Scott,

Firstly, where does GND2 come from? If the logic and the limit switches share the same gnd and/or 5V, using an opto does not buy you anything.


With your circuit, a high voltage pulse enters via the limit switch wires, it goes through the polyswitch and though the zener to the 5V rail. Because it is a fast pule, the polyswitch does nothing (it has to heat up to operate. polyswitches are only good for longer term overloads) and whatever is connected to the 5V rail gets zapped - not good.

What you need to do first is to use a RC network to soak up the nasties then use the transzorb to clip whats left of the spike. With MOVs and transzorbs you have to be aware of the energy ratings of these otherwise they will easily fail is misused. Because of your wiring, you are virtually guaranteed to have spikes on you input circuit, you would be best to do some measurements of the energy of these spikes to ensure your circuitry can adequately cope with it. With the above circuit I've described, the capacitor looks like a brick wall to the fast spike and the spike's energy is dissipated in the series resistor. That is why the others where asking about the response time required for your switch. Depending on the amount of energy to dissiapte, the resistor can be sized appropriately. The transzorb then has a fairly easy life just mopping up a small amount of energy and thus will ensure its reliabiliy. You can include the series polyswitch if you want to cope with longer term overloads like someone connecting 115VAC into the input.

A relay solves all your problems.......