Author Options:

Peltier thermoelectric heat pump, can it be used as Solar Generator?. Answered

Can a 40 mm square Peltier thermoelectric heat pump be used as a solar generator It operates on 13 Vdc @ 5 Amps. Operating at 13 Vdc, we measured the hot side at 190 Deg. F and the cold side at 55 Deg F. I have a 6 foot reflector dish and can add Reok led tracker. How many watts will it generate if it does?. Thanks for your help


The efficiency of a Peltier cell pile, for generation, is very lucky to exceed 5% unfortunately. Using it in a solar setup as you describe is a recipe for disaster for the cell. Any more than about 150 C and these things die.


we are designing a set-up for Peltier plates TEC1-04904. we will use a fresnel lens to heat up the hot surface and thus producing electricity. By adding a solar tracker, we hoped to maximize the heat needed for the plate to produce max power. Here is my question. Is this project can be sort of an alternate to solar panels?
Please email me at doncatulin@yahoo.com and help us. If you do so, we will surely put you in our acknowledgement page. Thanks in advance


Given that solar cell efficiency has been higher than 5% for over 30 years, and that the best current figures are close to 50%, it cannot.

If you were to use them as an alternative source of power, where will you put them and why?

I'd use them with low-grade heat, ie lower temperatures.

You have the possibility of high grade heat. You'd raise steam and power an engine for ~30% efficiency.

You MIGHT use standard metal piles for this, accepting the lower efficiency than the Peltier cell,(1..2% for a metal pile IIRC) but you have more heat.

I'm designing a project for an engineering degree, and I was wondering if you could tell me, could I use an array of thermo-electric modules to harvest solar energy, and relay it straight back into some fewer thermo-electric modules to generate a refrigeration effect? I hear that these modules are notoriously inefficient.

If I were to do it "solar" I'd look at ammonia absorption refrigeration instead, if efficiency was at all important.

Solar powered tec cooling by converting heat into electriciity ? forget it.

Excellent, thanks, absorption refrigeration was the proposed method in the first place, but my research unearthed TEC, so I thought I should investigate. I'll brush over it and forget it.

Thanks for the advice.

There are other methods. If I were you, I think I'd wonder if the Einstein-Szillard refrigerator would work.......

If it does, make sure I get an acknowledgement in your paper......


thanks for the info. i am planning a Solar Generator to produce 5kw daily, 24/7. Peltier cells are getting very cheap. 136.8W TEC Thermoelectric Cooler Peltier 12V is $6.50 landed. tegpower.com site is already selling Peltier cells rated at 25 watts with .6kw daily output. I've seen postings on forcefield of 3 watts output per cell with chimney mounted cells on forcefield, with temperature of 350 deg F. Efficient cells used by BMW is putting out 800 watts from exhaust waste heat at highway driving. i'll crunch the numbers and see wht it will take to put together a peltier generator to run 24/7 with 5kw daily output using off the shelf parts. I plan on building a large 20' dish solar collector with salt as the storage medium. There's a lot of discussion on asymmetric solar collector that is stationary and self tracking, simple and cheaper. What is the max potential on these cells with proper heat and chill water cooling?. A thermal water cooler can be easily build to draw some of the heat off the collector for cooling.

If we use your figures, say 136 W "cooling" and 3W TEGing, that's only TWO percent efficient. Looking at the TEGpower link, I don't think theyt are using single cells to get their 25W. How many kWh do you want ? 5 ? or 5kW for 10 hours , 50 kWh The BMW work uses new materials for TEGs that aren't commericially available yet. The higher the delta T the more efficient the process is.

