Regarding the new seven heatpipe tower coolers!

[lilhasselhoffer]

I'm pretty sure direct touch has never worked out well on performance regardless of whether you're cooling direct die or IHS.

Companies like CM can brag all they want about how good their heatpipe grinding tolerances are; fact is GPUs, laptops (where all are direct die) and IHS CPUs have all moved to some form of coldplate contact unless they are low end enough to warrant cost-cutting like direct heatpipe contact (ie. box coolers and low end low TDP GPUs).

So, this one blows my mind. Let me, for a moment, lead you down a thought process.

Heat plates are functionally a hollow cavity. One side touches the heat source, other side rejects heat to the environment, in the cavity is a liquid inside a low pressure environment. The idea is that with the correct pressure you can have the vaporization temperature for that liquid low enough that the cold side condenses and the hot side evaporates. The reason you want this is that the phase change energy for materials is huge when compared to the change in temperature. Cool?....Cool. That all checks out.
The problem is that with such relatively large dies, and uneven distribution of thermal energy, you get hot spots and condensation sites which are out of control, and getting that fluid onto the hot bits becomes a nightmare. Thus, when you cut open a modern hot plate it looks like a sideways heatpipe...so that they can channel fluids evenly once condensed. It's almost like the technology itself has become more advanced because simply slapping on a heat pipe didn't offer enough surface area, while a "vapor chamber" didn't really offer the consistency and control necessary to make sure you didn't have hot spots.


Now, thought experiment explained. By definition then, you're looking at heat pipes which are short and fat, or long but require their heat transfer sites to be warmer to support phase change dozens of centimeters away rather than a few millimeters.

Hopefully this makes senses. If not, allow me an anecdote. A car is a car. That said, I can own a truck or a standard car. The Truck is a great option if I want to haul around a boat...but it's going to use a lot more fuel. The car won't haul a boat, but it will be very courteous on fuel consumption. The truck is a heat plate, and the car is a pipe. Both the same technology, but built differently. The truck/hot plate transfers a lot of heat...not very far. The car isn't as powerful, but will definitely move that heat farther. This is why in practice we often see the vehicle fleet. The cold plate connects to a block and pipe apparatus. The plate is designed to transfer lots of heat fast, while the pipes are actually designed to phase change at a lower temperature so they can take that heat and share it across a much larger surface...thereby dissipating more heat by virtue of simply having more to send it out to.




If not otherwise clear, I'm agreeing with a lot of what was fundamentally said above.
1) Heat pipes are only as good as their construction.
2) Contact plates and blocks to force surface area connections with heat pipes are great....assuming they don't introduce large conduction problems by being thick.
3) A badly sealed or specified heat pipe is potentially a huge issue...and this was evidenced with the 7000 series of GPUs from AND that didn't have the right amount of liquid in them.
4) More heat pipes isn't better. A set of 100 pipes which cannot get hot enough to phase change the inner material is only as efficient as conductive cooling.
5) Your situation will vary. A cooler rated for 250 watts is not going to perform well with a 15 watt thermal load...barring the thermal mass allowing for a lot of energy to be dumped into it before it hits equilibrium. Likewise, a 100 watt rated cooler won't properly condense on a 250 watt load. This is why more cooler=/=better...but cooler rated for load=better.
6) Milled heat pipes generally have the problem that they are thick enough to mill, and thus not great on surface contact. Your mileage may vary, but I prefer shaped pipes with a friction fit. They generally favor thin walls for their malleability...but are just as prone to issues as anything else.

 

[AusWolf]

I despise glass side bonnets
I like stealthy builds and older retro looking cases, it's kind of infuriating how it's almost impossible to find cases without glass fronts or sides totally impractical and restrictive.

I don't disagree, although I'm not against a simple, modest, no-bling, no-flare gaming system behind a window, either.

For example, I'm really happy with my PC's looks with the all-black CPU cooler, non-RGB RAM, MSi Pro motherboard and AMD reference GPU. I'll post some pictures in the relevant thread once I'm back from holiday.

So, this one blows my mind. Let me, for a moment, lead you down a thought process.

Heat plates are functionally a hollow cavity. One side touches the heat source, other side rejects heat to the environment, in the cavity is a liquid inside a low pressure environment. The idea is that with the correct pressure you can have the vaporization temperature for that liquid low enough that the cold side condenses and the hot side evaporates. The reason you want this is that the phase change energy for materials is huge when compared to the change in temperature. Cool?....Cool. That all checks out.
The problem is that with such relatively large dies, and uneven distribution of thermal energy, you get hot spots and condensation sites which are out of control, and getting that fluid onto the hot bits becomes a nightmare. Thus, when you cut open a modern hot plate it looks like a sideways heatpipe...so that they can channel fluids evenly once condensed. It's almost like the technology itself has become more advanced because simply slapping on a heat pipe didn't offer enough surface area, while a "vapor chamber" didn't really offer the consistency and control necessary to make sure you didn't have hot spots.


Now, thought experiment explained. By definition then, you're looking at heat pipes which are short and fat, or long but require their heat transfer sites to be warmer to support phase change dozens of centimeters away rather than a few millimeters.

Hopefully this makes senses. If not, allow me an anecdote. A car is a car. That said, I can own a truck or a standard car. The Truck is a great option if I want to haul around a boat...but it's going to use a lot more fuel. The car won't haul a boat, but it will be very courteous on fuel consumption. The truck is a heat plate, and the car is a pipe. Both the same technology, but built differently. The truck/hot plate transfers a lot of heat...not very far. The car isn't as powerful, but will definitely move that heat farther. This is why in practice we often see the vehicle fleet. The cold plate connects to a block and pipe apparatus. The plate is designed to transfer lots of heat fast, while the pipes are actually designed to phase change at a lower temperature so they can take that heat and share it across a much larger surface...thereby dissipating more heat by virtue of simply having more to send it out to.




If not otherwise clear, I'm agreeing with a lot of what was fundamentally said above.
1) Heat pipes are only as good as their construction.
2) Contact plates and blocks to force surface area connections with heat pipes are great....assuming they don't introduce large conduction problems by being thick.
3) A badly sealed or specified heat pipe is potentially a huge issue...and this was evidenced with the 7000 series of GPUs from AND that didn't have the right amount of liquid in them.
4) More heat pipes isn't better. A set of 100 pipes which cannot get hot enough to phase change the inner material is only as efficient as conductive cooling.
5) Your situation will vary. A cooler rated for 250 watts is not going to perform well with a 15 watt thermal load...barring the thermal mass allowing for a lot of energy to be dumped into it before it hits equilibrium. Likewise, a 100 watt rated cooler won't properly condense on a 250 watt load. This is why more cooler=/=better...but cooler rated for load=better.
6) Milled heat pipes generally have the problem that they are thick enough to mill, and thus not great on surface contact. Your mileage may vary, but I prefer shaped pipes with a friction fit. They generally favor thin walls for their malleability...but are just as prone to issues as anything else.

