Re: ParPort output low - hi-Z or ground? opto to parallel

>
> I am looking at an external power supply thru a resostor on the
> isolated side of the opto. I can drive the LED with 10-15mA and
> attain isolation from the opto as the power supply and resistors

and

> LED side of the optos would be on a totally seperate circuit.
>
> On the parallel port side, It would appear that a non-buffered
> connection would have the port pins (10-13) connected to the opto

and

> the other side to ground.
>
> A state change of the inputs would result in a state change of the
> outputs.
>
> Would a pull-up or pull-down on the parallel port side offer any
> benifits ? and which would be best for the parallel port side of
> things ?
>
> Dave

I am assuming your reference to the isolated side is to the
transistor side? If you are using this as an isolated input to the
parallel port then the transistor side is connected to the parallel
port pins. The emitter (led) side needs to be driven by a separate
isolated power supply which can be any voltage as long as the series
current limit resistor is sized to set current to the correct
amount. Just divide the total LED volts minus the 1.2 nominal LED
volt drop by the target current in amps to get the resistor value.

On the transistor side you would connect the emitter to PC ground and
the collector to the parallel port pin input pin. A pullup resistor
of 4.7K to +5 (+3.3) would make sure that the pin goes high when the
transistor is off but may not be needed if there are already pullups
in place. Hook it up without the pullup and monitor the voltage on
on the collector with no signal on the opto led. If it has a solid
+4 or greater then you don't need a pullup.

I would not trust a parallel port pin to SOURCE 10 - 15ma. The
parallel port pins on a lot of PC mother boards are NMOS or CMOS and
are designed to be signal lines into a printer (high impedence). I
learned the lesson the hard way when I switched to a newer computer
on my engraving machine and it started doing strange and wonderful
things. In the PC world parallel port cables are supposed to be kept
to under 8' (I know, you can get more but its risky).

While I agree with Jon that running the led side with +5 on the anode
and +3.3 on the cathode side will work, that leaves about a milliamp
or so of "bias" current flowing when the opto is supposed to be off.
That's below the turnon threshold but it takes less noise to cause a
problem. For maximum noise immunity I would either run the led side
from the same voltage level as the parallel port or buffer the pin
output and shift the voltage level up to the +5 using open collector
buffers.