Oil cooled LED

[Walterk]

....you are just moving it quicker to the body ....You are just delaying the steady-state system temperature.

Heat travelling faster makes more heat leave the room to make space for new heat from the led, that's heat flux.
At first it shows by the time when the systems arrives at the steady state temperature,
and at second it shows by having a higher or lower steady state temperature.

 

[wquiles]

Heat travelling faster makes more heat leave the room to make space for new heat from the led, that's heat flux.

Yes, lower thermal transfer makes for a more efficient system. And we should always try to have as good thermal transfer as possible.

The problem is that now-a-days, in our eternal quest for "more power", the LED's is releasing significantly more heat than what the body/light can remove - that is precisely why the temperature of the body/light keeps increasing with time and why the efficiency/lumen output drops. Once you are in a situation where you are generating more heat than what the body can dissipate, the material used is no longer the important factor - the more critical factor is the surface area of the light. Since our flashlights have a fixed surface area, then we have an intrinsic limitation: the smaller the body, the hotter it's steady state temperature will be (assuming the LED is being driven at the same levels).

In the best case scenario, the blood of the person holding the light "is" the cooling system, unless:
- you are moving in air (such as in the case of a bike light housing with fins while the bike is moving). Here the effect is that of having a "fan" blowing colder air over the surface of the light. Of course when you stop moving, the temperature of the housing/body will go up again ...

or

- you are using the light under water (like in diving). Under water, the water surrounding the body/light can remove heat from the light much more efficiently than air or even the hand holding. Since for all practical purposes the thermal capacity of the water when diving is infinite (at least relatively to the puny power in a high power LED), the equilibrium point will give a significantly lower steady state temperature than outside of the water. The water basically helps us a lot to remove heat from the body/light much more efficiently.


Those of us old enough in the forum remember one of the very first LED lights from Surefire (L4 I think it was called): that tiny SureFire CR123 body used a 5 Watt Luxeon LED (on a star, thermally attached to the head which was also the heatsink - all one piece). You could not hold it in your hands after 3-4 minutes. I even remember creating a post here in the forum and asking if it was "normal"

And even today, with my Titanium Sunwayman V10R (again, a fairly small body), when using a 4.2V LiIon cell driving a 3watt Nichia 219 LED - if I leave it on a desk on HIGH for just a couple of minutes, it becomes "very" uncomfortable when I go grab it (and forget I left it on HIGH!).

Will

 

[saabluster]

Since our flashlights have a fixed surface area, then we have an intrinsic limitation: the smaller the body, the hotter it's steady state temperature will be (assuming the LED is being driven at the same levels).

I don't have time right now for in-depth responses but I have to say that this is not true. You can engineer around this. Just takes a little out-of-the-box thinking.

 

[Th232]

I don't have time right now for in-depth responses but I have to say that this is not true. You can engineer around this. Just takes a little out-of-the-box thinking.

Active cooling?

 

[saabluster]

Active cooling?

That's one. There is one other way though. Passive to boot.

 

[RoGuE_StreaK]

Active finning - think porcupine!

As things are going a little sideways here anyway, anyone know where to get small cheap fans for active cooling, say in the order of 25mm? Cheapest I've found have been one-off buys for $5+ each.
If I knew more about driving piezos, I'd try to engineer a piezo air pump - those suckers are expensive!

 

[wquiles]

That's one. There is one other way though. Passive to boot.

That would be cool. The more efficient we can make things the better

Will

 

[Steve K]

.... Since our flashlights have a fixed surface area, then we have an intrinsic limitation: the smaller the body, the hotter it's steady state temperature will be (assuming the LED is being driven at the same levels).
.....

I think I'd argue about the "fixed surface area" part of the statement. It's likely that there are practical limitations, since few people want to carry a flashlight the size of a baseball bat. Heatsink fins can be engineered to provide quite a bit of surface area, albeit at the cost of ruggedness and manufacturing expense. Adding a small fan is another solution, with obvious disadvantages.

In a lot of ways, it's similar to the problems of getting heat out of laptops. Ultimately, the most desirable solution is to get a CPU or LED with higher efficiency, thereby reducing the need to dump heat while improving battery life (or reducing battery size).

The development of LEDs that can run at 85C offer some advantages. I can see that they make design easier for fixed lighting (inside of buildings). I haven't run the numbers to see if it would help with flashlights, unless people are willing to wear oven mitts when using their flashlight.

Steve K.

