Weird stuff soldering copper Heat pipes question.

VidPro

Flashlight Enthusiast
Joined
Apr 7, 2004
Messages
5,441
Location
Lost In Space
This is a really weird question, i will try and break it down into parts. its IDEAS, and subject to major change, due to (as usual) i dont know what the heck i am doing.

8 or 9 6Watt RGB leds need heat sinking, and some sort of metal "art" tech art type of light fixture.
so i am trying to make a heat sink, that looks like some sort of faux art too, that is function over form.

i was thinking about attempting home-made heat pipes, some sort of liquid inside them that moves , using convection, this leads to a Lot of questions. If there is a moving liquid , even convectionally flowing, the heat gets moved away, as opposed to Held in the location, and its cheaper than way to much costs in copper.

First the possibility of the design, Remember CHEAP, available, and Home made in garage , not real shop.
---=---=---=---=---=-------
. . || . || . || . || . ||
. . || . || . || . || . ||
. . || . || . || . || . ||

copper PIPES (from home depot type) teir down off of T-junctions, connected to a top pipe that connects them all.
at the end of the pipe a simple solderable pipe cap, on the pipe cap the LED itself sits (at the ends of the legs dangling down)
angles of the pipes is adjusted for the lights location, to "aim" them, and then a plastic difusion box is placed around either the whole contraption, or each led item.


there should be a slight vaccume Sealed into the inside of the (heat) pipe, for 2 reasons, there is less air/oxygen to react to the metals, or the internal liquid, and vaporisation in a slight vaccume would occur at lower temps.

1) what to fill the pipes with? so it lasts 20 years.

1A) does oil convection flow, just like water can, with hot oil rising?
1B) how is getting oil on the pipe going to effect soldering in a seal?


2) wierd idea time, to create a VACCUME, a perfectly easy way to do that is to "steam" the insides, say using water , steam will replace all the air, then Seal this up, while its still steaming. When it cools down, a strong vaccume would form as the steam condences back to liquid. strong enough that it could even collapse the pipes.

2B) How would a soldering go when steam is still comming out, ya know trying to accomplish the SEAL, with the steam trick. soldering water pipes with water in them (as you know) doesnt occur till the water is GONE :)

3) a substance that has greater volitility , so it vapors faster at temperatures , but is not flamable, and can be purchaced (and tested) locally.

4) not an option, is what i have seen done before, where the vaccume is done with Valving, like air compression valves, any type of rubber plastic or whatever seal, that is changable, would not last long enough, besides the fact that i dont have a vaccume pump.

5) effects of various substances like 50-50 antifreeze, over TIME, via the metals themselves, ya know antifreeze becomes blob-o-freeze in 3 years , even closed water cooling systems (for computers) with the Right stuff in them, end up with terrible reductions of transfer at the CPU sink in less than 3 years. nasty junk collects around the water block on the cpu.

6) zero maintance of the substance internal to the pipes, one shot one time, get it right and hope that whatever coagulation, or conversion, or reaction that can or will occur, will not change it Much in 20 years. that is why i am thinking a 5W oil, but even oil has PH issues around metals.

so you get the idea? copper water pipes, turned into heat pipes AND a fixture at the same time, done with simple tools and garage shop guesses.
 
Last edited:
Well you'll have a hell of a time soldering it closed with water inside.
When heating the pipe assembly for the final joint any water inside will absorb the heat - needing more heat leading to steam - blowing the last cap off.
So solder the whole thing up dry and add a valve to evacuate air from it and add water or refrigerant.
In fact make the thing then get a HVAC mechanic to add a piercing valve, bring it into 15"Hg vacuum and give it a charge of R134a. If the Leds are pointing down it can act like a heatpipe. But not really because heatpipes use capilary action and your tubes are too big.

If you can draw your own vacuum you could buy a valve and can of R134a at an auto part store. Have fun.
 
i can draw my own vaccume if a straw or a aquarium pump will do the job :)

freon type substances would be a great idea, cost less than $3 to fill the thing with that. and it should last a Long time if any openeings are permenentally sealed, like welding closed any valving.

