New Waterblock Design Uses Porous Copper

H

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
Here's an interesting way of making a waterblock. This company has developed a unique waterblock that use porous copper instead of a traditional milled micro-channel design. The company claims this new design is up to ten times more effective cooling than waterblocks currently on the market. Thanks to Edward C. for the linkage!

VersarienCu™ is the result of innovative metallurgy research performed at the University of Liverpool’s Department of Engineering. It allows the creation of metallic materials with fine, open, interconnected pores which emulate structures that are commonly found in nature. This structural form permits thermal transfer levels that surpass anything that was previously possible.
Click to expand...
 
Makes sense, more surface area for heat to transfer through.. the pores would have to be very small though as to be filled in the TIM between the CPU and block. If its anything like the picture, though - that is just asking for air bubbles to be trapped in it and air is about the worst conductor of heat.
 
yeah, great heat transfer with a fluid, looks like piss poor heat transfer with a flat plane of metal like a heat spreader.

I'm thinking any advantage might be in a radiator, not the waterblock.
 
raz-0 said:
yeah, great heat transfer with a fluid, looks like piss poor heat transfer with a flat plane of metal like a heat spreader.
Click to expand...
...you do realize the side facing the CPU heat spreader would likely be milled flat, right? :rolleyes:
 
Unknown-One said:
...you do realize the side facing the CPU heat spreader would likely be milled flat, right? :rolleyes:
Click to expand...

So you are going to put liquid directly in contact with the CPU? Milling doesn't get rid of the holes you know. Also as the previous poster said, it doesn't look like it would prevent air bubbles.
 
raz-0 said:
So you are going to put liquid directly in contact with the CPU? Milling doesn't get rid of the holes you know. Also as the previous poster said, it doesn't look like it would prevent air bubbles.
Click to expand...

They would probably mill it flat and then solder it to a standard flat piece of copper as the base.

Looks interesting but I highly doubt 10x effectiveness. Modern waterblocks are already pretty damn effective.
 
leeleatherwood said:
They would probably mill it flat and then solder it to a standard flat piece of copper as the base.

Looks interesting but I highly doubt 10x effectiveness. Modern waterblocks are already pretty damn effective.
Click to expand...

If you did that, I suspect to get the flow through you want would require the system to run at greater pressures, and then you have to care about plumbing. I wonder how it would do for a phase change based heat sink. Now instead of tubes, you could have like a big foam filled T with lots of fins mounted to it and a very large internal surface area.
 
Unless there is another form of conduction between the porous copper it has little change of benefiting the normal HSF.
 
raz-0 said:
So you are going to put liquid directly in contact with the CPU? Milling doesn't get rid of the holes you know. Also as the previous poster said, it doesn't look like it would prevent air bubbles.
Click to expand...
Again, not a hard problem to solve.

Simply dip one side in a molten metal bath (probably nickel) to close it off, then mill it flat. :rolleyes:
 
The bottom could simply be sealed with molten copper... Or the treatment could be applied to the top half of an ingot, and sliced that way.
 
mmm... not so much. the melting temperature of nickel is 600 degrees higher than that of copper. Best just to melt a block on one side and mill the copper once cooled.
 
IMHO, make a huge steel form. Pour molten copper 1/4" of the way into the form.
Allow cooling.
Fill the remaining, say, 3/4" of the square form with the powder mixture and compact.
Fire @ 1100c as normal.
Slough off edges (iron contamination, if present) and mill top and bottom.

Huge sheets of this porous copper, capped with a flat base.
 
Arcygenical said:
The bottom could simply be sealed with molten copper... Or the treatment could be applied to the top half of an ingot, and sliced that way.
Click to expand...

It appears to be formed by a deliberately flawed version of metal injection molding. I don't know that that would work.
 
Pieter3dnow said:
Unless there is another form of conduction between the porous copper it has little change of benefiting the normal HSF.
Click to expand...

*sigh*

PHASE CHANGE. You know all the heat pipe HSFs out there. There's a liquid in them that turns to vapor and then condenses? I'm saying that at the very least if you replaced the empty space in the tubes at the top with an interference fit slug of this porous stuff (assuming that the porosity is enough for the vapor to migrate through the material), you could greatly increase the surface area of conductive material the vapor is exposed to. I don't know that with sticking to a tube design you wouldn't kill that advantage by only having so much surface areas shared with the fins that conduct the heat away from the tube.

However, I could see something where you build essentially a hollow bar and plate T filled with this copper stuff and the phase change liquid, and perhaps if the maximally porous version can get good airflow with a fan on it, adhere it to the outside wiht adhesive.

