Homemade Thermopile

You need more than a temperature differential-- you have to be able to force heat to flow through the thermopile, at an acceptable cost per unit of heat flow.

Even with exotic materials, efficiencies in the 3-8% range are typical.

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

I've been thinking about quietly buying up old smoke detectors for the Americium content, to build a basement reactor. A 50 gallon drum glowing at 300F saves heating bills. Seriously, a "so-called" thermopile needs an energy source. Anything useful needs Watts, a few hundred at least. Ken

Useless. Efficiency is what, parts per million with common materials? It's not even worth counting percent anymore.

Probably, a printed form would be best. Or given optimal materials (IIRC, NASA's RTGs use Si-Ge), lithography might be more appropriate. You want to integrate as many junctions as possible into a very small space. What you gain in thermal characteristics (in terms of heat loss) you lose in electrical resistivity, so you might as well minimize the material between junctions.

Probably, like, silicon and germanium. Oops...

Ya- junctions act like perfect voltage sources, so you're just talking resistor stuff here. The figure of merit of any two materials depends on the junction potential and their combined bulk resistivity. What length and thickness you want between junctions is up to you, but you might as well make it small. That means you need a steep thermal gradient, which means lots of power input, but also lots of current you can draw (depending on just how many junctions you've stacked). So you see, it makes up for itself.

Tim

-- Deep Fryer: A very philosophical monk. Website @

Hi. I'm building this:

It could easily run a generator. If you have maching capabilities, let's work together. Larry

An iron-copper thermocouple can be made from hardware-store stuff. It will give about 50 microvolts per degree C temp differential, and you can put multiple junctions in series to get more voltage. Of course, the resistance goes up as the heat source:sink distance goes up, and as you add thermocouples in series. Done just right, expect thermal efficiency on the ballpark of 0.5%. It's unlikely that you can harvest even a couple of milliwatts from the sorts of sources you mention. A

10 degree C temp differential is about 500 uV per couple, so you'd need ballpark 3,000 thermocouples in series to light an LED.

Older water heaters and furnaces use a thermocouple, in the pilot flame, to hold in the gas solenoid valve. I guess they use multiple junctions, too. 50 millivolts is typical here.

The iron will rust, too.

John

If you don't want to work together, check out:

Larry

The design parameter you should be considering are those that are available in the third world; that is the only place where the labor is cheap enough to do waht you are thinking of doing.

You need to pick an environment, such as in the sun on one side, in a river on the other side.

You will have to make a list of available materials and then determine the Seebeck voltage for each pairing. Start with materials that have large differences between Seebeck coefficients. Here are the Seebeck coefficients of some common materials in millivolts per degrees C at at 0 degrees C:

Aluminum 3.5 Iron 19.0 Lead 4.0 Carbon 3.0 Nickel -15.0 Copper 6.5

Source:

This is just a starting point. You need to find the figures for your target operating temperature, and balance cost against available voltage. I suggest doing your own experiments.

--
Guy Macon

Microvolts, actually.

John

snip

There were some made during WW2 for the SOE and resistance groups, These put out a low voltage and a half-decent current to charge batteries.

I have a reference to one that used 350 thermocouples (chromium-constantan) in a brazier heated by a good-sized charcoal fire. My guess is the efficiency was awful, but it beat having no power.

I wonder how their efficiency and voltage output compare to a modern Peltier cooler operated as a generator.

--
Regards,

John popelish

Didn't Voyager use thermopiles and plutonium? They still seem to be working

martin

What's three orders of magnitude between friends?

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Yeah, I shouldn't be so didactic, academic, advisory, donnish, edifying, enlightening, exhortative, expository, homiletic, hortative, instructive, moral, moralizing, pedagogic, pedantic, preachy, preceptive, schoolmasterish, sermonic, sermonizing, teacherish, teacherly, teachy, dainty, eclectic, exacting, fastidious, finical, finicky, fussy, nice, overparticular, particular, persnickety, picky, picky-picky, select, selective, or precise.

