20% Efficiency easy to manufacture Thermoelectric Material

[rhodomel]

http://www.technologyreview.com/news...lectric-power/

This could have wider range of application, such as fuel cells, ICE exhaust... by converting 20% of wasted heat to additional electricity from fuel.

[Ladogaboy]

This isn't really new. People have been experimenting with bismuth telluride for a while now, but the problem is that the efficiency gains still aren't worth the cost. A TEG system for a car costs, at this point, several thousand dollars, and the maximum efficiency increase that you can hope to gain is 10% (about the efficiency you gain from deleting the alternator). That will hardly offset the cost of implementation.

It does make more sense in an application like a hybrid or plug-in hybrid, but only on extended driving. Again, for the occasional road trip, one has to ask, would a several-thousand-dollar system be worth a 20% increase in fuel economy? I also doubt the cycling nature of the Volt's ICE would be the best choice to mate with such a system.

The price on these systems really needs to come down (if done by OEMs, that might be possible) in order to make them viable. Also, there are some other thermoelectric-thermomagnetic compounds/alloys that are more cost effective (though possibly slightly less efficient), such as a nickel-based compound that I've read about recently, but can't seem to find a reference for.

[DCFusor]

Yup, and BTW, check out the melting point of most peltier device materials that DO have "decent" efficiency - it's too low to even get power off a woodstove flue. (there are some insanely expensive ones that will stand high temps, but you can't buy them)

Let's say we have 10 of those devices they sell for beer coolers etc. They are about 70w input when used as a cooler (not counting the fan to keep the hot side cool so they pump heat at all). Each. Now, take ten of those, put the hot side right at the melt point (about 120c) and the cold side at freeze point (0c) and lets see - 2w apiece comes out. Or, 20 watts, from ten things 1.5" on a side, total, and of course, you need to keep that cold side cold and that hot side at or just below 120c (where they melt).

How do I know this? Tried it myself with some - I've been off the grid for a long time, and even a few watts when the sun doesn't shine is valuable to me. But the amount of work, the having to use an antifreeze-water mix to avoid melting them (the mix boils first to protect the device) and only having it work when it's well below 0c, and oh yeah, the fact that the heat exchanger pump draws more than 20 watts....interesting intellectual exercise. And at current *surplus* prices, those little suckers cost around $17 each plus shipping. I used them to make a fridge till I found out that the good old freon cycle is 3-5 times more efficient as a heat pump.

It's like when some jerk PhD comes into engineering and asks why they aren't done with this "intellectually trivial" exercise...It's amazing how many of them can't do the math on the details.

[Cord]

If people want to play with these devices for cooling an old desktop computer switching power supply works well with little ripple to power them.

You will need to short one wire on the plug that would have been in the motherboard to get it to start.

The devices are low voltage - high current
-----------------------------------------------------
I have one by my desk keeping the Dr Pepper cold.