My two cents: direct contact heatpipes are usually good with a thick IHS and/or a large, central CPU die. On the other hand, they're not so good when the die is small and/or offset from the centre, and the IHS is thin, thus the die(s) don't actually have direct contact with all the heatpipes.

 

[Frag_Maniac]

I don't know about you.. but I have thousands tied into my rig, I wanna see it without taking the side off

Well to that I say, once you see it for the purpose of looking at it's beauty, that never changes. So in that regard, once you've seen it, you've seen it. Assessing dust buildup though is not only functional for performance, it also dynamically changes from day to day.

We're basically taking the equivalent of looking at a fish aquarium, just for different reasons.

 

[tabascosauz]

So, this one blows my mind. Let me, for a moment, lead you down a thought process.

I'm sure a company as quality-focused as Noctua could do a better job with quality control on direct heatpipes than Corsair did with the A500. But at the end of the day it's still firmly a strategy for budget products. And the harder it is to make/ensure consistency and quality, the less likely it makes sense in a cheap product. Coldplates design haven't been incapable or lacking in any way, and seeing as direct heatpipes haven't proven themselves to be appreciably better in any way, I just don't see any reason or chance for direct heatpipes to make some sort of major comeback.

Now that Raphael has taken Ryzen's heat density challenges to an all time high, these cheap direct heatpipe coolers are pretty much instantly disqualified on Ryzens except APUs or 65W (76W)/heat output Ryzens - especially the ones with gaps between heatpipes. And even for APUs (where cooling has been traditionally easier and simpler), it remains to be seen what the thick AM5 IHS + higher power will do to Phoenix thermals on desktop. Intel is imminently moving to disaggregation with Meteor Lake while also shrinking to Intel 4 (iirc a contemporary to N4 or N5), so they are likely to start heading down that road as well.

It's almost like the technology itself has become more advanced because simply slapping on a heat pipe didn't offer enough surface area, while a "vapor chamber" didn't really offer the consistency and control necessary to make sure you didn't have hot spots.

The 7900XTX MBA thing didn't really demonstrate that much about vapor chambers in itself, since it was a glaring error in fluid content rather than an issue with contact.

Vapor chambers are showing up in laptops in a big way (Razer Blade 14/15/16(?), Zephyrus G14, Dragonfly Pro, Legion 16", etc.). Yes, quality control is still a thing, it's extra cost compared to heatpipes, and vapor chambers are not guaranteed to be better cooling, and not to mention there are even more packages/components that must make good contact, but it's clearly working out for the laptops that have it (Dragonfly especially).

 

[GordonFreemanInTheFlesh]

The more reviews I watch, the less I am convinced that any cooler is better than any other cooler really want to like HW Canucks cooler testing but their testing results are almost always curiously mixed up compared to most others I watch. They have the FS140 not only stomping the FC140 but every other cooler in the list; a quick look anywhere else (mostly Asian reviews since TR is popular over there) and the FC140 rolls all over the FS140 by an equally large gap..........

Another roundup review I saw recently had D12L thrashing the D15

I'm inclined to blame the noise normalized testing - everyone has different equipment and tries to pretend that they're the most scientific. Just show me what they can do as they are, if they're too loud I'll just swap the fans myself ffs

This here>
_"I'm inclined to blame the noise normalized testing - everyone has different equipment and tries to pretend that they're the most scientific. Just show me what they can do as they are, if they're too loud I'll just swap the fans myself ffs"_

I 100% agree with you, why don't they test heatsinks with only one particular type of fan that way we can see the actual difference where as how they're doing it right now it's fan testing more than anything and I am 95% sure that if we were to conduct tests on all 4,6 or 7 heatpipe coolers they would all perform within margin of error!!
Bottom line is it's the fan that makes the difference.

 

[claes]

I disagree that the fan is the bottom line, but they do affect performance and this is exactly why noise normalized tests exist. A fan running at 2400rpm on cooler A is going to look very different from cooler b’s fan that has a max 900rpm

Sure, using a standard fan is useful (I miss you SPCR, ty for trying GN, shout out to VSG), but only if you run it a fixed across coolers.

 

[Frag_Maniac]

I 100% agree with you, why don't they test heatsinks with only one particular type of fan that way we can see the actual difference where as how they're doing it right now it's fan testing more than anything and I am 95% sure that if we were to conduct tests on all 4,6 or 7 heatpipe coolers they would all perform within margin of error!!
Bottom line is it's the fan that makes the difference.

I totally get your point here, but reviewers often see it as comparing one whole product vs another, and that means with the fans they have stock. It would start getting overly complex if they were to swap out fans because everyone has their preferences on what they value most, cooling, low noise, durability, etc. Then it also becomes a toss up as to what each unit would end up costing after using whatever fan consumers want to use.

I'm a big fan (no pun intended) of coolers that come with fans that are effective at cooling AND low noise, AND reasonably durable. This is why I don't like some of Corsair's products because their fans are generally loud. I also don't like manufacturers pushing RGB so hard, because to me, it's unnecessary bling that doesn't even look appealing to me.

 

[joemama]

why don't they test heatsinks with only one particular type of fan

Because the majority of consumers will not consider changing the fan

 

[eidairaman1]

So, this one blows my mind. Let me, for a moment, lead you down a thought process.

Heat plates are functionally a hollow cavity. One side touches the heat source, other side rejects heat to the environment, in the cavity is a liquid inside a low pressure environment. The idea is that with the correct pressure you can have the vaporization temperature for that liquid low enough that the cold side condenses and the hot side evaporates. The reason you want this is that the phase change energy for materials is huge when compared to the change in temperature. Cool?....Cool. That all checks out.
The problem is that with such relatively large dies, and uneven distribution of thermal energy, you get hot spots and condensation sites which are out of control, and getting that fluid onto the hot bits becomes a nightmare. Thus, when you cut open a modern hot plate it looks like a sideways heatpipe...so that they can channel fluids evenly once condensed. It's almost like the technology itself has become more advanced because simply slapping on a heat pipe didn't offer enough surface area, while a "vapor chamber" didn't really offer the consistency and control necessary to make sure you didn't have hot spots.


Now, thought experiment explained. By definition then, you're looking at heat pipes which are short and fat, or long but require their heat transfer sites to be warmer to support phase change dozens of centimeters away rather than a few millimeters.

