Thermoelectric Solar Power


[Colin] has put together an instructable for a solar power generator that uses the thermoelectric effect instead of the photovoltaic (PV) effect. We have seen Peltier devices used in cooling cans, solder paste, backs, and hacked hard drives. This is the first hack we have seen where a Peltier device is used to generate electricity from heat, essentially running the device backwards. The thermoelectric effect is the same principle that is used to generate electricity in radioisotope thermoelectric generators used in deep space probes such as Cassini. What applications can you come up with to use the thermoelectric effect as a power source?

41 thoughts on “Thermoelectric Solar Power

  1. i wonder what kind of efficiency you can get out of this? would it be same as when running it normally? or do they function better as generators. I have to admit that I know very little about peltier.

  2. From the looks of it, not great. He says the theoretical maximum he could get get is about 2.3 watts, but never says if he actually gets there. He just says he got a peak voltage of 2.7.

    It is about the same as PV in terms of dollar-per-watt output, but this is much more difficult to maintain so it isn’t much competition in it’s current form. But like he said, this was a quick experiment. With refinement it could be put to useful work.

  3. How about using it on your cpu. Adapt it into the heatsink/cooling setup and use your cpu’s byproduct of heat to recirculate power back into the system in some way, even if it is just running a fan.

  4. Yeah it would be nice to see some real recorded data on the power output.

    It would be interesting to see an old satellite dish painted silver with the peliter element in the focal point, as much more energy could be captured this way.

    Somehow thermally connecting the ‘cool’ side of the peltier to the ground could give a better temperature differential.

    No point me trying it, I live in the UK :)

  5. Step it up to 5V and you should get around 300-400mA. Couple that with a USB port and you’ve got yourself a USB charger.

    This thing is a bit large though to be useful. If you’re anywhere without electricity, you’re probably hiking or backpacking, in which case a cheap solar cell would be more useful as a USB charger…

  6. I read somewhere that in the ’40s or ’50s that the Russians developed a radio that worked off a thermoelectric generator (whatever) hooked up to a coleman type lamp to run a radio, so mabye you could use this thing with a candle to run your Ipod during a blackout.

  7. I did a study on these devices in an engineering design class. We found that to offset costs in ~5 years, these would need to be placed in a temperature differential of about 2-300*C. These cells are usually ~.25″ thick, so maintaining that kind of temperature differential proved cumbersome and limited the potential applications. Still have a couple, they cool “soda” pretty quick and work great as cooling coasters.

  8. my thery was to take 5 or 8 peilters and sandwich them toghter every layer flipflopped from the last add power to half of the system to cool and heat put a heatsink on the top side and attach the other side to your exhaust system on your vehicle theroeticly if you had enough transfer you could take out your alternater and save fuel on power to weight

  9. You could put a heatpipe in the ground and run it up to the above ground peltier. Insulate the headpipe so that you can pull ~55deg ground temperature up without a pump. Now dont use a focused solar setup and instead do a vacuumed glass sandwich where the peltier is attached to a copper sheet and has an heat insulated backing and the heatpipe would touch the back of the peltier behind the heat insulator. on the front you would put some tiny spacers on the copper surface and put the glass on top of that. Use an epoxy and seal up the perimited between the glass and the insulated backing as the copper should be just a bit smaller so it fits inside. tap in a 1 way valve and pump the air out and/or fill the tiny space with Argon(which conducts heat very poorly vs air).

    Now you dont need any power to run a fan as the heatpump will try to push the heat into the ground. The copper plate will increase the surface area to collect more light and also not require the light to hit it directly. the argon/vacuum will help insulate the hot plate. the hot copper plate will likely get up to 200 degrees quickly giving a 65deg/200deg ratio which should generate a decent amount of juice.

    As long as you keep the temperature of the peltier within the specs of the device, it should run for many many years without fail. it is solid state and does not degrade by use, only by abuse and some natural decomposition of the materials from time.

    keep in mind that the act of producing electricity will actually reduce the temperature of the hot side as the heat is converted to electricity.

