A thermoelectric solar cell.

[Ruud]

The idea was to attach a peltier element to a solar box cooker and use the temperature difference to generate electricity. The heat storage is supposed to elegantly deal with intermittent sunlight, as opposed to PV cells, that immediately drop to near zero juice as soon as a cloud moves before the sun.

This is the design. The sunlight heats the heat storage. Hot air is transported out of the 'greenhouse' by the heat sink through the peltier into the waterreservoir. The seebeck effect will cause the peltier to output electricity.

Here's the inner box, created from cardboard, glue and aluminum foil.

These are the inner and outer box, insulated with pieces of newspaper. The heatsink I took from an old PC, is glued in place.

These pebbles will be the heat storage.

The pebbles are painted black, to get more heat from the sunlight.

Pebbles glued into the SBC.

Added a thermometer to the SBC. It takes away some sunlight from the heat sink, but that's okay. The heatsink is not supposed to directly generate heat anyway.

Two of the reflectors.

All four of the reflectors, covered with aluminum foil.

The picture frame I sacrificed to get the plexiglass windows.

One of the plexiglass windows.

First window glued in place.

Both windows in place.

Glued the reflectors to the box.

The piece of MDF is going to be the bottom plate, the little pieces of wood will be an adapter to make the solar cell tripod mountable.

Glued the tripod adapter to the bottom plate.

Bottom plate mounted on tripod, for easy 'manual solar tracking'.

About to glue the SBC to the bottom plate.

Some cleaning cloths, for insulation between the heat sink, the peltier and the cooler.

a few layers of insulation...

...then the peltier...

...and some more layers of cloth. An empty spam can is glued to the peltier and we're done!

Okay, now let's get this thing in action and rid the world of fossil fuels!

Unfortunately, I live in the Netherlands, the weather usually pretty much sucks here. Also, I'm at about 52 degrees North and it's winter here, so the sun isn't all that powerful right now. Still we saw some nice weather on the 23rd and 24th of februari so I was able to test the thing.

Here's me, testing the solar cell at the office. The office has double glazing, so that may have cost me 10% of the sunlight. Still the thing got pretty hot...

Unfortunately it wasn't a big success. The SBC worked fine and got to 70 degrees celcius (158F) on thursday and to 92 degrees celcius on friday (about 198F). The electricity output however was not very useful. We got to 0.082 volt at 0.023 ampere. Even when I cheated, by filling the cooler with ice cubes and cold water I only got 0.17 volt at 0.047 ampere.

Here's the solar cell at 70 degrees celcius.

(This is where it passes the 0.1 volt, after I cheated by filling the cooler with ice cubes.)

I'm not sure what went wrong. I had tested the peltier with two cups of hot and cold water and it performed al least ten times better then. Maybe the heat sink isn't that effective, or maybe the glue on the peltier is a bad heat conductor. I don't know. I might build an improved version some day, using some special heat conducting glue.

Anyway, here's some more tips, for people who want to build a thermoelectric solar cell like this one.

• My reflectors suck. A test with a laser showed that about half the sunlight bounces to the opposite reflector and back to the sun... Calculate the angles instead of guessing them like I did.
  
  
• My heat sink was bigger than the peltier. That may have cost me a heat. Next time I'll make sure the heat sink and the peltier are the same size.
  
  
• It's pretty stupid to have the heatsink at the back of the SBC, instead of in the ceiling, as the hot air is supposed to go up. (see improved design below).
  
  
• It's also pretty stupid to have the heat storage horizontal, instead of at a straight angle with the sunlight. (see improved design below)
  
  

Oh by the way, the intermittent sunlight thingy worked out pretty good. When closing the blinds it takes at least a few minutes before the voltage significantly decreases. Off course one could get the same result from a PV cell with a battery, but the peltier/pebbles is a way more low tech method, so who knows, they might be cheaper to manufacture.

Also, I apologize if this is thermoelectric solar cell story number 102394330. I haven't had time to read all of the postings on Otherpower yet.

[Ruud]

Woops.

Forgot to post the last image :

[Kwazai]

just a thought, but if its anything like the thermodynamic stuff (stirling cycle etc.) and convective heat exchangers then either the temperature difference is going to need to be real large or the surface area of the peltier device is going to have to be really large.

Mike

[Gary D]

Interesting project! You have handicapped youself by doing this inside(double paned windows). You should/ might be able to get a marked increase if you would take it outside and repeat the tests over a lunchbreak? I think the windows cut your potential heat gain by at least 20%.  Gary D.

[Stynus]

i think the plexiglass holds back the uv and ir light for a big part. and this light is needded for heating.

verder wel een vet project ;-)

[dinges]

Heheheh.

Before you know it, the Dutchies will take over this board

Nice post. Though I remember Peltier elements being horribly efficient as heat pumps. Driving them reverse won't make matters much better. Then again, there's only one way to really advance science; by proving others wrong.

