How do heatpipes work? Copper pipe filled with something, right? What are they filled with? Would water work? Are they pure copper or finished with something?
Heatpipes
- Thread starter Pudgygiant
- Start date
To the best of my knowledge heatpipes rely on a phase change to increase heat transport. So liquid evaporates at the hot end, travels as a gas to the cool end where it condences and flow back to the cool side as a liquid to repeat the process. Since considerable enegry is required for the phase change this should/could make for a very efficent heat transfer device.
Does it work? If the temperature is such that the gas never condenses then no, it will not work. The key would have to be getting the gas to the cool region and quarenteeing that it remains cool there.
Does it work? If the temperature is such that the gas never condenses then no, it will not work. The key would have to be getting the gas to the cool region and quarenteeing that it remains cool there.
So then one would have to figure out the approximate temps to decide which solution to put inside? And it has to travel UP, right? I.E. the cool side can't be on the vid card, with the hot side being at a point at the bottom of the case.
Yeah, the liquid would have to flow down to the hot area, I am not sure how to make it work with a video card where processor in mounted side ways???
From some of the stuff I have read it was not clear that they were actually able to transport as much heat with the gas/liquid phase change as simply having a mass of metal with a good temp gradient.
I think that the future is with water cooling systems. Not the current HS/fan combos, so if you want to design a future proof cooler think water.
From some of the stuff I have read it was not clear that they were actually able to transport as much heat with the gas/liquid phase change as simply having a mass of metal with a good temp gradient.
I think that the future is with water cooling systems. Not the current HS/fan combos, so if you want to design a future proof cooler think water.
Har de har har, just found this out researching this. The origin of calling diamonds "ice" is that they have an incredibly high thermal conductivity so they almost always feel cold to the touch.
OK, so this is the idea I had, this may be way of the mark, but... will a thick-guage, solid (as opposed to stranded), insulated pure copper wire conduct heat fairly well? If it would, you could put the heatsink completely somewhere else. I'm gonna have to try it (with an old crap p2 or something.)
The whole point of water cooling is to move the water! Moveing water can carry a lot of heat, much quieter then moveing air. At least a pump could be located at a distance.
Yes copper is a very good conductor of heat. I once saw pics of a porcpine HS made up of bristling copper wire. Not sure why you would need a HS if you had that much Cu wire it would be hard to transmit heat the entire length, ie the wire would BE your HS.
That sounds like a bit of an ambiguous statement. Is the thermal conductivity of heat pipes, as used in CPU cooling, really three orders of magnitude better than a copper heatsink? If so, why are heatpipe systems so big compared to traditional coolers?
cheers,
Andy
I believe that if you took a block of copper and a heatpipe, and put one side in boiling water, it would be a long time before you coulnd't hold the other side of the block... but the heatpipe would burn you almost instantly.
The heatpipe would move heat that much better - the copper block might need to reach a temperature difference of, say, 20C (arbitrary number) to move 100 watts through it, while the hot and cold sides of the heatpipe would be almost exactly the same temperature. (Sorry, I can't find sources for my numbers).
Over all, you won't get 1000x better temperatures using a heatpipe - you still have to put the heat somewhere once you've taken it away from the CPU.
Well, for one thing, if you had a foot-long copper heatsink, you wouldn't gain anything, because it will get only a negligible amount of cooling from the far away parts. With a heatpipe, you'd still have a lot of heat reaching the ends. Since with a heatpipe you can get benefits from a really large heatsink, you could use a bigger cooler and a slower fan (?).
You can make a horizontal heat pipe fairly easily.
If you don't know what you're doing, don't try this. You can get burned, both by heat and chemicals, as well as being exposed to bad fumes from soldering and stuff.
