electric cooling?

I was thinking about the Igloo plug into you car coolers. I think they use die-electric cooling? It`s where you have two differant metals and something inbetween...when you apply a small current one side gets hot and the other cold. Could a small (micro) version be made and used to cool a LS emiter...soldered to the cold side? I heard these use comparitivly little powerer? Maybe then run the LS at a higher millamp level...say...400-450 ma with bright results...any thoughts?

First, these Peltier devices use relatively high power. Second, they don't move much heat. Third, if you can get the heat out of the hot side of a Peltier cooler, why can't you get the heat out of an LS?

Seems like the best bet with these is to provide a good thermal path to a good heatsink (like an aluminum flashlight case, for instance).

I rather imagine that the Peltier will suck up more power than a couple of LSs, and be fairly useless for a portable flashlight device, because as an earier poster alluded to, you still have to get rid of the heat if it is an enclosed device.

If you really want to cool a LS fairly efficiently, you might want to check this thread out: http://www.candlepowerforums.com/cgi-bin/ultimatebb.cgi?ubb=get_topic&f=3&t=000801

It was started out as a joke, but it may have some application in real life, and is inexpensive.

According to the data sheets, freezing the LED to dry ice temperatures would significantly increase the light output (20% more at -20 degrees C). Talk about an extreme torch, battery AND dry ice powered.

But seriously heat is a major problem with the Luxeons and that's most probably why Lumileds have posted documentation on designs to keep the Luxeon cool. If I was to design an "ultimate" Luxeon torch, I'd take the Star/O, slap on heatsink compound and use a deep finned heatsink.

Then to ensure maximum heat is dissipated, I'd expose it to the atmosphere. A torch with hole behind the lens to expose the heatsink fins should do it. This way you can drive the Luxeon reasonably hard (near 350mA) without worrying too much about it overheating and dying.

Peltiers.... no. Not for (small) battery operated devices. It'll draw a LOT more power than the Luxeon itself and in general they'll also require a fan to assist in heat dissipation.

Hmmmm.

I remember (dust off the cobwebs) in my old solid-state class they discussed the this phenomenon. Not only can you use electricity to move heat, you can use the same peltier junction to turn heat into electricity! I don't know how efficient this is (probably not very), but this IS (I think) the technique that they use in nuclear-powered sattelites. Get enough fission going to generate heat, and pack it in a blanket of peltier junctions to create electricity.

All you need is one side of the wafer to be hot, and the other side to be cold. Then you *SHOULD* have electricity being generated.

Then, you can put this energy back into the batteries for a little "self-recharging."

Anybody want to try?

Try this for info and links to heatsinks www.peltier-info.com

Dans and Harrkev`s idea might not be so insane after all...check this site and click on the 2.5 watt at 3.3 volt generator. www.hi-z.com they talk about lost heat recovery...converting the waste heat back into usable power...so batts run longer...Star Treck stuff...beam me up Scotty...I need too replace my 123 batt...after running this Surfire for 2 years it`s dead...you would think they would have a better batt.

Those converters require a pretty hot
surface to generate useful energy.
Also, the other side has to be kept
relatively cool in order to get the
temperature differential across the
junction.

Also, using a peltier device to cool
something still requires a heat sink.
It would eat up lots of power too.
On the other hand, if
you wanted to set up a fairly elaborate
heat sink, you could in fact cool the
LS down to lower temperatures then
possible with normal ambient air and just
a heat sink.

I would be pretty sure
that two LS's driven at less current would
accomplish more output at less power then
trying to cool one single LS with any
technique. That would be cheaper too.
I would also be pretty sure that the same
two LS's would put out more light at less
power then trying to convert the heat back
to electical energy and recharge the batteries.

--Al

Hmm...well what I was thinking is...to get 2.5 watts at 3.3 volts they want a temp diff of 365 degrees...now...if we cut the temp diff in half...182...I don`t know...would we get 1.25 watts at 1.65 vlts...I`m sure it`s not that easy of a convertion but would be interesting...BTW...would a Luxeon light with 2.5 watts/3.3 volts...or not even light?