Looking forward to any updates.

Peter (absoluut GEEN Nederlander

[dinges]

Oh nee! Nog eentje! Administrator: act now before it's too late!

Actually, plexiglass should allow more light to pass through. IIRC, plexiglass allows most UV to pass through (whereas glass doesn't). IR won't get passed much by either, I think. Not really sure about that last one though.

If you tried heat-conductive paste instead of glue (any electronics shop will be able to supply this), your thermal resistance between Peltier element and heatsink (in this case: heatsource) might get an order of magnitude less.

I wonder what happens when you replace that Peltier element with a small PV panel. I think output would increase pretty much (though it might heat up as well; which is not good). Plus you can experiment with a small batch solar water heater (replace heatsink with a small black container of water).

Peter.

[Stynus]

i just had a nice idea make 4 or a bit more heatpipes and a few collectors and make them al on 1 peltier. so you got more heat on to the peltier.

an other idea is to use a lens and hang the peltier in the focuspoint. or a sanded sattelite dish in the firepoint.

en peter ik ben belg genen hollander

[dinges]

Belg... Waarom krijg ik nu plots een gevoel van medelijden 'k zenne kiek geen Hollander maar Nederlander. Of anders desnoods Brabander...

Where do you live? Live myself 15 km NE of Antwerp. Used to work in Turnhout for a couple of years (living in Vlimmeren/Beerse). Definitely want to go back there.

As far as Peltier elements go, I think there are better routes to generate electricity than these. They're notoriously inefficient, even more so when driven reverse (thermal energy input, expecting electrical energy output). Your ideas might increase output, but I think with less effort it would be better to use the energy to generate heat for water or the house, not electrical energy. For the last one, you also get all kinds of storage issues.

Myself, I'm slowly starting to work on solar batch heater. I think in these simple devices are the real energy saving gains to be had (saving in gas-bill). But electrical energy is nice as well of course. Still have a small (20W) PV panel that I have to install in the near future. Nice hearing from others in the area!

Peter.

[DaveW]

I did this experiment 10 or 12 years ago. I have some notes that might help. First, using a PED instead of a Seebeck device is an uphill battle. A PED is optimized to move heat. What you have built is a good hotbox, could be used to capture low level heat and move it indoors. What you need is a focal point box, to focus the heat onto a small area and place your input heat sink there.



  Something like this only not as battered. Or better yet, a spun mylar reflector from an old traffic signal head. The heat sink should be small and copper. Copper transfers heat about 60% better than aluminum. Center the heat sink in the focal area. To insulate the PED and provide support, something like micarta or Dupont vespel SP-1 should be used. Either will give good insulation and mechanical support. I use sil-gap 2000 (from Bergquist) instead of thermal compound, it provides 5 times the heat transfer. Don't forget to use nylon 66 fiber filled or vespel screws to capture the PED between the heat sinks, regular nylon screws will powder out in just a few days from the heat. A precise fit from heat sink to insulation to heat sink is critical.





  I machined my insulation to hold the PED centered and to feed the wires out. Use belleville washers to hold the proper tension, with the device heating and cooling and expanding and contracting it is easy to break something without them. For the back heat sink, copper with plenty of rib area is nice. Be sure to allow for plenty of air flow and have at least 5 times the back heat sink area as the inner one. A small fan will probably be required to maintain the temperature differential needed to make it all work. Carefully done, you can see several hundred degrees differential across the device, and I was able to get about 6 watts from a 40 watt PED. Of course this thing went on to become a cooler for a CCD imager mounted 90 feet in the air, it would have been far to expensive to make to really generate any usable voltage from it.

[Ruud]

Ha ha, dat internet begint steeds meer op een vakantie in Spanje te lijken, je komt er overal Nederlanders tegen.

* Switching from Klingon back to English *

@Mike: You're right about needing a way larger surface area of peltier. To power a household you'd need a large panel with several hundreds of these cells. Of course the cells would be quite a bit smaller than this prototype. (and they'd have to work! )

@Gary: if only my lunch breaks were that long! It needs about an hour to heat up to 90 degrees C, but you're right, I'll do the test outside some day.

@Peter: I agree that peltiers are by far not as efficient as PV cells when it comes to 'square metres of sunlit surface/Watts', but I was thinking more in 'dollars/watts'. Like maybe, now that peltiers are being mass produced for cooling CPU's and stuff, they'll become cheap enough to build solar panels that cost less than the PV's.

@DavdW: Thanks for the excellent tips on the heat sinks and insulation. About the hotbox, one of the reasons I went for an SBC instead of a focal point thingy, is that it wouldn't necessarily require tracking to work. I could just point the thing south and it would work 'sort of okay' throughout the day.

Also, I'm having a hard time trying to figure out what a PED is. Where do I get those?

