Why are heatsinks made of the materials that they are made out of?

[Howard]

Most aluminum heatsinks are made of 6035 alloy...
209 W/m-K

But, pure aluminum has a thermal conductivity of 243 W-m/K. Pure aluminum is very soft, so machining/forging/extruding wouldn't be a problem.

Synthetic diamond has a thermal conductivity of 2000 W/m-K. Imagine a heatsink made of that...

 

[Evadman]

There is a fine ballance between price, durability, and W-m/k. That fine line must be walked.

If you want to make a Diamond HSF, LMK so I can steal your computer

 

[Howard]

Who's going to drop their heatsink on the floor? Anyway, it's not soft enough not to be able to endure some rough tralve.

 

[zephyrprime]

I was reading a book on alloys that other day at my university's library and I learned all sorts of stuff about alloys that I didn't know before. I found out that alloys often have ingredients in them that don't improve their mechanical properties but instead improve their workability properties. For example, ingredients may improve flow or the smoothness of the surface or reduce the tendency to oxidize while being worked with. Sometimes chemical species are added to reduce sticking to the mold.

I don't know why 6035 alloy in particular is used for heatsinks.
Some of the metals in the fancy alloy you mentioned contained some precious metals I think. I'm pretty sure that gallium is a precious metal.

Also, most heatsinks are extruded and only minimally machined. Machine soft aluminum may be a lot easier than machining steel but I imagine it is still relatively costly to do so for a component that only cost $20-$30.

 

[Howard]

Keep in mind that some are forged.

Actually, the alloy I linked to is actually pure aluminum - those substances are just like [May contain traces of peanuts].

 

[CTho9305]

<< There is a fine ballance between price, durability, and W-m/k. That fine line must be walked.

If you want to make a Diamond HSF, LMK so I can steal your computer >>



I thought synthetic diamonds were worth much less than real ones and usually have defects that result in colors like yellow/blue/green



<< Color Clear Impurities (i.e. Nitrogen) and irradiation can change color to yellow, green, blue, pink, or brown. >>



aha, i was right

rofl@your sig

 

[CTho9305]

what are soda cans made of?

 

[mrzed]

I thought they were pretty much pure aluminum, but it says here that a small amount of manganese is added for strength.

Anyone here ever made a soda can heatsink?

 

[Howard]

You know, instead of using copper heatspreaders, we could just use all-copper heatsinks (if you take a pure aluminum heatsink and compare a pure-copper heatsink of the same design, the Cu one would cool better in the same environmental variables) and add diamond to the bottom using either CVD or PSD. Costly, but very efficient.

[Chemical Vapor Deposition, Plasma Spray Deposition, preferably the latter since it can produce layers up to 50 times thicker than the former]

 

[itrfuzzylogic]

where can i buy a Synthetic diamond heatsink?

 

[Shalmanese]

<<

<< There is a fine ballance between price, durability, and W-m/k. That fine line must be walked.

If you want to make a Diamond HSF, LMK so I can steal your computer >>



I thought synthetic diamonds were worth much less than real ones and usually have defects that result in colors like yellow/blue/green



<< Color Clear Impurities (i.e. Nitrogen) and irradiation can change color to yellow, green, blue, pink, or brown. >>



aha, i was right

rofl@your sig >>



Yes, they are far cheaper, something like in the hundreds of dollars per carat rather than thousands IIRC.

still not enough to be viable for a $20 HSF though

 

[rgwalt]

Machining heat sinks is cost prohibitive. Most sinks are extruded, poured, or forged. Also, molten aluminum is easier to work with than molten copper. I forget why, but I remember that fact from high school metals class.

Another thing to look at is the contact resistance of a material. Different substances have different contact resistances for different values of surface roughness and applied pressures.

Ryan

 

[RossGr]

Casting Al is a piece of cake, low melting temp, well compared to copper anyway! When you combine the ease of manufacture, cost of materials and effectifyness at cooling I bet that the copper is not all that cost effective. Copper has a very high melting point, it is nearly impossible to machine and is expensive. Since much of the effectiveness of a HS is in the geometies it may be that the flexabity of Al can make it a better HS in the end.

 

[ttn1]

You have to look at more than just the thermal conduction. You have three types of heat transfer: conduction, convection and radiation.

Copper is very good at conduction, while aluminum is very good at convection. At the CPU to heatsink interface you want very good
conductive heat transfer. At the fin to air interface you want very good convective and radiant heat transfer. Convection is greatly improved
by increased surface area.

All in all, there is way more to it than just high conductivity. If that were the case, why not make solid gold heatsinks. There is also the
economics of manufacture and availability of material.

 

[Mingon]

<< copper - it is nearly impossible to machine >>



I take it you have never machined copper then, its a pain but not impossible, just mind the swarf - ouch. The majority of mass produced heatsinks are extruded, this is a bit harder with copper and requires more extrusion pressure. As for casting a heatsink - I doubt they would be cast the design of heatsinks doesnt not lend itself to casting. Dont forget that pure aluminium (correct spelling you hethen's ) is more costly and probably not worth the effort.

 

[Agent004]

<<

<< copper - it is nearly impossible to machine >>



I take it you have never machined copper then, its a pain but not impossible, just mind the swarf - ouch. The majority of mass produced heatsinks are extruded, this is a bit harder with copper and requires more extrusion pressure. As for casting a heatsink - I doubt they would be cast the design of heatsinks doesnt not lend itself to casting. Dont forget that pure aluminium (correct spelling you hethen's ) is more costly and probably not worth the effort. >>



you are the one missing the point here

He said nearly

 

[Mingon]

No I am not missing the point, its nowhere near impossible, tricky yes, but in no way as awkward as some materials. I trained for four years as a mech engineer and used to do a lot of work with copper - in milling, flow turning and conventional turning and to a high degree of accuracy i.e 0.0005".

 

[Fandu]

Copper's not too bad to work with. The biggest problem is just trying to hold it while your working on it. It's quite easy to damage in a vise, etc because it's so soft. As far as 'machineability', it's pretty close to 6035. AFAIK the biggest reason that we don't see too many pure copper heat sinks is cost, and the second is it's weight. Copper's alot heavier, thus costs quite a bit more to ship around the world, and requires better retention brackets on the mobo's.

 

[Howard]

Um, before you start pissing up because of non-British spelling, it's spelled 'heathen'.

 

[Mingon]

I know I was being ironic

 

[robogk]

Hello Newbie here.
To say machining heatsinks is cost prohibitive is not entirely accurate. All heatsinks except for the cheap "spring like types" will need machining its impossible to fit stud devices without drilling and tapping and when fitting modules the surface of the heastsink should be skimmed to maximise thermal conductivity and even then a fine layer of thermal paste should be applied.

 

[soccerballtux]

what are soda cans made of?

soda, omg like so 'duh'....

 

[greenhawk]

FYI, bumping a thread from over 12 years ago.

 

[TurtleCrusher]

I don't mind a thread this old being bumped. It's interesting information to me.