On Mon, Sep 24, 2001 at 01:57:43PM +1000, David Duffy wrote:
> Dave wrote:
> >I am wondering if it is possible to multiplex a push button switch
> >input and an LED output on the same I/O pin.  My idea is that since
> >the switch input only needs to be sampled every 2ms, it can be set as
> >an output for an LED when not sampling the switch.  Also, when the pin
> >is switched to input to sample the switch, the time taken to do so is
> >so minimal, that the LED would not even flicker.  I figure the I/O pin
> >only needs to be set as an input for just a few assembly instructions.
> >Hardly noticeable, right?
> >
> >So, I've got a couple questions.  First off, is the idea described
> >above even possible?  And finally, what would the resulting circuit
> >look like?  I've come up with this:
> >
> >               LED   R1
> >            +--|<|--/\/\----o +5V
> >-----+     |
> >  I/O |     |
> >  Pin +-----+
> >      |     |          \
> >-----+     |           \
> >            +--/\/\--+   +--o Gnd
> >                R2    SW1
> >
> >However, I cannot seem to come up with R1 and R2 values that work when
> >the switch is closed, and the pin is setup as an input.
>
> Yes, this will work just fine if you delete R2 and make R1 about 330R.
> Use the TRIS register to put your i/o pin into the hi-impedance state
> to read it. Leave the i/o bit cleared when as output to make it like an
> open-collector output in parallel with the switch. I do this all the time.
> The LED will illuminate whenever you press the switch but in most
> applications this is not really a problem. Call it an LED test feature!

Hmm... thanks for the suggestion, but I really wanted to keep the LED
independent from the switch.  In this application it's not possible to
have the LED turn on and off with the switch.  Is this possible?

-Dave

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.