I'm a little confused about Heat-Pipes, school me pls

[nzimmers]

alright, I *think* I have a good idea how a heat pipe works.....

A) The thermal source on the bottom (cpu) heats up the fluid inside the tube, whatever that may be.

B) fluid changes to a gaseous state and travels up to the top end of the tube

C) Top end of the tube is colder, thus condensing the fluid, which drips down to the reservoir at the bottom.

not the most accurate description, but in a nutshell, that should be the basic concept. I feel pretty confident that gravity plays a role in this process, and a heat pipe (or passive cooling) would not function in the absence of gravity.

BUT.... here we have this ....

image

Now.....help me here, are the heat pipes placed horizontally actually working the way they are meant to? how is the re-condensted fluid gong to find its way back to the thermal source? what's to stop it from pooling mostly on the cooler end and allow heat to build up on the chip?

I'm guessing that the heat pipe in the picuture above isn't working properly, and that the thermal conductivity of the copper is doing most of the work here

Also, just wondering, how do you guys thing a heat pipe would perform compared to a solid rod of copper the same diameter of the heat pipe?
And, how do you think a heat pipe made of glass would perform (same diamter as the heat pipe in the picture......

As you can tell, I am not totally sold on the application of heat-pipes in the horizontal, or in in the application of personal compters..... just because ones CPU isn't overheating doesn't necessarily mean the heat pipe in there is actually performing the way it should, could very well be a good old case of efficient conduction

someone please set me straight

[frostedflakes]

Based on what I've read, gravity heat pipes are used rarely in computer cooling applications because of the disadvantage listed in your post -- their effectiveness is greatly decreased when mounted sideways or upside down.

The heatpipes used in most cooling products contain a wick inside that pulls the condensed liquid back to the heat source regardless of orientation. I couldn't explain all the physics behind it, but if you want more information search capillary force on Wikipedia.

Also I do not believe the transfer of heat through the gas relies much on gravity either. From what I understand, the gas will move from an area of higher pressure (i.e. higher temperature) to lower pressure (i.e. lower temperature), so again effectiveness should not be significantly effected by orientation.

Just my basic understanding, though, I may not be doing the explanation much justice. Definitely check out the heatpipe article at Wikipedia for more detailed information. They might also have something on heatpipes at HowStuffWorks.com.

[nzimmers]

yes, I see, capillary action would definitely make a difference, although capillary action works pretty much without regard to the orientation of gravity, and in that case should work upside down shouldn't it?

[frostedflakes]

I believe so. Effectiveness may be reduced slightly without gravity to assist, but I don't think the difference in performance between a wicked heatpipe upside down and right side up is significant. Same goes for it in a sideways position.

[nzimmers]

I am curious about one thing, it would be nice to make a comparison between a heat pipe and a solid copper rod (both being the same diameter) and see just how much more a efficient a heat pipe is.........


any thoughts?

[Rusty075]

Frosted is exactly right. Unwicked heatsinks are basically never used in computer parts now. The wicking material is a sort of silicon powder, and is sprayed on the interior of the pipe. One thing to understand is that there's really very little liquid inside these pipes. If you crack one open nothing is going to come spilling out. Even in heatpipes as large and long as those in the picture you are talking about tenth's or hundredth's of an ounce of actual fluid inside. The rest of the interior is filled with inert gas at less than atmospheric pressure. (a partial vacuum lowers the vapor point of the fluid)



A heatpipe is up to a couple of thousand times more efficient at conducting heat than a solid copper rod would be. For example, Thermacore, one of the big manufactures of heatpipes advertises a typical thermal conductance of 50000W/mK for their run-of-the-mill pipes. A copper rod would have a conductance of 401w/mK.

[albatros_la]

There are a lot of different types of heat pipes out there. The materials used to build them, the fluid type and the wick type play a big role in the heat transfer capability and they're varied with respect to the work temperature range. I've worked on heat pipes for a certain period and I can assure that the ones used in electronics are mainly (metal) mesh wick heat pipes. The sintered powder wicks are not so popular because of their high cost and the fact that this technology nowadays lacks in terms of reliability. An heat pipe without mesh in it it's simply called a thermosyphon and could work well only if the evaporator (let's say: the heater extremity) is placed downwards while the condenser (... the "cold" extremity) is placed upwards. This means that heat pipes have been invented in order to transfer heat from a point to a lower point. This principle works amazingly in space, where there's not gravity, and really less well here on earth. In normal gravity conditions a short heat pipe (like the ones used in electronics) work slightly well because on a short track the capillary pump effect which carries the cold fluid from the condenser to the evaporator through the wick is stronger than the gravity effect which tends to keep it down. Longer heat pipes (more than one feet long) works really bad even if they are slightly tilted. So gravity helps us only if we make heat pipes work like thermosyphons (but in this case, there's no reason to spent money buying the first ones instead of the second ones...).

