Making heat pipes.

Hi guys,

I'm interested in experimenting with home made heatpipes. I used the search but didn't turn up much.

At the moment I'm considering soldered copper pipe construction with a valve that will allow filling with water, alcohol or propane.

My main problem at the moment is finding a suitable wick. It needs to be something that will withstand soldering temperatures without deterioration.

I was initially considering using this fine aluminum mesh I had, I thought that it would wick fairly well when rolled up tightly several layers thick inside the copper ppe. However, this would not now appear to be the case after some experimentation.

Another desirable quality of the wick of course is that it should have enough rigidity or spring or whatever to stay in place inside the tube. I had considered using a cotton based fabric, and soaking it in water to protect it from soldering damage, but thought that it may well "wilt" inside the tube and block it.

All I can think of at the moment is buying some glass fiber roving anfd riolling that up inside the tube.

I had considered fine copper wire cable shielding, but that is difficult to get the right size, and the right thickness, and might have the same wicking deficiencies that the aluminum mesh did.

So does anyone have any experience or ideas to share?

Thanks for any help,

Road Warrior

Fiberglass will work. It can withstand pretty high temperatures without melting, and if properly woven has good capilarity properties.

Keep us posted.

Doh, I just thought, I could use a single layer roll of the aluminum mesh to keep anything else in place.

When you start getting down to it, PICS WE WANT PICS!!!

It would rock to have any heatpipe you want.
There are places where heatpipes could be used, but it's hard or amlost impossible to find proper ones.
We could use heatpipes for waterblocks or maybe even phase change parts.

Well, a heat pipe is a phase change device.

I don't think a wick is an absolute necessity. Check out the home-made heatpipes over at benchtest.

http://www.benchtest.com/heat_pipe1.html

Well, this particular one works because it's basically standing up and the liquid falls back to the hot end by gravity. If you want it to work in any other position, you need a wick. Or a pump, something.

stamasd said:

Well, a heat pipe is a phase change device.

Click to expand...

Yeah, I'm fully aware , But I meant to use heatpipe as some helper for Active PhaseChange rigs, for example to increase size of CPU block.

btw, I'd still try to get a wick. Maybe some vibrating pump would be nice, but for heatpipes I wouldn't use normal pump as it becomes to be normal phase change cycle.

Doesn't heatpipes require a bit more sophisticated liquid than water or alcohol ?

Hmm... Maybe propane under the exact right pressure could work.
One needs to find the sweetspot pressure when propane juust liquifies and can do it's gas and liquid magic....

I would chose a liquid with a boiling point just above ambient. I.E. something that will boil at 40C or something, that way I wouldn't have to deal with all this pressure. Then it's more about finding the correct amount.

I'm not an expert in heatpipes, but, doesn't it take two different densities of the wick?

One outer layer that will transport the gas and another that will "wick" back the fluid (in the middle of heatpipe)?

Obviously there's a need to have right fluid.
But with that alcohol - if I'm not wrong, isoprophyl alcohol (I don't know name in english) is used in fridges and could work in heatpipe.

And pressure & fluid stuff - there's a need to find that "sweet spot", and exactly one for conditions under which it will work. A problem might be making heatpipes working at temparatures <ambient, as when turning system off, it would need to resist pretty high pressure.
"Just over ambient" - I'd rather name it "just below heatsink temperature".

Wick - one wick can be just free space, when just installing other

Nope, trust me - you don't want isopropyl alcohol in there. I can't imagine what it would do in a fridge. Isopropyl and all of the higher alcohols have high boiling points and low vapor pressures, exactly the opposite of what you want. Isopropyl alcohol takes a LONG time to evaporate, and thus the phase change effect is too weak. If you don't believe, do the following: rub ethyl or methyl alcohol on one of your hands, and isopropyl on the other hand. Which one feels cooler?

Ethyl alcohol is marginal too, as its boiling point is a bit high as well (78 degrees C), but it has a higher vapor pressure so it could ultimately work somewhat. Much better is methyl alcohol (BP=64 degrees), acetone (BP=46 degrees) or pentane (BP=36 degrees).

A warning, all of these are volatile and when the vapors are mixed with air they're largely explosive, so watch the sparks when filling your tubes. If you want a liquid that's not explosive, try chloroform (BP=61) or methylene chloride (BP=39), but these are not good for your health. They're very strong solvents as well and chew through plastics like you wouldn't believe it.

Water will work if it's at low pressure so it boils more readily.

Anyone seen the science experiment where you take a thin can, part fill it with water, let it boil off until the can only contains a little water and pure steam, then seal the can?.... then as it cools the can is sucked flat (Actually crushed flat by atmospheric pressure) . Well for my water experiment I'm planning a variation of that. I will calculate (probably empirically) the amount of water I need in the pipe and rig a balance that will balance the heatpipe with the right amount of water in it. Then I will fill the heatpipe with a larger amount of water and apply a blowtorch to getit boiling vigourously replacing all the air inside with steam, then when enough has escaped the balance will tip and I will quickly seal the heatpipe. Then as it cools, hopefully, there will be very little if any air and some of the water will be forced to remain in vapour phase. The liquid water remaining should be under low enough pressure that it will now boil at a temperature far below normal boiling point. Around 30C would be great.

