Copper Sponge Silent PC Prototype

[DrBorg]

This is totally one of those "It Depends..." kind of question.

If you could care less what it Weighs vs. Cost, Copper is Best.

If you could care less what it Costs or Weighs, Silver is Even Better.

If you want to Engineer a solution for a given Power Dissipation, and Control both Cost and Weight, you use Copper and Aluminum.

It all depends on what you pay for.

I'd love to build you a silver heatpipe/heatsink solution for a 100% markup.

 

[sfsuphysics]

That's not obviously true: aluminum has a specific heat capacity 2.5x copper.

Since this looks so fun I'm going to jump in too

Actually Ducman's right, copper is about 3 and a third times more denser than aluminum, as a result it can physically hold more energy in a similar volume. Now being as a copper atom doesn't have 3.3 times as many nucleons this means that in any given volume there are going to be more copper atoms in a sheet/fin/whatever than there are in the same shape of aluminum. Less atoms means less capacity to hold onto energy, so the copper pot that is 300 degrees will have more energy in it than the same volume of aluminum pot that is 300 degrees, so the aluminum will cool quicker because there's not as much energy in it to dissipate.

Now that said the copper heatsink will not get as hot as an aluminum heatsink with the same amount of energy put into it, I've seen arguments that touching it tells you how well it's cooling (it doesn't). Now the hotter aluminum heatsink will dump that energy more quickly to the environment as a result, because power transfer is a function of temperature to the forth power... i.e.a hotter object will radiate MORE energy (Stephan-Boltzmann Law). Ipso facto the aluminum will dump the energy faster into the environment. The downside is that if you have insufficient airflow the temperature between the fins will also get hotter and that radiative power will drop as the temperature approaches the temperature of the fins, but that's true for copper too.

 

[Ducman69]

Now that said the copper heatsink will not get as hot as an aluminum heatsink with the same amount of energy put into it, I've seen arguments that touching it tells you how well it's cooling (it doesn't). Now the hotter aluminum heatsink will dump that energy more quickly to the environment as a result, because power transfer is a function of temperature to the forth power... i.e.a hotter object will radiate MORE energy (Stephan-Boltzmann Law). Ipso facto the aluminum will dump the energy faster into the environment. The downside is that if you have insufficient airflow the temperature between the fins will also get hotter and that radiative power will drop as the temperature approaches the temperature of the fins, but that's true for copper too.

Remember that copper conducts heat about twice as well as aluminum though, meaning that it can more readily transfer the heat from the CPU throughout its mass, and it absorbs and dissipates that heat within ITSELF (of which it can accept more, in order to deal with spurts of high CPU usage) away from the CPU better. Regarding transferring that heat to the air, if the aluminum has only been able to radiate that heat within its own mass half as well, its quite likely that there will be a large delta between the base of the heatsink and the tips of its fins compared to the copper. That would give the copper a higher average temperature to surface area since more heat reaches its extremities, and transfer the heat faster to the air.

 

[PhotoBobBarker]

This isn't even taking into account the finish of the Aluminum/copper.
For s solid to air HS, the WORST possible finish is nice and shiny. The best is a matte black finish. The fact that you can anodize Al with a black finish also helps it's case against Cu.

 

[jimmyb]

Your analysis is pretty flawed, and I am not a physicist.

Actually Ducman's right, copper is about 3 and a third times more denser than aluminum, as a result it can physically hold more energy in a similar volume. Now being as a copper atom doesn't have 3.3 times as many nucleons this means that in any given volume there are going to be more copper atoms in a sheet/fin/whatever than there are in the same shape of aluminum. Less atoms means less capacity to hold onto energy, so the copper pot that is 300 degrees will have more energy in it than the same volume of aluminum pot that is 300 degrees, so the aluminum will cool quicker because there's not as much energy in it to dissipate.

Heat capacity is generally a measured value. You can go and look these things up rather than speculating about them. Aluminum has a *specific* heat capacity ~2.5x that of copper; copper has a *volumetric* heat capacity ~1.4x aluminum.

These are the actual measured capacities for these materials to hold heat.

Now that said the copper heatsink will not get as hot as an aluminum heatsink with the same amount of energy put into it,

Unless you've gone and calculated this I doubt you can say with certainty; it involves solving a differential equation for the steady-state of the system. Copper will conduct heat out of the system more readily, however it also has a lower specific heat capacity and therefore will be hotter for an equal amount of energy.

i.e.a hotter object will radiate MORE energy (Stephan [sic]-Boltzmann Law).

