anvilmark writes "ABCNews has an article about a new carbon based thermal conducting foam. Very pricey to produce but has 4-5 times the efficiency of copper at 1/5th the weight of aluminum. ORNL technical documentation available here and here. Sounds like the perfect heat sink shim to me."
Re:Dumb question
by
NanoProf
·
· Score: 3, Informative
If you heat it up hot enough in the presence of
oxygen, it will burn. So will a diamond, for that
matter. The 3000 degree heat treatment
that they use during synthesis must be done
in an inert atmosphere. Coal burns more
easily than diamonds, I'd guess, but both
are carbon with fairly similar heats of
combustion. Local carbon bonding comes in
two main forms, sp2 (triangular) and sp3 (tetrahedral) bonding; the binding energies of
the two are quite close and both will happily
burn in the presence of oxygen, at high
temperatures.
-- Curtains for windows?
Re:shim..sink - what's the difference?
by
Anonymous Coward
·
· Score: 2, Informative
A shim is a small plate composed of thermally conductive material that you put between an FCPGA CPU and the heat sink. It has holes for the chip and the other parts of the CPU that rise above the rest.
The idea is to increase the amount contact surface area between the CPU and the heatsink.
Re:Goofy comparisons
by
GigsVT
·
· Score: 4, Informative
Uh, wrong, try again.
Where did you find this "fact", the encyclopedia retardica?
Alumina, i.e. sapphires et al, have high thermal conductivity, and yet are almost total insulators. QED.
Please, please, try to check your facts. We all make mistakes, but I have seen so many totally wrong posts in this article that it is depressing me.
-- I've had enough abrasive sigs. Kittens are cute and fuzzy.
I am utterly amazed....
by
dr_db
·
· Score: 5, Informative
That so many posters confuse a heat conductor with a heat pump.
Come on - this will not be "keeping your fridge colder" or "cooling your drinks". It will just make whatever it's attached to move to the ambient temperature faster. Wrap it around your fridge and you will have sour cream in your milk, etc. Or else the coldest kitchen around.
Either it's a brain dead friday, or the collective IQ of Slashdot is lower than I assumed over the last few years.
Re:I am utterly amazed....
by
shaar
·
· Score: 3, Informative
uhm... a fridge is a heat pump that takes heat from the inside and moves it outside. so wrapping this around your fridge (back of the fridge in particular) will give you a really _hot_ kitchen, and a really really cold fridge (given no hardware failure). The body of the fridge would do a decent job of insulating the outside from the inside, the conductor will just help the heat pump.
For somebody lambasting posters for not having a good grasp of heat transfer, you sure didn't spend much time thinking about it.
Re:I am utterly amazed....
by
sweet+reason
·
· Score: 2, Informative
wrapping this around your fridge (back of the fridge in particular) will give you a really _hot_ kitchen
it might make the fridge's heat exchanger slightly more efficient, but that just means the compressor won't need to run quite so long to cool the fridge, which in turn means the kitchen will get less hot, since the heat given of by a fridge comes mostly from the motor. the heat being moved from the inside came from the outside in the first place.
-- Everything should be made as simple as possible,
but not simpler. -- A.E.
Re:shim..sink - what's the difference?
by
Osty
·
· Score: 3, Informative
A shim is a small plate composed of thermally conductive material that you put between an FCPGA CPU and the heat sink. It has holes for the chip and the other parts of the CPU that rise above the rest.
Nope. A shim is generally not thermally conductive (and better damn well not be electrically conductive...), since it doesn't matter whether it is or not.
The idea is to increase the amount contact surface area between the CPU and the heatsink.
Again, wrong. The idea is not to increase the amount of contact surface between the CPU and the heatsink, as this would be impossible to do, unless you made the CPU itself larger -- heat only radiates off of a CPU from the little rectangular core in the middle; the ceramic surrounding the contact point has little to no thermal conductivity. Instead, the idea is to give the heatsink a larger area to which to apply pressure. This means it's going to be more difficult (though not impossible) to chip the CPU core if you're using a shim than if you're not. Shims only became popular with the Athlon Thunderbird chips that were trivially easy to break with a sloppy heatsink install. Since those shims were made out of copper (bad! that's electrically conductive, which means you could very easily short out your CPU), many of the more clueless overclockers instantly thought "copper == cool", and thus assumed that shims were another way to lower their CPU temps by a couple more degrees. They were wrong.
