copper vs aluminum

[winewood]

bobhead,

I looked for the article, but I couldnt find it. Could you point me in the right direction? Is it in the forum??

Thanks

[BillA]

any, and all, of Dave's articles should be read

http://www.amdmb.com/search.php?search=dave+smith

before posting ??

[bobhead]

front page of overclockers.com, basicly it stated
Myth: aluminum dissipates heat better than copper. I think thats what it was

it is a myth, they use aluminum pinned or finned heatsinks cuz they are lazy and its much easier to do and cheaper

[winewood]

bobhead..
I have looked for 20 minutes in forums and front page. Could you please provide a link. OR pm me the link, and I will post it. Thanks.

billA, wonderful link. That dave guy knows his stuff. I do believe I have learned a ton in the last 20 min!

[bobhead]

sorry, it wasnt on the front page anymore but in the WATERCOOLING section of overclockers.com

one of these 3 articles
http://overclockers.com/articles638/
http://overclockers.com/articles654/
http://overclockers.com/articles655/
and If you dont see it for some reason, I really dont know

[winewood]

I have read all the posted links, and performed hours more searches on this, and fail to see a statement refuting my understanding as a myth.

The closest was..
http://www.theinquirer.net/?article=6907

An expert tells us: "With heat sinks, aluminum is cheap and light, but copper is better thermally. In Watts per degree C per square centimetre, copper = 4.83 and Aluminium = 3.0

"Thermally" being the term used does not refer eplicitly to the dual nature of thermal conductivity or dissipation to my understanding. It does refer to a process in which heat removal gains were improved in a patent of using BOTH Aluminium and CU in this article *LINK*.
My belief remains that copper is the better conductor, and draws the heat into itself, while alum. can shed the heat more efficiently instead of retaining it. Please follow the *LINK* and explain why IBM wouldn't use just the "thermally better" copper instead of the combination of the different metals. Also view the patent here..
Switech , and the user BorgBob has some great usages of this technique that I am following carefully.

I appreciate all links on this, as they helped me understand the entire process better. (I read every one!)

[BillA]

ok, I'll waste some time
since you have read, and claim to have also understood, all the foregoing:

what is this 'heat shedding' property called ?

it's name ? (and a link to where it is defined please)
and how is it measured ? (an ASTM, BS, or DIN procedure will suffice)
and what then are the respective values for aluminum ? and for copper ? (whose test data are you citing ?)

no sales/marketspeak bullcrap
and no opinion (as yours is not of interest)
here we seek engineering data

[jaydee]

I am not really qulaified to answer this but i will point out what I see. First off you are comparing water cooling setups agains air cooling setups. This is your first mistake.

Think about it for a minute. Why do they put aluminum pins in the copper? To remove the heat that is stored in the copper. Now why do we use water? To remove the heat that is stored in the copper.

So basically what I am saying is water is replacing the attached aluminum in a water cooling system. Water is far superior to aluminum in removing heat from the cooper. So why would you add aluminum for no reason?

For the theory that al removes heat faster than copper is irrelevant in water cooling as you are using water in place of the aluminum.

[freeloadingbum]

Copper and aluminum can't shed heat any faster than the air or water can absorb it. As a result, they shed heat the same, regardless of their potential

[winewood]

<potentially argumentitive non constructive post edited out>

[jaydee]

http://www.efunda.com/materials/comm...tlProp=Thermal

I think all the answers can be found there. Amazing what you can find at www.google.com in 2 minutes. Note the link that had electric properties. It shows a resistance number that I think answers your question but I am not sure if higher resistance means it disapates more or less energy (heat) than the lower resistance. I would think less resistance means it flows better meaning it disapates faster but not sure.

Of course I can totally be wrong with this aswell. :shrug:

[winewood]

Jay.. thanks!
If Im reading this right,

A: copper is a better conductor electrically.
B: Alu expands and contracts slightly more than CU with temperature variances.
C: Copper draws energy into itself faster than alu.

However, I am wondering why manufacturers use a combination of these 2 metals instead of just solid CU? What advantage are they achieving buy this combination. With such small quantities used?

http://www.tp-e.com/TPEPres.pdf page 8
This shows the most efficent forced convection heat sinks to have a copper base with aluminium fins.

