Wow! Check out this SUPER-Enthusiast ATI mobo by Sapphire
- Thread starter OICAspork
- Start date
Technically, what I'm describing is that just because one has a great big hunk of copper doesn't mean one gets superior performance out of it.
I really think that those that want to discuss the value of what material is used in the top of a water block should be posting in hardware....that being said, there is no difference between any material that makes up the non contact area of a waterblock in terms of cooling.
Well you are probably right in that the discussion should be else where. But copper does conduct heat more quickly than many other materials as was shown by the heat conductance. And I relaize that plain water is not used, but they did not have a listing for water+redline, or anything. Maybe antifreeze has its conductance listed somewhere and I bet it would be close.
You may be right in that the top material is not significatnly important, but to say it makes no difference is incorect. In a design like that swiftech waterblock I showed, if the top is welded to the bottom, instead of having a rubber gasket then the heat will conduct to the top and into the liquid.
Still I think a heat pipe on a giant scale is a neat idea, if I have money to burn in the future I will do it just for fun, maybe isopropyl alcohol would be better though just b/c it is easy to get ones hands on and not as explosive as many other low boiling solvents. One might simply end up with water in the the water block though unless pure isopropyl was used, in chich case we are once again facing a scarcity of orignal product, and/or a hazard...
Another random idea pops unbidden into my head, computer moonshine
edit: I found the thermal conductance for antifreeze.
Mercury 8.3
Water 0.67
3M Flourinert FC-43 0.66
Methanol 0.25
Glycol, Antifreeze 0.25
Ethanol 0.14
Liquid Nitrogen 0.14
http://www.koolance.com/technical/cooling101/002.html
So it is actually significantly worse, indicating that less antifreeze not only will increase the flow rater by decreasing viscosity, but also increase the rate of thermal conduction...
You may be right in that the top material is not significatnly important, but to say it makes no difference is incorect. In a design like that swiftech waterblock I showed, if the top is welded to the bottom, instead of having a rubber gasket then the heat will conduct to the top and into the liquid.
Still I think a heat pipe on a giant scale is a neat idea, if I have money to burn in the future I will do it just for fun, maybe isopropyl alcohol would be better though just b/c it is easy to get ones hands on and not as explosive as many other low boiling solvents. One might simply end up with water in the the water block though unless pure isopropyl was used, in chich case we are once again facing a scarcity of orignal product, and/or a hazard...
Another random idea pops unbidden into my head, computer moonshine
edit: I found the thermal conductance for antifreeze.
Mercury 8.3
Water 0.67
3M Flourinert FC-43 0.66
Methanol 0.25
Glycol, Antifreeze 0.25
Ethanol 0.14
Liquid Nitrogen 0.14
http://www.koolance.com/technical/cooling101/002.html
So it is actually significantly worse, indicating that less antifreeze not only will increase the flow rater by decreasing viscosity, but also increase the rate of thermal conduction...
Sxotty, the difference in top materials is really insignificant. And while just plain distilled water is definately a better conductor than antifreeze....believe me you don't want to use just water..... one of the reasons for additives, like glycol is to increase the boiling point....hell, you'd hate to have water boil in your cooling system! You might not have a system - both cooling and computer - left! Like just about everthing in life, it's all about tradeoffs.....
Yeah, don't overdo the coolant mix. Plain water is way more effective, but it is naturally harsh and can have evil things in it (water is the universal solvent, after all). So, a little engine coolant is necessary unfortunately.
I'm still waiting for the "mad OC'ers" to appear from the coming mercury coolant craze!
I also find it hard to believe that copper wouldn't be far better than aluminum for that water block. At least for the surface in contact with the CPU. You want to get that heat out as best as possible. Hell, pop off the heatspreader too if you're a real man. I bet you get another few degrees off that CPU without that POS aluminum heatspreader with substandard goo on the other side.
I'm still waiting for the "mad OC'ers" to appear from the coming mercury coolant craze!
I also find it hard to believe that copper wouldn't be far better than aluminum for that water block. At least for the surface in contact with the CPU. You want to get that heat out as best as possible. Hell, pop off the heatspreader too if you're a real man
. I bet you get another few degrees off that CPU without that POS aluminum heatspreader with substandard goo on the other side.Mintmaster
Veteran
Sxotty, there's a very simple test to see if you're right.
If you have an all copper water cooling block on your CPU, is the top warm or cold?
We know the bottom is warm, as you can just check the temp of the CPU. Since the water is cool at the interface, there is a temp. gradient on the part of the block between the water and the CPU. If the top is cold, i.e. almost the same temp as the water, then there is no temp. gradient.
