making Frankenstein Intel coolers out of different models ?

A

Anonymous

I possess a bucketload of Intel heatsink-fan coolers.
Various models. I notices that the copper-core jobs have fans with
lower current. Thus I speculate these were made for lower noise?
Now if I take the copper-core heatsink with the highest fins, and
put on it the fan with most current, then would I get the best
cooler? Maybe one that could handle a Pentium D or Extreme Edition?
 
P

Paul

Anonymous said:
I possess a bucketload of Intel heatsink-fan coolers.
Various models. I notices that the copper-core jobs have fans with
lower current. Thus I speculate these were made for lower noise?
Now if I take the copper-core heatsink with the highest fins, and
put on it the fan with most current, then would I get the best
cooler? Maybe one that could handle a Pentium D or Extreme Edition?

Heatsink designs can be considered to come in different "classes".

1) Pure aluminum. Typically only good for low power processors, like
say 35W or less. Used ten years ago. Even if you make the cooler
extremely large, or even solid aluminum, the heat doesn't travel
well enough, to make good usage of the metal mass and surface area.

2) Aluminum with copper core. The copper functions as a heat spreader.
But the spreading angle isn't really good enough to make a superior
cooler. The heat still has trouble getting into the fins. The ends
of the fins remain cool to the touch.

If you increase the air velocity over the fins, eventually you
reach a point of asymptotic improvement. Around say 800 LFPM,
you're not getting any additional cooling. You can make the fan
scream, but it isn't helping. (They use the screaming fan method,
in rack mount server computers :) )

3) Current generation after market coolers use heat pipes. These
are more effective than solid copper, because they involve an
active transport mechanism (fluid to vapor phase change). The
heat pipes are a virtually perfect transport mechanism. The
heat pipes run up through the fin assemblies, making better
usage of the fins than in (1) or (2).

If you're cooling a Pentium D or EE of some sort, look for a
heatpipe cooler.

To compare rough numbers, a conventional cooler (2) will be around
theta_R of 0.33C/W. That means, if you run a 100 Watt processor, the
case temperature of the processor rises 33 degrees C.

A heatpipe cooler, with a large enough fin assembly, can achieve
somewhere around 0.15 to 0.20 C/W thermal resistance. That means,
with the same 100W processor, you're looking at a 15C to 20C rise
in CPU case temperature. A heatpipe cooler gives you the option, to run
a slightly higher CPU temperature, while turning the fan down
so it is inaudible.

This isn't the whole story to cooling. You've solved the local heat
problem at the CPU, but you also have to exhaust that hot air
with the computer case cooling fan. In some cases, you may find
the computer case fan is louder than your CPU fan. But if you
turn the computer case fan way down, the temperature inside the
computer case rises, and that will also result in the CPU getting
hotter. Both fans are necessary.

If you had $1000 to spend, Zalman used to make a computer case filled
with various kinds of heat pipe, which allow convection cooling of a
computer. But that isn't practical once the computer is dissipating
more than a couple hundred watts. As far as I know, they've stopped
making that, since there isn't much of a market for it. Such computer
cases are popular in recording studios, but for home usage, not many
people can justify such a high price.

http://www.xbitlabs.com/articles/cases/display/zalman-tnn500af.html

http://www.xbitlabs.com/images/other/zalman-tnn500af/opened-b.jpg

Paul
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top