The past year I've been building a crude circuit etching device, but I recently finished and started doing other things...
Last Tuesday I was hanging an old clock above the toilet in my bathroom, until I slipped, fell and hit my head on the porcelain.
When I came to, I designed this:
A very simple circuit that uses two wires from the Arduino to boost input voltages of 1-12Vdc up to +50Vdc (approx. 1/4W). The Arduino regulates the output, keeping it constant under varying loads. It's designed more as a proof of concept, than to power anything specific. But, as-is it could probably handle a small LED backlight or other lower power device. It also doesn't (completely) tie up the Arduino, so other tasks could also be going on.
The circuit and software both have room for improvement. I was able to get nearly 1W by using a darlington pair.
I haven't put a link in the interfacing with hardware section yet (if that's where it belongs). I'll wait until somebody here volunteers to build it, and lets me know how it works for them.
Very cool. If one was to swap the direction of the diode D1 and cap C1 would it be able to generate negative voltages? That would be handy for analog op amps where one might need a negative rail.
Edit: never mind, the Aurduino A/D can not deal with negative voltages on it's A/D pin so it couldn't read the feedback voltage, Duh..
For a negative booster, there are some pretty big differences. The transistor switches the input voltage, and the inductor is instead always tied to ground. The rest of the circuit is also goofy, with the diode between the transistor's emitter and the positive side of the inductor (the side not tied to ground). And the diode and capacitor polarities are reversed. I also think an NPN wont cut it. It'll be necessary to use a compound transistor (sziklai pair) to control the input voltage.
But, like you mention, regulating it will pose the biggest problem.
In the source code I put a note about using a depletion mode mosfet to convert negative Aout to a positive voltage. I don't know if it'll work well.
There might be a simpler way, too. Some addition to the voltage divider and a pull-up resistor to +5V. I'll have to think about this.