Cooling MCP "Crater" Waterblock Concept

Toying around with an idea I've been wanting to get out of my head for a while. I don't have any current plans to water-cool any of my computers, but I wanted to try my hand at designing a waterblock, and get some pratice using CAD software.
The first revision was a really rough model of what I basically wanted to accomplish. A block made without CAD/CAM equipment that would provide lots of turbulence without too much restriction.

Second revision lead to a more evenly sized passageway between all of the fins. as well as relocation of the outlets to the top of the block.

And then I switched to CAD software so that I could start implementing all of the toolwork in detail. Here's the third revision of the base.

The first concept was to plug the center of the block and collimate the inlet into 24 smallet jets. I don't know what I was thinking there because that would create a huge dead spot.

So everything was edited to move the jets (31 of them total) to the middle of the block with matching dimples in the center of the base. I also went back to a single outlet.


The revisions I have in mind at the moment are adding more steps to the underside of the "collimater" and potentially a different inlet shape. If possible I could also make a custom milling bit so that the dimples in the center of the block could be cut as rings instead.


Socket 462/478/754/939/LGA775/940

I need and am open to criticism because I've never designed anything like this before.

 

Hmm. A thought is that after the "impact zone" additional heat removal fins/channels will be superflous. After all, you're just cooling the IHS not the core. For a TEC block, that could be godsend, as the entire element would now be getting equal cooling. One thought though... where's the outlet going to be?

 

The outlet is much easier to see in these images.

Revised mounting bracket which can be attached to the underside, or topside of the block.



The 478/775 parts can also be removed.

 

Outlet position perfect. But still, the value of the outward spinles in normal use is unfortunately questionable. Looks cool yes, but i'm forseeing this block turning a laing d5's flow into that of an aquarius II's...

 

The restriction definitely isn't that bad. The channels in the base are 1.3mm wide and there are more than twice as many of them as similarly configured blocks. I also avoided using narrow inlet nozzles (31 @ 0.63").
The whole point of this block is high turbulence with low restriction. The solution as far as I can see it is keeping the volume of the path of travel the same but adding enough surface variation to reduce the boundary layer as much as possible.

 

ah. My bad. i was thinking the design forced water through the channels than out, rather then hitting the base, cascading up the sides, and being pulled in and around the channels then out

 

Yeah, here's the flow diagram.

x 24

The white part on the left side of the image may or may not be at all vital to the functioning of the block. I'll have to ask a friend or do some simulations to see if it has any effect.

 

only issue i see is that the flow diagram only involves one section of the block's interior. How will you ensure water stays moving equally over the other areas?

 

How about a spirel (sp?) effect? basically have your fins spirel deeper into the block in the direction you want the flow you go. Seems to me that that would utilize all the surface area.

 

concur, dont really see the extra channels outside the jetpipe bits being much use really
looks very tho

 

I improved the underside of the collimater so that it's countour matches the base.

I don't have to coax the water to move away from the center towards the edges, the pressure difference will do that for me. I did however design the collimater so that the outer nozzles are pointed ever-so-slightly away from the center.

That coupled with the white "exit ramp" part should theoretically keep the pressure fairly even all the way around the edge of the base.

Yeah, but that would reguire CAD/CAM, which I would like to avoid for cost reasons.

I'm experimenting with different base design. This one involves far less drilling and gets most of it's shape through blade work.

 

What about just taking the design you started with, and basically instead of turning your "grooves" as you did and do a stair step design . Basically start with one line, go either way and drop the height of each one starting at .005 mm , then .010 mm .. think you get the idea (i hope)

 

Decided to try something other than pins or fins.

 

I've been reading through pro-cooling and I did some more thinking on how to make my optimal design with the tools i have, rather than doing the inverse.

I came up with a different base somewhat inspired by native american patterns. This one will have a "+" shaped inlet nozzle to match up with the start of the channels.

I may revise it to have 45-degree bends in the channels rather than 90-degree.

 

Outside dimensions on this design are: 2.5" Diam. x 1.25" Height .25" base thickness (from underside to top of fins) and will use 1/2' ID tubing.

Hooray! I finally managed to figure out a 45-degree pattern that has the correct spacings. The new design has 10 more channels, with 52 total. All at 1.4mm width.

The collimater has 52 angled nozzles (13 different angles), which I'll have to drill by using a combination of an angle vice and a drill press.

The jets in the center will have the most turbulence do to their entry angle but the center is contacted heavily by the majority of the nozzles.

And I need to design the mounting bracket still, which will bolt flush with the top of the waterblock.

I've cherry-picked some of the best features I've seen from proven heatsinks and combined them into one block. Provided the machining is accurate to the model this could be a very effective block.

I was going to name this "Squaxin" but I think "Nazca" fits better.

 

These designs are so awesome, I love the circular design, so neat.

Will you ever be making any of these?

Just a prototype to see the results, you never know you may have the new Storm G4 in your hands "cha ching".

JCBeastie

 

Name it CaptainSlug . And if you come to actually make it, i'll be the first to buy

 

Had to start over again. I need to keep all of the channels uniform when it comes to restriction and that requires chaning their pattern again.

 

I think you are looking at designs that are way over-complicated and that look like you may be sacrificing a lot of flow and a lot of money for 1-2 C. Try a simpler design that you will be able to save time building. I'm not trying to discourage you, it just looks like there's greater money going to be spent making the block than is worth the few degrees from a complicated design.