Liquid cooled XM-L Eveready 2D sleeper

I am using the old driver that came with the same flashlight as the led.
I tried measuring the short current and open voltage of some old alkalines and figured, if they act like a constant voltage in series with a resister, I might get around an amp with two of them if the driver gets the most out of them. The new ones may do a bit better.

No, sorry, LEDs don't work like that. If your two alkalines have a voltage under load of 2.9v, and the LED needs 3.2v to run at 1 amp, you'll never shove 1A thru it no matter what kind of theoretical max current your power source is capable of. You could even replace the alkalines with a 50A DC power supply, and at the same voltage the current thru the LED will be exactly the same.

It depends on the driver. Sipik 68s work on 1.5 volts and below. All I know for sure is that the driver can’t increase the power. This one may not increase the voltage, but I don’t know that yet.

The expansion rate of liquids is insignificant compared to gases. Once those tiny bubbles begin to form, you may be creating a potentially dangerous situation. Liquids aren’t easily compressed so when the alcohol boils, pressure builds and something has to give. A few tiny bubbles shouldn’t be a problem at first but if they stick to the hot LED (instead of floating away, cooling and reverting back to liquid), those innocent looking bubbles having no room to expand, become highly compressed. Aiming the light downward would force the bubbles to cling to the LED creating a runaway situation and if your using a 2mm glass lens, it’ll probably blow out. If you really expect it to get hot enough to boil you should be experimenting with a thin plastic lens so any expansion can be observed before ending in disaster.

Using water would raise the boiling point and also transfer heat better so I was wondering why alcohol? Additives can be used to give water better anti-whatever properties. In either case, your going to end up with a well insulated flashlight :frowning: .

I can’t see why you think there is any problem. First remember that there is only a couple of ounces of alcohol and no source of ignition, so the worst that can happen is for the cooling to fail and the led or driver to be destroyed. Then, at this point anyway, I have only two alkaline cells for power, as discussed above, so if there is vapor at all not much of it will collect around the emitter when the light is pointed down and the cooling will still be adequate. If I or one of the others advances to lithium power and a more powerful driver, I am thinking that the whole light should be full of liquid, maybe with holes in the reflector if it is all the same liquid. Then it will cool well in any position. In this case, the whole light is plastic including the window, so there is lots of flexibility for expansion.
I agree that a higher boiling liquid will be needed for some high power and a ridged case. Water seems wrong, especially if it were exposed to voltage when the light is switched off, because it hydrolizes at around three volts. Low electrolyte content water might work and would, as you say cool better, but it might pick up or generate ions in some way. As discussed above, the alcohol has water in it, but I have measured the current under similar conditions and found it small.

My concern was that you hydraulically locked a container using a liquid that evaporates at a temperature lower than the heat source applied to it. Bubbles indicate excellent cooling through evaporation but they’ll also raise the pressure. Higher pressure raises the boiling point and without that phase change, you’re relying on the poor thermal conductivity and convection currents within the liquid to carry the heat away until temps rise even higher. You basically have the first stage of a heat pipe but without a proper thermal path to the outside world most of the heat will get stored so it’s more like a pressure cooker. A flexible lens can relieve pressure but the rigid glass will force it to build. Both can hurt you if they fail but the glass will be under higher pressure, sharper and without warning. It may not fail spontaneously, but an accidental bump might be all it takes.

I am thinking that a ridged light should have some air in it to allow for some bubbles, but more important to allow for thermal expansion of the liquid which will have a higher coefficient of thermal expansion than the aluminum and glass will. The temperature will vary both because of the heat generated by the light and because of the weather and environment. The air will compress with only moderate pressure change.

I was thinking the same thing. An air pocket would be a great idea especially if the liquid gets sloshed around enough while in use. Though, you’d probably get some horrible artifacts. You could also drill a tiny hole in the reflector and add a little accordian type bladder. It wouldn’t need to be very big because liquids build great pressure when heated but relatively little in volume. Your idea of using the area outside the reflector also would be even better and a small air pocket won’t find it’s way into the reflector if the passages are drilled in the right places. Realistically, liquid in the lense area would cool better than air but there are too many drawbacks. A heat pipe would be more practical.