Re:Double braindead, it seems
by
jaoswald
·
· Score: 3, Informative
Sorry, try again.
Conduction through a heat conductor can be represented with the thermal equivalent of Ohm's law. Warmth of a soft drink above room temperature is equivalent to a charged capacitor, where you can consider the room to be ground.
In your drink experiment, a drink warmer than room temperature will equilibrate to room temperature eventually. The speed with which it will equilibrate with a time constant
tau = RC
where R is the thermal resistance, C the heat capacity of the soft drink.
Lower R (better thermal conductivity) means the time is faster. However, when all is said and done, the drink and the room are all the same temperature, and that temperature does NOT depend on the thermal conductivity. It depends on the relative heat capacities. Given that the room is much bigger than the drink, its heat capacity is much larger, so the change in room temperature is negligible. (The amount of heat in a warm drink is the amount of heat in an infinitesimally warmer room.)
The only thing that could get signficantly hotter is a cold drink in a warm room.
I think you need to study a bit harder, Mr. PhysicsGenius.
Re:Goofy comparisons
by
Anonymous Coward
·
· Score: 1, Informative
No, diamonds have an incredibly high thermal conductivity. If it's heat capacity you are after you should try water.
Re:Not all that hot...
by
jmichaelg
·
· Score: 4, Informative
....PURE copper runs 350-400 W/m-K...pure copper isn't usually used, an alloy is.
I was talking about copper alloys. At room temperature, Both copper 110 and 101 alloys have a thermal conductivity of 391 W/m-k. Phosphorous laced copper alloys will drop you down to around 380. The only reason I happen to have these numbers is I'm currently working on a heat sink.
The news article that got this thread going had so many inaccuracies that I'm prone to think that a marketeer at Poco got somebody at ABC News all excited with hype. Given the foam's poor thermal conductivity, I seriously doubt the national security agency is using it as a heat sink unless, possibly, it's on a satellite. But if that were the case, Poco would have been nda'd and the story wouldn't have made light of day. The story smells of marketeer-speak.
You're right about the density uneveness. There are several elemental foams available that have very uniform density. You can get metal silver foams for applications where surface area is very important. John Carnack (of doom fame) has been playing around with silver foams as a catalyst for hydrogen peroxide to drive his rocket.
However, as a heatsink, foams don't fare well because heat transfer is partially a function, not of surface area as you assert, but of the cross-sectional area perpendicular to heat flow. Foams have lots of surface area which is nice for catalysts but have lousy cross-sectional areas which is what is needed to transfer heat from one edge of the foam to the other. Once the heat is spread out over a heatsink's mass, THEN the heatsink's surface area comes into play. Foams suffer as heatsinks because they can't move heat well from the primary hot spot to their extremeties.
Having said all that, there's some experimental work going on with carbon heat sinks that are configured in standard heatsink geometries. Anandtech's Cebit report shows a few pictures of some carbon heatsinks. Carbon is attractive, because as an element, it does show promise. As a working material, it's difficult. If carbon nanotubes ever get out of the lab, there'll be a huge change - they've got great thermal conductivity - somewhere in the thousands of watts.
Slashdot Mirror
If you heat it up hot enough in the presence of oxygen, it will burn. So will a diamond, for that matter. The 3000 degree heat treatment that they use during synthesis must be done in an inert atmosphere. Coal burns more easily than diamonds, I'd guess, but both are carbon with fairly similar heats of combustion. Local carbon bonding comes in two main forms, sp2 (triangular) and sp3 (tetrahedral) bonding; the binding energies of the two are quite close and both will happily burn in the presence of oxygen, at high temperatures.
Curtains for windows?
A shim is a small plate composed of thermally conductive material that you put between an FCPGA CPU and the heat sink. It has holes for the chip and the other parts of the CPU that rise above the rest.
The idea is to increase the amount contact surface area between the CPU and the heatsink.
Uh, wrong, try again.
Where did you find this "fact", the encyclopedia retardica?
Alumina, i.e. sapphires et al, have high thermal conductivity, and yet are almost total insulators. QED.
Please, please, try to check your facts. We all make mistakes, but I have seen so many totally wrong posts in this article that it is depressing me.
