Advice on "My" design idea..

[Myth]

Hi

I have been considering a "new" design for at block the past week, and would like to hear your opinions about it.

It's basically a Cascade and Pin Fin combination, and the idea was to archive a good combo block working at both low and high flow rates.

I'm sorry for my lacking painting skills, but hope you can get the idea from my attached drawing..

The gray dots are groves, and the black are fins.

I'm thinking of making a top with "nozzels" or "jets" so the water always hits the groves.

Would this work?

[Senater_Cache]

It would work in the sense that it would let water pass through itself. The performance you must test yourself with a prototype

[jaydee]

The majority of those pins are useless. Look at the MCW6000 design for a revision or my R-Type block.

[Myth]

I guess my drawing is also real bad.. On the block, there isn't enough room for that many anyway.

[Techmasta]

You could draw it on a piece of paper and scan it.

[Myth]

I'm trying desperatly to draw the sucker in a CAD proggy i found, and also to use real sizes so ppl can get the general idea of it..

[Myth]

Now i got a revised concept drawing done by a friend, in this prototype we have removed a lot of "fins" and made the body round.. thinking of making a larger block with 2 rounds of fins and then the "flow shield"

Think this would work?

Any design suggestions are welcome!

[Dieter@be]

Yes it "works"...ofcourse it works (it will cool down your chip), but the question is: "how well should it work?"
The concept is pretty well imho (although if the "grooves" part is above the core, then the fins wont dissipate much heat).
But i think it's very important that you can "guide" the water into the grooves, because -unless you use some kind of nozzles, the water that needs to get away will disturb the incoming water so performance will drop. (that's what i think ofcourse)

[Myth]

The reason for the fins was that i hoped that would help getting some of the heat from especially A64 and P4's with heat spreaders away, die cooling would require a large area on those i belive..

I was planing on jets aiming at the groves, or a "nozzle" doing the same..

[Dieter@be]

Oh ok then it should be ok, although it the cascade design is really hard to make (i believe it was jaydee who made 9 nine prototypes and then decided to stop, also the "real" cascade had many many prototypes)

[jaydee]

Originally Posted by Myth

The reason for the fins was that i hoped that would help getting some of the heat from especially A64 and P4's with heat spreaders away, die cooling would require a large area on those i belive..

I was planing on jets aiming at the groves, or a "nozzle" doing the same..

Well there is some useless stuff there. Those pins really can be removed if the jet impingment is used correctly. No real reason to aim jets at the holes as the gap between the jet and those cups will get filled with water and the jet stream will not even make it to the cup. You need to put the jet INTO the cup to make it really worth while. Or you can replace those cups with pins instead. As mentioned it will work, just a matter of how well. Hard to say much without a prototype or major thermal and liquid dynamics training. I have been one to trial and error with working prototypes but most people want a good block their first run. Hard to do without the education to do such a thing.

I would say make the jets Cascade style and get rid of the pins and that divider deal. Water will find it's way out without it.

[scopEDog]

If i might add something...I see that you have xp/p4/a64 mounting holes. The problem i ran into with my prototype was that on an xp board the other mounting holes would interfere with things around the socket (ie. capacitors, etc..). I've found it tough to come up with an all-in-one mounting solution for multiple cpus.

[Myth]

#11

We have talkeed about the same problems regarding the "holes" that you point out, and have decided to drop them and go for a different design. Also we decided to simply use some time making prototypes so we can get some real experience before hopefully ending up with a design that will have the wanted performance.. Thank you for pointing out the errors in the design.. Guess i was so caught up in the original idea, that i forgot some of the basics :-(

#12

The problem of getting into the way of cap's is one we are aware of, it's a pain in the a** because it will have to be a really high solution to fit abouve those.. I have an idea, but will need to make some messurements and tests to see if it will work..