An interesting read!

[Hoot]

If you have not already done so, please set aside approximately half an hour and read the excellent article on the front page about cooling. This should be mandatory reading before a person is allowed to post a question about cooling. Certainly one of the best breakdowns of cooling science for the PC owner I have ever seen. Read it and take the time to digest the content.

http://www.amdmb.com/article-display.php?ArticleID=105

Hoot gives this article a big "Two Thumbs up"

Hoot

[TranceBear]

This was my first place I went when I was having problems, I also highly recommend it!!

[Thelemac]

I like this article because it gives good information in an understandable manner. It also taught me something new...how air flow in the case helps. I knew it helped, but the actual physics of it weren't that clear to me. Now I know, and can better plan my airflow. :)

[Owenator]

I liked the article, a quick read for me because I have a Masters in Mechanical and Nuclear Engineering and focused on two-phase fluid systems (steam and water). I took courses in Computational Fluid Dynamics (CFD) among others and this rekindled my interest. I have some software at work (GOTHIC)s that calculates heat transfer through systems, we use it for steam line break accident analysis but it could be adapted to PC cooling no sweat. I will talk with the other engineers and set up a model of a PC case and see what kind of numbers I can get. I will use my home PC as a test case, this could be fun to do at work and my boss might even think I'm working, HA!!

O

[Owenator]

Originally posted by Would71
aaaaaaaaaaahhhhhhhhhhh, I bet you took that same test I did.. the nuke test for the Navy. I almost went into their nuke program, but decided I didn't wanna commit to 8 years of my life to one thing. :D Although, I am still fascinated with their systems to this day. One of the best examples of heat transfer in action, a ship's nuclear reactor.

Would71,

I was in the Navy ROTC alternate program for two days when I started college but when the showed me the 'dixie cup' hat and 'bell-bottom' pants I had to wair as a first year I thought I would look to different from the rest of my fellow college students so I opted out! I did go on to get a 'Nuke' degree just focused mainly on civilain uses. Those ship based reactors re cool though, they can go 20 years without refueling, pretty good mileage!

I actually work at a civilian/commercial nuclear power plant right now. It's basiaclly the old electric utility business with a new competitive bent.

I got the software this AM and I am going to start learning it so I can build a heat transfer model of my home PC. Wish me luck!

O

[Pitspawn]

I liked the glossary, cleared up a number of descriptions that I wasnt sure about. All the science backed up with proof is quite impressive too, a good read.

Thanks for linking them to overclockers :)

[Thelemac]

The heatsink is steady state because it isn't going anywhere. The fluid cooling the heatsink is, therefore it is not a steady state.

[cjtune]

Originally posted by Thelemac
The heatsink is steady state because it isn't going anywhere. The fluid cooling the heatsink is, therefore it is not a steady state.

A steady state system is defined as a system whose properties at any point (density, temperature, velocity, pressure, whatever) does not vary with time. Both the heatsink and the air cooling a heatsink will not be steady state for the first couple of minutes or so until the heat capacity of the HS metal is saturated with heat (temperature peaks out), but after that, till you turn off the power supply, it is steady state.

[BillA]

a worthwhile article
and as Hoot suggested, should be required reading before posting

some observations on the difficulties with the content of "technical" threads

the hugely disparate background of the individual posters virtually guarantees that 1/2 the posts will be (technically incompetent) babble, and 1/2 the posts will be almost incomprehensible to those without a technical background and/or education - and there's plenty of overlap in the two 1/2s.

I've no prejudice against learning, its an ongoing universally applicable process without end; and certainly discussion is a means of becoming aware of new "facts" and exposing errors in our "understanding".
And I'm a very strong advocate of using the "correct" word to describe something, even if it does have 5 syllables.

But it can be SO vexing to describe something that is a part of the most basic engineering or materials science coursework, and then have to do "battle" with a teener on a typewriter hell-bent on "proving" his opinion.

In engineering, the "truth" (with a grain of salt, eh) is described with reproducible numbers, not the majority's opinion on a public forum.

be cool

[Colin]

BillA,

Good to see you posting here again. What you just described hit the nail on the head. The reason we don't have more qualified engineers and scientists posting is the "ignorance factor." Why would an expert in his or her field want to waste their time proving the facts over and over to an ignorant person using high school debate tactics?

Here is another good article (http://www.pc-workshop.net/articles/heatsinkphys1.shtml) on the subject that Nevin dug up.