Hopefully this makes senses. If not, allow me an anecdote. A car is a car. That said, I can own a truck or a standard car. The Truck is a great option if I want to haul around a boat...but it's going to use a lot more fuel. The car won't haul a boat, but it will be very courteous on fuel consumption. The truck is a heat plate, and the car is a pipe. Both the same technology, but built differently. The truck/hot plate transfers a lot of heat...not very far. The car isn't as powerful, but will definitely move that heat farther. This is why in practice we often see the vehicle fleet. The cold plate connects to a block and pipe apparatus. The plate is designed to transfer lots of heat fast, while the pipes are actually designed to phase change at a lower temperature so they can take that heat and share it across a much larger surface...thereby dissipating more heat by virtue of simply having more to send it out to.




If not otherwise clear, I'm agreeing with a lot of what was fundamentally said above.
1) Heat pipes are only as good as their construction.
2) Contact plates and blocks to force surface area connections with heat pipes are great....assuming they don't introduce large conduction problems by being thick.
3) A badly sealed or specified heat pipe is potentially a huge issue...and this was evidenced with the 7000 series of GPUs from AND that didn't have the right amount of liquid in them.
4) More heat pipes isn't better. A set of 100 pipes which cannot get hot enough to phase change the inner material is only as efficient as conductive cooling.
5) Your situation will vary. A cooler rated for 250 watts is not going to perform well with a 15 watt thermal load...barring the thermal mass allowing for a lot of energy to be dumped into it before it hits equilibrium. Likewise, a 100 watt rated cooler won't properly condense on a 250 watt load. This is why more cooler=/=better...but cooler rated for load=better.
6) Milled heat pipes generally have the problem that they are thick enough to mill, and thus not great on surface contact. Your mileage may vary, but I prefer shaped pipes with a friction fit. They generally favor thin walls for their malleability...but are just as prone to issues as anything else.

Its like a passive ac system (heat exchanger without a compressor)

 

[Mussels]

The baseplate is all that matters most of the time, like how a pissy little 120mm dark rock slim can keep up with 240mm and 360mm AIOs on AM4 and AM5 upto around the 150W mark - just because it's got a quality baseplate to transfer that initial heat out.

Then you get coolermaster who use the same name (CM 212) for a dozen products with everything from solid copper to heatpips with 1mm+ gaps that confuse people as to WHY they perform worse (It's the crappy contact over the hot parts of your CPU)



Older CPU's with single temperature sensors could be misleading, as a CPU could report itself as running cold while it may have massive differences between cores.
New CPU's are more sensitive and more accurate with reporting so quality standards are being forced to go in the right direction.

Heat plates are functionally a hollow cavity

That's a vapor chamber, a heatplate is solid metal. They're just a flat heatpipe, and again - a cost cutting measure. Solid copper is the best choice you can get, with heatpipes being a lot lighter and cheaper to make and coming CLOSE in performance - it's why the best heatsinks don't rely on just heatpipes.

My two cents: direct contact heatpipes are usually good with a thick IHS and/or a large, central CPU die. On the other hand, they're not so good when the die is small and/or offset from the centre, and the IHS is thin, thus the die(s) don't actually have direct contact with all the heatpipes.

Direct heatpipes were always a cost cutting measure, but it wasnt a huge issue on the older, larger CPU dies. Now that CPU's concentrate their heat in small areas, the odds of having some of that gap-space over part of a die is basically guaranteed to happen



Theres a reason why all the best coolers are the best (air, water, AIO), and it's because they've got one thing in common: Solid metal baseplates to get the instant heat transfer, and a decent method of transporting that heat out to the fins.

NH D15

The dark rock slim that i love to bits
be quiet! Dark Rock Slim Review | TechPowerUp

The original CM212 that earned the reputation CM have been riding on for years - theres a very good reason why the new designs perform worse than the OG

One of the biggest differences between the CM212 variants is the baseplate is either direct heatpipes, heatpipes with alu in the gaps, and they change the material around the heatpipes between copper and alu, with some of them having actual decent sized chunks of metal right there on the heatpipes to get the heat transferred out - because the heatpipes can't keep up with plain old direct metal.

The fancy 6 heatpipe version isn't better because of the extra heatpipes - it just needed them to support the weight of two seperate fin stacks. It performs better because it's got a better base.

 

[claes]

I wanted to correct that post too but then I saw the wacky car analogy and the mostly correct conclusions and shrugged

 

[joemama]

That's a vapor chamber, a heatplate is solid metal. They're just a flat heatpipe, and again - a cost cutting measure. Solid copper is the best choice you can get, with heatpipes being a lot lighter and cheaper to make and coming CLOSE in performance - it's why the best heatsinks don't rely on just heatpipes.

Shouldn't the vapor chamber be better at transfering heat? I would imagine the vapor chambers to be more expensive due to more complicated manufacturing processes

 

[lilhasselhoffer]

I'm sure a company as quality-focused as Noctua could do a better job with quality control on direct heatpipes than Corsair did with the A500. But at the end of the day it's still firmly a strategy for budget products. And the harder it is to make/ensure consistency and quality, the less likely it makes sense in a cheap product. Coldplates design haven't been incapable or lacking in any way, and seeing as direct heatpipes haven't proven themselves to be appreciably better in any way, I just don't see any reason or chance for direct heatpipes to make some sort of major comeback.

Now that Raphael has taken Ryzen's heat density challenges to an all time high, these cheap direct heatpipe coolers are pretty much instantly disqualified on Ryzens except APUs or 65W (76W)/heat output Ryzens - especially the ones with gaps between heatpipes. And even for APUs (where cooling has been traditionally easier and simpler), it remains to be seen what the thick AM5 IHS + higher power will do to Phoenix thermals on desktop. Intel is imminently moving to disaggregation with Meteor Lake while also shrinking to Intel 4 (iirc a contemporary to N4 or N5), so they are likely to start heading down that road as well.



The 7900XTX MBA thing didn't really demonstrate that much about vapor chambers in itself, since it was a glaring error in fluid content rather than an issue with contact.

Vapor chambers are showing up in laptops in a big way (Razer Blade 14/15/16(?), Zephyrus G14, Dragonfly Pro, Legion 16", etc.). Yes, quality control is still a thing, it's extra cost compared to heatpipes, and vapor chambers are not guaranteed to be better cooling, and not to mention there are even more packages/components that must make good contact, but it's clearly working out for the laptops that have it (Dragonfly especially).

So...to the first point I disagree. I'd assume you've seen tear-downs of the Radeon GPU heat plates. To the other comments, I use the term heat plate instead of vapor chamber because a true vapor chamber has nothing inside of it. The heat plate in the Radeon GPUs demonstrated that even if they had the correct amount of fluid, there's still a requirement to channel the condensate to such a large plate evenly.
Put less obtusely, let's look at a frying pan. The high end ones are designed with ridges and discontinuities, because when you have an uneven heat source you need to make sure the conduction is spread out evenly. The design issues then should be obvious...but let me state it. A large area, like a processor, with a few finite heat sources, needs to channel condensates evenly onto the hot side so energy can be pulled evenly from the source...or you develop hot spots and basically destroy performance .