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by MrAl:
if you wanted to set up a fairly elaborate heat sink, you could in fact cool the LS down to lower temperatures then possible with normal ambient air and just a heat sink.<HR></BLOCKQUOTE>

That's true. There are three ways to get rid of heat: conduction, convection, and radiation.

Most conventional heat sinks concentrate on just the first two. But a good blackbody radiator could cool the thing down even further.

Recently, I noticed a white powder forming on the 50-meter blackbody radiator I had attached to my LS/O. At first I thought it was anthrax, but was relieved to find it was only frost

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Ken B:
Hmm...well what I was thinking is...to get 2.5 watts at 3.3 volts they want a temp diff of 365 degrees...now...if we cut the temp diff in half...182...I don`t know...would we get 1.25 watts at 1.65 vlts...I`m sure it`s not that easy of a convertion but would be interesting...BTW...would a Luxeon light with 2.5 watts/3.3 volts...or not even light?<HR></BLOCKQUOTE>

That company recommends that you contact them
for other temperature data, which probably
means it's not a linear function and output
decreases quickly with decreasing temperature, but i wont say too much untill
someone contacts them and possibly tests one
of the devices at a lower temperature diff.

Any idea what the price of the 2.5 watt model is?

Not sure what you mean about the 2.5watt
3.3 volt luxeon, but an LS at 3.3volts and
350ma consumes only about 1.2watts.


Another interesting (?) aspect about using
a peltier to cool an LS is that if the
temperature of the LS is brought down to
very low temperatures, then that means the
heat has to go somewhere else meaning the
other side of the peltier junction is
going to get pretty hot and even a heat sink
might get too hot to hold in the hand
comfortably. If the heat is conducted to
the metal flashlight body, i would bet it
gets pretty hot.

Al

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Daniel Ramsey:
I have been aware of a procedure called "cryo stress relieving" it involves immersion in liquid nitrogen for several hours. It is used to realign the moleculer structure in key high stress parts such as high performance connecting rods, rifle barrels, crankshafts. It settles the grain in machined parts and often induces longer life in ordinary nylon and the such. If we were to subject a Luxeon to this process could we enchance its output? As a mechanic and a gunsmith I know of this process and it has some merit. Any thoughts on this? Or am I getting too far off of the subject of "keeping the cool"?<HR></BLOCKQUOTE>

As far as I understand it, cryo treating is mainly usefull in a specific material range, that is, high carbon steels, mostly stainless, thus, its application in the cutlery and gunsmithing fields. I don't know that it'll do much to the materials in LEDs. In addition, cryo treating contributes to wear resistance and strength, but the problem we're looking at is one of heat dissipation.

I don't know if cryo treating would do anything to Luxeons either way, but i suspect it's not what we're looking for.

Joe S.

I have a feeling that cryo treating would be VERY detrimental to a LED of any kind, but especially a LS.

I used to have some of my hand-made knives cryo-treated, and it it my understanding that it is essentially for items that are of a homogeneous material.

The different co-efficients of expansion between the PC board, wire traces, semi-conductor materials, the "plastic" that encapsulates the semiconductor etc. would probably at least break the junction connections, if not completely tear the LED apart.

(I'm talking about liquid nitrogen cryo-treating causing the destruction, not the older, less efficient, less cold "dry ice"/alcohol cryo-baths that some knife-makers used to experiment with.)

Mr Al, do you think you (or some other inteligent member) could e-mail that company and explain to them what the goal is and ask them if the 2.5 watt one could be used to recover some of the lost power? Do they make one smaller? Could one be made thats round to fit a conventional flashlight body better? And maybe how much it is? I saw they had a click there...something like..."tell us what you want to do"

I`m sure they would need some info about the Ls heat charateristics...that might be found in the lumiled site. I would do it myself but honestly, I don`t think I would be the best choice for this "research and discovery mission". I`m way behind most of our members when it comes to understanding stuff like this.

I`m still thinking that...well, the heat has to be dealt with any way...if the cost isn`t to great and it`s fesable...why not put the heat to work...back into the system...might help the batt run much longer. Do step-ups get hot...could they be attached to the top plate as well...and produce even more usefull (convertable) heat?