    I would imagine that cooling with the heat pipe will be more effective than the small fan pressing environmental temperature air over a heatsink.

  10. Thermoelectics are an area of quite a bit of interest at the moment, I’ve a few side projects dealing with them but they are currently horrifically inefficient (2-4%), but new materials are in the pipeline which should up it a lot. They’re worse in generation than power use and require fairly complex load balancing to work well – draw too much current from them and their voltage drops, their apparent thermal conductivity rises and they start to reduce their own temp difference and produce less power. Likewise too low a load and you generate plenty of voltage but at low current so low power again. Ideally you need to monitor and control the load and temperatures for peak power throughput at any given temp difference.

    Various car manufacturers like BMW are already working on fitting them as a replacement for an alternator in vehicles, reducing crank losses and recovering waste heat.

  11. I once wondered about using the thermoelectric effect to recover some of the waste heat from a car engine, but neither the oil or coolant loops in a regular engine get hot enough for this to be remotely efficient. The exhaust manifold might though.

  12. Man, I was working on this stuff 15 years ago, this brings back memories. My wife still laughs about my experiments on the kitchen counter when we were dating. I guess if she didn’t run away when I was geekin’ on that she was good to keep!

    I remember looking into automotive cooling (needed way too many Peltier’s and current to get enough BTUs), automotive electric production, even had a few can coolers in the works way before I saw them on the market. Shoulda done it myself…

  13. In smaller cooling applications it might make a very nice analog self-regulating cooler – the hotter it gets the more it’s cooled… but everyone seems to agree that it’s not terribly efficient – so car engines, homes, etc might want to consider something more realistic….

  14. Oh it’s been used already in another hack to power a PC fan using the heat from a small burner cut into a food can. The fan blew on the cold side to extract the heat faster, increasing a bit the efficiency.

    Here’s a way to build a motor powered entirely using the Seebeck effect : cut a dozen 10cm strips of copper and brass wire, and twist them together at their end so as to make loops of wire, one half copper, the other half brass. Make the loops elongated, with the twists far apart. Cut a disc out of cardboard and glue the wire loops all around it, one twist towards the center, the other twist sticking out of the disc. Always keep the copper/brass wires oriented the same way. The loops must have as much area as possible without touching each other. You should get something star-shaped, with 12 wire twists sticking out. Now take a pin and balance the rotor on top of it, so that it can rotate freely. Put a magnet under the cardboard disc, with its magnetic axis perpendicular to the disc. Lastly, bring a small candle so that the flame touches the wire twist sticking out just next to the magnet. After a while the rotor should start turning.

    Here’s how it works : the copper/brass twists form a thermocouple. When a temperature difference develops between the two twists, a voltage potential is created. The voltage is very small but since the two wires are basically in short circuit a relatively high current will be generated. This current will interact with the magnetic field and generate a Lorenz’s force, perpendicular to both the wires and the field, making the wheel turn. The next loops comes in contact with the flame and the cycle repeats.

    To make a better contact between the wires you can solder them. The nature of the solder is not important. A copper-solder-brass junction is equivalent from a thermoelectric point of view to a copper-brass junction.

    I hope this was intelligible. I’ll probably put some plans for this on my website one day when I feel less lazy than usual.

  15. Thermoelectrics are a niche tech. Careful application of the devices can be cost effective compared to other power supply systems. One mundane common non-peltier device is the safety thermocouple gas valve. With a less common similar system called a “power pile” thermostat. The circuit called a “Joule Thief” and some thermoelectric devices can power low watt systems like radio connected sensors or short range repeaters.

  16. This isn’t thermoelectric per say, more like thermomechanic. I see a few commenters here have mentioned cars and reclaiming heat from the engine.

    The easiest and to date far most efficient way of doing this? Add another step to the combustion cycle where water is sprayed into the chamber!

    Fuel and air is sucked in, compressed and explodes. The piston comes back up pushing out the exhaust (at which point the engine will be at ~1600 degrees Celsius). Insert a spray of water and some of the waste heat is spent flash-boiling it pushing the piston down again.