I've had enough abrasive sigs. Kittens are cute and fuzzy.
That so many posters confuse a heat conductor with a heat pump.
Come on - this will not be "keeping your fridge colder" or "cooling your drinks". It will just make whatever it's attached to move to the ambient temperature faster. Wrap it around your fridge and you will have sour cream in your milk, etc. Or else the coldest kitchen around.
Either it's a brain dead friday, or the collective IQ of Slashdot is lower than I assumed over the last few years.
Nope. A shim is generally not thermally conductive (and better damn well not be electrically conductive ...), since it doesn't matter whether it is or not.
Again, wrong. The idea is not to increase the amount of contact surface between the CPU and the heatsink, as this would be impossible to do, unless you made the CPU itself larger -- heat only radiates off of a CPU from the little rectangular core in the middle; the ceramic surrounding the contact point has little to no thermal conductivity. Instead, the idea is to give the heatsink a larger area to which to apply pressure. This means it's going to be more difficult (though not impossible) to chip the CPU core if you're using a shim than if you're not. Shims only became popular with the Athlon Thunderbird chips that were trivially easy to break with a sloppy heatsink install. Since those shims were made out of copper (bad! that's electrically conductive, which means you could very easily short out your CPU), many of the more clueless overclockers instantly thought "copper == cool", and thus assumed that shims were another way to lower their CPU temps by a couple more degrees. They were wrong.
Sorry, try again.
Conduction through a heat conductor can be represented with the thermal equivalent of Ohm's law. Warmth of a soft drink above room temperature is equivalent to a charged capacitor, where you can consider the room to be ground.
In your drink experiment, a drink warmer than room temperature will equilibrate to room temperature eventually. The speed with which it will equilibrate with a time constant
tau = RC
where R is the thermal resistance, C the heat capacity of the soft drink.
Lower R (better thermal conductivity) means the time is faster. However, when all is said and done, the drink and the room are all the same temperature, and that temperature does NOT depend on the thermal conductivity. It depends on the relative heat capacities. Given that the room is much bigger than the drink, its heat capacity is much larger, so the change in room temperature is negligible. (The amount of heat in a warm drink is the amount of heat in an infinitesimally warmer room.)
The only thing that could get signficantly hotter is a cold drink in a warm room.
I think you need to study a bit harder, Mr. PhysicsGenius.
No, diamonds have an incredibly high thermal conductivity. If it's heat capacity you are after you should try water.
....PURE copper runs 350-400 W/m-K...pure copper isn't usually used, an alloy is.
I was talking about copper alloys. At room temperature, Both copper 110 and 101 alloys have a thermal conductivity of 391 W/m-k. Phosphorous laced copper alloys will drop you down to around 380. The only reason I happen to have these numbers is I'm currently working on a heat sink.
The news article that got this thread going had so many inaccuracies that I'm prone to think that a marketeer at Poco got somebody at ABC News all excited with hype. Given the foam's poor thermal conductivity, I seriously doubt the national security agency is using it as a heat sink unless, possibly, it's on a satellite. But if that were the case, Poco would have been nda'd and the story wouldn't have made light of day. The story smells of marketeer-speak.
You're right about the density uneveness. There are several elemental foams available that have very uniform density. You can get metal silver foams for applications where surface area is very important. John Carnack (of doom fame) has been playing around with silver foams as a catalyst for hydrogen peroxide to drive his rocket.
However, as a heatsink, foams don't fare well because heat transfer is partially a function, not of surface area as you assert, but of the cross-sectional area perpendicular to heat flow. Foams have lots of surface area which is nice for catalysts but have lousy cross-sectional areas which is what is needed to transfer heat from one edge of the foam to the other. Once the heat is spread out over a heatsink's mass, THEN the heatsink's surface area comes into play. Foams suffer as heatsinks because they can't move heat well from the primary hot spot to their extremeties.
Having said all that, there's some experimental work going on with carbon heat sinks that are configured in standard heatsink geometries. Anandtech's Cebit report shows a few pictures of some carbon heatsinks. Carbon is attractive, because as an element, it does show promise. As a working material, it's difficult. If carbon nanotubes ever get out of the lab, there'll be a huge change - they've got great thermal conductivity - somewhere in the thousands of watts.