Liquid Cooled LED?

Hello to all!
New here and it seemed appropriate to stop the lurking (a week now) and join this forum for some advice.
I've built several headlamps, knives and the like for hunters in the Rockies over the past decade... Only recently has the Super high output (anything over 300Lum) LED bug given the big slap in design profiling. Have decided to do something a little beyond the realm of conventional design and work within a design of power efficiency.

Why hasn't anyone developed a liquid cooling system for HI-LUMEN leds? or is there?oo:

As it stands now, the best we get for heat dissipation is a small chunk of aluminum, maybe the reflector, and possibly the flashlight body... and even then .. the thermal transfer from improper fitted pieces defeats the purpose of having multiple points of heat sinking!

If this thought has passed through your grey matter at one time or another, I've got full scaled CAD recipes in any language you want. So far it's just a headlamp front light module, but can be easily reworked into almost an format (regardless of size restrictions).

Please, PLEASE, attack me with your ideas if you are up for a collaboration!

Heat pipes come up now and then. Circulation of a liquid to effect good cooling would be one roadblock. I really like the form factor of two C cells in parallel. Such a design might work for liquid cooling because it tends to hold the light in a certain position.

I don't think it will ever come about at a commercial level as increased efficiency and running them at design amperage solves the problem for "normal" people.

The P7 for example seems to have a low setting on most of the lights that is about equal to hign on my ancient three luxion light...

It could work, but active liquid cooling isn't any more necessary for a flashlight than it is for a snazzy gaming PC, although I guess hand-held lights (once the light gets hotter than 100F or so, of course) and divelights are technically liquid cooled. Still, the main idea is that, for a flashlight, integrated liquid cooling is prohibitively complicated, expensive, and maintenance-heavy. An easier alternative would be forced-air cooling (i.e. a fan, which many lights use, although most of them are incans) or, even easier, passive cooling, which is the chunk of aluminum/copper/etc. you mentioned.

If you want to do it anyway, though, please post lots of pics.

Oh, and :welcome:

I have no idea how the mechanism works, but the high-frequency drive unit is my floorstanding speakers is described as 'Ferrofluid Cooled'. What does that mean? And can we use it?

Actually it is very interesting stuff. Look under 'Heat transfer' in Ferrofluids. Looks like another use for those little charging magnets.

One workable solution to liquid cooling for a portable hand-held flashlight would be to have a portion of the body of the light surrounded by a membrane filled with a chemical solution that absorbs heat quickly. This membrane would have two, one-way valves to insure that when the user squeezes the membrane that the pressure will force the liquid in one direction only, and a somewhat wide and flat spring to expand the membrane after it has been squeezed, making vacuum to pull the liquid back into the membrane. The out-bound valve would be connected to a thin walled hose that would be integrated into a copper heatsink (or perhaps the heatsink has pre-drilled pathways for the liquid to follow) so that it could absorb as much heat per pump and be able to circulate quickly enough.
This could eliminate costly pumps and fans, and would be relatively cheap to produce and maintain. All the user would have to do is squeeze and release the membrane (i suppose just tightening and loosening grip on the body of the light) to pump the fluid, and that could be sufficient enough to aid passive cooling to a degree that it stabilizes temperature effectively.

It's just an idea at 2:25am, so I'm sure there are some things I'm overlooking (like how the liquid will release the heat it has absorbed fast enough). But you asked for ideas, and this was the first thing to come to mind.

If you want to experiment and go through the trouble of building a liquid-cooled light, one reason I can think of to justify this type of system is a lot of high-power LEDs in a small space. Say, 50 Luxeon K2s mounted to an aluminum manifold. Just keep pumping coolant through the manifold and the "head" with the K2s mounted on it will be relatively small, allowing you to move the heatsink to a remote location.

(Only saying K2s is because they're dirt cheap and pretty bright for the $.)

Well, you don't even need liquid cooling when you do something really silly, like mounting 70 LEDs onto a single aluminium slab :devil:

no one as of yet has came up with practical design for liquid cooled flashlights.
Theoretically not only it could be done, but it SHOULD be done.
LED performance increases with the drop in temperature while battery pergormance drops with the drop in temperature...so why can't we provide a "binary" transfer medium to heat the batteries while keeping the LED cool?
thread that the idea originates: [Written by the big G almost 4 years ago]
A paradox to thermal relief??!?!?]

water has high heat capacity, but it also has properties that may be detrimental to the integrity of the assembly. I've once thought of something like a sterling engine...that forces a liquid from the heat source [LED] to the "heatsink tailcap" and returns to the bezel by gravity. the light, if created, should function like a regular military angle head mounted with the battery region above the bezel.
Since I'm not a machinist or a machinist's son...the idea somewhat dissolved itself on the drawing board.