Sorry.

John

Why not attack the problem by starting with the available heat source and then finding the technologies that best suit the delta T and heat flow. Also, consider what your system requirements (electrical load, weight and size cinstraints, reliability) are.

Thermopiles fit a very small set of potential applications.

--
Paul Hovnanian	paul@hovnanian.com
-----------------------------------------------------------------------
Procrastinators: The leaders for tomorrow.

A friend of mine ran a primary school team that entered a model boat contest with a 60W Peltier device that had ice one side, boiling water or candle heat (I forget) the other, and drove an electric motor. They won the contest hands-down, ten seconds to cover 10m. Might not have been efficient, but it generated good power for a short period.

On a sunny day (Mon, 03 Dec 2007 11:26:55 +1100) it happened Clifford Heath wrote in :

OK I have a 'Zibro' air dehumifier. It has a peltier in it, and runs one some separate swicthmode that outputs

13.6V 5A DC. Been running for days.. I disconnected the DC plug from the switchmode, measured the open voltage on t connector from the Peltier : 1.2V measured the short circuit current (amp meter ) .45A

Not bad!!!!!!!!!! Because it is handwarm on the hot side, and about 1 C or so at the cold side, not even freezing, just cold enough to create condensation.

Indeed I think that if yo uuse ice and a falme it can run a small electric motor. Now how's that for quick test :-)

THANKS FOR ALL YOUR IINPUT

To clarify the concept, what I am trying to do is see if there is a low tech, relatively cheap method of utilizing the UNLIMITED FREEenergy represented by the low-grade temperature differentials between natural occurring sources like ambient air and ground water.

I know that thermopiles are not very efficient. But since the energy is free and I am not looking to manufacture and sell anything (hence the "homemade") the labor is free, the only cost consideration is the material costs. If the material costs can be made low enough, a viable alternative may be achievable even with very low efficiencies.

In the example shown below, alternating layers (of equal thickness) Iron Foil (folded at one end), Paper (insulation), and Aluminum Foil are utilized. As many layers as needed could be added. Insulated bolts would compress the layers to insure the junction contacts. The whole assembly could be dipped or coated in plastic to seal out air and moisture.

XXX XXX WWW W = Copper wire XBX XBX W XXXXXBXXXXXXXBXXXXXXWX X = Thin plastic coating to prevent oxidation XIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIWX I = Iron foil folded at one end H XIIPPPPPP PPPPPPPPPPX` P = Paper J XAAAAAAAAAAAAAAAAAX C A= Aluminum foil XPPPPP PPPPPPPPPPPIIIX J HJ = Hot Junctions XIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIX H XIIPPPPPP PPPPPPPPPPX CJ = Cold Junctions J XAAAAAAAA AAAAAAAAAX C XPPPPPPPP PPPPPPPWIIX J XXXXXBXXXXXXXBXXXXXWXIX B = Compression Bolts XB X X BX W XXX XXX WWW

In a given system, the voltage differential is strictly a function of the number of junctions and the temperature differential. What I want to know is how to calculate the watts.

If I hold the temperature differential constant:

Is the wattage produce the same or is it dependent on the number of junctions? With the same number of junctions, is the wattage the same if I double the thickness of the layers or make the layers 3 inches by 30 inches instead of 3 inches by 3 inches?

For those of you that are interested in early thermo-electric devices

At somewhat lower efficiency (past the halfway point IIRC) by now. Partly due to the decay of the Pu238 (note it's not the more infamous Pu239 of weapons fame, nor Pu244 or so of reactor grade fame), which is quite rapid.

The largest RTG NASA has sent up was on Cassini, IIRC. Remember people were making a big stink about the 80-some odd pounds of plutonium on launch?

Tim

-- Deep Fryer: A very philosophical monk. Website @

motor.

I believe you can buy Peltier powered fans that work in stovepipes, so the ice may not be necessary. ;-)

Best regards, Spehro Pefhany

--
"it\'s the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com