 

[blasterman]

Back to the liquid thing............I'm pretty sure aluminum or copper has orders of magnitude better thermal conductivity than most liquids, excluding some weird things like mercury. So, I would guess that if you immersed a star mounted emitter like an XP-G or XM-L in liquid, and hit it with enough current, you'd likely boil the liquid and quickly kill the emitter. Also going by the chart (below) regular water has better thermal conductivity than paraffin, or any petroleum based liquid for that matter.

http://www.engineeringtoolbox.com/thermal-conductivity-d_429.html

Flashlights share a lot of the same thermal problems that standard light bulb retrofits have, and the solution in both cases seems to not deviate much from a standard cylinder. A big difference though is LED light bulb retrofits, and even higher end fixtures tend to use lots of 1watt and smaller LED's to help spread the load.

Obviously adding fins helps with convection area, but the positive thermal effect of fins limits very quickly in a passive cooled environment. Plus, the fins need to be thick and preferably radiate from the entire surface area of the immediate slug or plate catching the initial heat. Not sure why custom flashlights can't be CNC'd with thicker and broad fins immediatley behind the LED slug (??) or is it an aethestic thing?

 

[AnAppleSnail]

Passive cooling fins are quite different from what you guys are envisioning. For the most part, a passive heatsink cools by conduction/convection to air and radiation to things.

1. Heat air. The air is then at the flashlight temperature and cannot absorb more heat, until... that air rises away, drawing in cooler air. This requires open air, much like a tailstand on a tabletop. Conventional design includes stubby fins with the vanes pointed 'up' in normal orientation. Unconventional design would be a hyperboloid tower.

2. Radiation. Radiation cools directly to surrounding surfaces, and is mostly based on (Tlight - T(walls, floor, ceiling, sky)). Fins DO NOT help with radiation, except to increase the diameter of the radiator. Radiating is, well, radial. Fins sticking out face each other and cannot radiate to each other, so radiating heat is best done by a few long vanes. Conventional design includes high-emissivity coatings (Black anodize, copper, etc), 'bold' cooling features (Big cooling ribs, etc). Unconventional design would be to attach a matte-black half-inch-thick plate to a flashlight's heat rejection.

Passive cooling without the help of a hand is limited by heat rejection. This is mostly improved by:

More conduction to air (Higher surface temperature)
More radiation to air (Higher emissivity)
More conduction to non-air things (Hand, water, fans)

 

[AnAppleSnail]

Back to the liquid thing............I'm pretty sure aluminum or copper has orders of magnitude better thermal conductivity than most liquids, excluding some weird things like mercury. So, I would guess that if you immersed a star mounted emitter like an XP-G or XM-L in liquid, and hit it with enough current, you'd likely boil the liquid and quickly kill the emitter. Also going by the chart (below) regular water has better thermal conductivity than paraffin, or any petroleum based liquid for that matter.

Paraffin coolers don't work quite that way. Consider an ice flashlight, 300 joules per gram of melt. If I want a flashlight to run for (t) seconds at (W) power, I can plan on a safe thermal load. But the ice heat-sink (A true heatsink, not a block of metal that absorbs some heat and radiates it) will NOT give me great runtime - it's only for sprints.

Suppose I modify my Torchlab H3 triple with 30cc of ice (About 1 fluid ounce). When it has sat in a freezer overnight, this gives me (300*30) joules of melting before the flashlight reaches 1 degrees C. That is 9000 watt*seconds of heat that can be absorbed - or I can run it at (3.7v * 4.5A) = 15W for (9000/15) about 10 minutes before the ice will all melt.

These paraffins are chosen for practical melting points. They may melt at 50C, with a lower heat of fusion - so that in peak running the heat is absorbed to be released later.

 

[deadrx7conv]

Liquid cooling isn't too feasible for small or handheld lights. Its not worth even talking about it. For thought, research sodium filled engine valves and CPU heat pipes. But, I do notice that the head of my flashlight is seriously hotter than the tail. So, if there is a way to move the heat across the entire flashlight, we might benefit with longer run times before not being able to handle the flashlight.

Where liquid 'centralized' cooling would benefit is either multiple lights, or very high power lighting, when neither have a surplus of heat sinking area either due to a lack of space, or by trying to make the light visually appealing.

A tiny heat sink with coolant tubes might be easier to integrate into wall/ceiling lighting. Can you picture a single small fan cooled rooftop radiator with a loop of antifreeze:water running to:fro all LED ceiling/wall/floor fixtures? No more worrying about trying to hide those 5lb aluminum finned heat sinks in your interior decorations!

Another example would be a 100w LED mounted on a 1/2" pole. Would you want the basketball sized aluminum finned heat-sink on the top of the pole? Or, tuck a couple cooling lines up the internals of the pole with a radiator/sump/pump mounted elsewhere. The thought of multiple 8000 lumen lights, with no noticeable heat sinks, across a large yard, mounted remotely or on small diameter poles, could be interesting for someone.