Gransee, if you`ve already gone down this road and know it`s rediculus...please jump in.

sounds like time to COMBINE
how about the flashlight coffee warmer

remember a small finned heat sync can be cut trimed and hacked into about any shape regardless of the shape it was in.
If the synk that is there will maintain the led in air, a simple scr sync from radioshack would doubble that quick.

I think its funny how you are looking to sync at 350ma, i was uhh stupidly running these at 500-750 without any problem (dont do that at home), using air as a heat sync, well untill the carpet melted

the white lexeons are pretty rugged with thier OWN sync, that is my opinion.

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by Ken B:
I`m still thinking that...well, the heat has to be dealt with any way...if the cost isn`t to great and it`s fesable...why not put the heat to work...back into the system...might help the batt run much longer. Do step-ups get hot...could they be attached to the top plate as well...and produce even more usefull (convertable) heat?

Gransee, if you`ve already gone down this road and know it`s rediculus...please jump in.<HR></BLOCKQUOTE>

I just started reading this thread. Looks like a fun topic.

Yes, the issue of heating has been a design issue for our flashlights, especially the Arc-LS.

Like Al said on the first page, the peltier junction (or any other type of active cooling for that matter) would draw more power than what is gained in brightness by using it.

Peltier junctions are not terribly efficient. They are usually used because of their reliability (no moving parts) and simplicity (no coils, compressors, etc).

True, a thermocouple can be used to generate electricity by having heat "flow" through it. But the efficiency is quite poor. The heat "flow" required is high compared to the small amount of power produced.

As you add more active components that draw power, you increase the total amount of heat produced by your flashlight. That heat still must be conveyed to the surrounding air. It cannot be consumed by some system in the flashlight.

Even a system where a material (gas, solid, etc) is stored in a canister and absorbes heat as it changes state (boils off like liquid nitrogen, etc), has it's heat removed by another system.

You could build a flashlight that was powered by a CO2 canister. The expanding CO2 would absorb the heat of the LED while the resulting expansion would spin a turbine that generated the electricity to run the LED. But the resulting flashlight would be mechanically complex and not very bright for its size compared to other solutions.

Trying to increase the efficiency of the LED by actively cooling it will end up using more energy than you save.

There is a more efficient way.

The LS is one of the most efficient LEDs available on the market right now. All LEDs have heat problems, but the LS is also one of the largest LEDs in existence so it's heat problems are more obvious.

As you know, LEDs typically become appreciably more efficient when you under drive them.

I would recommend using 2 under driven LSs where the application calls for 1 actively cooled LS. Combine this with a compact and well-designed passive cooling system (like heat sinking the LED to the flashlight housing). Your system as a whole will produce more light for given power input. This translates into higher system efficiency.

The external radiator (required by any cooling system) will also be a lot cooler to your hand since the system total is more efficient.

Another consideration in your flashlight design is to make sure the battery is not noticably affected by the excess heat in the system. For most designs, this will not be a significant issue however.

Flashlights of the future will be LED or some other semiconductor variation. They will produce more light than today's LED flashlights because the light source is more efficient. They may also produce more light because of new power supply (battery) designs (fuel cells, etc). But my money is on the lights become brighter primarily because of the increased efficiency of the LED itself.

This efficiency can be realized by
<UL TYPE=SQUARE><LI>1. Under driving the die (The die would be oversized for the flashlight)
<LI>2. Actively cooling the LED
<LI>3. Improving the die composition to make it more efficient (new LED materials)[/list]

Item 3 by itself is most cost effective, item 1 combined with item 3 is most efficient but more expensive up front. Item 2 by itself or combined with item 3 is expensive and lowest in efficiency of all combinations.

Most LED flashlights of today don't even use the above list. Efficiency is sacrificed to lower cost for a given brightness (read: overdriven). In that mode a white LED flashlight is dimmer, still more expensive and less efficient than a good incandescent. What they do have going for them right now is ruggedness, good reserve output, beam quality and novelty.

Now you understand why the Arc-LS was designed to be underdriven in it's standard load-out. It is the first of a new generation of LED flashlights. As even newer LEDs come out, we will finally be able to offer a flashlight that is better than incandescents in every way.

Peter Gransee

Damm, well o`l well, thanks for putting the matter to rest. It was a fine dream. Good night everyone.