Heatsinks: Mass versus fin density (and/or thickness)

[CurrentlyNotHere]

Hi. I'm modding my video card, and I had a question or two. I have a rather large collection of heatsinks (probably 30+) , and I was wondering which ones would be better. Ones that have the larger overall mass (I have several MASSIVE server heatsinks with thick fins), or heatsinks that have thiner fins, and/or less mass as well.

(I already know copper is better than plain old steel/aluminum, so that is not an issue since I don't have any copper around.)

Anyone any ideas?

 

[OSUguy98]

surface area is your friend..... the more surface area you have (that air can easily flow across), the better heat transfer (cooling) you will have.....


Keep on Folding!! For the [H]orde!!

 

[ZOMGWTFBBQ]

The general rule with heatsinks IMO is the more surface area the better, mass of the fins doesnt weigh in nearly as much as the amount of surface area they provide.

Damn OSU lol, posted while i was writing

 

[CurrentlyNotHere]

So, what matters is the overall surface area of the fins (or is it the surface area of the bottom/plate of the heat sink?)

So, would it better if I used one large, or several larger pieces for VRAM/GPU, or should I keep it minimal (heatsinks wouldn't cover any of the PCB space, just the VRAM ICs and the GPU itself)

Also, mass doesn't mean anything?

 

[jon67]

The surface area of the heat exchanger matters because the heterogeneous heat transfer step (metal-to-air) is normally rate-determining (limiting), even with high air flow rates.

If you replace the Copper or Aluminum in your heatsink with a low thermal conductivity metal, e.g. Plutonium, heat transfer through the metal may become the limiting factor.

Increasing the mass of a heatsink without increasing the surface area will reduce CPU temperatures during short periods of CPU load due to increased heat capacity, but the steady-state CPU temperature at continuous load will not be affected.

 

[CoachB]

Mass would affect how long it took for the CPU/HS to reach temperature equlibrium. After that surface area rules.

A member over at the Rebels Haven forum tried running his water cooling system thru a 5 gallon tank hoping to use the large mass of water to avoid using a radiator. It worked fine for a while but the water eventually became too warm and he had to install a radiator (i.e increase the surface area) anyway.