Calculating heat sink effectiveness

The alternative to direct air cooling

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
James Long
Posts: 11
Joined: Sat Apr 09, 2005 12:36 pm

Calculating heat sink effectiveness

Post by James Long » Sun May 08, 2005 2:03 am

I’m interested in making a passive/semi passive watercooling setup as others on this forum have described before.
As usual, I have come up with a case design, but, since im not a mechanical engineer, im stuck trying to figure out whether the heat sinks will be good enough.

Basically, the case is a 360mm cube with a 80mm deep heatsink on each side, being cooled by two rows of 3 80mm (nexus) fans. The first idea was to use 120s on each side, but the resultant width (450mm cube) is a little bit big for my taste.

Ok:

The total area of the heatsink surface would be about 8.16m^2
With total air volume of about 0.02074m^3

Using some rough calcs, I get a thermal resistance of 0.186 deg/W.
Using Rth = Tr/P

I get a temp rise of 56 deg for a system dissipating 300W heat.
This is assuming only convection cooling, so how do I find the temp difference for air assisted cooling?


Basically, I’m trying to see if the case has enough cooling power to cool a cold water loop inside a totally (air) sealed case (with internal fans). Obviously this means the power dissipation is up to around 1500+W taking the 4 226W pelts into account.

Any ideas?



I realize this is hardly a normal quiet system, but I’m just interested to see if below ambient cooling can be done whilst keeping noise low and the people here are generally knowledgeable in such matters.


Heres a pic:

Image

James Long
Posts: 11
Joined: Sat Apr 09, 2005 12:36 pm

Post by James Long » Sun May 08, 2005 12:05 pm

Ok, I should have done some more research:


Commercially avaliable heatsinks, extruded fin 360x360x43:
A pair of these facing one another with three nexus 80s running half full power (allowing for the resistance of the sink and some quietening),
means a heatsink temp of 60 degrees for a load of 2000W!!

or more sensibly,
a rise of about 10 degrees for a normal watercooled setup.

Or, even better, a rise of 35 deg for passive convection cooling (of 200W dissipation).


All of these calcs are of course very rough.

Post Reply