Its the cost of an evacuation, and the fact that even Car AC systems dont stay sealed for 20 years. usually the failure is from movement, but then even the house systems dont stay up for 20 years, usually. i would be taking a risk on the valve items holding, and at fair high pressures when heated with 50watts of heat. ya see, nobody is going to say, Hey its not putting out cold air, fix it. it will just quit working right and 10 years later , i wont even remember how it was made :)
i would want to avoid any relying on plastic, neopreme, rubber, they all break down in time faster than lead and copper.

that might work even without a full evacuation, as there wouldnt be icing up in a expander thing. some way to just stuff a 134A valve on a copper pipe, then i could do it at home. Where the heat is from the led, the fluid would freak out, and run away, which should rotate the fluid in there pretty well, just like a boiling pot of water doesnt go above 210* , oops untill you put it in a pressure cooker, argggg.

i really have to do it at home, because it wouldnt be the first $100 i would get stuffed for a HVAC, but the problem is things might not be exactally right the first time. and going back and forth to HVAC guys, would be like throwing money in the air and running through it :D

Mabey there already does exist a copper pipe end fitting, that will match up close enough to a freon valve of some sort, i could just solder the threaded fitting to the end of the top pipe. there is always some crasy adapter things of different sizes, as long as i use Standard parts that get used everywhere (and are available locally).
 
Last edited:
i can draw my own vaccume if a straw or a aquarium pump will do the job :)

No that wouldn't get into a vacuum, I don't think you'd get below -14psi which is when you start drawing a true vacuum.
0psi on a gauge is 14.7psi absolute(psia) thats the pressure of the atmosphere on everything.
Your lungs aren't capable of drawing down that far. Aquarium pumps won't do it either sorry.
You'd need a positive displacement pump, you could make one out of a engine cylinder turned by another motor.
Check this out
I'm actually one of those A/C guys but I just operate equipment the company won't let us repair it they call in a contractor :shakehead hence a lack of tools to put the training to use with.
 
what percentage of 134 do you figure would be appropriate ?
filling it by 1/5th 1/4th or 1/3rd (liquid form)? say if i didnt get a evacuation. of course it probably needs a safety valve :) but if i HARD test it, away from humans, i can see if there is any bulging , before putting it in application.
 
Last edited:
No that wouldn't get into a vacuum, I don't think you'd get below -14psi which is when you start drawing a true vacuum.
0psi on a gauge is 14.7psi absolute(psia) thats the pressure of the atmosphere on everything.
Your lungs aren't capable of drawing down that far. Aquarium pumps won't do it either sorry.
You'd need a positive displacement pump, you could make one out of a engine cylinder turned by another motor.
Check this out
I'm actually one of those A/C guys but I just operate equipment the company won't let us repair it they call in a contractor :shakehead hence a lack of tools to put the training to use with.
A human can draw up to ~15-20inHg with effort.

Or attach something to the intake of an air compressor or the intake manifold of your car.

On earth, no vacuum will be able to collapse copper tubing.

If you get it to work initially, I bet that it will keep working for a very long time.
 
what percentage of 134 do you figure would be appropriate ?
filling it by 1/5th 1/4th or 1/3rd (liquid form)? say if i didnt get a evacuation. of course it probably needs a safety valve :) but if i HARD test it, away from humans, i can see if there is any bulging , before putting it in application.

Well if you give it a 1/3 in liquid and then flip the can and blast in some gas while venting air from the other end that could be the best you'd get without an evacuation. Of course then theres 2 valves on it, more places to leak over time.
 
Well if you give it a 1/3 in liquid and then flip the can and blast in some gas while venting air from the other end that could be the best you'd get without an evacuation. Of course then theres 2 valves on it, more places to leak over time.

which is heavier? the liquid turning to vapors/gas fast enough , would push out stuff, even out the same valve. how would air get trapped in it, if i releace some gas back out, probably the dang moisture that would be what gets trapped, hmmm, and that would be bad, the moisture :-(.
 
You could give Newbie/Jarhead a shout. IIRC, he did some messing around with heatpipes (including making his own). You can get his Instant Messaging link in his CPF profile.

Paul
 
OK.

You have a very complex system here that may not need to be so complex. (and maybe it does!)

First, you need to determine just how much heat energy you need to reject (Watts is a good unit to work in). This is WASTE watts, not the wattage of the lamp- a lot is light, but there is heat as well...

Then you need to determine if you can sink that heat away with a simple air heat exchanger (heat sink- finned aluminum or finned copper for example).