[Thelemac]

Originally posted by cjtune


A steady state system is defined as a system whose properties at any point (density, temperature, velocity, pressure, whatever) does not vary with time. Both the heatsink and the air cooling a heatsink will not be steady state for the first couple of minutes or so until the heat capacity of the HS metal is saturated with heat (temperature peaks out), but after that, till you turn off the power supply, it is steady state.

Hm...guess I didn't really know the definition...just figured that a steady state was one where there wasn't change. Since the fluid is, in fact, changing, I thought it was not steady state.

I guess my point really was that the fluid is moving on, whereas the heatsink isn't. So if the fluid can store more heat before it leaves the heat source, then it is going to be able to get the object cooler. Does that make sense to you, cause I'm not sure it does to me. I think I'm too tired to be arguing this. Oh well. :)

[CalCoolage]

Originally posted by Thelemac


I guess my point really was that the fluid is moving on, whereas the heatsink isn't. So if the fluid can store more heat before it leaves the heat source, then it is going to be able to get the object cooler. Does that make sense to you, cause I'm not sure it does to me.

Sounds right to me. It is the reason water works better than air, I would think. The guy that wrote the article simply said the fact that alumnum gets rid of heat better than copper is not applicable because the heat sink is at a steady state, which explains nothing. With respect to heat, the molecules of the sink and the air are both in an equilibrium. So is the motion of the fluid at the molecular level. Where does it happen that the heat capacity of the fluid is important, but the heat capacity of the heat sink means nothing?

Looks like I am going to have to figure this out myself, or else never get an answer. Provoking the techs here can't seem to get them out of hibernation.

If I do figure it out, I will no doubt get flamed as an egghead if I attempt to explain it. Probably no danger of figuring it out though.

[CalCoolage]

Originally posted by BillA
the desirability of high thermal capacity coolants needs no explanation here.

be cool
WARNING! Do not read the following message under any circumstance. The Surgeon General, the FDA, the EPA, the NIH, the Department of Agricuture and the Supreme Court have determined that it wll completely rot your brain. OTOH , the Congress thinks rotted brains are just great and will help them at election time.

I guess it is correct to assume that it is the larger amount of heat that water, or anything, picks up that makes it a better coolant, not something like viscosity or surface tension.

I only used heat capacity because that is something that is available and accepted. I only contrasted the argument for aluminum/copper with air/water with to avoid trying to figure out the details. How to quantify the molecular forces and the thermal gradient at the HS-water boundary, I could not guess. But the amount of energy each molecule gives up or takes up would seem to be what results in a substance having a particular specific heat. And the number of molecules involved, the density, would seem to be equally important in the rate of energy transfered. From this you could get a volume of heat passed by the HS (and an equal amount of heat taken up by the air) for a particular temperature difference/gradient. But we need to know the reverse. We know how much heat is passed. It is identical for all HSs however bad. What is the temperature difference for a given amount of heat? The temperature drops more near the HS/air boundary for aluminum than copper. Aluminum gives up its heat better in this sense, but is it good or bad or perhaps canceled by some other detail that has not been mentioned?

Having got this far, I say it is a bad thing. A larger temperature drop is equivalent to a larger apparent thermal resistance. It would be like putting some insulator in the place of the aluminum/air boundary. You can see that the better insulator would be the one that produced more temperature drop.

Some confusion is introduced because of the analogy to electical circuits, which should help we who are trained in electonics, but we are more used to thinking in terms of constant voltage (= temperature ) sources and not constant current (= heat) sources. An electical circuit translated from a heat map would be very difficult to interpret without doing the numbers.

Looking at it from the coolant side, water is way better than air. The heat capacity of water relative to air, I think, is very large but for aluminum versus copper is modest.

(brain rot continued in part 2!)

[CalCoolage]

Originally posted by CalCoolage

WARNING! etc.
(part 2!)

To sum up, the material with the lowest heat capacity tranfers its heat the slowest, not the fastest. A block of aluminum cools off faster than copper, even though it gives heat off at a lower rate, because it contains less heat to start with.

I would like to have produced an argument that heat capacity is irrelevant, but I think that must be impossible because:

1) It is this difference/gradient which makes the transfer happen at all.

2)Putting water in the place of air is a very big deal.