Other methods of thermal transfer such as peltier chips have also been discussed, but due to the inherent property of high current consumption peltier junctions have that the idea too have been canned and tossed

Instead of liquid cooling, which would be too complex. How about finned heat sinks with vent holes in the flashlight body. I'm envisioning something
along the lines of current assault rifles where the barrel is ribbed for greater surface area, then a vented shroud where the foward hand grip is.

When machine guns first came out they too were watercooled to deal with
over heating barrels. Now almost all aircooled.

toadman said:

Instead of liquid cooling, which would be too complex. How about finned heat sinks with vent holes in the flashlight body. I'm envisioning something
along the lines of current assault rifles where the barrel is ribbed for greater surface area, then a vented shroud where the foward hand grip is.

When machine guns first came out they too were watercooled to deal with
over heating barrels. Now almost all aircooled.

Click to expand...

Been done check out Raidfire Spear

An ion drag pump placed directly behind the die would do the trick, and would be so small and inconspicuous that all you'd need is a tiny hole somewhere within the sealed part of the head (so the flashlight would still be waterproof) to add miniscule amounts of water with an eyedropper from time to time. The LED itself would look slightly larger on the bottom, other than that it would look like a completely standard flashlight.

And all you have to do for one of these lights is wait until the technology improves; current ion drag pumps require ~60 volts (!).

CAD Rendering pix...

Well at least close to what i had in mind... These are converted from ACAD to FormZ and then made a few annotations to give a better idea.
Some prelim notes:
- all spacers, main body construction, pivot bulkhead, and face plates are stock 7075
- all screws are stainless (yes, the REAL weight behind everything) allen and will be countersunk into the plates
- All sliding areas (eg - pivot lock spring etc. ) are brass inserts.. but still not overdone
- The 4- 5mm LED's will be removed and 2 "almost identical" reflectors will take their place. We're going for 3 lux rebels trimitters here...
- initially the center beam will focus to a VERY tight spot (maybe 15-20 deg.) with the reamining two, towards the edges of the module, set at 60 deg and 90 deg. Going for a long elliptical light pattern with all three of them on.
- initially, i planned on adding the switch to the front, but i've come across a few very unique touch field modules in a very small size that will be incorporated into the battery housing area. Less electronics in the front will reduce the heat buildup as well.

And some pix:










Tried to keep them as "non-confusing" as possible, but if you have any questions about dimensions, weights, CNC tolerances etc etc... let me know

....and some "post-liminary" notes:
- I'm leaning towards am ethylene glycol (that's radiator fluid to all the young ones out there) as the coolant... it seems to work well in cars, and is non-corrosive to anodized AL
- All of the o-rings and seals which are usually rubber will be a poly-nylon based butyl. This has ZERO reactivity with ethylene and provides a slightly firmer seal.
- I may be well worth the while to mount all of the LEDs directly to the aluminum... in which case, i would opt for XR-E Q5's and run 10 in a sequence of 4 in the middle and 3 on the top and bottom of those.
- uh, my eyes are hurtin from converting all of these over to the new formZ layout so i have got to take a break.

.... and i was watching survivorman the other day when this idea popped into my head! He was boiling water inside of a plastic liter bottle (like coke or something). as long as the water stayed in contact with the bottom of the bottle, and only the flames touched it, the botle would not melt and could sterilize the water!
If the cooling efficiency of water can change it's core temp without plastic.. just by being incontact with it, imagine what we can do with a micro cooling unit for LED's.

next stop: 0deg Kelvin! oo:

traplight said:

Why hasn't anyone developed a liquid cooling system for HI-LUMEN leds? or is there?oo:

Click to expand...

Probably because there are no reason to do it!

Your have to get the heat away from the led, a aluminum flashlight can easily get the heat from the led to the body of the the flashlight, using a liquid cooling would not be a big improvement on that.
Getting the heat away from the body is another problem, this can be done by airflow or a hand holding the flashlight or a water tube to a external radiator, but who would want that solution?

Haha the ion pump.. thats a great idea.. and ferrofluid cooling in audio tweeters (except even that requires small amounts of mechanical movement and we're dealing with solid state)

Yes, HKJ people have always been consciously happy with inefficiency, our capitalist ways work upon that principal that "inefficiency creates employment".. and we can't have that any other way.. or can we?

