My design goal is to construct a portable electric generator that is powered by waste heat, such as from a camp fire or stove. This is designed to be an alternate to solar panels, relying on sunshine, which is not always available. Heat, however, can be generated on demand. The technology used here is identical to that used by NASA on all deep-space probes. The RTG (Radioisotope Thermoelectric Generator) uses heat generated by the decay of radioactive plutonium 238 (non-weapons grade), which in turn, is converted to electricity. Of course, I can't get plutonium, non-weapons grade or otherwise. If I could, I'd have traded in my Jeep for a DeLorean by now. But the thermoelectric generator chips are easy to obtain.
I bought some HT4-12-30s from Melcor. I wired four in series, and placed this assembly on a griddle on my stove. After a quick test to determine polarity, I wired in a green LED to give me a visual indication. It was exciting to watch the LED start to glow dimly, then brightly. I was planning to buy some more to make a module large enough to generate practical amounts of power. But I bought these chips from Melcor's clearance store. I decided to look for other sources before committing to buy from a limited supply. And I found what I was looking for. Hi-Z makes various modules, and after a bit of consideration, I decided to go with their 20 watt unit (HZ-20). It also had the added benefit of taking up less area than that needed for enough Melcor chips (For the record, they do make larger modules.) to generate the same power.
Now that I have selected a TEG, I have to design a mounting platform for it and the power electronics, all with unique design challenges. First, the voltage output of the TEG is lower than my target of 12 volts. I will have to build a DC-DC converter to step up the voltage.
Next, for this module to work, one side has to be hot, and the other side as cool as possible. My original idea was to sandwich the module between two plates of metal. Copper was my choice for the hot side, and aluminum for the base. But now, how to hold everything together? In a module designed to cool my telescope mirror, this was not a problem- nylon screws were near-perfect both mechanically and thermally. Here though, nylon is unusable. The hot side is rated at 250 degrees celcius. That's even hot enough to melt 60/40 solder (another possible problem). I solved the problem by finding a source for teflon screws. I originally planned to separate the two plates with ceramic tiles, but even those conduct heat (after all, the insulators on the chips is ceramic). So, I will simply seal the edges around the two plates with a high-temperature RTV silicone rubber. Heat is also a problem for wiring- insulation isn't usually required to survive such extremes. Once again, teflon was the answer- I found teflon-coated wire for the module.
This module is intended to be heated from the source by radiation and convection, rather than by conduction. To maximize absorption, the hot face side needs to be as close to a blackbody (In physics, a blackbody is a surface that is both an ideal absorber and emitter of energy.) as possible. So, this side will be black anodized or painted with a high-temperature black paint. To minimize heat getting to the cold side, the plates are separated by air or poor heat conductors (teflon, RTV). Heat transfer by radiation is possible, and to minimize this, I am considering polishing the facing aluminum side, and silver plating and polishing the copper side. The opposite of a blackbody is a surface that is both a poor absorber and poor emitter- a reflector. I have a simple way to test how close (or not) a surface is to being a blackbody. In a darkroom, illuminate the test surface with infrared LEDs, and using a digital camera (the sensor is made of silicon, which is very sensitive to infrared), observe how strong the reflections are.
So far, my first tests have been a mixed bag of sucess and failure. The HZ-20 has not put out full voltage yet, and that has largely been to blame on the poor thermal interface. The metal I've been using is just stock metal plate, which is less than perfectly flat. Before I will continue with the HZ-20, I will have to machine all pertinent surfaces to 0.001" tolerance (which Melcor also recommends). Another problem is my "griddle" design. The assembly is supported over the heat source, with the collector side down. After a while, the heatsink is hot, and power output drops. The Melcor chips have been more successful to date, but heat elimination is still a problem.
After my failures, I came up with two new designs- the "funnel", and the "Christmas tree". The "funnel" is basically an inverted "griddle" with a heat-collecting boom. The boom, and not the HZ-20 TEG module, is over the heat source. With the "Christmas tree", the TEGs will be in groups of 4 HT4-12-30s, mounted vertically on a hexagonal base. This base will either be seated on a small cookstove, or will have the burner pirated from one incorporated into it.
I have continued my experiments, and have not had the results that I've hoped for. With the Hi-Z module, I just can't get it hot enough on the stove, and it requires compressive loading of 200 PSI to make it effective. I will also have to come up with a different mounting scheme- it will be hard to reach 200 PSI with teflon screws. Heat is also likely the problem of the Melcor chip's lackluster performance.
I wanted to see how effective having the hot-side face of the heat sink reflective versus blackbody. I constructed a fixture with a plate and temperature sensor (LM335Z) that I could suspend over a hot plate. When the blackbody face was exposed to the heat fpr eleven minutes, I measured a temperature gain of 26.1 degrees celcius. With the reflective face was exposed, I measured a gain of 11.5 degrees. To further disrupt heat transferance, I placed a sheet of 1/8 inch thick teflon between the hot plate and test fixture. Temperature gain was reduced to 5.1 degrees.
My design goal is to construct a portable electric generator that is powered by waste heat, such as from a camp fire or stove. This is designed to be an alternate to solar panels, relying on sunshine, which is not always available. Heat, however, can be generated on demand. The technology used here is identical to that used by NASA on all deep-space probes. The RTG (Radioisotope Thermoelectric Generator) uses heat generated by the decay of radioactive plutonium 238 (non-weapons grade), which in turn, is converted to electricity. Of course, I can't get plutonium, non-weapons grade or otherwise. If I could, I'd have traded in my Jeep for a DeLorean by now. But the thermoelectric generator chips are easy to obtain.