First of all, I would recommend using fairly thin tubing, probably copper since it is easy to work with and can be soldered easily. Once you get the casing (1/2" copper pipe and 2 pennies work well for a simple experiment), get some very fine mesh metallic screen that you can roll up into 2 or so layers and extends the length of the copper tube. This mesh allows wicking/capillary action to move the cooled liquid from the cool end back to the heated end where it can evaporate again. Next, find a way to solder a small tube on one end so that you can put your phase change medium in the pipe once it is sealed and draw a vacuum on the whole pipe. When I did this, it was a small diameter brass tube about 1" long that was soldered into a hole drilled into one of the pennies. Anyway, using a small length of rigid rubber tubing and a large syringe (without the needle part), put your phase change medium in the pipe. The liquid used inside the pipe should be chosen based on the operating temperatures so that you will see a phase change. No phase change = crappy heat transfer, far worse than solid copper. I believe I used some solvent like acetone, though you could probably even get away with something like alcohol or maybe even water if you are operating at high enough temperatures (whatever it is, it should be non-corrosive to anything inside the pipe). The amount of liquid put inside should not be so much that you have the heat pipe even half full. Before you solder on the last cap, perhaps you can hold the pipe horizontal and partially plug one end so that you can fill it with water to see how much volume it takes to get the desired level inside the pipe. It will probably take some trial and error, but you should aim for just enough liquid to keep the screen inside wet, for a 1/2" pipe, maybe only 1/16" of liquid in the bottom. Like I said, trial and error, or you can do more research if you'd like. Once you have the liquid inside, you need to get as much air out as possible so that you only have liquid and vapor inside. Do this by either using the syringe and drawing air out, then sealing the tube somehow or with a vacuum pump if you have one. Once the end is sealed, you should check for leaks and performance before using this on any electronics, not to mention somehow figuring a way to attach it to whatever you are trying to cool.
Some notes:
A rectangular pipe, while harder to find and construct, would be easier to attach to a processor or similar heat source.
If your heat pipe fails during operation, it will have very poor heat conductivity and if on a processor, you will probably fry it.
Elevating one end of this type of heat pipe too much results in poor performance.
Yes, this heat pipe does work, yes I've tried it, no I personally wouldn't make one for my computer.
Some other answers based on questions posed already:
Using copper wires to transfer heat to a remote heat sink is a bad idea and, even with a large gauge wire, wouldn't function well. Conductive heat transfer is based on area, material, and distance (thermal resistance proportional to HL/A) where H is a material property, L is the distance and A is the area. Wire lacks the area of a heat sink base and has a long distance to travel compared to the distance between standard heat sink and fin surface.
Lots of people want to make a huge number of thin, long fins to dissipate more heat. In conductive heat transfer from a fin, there is a maximum effective length based on cross sectional area, geometry, material, and surface area. The amount of energy flowing through the fin from the heat source is limited by the fin cross sectional area (and material type), while the energy flowing out of the fin via convection is related to the surface area and ambient air conditions. It is very easy to construct a long thin fin that can dissipate much more energy than it can conduct through it.
As you increase distance between the heat source and the dissipation surface, you increase the thermal resistance and the heat source will be hotter for a given amount of energy dissipation.
If you don't know what you're doing, don't try this. You can get burned, both by heat and chemicals, as well as being exposed to bad fumes from soldering and stuff.
