Dumb yet technical question: Heatpipes

[janas19]

FWIW, I emailed a couple heatsink manufactures to see what they would say. CoolerMaster is the only one to reply so far. Per "Jarad", all of the heatpipes used in their heatsinks use pure water as a working fluid. The 212 line might be good, but it's not that good - if they're using real heatpipes, it's pretty clear everyone is.

You can either take their word for it, along with the heaps of evidence that's been presented here, or you can keep wearing your tinfoil hat.

At this point, you're trying to argue that the entire CPU heatsink industry is actively lying to consumers, including manufacturing objects that exactly match authentic heatsinks in every way except the presence of a fluid inside, to the detriment of heatsink performance. That's a pretty substantial claim, and I think it's fair to place the burden of proof on you. Do you have any evidence to back up your claims, bearing in mind that absence of evidence is not evidence of absence?

Unless you have something else, something new and verifiable, I'm done here. I've got other brick walls to beat my head against.

Well, I certainly appreciate you being proactive and e-mailing the manufacturers. Following up and posting their replies, and perhaps any datasheets or visual representations of the working fluid used, would certainly be helpful. Remember, my mind is not made up, nor did I ever make that claim. I only said I have good reason to be skeptical.

Beyond that, I feel that, since I am trying to be as fair and accurate as possible, I would disagree that the "burden of proof" rests on me since I did not make the CLAIM of selling a "heat pipe." I'm pretty sure that burden would, in fact, lie with the manufacturer who sells the devices, and (as with Thermacore) if they do incorporate a working fluid in their heat pipes, it would be rather easy to provide a datasheet confirming this.

Again, I'm not trying to attack anyone or make accusations, I'm only looking for the truth. I'll be happy to engage in debate with anyone who genuinely wants to know the truth, but I'm not going to engage in personal attacks or unfounded accusations.

 

[jraskell1]

I'll reiterate the end of my previous post.

If Heat Pipe coolers did not function as actual heat pipes, then they simply would not function as coolers, at all. There is simply not enough material present for them to function as a standard thermal conductance cooler.

 

[birthdaymonkey]

If this theory about the non-heatpipe heat pipes were true, then surely some manufacturer would simply release a high end CPU cooler without heatpipes: for example, a device with a solid piece of metal connecting the CPU interface area with the fin array.

If the faux heatpipes, which janas19 claimed are simply pipe-shaped heatsinks, conduct heat like any other heatsink material, then surely bigger faux heatpipes would be better than the smallish ones we normally see on top performing heatsinks. In other words, a solid block of copper would be better than a slender tube of copper at moving heat away from the CPU to the fins.

So why doesn't someone release a heatsink with such a design? Well, it could be 1) a conspiracy of non-competition involving every maker of computer cooling solutions in the world. Or it could be 2): that computer heatpipes actually conduct heat away from the CPU to the fin array better than a normal piece of metal would.

If we accept that (2), being the simpler explanation, is true, then the reason for this must be that A) heatpipes actually do work according to physical principles as commonly understood, or that B) it's just a coincidence that they move heat so well, and nobody really understands why, or that C) there is a secret technology at work whose real mechanism is being suppressed for marketing reasons.

 

[know of fence]

I'll reiterate the end of my previous post.

If Heat Pipe coolers did not function as actual heat pipes, then they simply would not function as coolers, at all. There is simply not enough material present for them to function as a standard thermal conductance cooler.

Yeah it was a well made, convincing argument. I must admit that that (never having held a heatpipe cooler in my hand) for a while I thought that heat pipes were solid copper and it was only the superior heat conductivity of Cu that allowed them to work as such. It turns out heat pipes are THE technological answer to all passive cooling woes.

However there are plenty of fascinating questions from this discussion.
- Why is so little fluid used inside the heapipes, shouldn't all the wick be completely drenched?
- Wouldn't heat pipes actually benefit from having the motherboard lying down, rather than on the side, like in most towers? Then not only capillary action but also gravity would aid to bring cooler water back to the cpu. This effect may even be (easily) measurable.
- How does pressure change inside the pipe during working CPU temperatures?
- At what temperatures will heapipes rupture, if left uncooled? http://www.silentmods.com/section2/item287/part14

 

[QuantumPion]

