Basic Thermodynamics Quiz (Heatsink Materials Discussion)
Missileman
Orlando, Florida Icrontian
Okay - You have 3 bars of metal :
1 iron
1 copper
1 Aluminum
They are all at 500 degrees.
Question 1: Which metal burns you the most?
Question 2: Which metal burns you the least?
Question 3: Which metal burns you the deepest? (for lack of a better word).
Air cooling or Water cooling. Which is better from a thermal stand point?
Answers and explanations to follow
1 iron
1 copper
1 Aluminum
They are all at 500 degrees.
Question 1: Which metal burns you the most?
Question 2: Which metal burns you the least?
Question 3: Which metal burns you the deepest? (for lack of a better word).
Air cooling or Water cooling. Which is better from a thermal stand point?
Answers and explanations to follow
0
Comments
Copper has the ability to accept heat easily and spead it uniformly throughout it's mass. It does this because it likes to keep it's heat. The spreading effect is why it is used on the bottom of the cooking pans. The quick heat pull is why the inserts in the heatsinks. Pure copper heatsinks require a higher air flow to cool a cpu. This is the metal that burns you the least.
Aluminum has the ability to get rid of heat very fast. Most alloys cool faster than they will heat up. This metal will burn you the worst since it is trying to get rid of heat the fastest. The bad part in heatsinks is that it doesn't spread the heat around its mass readily making it prone to hot spots that won't accept anymore heat easily. Cooling fins are best made out of aluminum.
Iron has the ability to accept heat slowly, spread it evenly throughout, and give off heat at a slow steady pace. This metal will burn you the deepest. It is one of the reasons why cookware is cast iron. Cooks the evenest without hot spots. Also why you don't see iron heatsinks.
Which is better AIR or water. Trick question. Both are using the air to cool the cpu. It is the transport medium that is different.
When you are trying to cool your system you are trying to adjust the thermal gradient. Thermal energy is always trying to equalize itself. The higher the difference in the temperature means a faster heat movement. All systems will reach a thermal saturation or balance point. In air cooling the heatsink gets to final temp fairly quickly. Then you only need to wait for the case temp to come up to it's final readings. Of course as the case temp comes up so do your temps because the thermal gradient is flattening out.
In water over air systems you carry the heat outside using the water. Then the water transfers the heat to the air to be carried away. Case temp doesn't mean anything because the water is whats transporting the thermal energy. A water system takes longer to reach saturation because of the volume of water. It works better than air because the radiator has a larger amount of surface area exposed to transfer the heat to the air so it in effect transfers more heat away faster. As water temp rises the temperature gradient flattens and system loses its advantage. Size of radiator, air flow, and water flow rate all effect this, but all systems will hit saturation eventually given enough heat input.
I'm not an expert on thermodynamics by a long shot, but I did instrument rocket motors and such for testing and do data analysis. I knew just enough to be dangerous. There were speciallized engineers to look into thermal problems. It's a very complex and confusing subject. Hopefully this makes sense to a few people.
I have always considered this idea that copper absorbs heat better than aluminum, but aluminum radiates heat better than copper does to be misleading. I may be wrong, but to the best of my knowledge, this is simply not the case.
What IS correct is this:
Copper conducts heat better than aluminum does; that is, copper moves heat faster than aluminum does. This myth that aluminum radiates heat better than copper probably got started by someone who noticed that an aluminum heatsink cools down faster than a copper one does.
This is true, but it has nothing to do with aluminum radiating heat better than copper. It has to do with mass. A given volume of copper will weigh a great deal more than the same volume of aluminum. Consequently, the copper cannot react as fast as the aluminum can. That's all. It has nothing to do with radiating heat better, or worse, or whatever, just the fact that since there's less "stuff" in an aluminum block, it'll change temperature faster than the same size copper block will.
At least that's my understanding.
"Hey Thrax, why won't my Computer boot?"
<i>"You're violating the second law of thermodynamics"</i>
"Memtest has errors, what should I do?"
<i>"Stop violating the second law of thermodynamics, and replace the memory."</i>
"Why do boys have peepees?"
<i>"The second law of thermodynamics."</i>
okay no, this one is definitely the third law of thermodynamics
All materials have a physical property known as specific heat(SH). This is a measure of the rate that it absorbs and releases heat energy. Fresh water has a specific heat of 1.00. This of course is the reference material. Air has a SH of 0.25 which means it transfers heat only 1/4 as fast as water.
