Overclock.net banner

[Hexus] Silent Power copper foam cooled PC prototype details and pictures

2K views 33 replies 22 participants last post by  145252 
#1 ·
Quote:
We covered the launch of the fanless 'high-end' copper foam cooled PC project by Silent Power at the end of last month. Now this fascinating crowdfunded project has been fleshed out with actual photos of the prototype (rather than renders), more details about how the heat buffer works with the copper foam and news that a DisplayPort connector has been added to the PC's specifications.
Source

I've been looking for the most silent and powerful rig for ages but this is really interesting. The less fans in a PC, the better.
 
#3 ·
I still believe that aluminum fins would perform better, look nicer, last longer, and easier to maintain.
 
#6 ·
Quote:
Originally Posted by Mad Pistol View Post

This actually makes a lot of sense. There are tons of heat-exchange points. It's porous which means air should move through it easily.
1) The conduction along the cooper cross-section is basically through thin metal wires.... This will result in a high thermal resistance so the heat won't be able to effective reach the outer layers of the foam.
2) There is no movement of air through convection within the foam. The warm air will have a harder time getting out.

A vapor chamber base with aluminum fins will probably perform way better.... I mean look at the tens of thousands of industrial designs already in existence. I could be wrong but I doubt it here.

How do they plan to bond the copper plate to foam? This is important for conduction.
 
#8 ·
Quote:
Originally Posted by phill1978 View Post

cant they at least put one fan in? just to be safe. its not like a 1000rpm 140mm fan is audible from 10cm away in a closed case
If it's engineered to be passive then it should be able run passively in the stated environment.

What's there to be safe about? CPU and GPU will throttle and shutdown.
 
#9 ·
Conductive heat loss through a cylinder :
q = 2*Ď€*k*(ti - to) / L

where
q = heat transferred per unit time per unit length of cylinder or pipe (W/m, Btu/hr ft)
k = thermal conductivity of the material (W/m.K or W/m oC, Btu/(hr oF ft2/ft))
to = temperature outside cylinder (K or oC, oF)
ti = temperature inside cylinder (K or oC, oF)
L = lengh of cylinder

The surface area maybe large.... but the cross-section of the wires are small. Which is better at cooling a 100w load.... a strand of 100ft of 28AWG wire or a 5ft of 8AWG wire?

The issue with thin wires is that the heat may not be able to effectively travel the length of the wire to gain the benefits of the increased surface area.
 
#10 ·
Quote:
Originally Posted by DuckieHo View Post

1) The conduction along the cooper cross-section is basically through thin metal wires.... This will result in a high thermal resistance so the heat won't be able to effective reach the outer layers of the foam.
2) There is no movement of air through convection within the foam. The warm air will have a harder time getting out.

A vapor chamber base with aluminum fins will probably perform way better.... I mean look at the tens of thousands of industrial designs already in existence. I could be wrong but I doubt it here.

How do they plan to bond the copper plate to foam? This is important for conduction.
The reason I think this will work is because given the metal type, copper is a better conductor of heat than aluminum.

http://www.engineeringtoolbox.com/thermal-conductivity-metals-d_858.html

Copper has a natural convection property where heat will move from the area of highest heat to areas that are cooler. This is why copper is used in heatpipes. However, copper is too heavy to be used effectively in heatsinks. That's why most are aluminum.

Because the copper is porous in this design, it should actually have far more surface area than a conventional array of aluminum fins. I think that's the reason it supposedly works so well.
 
#11 ·
Quote:
Originally Posted by Mad Pistol View Post

Quote:
Originally Posted by DuckieHo View Post

1) The conduction along the cooper cross-section is basically through thin metal wires.... This will result in a high thermal resistance so the heat won't be able to effective reach the outer layers of the foam.
2) There is no movement of air through convection within the foam. The warm air will have a harder time getting out.

A vapor chamber base with aluminum fins will probably perform way better.... I mean look at the tens of thousands of industrial designs already in existence. I could be wrong but I doubt it here.

How do they plan to bond the copper plate to foam? This is important for conduction.
The reason I think this will work is because given the metal type, copper is a better conductor of heat than aluminum.

http://www.engineeringtoolbox.com/thermal-conductivity-metals-d_858.html

Copper has a natural convection property where heat will move from the area of highest heat to areas that are cooler. This is why copper is used in heatpipes. However, copper is too heavy to be used effectively in heatsinks. That's why most are aluminum.

