CPU Cooling

The Borg

Senior member
Apr 9, 2006
494
0
0
Hi there,

I have been playing with water cooling and overclocking for quite some time now. I have come to realise that increasing the number of external heat exchangers and fans does not matter after a certain point. This lead me to think that the limiting factor is the amount of heat that can be removed from the processor itself.

This can be a combination of a number of factors:
1. Heat transfer abilty - therefore use copper, AS5, lapping, etc
2. Medium used - water, anti-freeze, etc
3. Velocity of fluid through heat sink
4. Actual design of the heat sink inners.

Again, this lead me to think: Why can the actual processor chip not be constructed in a way that allows better contact with the heat removal device? Has anyone realised that all these fancy heat sink and water cooling setups only cool ONE SIDE of the die?

Appart from the obvious problems of connections to the processor, why not have the die standing on one edge and a heat sink able for fit around? Instead of a nice flat surface to mate with, how about a proper male / female setup?

Any ideas?
 

VinDSL

Diamond Member
Apr 11, 2006
4,869
1
81
www.lenon.com
Originally posted by: The Borg
Again, this lead me to think: Why can the actual processor chip not be constructed in a way that allows better contact with the heat removal device? Has anyone realised that all these fancy heat sink and water cooling setups only cool ONE SIDE of the die?

Any ideas?
Thermalright IFX-14 :D

I don't see why someone couldn't come up with a water block setup that would do the same thing...
 

aigomorla

CPU, Cases&Cooling Mod PC Gaming Mod Elite Member
Super Moderator
Sep 28, 2005
21,135
3,674
126
Originally posted by: The Borg
Hi there,

I have been playing with water cooling and overclocking for quite some time now. I have come to realise that increasing the number of external heat exchangers and fans does not matter after a certain point. This lead me to think that the limiting factor is the amount of heat that can be removed from the processor itself.

This can be a combination of a number of factors:
1. Heat transfer abilty - therefore use copper, AS5, lapping, etc
2. Medium used - water, anti-freeze, etc
3. Velocity of fluid through heat sink
4. Actual design of the heat sink inners.

Again, this lead me to think: Why can the actual processor chip not be constructed in a way that allows better contact with the heat removal device? Has anyone realised that all these fancy heat sink and water cooling setups only cool ONE SIDE of the die?

Appart from the obvious problems of connections to the processor, why not have the die standing on one edge and a heat sink able for fit around? Instead of a nice flat surface to mate with, how about a proper male / female setup?

Any ideas?

because like everything in the world $$$ dictates standards. It it costs more, with only 10% complain rate, why double the production cost to meet with the 10%.

Sorry the industry is all about greed.

 

VinDSL

Diamond Member
Apr 11, 2006
4,869
1
81
www.lenon.com
Originally posted by: aigomorla
It it costs more, with only 10% complain rate, why double the production cost to meet with the 10%...
Bwahahaha!

You're officially on my 'Buddies' list now... :D
 

Late4Dinner

Member
Apr 3, 2007
42
0
0
I have some serious doubts about that IFX-14. I'm assuming it mounts to the back of the MB (I mean, where else could it go?), but how much CPU heat could possibly be transferred from the core, through the pins, through the socket, through the PCB, to the back of the MB? My guess is not much. To get the most out of something like that you would probably want to fill the socket with a 100% non-conductive thermal compound (straight silicone, AS Ceramique, etc.) prior to installing the CPU and hope it doesn't liquefy under load and drain out.

What might be a superior option, for those crazy enough to try such a thing, would be to form a waterblock around the socket. By epoxying a plastic box to the MB (making the MB the bottom) so as to enclose the socket and plumbing it with coolant lines, your CPU would be completely immersed in coolant. Obviously you would need to use a non-conductive coolant such as mineral oil.

Would anyone around here be crazy enough to try such a thing? I wouldn't be surprised. Lemme know if it works.