Aluminum has a SH of 0.22. Almost the same as air. This means if you held your hand above an open flame the air would transmit heat to you slightly faster than if a block of aluminum was between you and the flame.
Copper on the other hand has a SH of 0.09. Meaning that it wants to block the heat. It does this by holding the heat energy and spreading it acrossed its given mass.
Iron on the other hand has a SH of 0.11. About half way up the scale.
The formula for this heatsink stuff is : Heat gained or lost = Mass x Change in Temperature x Specific Heat.
Because you are dealing with mixed materials it complicates the issue, but basically copper carries the heat out over the complete mass before radiating it away. This means copper heatsinks need more mass/surface area to match the efficiency of an aluminum heat sink.
Density has nothing to do with it. It just happens that copper weighs more than aluminum.
Have you noticed how your system temp seems to rise to a certain point and then it seems to stabilze? This is because you have reached the configurations saturation point. The heat sink mass has reached its temperature balance with the surrounding air and is now transfering heat as fast as it is produced by the source.
Why do you think aluminum cases came about? If you had 2 identical cases, 1 steel (basically iron), and 1 aluminum. The aluminum would run 2-3 degrees cooler to the air inside cause it radiates heat away from the inside faster than the steel one. Engineers are looking for every degree they can suck out of these systems cause you are dealing with a lot of heat from a small place that you are trying to get rid of. Every degree you can cool the air inside the case goes directly to lowering the temperature of the processor.
The ratio of mass to surface area is the key. More mass postpones saturation, hopefully long enough to allow the surface area to temperature gradient to work faster than the heatsource. Moving more air/water acrossed the surface also postpones the saturation point. Faster fans/pumps affect it, but it still boils down to an energy build up that has to be disipated.
Geeky - I stand by my previous answers. It is not a myth about aluminum transferring heat faster than copper. It is fact as you can see from the figures above.
Here's my question: if Iron has the ability to spread heat uniformly throughout, why don't I burn my hand on the handle of the Iron skillet when I take it off the burner?
If you used cast iron to cook with you'd get it quick. It's the saturation point/mass thing. It's surface area is radiating faster than the heatsource so the heat stops filling up. Increase the temp of the heat source enough and eventually the whole pan will become red hot.
So that's why the pan handle would burn my hand if I put the skillet in the oven, but not from only cooking for a short time on the burner, right?
Outstanding explanation!
Well the oven thing is bit different. In the oven the temperature gradient is reversed. Your putting heat in over the whole surface area of the pan (including the handle) so it has no choice but balance its temp with air in the oven.
That's all this stuff is. A balancing act. Everything wants to become the same temperature. Thermal energy is the same as electricity. It all wants to go to zero state (balanced). Electricity is electron flow and this is heat energy, but it all works the same way. The basic physics remain about the same. Everything has a resistance to heat or electron flow. Flow is always in the direction of higher energy state to lower energy state. The material you have to flow through determines how fast the flow goes.
Anyway... I never took formal welding class and can't give you the numbers to match Missleman, but what I CAN say... Aluminum CAN be cut by an oxy/acetylene cutting torch, just very, very poorly. Basically, you spend more time heating it up than you do cutting... not worth it. Steel on the other hand.... Spend a few seconds getting a nice hot spot going and , your off cutting away.
For aluminum, we had to use a plasma cutting torch.... THAT ROCKS!!!!! you can cut Aluminum like butter... and I mean butter with a REALLY hot knife.... it's really awesome. You can draw things, write your name...
Why? Aluminum cutting with an oxy acetylene is pointless. Was the temperature of the flame is not hot enough to melt the Aluminum... yes and no. The aluminum was dissipating the heat almost as fast as the heat was being applied. If the acetylene was hotter, the aluminum would cut a bit better...
The plasma cutting is much hotter, uses electricity, and pin points the heat in a tiny spot.
One question, tho. You said that:
That doesn't make any sense to me.
Every test I've seen on the aluminum vs. steel cases shows no difference in cooling performance between the two. While you're right that in theory, an aluminum case would cool better, in practice, if the case has decent airflow, that's not the case. The air shouldn't be staying in the case long enough to heat up the case itself, which is why with a properly cooled case, aluminum vs. steel, or even acrylic, should make no difference.
