What do you think of my waterblock design???

NOTE: This idea was originally inspired by the article "Recycled Heatsink Water Block" written by Joe Citarella of overclockers.com.

The idea that that an air-cooled heatsink can outperform a water-cooled designed heatsink when put in a water-cooled environment stuck in my mind. Because air doesn't conduct heat very well the design of an air-cooled heatsink is based more around cramming as much surface area as possible into a small package while still allowing good air flow. Hmmm.... My MC-462 has alot of surface area....

I started by drawing the heatsink...


I then figured out what pins I would need to remove in order to enclose it all...


As you can see I also had a vague idea of what the enclosure would look like. I liked the idea of having the intake/ output raised so water could spread out in between the pins. The middle would be milled to less than 1/16 above the pins so the water is forced through them.

Here's the design a little further along..


I added the tabs so it could be bolted to the base as well as a slot for a 3" 1/16 diameter o-ring (AS568-041).

Here's a pic of what it might look like when enclosed...



Pros
Lots of surface area!
Good flowrate
Cheap! (I've got all the parts)\

Cons
Lot's of work
Not to sure how it's really gonna work
I'd be hacking my Swifty!!!

LMK what you think!!!

One thing I see I would think about is the fact the water like most things will take the path with the lest resistance. You might want to be sure your design will move water out of the bottem of the block and not just across the top where the ports are.

Dont get me wrong it looks great its just an idea after I gave it some thought

Keep it up

Thanks for the feedback and I agree. I tried to design it so the ports were higher so that the water could spread out among the pins. If you take a look in the middle of the enclosure you'll notice the it is less than 1/16 (more like 1/32) away from the pins, which will force the water down and through the pins. I even thought about sealing w/ a gasket between the middle of the cap and the top of the pins so the water has no choice but to go down.

Whoa Talk about planning things before you do them!! Totally impressive presentation of you idea.

Like Marshall McLuhen said " The medium is the message"

someone has some cad skils

Some stuff I forgot to add...
I had planned on milling the base to .200" (the same as the MCW462). The groove for the o-ring is also planned. I have found that the o-ring I'm using needs to be compressed by 10% and I shall make the groove accordingly. I had planned on using aluminum for the shroud, but powder coating it to avoid the "battery effect". I also will have a sacrificial anode in my reservoir.

Btw, I've never powder coated anything before, is there a special procedure to this?

Ok I had some trouble seeing the clearance in the iso view.

I still see hot water spinning away near the bottom of the block. maybe am wrong. I will need to give it more thought. Wish I had the cad skills.

Oh if I could only draw what am thinking I will give it a try later I guess.

Lets keep up the ideas

Do you think I should remove some pins to increase flow?

one thing I would change is where the intake is

I would put it right above where the cpu sits so its always going to get new water

Honest Answer

I really like your complete design. You went for the most thourough and perfect design, and you acheived it.

The problem is, it is much more than is needed. Engineers are two things, cheap and lazy. Do exactly enough, and no more.

Check out this article under WATER COOLING at overclockers.com:

Danger Den Pin-Fin Waterblock"
Joe Citarella - 1/29/01

Awsome Idea (i was planning to do this once) one problem you would have to deal with. the removed pins would either have to be pulled out (this could prove quite difficult) and the hloes filled and smoothed FLAT!!!! or cut and smoothed FLAT!!!! either way you need it very flat to avoid leaks

Re: Honest Answer

BubbaJoeLouis said:

The problem is, it is much more than is needed. Engineers are two things, cheap and lazy. Do exactly enough, and no more.

Check out this article under WATER COOLING at overclockers.com:

Danger Den Pin-Fin Waterblock"
Joe Citarella - 1/29/01

Click to expand...

I've checked out that article before. That's the one where he makes the pin-fin block and it does no better than the regular one right?. Yeah, it's probably just a pipe dream I ordered a Spir@l waterblock from D-Tek and I'm #64 in line so I thought in the meantime I could play w/ this. It was fun designing it anyway. You can probably tell I'm not an engineer... (I'm a stubborn painter).

Here's a good description of the flaw in my design...

Interesting idea, but not the most efficient. Here's why:

...The convection coefficient of air is much, much lower than that of water. This is the only reason why you require so much surface area for an air-cooled heat sink. With water cooling, much of that area goes to waste. You can think of the overall thermal relationship as an electrical circuit with a few resistors in series. Rtotal = Rcontact + Rconduction + Rconvection. Contact resistance is a function of smoothness and compound. Conduction resistance is a function of material and geometry. Convection resistance is a function of air velocity, surface area, and geometry.

Here's the rub. If any of the three resistances is too high, it makes no difference how low the others are. Overall resistance can be no lower than the smallest of the three. The task, then, is to reduce each resistance to an acceptable value such that no single one causes a problem. With air cooling, the only way to do this is with sufficient air volume/velocity and copious surface area. Using the same surface area with water greatly reduces the convection resistance, but does nothing for the remaining two.

You would do just as well or better to get the water closer to the die.

Now, if you want to get really fancy and have a neat back-up system, you could do the following. Place a conventional hsf combo atop a good water block. If you rig a feedback signal from the pump, you could run the fan on the conventional hsf in the event of a pump failure. This would most certainly not keep your computer stable, but may very well preserve your processor if the pump fails. (This is easier than it appears.)

Also note, there's nothing truly wrong with your general idea. It will work and work well. It's just that it won't be any better than a good water block with no pins (and it'll be a lot bigger).

-myv65
(over at amdmb.com)

dave is a damn good guy to run this stuff by. a very knowlegeable and helpful individual. has troubleshot many a design of mine and schooled me on lots and lots of stuff. definately one of the good guys.

Yeah, that was a humbling experience... In 30 seconds he tore apart an idea that took me many hours to come up w/. I like that though... you gotta learn somehow.

*chuckle*

oh yes. i know that one quite well. he's the one thats been drilling me in Thermodynamic Theory for the past 4 months helping me iron out the details of the Core Project.

humbling... you betcha. great guy to know. i've learned more in one email from that guy then i had from months of being on the forums. you just have to have a pseudo-grasp on the language he's using to break it down for you in the first place - that and the ability to be able to manually put it into the construct you're working within so it all makes sense and you understand the applicability of what he's telling you.

one very smart cookie. no two ways about it.

btw Hatchet

outSTANDING layout and diagraming of your concept. my hats off to you. really, you've done a fantastic job.

what version of CAD were you using curiously?

surface area is the answer to everything

Lol, and I thought the answer to everything was 42...

r0ckstarbob said:

what version of CAD were you using curiously?

Click to expand...

Autodesk Architectural Desktop 3.3.