Then, if that is not enough, you may have to go to forced-air cooling of the heat sink (add a muffin fan to the fins).

If this still is not enough, the best free-standing solution (in my opinion) is to use those CPU cooling devices. They are very efficient, cheap and use little power. It seems obvious that this will be a home illumination device; not something you will carry around with a battery pack.

Back to the current idea of using 134a refrigerant. The idea is sound, but executed in the manner you propose is not. 134a (and all refrigerants for that matter) work on a decompression cooling effect (which has a fancy name I cannot remember right now). This means you need a compressor, evaporator and condenser. The evaporator gets cold as the liquid expands into a gas (actually it absorbs heat) and then it is compressed (further heating) and air passes over the condenser to condense the hot gas back to cool liquid as the heat is removed. Notice that all the cans of 134a sitting on the shelf at the store do not ice up? that is because they are steady-state. Shaking them (to force expansion of the gas) causes them to cool, which you feel in your hand. Your steady-state design will not remove any heat.

So, with that out of the way, you can forget about the vacuum pump, which you wouldn't need, even if the static 134a idea would work. If this would work, all you need is two valves, one at each end of the system, and you purge it out with the 134a. This will exclude all the other gasses and leave just the desired gas. Using steam will leave water behind (which is hydrogen and oxygen) and trying to draw a vacuum with no replacing gas (purge gas) will simply have air (nitrogen-oxygen-others) under a vacuum. Having 'low pressure' (as we refer to it in low pressure plasma systems), does not mean there is no oxygen. You have to pull vacuum, and then back-fill with an inert gas (we use Argon typically) and then repeat. After three evacuations, it is usually at a low enough oxygen content to begin operations...

Back to the heat pump for the CPU: you will notice a continuous flow of liquid. The CPU heats the liquid, which is drawn away, and then it is cooled and returned. I suggest this method as the best bet if you need to pull that much heat.

For liquid (particularly in copper tubing) I suggest de-ionized and/or distilled water. Available at the grocery store for a dollar a gallon or so. With the other minerals out of it, it will not react with the copper for over 10 years. Do remember that the solder has many elements in it that may react with the 'clean' water to form other compounds, so I suggest silver braze for all connections, if you must braze, but there is another option to brazing I will cover below...

Try this idea: Use a solid copper or aluminum heat sink 'stick': Imagine a 1" diameter rod (solid) with a coil of copper tubing (1/4 inch) wrapped around it. Wrap it around a 3/4 inch diameter rod, so when it releases tension, it will fit snugly on the 1 inch rod. Pot the coil with thermal epoxy on the rod (or solder it for an 'intimate' contact). Make all connections from this coil to your heat pump with Swagelok fittings (compression fittings)- either brass or stainless steel. These won't leak, exclude any foreign material in the system, and allow you to take it apart for changes or repairs. You need two circuits (hot and cold for arguments sake). The cold circuit has to be manifolded to the hottest part of all the coils (close to the lamp), and the output of all the coils should be manifolded together as well. These now common-rails should be connected to the heat pump.

Even simpler still: get a water pipe tap for a refrigerator water on the door hook-up. These use 1/4 inch polyethylene tubing (food grade) cheap at Homo-Depot. Hook this to your coils, and the other end to the drain. Small a/c or d/c solenoid valves in the inlet side open when the light(s) are turned on. No cooling problems there- city water runs around 60 degrees F...
 
I typed a nice response with tons of ideas for you.

Then the worthless server that hosts this board dropped the connection (again).

Send me an e-mail at jim[dot]leach[at]sulzer[dot]com for information on your system.

What you have in mind won't work the way you think it will, and a vacuum pump won't do what you want it to do either (but is not needed at all anyway).