However there are plenty of fascinating questions from this discussion.
- Why is so little fluid used inside the heapipes, shouldn't all the wick be completely drenched?
- Wouldn't heat pipes actually benefit from having the motherboard lying down, rather than on the side, like in most towers? Then not only capillary action but also gravity would aid to bring cooler water back to the cpu. This effect may even be (easily) measurable.
- How does pressure change inside the pipe during working CPU temperatures?
- At what temperatures will heapipes rupture, if left uncooled? http://www.silentmods.com/section2/item287/part14

1) No, if the wick was drenched than it would not wick properly. Ever overfill a zippo lighter?
2) No, the capillary force is much greater than the weight of the fluid, orientation has negligible effect
3) PV=nRT
4) longitudinal stress for a thin-walled pressure vessel:
sigma = P*r/t
copper yield strength = 70 MPa
use ideal gas law to substitute for P
sigma = (nRT/V)*r/t
solve for T:
T = (sigma * t * V)/ (r * n * R)

So given the the volume, radius, wall thickness of the heatpipe and mass of water contained, you can calculate the failure temperature (assuming it is uniformly heated and there are no defects, etc).

 

[kevinsbane]

However there are plenty of fascinating questions from this discussion.
- Why is so little fluid used inside the heapipes, shouldn't all the wick be completely drenched?
- Wouldn't heat pipes actually benefit from having the motherboard lying down, rather than on the side, like in most towers? Then not only capillary action but also gravity would aid to bring cooler water back to the cpu. This effect may even be (easily) measurable.
- How does pressure change inside the pipe during working CPU temperatures?
- At what temperatures will heapipes rupture, if left uncooled? http://www.silentmods.com/section2/item287/part14

1. You want there to be enough liquid for it to work, but not enough to drench the wick. As you increase the amount of liquid in any given heat pipe, the heat pipe responds more slowly. IE, it takes longer to boil 10 mL of water than 1 mL of water. Let's say it takes 10x longer to start to boil 10mL of water than 1mL. But meanwhile, before it boils, no heat is being transferred away from the heat source! So therefore, all you have is a pool of liquid heating up slowly at the heat source. The end result is that you hit higher temperatures faster. In essence, by increasing the amount of liquid in a heat pipe, you raise the operating temperature of that heat pipe. It won't begin to "work" until it hits that point.

Thus, with less liquid, the heat pipe will respond more quickly. If you have just enough liquid that the instant the pipe heats up just a bit, it evaporates part of the liquid, then the heat pipe is operating optimally.

2. Gravity would help, but as far as I know, most heat pipes simply don't have the quantity of liquid where it would make a difference. Referencing the previous posts, it seems most computer heatsink pipes don't have enough liquid in them for liquid to drip out. Considering the texture of the interior of the heat pipes, I would suspect that surface tension has a far greater strength relative to gravity, and thus gravity would not be a factor.

3. The boiling part of the heat pipe has a high pressure, and the condensation portion of the heat pipe has low pressure. In between, wind. A boiling kettle will expel steam, correct? Collect this steam, recondense it, and allow the water to flow back into the kettle. That's a heat pipe.

4. Heat pipes don't really have enough liquid for this to be an issue...

Beyond that, I feel that, since I am trying to be as fair and accurate as possible, I would disagree that the "burden of proof" rests on me since I did not make the CLAIM of selling a "heat pipe." I'm pretty sure that burden would, in fact, lie with the manufacturer who sells the devices, and (as with Thermacore) if they do incorporate a working fluid in their heat pipes, it would be rather easy to provide a datasheet confirming this.

I would disagree with your disagreement. Because it is wrong. You do not need to actually be *selling* a heat pipe (or even just claiming to be selling a heat pipe) to be required provide some burden of proof. The heatpipe manufacturers' claims make sense. And their claims are borne out through empirical testing: cut a heatpipe, and the heatpipe stops working. You, on the other hand, have made an extraordinary claim: that they are lying, and that something else (unknown) is at work. As the saying goes, "Extraordinary claims require extraordinary proof.". Having made an extraordinary claim, you must provide extraordinary evidence. As of yet, all I have read is "But I can't see any liquid, therefore there must not be any".

 

[Zap]

What does all this tell me? The conclusion of my investigation is that CPU heatsink manufacturers claim that their heatsink pipes are ?heat pipes,? however no evidence I found has proved this claim. The evidence, in fact, strongly suggests that CPU heatsink pipes contain no liquid, and hence are not true ?heat pipes.? My belief is that CPU heatsink manufacturers use the term ?heat pipe? as a marketing gimmick, mainly.