My heatpipe could of course be crushed flat by atmospheric pressure instead, in which case I'll have to experiment with thicker walled tubing, or with letting it cool a little before sealing.

If water doesn't work very well then I'll try alcohol. After that I might try butane or propane, but am thinking that you need positive pressures with those. Water is most desirable because it has a higher specific heat capacity and is less dangerous should leaks occur.

My thinking about application is that I might initially use forced air cooling on a regular heatsink clamped to the hot end, with the other end on a CPU. My feeling is that you might be able to treat it rather like a peltier in that you might be able to run with a higher heatsink temperature, yet still have adequate cooling at the cold side. However I guess as the temperature of the system increases you would get the pressure increasing and the boiling point rising. Also using huge sinks found in amplifiers and such, passive cooiling might be viable.

I had thought also that a heatpipe would offer a diversion in water cooling, that might be more efficient and allow high flow velocities. Because instead of water blocks that impede flow you could have relatively greater flow through cooling coils wound around heatpipes protruding through the case and have a much shorter and smoother circuit, and the dangerous end of the water would be well away from your expensive bits. Actually with a less restrictive flow and greater control over the layout of your system, then pumpless watercooling could be possible!

anyhoo, thanks mota for that benchtest link, very interesting.

Road Warrior

To get water boiling at 30 degrees you need to get the pressure down to 30 torr (that's 1/25 atmosphere). Can you get a seal good enough? I mean in the long run.

Well, you'd notice a slow drop off in performance and do something about it. As far as sudden catastrophic failure goes, I think you'd probably have the same risks as a pump going out while watercooling, the copper tube itself should allow enough conduction for a CPU to avoid frying instantly. So if you implemented thermal shutdown in software, or used the CMOS settings for that, I think you'd be safe enough. The valve I will be using will be a fine threaded, large head, brass screw tapped into the end of a copper pipe cap. I might silver solder a bit of plate on the end of the cap so there's more for the thread to bite, and it will have a neoprene washer. I may well also tin the threads with solder, so that they actually seal when force tightened instead of relying wholly on the washer. The ends will be silver soldered pipe caps. 30 Torr to me doesn't seem all that high of a vacuum to maintain, if we were talking another power of 10 I'd be more concerned, though there is of course a concern that the copper pipe could exhibit a degree of porosity.

Hmmm thinking though, it's fairly easy to make a system tight to air at a pressure of 2-3 atmospheres, and we are talking a pressure differential of max 24/25 of an atmosphere.....

I prefer empirical experimentation over theory, because since I don't have any precision instruments or equipment, I could design this thing theoretically perfectly, but managing to attain that design in the flesh would be highly improbable without more sophisticated tools. However, it is probably about time I ran some figures, just to see which ballpark I want to play in.

Something concerning me at the moment however, is that the coolant charge that might be most theoretically efficient could run the risk of running dry on a high load, allowing temps to suddenly rocket before settling again, on a CPU this would be really bad. Therefore a less efficient amount of coolant charge, such that the temperatures might on average be a degree or two worse, but that it will not boil dry might be more in line with what is desirable for a CPU cooling situation. Have to think about that though.

The ends of the tube don't have to be pipe ends you know... you can have a little tank to store extra coolant for temp spikes. It'll make the construction more complicated but probably worth it.

Also, neoprene+propane=bad. The propane will destroy any neoprene rubber after a while. Keep the rubber out if you're planning to go for organic solvants.

What I would probably try is the following. Are you familiar with the solder paste that Radio Shack sells? It's called Solder-Weld, and it contains fine particles of a tin-silver alloy with a fairly low melting point. You apply it to the place where you want the soldering to take place, then you heat with an iron, lighter or other source of heat (I use a butane mini-torch) and it melts in place. I would use that to solder the threads in place for sealing purposes.

Link to Solder-Weld

Instructions

mmmm - water heatpipe would be cool, but it would be hard to get proper pressure.

hmmm. i wonder if you can make these with insulated, braided copper tubing. make a flexible heatpipe. This thread is great reading guys, thanks. I have a similar project for my shuttle pc project later on.

Flexible heatpipe is a challange. Water heatpipe would have to work on pressure far less than ambient. It's relatively easy to keep things under high pressure, but low - it's definietly hard when you want to have thing flexible. It would need to be similiar to tubing for these "pulling pumps".

as for a wick system, for use with water, would it not work to saturate your wick material with fluid, and then freeze it before integrating it to your pipe and sucking out the air? this way, you can work with your working fluid as a solid while its within the wick. and if the wick is fully saturated, is this not enough working fluid? or do you want enough to keep the wick saturated and still have some left over?