[Stynus]

the peltiers themself are not inefficient but the insolation arround it is because it is verry thin.

and to use it in "reverse" you need other materials inside the peltier i think.

like in a thermocouple.

en ik woon in Lommel (noord limburg)

[DaveW]

Sorry, Did not mean to be speaking Ferengi when the official language here is Klingon. PED or Peltier Effect Device rather than SD or Seebeck Device. A PED is made up of mostly BiTe, or bismuth telluride, while a SD is mostly PbTe or SiGe - lead telluride or silicon germanium. Melcor in New Jersey makes and carries a good line. The difference can be a lot, I was able to get almost 38 watts from a 40 watt SD, against the 6 or so I got from the PED. And tracking is a given, without tracking the output falls off very quickly. As I said, your money would be better spent on making a low level hot box, moving heat from the sun indoors in the winter. The best use I have found for large PEDs is as dump loads, with a simple relay reversal they can either heat or cool the battery house, but you need really large heat sinks to make it work - 8 foot X 1 foot X 6 inch heat sinks for a staged dump made up of 4 270 watt PEDs.  

[Kwazai]

I had seen the copper plate/salt based diy solar cells- is there a diy peltier or seebeck device, possibly something that could be produced really large ,really cheap?

Mike

[Ruud]

Thanks Dave. The element I used is a 17 watt BI2TE3 (I suppose that's bismuth telluride?) thingy. So I guess that makes it a PED. Anyway, the element worked pretty well before I glued it into the hotbox. Using a cup of hot water it yielded enough energy to run a small 'solar motor' that requires at least half a volt at 100ma.

I think that the problem is in the way I used the element, rather than the element itself.

[Ruud]

@Mike: Sure, you can make a Seebeck device out of any two metals. Here's one I made from copper and iron wire. It's not all that efficient though. It uses enough candles to power an airplane but only got to 0.045 volts...

[dinges]

Lommel... no, that would be too easy, like stealing candy from a child

Perhaps it depends a bit on your definition of (in)efficiency. I haven't got exact numbers, but I seem to remember that efficiency of Peltiers was the same or less than those of PV panels (10-15%) which is not terribly efficient.

If Ruud were to make a calculation of how much energy he 'caught' (frontal opening of his collector), assuming an insolation of, say 1000 W/m^2 (though, in the current weather & Ruud's setup (inside, behind glass), it may be closer to, say, 600W/m^2? Just guessing here), and how much he gets out... I think it would not be measured in procent but promille.

That same amount of energy could probably almost 100% efficient be used for heating water.

Peter.

[Kwazai]

when I first read about them I was wondering if you could take something like steel sheet buried in the ground (crawl space size) and aluminum sheet(film) in the attic and get any power out of it. Is their some thermocouple type math buried somewhere on the net?

Mike

[Chagrin]

I would've made the inside of the "hotbox" black instead of using the foil. Seems like you'd want to avoid reflecting light back out of it. Maybe add a few pieces of foam around the sides of it to keep the heat in the box.

Also, mylar is almost twice as reflective as aluminum. You can find it cheaply as gift wrapping paper or maybe buy a bag of potato chips. Eat the chips, turn the bag inside out

[wgrube]

Hi, I liked your idea, I'm currently working on something like this. Let me give you a few tips to improve your idea. First, it is better to use a parabolic mirror instead of the card board solar oven (I like these, but they aren't good enough for this purpose), there will be no glass (or plexyglass) blocking the rays so you get more heat concentrated on the peltier element. Ok, this will work only while you have sunlight, but it is more efficient. Second, thermal transfer paste! Any glue will block heat transfer. Even toothpaste can be used, but it is not so good. Third, aluminium is a bad heat conductor, copper or even silver is better. Build your heatsinks with copper and you'll have better heat transfer.

There are some ideas to build a parabolic mirror, like in:

http://www.greenpowerscience.com/PARABOLICMIRROR.html

You can also use a small parabolic dish from a satellite tv installation and coat it with a reflecting material.

To cool the other side of the peltier element with water you can use small copper tubes soldered together to make a small heat exchanger that will be connected to a water reservoir kept in a cool shaded place.

I found a video on the web, where a peltier element was attached to two aluminium heatsinks (one black painted for heat collecting) and the results where similar to your experiment by just putting this basic assembly into plain sunlight, 80mV or so without concentrating the sun rays.

http://www.google.com.br/url?sa=t&source=web&oi=video_result&ct=res&cd=2&url=htt
p%3A%2F%2Fwww.metacafe.com%2Fwatch%2F1952443%2Fpeltier_solar_generator_1%2F&ei=H_A6SsbZNsmEtwev1
enbDA&usg=AFQjCNHwGUOP57ttPuazL6fTmpIErEz2rQ&sig2=aeKyZI5qDTrgu96uBD8uAQ

I hope these few tips will help you.