You have come to the wrong conclusions due to one fact, and that is if it were NOT a heatpipe, those heatsinks would not work as well as they do. Not even close. See below.

You can draw your own conclusions.

My first conclusion after seeing that wall of text was that you have too much time on your hands. Forgive me if I didn't follow all the links.

I'll reiterate the end of my previous post.

If Heat Pipe coolers did not function as actual heat pipes, then they simply would not function as coolers, at all. There is simply not enough material present for them to function as a standard thermal conductance cooler.

This. The "heatpipes" in huge tower heatsinks HAVE to be heatpipes because if they were not the heatsinks (heat exchangers) would not work any better than the stock cooler, and probably worse.

The heatpipe manufacturers' claims make sense. And their claims are borne out through empirical testing: cut a heatpipe, and the heatpipe stops working.

This.

 

[The Boston Dangler]

2. Gravity would help, but as far as I know, most heat pipes simply don't have the quantity of liquid where it would make a difference. Referencing the previous posts, it seems most computer heatsink pipes don't have enough liquid in them for liquid to drip out. Considering the texture of the interior of the heat pipes, I would suspect that surface tension has a far greater strength relative to gravity, and thus gravity would not be a factor.

not true. this is a pic of one of my coolers - the manufacturer states mounting with the bends facing upwards will reduce performance. the capillary action isn't enough to draw liquid up and over to the heat source. this is not unusual for pc heatsinks, ymmv.

 

[janas19]

This is where I eat crow.

AAAAAGH I WAS WRONGGG ARRRRRR GODD$%%IT MOTHER&$&&* SON OF A B*$$H.

God dang it. I think I was wrong. How did I not believe the pipes on cpu coolers were not heatpipes? Lol. Welll, I was operating under the assumption that a heat pipe has to be under vacuum (not always, but Wikipedia says typically) that a larger amount of fluid would be needed (I understand a wick would only need to be moist, not dripping. I understood this all along. I just thought there would need to be a great amount of liquid (thought 8-9 mL) within the pipe to cool properly- not understanding that as soon as the vapors condense, the cycle begins again almost instantaneously. Seems even a small amount eg 5-6 mL per pipe would suffice).

Apparently, in my research of phase change coolers, with compressors and refrigerants, I became biased as to how a"phase change cooler" should look and operate - never considering that vapor change can take place in a much more passive manner. I was an idiot guys. I'm sorry. I didn't mean to be a troll, I was just confused...

I do still have one question, if anyone is willing to help: I note that many of the larger tower coolers with the heat pipes can keep a CPU operating temp to 60-70 degrees. My question is, at this temperature, how does the water change into vapor? It would seem the process would somehow stop at that point.

Thank you.

 

[janas19]

1. You want there to be enough liquid for it to work, but not enough to drench the wick. As you increase the amount of liquid in any given heat pipe, the heat pipe responds more slowly. IE, it takes longer to boil 10 mL of water than 1 mL of water. Let's say it takes 10x longer to start to boil 10mL of water than 1mL. But meanwhile, before it boils, no heat is being transferred away from the heat source! So therefore, all you have is a pool of liquid heating up slowly at the heat source. The end result is that you hit higher temperatures faster. In essence, by increasing the amount of liquid in a heat pipe, you raise the operating temperature of that heat pipe. It won't begin to "work" until it hits that point.

Thus, with less liquid, the heat pipe will respond more quickly. If you have just enough liquid that the instant the pipe heats up just a bit, it evaporates part of the liquid, then the heat pipe is operating optimally.

2. Gravity would help, but as far as I know, most heat pipes simply don't have the quantity of liquid where it would make a difference. Referencing the previous posts, it seems most computer heatsink pipes don't have enough liquid in them for liquid to drip out. Considering the texture of the interior of the heat pipes, I would suspect that surface tension has a far greater strength relative to gravity, and thus gravity would not be a factor.

3. The boiling part of the heat pipe has a high pressure, and the condensation portion of the heat pipe has low pressure. In between, wind. A boiling kettle will expel steam, correct? Collect this steam, recondense it, and allow the water to flow back into the kettle. That's a heat pipe.

4. Heat pipes don't really have enough liquid for this to be an issue...

Makes sense.