The use of heat pipes in "low end" solutions is a joke. Allow me to state that I've yet to see a "low end" 200+ watt TDP chip...but finding a 250+ watt air cooler composed of multiple heat pipes is pretty easy. As an example, amazon listing: BeQuiet 250 watt cooler
If you're going to claim that the disparity is substantive...then I'm not sure you and I read from the same dictionary. I'd call that a premium solution...which has brethren dating well back into the days when Intel stopped including garbage coolers with their enthusiast tier offerings...let alone the professional tier stuff. I remember the Zalman silent monstrosity...about 2 pounds of copper that sliced flesh left and right while cooling anything from the Core2 lineage (a predecessor to this: Zalman cooler). The engineering that made that premium, as well as the materials, are what made it capable.


Final bit...because this is important. Direct contact heat pipes and vapor chambers are literally the same technology. Phase change cooling. The statement that one is better or worse is pretty backwards...as I thought my car analogy showed. If you cannot see that is the case, then let me try one last time. Let me also do some logistical stretching, and ask basic questions.
1) When was the last time you saw a solid cooler?
2) If vapor chambers, heat plates, and heat pipes are all the same tech then why do we look at them differently?
3) Why would a laptop need a heat plate and pipes, as the examples you cited show?

Let me answer. For me, 1 is outside of the computer industry, where a couple hundred degrees F is the delta. At this point you have issues with phase change...but if your target is no more than 90 C, water boils at 100 C, and you can pull a minor vacuum to get that boiling point to 70 C then it makes sense to do a heat pipe with very cheap water. Application matters, but I haven't seen any computer hardware without some form of heat pipe in years...barring ultra low power items like a PI. Those solutions that do use just solid state cooling suck...as anyone who has recycled a stock Intel cooler will attest to. Conduction is bound by distances...so is less efficient as power draw increases...which brings us to why heat pipes rule the roost.

Regarding question 2, marketing. Phase change cooling is not sexy. Vapor chamber is a cool buzz word. Sapphire I believe attached it to a "Toxic" brand, and it was a selling point. Thing is, this tech goes back decades. It was used in electronics since they started pulling enough heat to matter...and the things were almost always paired with a gigantic cooler block. Phase change cooling is great because it doesn't suffer from having to physically connect two points, and thus have extreme barriers with physical distance. IE, conduction is what actually transfers the thermal energy into the ambient...but the surface area of the fins to the pipes is hundreds of times greater because of that inefficiency.
So...why not just call everything a heat pipe? Well, heat plates don't do distance. They dump energy fast, but the vapor chamber is generally very thin. It's done this way to dump energy in a plane. Heat pipes generally dump in a single direction, whatever the pipe flows in. In this way you can choose to either dump into an immediately attached fin cooler, or move the heat from point to point.

Regarding 3, it's partially answered in 2. Open whatever laptop you have, and you'll find at least a few heat pipes taking energy from components to wherever you can shove a fin block and fan into. I'm looking at an Aorus model now that has a small heat pipe from the NIC and GPU which curls to one side of the unit, and another pipe leading from the CPU to a different set of fins. When I ripped it apart there was a small vapor chamber on the CPU....leading into those heat pipes. Why? Suck it from the CPU, dump it into a transfer conduit, and dump that to the surroundings by a fin block that you otherwise couldn't fit next to a hot component. It's almost like how a CPU block in water cooling sucks at cooling, and most of the actual heat transfer is inches to feet (or centimeters) away at the radiator...because fluid flow and transfer is the same principle as phase change cooling but less efficient only by virtue of phase change energy being much higher than heat capacity to change a liquid a single degree.
Yeah...from an engineering standpoint this is all fluid flow from point to point, energy transfer, and return. It's almost like none of this is complicated...and if we wanted to we could do the opposite. Let me qualify. You could take something like say, ammonia, and place it inside the same type of system. That, or say Freon using the same phase change cycle...aided in part by pressure changes forcing phase changes in a relatively adiabatic environment to forcibly transfer phase change (then heat transfer as pressure is relieved and the matter phase shifts back by absorbing ambient heat)...or in laymen's terms have A/C work.


If it isn't clear, the terminology here is all artificial. Phase change cooling is nothing new, and heat pipes versus heat plates, versus vapor chambers is all an artificial construct to simply have a fluid vaporize to absorb heat on one surface and condense to transfer that heat to another surface. Whether it's a relatively short range along surfaces, or a less efficient longer range to move it out of an area, it's all about how the engineers design it and how the manufacturer builds it. Saying one is premium and the other is not is...fundamentally misunderstanding that the technology is the same.
My Geo runs on an ICE, and so does my truck. Call me when you've got an engine that runs on concentrated moon water...because otherwise you're just deciding to pretend that these things aren't simply a factor of tuning performance...which in itself is a joke. Trying to tune performance when you can see 65-125 watts of load, varying thermal generation by point, and varying contact area and quality (see: the Gamer's Nexus investigations) makes this an unwinnable challenge. If you want to disagree then explain to me why I can spend hundreds on a liquid cooling setup...and have it beat hands down by a cheap air cooler. Please note, this is something like a 65 watt tdp...where the problem becomes having such a small delta between the water and CPU so that the water to ambient delta basically has nothing to conduct. The phase change cooler instead has a much easier time, because it doesn't transfer heat via conduction.

 

[claes]

What the hell are you talking about?

The difference between a heat plate/base plate and a vapor chamber is that the former is a solid hunk of metal and vapor chambers are filled (partially) with liquid (ie not empty). There is no phase change in solid metal, just conduction.

Heat pipes and vapor chambers are not at all the same. Vapor chambers offer phase change across the entire coverage of the heat source, heat pipes can only conduct heat where they make contact with the heat source.

There have been a few cpu coolers that use vapor chambers as their base plate and they don’t seem to offer any improvement over solid metal. I haven’t seen any real comparisons with GPUs, but given how many high-end GPUs forego vapor chambers for a traditional base plate and perform just as well as those that do use vapor chambers I’m gonna say the difference is negligible.

Sure, a heatsink with a higher capacity will definitely see diminishing returns on a lower wattage chip. It’ll also outperform a heatsink with a lower heat capacity on a lower wattage chip, as dozens and dozens of reviews have shown. Sounds like a you problem for spending too much cooking a 65W chip.

IDK what you consider a high-end frying pan but I would encourage you to throw away anything with ridges. The best pans are flat and solid with as few irregularities as possible. This is just a wild claim. Any inconsistencies in heat distribution are a problem of the heat source, as dozens of tests on high-end convection have shown.