The lifespan of an LED is directly proportionate to the amount of heat that can be dissipated from the phosphor cup and the anode. When a manufacturer gives a suggested diode lifespan based upon INCREASED forward voltage and current, they have calculated (with standardized available consumer/manufacturer application non-withstanding) that the led will have just that, a lifespan.

If the LED is hot, then it's knocking years off of its life. Closer proximity cooling will guarantee longer life and a more stable operating environment. There's a reason why CPU manufacturers are trying to reach a 0deg operating environment for solid state electronics.

Basic Rebel data (can download the entire pdf spec sheet from www.luxeonstar.com):
Lumens- 240 lm @ 350mA, 435 lm @ 700mA
Max Continuous Current- 1000 mA
Forward Voltage- 9.45 Vf
Dominant Wavelength- 6500K
LED Type- Tri-Emitter
Radiation Pattern- Lambertian
Weight- 5.5g

We're talking 3 of these campers possibly running at the same time. The ranges for output being (math when it comes time ) 60lm, 140lm, and 200lm per tri-mitter! and that's just going with the basic 350mA current... and not including another probable bypass for a jump to 750mA (x3) for short periods.

All tweaks aside, i guess what i'm trying to say is this:
Instead of trying to squeeze every last bit of light out, i'm trying to 1)increase the longevity of the diode, and 2) trying to give the longest runtime.
These 2.5 hour 100lm runtimes are not going to cut it:sigh:

I'm envisioning an ultra-powerful, overdriven array of emitters, cooled with glycol-water and a massive thermoelectric heat exchanger.

How about phase-change cooling? Some PC overclockers do it. I wonder how much current a Rx-bin XR-E will handle at ~-195℃. It shouldn't be too hard to rig up an evaporative liquid nitrogen cooling block.

FWIW, the Prius has liquid cooling for its motor controller (as does the Ford Escape)

HKJ said:

Probably because there are no reason to do it!

Click to expand...

There is every reason to do it if you are trying to run LEDs(especially multiple LEDs) hard.

HKJ said:

Your have to get the heat away from the led, a aluminum flashlight can easily get the heat from the led to the body of the the flashlight, using a liquid cooling would not be a big improvement on that.

Click to expand...

Alumnium works great in normal lights at normal power outputs, but, relative to other solutions alluminum is not all that great for transfering the heat from high power outputs.

HKJ said:

Getting the heat away from the body is another problem, this can be done by airflow or a hand holding the flashlight or a water tube to a external radiator, but who would want that solution?

Click to expand...

This is a good point. Another thing that can help get rid of waste heat is through radiation. Using black anodizing or a high emissivity paint, such as Krylon "High Heat"ultra flat black, can help dramatically reduce the body temperature.

Ultimately I think the best solution here are heatpipes. They work like the water cooling device traplight wants but are far more easy to employ.

saabluster said:

Alumnium works great in normal lights at normal power outputs, but, relative to other solutions alluminum is not all that great for transfering the heat from high power outputs.

This is a good point. Another thing that can help get rid of waste heat is through radiation. Using black anodizing or a high emissivity paint, such as Krylon ultra flat black, can help dramatically reduce the body temperature.

Ultimately I think the best solution here are heatpipes. They work like the water cooling device traplight wants but are far more easy to employ.

Click to expand...

Your have to look at the total path of for the heat, i.e. from led to air.
This includes many steps:
From the die inside the emitter to the surface of the emitter.
From the emitter surface to the emitter mounting
From the emitter mounting to the surface of the light (This is usual multiple steps).
From the flashlight surface to the air (or to a hand holding the flashlight).

All of these steps must have a low heat transfer resistance, but there is no reason to make a single one of them much lower than all the others. I.e. if the total value is 20 K/W (Temperature in emitter increases 20 kelvin for each watt), changing one item from 1 K/W to 0.7 K/W would not really help the total.

Stuff like heatpipes do not even do that, they just moves the heat to another location, but in a flashlight your do not really have many other locations for the heat.

I believe the best way to move heat around inside the flashlight is to mount the led on copper and that copper is in good contact with a tick aluminum flashlight head and body, this way your will get a low K/W from the emitter to the flashlight body. The next task is then to design a body with cooling fins.

Comparing the cooling in a flashlight to the cooling in a computer is not valid, in a computer your have a big heatsink with a fan, this give your a very low K/W (below 1 k/W) for the heatsink to the air. To get advantage of that low value, the path from the cpu to the heatsink must also have a very low K/W.