Speak with you soon...:wave:
 
Your steady-state 134 design will not remove any heat.

right, within the closed system of the pipes themselves the heat will not leave magically via the refrigerant. it has to GO somewhere.
with a heat pipe system, all i have to do is move the heat from the small location, and have it get all over the whole system, from there passive cooling will exist over the larger ammount of copper.

so all i have to do is move the heat from the 1inch square , to the 14 inches square of the whole rest of the surface, then passive air cooling will be enough.

i Have and could use CPU heat sinks, but they are clunkey, not artsy like a metal sculpture, a FAN is not an option, i hate fans, they are the total answer to cooling, as passive cooling without rapid air removing takes WAY more space, and weight of sink, but the last thing i want is more fans going off in this place, much work has been done to reduce the noise of computer fans, and i have huge passive GPU sinc, and that didnt even work without a fan.
and no active pumps going on, i dont want a light to sound like a fridgerator, if freon type system, wont work without a compressor, then it wont work at all.

somehow with the right liquid behind the copper there is a magic movement, of the heat, within the system.
even if that movement is vaporising water, which recondences all over the rest of the area of the pipes, this works without capilary action, as long as the heat source gets next to the liquid.

without capilary wicking, that does become a bad idea, if somone trys to operate it upside down :-( hmm didnt think about that.
capilary liquid transfers dont HAVE TO be inside small pipes, i can use a rolled up paper towel, and have water capilary quite a ways, not that i would PUT a paper towel in it, just that Big fat tubes, could have capilary items in it to transfer liquids.

which brings up the other 2 possibilities, which i dont like as much.
use a ALL copper sheet, and formulate some sort of art from it, without any heat piping. and that seemed to Lame , to be testing anything neeto.
or
formulate the copper sheet into box frames, where the air internally will "chimney" up through the openings, and heat will move via convection, and be increased in speed using the chimney effect.
and that is when i thought about using liquids, and just regular available water pipes, as the water pipes are already formed, and could be assembled like tinkertoys (about my speed)
 
Last edited:
Well, if you want to try air convection via square aluminum tubes, Homo-Depot has those too. (you never indicated the watts of heat you need to reject from each light)

I made a stereo rack out of them, lag-bolted to 3/4 inch white oak boards. Fairly thin wall, so rather light for thir size. Probably 2024 or some other low-grade alloy...

If those are kept in a vertical (or near vertical) position, the combination of conduction and convection will move the heat.

BUT, you have to keep one thing in mind with these type of systems: If the media gets overheated (aluminum square tubing rises to near emitter temp) they will decrease in effectiveness.

I suggest a 1" round bar of copper with the emitter attached intimately to that (silver paste, or whatever, to get heat into that copper) and slot the back side 1" deep, 3/4 inch wide (assuming using the above 3/4 inch aluminum tubing) and attach it with a copper or aluminum bolt. This way, the tube remains vertical, but the light can be moved up and down. Make a swiveling base for the square tube, and it can rotate as needed. (apply silver paste at joints to keep conductivity high)

In fact, make the tubes about 6 feet long, and have the bottoms close to the floor (but they have to be open to let air in). The tops will be about your head height. Mount the lights somewhere in the middle; perhaps at various heights to add a bit of interest. Combine the flexible 'up and down' movement of the individual lights and it should be a cool looking light.

Wrap the bottom area with expanded aluminum (also at Homo-Depot) to prevent pets and children from touching the aluminum tubes. I think after a while the entire tube will be warm, perhaps hot to the touch, well below the emitter (and hotter above)...

That should look cool, and work fine for relatively short periods. Only empirical testing (actually building and running it) will tell if it will be a viable solution...
 
there is 8 or 9 items, with about 5 watts of waste heat EACH item to remove.

so your saying use copper rod, as the inital heat sinc, and "insert" that into the square aluminum? kind of a head at the end, inside the chimney of square aluminum?
 
here is the problem I (myself) have with big heatsinc BLOCKS, they seem to HOLD heat really well. like a big clunk of copper, moves the heat throughout the copper alright, but the whole big hunk is then heated pretty well.

With the LED, the only "need" is to get the peak heat from the base to Anywhere else :) , that is where i thought that a "liquid" that could somehow move, is better than a big clunk that is solid and doesnt move.

so although copper might transfer heat faster through it, the copper is stuck in place (solid), and becomes a big ball of heat, wheras a liquid is not stuck, IF it will move away when heated.

so with thin walled copper, and a moving transfer behind it, the first copper item (the one where the led is) stays cooler than with a big clunk of hot copper. That is the assumption anyways:)
 
Last edited:
For liquid (particularly in copper tubing) I suggest de-ionized and/or distilled water. Available at the grocery store for a dollar a gallon or so. With the other minerals out of it, it will not react with the copper for over 10 years. Do remember that the solder has many elements in it that may react with the 'clean' water to form other compounds, so I suggest silver braze for all connections, if you must braze, but there is another option to brazing I will cover below...

i dont have hot enough equiptment for brazing, just home propane torch, last cheap brazing kit with solidox i had was a pain, and it is dead and in the trash, and the good stuff is to much money for to little use.

i was planning on just lead acid type soldering?? with the parts interfitting.
can i silversolder with a cheap propane torch? and then there goes the FLUX stuff getting into the parts again.
plus if i braze the thing, there is going to be more cleanup and polishing, and real work :) to have to do, assuming i dont put holes in things from bad work.
 
