LED+Peltier Cooling

Has anyone thought about cooling a LED with a Peltier Element? Is of course reducing the battery run time, since the power consumption of the peltier element is higher than the transported heat. But this could be used to overdrive LED's harder, with still good LED efficency.

Uh, not to sound pessimistic (who wouldn't want a supercooled, overdriven LS), but the battery pack would have to be absolutely insane (SLA types) to feed both the peltier AND the overdriven LS (and even at that, I'm not sure if it'll work), and then there's the issue of heat dissipation, which if I'm not mistaken, is agonized by the heat generated by both the LS AND the peltier...

I don't see it working anytime soon on flashlights as we know it now, not with what current peltiers work...

Some modders here have experimented with liquid-cooled LSs before (I think McGizmo?), and that's more viable of an option than peltiers...

The energy needed to drive the Peltier cell is roughly from five to ten times more than is needed to drive the LS. Plus, the Peltier cell needs heatsinking... You cannot power such a gizmo with AA batteries, nor you can think of making an handy one. At the current state of the technology, it is wiser to stack two or three underdriven LS...

Anthony

I've used these from time to time. They do generate a bunch of heat on their own, that is 'the hot side' gets plenty hot without pumping any heat at all. And we're already fighting a heat problem, right?

To get more than about 10 degrees C drop you need to put them in series. The photo diode cooler I last worked on used such a set up to get nearly 20 degree C drop but used forty elements backing 6 to do it. It dumps heat into cooling water. It draws about 18 Watts as I recall and 'pumps' about .1 Watt from the array.

The elements are electrically in series, but thermally in parallel.

Probably not at all practical for flashlights. Best to work on the traditional solutions, better heatsinking, more surface area, moving the air. A small fan would do far more good, IMO.

Doug Owen

I don't know about the one you were using...but there are many peltiers that have about a 70c temp. difference. I have one that I made a pop can cooler out of...only problem is it uses 80 watts.

[ QUOTE ]
zmoz said:
I don't know about the one you were using...but there are many peltiers that have about a 70c temp. difference. I have one that I made a pop can cooler out of...only problem is it uses 80 watts.

[/ QUOTE ]

Wow, a seventy degree delta. That make it what, fourty below inside? The back of my refrigerator gets very hot, but it's the beer inside I'm interested in....

Doug Owen

Yeah...I beleive the biggest one is 226 watts, and I assume that the temp difference is even higer. I got it to -36F with the hot side being cooled in ice water. With just a fan on the hot side it gets to about 20F, then there is a peice of aluminum on top of that the size of a pop can. That's where I keep the "beverages" while computing. /ubbthreads/images/graemlins/wink.gif

http://www.coolchips.gi/

Give them a few years.

I'm going to hunt for a Peltier and a CPU waterblock, and put the radiator in the freezer, and see how much juice I can run a 5w Cyan at. Might take a while to gather the parts, unfortunately.

You guys don't want to see the images my mind is coming up with imagining these "flashlights" /ubbthreads/images/graemlins/ooo.gif

(Back when I was in photorecon we had Peltier's in some of our "portable" processing gear and used this really neat cryogenic unit about the size of a stainless quart Thermos to cool some of our detectors (to just a few degrees above absolute zero). Of course the power draw was absolutely insane ... )

[ QUOTE ]
Tomas said:
You guys don't want to see the images my mind is coming up with imagining these "flashlights" /ubbthreads/images/graemlins/ooo.gif



[/ QUOTE ]

Hey, Liquid Nitrogen is cheap. First we get us a false bottom D cell, fill it up, put in a small pump.....

Doug Owen

Heh. I'd just use good ol' Prestone radiator fluid...

Well, you can always go get a 55-gallon dewar for storing/transporting lots of LN2, so you always have some on hand....

But, even more interesting is the possibility of using the peltier in the opposite direction /ubbthreads/images/graemlins/wink.gif Using the heat generated by overdriving the luxeon to generate power due to the temperature differential and using that power to run the luxeon. The hotter it got the brighter it would run!


"Lisa! In this house we obey the laws of thermodynamics!"
-Homer Simpson

/ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/yellowlaugh.gif /ubbthreads/images/graemlins/cool.gif

[ QUOTE ]
James S said:
But, even more interesting is the possibility of using the peltier in the opposite direction /ubbthreads/images/graemlins/wink.gif Using the heat generated by overdriving the luxeon to generate power due to the temperature differential and using that power to run the luxeon. The hotter it got the brighter it would run!


"Lisa! In this house we obey the laws of thermodynamics!"
-Homer Simpson

/ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/yellowlaugh.gif /ubbthreads/images/graemlins/cool.gif

[/ QUOTE ]

There is already something like that James, its known as a thermocouple.

Dan

Dilettante posted about something simular a while back... Thermoelectric Cooler/Generator /ubbthreads/images/graemlins/wink.gif

D.T.S. • Thin Film Thermoelectric Generator Systems GmbH

An interesting concept, but not enough power yet for LED use /ubbthreads/images/graemlins/frown.gif /ubbthreads/images/graemlins/icon23.gif

Not to mention you have to observe the laws of conservation of energy. I guess, if you captured the light coming out of the LED as well as the heat, and had a 100% converstion efficiency...but at the same time, you would not be able to observe the light being emitted, or use it for any practical purpose, or even measure it. So, on a different note, you could build a "black box" with an LED, thermocouple, and photocell inside and call it a perpetual motion machine, claiming that opening the box or measuring the LED's output will cause it to cease to function...well knowing, in fact, that it never stays lit at all. /ubbthreads/images/graemlins/grin.gif

On a more serious note, the thermocouple relies on a delta of heat to generate the electrical curent, much like a water turbine relies on a pressure differential...given that the higher operating temperature of the LED would cause it to get dimmer, you may be losing ground instead of gaining it. /ubbthreads/images/graemlins/icon23.gif Conversely, the LED's efficiency gets better with a decrease in temperature. I think it's either the blue or cyan Luxeon that has a sharp rise in output as ambient gets colder, whereas the other types have a more linear slope.

As another note, trying to use a peltier to generate heat from an LED will simply kill the LED. You need a fairly steep deltaT, and the efficiency sucks.

There IS a commercial product that I found interesting. High-temperature thermocouples used to line the combustion chamber of a boiler, and they generate power to run water/fuel pumps and the control circuitry. Again, another Popular Science/Mechanics tech article.

You can actually use a peltier to measure the output of light from a laser at least. The peltier then should have a black surface to absorb all light the power output then can be calculated via the output current of the peltier. This is how many laser power meters are run. I am not sure how effective it would be for measuring lux output, but it would be a fun experiment. I have been playing with a peltier I got out of a coleman cooler abandon on the side of the road. It has been fun, with no load @ 12vdc it pulls 4.5-5 amps so ~60 watts. You can actually stack peltiers to get a lower temps. Well a bit off topic but I hope that somebody finds it interesting.

Peace,
Adam

I bought a little 1.5" square peltier that runs off 10-15 volts...lots of fun to play with. /ubbthreads/images/graemlins/grin.gif

I fitted the hot side with a huge CPU heatsink/fan and the cold side with a smaller one, then installed it in an insulated cardboard box. I was able to chill three Coke cans in about 15 minutes! /ubbthreads/images/graemlins/eek.gif Not quite what I would call refrigerator-cold, but cooler than room temperature. /ubbthreads/images/graemlins/cool.gif

I have a peltier that runs off 12v.
By mounting it on a Pentium Pro heatsink and connecting it to a 6v power supply, the cold side gets covered with frost in a few minutes.