I can’t even address the rest… tl;dr and loaded with contradictions

 

[tabascosauz]

So...to the first point I disagree. I'd assume you've seen tear-downs of the Radeon GPU heat plates. To the other comments, I use the term heat plate instead of vapor chamber because a true vapor chamber has nothing inside of it. The heat plate in the Radeon GPUs demonstrated that even if they had the correct amount of fluid, there's still a requirement to channel the condensate to such a large plate evenly.
Put less obtusely, let's look at a frying pan. The high end ones are designed with ridges and discontinuities, because when you have an uneven heat source you need to make sure the conduction is spread out evenly. The design issues then should be obvious...but let me state it. A large area, like a processor, with a few finite heat sources, needs to channel condensates evenly onto the hot side so energy can be pulled evenly from the source...or you develop hot spots and basically destroy performance .

The use of heat pipes in "low end" solutions is a joke. Allow me to state that I've yet to see a "low end" 200+ watt TDP chip...but finding a 250+ watt air cooler composed of multiple heat pipes is pretty easy. As an example, amazon listing: BeQuiet 250 watt cooler
If you're going to claim that the disparity is substantive...then I'm not sure you and I read from the same dictionary. I'd call that a premium solution...which has brethren dating well back into the days when Intel stopped including garbage coolers with their enthusiast tier offerings...let alone the professional tier stuff. I remember the Zalman silent monstrosity...about 2 pounds of copper that sliced flesh left and right while cooling anything from the Core2 lineage (a predecessor to this: Zalman cooler). The engineering that made that premium, as well as the materials, are what made it capable.


Final bit...because this is important. Direct contact heat pipes and vapor chambers are literally the same technology. Phase change cooling. The statement that one is better or worse is pretty backwards...as I thought my car analogy showed. If you cannot see that is the case, then let me try one last time. Let me also do some logistical stretching, and ask basic questions.
1) When was the last time you saw a solid cooler?
2) If vapor chambers, heat plates, and heat pipes are all the same tech then why do we look at them differently?
3) Why would a laptop need a heat plate and pipes, as the examples you cited show?

Let me answer. For me, 1 is outside of the computer industry, where a couple hundred degrees F is the delta. At this point you have issues with phase change...but if your target is no more than 90 C, water boils at 100 C, and you can pull a minor vacuum to get that boiling point to 70 C then it makes sense to do a heat pipe with very cheap water. Application matters, but I haven't seen any computer hardware without some form of heat pipe in years...barring ultra low power items like a PI. Those solutions that do use just solid state cooling suck...as anyone who has recycled a stock Intel cooler will attest to. Conduction is bound by distances...so is less efficient as power draw increases...which brings us to why heat pipes rule the roost.

Regarding question 2, marketing. Phase change cooling is not sexy. Vapor chamber is a cool buzz word. Sapphire I believe attached it to a "Toxic" brand, and it was a selling point. Thing is, this tech goes back decades. It was used in electronics since they started pulling enough heat to matter...and the things were almost always paired with a gigantic cooler block. Phase change cooling is great because it doesn't suffer from having to physically connect two points, and thus have extreme barriers with physical distance. IE, conduction is what actually transfers the thermal energy into the ambient...but the surface area of the fins to the pipes is hundreds of times greater because of that inefficiency.
So...why not just call everything a heat pipe? Well, heat plates don't do distance. They dump energy fast, but the vapor chamber is generally very thin. It's done this way to dump energy in a plane. Heat pipes generally dump in a single direction, whatever the pipe flows in. In this way you can choose to either dump into an immediately attached fin cooler, or move the heat from point to point.

Regarding 3, it's partially answered in 2. Open whatever laptop you have, and you'll find at least a few heat pipes taking energy from components to wherever you can shove a fin block and fan into. I'm looking at an Aorus model now that has a small heat pipe from the NIC and GPU which curls to one side of the unit, and another pipe leading from the CPU to a different set of fins. When I ripped it apart there was a small vapor chamber on the CPU....leading into those heat pipes. Why? Suck it from the CPU, dump it into a transfer conduit, and dump that to the surroundings by a fin block that you otherwise couldn't fit next to a hot component. It's almost like how a CPU block in water cooling sucks at cooling, and most of the actual heat transfer is inches to feet (or centimeters) away at the radiator...because fluid flow and transfer is the same principle as phase change cooling but less efficient only by virtue of phase change energy being much higher than heat capacity to change a liquid a single degree.
Yeah...from an engineering standpoint this is all fluid flow from point to point, energy transfer, and return. It's almost like none of this is complicated...and if we wanted to we could do the opposite. Let me qualify. You could take something like say, ammonia, and place it inside the same type of system. That, or say Freon using the same phase change cycle...aided in part by pressure changes forcing phase changes in a relatively adiabatic environment to forcibly transfer phase change (then heat transfer as pressure is relieved and the matter phase shifts back by absorbing ambient heat)...or in laymen's terms have A/C work.


If it isn't clear, the terminology here is all artificial. Phase change cooling is nothing new, and heat pipes versus heat plates, versus vapor chambers is all an artificial construct to simply have a fluid vaporize to absorb heat on one surface and condense to transfer that heat to another surface. Whether it's a relatively short range along surfaces, or a less efficient longer range to move it out of an area, it's all about how the engineers design it and how the manufacturer builds it. Saying one is premium and the other is not is...fundamentally misunderstanding that the technology is the same.
My Geo runs on an ICE, and so does my truck. Call me when you've got an engine that runs on concentrated moon water...because otherwise you're just deciding to pretend that these things aren't simply a factor of tuning performance...which in itself is a joke. Trying to tune performance when you can see 65-125 watts of load, varying thermal generation by point, and varying contact area and quality (see: the Gamer's Nexus investigations) makes this an unwinnable challenge. If you want to disagree then explain to me why I can spend hundreds on a liquid cooling setup...and have it beat hands down by a cheap air cooler. Please note, this is something like a 65 watt tdp...where the problem becomes having such a small delta between the water and CPU so that the water to ambient delta basically has nothing to conduct. The phase change cooler instead has a much easier time, because it doesn't transfer heat via conduction.

The vapor chambers in the aforementioned laptops aren't there for contact in a heatpipe system. Literally the entire cooling system is one vapor chamber. My G14 is the same way.



Not sure where all the car analogies are coming from

 

[Mussels]

Shouldn't the vapor chamber be better at transfering heat? I would imagine the vapor chambers to be more expensive due to more complicated manufacturing processes

Transferring a small amount of heat, yes. Exactly like a heatpipe.

Storing heat? no. Less mass means less ability to absorb the heat, which means it's not as good for little spikes - things like how ryzen CPU's spike up in temperature with single threaded burst loads.
It takes time (not much, but still time) to boil that vapour and recondense, which delays the onset of the cooling performance briefly, and it relies on however much solid metal is present for that time period.
have no idea how long that is as it's probably less than a second, but for CPU's that poll a thousand times a second that's still slower than they are.