The idea behind a heat pipe has nothing to do with convection and everything to do with phase change. It works because on the warm end, the fluid evaporates which absorbs enormous amounts of heat (heat of vaporization) from your heat source. This is the same as your body sweating to cool off. On the cold end, the that heat is dumped away from your LED/CPU/etc... when the vapor condenses (heat of condensation). Ever notice how quickly your beer gets warm in humid environments? This use of phase change transports MUCH MORE heat than if the pipes were filled with fluid and relied on convection.

For this reason, oil will be useless since it won't readily evaporate at these temps and refrigerant will be also worthless since it won't readily condense unless you have your pipes pressurized which is making it overly complex for no good reason.

For the temps we are dealing with here, ammonia is sometimes used but not a good idea here because of corrosion issues with copper. Ethanol or water are good candidates and commonly used.

No need to completely evacuate the lines, again this is just making it more complex than it needs to be. Just solder one end of the pipe to seal it off, and put in your fluid (I would just go with de-ionized water with a little anti-freeze) on the other end. After the fluid is in, sweat solder most of the other end, heat up the heat pipe until you get a little vapor coming out of the unsoldered gap, then solder it completely to seal it. Since most of the gas expanded and left the pipe when you heated it, the pipe will still be under a partial vacuum under room and operational temperature. That's all the evacuation you need. At 0 psi, the boiling point of water will actually be too low to be effective in a heat pipe for our temp range.
 
Last edited:
The idea behind a heat pipe has nothing to do with convection and everything to do with phase change. It works because on the warm end, the fluid evaporates which absorbs enormous amounts of heat (heat of vaporization) from your heat source. This is the same as your body sweating to cool off. On the cold end, the that heat is dumped away from your LED/CPU/etc... when the vapor condenses (heat of condensation). Ever notice how quickly your beer gets warm in humid environments? This use of phase change transports MUCH MORE heat than if the pipes were filled with fluid and relied on convection.

For this reason, oil will be useless since it won't readily evaporate at these temps and refrigerant will be also worthless since it won't readily condense unless you have your pipes pressurized which is making it overly complex for no good reason.

For the temps we are dealing with here, ammonia is sometimes used but not a good idea here because of corrosion issues with copper. Ethanol or water are good candidates and commonly used.

No need to completely evacuate the lines, again this is just making it more complex than it needs to be. Just solder one end of the pipe to seal it off, and put in your fluid (I would just go with de-ionized water with a little anti-freeze) on the other end. After the fluid is in, sweat solder most of the other end, heat up the heat pipe until you get a little vapor coming out of the unsoldered gap, then solder it completely to seal it. Since most of the gas expanded and left the pipe when you heated it, the pipe will still be under a partial vacuum under room and operational temperature. That's all the evacuation you need. At 0 psi, the boiling point of water will actually be too low to be effective in a heat pipe for our temp range.

Wikipedia has a very nice graph of water's boiling point v. ambient pressure.

Water_vapor_pressure_graph.jpg
 
Something that you might consider is to solder together the whole thing empty. At one end of your top pipe you have a threaded fittting which is uncapped. You put in your working fluid and heat the entire assembly. The fluid expands, vapor starts to come out the threaded opening. You apply pipe dope or teflon tape on the threads and cap it mechanically. Screw it down good. As the unit cools the pressure inside drops and your working fluid hopefully will evaporate easily. I was thinking alcohol might be a candidate for your working fluid. The problem with my setup would be to not set it on fire when you heat the assembly. That would be bad. I don't know how you arrange for each arm to have some fluid and none of them to be dry.

As a side note, while you are soldering you could solder copper wires to the pipe. Fan them out artistically and each wire now becomes a lesser radiator for the main assembly.
 
Top