No bugs to work out. There are plenty of CPU liquid cooling solutions. There are plenty of automotive type cooling(antifreeze/oil/ATF/PSF...) And, many houses have run radiant or forced hot water type heating systems for years. You'll just need competence to make it leak and sweat proof.

So, whatever happened to Switch Lighting and EternaLeds Hydralux?

 

[blasterman]

Paraffin coolers don't work quite that way. Consider an ice flashlight, 300 joules per gram of melt.

I think we're talking about two different applications. I was referring to LED retrofits like those below that use liquid filled capsules that do *something*, but we're not sure what it is:

http://www.switchlightingco.com/
http://www.candlepowerforums.com/vb...-LED-Globe-A-Shape-Bulb-Eternaleds-HydraLux-4

The idea of using state-change to absorb heat in short run applications sounds totally feasible to me. Anybody who's taken HS chemistry understands how this works when you graphed the amount of energy it took to melt ice, or wax, or mothballs, etc. The one caveat I would though, and I'm sure you've thought of it, is once you converted the solid to a liquid you would then have to wait longer for the now liquid to realease it's energy and resume a solid state -vs- using a conventional design.

Fins DO NOT help with radiation, except to increase the diameter of the radiator

Not entirely true....even in a vacuum fins aid in radiating thermal energy away from the source. Best example is cooling fins on the RPG used by space probes. But yeah, I get what you're saying and agree. I'm constantly telling people to avoid high performance CPU coolers for passive cooling Bridgelux because the thin and dense fin arrangements actually trap more heat than short, stubby ones. For flashlights though my visual would be a wide collar of thick fins radiating with a square length to height ratio around the LED contact area. This is very similiar to some LED retrofits and 1watt laser pointers I'm seeing. This allows you to use both convection and hand thermal transfer. While a stationary flashlight obviously has the convection problems you mentioned the reality is that most of the time you're moving a flashlight around, etc. Not sure if the fins would yield a 1% improvement or 20% improvement, but at worse it's extra mass to absorb heat.

What would be really cool is if you were able to use carbon fin heat piping on a flashlight, and then polish it.

 

[idleprocess]

That's one. There is one other way though. Passive to boot.

Clever use of a heat pipe comes to mind - helps move the heat away faster than sinking it into a block of metal - but only achieves so much.

Where liquid 'centralized' cooling would benefit is either multiple lights, or very high power lighting, when neither have a surplus of heat sinking area either due to a lack of space, or by trying to make the light visually appealing.

A tiny heat sink with coolant tubes might be easier to integrate into wall/ceiling lighting. Can you picture a single small fan cooled rooftop radiator with a loop of antifreeze:water running to:fro all LED ceiling/wall/floor fixtures? No more worrying about trying to hide those 5lb aluminum finned heat sinks in your interior decorations!

This has occurred to me as well, although I suspect it would not be practical on the small scale of most residential / commercial arrangements where fixture size isn't a problem so long as the per-fixture power is reasonable. Hobbyists will likely just use a fan if heat removal / sink limitations are real issue ... commercial applications will use highly-engineered thermal design.

Worked really well for the flying searchlight, albeit the fluid seemed largely to be for speed of heat removal and aerodynamic convenience.

So, whatever happened to Switch Lighting and EternaLeds Hydralux?

I barely remember Hydralux ... Switch Lighting is going on 12 months past their original estimate on retail availability, although they seem a great deal more serious about launching a generally-available product.

 

[saabluster]

Clever use of a heat pipe comes to mind - helps move the heat away faster than sinking it into a block of metal - but only achieves so much.

OK. So you have the implement. Now what is the implementation? Sorry if I'm being coy. Just trying to stimulate the minds here rather than just blurt something out. Although it is true that what I am referring to "only achieves so much"(that applies to pretty much everything) it achieves much more than what I think you are thinking.

 

[Th232]

Heatpipe stretching to the other end of the flashlight, normally the coolest part? That will help evenly distribute heat instead of having some kind of a thermal gradient from one end to another (how steep that gradient is to begin with is another question). That combined with finning the entire body would be interesting to see, as would holding it. Assuming you've designed the fins correctly and with enough space around them. Might actually end up more like a pin type heatsink than one with fins.

I've never seen a flashlight where the body was completely finned. Maybe it should be called the Porcupine.