    The only problem is lubrication. Oil and water isn’t exactly known to work well together.

  17. You can surround the car engine with peltiers and use it to charge the battery. Of course I do not know how many peltiers you will need even if it will be sufficient to charge the car battery at this stage, but if it appears to be possible, you can get rid of the alternator, which draws 4-5 HP from the engine at least. this will bring you some gain with the power. Then of course it is necessary to compare the power gained to the power lost due to the added weight.

  18. we all hate mechanical parts because they wear, but there is nothing as efficient as heat pump, it strange that there isn’t much project about it, heat pumps perfect for camping it can work off solar or camp fire

  19. @ PlastBox

    I read about that in an issue of PopSci. The man who suggested it stated that the lube problem wouldn’t apply much to diesel engines( most additives of diesel fuels are oil-based).

    As for the article, how about using a air-cooled radiator coil to disperse the heat? Point the fan in its direction and keep it wrapped up in a box to shield it form the sun, you should get a better output.

  20. @Plastbox – The material and lubrication problems are vast, and the engine itself isnt anywhere near that sort of temperature, only the gas within it – the metal surfaces are kept surprisingly low.

    BMW are, if you read the tech papers on it, indeed working on exhaust manifold mounting the peltiers. The problem then is cooling the cold side – as with all heat engines you still need to get the cold side as cold as possible for energy generation.

  21. Harold:
    Considering the fact that all energy on earth is indirectly some form of solar energy, I wonder how inefficient solar cells actually are.

    Consider fossil fuel.. Prehistoric plants and algae use photosynthesis to get energy from the sun. They in turn are eaten by animals of ever higher standing on the food chain. A very small number of these are actually fossilized in a way that yields raw oil and gas which we can use.

    Compared to a consumer-level 14% efficient solar panel I’d say they are doing a quite marvelous job at converting solar energy to working power.

    And of course, thermoelectric plants using molten salt etc. are far, far more efficient than regular solar panels, and are capable of producing power through the night from all the heat stored in the melted salt.

  22. Well, this is not new, I have some thermo-electric coolers, which are these silicon and ceramic sheets, they are very good at cooling or generating heat, but backwards… Not that good, for instance, if a Peltier effect cooler is used to generate energy, from a 100W heater/cooler you will only get around 10W (at max. and if lucky)…
    In the energy generation field they are good not for dedicated generation but for energy recycling. They are being used in some cars, around the engine to get all the unused energy (heat) emitted by the engine. This can make the engine run a lot better because like this the exceeding heat can be absorbed by the TEC and converted to electricity. But they are no match for solar panels and less for wind turbines/generators. They can also be used to absorb the heat behind a solar panel too, for instance :) Solar panel life improved and energy generated :)

  23. Focused sunlight can destroy a Peltier module with excessive heat.
    Instead, how about this? Sandwich the Peltier between 2 squared, metal reservoirs. Fill cool side with water.
    Next, heat up a cup of mineral oil with the focused sunlight. Use a thermometer to monitor the temperature of the oil as it heats. When the mineral oil reaches the maximum working temperature of the Peltier module, remove it and pour the hot mineral oil into the hot-side reservoir. The Peltier module should generate electricity with absolutely no danger of over heating. If a greater amount of mineral oil is heated and the excess is stored in a thermos, the Peltier can continue working long after that initial heating by simply replenishing the hot-side reservoir with fresh hot oil from the thermos each time the hot-side reservoir cools.
    Haven’t seen this done yet, but it makes a lot of sense that it will work.

  24. A few years back I had a little thought.
    Thermoelectric generator using natural temperature difference above and below ground. Gold monolith 50m below ground level with a grid of peltiers on top, then with a matching gold monolith on top. Each monolith say 50m tall by 1m X 1m wide.
    Using the temperature difference to create a thermal electric generator to create electricity. Probably would be a bit expensive but would like to know if the theory would work.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.