Thanks for all the replies. I got it working, and decided to go take a break for a while. I didn't go back and try the 4n35 opto again, I plan on doing that really soon. I don't know what I was doing when I thought I was only getting 2V for highs coming out of the Uno. Too much stress and my 5V bench supply decided to fail and take my Uno and a few other chips with it the night before. That and too little sleep and too much stress I guess. Oh, and I'm rusty on some aspects of circuit design, just like the 20 year old power supply that failed.
First off is the wireless. I agree, but the design requirements are no RF. Wireless would solve many other issue we have.
The main reason for converting to 12V was to not be using the 5V clean shielded supply as an output anywhere, or expose the output pin to the environment. Everything else is optoisolated, I will figure this out later as I can't use a isolated 5V supply with this circuit. This signal wire will go a few feet along a pair of 4 gauge wires that switch 150A, and the whole 12V supply and other signal wires come through a 15ft track with those 4 gauge wires.
This circuit i attached works. It does provide some isolation for the 328P plus a little with the mosfet. My down fall on the previous circuit was that I didn't use a resistor to pull the base of the 3906 back up to shut it off quick. Trying to saturate the 3904 was also mistake. This might be the whole reason the 4n35 didn't work.
Correct me if I am wrong but the circuit I attached is more immune to noise since it is a current, at least it protects the data in pin on the strip. And shields the pin of the Atmel.
Qdeathstar:
Also, chips with a clock line, like ws2801s are more noise resistant, and you can slowdown their click speed...
Good idea, sadly too late. We'll discuss the last guy at a later date.
EDIT: Found the 2V output issue. Scope probe that reads reads right on dc voltages, but not always at higher frequency.