Heat flow is proportional to temperature gradient (can't remember what the law is called). I assume the answer to my question is "cold" (unless you have very low water flow), and thus there is very little heat dissipation around the sides and from the top of the block.
You can quote all the numbers you want, but I guarantee you that it makes very little (read: not likely measurable) difference that the top isn't copper.
If you have an all copper water cooling block on your CPU, is the top warm or cold?
We know the bottom is warm, as you can just check the temp of the CPU. Since the water is cool at the interface, there is a temp. gradient on the part of the block between the water and the CPU. If the top is cold, i.e. almost the same temp as the water, then there is no temp. gradient.
Heat flow is proportional to temperature gradient (can't remember what the law is called). I assume the answer to my question is "cold" (unless you have very low water flow), and thus there is very little heat dissipation around the sides and from the top of the block.
You can quote all the numbers you want, but I guarantee you that it makes very little (read: not likely measurable) difference that the top isn't copper.
It doesn't matter anyway. And no I don't have an all copper waterblock, mine is copper and aluminum, but there is a gasket in between so it is irellevant. I was just stating what is common sense, and I already agreed that the difference may be very minor but it should still theoretically be their. We are talking about completely different reference frames though. I am thinking about differences detecatable with a bomb calorimiter, and you are talking about degrees C.
Furthermore I quit talking about that awhile ago since it is irrelevant to the goal at hand, which is making significant decrease in temperature, but continue to carrry on the conversations amongs yourselves if you insist. I was talking about heat pipes, and ethanol, mercury and other non traditional things.
Like I said I think that a huge heat pipe type setup oculd be very effective, but the pressure gradients might be too large, have to use some sort of braided tubing like on gas lines or something...
Furthermore I quit talking about that awhile ago since it is irrelevant to the goal at hand, which is making significant decrease in temperature, but continue to carrry on the conversations amongs yourselves if you insist. I was talking about heat pipes, and ethanol, mercury and other non traditional things.
Like I said I think that a huge heat pipe type setup oculd be very effective, but the pressure gradients might be too large, have to use some sort of braided tubing like on gas lines or something...
So you're saying that having copper instead of aluminum conducting the heat away from the core (as in in contact with the die) is not a significant improvement?
Why do AMD and Intel make OEM heatsinks with copper bases then? You can't tell me the entire copper heatsink craze is a prank to make money. I also have a hard time believing that water blocks are so much different than heatsinks. Instead of air, they use water as a medium. They still need effective, efficient surface area.
On another note, it's funny how many video cards have fake copper heatsinks. I couldn't believe what I was looking at when I saw a friend's MSI 6600 GT. The damned heatsink was copper anodized aluminum.
Does anyone use mercury or ethanol? Lol. That would be interesting.
Why do AMD and Intel make OEM heatsinks with copper bases then? You can't tell me the entire copper heatsink craze is a prank to make money. I also have a hard time believing that water blocks are so much different than heatsinks. Instead of air, they use water as a medium. They still need effective, efficient surface area.
On another note, it's funny how many video cards have fake copper heatsinks. I couldn't believe what I was looking at when I saw a friend's MSI 6600 GT. The damned heatsink was copper anodized aluminum.
Does anyone use mercury or ethanol? Lol. That would be interesting.
No swaye you misunderstand.
I wouldn't know if a pure copper waterblock would be significantly better b/c mine isn't pure copper. The base is copper the top is aluminum. And I have plenty/too much antifreeze in there... I had so much that it started to have little white solids floating around
I wouldn't know if a pure copper waterblock would be significantly better b/c mine isn't pure copper. The base is copper the top is aluminum. And I have plenty/too much antifreeze in there... I had so much that it started to have little white solids floating around
IIRC the heat transfer formula (for a brick) is basically:
H=KxAx(T2-T1)/L
where:
H= heat conducted
K= constant dependent on material (for SI system Cu=.92 Al=.48 )
A= section area
T2 and T1= temperatures in the different interfaces
L= length
Now the important part is that copper can transfer twice as much the amount of heat (see the K value) for the same other values (A,T2,T1,L)
Other K values: iron .1 glass 0.02 wood .002 air .00057
This is what I remenber.
I think copper when used properlly can help conduct the heat a lot.
H=KxAx(T2-T1)/L
where:
H= heat conducted
K= constant dependent on material (for SI system Cu=.92 Al=.48 )
A= section area
T2 and T1= temperatures in the different interfaces
L= length
Now the important part is that copper can transfer twice as much the amount of heat (see the K value) for the same other values (A,T2,T1,L)
Other K values: iron .1 glass 0.02 wood .002 air .00057
This is what I remenber.
I think copper when used properlly can help conduct the heat a lot.