• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Direct die wc on a P4?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Sssnake

Registered
Joined
Dec 3, 2003
Location
sweden
Just had this crazy idea, that i wanted to share with u all.

Now direct die watercooling has been a pain with amd cpu´s aswell as older intel ones. But now that intel has been kind enough to cover the die with a peice of copper, maybe one could try to watercool the IHS directly?
My idea was THIS:
Get a peice of plexi/lexan thats just as big as the cpu,make a inlet at the top,possibly with small drilled jettubes,make 2 or mabye even 4 outlets down near the base, mill out a perfect fit with the cpu at the base. Use something like marine goop or any other good water-resistant glue to bond the acrylic wb to the cpu. Maybe apply some 2-component epoxy on the outside, to make a good seal.
If all goes well u would have a jet-imprigment (allmost)direct die waterblock.(without the minicups, unless someone is crazy/brave enough to drill them into the IHS).

So what do u all think? Pure fantasy/wishful thinking or might it actually work?
Does anyone know anyone else who have tried this?

thanksss
 
I would definately drill some small cups or make some channels on the IHS or you may get worse temps than a good waterblock, simply because there is not much (realative) surface area on the IHS compared to a good waterblock for the water to interact with. So, if you created some surface area increasing structures it seems like it could work well.
 
squeakygeek is correct, you need the surface area and everything for it to perform, the surface are thats impossible to get out of that thin of metal.

your better off removing the heatspreader and using a waterblock on the core directly or just on the heatspreader.


Jon
 
WAIT

WAIT


drilling the IHS, or even machining it, i've never thought about that


the thing IS as thick as the cascade bottom right?


this could work, especially on the prescott and the M0 cores


why would the water block be better, assuming fine channels or cups are cut or drilled into the ihs itself?
 
fafnir said:
the thing IS as thick as the cascade bottom right?

Nope its not. Its nowhere near thick enough to do anything like your suggesting and anyway if you take all the time to mod the spreader you may aswell take it off completly and use a block straight on the core. That would get better results.
 
Well the whole point was to uyse the IHS so i would get the typical problems ppl have gotten with bare die cooling b4. Maybe one could use a needle file or a dremel to make a kind of grid onto the IHS? If i do that and get enough water speed(usinfg the jet tubes and 4 outlets) maybe? This is just a thought tho....
It has been fun getting ur opinions on the matter :)
thanksss
 
hehe, 1 think i kinda overlooked... Might be prudent to seal up the litle hole in the corner of the IHS. Didnt think about that...
 
One thing that you should bare in mind is that the IHS needs pressure on it to contact the die underneath. It relys on pressure from the hsink or waterblock to deflect it slightly so it makes full contact with the die. So if you direct die you dont have that pressure and the IHS will not make full contact which will lead to a huge increse in its thermal resistance and high temps.

Take a look at procooling.com forums. There is a large thread covering this issue.
 
its just that i've been trying pretty hard trying to pop a prescott heatspreader safely and its a no-go so far,

was just exploring the idea, anything that might work
 
yes i have thought of that, just now actually. One could make some kind of retention mechanism as the ones used on normal waterblocks. a solid block of acrylic, even tho its preforated slightliy, should be enough to put pressure on the ihs, provided one mills out a small shelf/ledge inside the block, wich would rest against the top of the IHS´s outer edge, dont u think?
 
Humm that may help a little but the pressure relly needs to be over the die for it to work which ofcorse is not possible witch DD. Then again maybe it is not as big problem with presshots due to the die being bonded to the IHS?
 
Sorry couldnt resist. And no im probably not gonna do this. It was just an idea i had, that i liked to get some feedback on. so... thanks .
 
Back