I need to make 5kw daily to be energy independent. I have PV and Vawt with average 2kw daily output. I know that these thermal electric cells are low effeciency but the suit my energy best and the cost per watts will be less than $1/ watt for DIY A continous low watts generator will allow me to use existing equipment and triple capacity. A 250 watts input over 20 or 24 hrs will hook up to my 45 amp current charge controller. Say I can get 2.5 watts per peltier cell, 250 watts will cost $650, $300 for 3 mil 55" x 150' 98 per reflective Mylar, $500 for plumbing, wood frame, dc pump, salt, etc. Some salvage parts and insulation from the landfill. I'll be at $1500 for 5kw output per day. If I go with PV it will be $6000 for 1200 watts + $4000 in moot charger and inverter upgrade for 5kw daily output. Help me out. How is multiple cells assembled like the tegpower panels. Also, what storage volume do I need for 3 days of power from molten salt. I have biocomposite to make storage tank and tons of fiberglass insulation from salvage water heaters to make quadruble insulation. What will be an ideal working fluid to move heat to storage from the concentrators. I will have 150' of panels 55" wide, probably parabolic with tracking or stationary asysemtry. I will use a DC Teflon coated pump to circulate fluid. Will a truck heat exchanger be suitable, similar to the ORC setup. Please don't be negative because I'm in Puna Hawaii and there's tons of low tech setup that has been running for years

If you have a really dense source of solar power like that - a prime, high temperature source, I'm not sure I'd run TEGs. There are better methods, perhaps we can explore.

I still don't understand your output terms. 5kW is a rate of doing work - Joules per second. kWh or kilowatthours are a unit of energy use. If you only need 5kW for one hour, you have an energy requirement of 5kWh. Is that really all you need ? In a Hawaiian climate, I'd guess, on 10 hours of Daylight, that would translate to 500W of capacity in your power generation.

Tegs will really need to be assembled in very accurate flat panels for life and efficiency - I use precision Peltier cells very often,and we assemble them by lapping their heatsinks absolutely flat, before assembling them with silver compounds..

It is a bad idea to put too many into one assembly, because they have to be very securely clamped into their heatsinks. You'll have 200 to assemble.

Are you sure about using "molten salt" - Which molten salt ? Molten sodium chloride is 800 C. What material can you hold that in ?

You'll need 20 m2 of collecting surface to extract 500W by TEG - and the abiltty to reject nearly 16kW of waste heat.

You'd need about 4 to get 500 W from a solar electric panel. To meet your 5kWh requirement, you'd need 500W of solar panel, and 10 hours of light. The panels would cost about 2500 USD I think.

I'd be dead keen on semiconductor TEGs myself, if they were more robust, and more efficient. Metal junctions Tegs may be a cheaper method, and they WILL take a lot of heat. They are less efficient though !

You'd get a lot more efficiency with a stirling cycle engine powered by your storage bank.

Where can you get a stirling engine that will produce between 1 and 5kw?

I have been leaning towards Stirling for a long time. It's a bit too technical for my skill level but doable. A 3 kwe stirling was developed by infinia STC. Rinnai aslo is develping 1 kwe LFPSE module for Europe. Can a DIY build a 5-kWe Free-Piston Stirling Engine Convertor, similar to Foster Miller/ Auburn University. A 1-2 kwe FPSE will be sufficient for my needs. I will also need to find more economical and easily obtained materials to fabricate th FPSE, salvage yard has lots of different metals that can be melted nd fabricated in a backyard blacksmith shop. How much $ and what special equipment will I need to build one?. The FPSE can be low maintenance with lesser quality materials. Hot End Assembly The hot end assembly consists of the three heat exchangers: the heater head, regenerator, and cooler (see Figures 2 through 4). The heater head and cooler are both shell and tube heat exchangers, each with 1800 small diameter tubes (0.89 mm ID, 1.65 mm OD). The ends of each tube are electron-beam welded to the tube sheets. The heater shell, cooler shell, and tubes are constructed of Inconel 718. The regenerator is constructed of a porous metal matrix with 80% porosity (20% density) brazed to an inner and outer metal liner. The matrix is a sintered structure of randomly arranged metal fibers 0.022 mm in diameter. The matrix is type 316 stainless steel; the inner and outer liners are Inconel 718. Displacer Drive Assembly The displacer drive (Figure 5) pumps the helium through the heat exchangers. It consists of the displacer (dynamic component) and the post and flange (stationary component). The displacer consists of a central rod, a displacer dome and its support on the hot end, and the gas spring piston on the cold end. Two gas springs, one on each end, provide the spring stiffness needed to establish resonance. The forward spring is relatively weak (24% of stiffness) versus the aft spring (76% of stiffness). The rod provides the bearing support for the displacer drive. An internally pumped hydrostatic gas bearing is used in the SCA design for long life. All components in the displacer drive are alloy steel, with the exception of the displacer dome, which is Inconel 718 due to exposure of the dome to the hot expansion space. In the CTPC, the equivalent steel parts were fabricated from beryllium for its low density and high stiffness. Because of the additional losses in the gas springs suspending a heavy steel displacer, a penalty of approximately 3.5 points in efficiency is predicted. Since the bearings are not active during initial engine start-up, a compliant wear couple is implemented on the bearing and seal surfaces to tolerate transitory contact. The inside diameters of the cylinders are plated using Poly-Ond® and the outside diameters of the rod and pistons are coated with Xylan® 1054. Cold End Assembly The cold end assembly consists of the alternator assembly, pressure vessel, and joining ring (see Figure 6). The alternator assembly includes the linear alternator, power piston and cylinder, and the suspension system. A flexure support is used instead of a gas bearing to support the piston and plunger. The flexure provides radial and circumferential stiffness which resists the forces between the magnets and stator iron. The entire alternator assembly, including power piston and cylinder, will be supplied as a single assembled unit by Qdrive, Inc. (Troy, NY). The alternator is designed to deliver 5 kWe at a nominal stroke of 22 mm and 6 kWe at an overstroke of 24 mm, both at 85 Hz. To reduce lead time and expense, an off-the-shelf size was customized to obtain the power output, voltage level, and temperature capability of the design requirements. These modifications included slightly larger magnets, Hiperco laminations in the stator, samarium cobalt magnets, and custom windings. A performance penalty was accepted in lieu of an optimized custom alternator design. The current design is calculated to provide 88% efficiency at 5 kWe, versus approximately 93% efficiency with an optimized design The engine is fabricated out of Inconel 718 and alloy steel.