The airflow and the case air temp are effecting it. If the air in the case is cooler than the heatsink then yes even the cool fins transfer off heat energy. If you slow down the airflow, increase the heat output, or warm up the case air then the fins get warmer farther from the source. This is more apparent on copper than aluminum because copper conducts the heat through itself faster than aluminum. It's the temperature difference between the materials that effects the speed of temp change. The temperature gradient as it's called. If you blow cooled air into yur case your cpu temps would go down and the cooling fins on the heatsink would be colder. Doesn't mean they are not working. Quite the opposite. It means they are working carrying the heat away faster than the heat from the source can get out there. Increase the air temp and the transfer rate slows down meaning the heat can penetrate farther out the fin before being removed.
Your right about the cases. Proper air flow is the most important factor. Again it's airflow volume and rate. If you had 2 identical cases, with identical airflow patterns and heatsources, the aluminum would be cooler. The problem is you get all kinds of unseen swirling, deadzones and such going on inside a case. Plus you have other heatsources too (NB, GPU). They play a part in it also. Most times when you hear about the heatsink fan works better sucking than blowing or such it is because when the fan is blowing it is picking up warmer air coming off the GPU thus reducing the temperature gradient of the air going over the heatsink. Then you have fan deadspots in the center etc....
There are a million and one things that effect the cooling efficiency of a heatsink/fan. The basic rule is simple though. You want the coldest air possible traveling acrossed the fins. Keeping the air changed out in the case more often makes this concept easier. External radiator water cooling makes this even easier since the heat is being carried out on the water instead of the air. This keeps the temperature gradient inside the case higher plus the radiator being outside the case means it has a higher temperature gradient being that it is pulling cooler room air.
Air movement effects cooling because as the air warms around the heatsink fins the temperature gradient drops slowing the transfer. Moving fresh cool air/water in is key. Slow your heatsink fan down and watch how much warmer the heatsink fins get yet your cpu temp only rises a degree or two.
I hope all this makes sense.
Now, on the metals and engineering and construction of heatsinks - let's throw in some variables:
space between fins (air pressure needed to flow through)
joinng of fins to core (is it a loose crimp, or a weld with an metal alloy)
quality of workmanship (precision) of different metals joined together (such as cooper core surrounded by aluminum fins)
All materials have a group of thermophysical properties that we care about.
Emisivity and absorbance are not an issue. Our temperatures are too low for radiant heat transfer to be an issue.
Heat capacity, this is a measure of how much energy it takes to change the temperature of the material. The only way that this comes into play in computer cooling systems is the time constant of the system. If you build with high heat capacity metals (Cu) or have a lot of mass to heat (water tank) the system will take a long time to come up to its stable temperature. Once the system is stable with constant heat load and cooling this property has no role in heat transfer.
Thermal conductivity is what we need for cooling.
In order to picture this think of a cooling system, CPU/compound/heat sink/air.
Now, picture each of these as a resistor, all in series. There are more of course, since each of the interfaces has a resistance to heat transfer. In the overall picture the interfaces are by far the biggest resistances. And the boundary layer of air on the surface of the heat sink is the worst of all. The reason that high CFM fans help is that the higher velocity airflow results in a thinner boundary layer.
High tc metals let us build larger heat sinks and have a reasonable amount of heat reach the entire structure. Larger heat sinks let us overcome some of the limitations due to convective heat transfer and boundary layer effects.
You know that power plants have to condense a million pounds of steam an hour. That takes some heat transfer.
Being a metallurgist I like the materials side better. But most people need more help with the heat transfer part.
I wasn't trying to teach a class of engineers a class in thermodynamics. I was trying to explain basic principles in a way normal people can understand. Doing that means taking some dramatic license with the science. Actually Specific Heat is the thermal energy property of a given material. I quote "The specific heat of a substance is defined as the amount of thermal energy that must be absorbed or lost for 1 g of that substance to change its temperature by 1º C." This correlates to being a direct measure of heat absorption/emission rates.
Simply put - this means water is hard to heat and slow to cool. Copper heats quickly and cools slowly, Aluminum heats quickly and cools almost as fast. It changes nothing that we have discussed. Thermal conductivity rate for any given system is constant. The only part it plays is getting the heat to spread to the farthest surface area to make removal easier. Once a system is put together you can't effect the basic conductivity unless you affect one of the items we have been discussing.
You're right that conductivity is important for the design, but once the system is set the physical properties can no longer be changed. Yes the boundry layer is important, so is the fin joint connections, type of materials used. That's what the designers get paid for. I was simply trying to give the regular user a place to look when he's trying to figure out how to make his system run a bit cooler.