3) When I first started to pay attention to heat tranfer, I believed that moving air transfered more heat because more molecules per second contacted the heat sink. But I have been informed that that is not so, not if you hold the temperature of the air constant. The actual reason is supposed to be that moving the air faster reduces the thickness of the layer that sticks to the heat sink. At the surface, the air is alway stationary. This sticky layer is the main impediment to heat transfer. (In the air that moves, heat is carried away in the same way as if you put hot air in a box and drove it somewhere else.)

What this means is that heat is tranfered to the air by conduction, not convection, and fluids are the same as solids when it comes to analysing their thermal characterisics, once you determine the thickness of the sticky layer. Therefore, saying that it does NOT matter what the solid is, but it IS important what the fluid is, has to be wrong.

This leaves the problem of why some top notch designers use aluminum fins, with a copper base, on their "unlimited class" HSs. Some say it is to save money, which I find completely implausible. I also believe that combining two materials drives up the cost. Some say it is to save weight. But a couple of pounds is nothing for a screw mounted HS. I am quite sure it is because of a thermal advantage.

Here's an idea. Realistically strong, usable materials are alloys, and the numbers in tables are ususally for pure substances. Copper, I think, is particularly difficult to alloy without ruining its good conductance characteristics. (I'm making this up.) Heat sinks that use all copper, and which are among the best, often have something that supports the fins somehow, and the fins are not very tall. In order to be able to use taller fins and not use supports which interfere with airflow, they might use an aluminum alloy whose thermal transfer charcteristics nearly compensate for the lower conductance.

I think I've produced a theory which neither faction will like! :)

I'd love to see corrections.

[TUK101]

What an awesome article. This guy knew what he was talking about when he wrote this and just flat out backs it all up with facts. One other thing.......there were actaully peaple who thought that aluminum cools better than copper? Why do you think they stopped wiring houses with aluminum? Aluminum swells when heated and doesnt get rid of heat fast enough, plus when it cools it shrinks back and loosens any connectors.:eek:

[JetMech]

Originally posted by Thelemac


Hm...guess I didn't really know the definition...just figured that a steady state was one where there wasn't change. Since the fluid is, in fact, changing, I thought it was not steady state.

I guess my point really was that the fluid is moving on, whereas the heatsink isn't. So if the fluid can store more heat before it leaves the heat source, then it is going to be able to get the object cooler. Does that make sense to you, cause I'm not sure it does to me. I think I'm too tired to be arguing this. Oh well. :) That makes very good sense. As a matter of fact that is one of the benefits of operating your PC in an air conditioned room. There is no such thing as cold. There is a more heat condition and there is a less heat condition. Heat is simply friction caused by the molecules in a given medium striking against each other. This is also energy, and energy can neither be created nor destroyed; therefore the molecular movement is constantly transfering from the more heat medium to the less heat medium. This transfer will continue until both mediums molecules are moving at the same rate. This is the steady state. It is almost a myth because there are so many variable especially in a HSF situation. The fluid medium is constantly moving. It is the less heat medium and therefore the energy in the more heat medium is allowed to transfer ( it has no choice) to the lesser. Because the fluid is moving its state is constantly changing. Because the more heat medium is constantly giving up energy its state is not steady. I don't see how a steady state could ever exist inside a PC. The only way would be to liberalize the definition and that physics will not allow us to do. When we cool we strive to find that point in which the energy produced by the chip is absorbed by the sink and subsequently by the air faster than the chip can produce it. An impossible feat but one we aim for none the less.

[snn47]

cooper does have a benefit e.g. if it is used for the base plate.

Everyone just talks about the temperature generated by the heat and the surface area , never about the contact area of the heat source, and if the heat can be conducted to more then the core area.

A Peltier with 100W and a contact area of 50 * 50mm to pass the excess heat is different to the ~10*10mm of an Atlon.

Therefore the surface area of an heat sink is only one factor if the heat cannot be conducted fast enough through all parts of the heatsink.

[cld230]

I'm not sure if anyone has really cleared up this whole matter of whether copper or aluminum is better or worse or what the whole deal is with this stuff, so here is my shot at it...

Thermal conductivity - this is how well heat is moved around within a solid

Heat capacity - this is how much heat energy whatever you are dealing with can hold

The thermal conductivity of copper (400 W/m K) is higher than that of aluminum (237 W/m K), so copper will be better at moving heat once it is in the metal. If your surface is the same (same surface area, roughness, etc.) the convective heat transfer (heat getting into/ out of the solid) will be the same. So copper if you have identical pieces of copper and aluminum, the copper will transfer heat from your heat source (CPU) to your coolant (water or whatever) better. The advantage of aluminum as pointed out in the article is that the heat transfer relative to the mass is better, so your aluminum block is going to be a lot lighter.

