Direct Die AMD Watercooling

**TheDude** · 06-10-2003, 09:32 AM

Someone in another thread mentioned the possiblity of direct die watercooling (think it was PimpJack). Anyhow...couldn't stop thinking about it. I have some old gear lying around and think I will try it just to see what happens. I would think (if it works) that it would offer much better cooling. As long as everything is sealed good and tight with heat resistant gasket material, I don't see why it would not work? Here's a pic of what I will try:

**wdd1040** · 06-10-2003, 10:18 AM

...

**RichBa5tard** · 06-10-2003, 10:48 AM

uhm, you're going to watercool the entire cpu surfance? I think that will shortcircuit the bridges and components on top. The previous direct die coolers i've seen only cooled the die itself.

**Major** · 06-10-2003, 10:57 AM

Hey The Dude good luck !

Like RichBa5tard said, all of the direct die cooling I've seen only put water on the die, and not the chip itself. Maybe if you find something "permenant" to protect the traces and components on the top of the chip you rig will work. Maybe a liberal application of RTV silicon might stay in place and keep it dry.

Also direct die cooling requires a very high flow rate due to the small surface area.

**bowman1964** · 06-10-2003, 11:24 AM

good luck my friend....i have seen it done.it works pretty well.the angle of water inlet should keep the coolant forced on the die itself.
now be sure to remove all the oils from the cpu and then spray the top with spray clear paint to seal of the pores.then remove the paint from the die itself with paint thinner or something.sometimes it will get hot at first (on the top of the die)and peal off the paint impairing the coolant hitting the die.
then use a non hardening adhesive.the best i have seen it marine goop,sold at lowes ,walmart,etc.
this stuff is fully water proof,but will not tear when you pull on it like silicone will.i have streched a small piece before like a rubber band.it is awsome.that is the stuff i used to seal the temp probe in the top of your chiller tank.

good luck..it will work..

**PimpJack** · 06-10-2003, 11:32 AM

Awsome DUDE! I'm glad u will explore direct die. Take a look at procooling's forum (water/phase change cooling) and you will see some cool ways of using direct die cooling. I don't think anyone has done it on AMD so u will be the first in my book. P4's seem to be easier since they don't have any traces. If u seal the "block" tight I wouldn't worry about killing the cpu as long as traces are sealed. Also do keep the water stream close and directly on the die. Good luck and keep us posted.

**TheDude** · 06-10-2003, 11:49 AM

Thanks for all the great input and tips guys! Like I said...it's going to all be gear that I can afford to screw up if it happens. We shall see.

**antipop** · 06-10-2003, 12:04 PM

Here what some guy made ome while ago
http://www.hardforum.com/showthread.php?threadid=513957
I know it's [h] but the guy did made a good job

**TheDude** · 06-10-2003, 12:17 PM

Nice long link Antipop...I will read it all...should be some good info in there. Well, I had never heard of anyone doing it before, guess I should have known better! I don't think I have ever had an original thought in my life! Still going to do it tho...mine will be different than his. His is a P4...maybe I will have the 1st AMD as PimpJack says?

Bowman...good tips all my friend, but especially the marine goop...I was wondering what to use that would be up to the job. I do plan to cool the entire surface, sealing all the traces and as Bowman suggested sealing entire surface first. I imagine it will take some experimenting to get the flow and pressure just right. I will prep a dead chip first for practice before risking a good one. It should take some time to complete along with everything else going on in my life, but I will post any progress ( or failures).

**antipop** · 06-10-2003, 12:35 PM

In fact when i search the forum, i found some other thread about direct die cooling but i didn't read them, i just remebered that one. Try to search for direct-die
You should find something that will seal the bridge and don't risk to be removed by the pressure of the water. Keep us informed on how it goes

**gobbo** · 06-10-2003, 01:44 PM

Go for it Dude!!

Lets see what the results are from it!

**saaya** · 06-10-2003, 02:55 PM

i dont think itll work very well since the cpus surface is really small.

good luck nevertheless

!

