Heat distribution Problem

I am a computer nerd. To maximize my computer speed, I need to minimize heat on the processor. I have a 3"x3" ceramic chip, attached to a like sized aluminium heat sink (fans and fins and such). The problem:
almost all the heat is generated in a 1" by 1" raised (a few microns)metal square centered on the chip. Therefore, the aluminium heatsink has an extremely hot center, and though the heatsink is cool to the touch, the chip center is still about 53 deg. Cel. Not good. What I need is something that will dissipate the 1" square of heat, to the entire 3" square base of the heatsink. I was thinking of a copper plate sandwiched between the heatsink and the chip, will that work? Is there a better solution?

 

[svanels]

Put in some more fans, make some room for adequate ventilation. The 40 pin flat ide cables block the circulation of air. Take a sharp knive cut the strings in groups of five and ty them together with isolation tape. A 2" wide cable will be come 1/2" wide.
Handy if you have a lots of drives or cd-roms

Regards
Steven van Els
SAvanEls@cq-link.sr

 

[MortenA]

I saw i "Toms harware" (

http://www.tomshardware.com)

a design for a water cooled processor.

Check it out yourself.

Best Regards

Morten

 

[hessian]

Why do you think 53C is hot for the chip case?
Wimpiest commercial chips are designed to operate at max 70C case. NOTE: you are NOT measuring either the chip temperature or the case temperature if you are reporting the "heat-sink" (spreader_convector) temperature. How do you have it attached to the "chip" (case)? There are various techniques, including thermally "conductive" (relatively speaking) adhesives, pads, and thermal grease. The most important consideration is how thin (!!!) the interface is, and whether it contains air. (If you didn't use a clamp and thermal pad or grease or adhesive, you have air). Air is a terrible heat conductor (.0007 W/in^2 /C), and thermal grease and adhesives are not a great deal better.
Your large package has only a small area actually in contact with the relatively small chip inside. If you are seeing a (measured) large delta-T from the center of the "heat-sink" to its edge:
1)you probably have reasonably good contact to the case
2)you don't have a non-uniform/poor convection problem,
3)you have a spreading problem, which would be best addressed with a copper layer or copper "heat-sink" IF you have good enough flatness and thermal interface materials thin enough to not make the problem worse.
Thermocouples are much better than touch to determine this.
IF the thermoucouples are actually measuring what you want to measure.

 

http://www.liquicool.com