Stupid WB Idea?

[CheeseBall]

Here is the idea. It is very simple and practical. But since no one has done anything like it maybe before maybe there's a reason. You are the experts.

First, I was thinking about 2 things: simple to make. And I was also thinking, "with every other waterblock your only using the bottom half, not much heat would be transfered to the top half" (especially if it's plastic). I came up with this: machine a piece of copper to about 2.5"L X 2"W x 1"H(whatever), then drill 2 wholes, side by side, centered, all the way through the whole legth.

Here's a nice drawing or somethin: (view from top)
_________________
| |
|--------------------------|
one hole Water in from a T
|--------------------------|
|--------------------------|
other hole Water in from a T
|--------------------------|
|_________________|

After going through it would rejoin with another T.

The advantages of this design are:
1) Most heat will travel below and right up between the
holes/water.
2) The heat can also travel above the holes/water since it's
one solid piece of copper.
3) easy.
4) Water gets in and out quickley

There is not a whole lotta surface area and resistance so maybe it's going to suck. Let me know what you think. And if this idea is infact the most retarded thing you've ever heard of, please give me suggestions. I will have access to lathe, mill, cnc, and I'm pretty good with the manual stuff, and I can get help with CNC

[LiquidRulez]

No..I dont think its retarded. I say go ahead and try it, and see what you get.I would put some sort of recesses in the base at the bottom of the internal side, to create more surface area and turbulance to aid in heat transfer to the water.
Youre just going to suffer like the rest of us and try out your ideas........
Thats the whole process in a nut shell basically....trial and error

Havr fun !

[koslov]

I think you're thinking of a cross-drilled waterblock. These were very popular during the "early years" of watercooling. The design was perfected, most say, by xjinn.

[CheeseBall]

There's not really a practical way I could put recesses internally and still keep it one solid piece of copper. Do you think having one solid piece (no top and bottom) will help a good amount or not? I might go ahead and make this one then make a different block w/ a small center spiral (water coming in from above cpu), pins over die. I'll post a drawering (drar-ring if you perfer. I do) later.

Drar-rings to the rescue! What show or movie did mike myers say drawering alot? I think it was SNL and he was a lil kid in the tub. Pretty funny.

[msv]

http://www.procooling.com/reviews/ht...art1-1_2_0.php

Cool Computers made a WB approximately Your described design, and it performed quite well.
Mikael S.

[bigben2k]

The Xjinn block:



In your implementation, you have 2 channels. You might consider a different channel arrangement: most of the heat will travel up the center mass, leaving a big difference in temp between the coolant, and the die contact area. This is why baseplate thickness is so important.

Xjinn's is overkill. Thermal simulations have shown that the top 2 rows don't add any significant cooling, and the middle one barely does anything at all.

[jaydee]

I always had a theory about cross drilled blocks. Why not TAP the holes afterwards? Would create some serious turbulance. I know this is near impossible in a Xjinn block but with biggers holes.....

[jaydee]

I just had a hell of an idea! Why not drill a hole right down the center (offset to sit over the top of the CPU if needed) of the peice of copper from left to right, tap the hole to create micro channels and then drill three holes strait down to meet it from the top for a single inlet and duel outlet configuration!!!! Holy s$#% I think that would work well!

[CheeseBall]

Not a bad idea atall, the tapping would probably help alot, and if I drilled through the top maybe I could make a few pins there. I would drill part way through the top of the block, take an 1/8" endmill and machine down as far as I want to go leaving somes pins over the die, then cross drill it from each side up to where the pins start. Get it?

Do you think having 1 inlet on top, and 2 outlets on either side would be better then having 2 cross-drilled holes placed right over the top and bottom of the die so most of the heat goes right between the 2 holes? The holes as I think you would understand would be fed by a T and come together w/ another T.

One last thing, BigBen I don't totaly understand xjinn's design. About "Swiss Cheese" effect, I don't see why there's holes in the water channel. Water would leak out. I know I am missing something, what? Does he have 2 channels? Why are they so far apart? Also, if that is another channel the other one dosn't look swiss cheesed.

Thanks,
Bryan

[jaydee]

For best convection you want the inlet right above the core and the cool water pushing down. You have to think backwards a little. Don't think of the heat from the CPU rising, think of the coolness of the water spreading downward towards the core.

Here is drawing of what I am thinking about doing for my REV. 3.0 block.

[bigben2k]

Cheeseball: the pic only shows a drilled block. The big holes are meant to be inlet/outlets. The coolant flows from one side to the other, through the small holes. Also, the tiny holes on the side of the block are capped (with a copper plate).