TL;DR:

• It can transfer heat, but it needs time to boil and recondense to do so - plain metal doesn't.
• Heatpipes can also be saturated, they can reach a point they can't get any hotter and just stay 'boiled' and no longer cool well
• Locked rotations mean you can't turn the cooler to improve cooling

Dual CCX ryzen 7000
OG hyper 212, perfect copper contact on IO die and both CCXs with heatpipes to transfer heat away from that copper plate

Then try with one of the cheap monstrosities
Theres a good 10+ variants, this is the worst one for obvious reasons
Official photo on the left, and one Chad from tweaktown took of a different revision (It's weirder when you have these people on social media for some reason)
Notice how the slightly newer version has a more flush finish - the problem is they kept going back and forth with the bases, so you never know what you'll get and quality control can be lacking.

These arent to scale because i don't care that much, this is just an example since the actual physical layout varies between every single CPU generation, and how it's mounted with every direct heatpipe variant from every brand.
Some heatpipes are useless without that baseplate to transfer heat to them, and some CPU cores are lucky to even have aluminum above them because they've just got that airgap instead like the worlds lumpiest stock cooler.

Change this to a single CCX setup and it's possible to have the entire 140W of heat from CCX and IoDie on one heatpipe, without that baseplate to spread the heat to the other heatpipes they don't help at all.

Dual CCX is worse if you run unlocked PBO, but with PBO limits active a single CCX will be much worse with half the heat transfer area.

When you see how older CPUs were designed you can see how this design worked well enough - on some sockets rotation was locked to have them face a certain way to help, but overall you had decent contact with the hotspots.
Delidded 6700k
(again, not to scale. I don't care. They had some wiggle room and slid around a little anyway)
If four of the 6 heatpipes had contact and the other two helped with the passive bleedover, sure - that worked. Doesnt work on a setup where one heatpipe has all the heat.

If i could find flat images of these coolers and actual measurements i could do better examples, but these companies love hiding the baseplates of the shittier coolers - they only show off the good ones or use CGI renders of the ones that always arrive looking like a transformer used them as a chew toy.



Heres an attempt at using a delidded ryzen 7000 with transparent images, since i know someone will argue
The angle is slightly off because the image was angled in the original video i snagged it from, and the heatpipes are actually slightly smaller than this in reality - but the mounting holes lined up more or less accurately so i kept it that scale.
This orientation is how the cooler has to be mounted on AM4/5, you can't rotate it. One heatpipe for two CCXs.


Rotated to how it looks once its in the socket, and matches how they've performed for me in real life on single CCX chips:

At best you could end up with 2 pipes on each CCX with this setup, but they're sharing with the IO die and very off-centered - meaning only half the heatpipe is even being used. It's not travelling easily or fast from the bottom of this image, up, then a 90 degree angle up the heatpipes - it's designed to have the heat in the center to spread the pressure of boiling equally in both directions - off center heat means unequal pressure, and less than perfect use of those heatpipes. It wont reach all the way to the top of the 'far' pipes via the gas/liquid, only by travelling along the copper itself.

AM4 is oriented this way in the socket too:
It does slightly better, but you're getting only two off-centered heatpipes on a single CCX chip. It may work okay with a 5900x at 140W, but it'd choke massively with a 5800x at the same wattage.

 

[AusWolf]

Transferring a small amount of heat, yes. Exactly like a heatpipe.

Storing heat? no. Less mass means less ability to absorb the heat, which means it's not as good for little spikes - things like how ryzen CPU's spike up in temperature with single threaded burst loads.
It takes time (not much, but still time) to boil that vapour and recondense, which delays the onset of the cooling performance briefly, and it relies on however much solid metal is present for that time period.
have no idea how long that is as it's probably less than a second, but for CPU's that poll a thousand times a second that's still slower than they are.

TL;DR:

• It can transfer heat, but it needs time to boil and recondense to do so - plain metal doesn't.
• Heatpipes can also be saturated, they can reach a point they can't get any hotter and just stay 'boiled' and no longer cool well
• Locked rotations mean you can't turn the cooler to improve cooling

Dual CCX ryzen 7000
OG hyper 212, perfect copper contact on IO die and both CCXs with heatpipes to transfer heat away from that copper plate
View attachment 305410View attachment 305412


Then try with one of the cheap monstrosities
Theres a good 10+ variants, this is the worst one for obvious reasons
Official photo on the left, and one Chad from tweaktown took of a different revision (It's weirder when you have these people on social media for some reason)
Notice how the slightly newer version has a more flush finish - the problem is they kept going back and forth with the bases, so you never know what you'll get and quality control can be lacking.
View attachment 305413View attachment 305419


These arent to scale because i don't care that much, this is just an example since the actual physical layout varies between every single CPU generation, and how it's mounted with every direct heatpipe variant from every brand.
Some heatpipes are useless without that baseplate to transfer heat to them, and some CPU cores are lucky to even have aluminum above them because they've just got that airgap instead like the worlds lumpiest stock cooler.

Change this to a single CCX setup and it's possible to have the entire 140W of heat from CCX and IoDie on one heatpipe, without that baseplate to spread the heat to the other heatpipes they don't help at all.
View attachment 305417View attachment 305418

Dual CCX is worse if you run unlocked PBO, but with PBO limits active a single CCX will be much worse with half the heat transfer area.

When you see how older CPUs were designed you can see how this design worked well enough - on some sockets rotation was locked to have them face a certain way to help, but overall you had decent contact with the hotspots.
Delidded 6700k
(again, not to scale. I don't care. They had some wiggle room and slid around a little anyway)
If four of the 6 heatpipes had contact and the other two helped with the passive bleedover, sure - that worked. Doesnt work on a setup where one heatpipe has all the heat.
View attachment 305416


If i could find flat images of these coolers and actual measurements i could do better examples, but these companies love hiding the baseplates of the shittier coolers - they only show off the good ones or use CGI renders of the ones that always arrive looking like a transformer used them as a chew toy.



Heres an attempt at using a delidded ryzen 7000 with transparent images, since i know someone will argue
The angle is slightly off because the image was angled in the original video i snagged it from, and the heatpipes are actually slightly smaller than this in reality - but the mounting holes lined up more or less accurately so i kept it that scale.
This orientation is how the cooler has to be mounted on AM4/5, you can't rotate it. One heatpipe for two CCXs.