[jaganath]

This principle works amazingly in space, where there's no gravity

There is gravity in space, just not as much of it.

[PPGMD]

There is gravity in space, just not as much of it.

Gravity is relative, but black holes are absolute.

[spookmineer]

A heatpipe is up to a couple of thousand times more efficient at conducting heat than a solid copper rod would be. For example, Thermacore, one of the big manufactures of heatpipes advertises a typical thermal conductance of 50000W/mK for their run-of-the-mill pipes. A copper rod would have a conductance of 401w/mK.

A thousand times... I'm not sure what to believe, How do heatpipes work? say:

As most heat pipes are constructed of copper, an overheated heatpipe will generally continue to conduct heat at only around 1/80th of their original conductivity.

I don't know how they concluded this.
Either way, in both cases a defective heatpipe will lose its function completely.

[Rusty075]

I suspect that "OCModShop" rectally extracted the "1/80th" number, but yes, a failed heatpipe would basially become useless. But a heatpipe failingis not really something to worry about. Short of mechanical damage, there's not many ways you can destroy a heatpipe....they generally have about a 20 year MTBF.

[albatros_la]

There is gravity in space, just not as much of it.

Yep, there is. Sorry: I'm used to speak about space in terms of no gravity as micro-gravity conditions. Call it bad habit!

spookmineer: that's right. An OVERheated heat pipe is in fact useless! When you exceed the maximum heat transfer rate (which varies with respect to the tilt angle - here on earth -), different working limits can occur (there are several ones: boiling limit, evaporator dry out, etc...). Each limit, in a different way from the others, makes the heat pipe totally ineffective!

Rusty075: you're right: heat pipe is a durable component. Failure can mainly occur in case of bad bending of the device. This could make air to leak into the heat pipe through micro-cracks (some times really hard to inspect) and mess up the entire device.

[nzimmers]

A heatpipe is up to a couple of thousand times more efficient at conducting heat than a solid copper rod would be. For example, Thermacore, one of the big manufactures of heatpipes advertises a typical thermal conductance of 50000W/mK for their run-of-the-mill pipes. A copper rod would have a conductance of 401w/mK.

okay, um, there's alot of different kinds of heat pipes though, I find it hard to believe that the heat pipes commonly used is personal computer applications have 50,000W/mK thermal conductance. But that's just my gut reaction.....

here's a qestion though - would the heat conductance of a copper rod (as you mentioned about 401w/mK) be enough for an average CPU with a TDP of about 45W ?

[jaganath]

here's a question though - would the heat conductance of a copper rod (as you mentioned about 401w/mK) be enough for an average CPU with a TDP of about 45W ?

Not a rod, that's a bad shape for dissipating heat, not enough surface area, but sure, there are plenty of all copper heatsinks that don't use heatpipes, for example the Zalman CNPS7000B-Cu, Zalman claims it can cool a P4 Prescott 540 (84W TDP) so 45W should be no problem. I would imagine that a heatpipe heatsink with similar surface area would perform better though.

[nzimmers]

if you take a look at the image I included in my original message, I wonder how swapping out the heat pipes for a copper rod that is attached to the side mounted heatsinks would perform - that's the example I had in mind

[jaganath]

if you take a look at the image I included in my original message, I wonder how swapping out the heat pipes for a copper rod that is attached to the side mounted heatsinks would perform - that's the example I had in mind

You can try it, but don't blame me if you fry your CPU.

[Zinj]

I'm not an engineer or a physicist, but after reading some of the material on thermodynamics at the Engineering Toolbox (www.engineeringtoolbox.com), I think you'll run into a problem. Copper has very high conductance, which is to say it passes heat through itself quite well. About 400 w/Mk as someone posted. In fact it's much higher than water (.58 w/Mk), even though water is commonly used in radiators.