Heck, even without vapor phase, if the really porous stuff can have air pushed through it by a quiet(ish) fan, you just used a reasonably thermally conductive adhesive to secure it to a fat copper plate and put a fan on top.

I see it ahveing great applications for it increasing surface area exposure to liquid and air. BUt in water cooling, I suspect the limiting factor at the die side is how much water you can shove past it, which this doesn't look like it facilitates.
 
So you are going to put liquid directly in contact with the CPU?
Click to expand...
Direct die water/liquid cooling has been done so it's not necessarily a problem, but in the case as some have suggested it would not be necessary.

Unless there is another form of conduction between the porous copper it has little change of benefiting the normal HSF.
Click to expand...

This manufacturing technique is not intended for a standard HSF, it's for a waterblock.
http://www.versarien.com/versarienpc/

As for actual performance and measuring effectiveness, you always have to keep in mind the basics. If you are not using phase change, peltiers, water chillers, or some other form of active cooling like LN2, then in the end water cooling can not take you below the ambient temperature of the surrounding environment.

Your basic water cooling system can't cool a system below room temp. If you bury a heat exchanger below the frost line you can't cool the system below the temperature of the surrounding ground at that depth, etc. Therefor, there is a physical limitation or threshold that you can't beat. With this in mind if your room, where your system is, has a nice comfortable temperature of say 75F, then you will not get your watercooled components below this temp without changing the conditions with active cooling. Cool your room better, lower the ambient, the system runs cooler but your electric bill is higher. Everything has a cost.
 
flow shouldn't be problem , if anyone remembers back in old carb days they used to sell this type of porous fuel filter instead of the paper ones .
They worked real well and better IMO to the paper ones, at least as far as long life .
 
I don't even understand where the "flow thing is even coming into this, it's not like the fluid is being pumped through the material, it's being pumped against the surface of it.
 
I am surprised nobody has mentioned it yet, but wouldn't the small pores tend to clog up pretty easily and quickly with any oxidation caused and/or contamination in the cooling liquid.

Seems like a fairly bad choice for water cooling, even with glycol mixed in to prevent corrosion and growth of organisms.

You would know what I mean if you have ever seen what happens to an automobile radiator after a while.

The easiest way to drop temps in water cooling systems is to use a surface tension reducer. This makes the water be able to transfer more heat.

Works great in automobile applications, as well as pretty much any other water cooled systems.

One version is called Water Wetter.
 
cyclone3d said:
I am surprised nobody has mentioned it yet, but wouldn't the small pores tend to clog up pretty easily and quickly with any oxidation caused and/or contamination in the cooling liquid.

Seems like a fairly bad choice for water cooling, even with glycol mixed in to prevent corrosion and growth of organisms.

You would know what I mean if you have ever seen what happens to an automobile radiator after a while.

The easiest way to drop temps in water cooling systems is to use a surface tension reducer. This makes the water be able to transfer more heat.

Works great in automobile applications, as well as pretty much any other water cooled systems.

One version is called Water Wetter.
Click to expand...

Except a car radiator and PC watercooling are thermodynamically two entirely different animals. Wetters really aren't needed in PC setups as the coolant doesn't get anywhere close to thermally saturating the coolant unless there's a pump failure...and in that last case, your problem is not the coolant.

Might be neat to try...but long term I see break down/off of the porous copper being a problem.
 
First off, water isn't really the best coolant to use, there are better choices if more expensive. These other coolants do not have the downside of water or water and additive mixtures.
 
Except a car radiator and PC watercooling are thermodynamically two entirely different animals.
Click to expand...

You know this really isn't a correct statement. I know what you mean, the problem is the way you are saying it. They are the exact same principles applied given a different set of conditions.

Same animals, different zoos :D
 
How does galvanic corrosion enter into it?
Click to expand...

I think he is suggesting that many conventional waterblocks might "suffer" from this effect already.

Actually, if you are just dipping the base that mates with the processor's heat spreader, water never comes into contact with the nickel anyway so galvanic corrosion doesn't come into the equation at all, at least not from the nickel.

Some of you guys got me wondering if you have ever done any watercooling to begin with.
 
Actually, if you are just dipping the base that mates with the processor's heat spreader, water never comes into contact with the nickel anyway so galvanic corrosion doesn't come into the equation at all, at least not from the nickel.
Click to expand...

I must amend this, if the material is so porous that water can move through the copper material and reach the nickel base then yes, it could cause an issue.
 