First of all, I would recommend using fairly thin tubing, probably copper since it is easy to work with and can be soldered easily. Once you get the casing (1/2" copper pipe and 2 pennies work well for a simple experiment), get some very fine mesh metallic screen that you can roll up into 2 or so layers and extends the length of the copper tube. This mesh allows wicking/capillary action to move the cooled liquid from the cool end back to the heated end where it can evaporate again. Next, find a way to solder a small tube on one end so that you can put your phase change medium in the pipe once it is sealed and draw a vacuum on the whole pipe. When I did this, it was a small diameter brass tube about 1" long that was soldered into a hole drilled into one of the pennies. Anyway, using a small length of rigid rubber tubing and a large syringe (without the needle part), put your phase change medium in the pipe. The liquid used inside the pipe should be chosen based on the operating temperatures so that you will see a phase change. No phase change = crappy heat transfer, far worse than solid copper. I believe I used some solvent like acetone, though you could probably even get away with something like alcohol or maybe even water if you are operating at high enough temperatures (whatever it is, it should be non-corrosive to anything inside the pipe). The amount of liquid put inside should not be so much that you have the heat pipe even half full. Before you solder on the last cap, perhaps you can hold the pipe horizontal and partially plug one end so that you can fill it with water to see how much volume it takes to get the desired level inside the pipe. It will probably take some trial and error, but you should aim for just enough liquid to keep the screen inside wet, for a 1/2" pipe, maybe only 1/16" of liquid in the bottom. Like I said, trial and error, or you can do more research if you'd like. Once you have the liquid inside, you need to get as much air out as possible so that you only have liquid and vapor inside. Do this by either using the syringe and drawing air out, then sealing the tube somehow or with a vacuum pump if you have one. Once the end is sealed, you should check for leaks and performance before using this on any electronics, not to mention somehow figuring a way to attach it to whatever you are trying to cool.
Some notes:
A rectangular pipe, while harder to find and construct, would be easier to attach to a processor or similar heat source.
If your heat pipe fails during operation, it will have very poor heat conductivity and if on a processor, you will probably fry it.
Elevating one end of this type of heat pipe too much results in poor performance.
Yes, this heat pipe does work, yes I've tried it, no I personally wouldn't make one for my computer.
Some other answers based on questions posed already:
Using copper wires to transfer heat to a remote heat sink is a bad idea and, even with a large gauge wire, wouldn't function well. Conductive heat transfer is based on area, material, and distance (thermal resistance proportional to HL/A) where H is a material property, L is the distance and A is the area. Wire lacks the area of a heat sink base and has a long distance to travel compared to the distance between standard heat sink and fin surface.
Lots of people want to make a huge number of thin, long fins to dissipate more heat. In conductive heat transfer from a fin, there is a maximum effective length based on cross sectional area, geometry, material, and surface area. The amount of energy flowing through the fin from the heat source is limited by the fin cross sectional area (and material type), while the energy flowing out of the fin via convection is related to the surface area and ambient air conditions. It is very easy to construct a long thin fin that can dissipate much more energy than it can conduct through it.
As you increase distance between the heat source and the dissipation surface, you increase the thermal resistance and the heat source will be hotter for a given amount of energy dissipation.
Wow zeronine8, very in depth, nicely put. I'm gonna have to experiment with this, once I get some coinski. Would a liquid that would change phase at a very LOW temperature (say, 70 degrees instead of 100) work? If so, why do they not just fill them all with that?
And would a CPU block filled with the aforementioned solution work? At least aswell as a water block?
And would a CPU block filled with the aforementioned solution work? At least aswell as a water block?
Say you have a processor that you want to run at 70C, then I'd find a relatively harmless liquid that evaporates at 60-65C. The key is that you have to be sure you can keep the opposite end of the heat pipe below the boiling point (like maybe ~50C) so that you ensure phase change back to liquid. Heat pipes really only move heat from one end to the other, you still need some sort of heat sink to dissipate it to the atmosphere (or water cooling or anything). I'm not that familiar with water cooling apparatus, however if you are suggesting that you want to use a standard water cooling setup with a phase-change liquid instead of water, it won't work (at best you'd probably end up on par with water cooled results for a lot of extra hassle). The two methods of heat transfer are completely different. I'd recommend google to find some heat pipe theory, you should find some sites that discuss the energy transfer via phase change as well as some basic design principles. Most of the stuff I have is in textbook or research experience form, not really online
.098,
Execellent posts! Thanks for the specific info. From your details it appears that my general remarks were not to far off the mark.