**KnightElite** · 06-10-2003, 03:23 PM

Actually, I've read about this being done a while back. In fact, one watercooling company sold blocks specifically for direct die AMD watercooling. Things to remember though, are to make sure you use something to cover up the CPU surface other than the die itself... to make sure you don't short out any of the bridges or whatever. Second thing is that eventually the CPU will erode away, and will stop working, but you get good performance up until that happens

.

**TheDude** · 06-10-2003, 03:58 PM

Originally posted by saaya
i dont think itll work very well since the cpus surface is really small.

good luck nevertheless !

True the surface is small, but consider that you are not trying to transfer heat thru a block and thermal grease but directly into the water. If I can get everything right, it should cool better than a block.
And I realize all the bridges and traces have to be and stay sealed. The water may erode the chip after a while, but I imagine that would take a very long time. Anyway...this is all just for fun.

**Major** · 06-10-2003, 04:06 PM

I wonder how much room there is too play with on the die, maybe drill some small dimples in the surface of the die? hehe

**TheDude** · 06-10-2003, 04:21 PM

Originally posted by Major Slaughter
I wonder how much room there is too play with on the die, maybe drill some small dimples in the surface of the die? hehe

I think it would crack pretty easy.

**saaya** · 06-10-2003, 05:13 PM

Anyway...this is all just for fun.

hehehe yeah thats true, if you loose a chip just get a new 1700+ for 40$, that probably even ocs higher than the other one you just lost... i LOVE amd

**NoFuture** · 06-10-2003, 06:39 PM

Why don't you glue an small heatsink (ramsink?) on your core, I think that would improve performance a lot!

**TheDude** · 06-10-2003, 07:31 PM

Originally posted by NoFuture
Why don't you glue an small heatsink (ramsink?) on your core, I think that would improve performance a lot!

Actually....I was thinking of trying something along those lines...something like that would produce some turbulence with the fins. I tend to think that just a direct flow over the die itself will prove best. I want to try it without anything first to see how it does, but that's a good idea. I might stick one on there to see if there is an improvement. That would mean using 2 chips instead of one because I doubt I will be able to separate the chip from the top block without damaging it. That means also that I need two identical chips to judge any improvement. The cost just doubled.

**k1114** · 06-10-2003, 08:52 PM

You're absolutely insane, and I mean that in a good way

**Major** · 06-10-2003, 08:57 PM

Originally posted by k1114
You're absolutely insane, and I mean that in a good way

I agree, In a good way !

**IFMU** · 06-10-2003, 09:43 PM

This is something thats been done before. Ive seen it quite a bit actually when I first started looking into water cooling. Its been some time since its been really looked into though, to many possible drawbacks from it. Cant recall all exactly.?

Ok, first thing I can say that folks would use clear finger nail polish on the bridges. Beautiful sealent. Thin, but still worked.
I dont think that anyone would actually put something on the die itself. Water would go directly onto it. Nothing to inhibit heat removal from the die to the water.
The biggest drawback I recall folks had with direct die cooling is that it was so difficult to attach the "block" to the CPU itself. Very hard to attach.

Now, this is something I had actually started thinking about a lil bit off and on recently. One thing I think would really help with it would be to use a type of setup like Cathars new design with the Cascade. Use the jet style spray to hit directly on the die itself.

Def something Im gonna watch here.

**antipop** · 06-11-2003, 01:16 AM

Dude, you'll also want to start the water before the comp, otherwise you'll just burn the chip
The heatsink is not a really good idea IMHO. Direct water would be more efficient

**TheDude** · 06-11-2003, 04:07 AM

Dude, you'll also want to start the water before the comp, otherwise you'll just burn the chip

I will try to remember that one

**DR. YT** · 06-11-2003, 04:17 AM

ive also heard after prolonged use the chip will eventualy absorb the water. and i beleive theres a thed someware about using a spray style block.

Thread: Direct Die AMD Watercooling

Thread Tools

Search Thread

Rate This Thread

Display

Direct Die AMD Watercooling

Bookmarks

Bookmarks

Posting Permissions