Overclock.net banner

1 - 20 of 60 Posts

·
Registered
Joined
·
40 Posts
Discussion Starter · #1 ·
This is my proposal for the next step in heatpipe cooler performance. I suspect that facing area is the primary performance limitation on current premium heatsink designs. This design is the most direct way to step things up.




Advantages:
  • Twice the facing area of the best current designs
  • Allows unrestricted ambient intake and exhaust
  • Optimal heatpipe orientation for gravity feed
  • Reduced requirement for clearance and case width
Challenges:
  • Incompatible with many (all?) current cases - this one's a doozy :)
  • Variable CPU placement dictates flexible mounting design - could be tricky
  • Additional weight means case mounting is a requirement
Notes:
  • Most obviously the heatpipes aren't optimally placed in this image. I have limited modelling fu and only did enough to describe the concept.
  • A horizontal placement of the radiator would possibly allow compatibility with some top rad-ready cases but at the expense of a longer heatpipe run and much reduced gravity advantage.
Thoughts:
Phase change heatpipe coolers already often outperform watercooling units. The water cooling advantage is flexible placement allowing a larger radiator. Given a large ambient facing area a heatpipe design could outperform all current closed loop units on the market.

I'm still curious about the performance impact of longer than usual heatpipes but decided that a gravity flow design is likely to mitigate.
 

·
Registered
Joined
·
4,172 Posts
if it was possible it would weigh a ton and cost a fortune a water cooling loop would most likely be cheaper.

as you said al cases are diff so making the pipes reach would be difficult if not impossible.
the n=only thing i think could work is the same as you have it but have the heatsink close to the top of the mb and case companies make a case so the heat sink can be screwed to the back of mb panel

edit: new idea. bring the pipes towards the front of the mb and through the heatsink length ways and turn the fans around
 

·
Registered
Joined
·
575 Posts
This looks interesting. What exactly do you mean when you say "facing area"? I think possibly creating a strategically coloured fin network could help with heat distribution, perhaps jet black dull fins where the heat pipes meet the radiator and some how design it so silver fins can reflect the heat upwards and draw it away from the black base of the radiator. Graphene would be a huge innovation in heat sink design because it is by far the best heat conductor.
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #4 ·
Weight: Yes it would have to be secured to the case. It still sometimes makes me sad that BTX didn't succeed as a new universal form factor. Fixed CPU position and mounting would have made this a much easier sell.

Cost: More than current designs but still less than watercooling and this would potentially out perform a triple rad water setup.

Front position: Long heatpipes working against gravity made this option unattractive for me. Also I like the idea of an unrestricted ambient flow for the radiator.

Facing area: Frontal area taking in fresh ambient air. For example the NH-D14 facing area is less than 140MM^2. The NH-U14S has less bulk and total surface area but it's larger facing area more than compensates in most situations. This design has more than twice the facing area of the NH-U14S.
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #6 ·
Thank you for the entertaining video. It set me on a tour of loop heat pipe designs. I think it's unlikely to be a viable product. Pumping losses from moving all that liquid around has to make it less efficient than a typical heatpipe. It also would be necessary to compensate for significant pressure changes, possibly more than most water loops can sustain.

Interestingly the most advanced loop heat pipes are capillary pumped and a little complicated and are an excellent solution in weightless environments yet they fall short of the performance of a simple vertical heatpipe when gravity can assist.
 

·
Premium Member
Joined
·
27,526 Posts
Quote:
Originally Posted by matman View Post

Interestingly the most advanced loop heat pipes are capillary pumped and a little complicated and are an excellent solution in weightless environments yet they fall short of the performance of a simple vertical heatpipe when gravity can assist.
If I understand you correctly, a conventional heatpipe cooler performs better on a horizontal motherboard than on a vertical motherboard?
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #8 ·
Conventional heatpipes use various methods to wick the condensate back to the evaporator. In very general terms the better the wick, the better the heatpipe performs without gravity assistance or even against gravity. However, a stronger wick is more likely to impede gravity return in a vertical pipe.

To answer your question, most heatpipe CPU coolers are well wicked and capable of functioning in most orientations, Even so all of them will perform a little better when oriented with the radiator above the block and many would stall out if installed on an upside down motherboard.

