Just a thought

Why aren't there any flashlights that charge the battery using the excess heat the emitter produces?

Thermoelectric charging should increase run time by at least 100%...

Is there any flashlight that does so? Not budget, even. :)

Because heat to electricity conversion is awfully inefficient and the implementation isn't trivial (at least not in a torch). Basically the steam engine principle is still the way to go in this manner (but using other liquids than H2O) afaik.

So what we would need would be a liquid that boils at a quitable temperature, say something like alcohol (there's probably a better choice).

So then we need to harness the power of the expanding vapor with a turbine and dynamo. That power then needs to be used either to help run the light or to charge the battery, so you'll probably need some kind of microcontroller in there to.

Then finally you need to condenss the vapor so it can be reused.

Certainly possible, but not easy and I've no idea how efficient it would be.

Surely a Peltier diode in reverse should be usable. To a limited extent, but at least it'll fit in a hand-held light. The thing is; we want to keep the temperature differential as low as possible which on simple thermodynamic grounds means efficiency will be very low.

We have a hard upper limit on temperature - the LED will die if it gets hotter than that. We also have a lower limit - what our hands can tolerate so we have a maximum temperature differential of under 100oC. I'd be interested to see if a Peltier could output current and voltage sufficient to actually put some charge back in a cell.

All heat to work methods are of limited efficiency when dealing with small temperature differentials. The physics of this is covered in thermodynamics of heat engines. The maximal efficiency of a Carnot cycle (theoretical max) is not really that high between room temperature and led temp.

Since you don't want to carry around a mechanical engine, you want the thermalcoupler that Don mentioned. It's just incredibly cost ineffective for a budget light and not really effective in our application.

They're actually used with deep space probes in these: http://en.wikipedia.org/wiki/Radioisotope_thermoelectric_generator, which have something more potent than an led as heating element.

Hey! No one told me you wanted a sensible option :P

I'd rather see a light with a solar panel built into the handle, then I could just leave it on the window sill all day and it would have a full charge every night. Heck, they do that with $2 garden lights at Walmart. At 5am they are still burning brightly.

But their not pushing XM-L outputs... Personally id rather be carrying another set of batteries, and have a solar charger separate to the light.

But they are charging an AA battery, doesn't matter what the battery is pushing.

Erm, most of those solar garden lights use 500mAh Ni-CDs, and they take 12-14 hours to charge fully.

The MIT guys may have made it a little more viable... http://www.gizmag.com/energy-efficiency-breakthroughs-at-mit-and-berkeley/8790/