Hi Rugged,
Your idea to make ATP is not bad at all, I'd say! Here's a very simplified description of how plants generate ATP:
There's an enzyme called ATP synthase that is embedded in a membrane. There's a pool of ADP and phosphate on one side of the membrane, and some of it stick to the enzyme. To get ATP, the ADP and phosphate must be welded together, and the cell uses a cunning plan to accomplish this. It actually creates an electric potential over the membrane, which attracts protons (H+ ions) to cross the membrane. The only spot where the protons can cross it, is through the ATP synthase enzyme. The flow of protons through the enzyme generates enough energy to weld the ADP and phosphate together.
When ATP is "used up", it gets split up again in ADP and phosphate. So it can be used again to generate new ATP.
So there you go, with some of the natural enzyme and a membrane your machine would already work! You could think of ATP synthase as a kind of energy converter: it converts electrical energy (the electrical potential) into chemical energy (ATP).
You also asked how fast this would work, and that's a more difficult question. The speed of the enzyme is not a problem, since enzymes typically work very fast. But of course, when the enzyme does not have a plentiful supply of ADP and phosphate, it can't do much work. Also, if there's a lot of ATP present around the enzyme, it will stop working. So your job will be to make sure the newly made ATP is transported away from the enzyme, to the place where the energy is needed, and to return the ADP and phosphate to the enzyme.
If you'd like more information, try to look up "ATP synthase" or "chemiosmosis".
Good luck!
Airmid.