I still despise using a ICE for this, but it's the most practical and feasable use we have right now. I'd much prefer H fuel cells (once we can cheaply seperate H2 and O2) or an alternative engine (higher efficency). An ICE is ~30% efficient at best , for several key flaws, like having to change direction (piston up-down) in its motion, which wastes a lot of kinetic energy. But for practical reasons, ICEs will have to do for this.
Your idea places a battery in between the cells (in wiring, at least), to handle the electrical and other losses, which is charged by an alternator powered by the engine. If you're interested, contact me (MSN or AOL- orangecrusader14 ) about my design to gain power to charge the battery (with luck, at a rate equal to greater than chemical or electrical losses from the cells) without having to use the ICE's alternator and sapping more power from it. It's very simple (one moving part, not counting the alternator, which would make it 2 moving parts, without gearing), and nearly idiot-proof, I believe. As long as the metal-acid reaction can generate a decent amount of heat (for a decent amount of time, even a minute (or less) should be suffucient) this system will work.
dkmacmillan, I'll give a crack at answering your questions.
The system, for safety, will require a bubbler, and a fairly large one at that (always want a small reserve, not simply on a produce-consume basis. Once the battery is disconnected (should be hooked up to car ignition at some point for a built-in on/off switch, or we can place one on the dash), and the cells cannot be charged, then the reaction will stop itself, and the active cell will continiue to produce hydrogen until the chems are used up. This excess hydroxy will be stored for the next time the car is started, since that takes up a bit more fuel than usual, and will give the system a bit of extra time to start up. It would be perfect to have the system shut down and start up with one cell fully charged, and ready to produce hydrogen, but I can't think of a foolproof way to hold off the reaction without wasting power...
All that's created in this system as a gas is hydrogen and oxygen, and no significant other things, except for what comes into the engine from outside sources (if needed). We can also place some filters, if there are any contaminants.
Cost is a tough thing to figure out. The chems are actually very cheap (rust, and a quicklime, both really cheap. The acid is a bit more of an investment, but all in all it should cost (considering all the components are cheap and easily available) no more than several hundred dollars, having to buy every single thing yourself. Getting a good acid-proof tank (getting a simple metal one coated is easy) costs a bit more, but these are all one-time costs, and the chems will rarely have to maintained or replaced.
Seeing as water costs far, far less than gas (or free if you have a source), fuel costs are near nothing. Pure water would be preffered, and although that's a bit more costly, it's still available for a few bucks for 20 L or so, a lot less than gas, going on a liter (or gallon) by liter basis. Initial costs are the biggest cost for this, and very little otherwise. Should be able to pay for itself within a few months, less than a year, I think.
This can be used in combination with a gas system, as a booster, or handle a large portion if gas use. The advantage of this over a usual system is that it isn't sapping power all of the time, and it creates its own as well. The chems are more or less a one-time buy, and have to be topped off or maintained rarely (hopefully), while electrolytes and other parts of a usual system have to be continually topped off and monitored, keeping levels at the right point. With this, you only add water, nothing else.