Because the copper is porous in this design, it should actually have far more surface area than a conventional array of aluminum fins. I think that's the reason it supposedly works so well.
The problem, though, is that not much air will travel through the "foam". It looks more like copper wool to me, but anyway. It's going to act much like a solid with regards to airflow, with the heat exchange mostly taking place along the outer surface of the foam.
 
#12 ·
Quote:
Originally Posted by Mad Pistol View Post

The reason I think this will work is because given the metal type, copper is a better conductor of heat than aluminum.

http://www.engineeringtoolbox.com/thermal-conductivity-metals-d_858.html
The thermal conductivity copper has nothing to do with the limitations of physics. Note the equation provided already includes thermal conduction. It's the cross-sectional area that will be the problem.

Quote:
Originally Posted by Mad Pistol View Post

Copper has a natural convection property where heat will move from the area of highest heat to areas that are cooler.
It's "conduction", not "convection" since that applies mostly to fluids. Virtually all materials conduct heat.

The issue is the geometry. Again... which would cool a 100w load better? A single 100ft 28AWG wire or a single 1ft 8AWG wire?

Quote:
Originally Posted by Mad Pistol View Post

This is why copper is used in heatpipes. However, copper is too heavy to be used effectively in heatsinks. That's why most are aluminum.
No, thermal conductivity of the two metals don't have that much to do with heatpipes. Heapipes use phase change and convection to transport heat... not conduction. The only conduction would be from the source, through the metal.... then from the gas, through the metal to radiator.

While weight is a reason for few copper heatsinks, the bigger reason is cost. Copper costs 3 times as much. Remember the Thermalright TRUE Copper? The regular TRUE was $75 while the copper version was $125.

Futhermore, the thermal conductivity of the metal is only part of the factors of a good heatsink. Design is far more important. The temp difference between the two TRUE was only 2C.
Quote:
Originally Posted by Mad Pistol View Post

Because the copper is porous in this design, it should actually have far more surface area than a conventional array of aluminum fins. I think that's the reason it supposedly works so well.
The copper itself isn't porous. It's more like a copper bundles.... heat still has to transverse the length of the wires.

Again, the issue isn't surface area.... My concern is the conduction of the heat along the copper threads.
 
#14 ·
Quote:
Originally Posted by DuckieHo View Post

-snip- Again... which would cool a 100w load better? A single 100ft 28AWG wire or a single 1ft 8AWG wire?
That question isn't really analogous to the topic at hand. The foam has superior surface area compared to traditional heat sinks, but is not created from a single strand of copper (i.e. the heat does not need to travel via a single cross-section to reach the outer parts of the foam). There are tens of thousands of inter-connect points between the tiny "wires" where they come in contact with each other and so heat dissipation is aided. instead of being impeded by thermal resistance like it would be in a thick block of copper. the cross-sectional surface area of the "wires" that make up the foam are irrelevant because the wires are all in contact and can transfer heat between one another. To that end, the question should be:

which would cool a 100w load better? 1ft of stranded 8 AWG wire or a 1ft length of solid conductor 8 AWG wire?

it's the same principle as "more heat pipes = more efficient heat sink" (i.e. more potential for dissipation of heat by increasing number of conduction routes)

of course, the foam is also not created of singular wires at all. it is a single, porous mass.
 
#15 ·
Looks like something that is used to scrub dishes
biggrin.gif
 
#16 ·
Quote:
Originally Posted by jchap1590 View Post

That question isn't really analogous to the topic at hand. The foam has superior surface area compared to traditional heat sinks, but is not created from a single strand of copper (i.e. the heat does not need to travel via a single cross-section to reach the outer parts of the foam). There are tens of thousands of inter-connect points between the tiny "wires" where they come in contact with each other and so heat dissipation is aided. instead of being impeded by thermal resistance like it would be in a thick block of copper. the cross-sectional surface area of the "wires" that make up the foam are irrelevant because the wires are all in contact and can transfer heat between one another. To that end, the question should be:

which would cool a 100w load better? 1ft of stranded 8 AWG wire or a 1ft length of solid conductor 8 AWG wire?

it's the same principle as "more heat pipes = more efficient heat sink" (i.e. more potential for dissipation of heat by increasing number of conduction routes)

of course, the foam is also not created of singular wires at all. it is a single, porous mass.
Of course, but think about modeling this design. The wires are not all connected though. To effectively transfer heat, they must be bonded to each other. Anything less than bonded wires transmits poorly. If they are too sparse, then they would have issues transporting the heat. However, if they are too dense, then they would have issues with radiating the heat to the air.