Uhhh the Stefan-Boltzman law is only applicable to black body radiation. Heat sinks are not black bodies.

Heat sinks will absorb and dump the vast majority of their heat using standard convection. The rate of change of energy (not temperature) is proportional to the temperature differential. You have to factor in heat capacities here to know exactly what's going on.

It's a question you might see on a first year physics exam, and you would fail if you just attempted to reason through without doing any actual calculations.

 

[PhotoBobBarker]

jimmyb; the only part that I would (only kind of) disagree on is that in engineering (not pure science) terms, "specific heat" can be referring to volumetric heat capacity OR mass heat capacity (depending on the units given).

As for the the second quote. Given the same mass, the Al would be FAR better than the Cu (given the same surface finish as well). If we are looking at the same volume, things start looking a little different (in the way of the Cu HS.

 

[SGA76]

I suspect that the use of aluminum for the fins is more related to cost and weight that anything else. The fins being further from the motherboard, the impact of extra weight is increased. This adds extra stress on the motherboard. Using aluminum simply helps to keep the weight lower and keep costs down.

Thermal conductivity of copper is double compared to aluminum.

Copper does conduct heat much better than aluminum.
When I worked in a foundry making aluminum castings we used copper shims to counteract shrinkage during a pour.
Works fantastic pulling heat, doesn't dissipate heat half as well as aluminum. Those shims would stay smoldering hot for hours after we broke them out of the sand.

 

[jimmyb]

Given the same mass, aluminum will have greater volume (a bad thing - heat must now travel further) and different surface area (maybe a good thing). I think it is difficult to say.

 

[Grahamkracka]

I wasn't arguing he said that, but agreeing with him that copper conducts heat better but aluminum dissipates heat faster than copper. The reason I was thinking that, is if you took a copper and aluminum pot out of the oven that were both heated for three hours to 300oF and set them on the countertop, the aluminum would cool off faster because its lower density. However, now that I think about it, you are right because the heatsink is only acting as a conductor, and the better conductor will transfer heat, and thus cool faster no matter what. I think I confused myself because the denser metal would actually just have more energy to dissipate. My bad. Yeah copper is better.

You are correct, the copper has a much higher density...and therefore can actually store more thermal energy per unit volume. Combine that with copper's superior thermal conductivity you have a substance that is superior in both drawing heat from it's source and transmitting it to a separate medium (air in this case).

 

[sfsuphysics]

Your analysis is pretty flawed, and I am not a physicist.

it involves solving a differential equation for the steady-state of the system.

It's a question you might see on a first year physics exam, and you would fail if you just attempted to reason through without doing any actual calculations.

Armchair physicist though? Since you claim to know all about the physics behind it and the types of tests you'd see them on (hint: not a first physics exam)

Heat capacity is generally a measured value. You can go and look these things up rather than speculating about them. Aluminum has a *specific* heat capacity ~2.5x that of copper; copper has a *volumetric* heat capacity ~1.4x aluminum.

These are the actual measured capacities for these materials to hold heat.

Describing how something works in laymans terms is not speculation, and IMO it's a helluva lot better for the majority as a descriptive than "Look it up in a table, there's a number for it!" that sounds like responses I've heard from engineering students "What formula do I need?"

Unless you've gone and calculated this I doubt you can say with certainty; it involves solving a differential equation for the steady-state of the system. Copper will conduct heat out of the system more readily, however it also has a lower specific heat capacity and therefore will be hotter for an equal amount of energy.

Yes but it has a higher volumetric effect you said so yourself in the table number you dug out

copper has a *volumetric* heat capacity ~1.4x aluminum.

I was talking heatsinks of equal size (volume), not equal mass. So yeah, the copper should be cooler with a given volume.

Uhhh the Stefan-Boltzman law is only applicable to black body radiation. Heat sinks are not black bodies.

Yeah they are, you're a black body, the seat you're sitting on is a black body, now are they perfect black bodies? No such thing exists, however every solid objects radiates heat like a black body (with maybe some exceptions like supersolids or what not that's a whole different realm of thermodynamics)

Heat sinks will absorb and dump the vast majority of their heat using standard convection. The rate of change of energy (not temperature) is proportional to the temperature differential. You have to factor in heat capacities here to know exactly what's going on.

And for heatsinks of equal volume, aluminum will be hotter, as a result the change in energy is proportional to the temperature differential (higher with aluminum), hence it dumps energy to the environment faster.