Rotated to how it looks once its in the socket, and matches how they've performed for me in real life on single CCX chips:

At best you could end up with 2 pipes on each CCX with this setup, but they're sharing with the IO die and very off-centered - meaning only half the heatpipe is even being used. It's not travelling easily or fast from the bottom of this image, up, then a 90 degree angle up the heatpipes - it's designed to have the heat in the center to spread the pressure of boiling equally in both directions - off center heat means unequal pressure, and less than perfect use of those heatpipes. It wont reach all the way to the top of the 'far' pipes via the gas/liquid, only by travelling along the copper itself.
View attachment 305424


AM4 is oriented this way in the socket too:
It does slightly better, but you're getting only two off-centered heatpipes on a single CCX chip. It may work okay with a 5900x at 140W, but it'd choke massively with a 5800x at the same wattage.
View attachment 305426

Exactly my point a couple of posts above. Direct contact heatpipes used to be cool (and still are with some setups, I guess), but it's best to forget that the technology has ever existed with Ryzen 3000 and newer.

 

[lilhasselhoffer]

What the hell are you talking about?

The difference between a heat plate/base plate and a vapor chamber is that the former is a solid hunk of metal and vapor chambers are filled (partially) with liquid (ie not empty). There is no phase change in solid metal, just conduction.

Heat pipes and vapor chambers are not at all the same. Vapor chambers offer phase change across the entire coverage of the heat source, heat pipes can only conduct heat where they make contact with the heat source.

There have been a few cpu coolers that use vapor chambers as their base plate and they don’t seem to offer any improvement over solid metal. I haven’t seen any real comparisons with GPUs, but given how many high-end GPUs forego vapor chambers for a traditional base plate and perform just as well as those that do use vapor chambers I’m gonna say the difference is negligible.

Sure, a heatsink with a higher capacity will definitely see diminishing returns on a lower wattage chip. It’ll also outperform a heatsink with a lower heat capacity on a lower wattage chip, as dozens and dozens of reviews have shown. Sounds like a you problem for spending too much cooking a 65W chip.

IDK what you consider a high-end frying pan but I would encourage you to throw away anything with ridges. The best pans are flat and solid with as few irregularities as possible. This is just a wild claim. Any inconsistencies in heat distribution are a problem of the heat source, as dozens of tests on high-end convection have shown.

I can’t even address the rest… tl;dr and loaded with contradictions

Let me go over this one time. A vapor chamber is a hollow cavity with a low pressure area. Inside said low pressure area is a fluid, whose boiling point has been depressed by that low pressure environment. Done, we are good?
Now, I call these a heat plate instead of a vapor chamber because they are not purely a plate to plate interface. You are more than welcome to use different terminology...but it's stupid. A vapor chamber had those structures inside of it to channel the condensate back to the evaporation plate because if they didn't exist you'd have hot spots. I'm hoping you can understand that. I term this a heat plate because it's designed differently from a pure vapor chamber.


Now the other bit. You seem to want to die on the heat plate actually being the contact block. Why use that terminology? Well, because it's not stupid. A contact block takes complex tubular geometry, and provides a conductive conduct path between the planar heat sink and the tubular heat pipe. You're welcome to whine that this is not how marketing sells it to you...but I can only suggest that when marketers dictate terminology then the whole world will be driving an SUX. If the Robocop line there goes above you, it's meant to convey that marketing doesn't have to be accurate. They have to sell us crap.


You...seem to not understand how thermal energy propogates. In instances where you have a point source, that is uneven, to create even heating you generally need to create discontinuities. You seem to not want to understand this as a pan analog, despite a 20 second search of amazon yielding dozens. Let me discuss this purely from the technology side. You have a per core temperature. You have a die temperature. You rarely have these temperatures uniform. Why? The measurement on the IHS is one temperature....right? Well, competent testing shows heat spikes in certain areas. It's uneven...so why does this matter for a heat plate? Please, call it a vapor chamber if you want. Those ridges inside the heat plate are there to make sure hot spots get fluid...otherwise if they evaporate everything they have to transfer energy by conduction through a gas...which is what borked the Radeon GPUs. Not enough fluid had the inside of the heat plate massively uneven in heating, which created a runaway thermal event despite the edges of the heat plate being relatively cool.
Now...let me short this, because you seem not to want to read. There are pure vapor chambers. The tech is simple, and for a relatively uniform load it's great. There are heat plates, which are vapor chambers which have internal geometry to make sure that planar heat from a source can be quickly translated away from the source and onto another semi-planar surface. These are heat plates...as you transfer from plate to plate. The terminology is simplistic...and free from marketing...because we are well beyond the age of simple vapor chambers. This is not a choice, but necessary because hundreds of watts of thermal energy in points is not going to have your basic vapor chamber effective...



Let me try one last simple illustration. I'm going to go with the pan. You have a cast iron pan over a burner...and it's great. It takes forever to heat up, but when it does you have even heat and it doesn't fluctuate. Huge energy input, high temperature, even heat. Now, you buy a new aluminum pan. It heats in 1/10th the time, and there are hot areas where the burner is concentrated. Is it garbage? No. It's rapid heat transfer, but it takes way less heat.
Now, you've got to steam some dumplings. Do you do it in a cast iron pan? No. The thermal mass is huge, the steam takes forever to build, and it's a mess. You do it in a thin walled steel pot or an aluminum one, because it's easier to heat. The water also acts as mediator for those hot spots.
Now, let's say you boil all the water our. The pan doesn't just suddenly get empty, right? It has certain areas which boil first...and this is easy to see. That uneven heating could have you burning one area and having another al dente. If you have such an uneven heating in a situation where you literally target one temperature for phase change cooling you've started the race with a broken leg.

 

[claes]

Now, I call these a heat plate instead of a vapor chamber because they are not purely a plate to plate interface. You are more than welcome to use different terminology...but it's stupid. A vapor chamber had those structures inside of it to channel the condensate back to the evaporation plate because if they didn't exist you'd have hot spots. I'm hoping you can understand that. I term this a heat plate because it's designed differently from a pure vapor chamber.

I prefer not to make up terms where the language already exists. What I think you are describing isn’t novel but incredibly common

Plate heat exchanger - Wikipedia

en.m.wikipedia.org

Now the other bit. You seem to want to die on the heat plate actually being the contact block. Why use that terminology? Well, because it's not stupid. A contact block takes complex tubular geometry, and provides a conductive conduct path between the planar heat sink and the tubular heat pipe. You're welcome to whine that this is not how marketing sells it to you...but I can only suggest that when marketers dictate terminology then the whole world will be driving an SUX. If the Robocop line there goes above you, it's meant to convey that marketing doesn't have to be accurate. They have to sell us crap.

The vast majority of consumer CPU heatsinks don’t use vapor chambers or “heat plates,” but instead just solid hunks of copper. Call it stupid or marketing if you want, but the only manufacturer that I know of that has played around with vapor chambers as the initial medium of exchange between the cpu and a cooler is coolermaster, with little success.

