![]() | ||
|
|
Random Nonsense / Geek Stuff All those random tech ramblings you can't fit anywhere else! |
![]() |
Thread Tools |
![]() |
#1 |
Cooling Neophyte
Join Date: May 2001
Location: seal beach
Posts: 1
|
![]()
ON DIE
I have started brainstorming for an on die cooler running mineral oil. It is going to be designed around a socket A architecture. Here's the points I've thought of so far... -A plexiglass casing (glue doesn't bond well to lexan) -a block with a close cutout around socket, bolted in through the AMD aux. holes, sealed with automotive gasket maker, and then double sealed with silicon sealant -an extra cutout close around the die with 3 holes... --1 on top, an intake, that shoots the oil right on the proc --1 on both sides of proc to make a good vortex and take away heat -a radiator of some sort -a pump of some sort -I was thinking of having it inline but someone please tell me their experiences or rumors of mineral oil (specific heat and other properties) (please suggest a good pump for oil) -Mineral oil for the substance because of its non electrical conductivity --someone please tell me if it'll erode aluminum piping or copper piping That's about it, pretty simple but I think it'll really work well as it doesn't have to worry about the heat lost as it cools all of the chip, not just that which transfers its heat successfully to the waterblock, peltier, aircoooler, etc. ANY COMMENTS, SUGGESTIONS, LINKS, ..., would be apreciated |
![]() |
![]() |
![]() |
#2 |
The Pro/Life Support System
Join Date: Dec 1999
Location: Denver, CO
Posts: 4,041
|
![]()
Been there done that.
Direct core cooling in inefficent. I dont have time to get into it thismorning but heres the basics: 1. there is a significant loss in surface area to cool, you just have the few cm^2 of area to transfer heat to the coolant. This is EXACTLY the same as running the core in open air with no HSF.. you CAN cool it but you will need an insanely high flow rate of coolant or air to cool it as efficently as with a waterblock. A Cu block acts as a heat spreader as its much more efficient at conducting heat then water of mineral oil. This allows the Cu block to contact more coolant per second, at more reasonable flow rates, and in turn cool better. 2. the leak risk on direct core cooling after a few power cycles are very real. it can be done, but I do feel more confident having a Cu block soldered together Vs. plastic gooped onto my core.
__________________
Joe - I only take this hat off for one thing... ProCooling archive curator and dusty skeleton. |
![]() |
![]() |
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|