 

[QuantumPion]

I do still have one question, if anyone is willing to help: I note that many of the larger tower coolers with the heat pipes can keep a CPU operating temp to 60-70 degrees. My question is, at this temperature, how does the water change into vapor? It would seem the process would somehow stop at that point.

Thank you.

The water changes into vapor in the same process as at any temperature below the boiling point - evaporation.

 

[janas19]

The water changes into vapor in the same process as at any temperature below the boiling point - evaporation.

Ahhh... *face palm*

Thank you. I must be totally confused now. Somewhere, somehow in my childhood some science teacher filled my head with the idea that liquids change into vapor only at boiling point. I'm confused, how is evaporation different?

*Edit* I see now - boiling point is where the vapor pressure is equal to atmospheric pressure. Evaporation is a surface phenomenon only. Heat of vaporation is what causes cooling.

Please forgive me, I went to a public school! I was a bad student! *sobs* :-(

 

[know of fence]

1) No, if the wick was drenched than it would not wick properly. Ever overfill a zippo lighter?
2) No, the capillary force is much greater than the weight of the fluid, orientation has negligible effect
3) PV=nRT
4) longitudinal stress for a thin-walled pressure vessel:
sigma = P*r/t
copper yield strength = 70 MPa
use ideal gas law to substitute for P
sigma = (nRT/V)*r/t
solve for T:
T = (sigma * t * V)/ (r * n * R)

So given the the volume, radius, wall thickness of the heatpipe and mass of water contained, you can calculate the failure temperature (assuming it is uniformly heated and there are no defects, etc).

4) Awesome reply :thumbsup: to a tongue-in-cheek question. with t being the radial thickness and r the inside radius. By knowing at what temperature a heatpipe blows up it's possible to calculate the amount of gas inside it. A special oven would be required for this, rather than the blowtorch used in the link. Though yield strength also depends on temperature and those numbers may be even harder to come buy.
To determine how much fluid is inside, it would probably be easier to weigh the thing, open up and dry and weigh it again.

3) Q: Pressure inside the heatpipe? --> Your A: PV=nRT
Ideal gasses never liquefy. This may be a problem, I'm trying to understand what the limiting factor of heat pipe efficiency is. Pertaining to the more general question of why heat pipes haven't killed water pump cooling completely.
It may be gas viscosity as the vapor moves to the cold part, gas viscosity also happens to increase with rising temperature.
Or it may be just copper-thickness of the pipe, as the latest designs all insist on directly pressing pipes against the CPU.

2) Q: Is it better to place CPU on the bottom?
It's a a quantitative question, though I have yet to stumble upon a review that can quantify the difference. But they do tend to mention positioning in passing.

1) Q: Why not more liquid? Your A: Because wick doesn't wick when overfilled.
A candle wick is always overfilled, yet it works, by pushing the melted wax to a place surrounded by air.
I don't like Kevinsbane's explanation either. We are dealing with surface film quantities either way. So let me rephrase the question:
If the wick is partially dry, what prevents the crucial part above the CPU of being that dry spot?

Early model heatpipe - water supply tube with repurposed solder wick sticking out.​

Regarding what is inside the heat pipe:
- Primarily it's water.
Water is a polar solvent which adheres to/wets the polar copper wick.
It has a convenient temperature range and high evaporation enthalpy. http://en.wikipedia.org/wiki/Enthalpy_of_vaporization

More Questions!
5) Do head pipes actually reach the boiling point?
A : sometimes
B : always
C : never
D : usually
- Best explanation I've found, yet: http://www.cheresources.com/htpipes.shtml

 

[janas19]

Thank you for the follow-up here, Know of Fence. I'm really enjoying your posts and the pics you provide, to help me understand heat pipes better. I'm thinking about building a heat pipe - mentioning using repurposed solder wick is just GREAT. Awesome find dude, really. :thumbup:

Just thought I would say that.

 

[janas19]

If the wick is partially dry, what prevents the crucial part above the CPU of being that dry spot?

- Primarily it's water.
Water is a polar solvent which adheres to/wets the polar copper wick.

Quick correction: The wick isn't about polarity. If polarity was the only property at work here, it would just make the water stick to the wick.

What causes the water to move down the wick is capillary action. It is the exact same property where, if you put just the edge of a dry paper towel in a small puddle of water, the water will move UP the towel, even against gravity. If part of the paper towel dried out, and there was still more water left, the capillary property would move the water back to the dry part. Capillary action is based upon permeability (tiny pores in a material) and surface tension, I believe.