You did (which is why I stickied the thread). However, you also managed to get the attention of some of the resident cooling nuts, myself included, and so you're now going to end up with a long-winded discussion on the various properties of aluminum and copper, whether you like it or not.
Now who is our resident engineer who can elaborate on precise implimentation?
We need more threads like this.
I saw an article on that topic a while ago.
The conclusion was that the case material didn't affect the temperature at all.
I'll dig up the article if you want to read it?
But aluminum is lighter and better for a LAN case.
EDIT: Found the article:
http://www.systemcooling.com/alum_steel-01.html
You are right and wrong. There was about a 0.5 degree lower on the aluminum. Of course in their test they monitored CPU temp and didn't monitor case temps. I would not expect a big drop on CPU temps since it is the heat source to start with and would have the least effect from the slight temp drop in case air temp. But you are right the benefit is almost nil. In theory it could be more but probably with the high air flow rate through the case it negates any advantage the aluminum would have.
Like I said, I'm not a thermodynamics guy. edcentric is probably more familiar with this low range stuff than me. I wouldn't even attempt to argue the physics of metals with him
I worked with rocket exhaust temps, re-entry vehicle temps, Hydraulic fluid temps. Lowest stuff was 1200 degrees. Most of the materials were all kevlar and carbon fiber composites which do not behave like a metal at all. We did stuff down to -70 F but we were only concerned with ambient temp and not looking for failure analysis data.
I'm not looking to offend anyone or start a war. Sometimes too much info is a bad thing as it just complicates the issue :banghead:
edcentric I hope you didn't take offense.
let us imagine we are going to test the absorption qualities of three textiles: felt, like under the floorboards of cars, paper towels, and terry hand towels.
now if you take one pound of each of these, and pour a bottlecap of water on, youll notice that the paper towels will have a biiit of lagtime before the cap of water is absorbed; the terry hand towel will absorb it immediatly, and the felt, willl slooooooowwwwly absorb the water, abeit completely;
Now if you dump an additional half cup on these three textiles, the terry will absorb the most, the fastest; the paper towels will have their characteristic slight lagtime, then absorb amost as much ad the felt underpad would absorb muuuuch slower that they, and additionally, it wouldnt spread thru the felt body as fast; and it would be harder, to put a fan on it and blow it dry than the other two; although ultimately you mighe be able to ake it hold as much wqater as either of the other two
this is very much like copper aluminum and iron or steel;
remember these products, gentlemen were NOT DESIGNED for our heatsinks; they have other primary uses such as high moldability and ductility for coopper, light weight and castability for aluminum, and high rigid strength for iron; these are products designed for OTHER uses, that we are using because 1 they work more or less, and 2 theyre available to us for general use; we can get them due to their abundent presence, stemming from their widespread uses in OTHER fields;
if you were in charge of some priceless document, and you had a pound of felt pad, a pound of paper towels, and a pound of cotton terry, on hand, and some guy came and splashed ur document with his drink, you'd reach for the terry cloth, because of it's ability to absorb a LARGE quantity QUICKER than the paper towels.. or felt
true, the paper towels (Aluminum)might actually take less work to dry(can dump heat off surfaces, once absorbed); but quickly getting the liquid off the document is the deal here, and quickly getting the heat off that chipface before damage occurs is the same situation; and so, since copper (Terrycloth)absorbs better, at its surface regions, than aluminum, although in its interior it might be a biiiit weaker at actually spreading the heat, we can just provide a biiit more surface area to blow the heat off of; the main thing with chip survival is to be able to absorb LARGE pulses of energy, faster; the surface atoms of copper absorb better, from another medium, than aluminum's surface atoms do, although the aluminum once it DOES have it inside, can transmit it BACK OUT of it's molecular structue a bit quicker; if the copper dumps heat say 15% less well than aluminum, we add 15% more surface area, and equalize that way; if one heatsink has fifty fins and another one has fifty seven, what's the diff to us, the user? The copper takes the heat IN a bit easier, and similarly sheds it more reluctantly; the aluminum finds it a bit harder to acept the heat IN, but, similarly sheds it with more enthusiasm; the iron would find it easier than aluminum to accept the heat, and, more reluctantly than copper, but would be MUCH more reluctant, to shed the heat it had absorbed.l hope i helped too.