I prototyped a quick circuit like this one. I just connected a 5V DC supply through a 1K resistor to turn the FETs on and off. The output of the FETs are connected through a 25W light bulb to a 120VAC source.

The circuit works in that turning the 5V source on turns the bulb on and jumpering the gate side of the 1K resistor to the negative side turns it off except it only turns one of the FETs off. I presume only one is being turned off because my scope shows one half of a sine wave.

So I'm thinking that either jumpering that resistor isn't sufficient to turn both off or I have a bad FET. Correct?

If I have a bad FET, and it's the positive going part of the sine wave, can someone tell me how to tell which FET is bad?

EDIT: Perhaps answering my own question. The gate to source voltage is listed as +-20V so I'm trying to drive it with a logic level of 5V and perhaps that's the problem? If so, why does one FET turn off?

 

So I run the circuit with just one FET in at a time and both behave the same; I can turn one side of the AC waveform off but not the other. The only thing I can think is I may not have the drain and source hooked back-to-back as they should be. I'll rewire it in again and see what I find.

In the meantime, if anyone notices anything, please reply.

 

You see that diode drawn inside the FET in your schematic? That's a body diode, and it really is there; it becomes a half-wave rectifier in your case.

 

Yes, I know that, but it doesn't explain why wiring the two FETs back-to-back isn't turning the AC off in both directions when I apply the gate voltage.

 

You didn't mention the condition of the light bulb in each circumstance - illuminated or not.

 

You mentioned a CRO - Perhaps some details of the entire set up would prove informative. Is the mains supply transformer isolated?

 

My test circuit just uses a 5V power supply through a 1K resistor for the digital input. The 25W light bulb lights full when I turn it on. Shorting the input to the gates to 'ground', the negative side of the 5V power supply on the resistor side causes the negative peaks of the AC waveform to go away on my oscilloscope and the bulb appropriately dims.

The AC for the light bulb and FETs are fed directly from the main, as is the 5V power supply.

 

Embarrassingly, I found the issue was with a misplaced wire. Haven't a clue how I missed it but there it was and my circuit now works.

 

Haha, sweet. Happens to the best of us.

Be careful, by the way. Loose wires are not compatible with mains-direct circuits.

 

To see the internal structure of CD4066 on the top left of the page 6.