stirling1.jpgStirling FSPE.jpg

Hey Steve, No I don't need 5kw/h stirling. A scaled down version of this engine to 1kw/h is plenty power. You have plans for piston type stirling that can run a 1kwe generator head?.

Sorry, no I don't have plans, but I think you'll find plenty of free piston engine sketches at least on line. You've got my interest piqued ! I've a couple of good stirling engine books, one at least has an FP design to play with in it. If you can handle pressurised helium (and its a very sneaky gas by all accounts) pressurising is the way to go with it, and the FP methods mean you can seal all your bits in quite easily, and still pick off power outside.

Hey Steve, Had found a source for Thermoelectric generation module,Kelk Ltd Japan has 50mm x 50mm x 4.2mm with 24w output, 7.2 effeciency. $30k yen each @ 50 minimum order. Ouch!, that's US$15 / tec watt. The lesser effecient heat pump peltier cells is US$1.60/ tec watt or 0.24 cents/ watt in PV terms. Anyhow, I've orderd 10 heat pump peltier cells to try out. I'll try and do an instructable on this project. I've also found some stirling plans that I'm considering, but a scaled down FPSE with 1 kwe output is what I desire. I'll keep you and sean posted on my findings. Lastly, do I need a TEC driver for power generation or a simple charge controller?.

Is this a good unit to buy? I want to play around with this. The site does not say its efficiency or I just do not know how to read it. Can anyone help? 

RESULT ! Well found. You need half as many cells this way. Microgen BUILD free piston stirling generators, just as you need, but I don't see how to actually buy one. You should just have a nice DC output from the cells, assuming you can keep them at an acceptable temperature. I'd put some kind of mechanical system on to pull the things out of the sun if they get too hot. Some kind of charge controller is going to be needed though.

That looks awfully expensive "electron beam welding" doesn't come cheap for a start ! I'll look for some links to Sterling engines I have. There are some really good books on them, and I think they'll scale up - and won't take a very highly equipped workshop. I own a scientific instrument manufacturing company, we work routinely with tolerances below 0.0001" - and a 5kW sterling to that description, for a one off, scares the hell out of me :-) So ARE you looking for 5kW for 10 hours ? Steve

re: lapping...Have you tried using the thermal transfer pads that people like Vicor specify for use in connecting their DC/DC converters to heat sinks? (just curious what you think of them or the general category of transfer pads for TEG use). Or do they result in unacceptable losses?

