves port as an output most of the time. This adds power dissipation (depending on input setting). Any comments? Any better ways to do this for a true bidirectional high current capable port pin. Russell McMahon _____________________________ >From other worlds - http://www.easttimor.com http://www.sudan.com What can one man* do? Help the hungry at no cost to yourself! at http://www.thehungersite.com/ (* - or woman, child or internet enabled intelligent entity :-))

Russell, How about this: Have one pin per ternay input PLUS a single additional pin. Feed your inputs directly into your ternary input pins, and then have a 10K pull-up from each pin to the single additional pin. To read your inputs, first pull all of them high with the pull-ups. Read the values. Then pull all low, and read the values. The ones that changed when you changed the pull-up state are tri-state. The ones which were always low are low, and the ones that were always high are high. Alternatively, use PORTB or other such port which has internal pull-ups. You can eliminate the additional pin, just toggle the internal pull-ups. Sean At 12:10 AM 2/26/00 +1300, you wrote: >Roger Hawkins <spam_OUTrdhawkinsTakeThisOuTwbmpl.com.au>,bill roberts <.....billyleebobKILLspam@spam@home.com>, ADR <infoKILLspamadr.co.nz> >To: .....PICLISTKILLspam.....MITVMA.MIT.EDU > >This may be old hat. >Anyone care to critique my thought processes. > | | Sean Breheny | Amateur Radio Callsign: KA3YXM | Electrical Engineering Student \--------------=---------------- Save lives, please look at http://www.all.org Personal page: http://www.people.cornell.edu/pages/shb7 EraseMEshb7spam_OUTTakeThisOuTcornell.edu ICQ #: 3329174

Reading ternary inputs. Sean suggested applying inputs to pins directly plus resistors from each input to an "extra i/o pin" So Yes - this works with the use of 1 extra pin plus 1xR per pin and no need for caps or rc timing. One algorithm could be - Extra pin hi Read input to VA Extra pin low Read input to VB VC = Va XOR VB - high bits are tristates Vd = Va XOR VC - high bits are highs Vd = Vb XOR Complement(Vc) - high bits are lows VC, Vd etc aren't necessarily actual registers- just intermediate steps in process. Efficiency Pins 1 2 3 4 Levels binary 2 4 8 16 Levels ternary 3 9 27 81 Levels as above - 3 9 27 Better than binary at 3 pins. Always 1 pin worse than true ternary (of course). Russell McMahon {Original Message removed}