I would think after a while the holes would plug up with crud.
unless u could keep the water super clean all the time. distilled plus something to kill bacteria, but still I could see these holes getting plugged or partially plugged. Am I crazy to assume?
 
People have experimented with metal foams for a while as heat exchangers, the dP across them tends to be huge, and I'd kinda assume this will be similar.

Two-phase micro-channels is where it's at for direct contact->liquid heat exchangers. Been around in research since the early 00's, but finally started being sold in systems in I want to say 2011. Of course you need to make sure your system is sealed better than normal liquid cooling stuff. Well, while at it, do a reduced pressure system and set the boiling point of the fluid at what temp you want the system to remain at :eek:.
 
Unknown-One said:
How does galvanic corrosion enter into it? :confused:

I suggested dipping it in nickel, which is commonly used as a plating to prevent galvanic corrosion.
Click to expand...

ANY two metals will corrode if they are in electrical contact AND electrolytic contact... Aka, both need to contact the fluid, and eachother electrically.

If a nickle plating chips, the entire coating will fail over time.
 
lcpiper said:
Actually, if you are just dipping the base that mates with the processor's heat spreader, water never comes into contact with the nickel anyway so galvanic corrosion doesn't come into the equation at all, at least not from the nickel.
Click to expand...

Well that doesn't make any sense... If the water doesn't come in contact with the nickle, well, the material would be waterproofed, and wouldn't need coating/dipping/plating at all!
 
coating/dipping/plating
Click to expand...

If the material will not let water pass through it to the base, then the base still needs resurfacing to achieve a better mate with the heat spreader of the processor.

Of course if material is that porous it would also leak coolant out of the sides as well, leaky leaky.

Something must be sealed in some manner as you just can't let them leaky like the fuel cells of the BlackBird. Unless you plan to fly your waterblock at mach 2+ :D

That is, unless you want to go all out with immersion cooling.
 
lcpiper said:
If the material will not let water pass through it to the base, then the base still needs resurfacing to achieve a better mate with the heat spreader of the processor.
Click to expand...

Oh right, I forgot what we were originally discussing!
 
Suppose you could also just dip it in more copper, or create a copper "dish" with the sponge material inside of it, and heat the assembly until it fuses.
 
Two-phase micro-channels is where it's at for direct contact->liquid heat exchangers. Been around in research since the early 00's, but finally started being sold in systems in I want to say 2011. Of course you need to make sure your system is sealed better than normal liquid cooling stuff. Well, while at it, do a reduced pressure system and set the boiling point of the fluid at what temp you want the system to remain at .
Click to expand...

Blakroth, this makes me ask if you are an engineer of some sort, did some work with thermal dynamics maybe?
 
lcpiper said:
Blakroth, this makes me ask if you are an engineer of some sort, did some work with thermal dynamics maybe?
Click to expand...

Grad student at Mizzou focusing on thermal/fluids area of Mechanical Engineering, my undergrad research was in two-phase interconnected microchannels way back in like 07.
 
Since the material is made by creating a mixture of copper and other materials and part of the process involves subjecting the material to 1000 degrees Celsius, I think it's safe to say something is being melted pretty thoroughly. Just have a copper block with a milled inset for the material, melt the material to create the porous core, set it into the milled inset, reheat and add a little pressure maybe to create a thermal bond between the core and the rest of the block.
 
Cool, well since the water in a typicla watercooling system never reaches boiling temperature it is not going though a phase change in a typical system.


do a reduced pressure system and set the boiling point of the fluid at what temp you want the system to remain at .
Click to expand...

I have never heard of an approach like this applied to keeping a component in a temperature range of say 32 to 45 Celsius.

I am open to hearing possibilities.

I am far from a thermal engineer.
 
lcpiper said:
I have never heard of an approach like this applied to keeping a component in a temperature range of say 32 to 45 Celsius.

I am open to hearing possibilities.

I am far from a thermal engineer.
Click to expand...

Well there's a little bit of annoyance to it, since it's not really the surface temperature but the energy output that makes the coolant boil (since it's just raising the temp of the coolant), but if you have the coolant boil at say 37 degrees, the surface won't get much warmer than that. With a device like a cpu though you run into problems, since you design it to boil at the max output really, so it will be single phase when the cpu isn't running at it's max power, so the temperature of the fluid won't rise enough to cause it to boil (we kinda call this flooding the surface in this case). Toss a temperature controller connected to a pump and it's a different story of course.
 
You must log in or register to reply here.
Back
Top