Post by thread:False HI on Pin, Trying to use pin for both input

I'm trying to use one I/O for both input and output (RA1 on a 16C84). Software - Every 8sec or so the pin is changed to an output and set high to drive an LED for about 200msec....then turned off and changed back to an input. Between the LED blinks I poll the input to check whether it's HI. To debounce, the input must be HI once and again 20msec later. I've got the weak pull-ups disabled. Hardware - The pin has a 470 ohm resistor and (Smurf-) blue LED to ground. Also a pushbutton switch and 10k resistor to V+. What should happen: If pushbutton not pressed - LED should just blink every 8 sec. If pushbutton is pressed - start other routine. What does happen: Other routine starts by itself. Even if 10k resistor is pulled out of the circuit and circuit left open (so pushbutton is not malfunctioning). I added a 1s delay between turning off the LED and reading the input and problem still occurs. (90% sure the delay is working properly.) If I ground the pin, the error does not occur (and LED does not blink). There is a 20msec delay between turning off the LED and reading the pin. Between LED blinks, the pin is read every 20 msec. It appears that a LED blink is not necessary to make the error occur. It's intermittent. :) At other times the other routine functions correctly when the pushbutton is pressed. If relevant, I'm working on a breadboard. Can anyone tell me I'm doing wrong? I suspect it's something simple but I haven't been able to figure it out. Dave Scott

Dave, Did you switch banks for each time you changed from in to out. Gordon ----- Original Message ----- From: Dave & Ann Scott <spam_OUTandscottTakeThisOuTLOGANTELE.COM> To: <.....PICLISTKILLspam@spam@MITVMA.MIT.EDU> Sent: Friday, August 13, 1999 12:37 PM Subject: False HI on Pin, Trying to use pin for both input & output > I'm trying to use one I/O for both input and output (RA1 on a 16C84). > > Software - Every 8sec or so the pin is changed to an output and set high to {Quote hidden}> drive an LED for about 200msec....then turned off and changed back to an > input. Between the LED blinks I poll the input to check whether it's HI. > To debounce, the input must be HI once and again 20msec later. I've got > the weak pull-ups disabled. > > Hardware - The pin has a 470 ohm resistor and (Smurf-) blue LED to ground. > Also a pushbutton switch and 10k resistor to V+. > > What should happen: > If pushbutton not pressed - LED should just blink every 8 sec. > If pushbutton is pressed - start other routine. > > What does happen: Other routine starts by itself. Even if 10k resistor is {Quote hidden}> pulled out of the circuit and circuit left open (so pushbutton is not > malfunctioning). > > I added a 1s delay between turning off the LED and reading the input and > problem still occurs. (90% sure the delay is working properly.) > > If I ground the pin, the error does not occur (and LED does not blink). > > There is a 20msec delay between turning off the LED and reading the pin. > Between LED blinks, the pin is read every 20 msec. It appears that a LED > blink is not necessary to make the error occur. > > It's intermittent. :) At other times the other routine functions > correctly when the pushbutton is pressed. > > If relevant, I'm working on a breadboard. > > Can anyone tell me I'm doing wrong? I suspect it's something simple but I > haven't been able to figure it out. > > Dave Scott

Dave, Try putting a pull down resistor (100K) from the pin to ground, or across the LED, as the I/P is floating - worse, the LED may be generating a (very very tiny) current from light. (I believe it can happen) I use the inverse (LED to vcc, push button low) of your circuit (with the extra resistor) on my graphic LCD checker:- http://ourworld.compuserve.com/homepages/steve_lawther/t6963tst.gif with no problems. hope it helps, Steve Lawther > {Original Message removed}

Yep. Here's turning off the LED: bcf PORTA, 1 bsf STATUS, RP0 movlw 0x00E ; 0E=0000 1110 ; RA0 is an output ; RA1, 2 & 3 are inputs movwf TRISA & 0x07F bcf STATUS, RP0 ---------- {Quote hidden}> From: Gordon Varney <varneyKILLspamCLAS.NET> > To: .....PICLISTKILLspam.....MITVMA.MIT.EDU > Subject: Re: False HI on Pin, Trying to use pin for both input & output > Date: Friday, August 13, 1999 12:39 PM > > Dave, > Did you switch banks for each time you changed from in to out. > > Gordon > > >

Problem solved.... :) (I hope.) > Steve Lawther replied: > Try putting a pull down resistor (100K) from the pin to ground, > or across the LED, as the I/P is floating - worse, the LED may > be generating a (very very tiny) current from light. (I believe it can > happen) I've only done a few quick tests but it appears to solve it. With the additional resistor it works correctly. Without it, I can trigger it with a flashlight! Also, never noticed the problem occurring during the night - always during the day. > I use the inverse (LED to vcc, push button low) of your circuit (with > the extra resistor) on my graphic LCD checker:- Is there any particular advantage to either arrangement? Thanks.....Dave Scott

> Van: Dave & Ann Scott <EraseMEandscottspam_OUTTakeThisOuTLOGANTELE.COM> > Aan: PICLISTspam_OUTMITVMA.MIT.EDU > Onderwerp: False HI on Pin, Trying to use pin for both input & output > Datum: vrijdag 13 augustus 1999 19:37 > > I'm trying to use one I/O for both input and output (RA1 on a 16C84). <Snip> > Hardware - The pin has a 470 ohm resistor and (Smurf-) blue LED to ground. > Also a pushbutton switch and 10k resistor to V+. > > What should happen: > If pushbutton not pressed - LED should just blink every 8 sec. > If pushbutton is pressed - start other routine. > > What does happen: Other routine starts by itself. Even if 10k resistor is > pulled out of the circuit and circuit left open (so pushbutton is not > malfunctioning). Hello Dave, You are testing the the I/O while the LED is Off. This means that the I/O line is floating (or rather High, due to the Chip-design). Reason ? Think of a LED as a Zener-diode. If it's ON, current flows. If it's OFF allmost _NO_ current flows (leakage only !), so it can't be used as a pull-down. Solution : Put a resistor parallel to the LED. How come I know this ? Simple: I stepped in the same pit-fall you did :-) Greetz, Rudy Wieser