I might add that HEAT OF FUSION is a critical parameter of your fluid, this is the amount of energy per gram of liquid required to drive the phase change. A large Heat of Fusion will mean you need to transport less liquid for a given amout of energy transfer. So if you are able to find several liquids which meet your temperature and non-corrosive requirements, choose the one with the higher Heat of Fusion. This number should be available for most common liquids. Goggle is your friend.
Execellent posts! Thanks for the specific info. From your details it appears that my general remarks were not to far off the mark.
I might add that HEAT OF FUSION is a critical parameter of your fluid, this is the amount of energy per gram of liquid required to drive the phase change. A large Heat of Fusion will mean you need to transport less liquid for a given amout of energy transfer. So if you are able to find several liquids which meet your temperature and non-corrosive requirements, choose the one with the higher Heat of Fusion. This number should be available for most common liquids. Goggle is your friend.
Soulkeeper
Diamond Member
- Nov 23, 2001
- 6,736
- 156
- 106
i got an sp-94 it works pretty good
depending on the orientation of it (verticle or horizontal) you get like 1 to 2c diff in temp
i personally don't think heatsinks get hot enough to truly use heatpipes to their full potential
just making the heatsink bigger or with more fins seems to make a larger difference
but i like the fact that new ideas are making their way into computer cooling
depending on the orientation of it (verticle or horizontal) you get like 1 to 2c diff in temp
i personally don't think heatsinks get hot enough to truly use heatpipes to their full potential
just making the heatsink bigger or with more fins seems to make a larger difference
but i like the fact that new ideas are making their way into computer cooling
The problem, I think, is that the entire system will stabilize at the boiling point of the coolant (assuming you have adequate cooling; if you don't, it'll stabilize at some higher temperature). If you use a coolant with a high boiling point, the heatsinks will get hot (and thereby more efficent), but the chip will be running hotter as well. A coolant with a low boiling point gives low chip temps, but low heatsink temps as well (and so you'll need a larger heatsink to dissipate the same amount of heat). At least I'm pretty sure. It's been a while since I did any thermodynamics problem sets.
Something that no one else mentioned. Heat pipes rely on low pressure. Water boils at lower temperatures at lower pressures. I read where heatpipes have been sealed with lower than sealevel air pressure inside. This makes the phase change take place at lower temperatures. You don't want to have to heat water to 212 degrees Fahrenheit to make it boil.
The heat pipe must also have some vertical displacement. It supposedly doesn't take much, just so long as the pipe is inclined with hot end down below the horizontal.
Yes, heat transported in moving water moves faster than the heat radiated thru a solid metal object.
I found a heatpipe from a Dell note book. the low end was flat for contact with a chip and the othe end had a wrap around aluminium heatsink on it. I think the heatsink was placed so a cooling fan would blow on it.
The heat pipe must also have some vertical displacement. It supposedly doesn't take much, just so long as the pipe is inclined with hot end down below the horizontal.
Yes, heat transported in moving water moves faster than the heat radiated thru a solid metal object.
I found a heatpipe from a Dell note book. the low end was flat for contact with a chip and the othe end had a wrap around aluminium heatsink on it. I think the heatsink was placed so a cooling fan would blow on it.
No, heatpipes don't neccesarily need low pressure. Its merely that theres no economical, safe liquid that boils at the right temperature at amospheric. In fact, heat pipes would work better under high pressure since theres more mass to move around.
One of the failings is that unlike generic heatsinks, if your big heatsink doesn't work for some reason, then temperatures can shoot through the roof. Not only that, you get absolutely no warning since it goes from cool to fry your CPU almost instantly.
One of the failings is that unlike generic heatsinks, if your big heatsink doesn't work for some reason, then temperatures can shoot through the roof. Not only that, you get absolutely no warning since it goes from cool to fry your CPU almost instantly.
I did infer that the heat pipes with low pressure would be water filled. I did find one interesting DYI article that used R134.
Homemade Heatpipes
Homemade Heatpipes
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 407
-
Facebook
Twitter