Buddy boy dont put your foot in your mouth. The reason they use aluminum is that its cheeper. Simple. They just engeenere it to be very good with alu. But if you engeeneer copper rite it far outperforms aluminum. BTW industry leaders that built servers that have up times of years use copper heatsinks. THey use low profile sinks with allot of thin fins and filters on the cases so they dont clog up. Copper displaces more heat than aluminum. Its in the cheemical properties section in your chem book. I sugest you look it up.
Emon...Thanks for your comment... Can you perhaps give me the details of the heat-sink comparison you are referring to? It would be very interesting to see how the test was conducted. I will then comment accordingly. When you say heat-sink, do you mean: 1. Heat-sink as in finned extended-surface machined/fabricated sitting on top of a CPU? 2. Water-block cooler sitting on top of CPU? 3. Finned radiator-type cooler? Each one has entirely different design rules. adTherm <www.adtherm.com>
IMHO, there has been lots & lots of hype about which material performs better - Al, or Cu... lots of passion... I have performed a number of studies of 'extended fins' in terms of fundamental Heat-Transfer principles... The results have been rather interesting, to say the least... Oh, in terms of background - I am a Consulting Engineer, by profession - studying towards my Ph.D. I have been deeply involved in the debate which raged in the Automotive Industry for many years... copper, or aluminium... Aluminium seems to have won there... My website will perhaps to elaborate in more detail. Computing is my passion, & I am thinking of moving into the Computer Cooling arena - in terms of product manufacture... Any feedback would be most welcome... positive, or otherwise... adTherm <www.adtherm.com>
Uh, no **** huh? If you read the post I was replying to, you'd probably have said something completely different. adTherm, I'll try to find a link soon. I remember seeing several comparisons from a manufacturer that had two identicle heatsinks, one copper and one aluminum, just different models. Off hand I know Zalman's radial design heatsink things...whatever the hell you call those, are in aluminum and copper. Well, there's one with a copper core and the rest aluminum, not sure if there is an all aluminum. However it's not the design you're talking about, and since it's so god damn huge there's probably negligible performance difference because both can cool just as effectively (but I have never looked at benchmarks for these, only guessing). I think there was an alu vs cu thread on Ars Technica that had a link to the article I'm thinking of, I'll check there. And when I say heatsink, I mean like this or this. Now I'm speaking of two identicle heatsinks, one being made of copper and the other aluminum. I have never seen an aluminum one in such a test outperform copper. If you're speaking of two heatsinks with identicle masses, then yeah, I can see aluminum winning easily. But that's not a very fair test then is it? I mean, when comparing sheer performance, ignoring anything like weight, which may or may not be a factor (PC cooling, bolt it into the motherboard, not a problem. Copper radiator in a car, a problem).
Thanks... Emon... Those are really neat pics. The first one is a typical rectangular extended fin (copper), the second is a 'pin-type fin' unit with turbulation ribs near the top of the fins (aluminium). A lot would be to do with how these units are cooled - the first looks as though it could perhaps have a cooling fan bolted to the top edge of the fins - I have seen a number of that design. The aluminium unit looks to be of natural convection type - ie. no specific cooling fan attached & relying purely on the air-pull through over the motherboard. adTherm...
For this thread I was assuming everything was talking about cpu or gpu waterblocks - not radiators. IMO There is little point using expensive materials for radiators - copper is good enough. Im also quite sure that a copper rad will beat any alu one, assuming the deisgn is optimised for such. However radiators are usually more fragile, because of the thin fins, I think brass acctually makes a better rad, as it would be stronger, although it will sacrifice some performace. The thermal conductivity is still close to alu, and has very few problems with corrosion. Done some research today on how I could make a diamond/copper base for a heatsink or thermal conductivity test. I am sure I wont get the chance using the cold press and hot press, they would take far to much time to set up and I dont know how either. Im not sure the hot press can even make them, as we probably havent got the right graphite molds. Another way I could do it is cold-pressing (a different press) at ~200 bar into a button (much easer on this press). This can then go straight into the furnace. To make it I would use copper and diamond, to press them. As it is cold pressed, the copper will be slightly porus, becuase there is no heat to fuse it all together. To solve this it can go in the furnace at 1000c (or max temp with no risk of melting) with some silver on top. Silver will then melt and get drawn into the porus gaps, hopefully giving a solid diamond/copper/silver base, good enough for a waterblock. Im sure the w/m.k will be better than copper on its own, and I could send it to be tested properly. Im going to find out more about this method, and see if theres any more that can be done with the avilable tools. adTherm I suggest you do some reading over at forums.procooling.com as theres a lot of watercooling developement/research over there, mainly in waterblock design (not radiators) but it might still be usefull for you.
kbn, actually, brass has a thermal conductivity of 100 w/mK, and aluminum's is like, 240ish I think? Brass is really a bad idea, it's more durable but it does suck quite a bit more than aluminum. Besides, what kind of a beating would you possibly give to a brand new radiator?
^^Thanks 'kbn'... nice link... I also have a number of ideas for potential water-block developments & improvements. I will simulate these & tune the ideas. ----------- In terms of equivalent heat-transfer: For water-blocks, the issue to consider is the following equation (Fourier eqn): q' = |dT| / (|dx|/(A*kw)) where: q' = heat transfer |dT| = T,hot - T,cold |dx| = wall thickness A = Plane area kw = wall thermal conductivity for the same area, thermal resistance is governed by the ratio (|dx| / kw). In other words, if you maintain (|dx| / kw)al = (|dx| / kw)cu = (|dx| / kw)diamond ... the same, then the water-block will perform exactly the same in terms of heat pulled off the cpu. It is called 'heat transfer through a plane wall', in heat-transfer books. adTherm...
If you're going by the numbers, then diamond no doubt. If you're going by cost/benefit analysis, then aluminum or copper
Certainly does... Cartridge Brass (70% Cu, 30% Zn) : k@300K = 110 W/m.k Phosphor Gear Bronze (89% Cu, 11% Sn) : k@300K = 54 W/m.K For brasses, k proportional to Cu/Zn ratio. Bronze has a different structure, hence thermal conductivity. adTherm...