You...seem to not understand how thermal energy propogates. In instances where you have a point source, that is uneven, to create even heating you generally need to create discontinuities. You seem to not want to understand this as a pan analog, despite a 20 second search of amazon yielding dozens. Let me discuss this purely from the technology side. You have a per core temperature. You have a die temperature. You rarely have these temperatures uniform. Why? The measurement on the IHS is one temperature....right? Well, competent testing shows heat spikes in certain areas. It's uneven...so why does this matter for a heat plate? Please, call it a vapor chamber if you want. Those ridges inside the heat plate are there to make sure hot spots get fluid...otherwise if they evaporate everything they have to transfer energy by conduction through a gas...which is what borked the Radeon GPUs. Not enough fluid had the inside of the heat plate massively uneven in heating, which created a runaway thermal event despite the edges of the heat plate being relatively cool.

Sure, some GPUs use vapor chambers as an initial point of transfer from the GPU to the cooler, as do many laptops. I do not know of any pans that do this because it’d be a terrible design (maybe you mean tri-ply steel, but there are no ridges — aluminum is used precisely because it thins and spreads evenly). Don’t know what your point is here.

Now...let me short this, because you seem not to want to read. There are pure vapor chambers. The tech is simple, and for a relatively uniform load it's great. There are heat plates, which are vapor chambers which have internal geometry to make sure that planar heat from a source can be quickly translated away from the source and onto another semi-planar surface. These are heat plates...as you transfer from plate to plate. The terminology is simplistic...and free from marketing...because we are well beyond the age of simple vapor chambers. This is not a choice, but necessary because hundreds of watts of thermal energy in points is not going to have your basic vapor chamber effective...

I understand, but your initial claim was that CPU coolers use vapor chambers to transfer heat from the CPU to the cooler’s heatpipes. The vast majority do not.

Let me try one last simple illustration. I'm going to go with the pan. You have a cast iron pan over a burner...and it's great. It takes forever to heat up, but when it does you have even heat and it doesn't fluctuate. Huge energy input, high temperature, even heat.

Idk of anyone who ever claimed that cast iron provides even heat. They are notorious for hot spots.

Now, you buy a new aluminum pan. It heats in 1/10th the time, and there are hot areas where the burner is concentrated. Is it garbage? No. It's rapid heat transfer, but it takes way less heat.
Now, you've got to steam some dumplings. Do you do it in a cast iron pan? No. The thermal mass is huge, the steam takes forever to build, and it's a mess. You do it in a thin walled steel pot or an aluminum one, because it's easier to heat. The water also acts as mediator for those hot spots.
Now, let's say you boil all the water our. The pan doesn't just suddenly get empty, right? It has certain areas which boil first...and this is easy to see. That uneven heating could have you burning one area and having another al dente. If you have such an uneven heating in a situation where you literally target one temperature for phase change cooling you've started the race with a broken leg.

I have no idea what you’re trying to illustrate here except for a lack of knowledge of the history of Asian cooking and their pans. Appreciate all of the condescension but assuming I am an idiot and making weak analogies that don’t even correlate to your claims is not very compelling.

 

[Ruru]

Freezer 50 is fine (I used one, put it to my 2nd rig as I put my main system back to custom loop) but it sucks that it doesn't use standard fans.

Personally I think that air cooling has somewhat reached its limits or at least it's close to it (unless you go nuts and make a way bigger cooler than what we currently have).

 

[Cheeseball]

Putting a bunch of heatpipes into coolers isn't a new thing.

Anyone remember the Scythe Orochi? 10 heatpipes. Performed well, but was not a chart topper.

View attachment 304612

The Orochi was primarily a passive heatsink, so it would not be a chart topper by itself. But if you got the one with their 140mm fan, it would beat the Tuniq Tower 120 back in the day. (which it should since its 3 heatpipes vs 10 lol)

 

[GordonFreemanInTheFlesh]

Freezer 50 is fine (I used one, put it to my 2nd rig as I put my main system back to custom loop) but it sucks that it doesn't use standard fans.

Personally I think that air cooling has somewhat reached its limits or at least it's close to it (unless you go nuts and make a way bigger cooler than what we currently have).

I don't see why we can't like stack heatpipes on top of each other (I know that is going to lose efficiency a lot though)!
I wouldn't mind paying 150€ for a cooler that has maybe 12-18 heatpipes!

A quad tower cooler maybe....

 

[Mussels]

Now, I call these a heat plate instead of a vapor chamber because they are not purely a plate to plate interface.

you cant use a name you invented for yourself, and expect it to make sense. use the official names or you invite arguing for no reason.
A vapour chamber is a hollowed out piece of metal (to save costs) with a fluid inside it
a heat plate is a piece of solid metal to absorb heat by direct surface area contact

I don't see why we can't like stack heatpipes on top of each other (I know that is going to lose efficiency a lot though)!
I wouldn't mind paying 150€ for a cooler that has maybe 12-18 heatpipes!

A quad tower cooler maybe....

because it wouldnt work?
What absorbs heat best is not the same as what transfers or dissipates heat the best

absorbs best: perfect contact with zero gaps of a solid dense metal with a high thermal conductivity, usually copper.
transfers best: copper, as far as common materials go. Copper heatpipes achieve this with less weight and material cost.
what dissipates best: metal fins with as much surface to air/water contact area as possible. Copper and aluminium work great here, alu because of it's lighter weight letting you use more fins before things break.


when you talk about stacking these things, remember what the fins on your heatsink already are - that's what they're doing. You're forgetting that the heat has to transfer from the source (the baseplate) up, so every centimeter it travels up there is less and less heat, so the further you get the less heat reaches there.

In order for heat to reach the highest levels of a cooler the lower parts have to be totally heat saturated to the point everythings already overheating - exactly what happens with those direct contact coolers where the cores and IHS are roasting, because the heat can't travel up to the next layer fast enough.

 

[sLowEnd]

I don't see why we can't like stack heatpipes on top of each other (I know that is going to lose efficiency a lot though)!
I wouldn't mind paying 150€ for a cooler that has maybe 12-18 heatpipes!

A quad tower cooler maybe....

It can and has been done. For me, the Xigmatek Thor's Hammer comes to mind.

Xigmatek Thor's Hammer HDT-S126384 Review

As you surmised, the stacking of the heatpipes didn't actually give it a leg up on the competition. The cooler still performed pretty well, but not because of the stacked heatpipes.

 

[AusWolf]

It can and has been done. For me, the Xigmatek Thor's Hammer comes to mind.

Xigmatek Thor's Hammer HDT-S126384 Review

As you surmised, the stacking of the heatpipes didn't actually give it a leg up on the competition. The cooler still performed pretty well, but not because of the stacked heatpipes.

Being on par with the CM 212 Evo with so many heatpipes isn't exactly what I'd call "performed well".