A heatpipe optimised specifically for vertical orientation will significantly outperform a heatpipe designed for orientation flexibility. More so as the length increases. My concept would use minimally wicked pipes and mandate orientation as shown in the OP.
 

·
Premium Member
Joined
·
6,034 Posts
Quote:
Originally Posted by matman View Post

Thank you for the entertaining video. It set me on a tour of loop heat pipe designs. I think it's unlikely to be a viable product. Pumping losses from moving all that liquid around has to make it less efficient than a typical heatpipe. It also would be necessary to compensate for significant pressure changes, possibly more than most water loops can sustain.

Interestingly the most advanced loop heat pipes are capillary pumped and a little complicated and are an excellent solution in weightless environments yet they fall short of the performance of a simple vertical heatpipe when gravity can assist.
the cooler shown in the vid has the same pump as a conventional heatpipe
biggrin.gif
the "pump" is simply the boiling action of the liquid inside. heat is transferred due to the heat of evaporation which is the exact same physics behind heatpipes. the only difference between that unit and heatpipes is the liquid used and the method of liquid return. heatpipes use wick to return the liquid to the heat source while the unit in the vid uses gravity to return the liquid. in theory the unit shown in the vid should work as well as heatpipes due to the special liquid they use inside. most heatpipes simply use pure water sealed inside at a reduced pressure to drive the boiling point to the desired temp for the application. whereas the liquid used in the cooler shown naturally boil at the desired temp. as for the "pumping loss", there isn't any
biggrin.gif
boiling the liquid is the "pump" and the more you boil the more heat is removed from the heat source. as for the pressure change, that's why they have to use a special liquid that is engineered for this application. the radiator condenses the vapor back to liquid thereby creating a low pressure zone that hot vapor naturally move towards which is proven effective since heatpipe physics uses the exact same principle.

personally i like the design of that unit in the vid. i think it has a great deal of promise if they can ever figure out a tubing solution that will last over 10 years. the primary downfall to that system are the tubes used. it can degrade over time from the many heat/pressure cycles it must endure. unlike copper pipes which are far more durable, the flexible tubing is the main concern in that unit.
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #10 ·
Quote:
Originally Posted by psyclum View Post

the only difference between that unit and heatpipes is the liquid used and the method of liquid return
I would say the main difference is the amount of liquid used. Heat pipes use a small amount of working liquid and transport the heat in gas state. This contraption contains mostly liquid which the gas must work against to reach the radiator (i.e. pumping losses).

Now that I think about it, we're actually looking at something much more like a typical water cooling loop with a very slow pump and less like a heatpipe. Nearly all of the heat from the gas state is just going to be transferred back into the liquid so it's not going to behave anything like the thermal resistance of a heatpipe.

I haven't been able to find any empirical testing of the setup in that video or anything similar. In the the absence of evidence except for the party trick shown I'm going to stick with my conclusion that a party trick is all it's good for.
 

·
Premium Member
Joined
·
6,034 Posts
/shrug.

it's simple physics... you boil liquid, it becomes a gas. gas is lighter then liquid so gravity will drive the gas bubble to travel upwards. THAT is the pump...

as for "pumping loss"... what exactly are you "losing"? the energy that is DRIVING the "pump" is the heat generated by the CPU... so "LOSING heat" is actually the job description for a heatsink isn't it?

as for the "speed" of the pump, it is directly related to the amount of heat that is boiling the liquid. low heat = slower boil, high heat = fast boil. as long as the radiator is able to dissipate the heat, the gas will be returned to liquid state as quickly as it's cooled down and gravity will pull the liquid back to the waterblock to be reboiled so there wouldn't be much of a pressure problem. you MAY run into a pressure problem if your radiator is too small or your fan is broken and not dissipating the heat fast enough and the additional gas would increase the pressure to ultimate failure pressure, but i'm sure that has been taken into consideration in the construction of the unit. this is also the reason of my concern as indicated in my previous post that the tubing is the primary problem of this system. maybe they can use some kind of braided tubing for reinforcement to withstand the elevated pressure over long lifetime but that would render the liquid invisible which would lose the "cool factor" but from a physics standpoint the science behind it is sound.

this is not the 1st time this type of principle is used. it's a close relative to the heatpipe principle and here is another similar concept that uses the same physics

https://www.google.com/patents/US5606341
https://www.google.com/patents/US7337829

whether you believe the concept works or not, it's up to you. but if you go by the "vids or it didn't happen" principle... then the vid has already provided
biggrin.gif
 