"which would cool a 100w load better? 1ft of stranded 28 AWG wire or a 1ft length of solid conductor 8 AWG wire?".... if the outer dimension is the same, the 8AWG. Each strand within the bundled of 28AWG will have an air boundary layer.... a low thermal conductivity layer.

Note that heatpipes are not used for cooling in themselves.... they are used to transport energy to cooling assemblies.

Take a piece of aluminum wool and place a lighter under it.
Take a piece of solid aluminum that is the same size and place a lighter under it.
How long does each take to get warm?

http://www.ergaerospace.com/Copper-properties.htm
Duocel® Copper Foam: Thermal conductivity = 10.1 W/(m*C)

http://www.engineeringtoolbox.com/thermal-conductivity-d_429.html
Copper: Thermal conductivity = 400 W/(m*C)

Again.... my thought is..... While there might be more surface area, the vastly lower thermal conductivity will bottleneck heat transportation and therefore make ineffective.

Found a paper on the topic... haven't had time to read it yet: http://www.virginia.edu/ms/research/wadley/Documents/Publications/The%20effects%20of%20topology%20upon%20fluid-flow%20and%20heat-transfer.pdf
 
#17 ·
Quote:
Originally Posted by DuckieHo View Post

Of course, but think about modeling this design. The wires are not all connected though. To effectively transfer heat, they must be bonded to each other. Anything less than bonded wires transmits poorly. If they are too sparse, then they would have issues transporting the heat. However, if they are too dense, then they would have issues with radiating the heat to the air.

"which would cool a 100w load better? 1ft of stranded 28 AWG wire or a 1ft length of solid conductor 8 AWG wire?".... if the outer dimension is the same, the 8AWG. Each strand within the bundled of 28AWG will have an air boundary layer.... a low thermal conductivity layer.
Of course, I was using "wires" only in response to other comments.

Marketing it as "copper foam" gives the impression that it is a single structure. In that regard, it is only similar to copper wool in appearance, because copper wool is comprised of strands whereas this copper foam sounds to be a single, porous mass.

It seems plausible in concept, but real world performance figures are needed to draw a more definite conclusion. Now that a prototype has been manufactured, we're one step closer to seeing those results.
 
#21 ·
Quote:
Originally Posted by DuckieHo View Post

Of course, but think about modeling this design. The wires are not all connected though. To effectively transfer heat, they must be bonded to each other. Anything less than bonded wires transmits poorly. If they are too sparse, then they would have issues transporting the heat. However, if they are too dense, then they would have issues with radiating the heat to the air.

"which would cool a 100w load better? 1ft of stranded 28 AWG wire or a 1ft length of solid conductor 8 AWG wire?".... if the outer dimension is the same, the 8AWG. Each strand within the bundled of 28AWG will have an air boundary layer.... a low thermal conductivity layer.

Again.... my thought is..... While there might be more surface area, the vastly lower thermal conductivity will bottleneck heat transportation and therefore make ineffective.
It's the convection that is reduced. Copper has the same thermal conductivity regardless of it's arrangement, it's a physical property. In the case of the "foam", the limiting factor is the heat transfer coefficient of air.

Heat transfer to the air is defined by Q= h*A*dT. The area is huge, so the biggest concern if this is passive is the convective heat transfer coefficient. It wont be 0, because the air will be changing densities and diffusing in and out of the foam, but it wont be anywhere near the magnitude of forced convection.

As for the point of the wires being bonded, if it's a true metal foam(https://en.wikipedia.org/wiki/Metal_foam), it wouldn't need to be bonded, since every copper strand is connected to multiple other copper strands. You just get a lot of void space for air permeation.

I'm interested in seeing how this performs.
 
#23 ·
Source
Quote:
The top of what its creators are claiming is the smallest high-end PC in the world looks like something you'd use to scrub pots and pans, but the intricate copper mesh provides a remarkable amount of surface area to dissipate heat channeled directly from the CPU and GPU-supposedly up to 500 times greater than a traditional heatsink using fins.
 
#25 ·
#26 ·
Quote:
Originally Posted by HybridCore View Post

I've seen this 2 times already in the news section. Please search the forum before you post a news article. Thank you.

Original thread: http://www.overclock.net/t/1499110/hexus-copper-foam-cooled-fanless-pc-seeks-crowdfunding
Repost of the original: http://www.overclock.net/t/1504517/hexus-silent-power-copper-foam-cooled-pc-prototype-details-and-pictures
Go search my title and you won't find those articles. Stop assuming stuff.

Also its a different article.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top