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

Waterblock design - comments?

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

ArE_eYe_SeE_kAy

Member
Joined
Jun 25, 2003
Location
Grand Rapids, MI
I've got a real bug to build my own waterblock since i have a cnc at my disposal and use AutoCAD everyday. I have looked at a ton of your home made designs that i've seen posted here in this forum and i'm dying to try this. I was hoping that i could get some comments on this preliminary design:

Waterblock-or.jpg


The top piece is 3/8" clear lexan with 3 holes. 1 in the middle for cold water in, and (obviously) the other for warm out. It's odd shaped compaired to the copper block because i'm planning on drilling a hole in each side out and not over the block to accept UV LED's.

It's a 1/4" thick 2-3/4" x 3" solid copper block. All the cutting is going to be done with a 3/16" bit. The blue channels are going to be 3/32" deep and the red channels are going to be 5/32" deep. Hopefully this will speed the flow through the longer channel and slow down the flow in the shorter channel giving them both equal time to accept heat.

Any comments will be real helpful.
 
someone else on this forum made a very similar block to that, yes its very restrictive. Its not the best design out there, I can tell you that. I think you might have made it much too complicated, There are very simple designs that will perform well. It is just how you use them.

You have all your structures far away from the core, only the walls around the inlet will do any cooling. All the cooling can be done within 2mm of the core.

it sounds like you got some kind of decent cnc? I would get ahold of some small end mills, 1/16" or so and design something with lots of surface area over the core.


Jon
 
I think it would make an exellent block if you use a Peltier, plenty of coverage for the entire thing, and with the channels being as wide as they are (not to mention 4 of them) that should make up for all the corners.
 
I'd suggest putting a spike in the middle of your centre inlet. I say a spike - if you imagine a pillar with the top rounded off or like the centre of a lemon juice dish, it should add a little extra to the performance if the spike is above the cpu core.
 
There is too much restriction in that block for simply cooling the core. It would be a better design if you were planning to use it with a peltier. For CPU cooling, the majority of the paths in this block are unneccesary.

To get the best performance, and with the mill you have at your disposal, you want to concentrate at increasing the amount of interaction between water and block in the area directly over the core. If the waterblock is to work well, then the heat will not migrate out from the core-area on the base very far at all and hence your concentration with surface area and waterflow needs to be in this area.

Hopefully this will speed the flow through the longer channel and slow down the flow in the shorter channel giving them both equal time to accept heat.

Now for the blue channel/red channel problem. First of all, your application will do the opposite of what you want to do with it. Making the blue channel smaller than the red channel will increase pressure and velocity in the shorter blue channel. The water will flow through the blue channels even faster and spend less time there.

Now the second problem here is this: the important factor is the volume of water that is traveling through your block, not the quantity of time the water spends in your block. You want fresh cool water continuously entering the block while the warmed water is continuously leaving your block. There is not a time requirement for water picking up heat - The only requirement there is, is for a temperature differential. Hence you want the largest temperature differential possible, which means you want to bring the cold water from the rad to the hot block as fast and often as possible. You do not want the water to spend more time in the block or radiator "absorbing" or "dispersing" heat.
 
Back