 

[RoGuE_StreaK]

I'd previously investigated doing that (heatpiping along the body to the rear), but couldn't find much info on DIY-level heatpipe construction; the wicking materials, and somehow filling and sealing the pipes under vacuum seemed beyond reach. But if I'm more than happy to be shown a way!
My latest redesigns have quite extensive finning, will have to finish them off to see how they affect machining costs - not to mention whether it'll actually do anything...

 

[MikeAusC]

Heat pipes are so cheap there isn't much point trying to match the performance of wick-lined Heat Pipes. Try Element 14 or Farnell.

 

[idleprocess]

OK. So you have the implement. Now what is the implementation? Sorry if I'm being coy. Just trying to stimulate the minds here rather than just blurt something out. Although it is true that what I am referring to "only achieves so much"(that applies to pretty much everything) it achieves much more than what I think you are thinking.

That's about as far as the thought went ... expertise in thermodynamics is something I lack.

My thinking was that even in situations where you have a massive heatsink, the heat will only travel so far - as many have witnessed with C/D maglite mods where only the head of the flashlight gets appreciably hot while the tail might be close to room temperature. 'Seems that if you were to transport that heat faster ala the phase change that heat pipes allow for, you could realize more utilization of your heat sink's mass and surface area.

As far as implementation ... not so sure. Simplest seems to have an interface to another heatsink in the tail, although clever implementation could perhaps use the heatpipe more like a passive radiator, with the heat dumped periodically along its length as opposed to at its terminus (my ignorance comes into play here - I don't know if they can be made to work this way). Another thought was some triple-walled battery tube arrangement - where the battery tube itself acts as the heat pipe - but that would be tricky to execute and likely troublesome since it would contain the fluid within a thin-walled construct that's also a structural part.

 

[saabluster]

Heatpipe stretching to the other end of the flashlight, normally the coolest part?

Although this idea does have merit it is not exactly out of the box. That is the obvious solution but ultimately not the best when trying to combine super high current in a small package without burning the user's hand. In fact it may, at certain current levels, do quite the opposite.

That's about as far as the thought went ... expertise in thermodynamics is something I lack.

My thinking was that even in situations where you have a massive heatsink, the heat will only travel so far - as many have witnessed with C/D maglite mods where only the head of the flashlight gets appreciably hot while the tail might be close to room temperature. 'Seems that if you were to transport that heat faster ala the phase change that heat pipes allow for, you could realize more utilization of your heat sink's mass and surface area.

As far as implementation ... not so sure. Simplest seems to have an interface to another heatsink in the tail, although clever implementation could perhaps use the heatpipe more like a passive radiator, with the heat dumped periodically along its length as opposed to at its terminus (my ignorance comes into play here - I don't know if they can be made to work this way). Another thought was some triple-walled battery tube arrangement - where the battery tube itself acts as the heat pipe - but that would be tricky to execute and likely troublesome since it would contain the fluid within a thin-walled construct that's also a structural part.

Well I don't consider myself an expert in thermodynamics either. I have done a lot of thinking and testing in that area though. At least as it relates to flashlights. The conclusion? That a great many lights today are being designed all wrong. There are two instances where you need to reconsider traditional flashlight design. 1When flux density reaches a level that the heat from steady state operation rises above comfortable user tolerance. 2. When creating a larger sized light but still want the flashlight form-factor.

Case in point. The Olight SR90. Now this is just my opinion mind you, but I think that light reeks of poor design. All they did was take a normal flashlight and hit 300% enlarge on their printer and call it a day. The result is a bloated beast. The scaling-up process of the traditional form-factor only works so far. At a certain point you have to rethink it.

The solution is something I call a "thermally decoupled heatsink". Instead of the whole flashlight being the heatsink only a small portion in my design is used for heatsinking. Even if the rest of the light is metal the heatsink is thermally insulated from it. Normally this would be a bad thing. In fact there are a few lights out there that by default have what appears to be a similar idea but when you scratch the surface a little more you will see that this idea is a bit more novel. At least in the flashlight world. Some lights are made primarily of plastic and so to get rid of heat they add a little heatsink. The end result is completely the opposite of what I am talking about though. They actually have to reduce power with this design and here I am talking about increasing power.

The key is to use a heatsink that in and of itself is a phase change device. It would rely not on heat conduction to a hand but radiation and convection. The higher the temperature difference between the heatsink and ambient the more heat can be disposed of and the more efficient these forms of moving heat become. The key of course is to do this while keeping junction temps down. Which is why the use of a phase change element is absolutely crucial to the design. I won't get into all the specifics of my design but I assure you this idea is not pie-in-the-sky and that it does indeed work. I don't have the resources to commercialize it but if Surefire or the like want to get with me we can make some magic happen.