As far as the heat capacity, the heat capacity of aluminum (900 J/kg K) is higher than that of copper (385 J/kg K). But, we have already decided that the copper block is going to be a lot heavier than the aluminum (all things being equal, about 3 times), the copper will simply hold more heat. Under normal operation, this is no big deal, but what it means is that the copper will take longer to heat up and cool down. If your pump breaks, the copper may save your CPU for a bit longer.

I'm not going to get into the water heat capacity versus air heat capacity argument, because I think that we all are convinced that water will cool things better than air.

[3DMike]

When trying to asses the performance of a new heatsink design, or even when designing one yourself the link below may prove very usefull!

http://sound.westhost.com/heatsinks.htm

[Maximus Nickus]

I think this works... http://sound.westhost.com/heatsinks.htm
:D

If it don't then type in in ya browser!!!!

[LS7_]

Great article.
LS

http://i2ff.com/nascar/rank.php?id=986360

[arhines]

Indeed, a good (although in some cases non-computer heatsink relevant) read. I found the altitude derating factors very interesting...I had always thought that living in colorado or some other high up place would limit your oc'ing ability, but didn't think it would be by that much!

[knuckledragger]

Has anyone ever tried an actual refrigerant system? Filled with the current automotive R135 or household R22 refrigerants?

Knuckledragger

[knuckledragger]

Originally posted by knuckledragger
Has anyone ever tried an actual refrigerant system? Filled with the current automotive R135 or household R22 refrigerants?

Knuckledragger
==============================
Again I ask, has anyone ever tried a cooling system that uses real refrigerant? Remember, ammonia is FAR superior to R135, R22 or even the old R12 as a coolant. Except for that little poisonous thing.

Just curious.

Knuckledragger

[Shepherd]

Be shere it works or your sol or go with VAport chill

[quadrophenic]

the following is some of the key facts i read in an article here on overclockers.com not sure who wrote it but it was very good anyway:

heat is energy... the cold molecules have less energy than the hot ones so the energy is distributed through the heatsink( or waterblock)

in convection cooling a fluid (air or water) flows over the heatsink and the energy from the heatsink will hopefully attach the a fluid molecule (air or water) and then be cooled or taken out of the case...

so what you are infact trying to do is pull the heat off of the cpu rather than put cold into it

now radiation is simply the metal 'radiating' the heat or energy off of the heatsink into the air without air or fluid... aluminum has a better ability to do this than copper but it is such a small part of the cooling process it is not really an issue since copper conducts heat much better than aluminum.

conduction is when you have a cold surface that pulls the heat off by direct contact (peltiers etc..)

just wanted to post this because it helps give you a real grasp of whats going on... i recommend reading the article (i will try and find a link to it)

the same person also wrote a great article on the physics of flow in your water system and explained laminar flow and turbulant flow and why one is better and how to be able to somewhat predict if a waterblock will have lots of laminar flow (which is bad)...

-quadrophenic

[Tweak]

Originally posted by knuckledragger
Has anyone ever tried an actual refrigerant system? Filled with the current automotive R135 or household R22 refrigerants?

Knuckledragger

I have tried to use a dehumidifyer. It worked great, but I blew up 2 motherboards because of the condensation. So here I am back at air cooling.

[elysium]

Here is a great article on the different heat sink fans cooling ability:

tomshardware.com's heat sink fan roundup (http://www6.tomshardware.com/cpu/00q4/001211/cpu_cooler-10.html)

Just use the marks to get to the conclusion.

Ely

[FTC]

Yep... really good place. Do not forget to download and try the 'performance calculator' in that place. It is a really nice little spreadsheet to calculate the theoretical C/W of almost any heatsing made of flat fins, taking into account materials, coating,...

Regards
FTC

[Nov4]

my first post... cool...

when u says oxidize u mean it stops looking clean-brown and looks like uuhhh some dirty metal colour? not very sure, the links didn't work on my comp. i got a mark of a fingerprint in this rainbow colour (sort of like an oil stain but its not) on the copper base of my volcano 6Cu. How do you go about PROPER oxidizing of that copper??? and how good is the oxidized over the normal?

[Mark Larson]

They need to fix their stylesheet - doesn't look quite right on Mozilla 1.0.1

Happened on every system i've tried their site on. The forums are fine.