However
http://www.flotherm.com/technical_papers/t275.pdf
Page 3 shows copper having an advantage when it is thin, but is almost the same is the thickness of the sample approaches 1 mm X .8 mm.

http://www.flotherm.com/technical_papers/t281.pdf
There appears to be a 7% variance per 50 W of power in an air cooled environment in favor of CU

The last link showed very little difference between CU and AL. The most advantageous benifits of CU are the highest performance, however I see that the change is very small and AL is 3 times lighter and much cheaper per pound.

However, for with a watercooled block with water reaching both the copper and aluminum fins or posts of the block.. and the fluid addition to absorb the heat.. I think the differences will be negligible. What is then left is the design of the material and the efficiency in which it can transfer the heat to the liquid.

Conclusion. Manufacturers use the most expensive material in the most crucial area of the chip. After that the AL and CU similarities thermally void themselves out, as surface area to fluid contact is what will be transferring the heat. This brings the waterblock maker to use copper for the base, and freedom to use whatever they need for the top, as long as they are bonded in a favorable way to transfer heat between the two.

jaydee116 is right!
What do you guys think?

[BillA]

except that no 'bonding' of any type is favorable compared to a homogeneous material

we have been here before; aluminum is cheaper,
and copper is better

[myv65]

Not a week goes by that this subject doesn't raise its ugly head. . .

The statement that "copper absorbs heat better and aluminum gets rid of it better" can be construed as true, but only when one really understands how things work. First, let's get one thing straight, the statement only applies to an extremely narrow definition. Second, it mainly applies only to air cooled heat sinks.

If you look at the top performing air heat sinks, many have a copper base and aluminum pins or fins. This has little to do with manufacturing cost and everything to do with cooling effectiveness. When you need to transport heat with a high flux (heat per cross sectional area), conductivity is the only thing that really matters. The die pumps out a lot of heat from an area smaller than a dime. In order to move this heat somewhere else with a low delta-T you need high conductivity. Copper bottoms always beat aluminum (and copper water blocks always beat aluminum ones).

Once you get the heat away from the die you need to get rid of it. With water cooling, this means a short trip to water. Since convection via water is pretty efficient (anywhere from 10 to 20 times better than air), you don't need a whole lot of surface area. Since you don't need much surface area, the blocks can be small and you don't need to make the heat travel far.

With air cooling, convection becomes the bottleneck. Now you've got a choice. You can use a little copper and make thin fins/pins, and create a lot of surface area. Alternately, you can use a little more aluminum and create not-quite-so-thin fins/pins. Because copper is so much more dense, you get a lot less volume of metal for a given weight. So if you want a heat sink that weighs in under 500 grams, you'll only get so much surface area out of copper.

What it really comes down to is a comparison of conductivity versus density. If you made a ratio using these properties, a higher value would equate to higher "heat transfer efficiency". When you do this with copper and aluminum, aluminum's ratio is higher. Basically you can make pins fatter in aluminum such that the delta-T of conductivity through the pin is lower than for copper, all compliments of aluminum's lower density.

Lots of folks simplify this (read: don't have a clue how it works) and say the famous statement that I quoted above. There's a very good reason why you see hybrid air heat sinks, and it's the one I've tried to explain here. If you get all of it, you'll also know why it does not apply to water cooling.

Edit for a footnote:

If you have two heat sinks or water blocks of identical geometry, the copper one will perform better (and weigh a lot more), period. If you have an equal mass of both materials, you will get better results from aluminum in air cooling because of the facts stated above.

[BillA]

thanks for the edit there
for if one removes weight and cost from the equation then copper is the obvious choice

(yea, I know - silver; gimmie a break here)

[morphling1]

myv65 how many times did you explain this
Someone should take that post and nail it on top and use flashing fonts, so no one would ask that question again

[chewyboy]

all right looking through the waterblockdesign/contruction forum i this is the first i see this in the topic of anything and i've yet to see it in another post. I hate to say it but the thread was going nicely till, well to be honest the last couple posts were extremely rude. I'm glad myv65 explained it but i have yet to see this asked all the time, and yes i did do some looking. plus if you go back to the original topic you will see it's not just about which is most uber it's hey i have a buttload of scrap (not really scrap just extra) aluminum laying around but i'd have to pay out the wazoo for copper if i'm doing this for testing and not wanting every single mhz out of it which is better. sorry i'm ranting a bit for a newb but read all the posts not just the topic and then say stuff like "Not a week goes by that this subject doesn't raise its ugly head. . ." that right there raises a big red flag IMHO. this is one of the most informative boards i've yet to find, please do not turn it into another slashdot.......