·
Registered
Joined
·
575 Posts
Quote:
Originally Posted by psyclum View Post

/shrug.

it's simple physics... you boil liquid, it becomes a gas. gas is lighter then liquid so gravity will drive the gas bubble to travel upwards. THAT is the pump...

as for "pumping loss"... what exactly are you "losing"? the energy that is DRIVING the "pump" is the heat generated by the CPU... so "LOSING heat" is actually the job description for a heatsink isn't it?

as for the "speed" of the pump, it is directly related to the amount of heat that is boiling the liquid. low heat = slower boil, high heat = fast boil. as long as the radiator is able to dissipate the heat, the gas will be returned to liquid state as quickly as it's cooled down and gravity will pull the liquid back to the waterblock to be reboiled so there wouldn't be much of a pressure problem. you MAY run into a pressure problem if your radiator is too small or your fan is broken and not dissipating the heat fast enough and the additional gas would increase the pressure to ultimate failure pressure, but i'm sure that has been taken into consideration in the construction of the unit. this is also the reason of my concern as indicated in my previous post that the tubing is the primary problem of this system. maybe they can use some kind of braided tubing for reinforcement to withstand the elevated pressure over long lifetime but that would render the liquid invisible which would lose the "cool factor" but from a physics standpoint the science behind it is sound.

this is not the 1st time this type of principle is used. it's a close relative to the heatpipe principle and here is another similar concept that uses the same physics

https://www.google.com/patents/US5606341
https://www.google.com/patents/US7337829

whether you believe the concept works or not, it's up to you. but if you go by the "vids or it didn't happen" principle... then the vid has already provided
biggrin.gif
Sorry, but what exactly are we arguing here? I love science :3
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #13 ·
I'm not sure why this Silverstone demo is being defended so strongly as the solution I'm looking for in the OP.
Quote:
"vids or it didn't happen"
Like most people who discuss CPU cooling performance I'll go with "temps or it didn't happen"

I'm not and haven't argued the science of "hot air goes up" but there's no data at all on how efficient that silverstone demo setup was. If it works very well why is there no product? Why are there no other products in any field using this very simple design to move heat? Most likely because it's not as effective as existing methods.

I also love science
youtube =/= science
patents =/= science (patents aren't even proof of functionality let alone production viability)
Quote:
low heat = slower boil, high heat = fast boil
Actually in a pressure sealed system more heat = higher pressure = higher boiling point = higher fluid temperature = less efficient heat extraction. The system will however reach equilibrium and behave pretty much like a slow moving water loop. Whether the effect of an evaporation phase will compensate for the lack of impingement turbulence and the fact that the hotplate is partially covered by thermally insular gas is not clear.
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #14 ·
Ok I think I have an analogy to illustrate pumping losses.

If you set up liquid nitrogen in a pot on a CPU it will boil and remove energy as nitrogen gas escapes. If you trapped some pressure produced and used it to pump some water around a loop that pressure would feed back into the system and reduce the effective cooling of the system.

A heatpipe produces a very effective phase cycle with very little pumping overhead. There is just no way this system can be as efficient while pumping a kilo of liquid around.
 

·
Premium Member
Joined
·
27,526 Posts
The coolant used is very important. For example the phase change of something like R134 will absorb / release many times more heat than water can.

Captherm MP1120 is also a phase change heat powered cooler
http://forums.bit-tech.net/showthread.php?t=267032
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #16 ·
Very nice. I hope they succeed and show everyone else the next step in CPU cooling.

Current premium coolers all seem to be just taking off the shelf heat pipes and slapping them between a block and a radiator. There's bound to be significant gains in truly optimising a heatpipe/phase change design for CPU cooling. These guys seem to be stepping up with some real R&D.

Hard to say from pictures or website but it seems to be a heavily custom vapor chamber heatpipe hybrid.

Regardless the technology though there is no substitute for moving more cool air through the radiator. I stand by my proposed model.
 

·
Premium Member
Joined
·
27,526 Posts
Was just looking at your OP and wondering... horizontal pipes through base with both sides going up to fin block would seem logical. Twice to cooling area from same heat area. This would also increase area available on pipe between fins and base for mounting.