The same property occurs in heat pipes. The wicks are metal, but they are porous (great pic, btw). As soon as the bottom part dries out, and more liquid becomes available (which it does, since the vapors condense in the cooler portion), the liquid will move back to the bottom, allowing the cycle to continue.

I think that would answer the first question too.

My only question would be this: On the bottom surface of the heatsink, does the wick touch the surface? Or is it above the heatsink, and the liquid falls (drips) on?!

 

[kevinsbane]

1) Q: Why not more liquid? Your A: Because wick doesn't wick when overfilled.
A candle wick is always overfilled, yet it works, by pushing the melted wax to a place surrounded by air.
I don't like Kevinsbane's explanation either. We are dealing with surface film quantities either way. So let me rephrase the question:
If the wick is partially dry, what prevents the crucial part above the CPU of being that dry spot?

Be careful on your terminology. A "dry" wick still has a layer of surface water, except in situations of absolute 0% humidity. It has a surface film of water that will be present regardless of how much you rub away at it. This film of water is a function of relative humidity, and if you manage to remove it mechanically, it will reform as some of the humidity in the air condenses on to that dry surface. This film is also not a constant thickness; an area with less humidity will have a thinner film than an area with higher humidity. This film of water does still contribute to the heat transport properties of heat pipes. It would be technically inaccurate to say that the film of water "moves" from areas of higher thickness to lower thickness, but in practise this is what it "appears" to do.

That aside, the thing preventing the "dry" spot right above the CPU is enough working fluid in your heat pipe. And that is determined by various factors, but generally 3 major variables: design operating temperature, heat source, dissipation rate. In general, the greater the design operating temperature and the greater the heat flux, the more liquid is needed. The lower the dissipation rate at the cool end, the more liquid is needed.

You didn't seem to like my explanation as to why there isn't more liquid, so perhaps a quick analogy would be illustrative as to why more is not necessarily better.

Imagine that a heat pipe is a highway between A and B, the working fluid are buses, and the people are the units of heat to be transported. People gather at A at a certain rate, and the goal is to get them to B as efficiently as possible. Now, due to the peculiarities of physics*, a bus can ONLY leave A when it is full of people. If there's only 40 people in a 50 person bus, it doesn't leave; it just sits there waiting for 10 more people. Now, imagine that 5 people a minute are arriving at A. It takes 10 minutes to fill up a bus, then the bus leaves to drop people off at B and another one comes. During those 10 minutes, people are just sitting around at A, getting angry that the bus isn't leaving. 1 min before the bus leaves, 45 people (45 units of heat) are at A.

Now, we double the capacity of the system: we double the size of the bus. (The physics behind this doesn't mean we double the number or frequency of bus arrivals). So now, we have a 100 person bus. Rate of people arriving doesn't change. So now it takes 20 minutes to fill up a bus. People have to wait twice as long! Not only that, 1 min before the bus leaves, 95 people are at A (95 units of heat)!

So what does this all mean? A given heatpipe is optimized for a certain set of environmental factors. Heatpipes also only operate at a certain temperature range. For a CPU heatpipes, this is the design tradeoff.

Adding more liquid increases the capacity of a heatpipe to transfer heat. At the same time, it increases the operating temperature; both the minimum temperature (where enough evaporation at the hot end takes place) and maximum (all the liquid in the heat pipe evaporates) is increased. What prevents a "dry" spot above the CPU is proper design and testing for the right amount of liquid for your design specs.

So, for your CPU: do you want your heatsink to have the operating temperature between 50-150C, and thus safeguard your CPU up to 150C? That is, it only really starts working @50C, but it keeps liquid at the hot area up to 150C. Or do you want the heatsink to start working @25C, but work only up to 100C?

5) Do head pipes actually reach the boiling point?

They can. Getting to boiling point would significantly degrade the efficiency of a heat pipe though. When both ends reach boiling, the heatpipe has failed and the hot end rapidly increases in temperature. For a system that is working properly, no.

*Heating 1 mL water from 20-30C = ~42 J (+42). Heating it further from 30-40C might take an additional ~42 J (+42), but then assume it competely evaporates, which consumes 2240J (-2240).

 

[janas19]

So, for your CPU: do you want your heatsink to have the operating temperature between 50-150C, and thus safeguard your CPU up to 150C? That is, it only really starts working @50C, but it keeps liquid at the hot area up to 150C. Or do you want the heatsink to start working @25C, but work only up to 100C?