[Cathar]

Originally posted by unregistered
thanks for the edit there
for if one removes weight and cost from the equation then copper is the obvious choice

(yea, I know - silver; gimmie a break here)

Well, cost isn't even that big of a difference either.

Copper/Aluminium stock prices

1 tonne of Grade A Copper is $1544 US
1 tonne of Primary Grade Aluminium is $1340 US

However, since copper's density is ~3x that of aluminium, it will cost nearly 4x as much for the same volume of it....

600g of copper costs $0.92 US
200g of aluminium costs $0.27 US

Neither amount will break the bank.

Consider the price of silver

600g of silver costs ~$100 US

[myv65]

@chewyboy,

Pretty much all the replies on the first page had already said:

1) Aluminum is fine, though not the top performer
2) Copper is better, but more expensive and more difficult to work with

The "ugly head" comment comes from cruising more forums than just this one and was not intended as a slight or insult toward you. I see nothing offensive in the remainder of the thread. If you thought otherwise, well then ya ought to read some of BillA's "explanations".

Anyway, you can find this topic discussed dozens of times among the more popular forums like overclockers, amdmb, anandtech, etc. Every time you'll see the same thing, namely people repeating the famous statement with nary a clue of what it really means.

What you'll find here are a bunch of people that are generally pretty bright, say it like it is, and make no apologies for posting a very straight-forward response.

For your particular situation, I think it's pretty obvious that you ought to at least start out in aluminum. Chances are somewhere between slim and none that your first attempt will be the holy grail of water blocks. If you're going to practice a little, by all means use what you have at hand. If you decide to switch to copper somewhere down the road for those few extra degrees, there's plenty of info laying around to indicate what design changes would be advantageous for copper vs aluminum.

[Joe]

Originally posted by unregistered
we have been here before; aluminum is cheaper,
and copper is better

Rofl, I am betting you could still dig up those threads where I was on the same argument with you about 2 - 3 years ago. hehehehe

Ohh the good ole days.

[BillA]

hey Joe, did you edit the history files ?
I did a search under "fu*king idiot" and came up blank
(plenty under "asshole" though - you are a revisionist !)
<joke, eh>

[chewyboy]

myv65, thx for your reply. as for reading other forums I’m pretty new to most of them, but have lurked for a bit. the forums you mentioned overclockers, amdmb, anandtech, out of those three i can honestly say that I’ve only read the anandtech one's and have since stopped reading them because of what I like to call the slashdot effect. What I consider the slashdot effect is in essence no matter what someone posts be it a question because they want to learn something or any statement be it true or not, heck even the statement of "2+2=4" on slashdot will get thousands of hate replys telling that person he is an idiot and to go learn math again. I've noticed that on anandtech's forums as well and well just don’t have the time in my day to care about them. I will however start browsing the other two as i feel they will probably bring to light some info that i do not yet have. As for doing this stuff, yes i am new at it so i will make mistakes. I also do not believe that I’ve asked any dumb questions on this board so far, and for the most part people have been nice enough to answer the questions with both good timing and proficient answers. Thanks for replying with the technical answer again in the previous messages.

[Joe]

Originally posted by unregistered
hey Joe, did you edit the history files ?
I did a search under "fu*king idiot" and came up blank
(plenty under "asshole" though - you are a revisionist !)
<joke, eh>

Umm nope no edits, but thats because the F word is blocked out in the forums so it was ****ing idiot maybe that search string will work ehehehe

[bigben2k]

And strangely enough, we've all failed to answer your question...

"How much do you really lose by using aluminum for your waterblock construction over copper?"

Answer:
It's going to depend on the design of the block.

If you compared the exact same design in both metals (which one wouldn't do), you might see a difference anywhere from 5 degrees C to 20 deg C.

The trick is that the baseplate of the waterblock will have a temperature that will range from hot (against the CPU die) to cold (touching the coolant). In copper, you can use a very thin baseplate (again, depending on the design). In Aluminium, you have to compensate for the thermal resistance by using a thicker baseplate, which increases the hot side temp (Did I get that right, guys?).

[BillA]

pretty close Ben

the hot side will be shedding heat faster though, so it will be a little cooler

couldn't resist, sorry folks