Isn't this why most coolers use double ended pipes?
 

·
Registered
Joined
·
40 Posts
Discussion Starter · #18 ·
You are making perfect sense. I was just worrying about keeping the pipes as short and vertical as possible. Thank you.

Considering how long it took me to make just a single pipe bend in Sketchup, drawing this revision is going to take me a while. It's significant enough that I definitely want to draw it again now.
 

·
Premium Member
Joined
·
27,526 Posts
You could just get a Zalman TNN 500AF
or a Zalman LQ1000 Z-Machine Hybrid Liquid Cooled Case (w/WB5+)
or the ThermalRight HSC-101
or Quiet PC Zeno H61 Fanless Media PC

There are others. At the 2007 and 2008 Computex show there were several passive cooling concept cases displayed
 

·
Premium Member
Joined
·
6,034 Posts
Quote:
Originally Posted by matman View Post

I'm not sure why this Silverstone demo is being defended so strongly as the solution I'm looking for in the OP.

Like most people who discuss CPU cooling performance I'll go with "temps or it didn't happen"

I'm not and haven't argued the science of "hot air goes up" but there's no data at all on how efficient that silverstone demo setup was. If it works very well why is there no product? Why are there no other products in any field using this very simple design to move heat? Most likely because it's not as effective as existing methods.

I also love science
youtube =/= science
patents =/= science (patents aren't even proof of functionality let alone production viability)
Actually in a pressure sealed system more heat = higher pressure = higher boiling point = higher fluid temperature = less efficient heat extraction. The system will however reach equilibrium and behave pretty much like a slow moving water loop. Whether the effect of an evaporation phase will compensate for the lack of impingement turbulence and the fact that the hotplate is partially covered by thermally insular gas is not clear.
yes, however, this is only true IF the heat is not being removed by the radiator. IF the heat is removed, the gas condense back into a liquid and reduces pressure and boiling point drops again. as mentioned, the speed of the pump is dependent on the heat generated by the cpu. high heat = more boiling of the liquid. as long as the radiator is capable of dissipating the heat produced, the pressure(boiling point) will not exceed the designed specs of the system. you seem to think the entire system is filled with water. this can not be true for the science to work in this type of system. the radiator portion of the setup is a vapor chamber where hot gas is converted back into liquid. so the total amount of liquid involved is simply the amount that is contained in the hotplate as well as the tubes. (just like heatpipes are not filled entirely with liquid either
tongue.gif
) the radiator portion of the system must remain a gas expansion/cooling chamber that becomes pressurized. the amount of pressure depends on how quickly the heat can be dissipated and converted back into liquid state. so the total amount of liquid in the system is not as much as you think if the science of the system is to work properly.

as for why i'm defending this.. i'm talking about the science behind the product. i can care less if silverstone or any other manufacture made the product
biggrin.gif
as Doyll has pointed out, captherm has a product that uses similar principle. why are the product not on the market yet? construction cost and generally accepted practice. just as heatpipes are once rarely used as heatsinks the need for that level of technology has made the market viable for the product. it's more expensive to make a heatsink with heatpipe and fins so they used to use large chunks of aluminum or copper since it's cheaper. but now the thermal demands requires more effective heat removal, heatpipes have become mainstream. the silverstone product simply banks on the acceptance/popularity of having a radiator in a system as well as the demand for quieter cooling system in the marketplace. whether the system is in fact quieter is yet to be seen (boiling liquid can be quite loud) what it comes down to is cost. in the past people are resistant to buying a heatsink that cost over $100. however as the popularity of the H100i and similar product grows, the cost barrier has been broken. a large radiator based cooler with its associated cost becomes a viable product for the marketplace. this is the reason why you haven't seen this type of product on the market till now, it's an economic reason, not because of the science.

the vid is shown as a prototype. they may not release the product depending on how well received the product is. however just because something is not a consumer product, doesn't mean the science behind the product doesn't work
smile.gif
consumers are inherently fickle and resistant to change. it's like the BTX specification intel tried to push out years back. it's a MUCH better thermal design then ATX, however the market decided that ATX will be here to stay
biggrin.gif
too bad because i preferred the BTX design since it allowed for more efficient airflow/cooling.
 
1 - 20 of 60 Posts
Top