The tradeoff I believe you are referring to is range. Adding more liquid will increase the overall cooling range of that liquid, and allow it to cool at a higher operating temp. However, as someone else pointed out, a bigger range means that it will be slower and hotter at the low end.

I heard someone say, to optimize a heat pipe, theoretically you want it to boil at your target temperature. Thus, if your target temperature was 60c, you would apply proper vacuum so that the liquid boiled at 60c.

This method would work if the temperature of the cpu was constant - say, for example, it was a heat plate with only one setting, and you wanted it to cool at one setting.

However, a CPU has a maximum operating temperature, and it does not always operate at that maximum temperature. The temperature will fluctuate according to load. So what I'm wondering is, how would you optimize a heat pipe to account for temperature variances? Do you want cooling at maximum temps, midrange, what?

 

[know of fence]

Discussions are suited to kindle an interest in these matters, but arguments and deductive reasoning can't carry one very far if there are so many unknowns involved. It's impossible to Sherlock Holmes this stuff. That's why they are the experimental sciences, the best thing one can hope for is to produce (measurement-) data or cite good sources.

Heat pipes look easy, but they involve

A. aggregate states, vapor pressure --> thermodynamics
B. streaming, viscosity --> rocket science
C. wick, working fluid --> surface physics, material science

A. Kevinsbane, you bring up the interesting calculations of how much energy it would take to evaporate a certain amount of water. And we know one CPU pipe is designed to handle 20 to 40 Watt (Joule/s) of heat. From this using 2240 kJ/kg (vaporisation enthalpy) we could calculate the minimum amount of water required to wick back every second on top of the hot part. 20 / 2257 = [J/s]*[g/J] = 8.8 mg/s = 8.8 µl/s

Of course this doesn't yet include temperature difference and heat capacity of water vapor (2.08 J/[g K]). We'd have to know the difference of temperatures between hot gas and the cold end. This should be added to vaporization enthalpy in the equation above.

It is of interest to note that even at a temperature delta of 80 degrees, heat capacity of vapor is still (166 J/g = 166 kJ/kg), negligibly small compared to the enthalpy of evaporation.

Also "relative humidity" is a meteorological term applying to water only for [present vapor pressure] / [vapor pressure at standard temperature and pressure]. There is no reason AFAIK to use humidity in this context. The gas state should be always saturated with vapor inside a heatpipe, which means, similarly to what you've said, there is a film of condensed water on every surface, except the hot ones of course, they are the reason we need a wick to soak enough water to the source. Unless of course there exists one super-wick which can dry air.

- "Vapor pressure" again is the most fundamental thing, it should be labelled on all chemicals. It is the reason for why things smell FFS.

B. The source I linked earlier states regarding a small notebook heat pipe:

The amount of fluid in a heat pipe of this diameter is less than 1cc. In a properly designed heat pipe, the water is totally contained within the capillary wick structure and is at less than 1 atmosphere of pressure.

A wick works even at a fraction of its minimal capacity (e.g. felt tip pen), I think I can appreciate why less is more, but the amount could also be controlled by the thickness of the wick. I concede your responsiveness arguments as true, just spare me more anthropomorphic analogies.

C. From the same source

The quality and type of wick usually determines the performance of the heat pipe, for this is the heart of the product.

A wick works by increasing surface and thus increasing surface creep. Also Kevinsbane's explanation of ever thinning liquid film is adequate I think if the attraction between Cu and water is stronger than between H2O molecules.
Now I'm assuming that the negative "free" electrons sitting on the outside of the metal make all metals polar and negative on the surface. The metal surface is attracted to to the positive side of the water dipole. This would also work with the polar glass surface (just like in the picture).

It is generally understood that a high surface tension increases the capillary effect, but it also can be too high. Surface tension describes among other things forces between particles. Those can be straight up EM-forces, like in the case of water, or rather complicated effects similar to those in the covalent bond / metallic bond. Mercury because of the latter has such a high surface tension (strong inter particle forces), that it doesn't at all wet most surfaces. No surface film - no capillary effect.

 

[The Boston Dangler]

i found this linked on frostytech.com:

http://www.1-act.com/advanced-technologies/heat-pipes/water-heat-pipes.php

 

[aigomorla]

A friendly reminder... seems like this thread has gotten a bit heated up.
Now i dont like playing cop, so please dont make me...

