Picking zero-crossing resistors for phase fired control

[The Master]

Hi, I've done several projects in the past using phase fired control to dim AC loads. I've always used 2 diodes from the supply transformer to give me a DC ripple then a 10K resistor to connect that into a PIC and a 100K pull-down resistor. I'm now making a new project using a PIC24FJ64GA002 which operates at 3.3V rather than 5V. When designing the circuit I realised that I've just been using 10K and 100K without giving any thought as to why and that they might not be suitable for lower voltage PICs.

Is there a calculation I can use to figure out what value the resistors should be based on the supply voltage and the maximum voltage/current rating of the PIC pins?

[Jerson]

A simple voltage divider equation is all you need.

Vout = Vin x Rbot/(Rtop + Rbot)

Since you're feeding into a 3.3V device, the V/2 point is 1.65V. So, you need to aim for a Vout between 1.65 and 3.3V to keep the PIC happy. At higher input voltages you might get the PIC to frown at you or even give up.

[Ioannis]

I always had a doubt if the secondary of a transformer is in phase with the primary.

From a test I made on an unloaded one, it seems that it is. But I want to know if loaded stil is in phase.

Also toroids and classic E-I core are the same regarding phase?

Ioannis

[The Master]

Ahh, I never thought of it as a divider before. I've always thought of it as a limiting resistor and a pulldown resistor but by the looks of things it doesn't even matter which resistor has the biggest value as long as the voltage comes out right.

I'll have a play around with it so I can understand exactly how the maths is working and I'll test it out on my scope first to keep the PICs happy.

Ioannis: I can see how a coil would give a slight delay but I am only dimming lights with it and as long as I keep the timings right then everything works as intended. None of the coils I've used have been drastically out of phase. Even if they were then it's only a matter of adjusting the delay between the detected zero-crossing and resetting the outputs.