Not the best size or configuration but it would be cool to build something from one of these J) :

Liquid cooling for a combo that's running at, at best, maybe 3 watts is the exact opposite of a sleeper. Since 'sleeper' is most commonly an automotive term, and one I'm quite familiar with, I have to say this thing is about as much of a sleeper as a Kia Sephia with a full body kit, triple rear wing, and a bone stock 1.5L & auto trans.

true, but 400 lumens from a light that probably produces 10 is a power output jump of 4000% and that kind increase in output would make any auto tuner giddy.
Keep in mind that if successful, he can continue to upgrade and can replace the driver for more watts, 3W is not the max output LEDs can do, its just what he is starting at, he is still ‘tuning’ it if you will

Will still be stuck with a max input voltage of no more than 3.2v open circuit, NiMh are 'capable' of higher current but the low voltage vs. the Vf is what limits the current, not the potential short circuit current from whatever cell is used. Oh wait, could use a boost driver... none of which are capable of any drive currents that would call for any exotic cooling at all.

So, if it were in a different light, with a different driver, and using a different power source, then this would be a really neat project.

Thanks Bort.

If it takes 2D, just about any Li-Ion could be used with a spacer.

I don't get it. I'm not saying the idea is horrible (though some here are), I'm saying THIS idea is horrible. You will learn exactly zero from it, because even if the cooling is completely ineffectual the light will still work just like it would have without the liquid. An experiment has to be able to fail before you can know if it worked or not. Just not leaking and not exploding doesn't count, as it wouldn't leak or explode without the liquid cooling.

It really is a dead end. The dome on the LED isn’t going to transfer heat to anything very well. In fact, anything other than gas inside the reflector is likely to absorb more radiant heat.

If you really want it to be a sleeper, take a hint from Match. Gut that thing and get yourself the biggest hunk of copper you can cram in there. Just leave enough space below it for a single Li-Ion.

way to offer constructive criticism, nothing you have mentioned is insurmountable, and besides, he might decide to use an mt-g on his future project with a boost driver, and if this fails then a more expensive one may fail for the same reason, one should make allowances for unexpected ideas because nothing ventured means nothing gained, and i don’t expect he will empty his bank account on this one.
If metal gets into contact with the liquid, the pill or star, it may provide a lot more thermal transfer then we were expecting, maybe someone wants to do the math on this?

Besides, i’m guessing you have heard the expression anything is possible if you have enough money, i heard it form a few auto tuner people i use to know

OK, unsubscribing, the s-m-r-t people have decided this is a wonderful project.

Guess I'm not s-m-r-t. My only hope is that this doesn't end in third degree burns and an insurance claim...

I often do things that are pointless.. I occasionally do things that are dangerous. I really, really try to avoid things that are dangerous and pointless. Marriage, for instance

Also un-subscribing...


me fail english, thats unpossible :bigsmile:

i’m not convinced its a great idea, but if done safely whats the harm in trying?
i’ve done what others would call silly experiments, and learned things from them

I would hope regardless of the outcome this doesn’t turn into an insurance claim, one visit to this thread by the adjuster and claim denied.

It is a very interesting idea, but THIS particular experiment isn’t even going to prove concept, its a total waist of time.

Want liquid cooling? Come up with a way to circulate liquid threw the typical metal heatsinc, the keys is circulation, not emersion in standing liquid. I cool 3 1000w HID’s in a 12x14 room using 100 gallons of water and they lower the temp (vs. forced air cooling) an additional 15* [my 2 55gal res’s are burried 6’ in the ground] Liquid coolin works and it works very well, but just using a few oz of liquid isn’t going to do anything. Take a regular un modded flashlight, hold the head in a 1 or 2oz shot glass of water, run it and measure the temp of the light’s head after a few minutes, now do the same thing again but without water and measure the temp after the same time, I bet the temp is the same. Now circulate the same amount of water threw a radiator and measure, I bet that way you see a difference that way.