Lets all be civil and try to be somewhat professional... well, adhere to the forum rules that is.

Basically this is a warning for those of you guys who are about to snap and start going postal... take a break b4 i make you take a break...

That is all

This has been a public broadcast from your friendly AT moderator... had this been a real test people would be issued ban's...

CnC Aigomorla

 

[janas19]

:/ Well calculations are great and necessary, but a jumble of calculations alone is just confusing. That's why I tried to offer that, for us people who aren't engineers

 

[deimos3428]

Sounds like it's only a matter of time before someone figures out how to combine a liquid cooling loop within a heatpipe, and do some sort of steam cooling loop. Has it been done already?

 

[know of fence]

:/ Well calculations are great and necessary, but a jumble of calculations alone is just confusing.

My last post was all over the place...
Posting calculations I was hoping a wiseacre will appear to point-out or correct my errors. I heard Lawrence Krauss say on Youtube, the desire to nail faults with other's work was what really drives the science community. Something like a forum peer review, wouldn't it be nice?

On that note:

In the previous post we established the amount of water per second (8 to 9 mg) it takes to cool 20W of heat using evaporation, which then is converted to mol quantity using water's 18 g/mol atom weight. With that it's possible to finally calculate the pressure change inside a heatpipe, with a couple assumptions.

Assumption 1: Hot water vapor acts similarly to ideal gas.
Assumption 2: The pipe has a volume of 1 cm³ or 1 cc.

Pressure according to the ideal gas law p = nRT/V is inversely proportional to volume, so a 10 times bigger tube will have a tenth of pressure. Pressure inside the heat pipe also is different from one end of the pipe to the other, so we can consider the 1cc Volume above the hot part as a sort of initial pressure chamber which then pushes hot steam to the other end.

p = nRT / V

(R = gas constant 8.3145 J/[mol K]; T = 378 K; V = 0.000 001 m³; n= 0.005 mol)

p = 1'571'441 Pa ≈ 15 bar ≈ 15 atm

Those 15 bars however are the initial vapor pressure above the hot parts, not total pressure inside the pipe, the speed of steam inside the HP will be considerable, which is likely to be the reason for why these pipes allegedly can transport heat up to 300 times better than a copper rod of the same diameter. The big pressure difference/gradient is the reason why gas viscosity isn't a deciding factor in the effectiveness of a pipe.

A self made HP will certainly work if the cooler is up and the heat source is down, but to get a heat pipe that works sideways, or even upside down solder wick may not be enough.
I indistinctly remember having burned my hand both stirring boiling water with a pipe and a even a drinking straw.

 

[technogiant]

This is a very interesting topic for me as heatpipes have become the bane of my computing life.

I've made a computer chill box using a striped down ac unit which cools the internal air temp to -30c at idle and -22c at full pc load.

I wanted to use standard air cooling components but found that heatpipes don't work at these temps.....the working fluid is almost invariably distilled water and it of course freezes and so does not move the heat through phase change as designed.

Solid metal conductive heat sinks have no problem for obvious reasons but are not as effective as a heatpipe design.

I tried a zalman 8900 which combines conductive and heat pipe cooling, this enables the conducted heat to thaw the heatpipes but there is a second problem with heat pipes, they are tuned to a certain operational temperature. The pressure in them is reduced to allow the water to boil at lower than 100c....from my experimentation I think they kick in at about 30c......below that temp the water does not boil and so does not transfer heat.

It's not too much of a problem for cpu cooling,I now have a zalman 7500 solid copper star flower cooler.

The problem is with current graphics cards, atm I have 2x gtx460 palit sonic cards which have solid heat sinks. But most modern cards now use heatpipes or vapour chambers which employ the same heat transfer method.

So I'm going to be struggling when I come to up grade.

I've toyed with the idea of changing the cooling fluid, butane liquifies at -1c and isobutane at -11c so would be good candidates....but would be an awkward job.....perhaps easier with a vapour chamber than trying to do all the convoluted heatpipes.

Does anyone here have any ideas?

 

[Cespenar]

One question was: How can liquid vaporize before boiling point. Not the exact words but for future reference try to remember that when heatpipes (things that pipe heat from point A to point B) are manufactured, a partial vacuum chamber or partial pressure chamber is used. Can't believe you haven't heard of them.
deimos, a watercooled heatsink doesn't need pipes. It is perfect as it is.