However I'm not saying aluminum is better than copper, as Ducman pointed out, getting heat away from the CPU is the primary concern, and how it moves outwards (size of heatsink) becomes a primary concern as the size of the heatsink scales upward. However you get aluminum the same temperature as copper, similar volumes, the aluminum will cool off faster which was the big hubbub of argument not whether or not a particular type of heatsink works better than another.

 

[Grahamkracka]

For what it's worth, this is beyond what I learned during my 1st year physics haha. We just brushed over the basic thermodynamic principles and were mainly tested on an idealized carnot engine during that part of the semester....so yea.

 

[Grahamkracka]

Given the same mass, aluminum will have greater volume (a bad thing - heat must now travel further) and different surface area (maybe a good thing). I think it is difficult to say.

To make an accurate comparison, both would have to be the same shape and size. Sure, an aluminum sink may be better than an copper block. But that's not what we're talking about

 

[Grahamkracka]

Those were the ones that didn't make it into their junior year. I was a tutor and most of my customers were the "memorize all my equations" type. Nobody actually knew how to derive shit.

 

[jimmyb]

(hint: not a first physics exam)

I've sat through a handful of first year physics exams as student, so while I can't speak to what exams you've been exposed to, I can certainly confirm that some basic thermal questions have shown up on them.

This sort of stuff wouldn't show up much earlier on account of requiring knowledge of differential equations, and it wouldn't show up much later because it is pretty simple.

Describing how something works in laymans terms is not speculation, and IMO it's a helluva lot better for the majority as a descriptive than "Look it up in a table, there's a number for it!" that sounds like responses I've heard from engineering students "What formula do I need?"

Your description was neither layman's terms nor accurate. The heat capacity of a material is *defined* as the amount of heat required to raise the temperature one degree (or Kelvin). Apparently you can attempt to derive this value using monte carlo simulations, but it's fundamentally an empirical value, so pointing to a table with a measured value is by definition more accurate.

Yeah they are, you're a black body, the seat you're sitting on is a black body, now are they perfect black bodies? No such thing exists,

You could call them gray bodies, but black bodies are by definition idealized. In any case, are you seriously trying to examine the EM radiation of a heat sink??? Heat sinks work primarily through convective heat transfer.

And for heatsinks of equal volume, aluminum will be hotter, as a result the change in energy is proportional to the temperature differential (higher with aluminum), hence it dumps energy to the environment faster.

Go check Newton's law of cooling again. There are other variables such as thermal conductivity (much higher in copper) which are going to affect the rate which energy is dumped - further to this, you have to scale the energy delta with the specific heat capacity to determine the actual temperature delta.

This is something you would need to do a bit of algebra to figure out.

 

[Grahamkracka]

This conversation is basically how my work day goes...only without donuts or frozen yogurt. Strange how that worked out.

 

[sfsuphysics]

In my defense, if I had donuts I wouldn't have jumped into this conversation

 

[Geef]

Some of you guys are way too paranoid about dust. Unless you have an environment where the dust is sticky (machine shop, smoker, etc...) you can still easily blow dust out of that thing.

Imagine living in a house with 3 cats, 3 dogs, on a farm(lots of dirt comes in). Two of the dogs are Mastifs that constantly shed all year round.

This copper sponge won't be in my house any time soon.

 

[xorbe]

That doesn't look effective to me.

 

[Grahamkracka]

In my defense, if I had donuts I wouldn't have jumped into this conversation

QFT

 

[SGA76]

I live on a farm, dust happens A LOT out here.

 

[Fenris_Ulf]

Well, according to my thermodynamics textbook (Introduction to Heat Transfer, Incropera and Witt), material has no factor on convective cooling (solid to fluid heat transfer). It's shape, fin efficiency, airflow, surface roughness, and temperature. Material does come into play with conduction (heat transfer through a solid) - its one of the key variables. And radiative convection? Pfft., at only a few degrees of difference the emissivity is not even measurable.

 

[Grahamkracka]

Well, according to my thermodynamics textbook (Introduction to Heat Transfer, Incropera and Witt), material has no factor on convective cooling (solid to fluid heat transfer). It's shape, fin efficiency, airflow, surface roughness, and temperature. Material does come into play with conduction (heat transfer through a solid) - its one of the key variables. And radiative convection? Pfft., at only a few degrees of difference the emissivity is not even measurable.

Convective heat transfer is a function of conduction AND fluid flow. Material does play a part in conduction as each material will have different a thermal conductivity. Given equivalent volume/shape copper wins, end of story.