The problem with the copper rod is that any heat built up has to be passed along a rod with a very limited cross sectional area. My guess is that that transfer is going to be quite slow until the temperature differential between different parts of the rods rises to what are likely to be unacceptable levels from your CPU's standpoint. The benefit of a heatpipe is that it's quite efficient at transferring that heat away. Water in a radiator setup is also quite efficient, not because water conducts heat well, but because it absorbs a great deal of heat and because it readily transfers it to other materials (convective heating). To make the movement of heat away from the CPU effective, a pump typically pushes the water.

Somebody please correct me if I'm off base with this.

Zinj

[star882]

Rusty075: you're right: heat pipe is a durable component. Failure can mainly occur in case of bad bending of the device. This could make air to leak into the heat pipe through micro-cracks (some times really hard to inspect) and mess up the entire device.

Actually, if it cracks, what really happens is the Freon leaks out. This will happen very quickly as there's only a tiny amount of Freon in a heatpipe. Freon is really what makes heatpipes work so well. The Freon is under about 200PSI of pressure under typical operating conditions, so even a tiny leak will cause it to fail fast.

[jaganath]

That may be the case for expensive heatpipes used on the Space Shuttle etc, but most consumer-level heatpipes just use ordinary water:

http://www.illuminex.biz/applications.html

Most heat pipes used in electronics systems are made of copper (due to its high thermal conductivity) and use water as the working fluid.

Computer Power User: Heatpipes explained

Heatpipes may contain fancy liquids (called the working fluid) such as ammonia or ethanol, but heatpipes created for CPUs generally use just plain, pure water as the working fluid. You might wonder how your CPU boils water into vapor before it vaporizes its own transistors, but if you remember your basic science lessons, you know that water boils at a much lower temperature when it’s in a vacuum. That trait makes water a useful working fluid at temperatures ranging from about 30 to 200 degrees Celsius.

[albatros_la]

The heat pipes usually used in electronics applications are copper-water heat-pipes. Freons boiling point is too low for those applications: in fact is used in devices with "lower" temperature ranges (generally, refrigerators). A copper-water heat-pipe has got in it ONLY water, if some air remains into it, it's considered like a bad made device. On space environments heat pipes are usually of the aluminium-ammonia type, cause aluminium is lighter and ammonia has a good chemical compatibility with it and a boiling point which is the best for those applications. I know this 'cause I've worked for a period for a satellite supplier projecting thermal management units, where heat pipes are a must have. However, those types of heat pipes are really different from the ones used on earth: they're groove type, which means that there's not mesh or sintered powder in them but the vessel itself is axially grooved. This type of heat pipes works really good in microgravity conditions but in gravity environment they work almost horizontally but fail even with really little tilt angles.

[spookmineer]

That depends on the pressure.

If water is used in heatpipes because the pressure is lower then 1 bar, causing it to boil at lower then 100 °C, you could also use freon at a higher pressure so the boiling point is at the needed working temperature.

Using water would be a lot safer though, using thin copper heatpipes with freon of over 10 bar (depending on the kind of freon) is not reassuring.

Most heatpipes (if I read correctly) use an inert gas to provide the needed (lower then 1 bar) pressure. In vacuum, the little water that is inside the heatpipe would evaporate completely.

[cmthomson]

Heat pipes used in commercial PC heat sinks almost universally use sintered wicks to work in flexible orientations, use water as the working fluid because of its astounding heat absorption as it boils, and use reduced internal pressure to lower the water's boiling point to about 40C.

[Kremmit]

if you take a look at the image I included in my original message, I wonder how swapping out the heat pipes for a copper rod that is attached to the side mounted heatsinks would perform - that's the example I had in mind

I'm pretty sure that if a plain old copper rod moved heat better than a much more expensive heatpipe, then that's what heatsink manufacturers would be using. Give it up, dude.

[ronrem]

This principle works amazingly in space, where there's no gravity

There is gravity in space, just not as much of it.

There is,in a sense,the sum total of the universe's gravity where ever you are-but being relativly near a relativly big hunk of it is what we refer to. In space,the gravity of half the universe is balaced by the other half

[jaganath]

In space,the gravity of half the universe is balaced by the other half

So why is the universe expanding? Genuine question (probably should be in off-topic, but what the heck).