Nothing beats metal to metal contact we've found. You can't afford to get a single hot spot under a cell or the whole thing goes to hell quite quickly. I've seen a new idea for "diamond" heat transfer compound, which we'll have a try with at some point. I build a test rig for lubricants that needs to reach -45 deg C - in a +40 Ambient - every little bit helps there ! The thing they never tell you about P cells is their efficiency drops like a stone below about 30 Deg C. hot side. You're VERY lucky to shift 10 W at -45 even with a "150 W" cell.

True enough. k is k. I was happy with the little pads though, since they squished down rather nicely and thermal goo always squished out, leaving me with doubts as to whether I was going have a void. Just was curious if you'd ever used them and if so, your experience. I'm impressed that you can get so low...staged? TEG/TEG/TEG...?

Yes, we had to stage them. I have a new system that I need to design high current drivers for that uses three stages of Russian Peltier cells to get there - Two BIG 150W cells, then a 150W cell, then a 60 W cell (because I had one) - thats 20 A per each at 24V, and one at 7 theoretically.

Sounds like fun. Serious fun. Jobs like that are hard to come by in my neck of the woods.

I just bought a used 12V 48W TEC and I would like to try making a TEG with my solar box cooker. The temperature inside the cooker gets up to around 250-300 F when the outdoor temp is 70 F. That would avoid the super hot temperatures of the parabolic reflector that could damage the TEC.

I would imbed the TEC in the rear wall of the cooker, so the external heat sink would always be in the shade. I would probably paint the heat exchanger in the cooker black, so it absorbs heat as much as possible.

Is there anything else I should know before trying this? What kind of load should I attach to the wires in order to measure the output? Would a 5W resistor do the trick?

I think you might get somewhere with this approach. Yes, a 5W resistor will be fine, and if you assume the usual grim efficiency of a TEG, something like 1/4W of electrical output at 12V means you can start with a 620 Ohm resistor as a load. I'd just use a 1K 0.5W potentiometer personally, set it as a rheostat, set the knob mid-range and try it. Measure the OC voltage, apply your load, and adjust it until the voltage across the rheostat is 1/2 the OC volts. Now the resistance of the 'stat is = to the internal resistance of your TEG AT THAT differential temperature. <br />

I’ve tried standard peltier modules with limited success and very short service life. They cannot hold up to the higher temperatures necessary for good power generation performance. I found a terrific supplier with both standard temperature and high temperature TEG modules designed specifically for power generation. They sell a lot of devices on eBay and you can also buy from them direct. The company is Thermal Enterprises and here is are links to a couple of their eBay items.



The parabolic dish is so simple. Such an eloquent way to make power.

This guy has some experiences with making parabolic dishes to use as power sources:



A parabolic dish power source can be compact too. A 24" dish is capable of making a few thousand watts of electricity on the fly.
The problem, of course, is converting that energy into electricity efficiently.

I love the idea of the solid state Peltier device converting power/sunlight from the parabolic dish. But most Peltier devices are inefficient at the moment, 5-15% efficient.

However, this company claims to have an answer to higher efficiencies, 70-80% of Carnot efficiency. Very useful to the parabolic dish setup.


Soon, these "solid state thermal electric generators" will be cheap and plentiful.

And Soon, efficiencies will approach 100% - Research "AeroGels"

I envision a 36" parabolic dish on the roof of my RV, tracking the sun as I camp. Providing 3500 Watts of power on the fly, and charging my 8 deep cycle batteries.

Here are more extremely interesting links:








That's a question I've been meaning to test ever since I first learned about Peltier devices, but honestly, I never did. Since a Peltier device is, at least in theory, a dense array of thermocouples, it *should be capable of generating electricity by creating a temperature differential across the hot and cold plates. If I can be so bold...lol, why don't you do an experiment and let us know how it goes...I know that at least one unnamed person would be interested in the results... To set up the experiment, you'll need a cold sink (water bath) to immerse the fins on the cold side, and a hot sink for the other side...that's a little trickier, but I suppose you could paint the heat sink with flat black paint or devise some other clever method of coupling the heat sink to the heat source (sun). Hmm...try it reversed as well, (hot side in cold sink, cold side in hot sink) just for shs and giggles to see what the difference in output is... Avoid over heating the device, since they have maximum operating temperatures...And presuming inverse logic on the cooler's operation, a large differential would likely be needed for the most efficient generation. best of luck!