Base thickness on a specific block

[1398342003]

This is the base design of a WB I want to make.

All I really need is a reference thickness. Dimensions are 80MM square, 38MM deep, the pins are 5MM.

Please bear in mind that this is 6 months old, and one has been changed slightly, but not significantly, and another has changed radically.

I need my website PW but I forgot it. (I'll look for it to give updated pics)

/EDIT the large dots are Socket A mounting holes. The smaller bolts are for the cap/lid.

[Tuff]

copper=thinner the better...Aluminum..has to be thicker.

Tuff

[Monster811]

On the block design:

1) Remove the first and last 3 rows of pins, all they do is resist flow.

2) put the outlet on the opposite side from the inlet, using those long channels creates unneeded flow resistance.

[8-Ball]

Originally posted by Tuff
copper=thinner the better...Aluminum..has to be thicker.

Tuff

Not a general rule, and often not the case. It depends very much on the style of waterblock.

Cathar managed to utilise a 1mm baseplate only by creating an extremely efficient design and combining this with a centre inlet forcing water down onto the base, enhanced by jet impingement.

Bigben2k seemed to know quite a bit about it and should be able to explain better if he happens to see this.

8-ball

[1398342003]

Sorry, I forgot to mention that the construction material is copper.

1) Remove the first and last 3 rows of pins, all they do is resist flow.

2) put the outlet on the opposite side from the inlet, using those long channels creates unneeded flow resistance.

1) Do you mean the first 1 and last 3 or the first 3 and last 3?
2) The PSU is too close to the block for that, and what with the depth, it will not restrict flow, but ease it.

[bigben2k]

Originally posted by 8-Ball
Bigben2k seemed to know quite a bit about it and should be able to explain better if he happens to see this.

Flattered, I would attempt, if it was relevant. I almost passed over this thread without reading it: I enjoy creative usernames

First, what was meant about the pins is indeed the first three rows, and the last three rows. You might want to take a look at the AMD specifications, for core measurements.

This block is another good attempt, and will indeed work well. Where it is limited, is again, in that the flow is crosswise, not a central inlet, which concentrates the heat removal from the pins, and not the baseplate, where most of the heat will come from.

Since the flow is quite directional, I don't see any reason to pass over the idea of fins, instead of pins (which would be just as easy to make). In fact, the whole block dimension could be reduced significantly, by having a single outlet on the opposite side. It would also reduce the flow restriction significantly.

This design is very large, I suspect because of the unusual inlet/outlet configuration. If I didn't know better, I'd say this is for a rackmount PC, where the height is severely limited.

[1398342003]

The username was a temporary form of personal ID, now obsolete.

If I were to use fins instead of pins they could be more dense, and easier to make.

The outlets do need to be pointed in the same general direction as the inlet, as said, pressed for space.

How about angling the inlet 45 degrees?


Just wondering, but what would having a central inlet and 4 corner outlets, with fins, do to performance in the block. (like a LARGE version of Bladerunner's WB)

[GTA]

That would make a very very nice GFX block, as the inlet, as both outlets kind of go into the side of the block.

Thats probably what you were doing it for I presume?

[1398342003]

Actually it's a CPU block. Look at the dimensions. 80MM Square, or 3.14 inches. I was going to make a GPU block, but CPU cooling is more important to me.

[1398342003]

Would that be better? (With more fins that are shorter and thiner than shown.)

[Monster811]

Does it have to be low-profile?

[1398342003]

Yes it does.

[LiquidRulez]

What are planning on mounting this on??

I dont think it will be a great performer at all.
I think the design is pathetic and quite boring.
Keep trying though.

[8-Ball]

Originally posted by LiquidRulez
What are planning on mounting this on??

I dont think it will be a great performer at all.
I think the design is pathetic and quite boring.
Keep trying though.

A little harsh maybe?

It's an interesting problem really, as people have become so obsessed with microfins, jet impingement and centre inlets that they can't see any other way of doing it. (Mind you, that's probably me being a little narrow minded at this point.)

In case you hadn't noticed, this has to be a low profile design which immediately limits your options, as it has to be a side entry design.

Perhaps some positive suggestions?

8-ball

[hara]

Maybe a #rotor style block would help you as it doesn't need a centre inlet for optimal performance.

[scopEDog]

Originally posted by 1398342003
Just wondering, but what would having a central inlet and 4 corner outlets, with fins, do to performance in the block. (like a LARGE version of Bladerunner's WB)

You mean something like this? Maybe it'll give you some ideas..

[bigben2k]

Originally posted by 1398342003


Would that be better? (With more fins that are shorter and thiner than shown.)

I like this design best, because the barbs can be almost down to 0 degrees, aka flat with the base.

If you used 3/8 tubing, you should get a very nice overall fit, while keeping it all very slim.

Also, with this design, you can keep the baseplate down to 2 or 3 mm, with ~4 mm tall fins, keeping it all under 1/8" thick (except the barbs), and top

[1398342003]

Originally posted by scopEDog
You mean something like this? Maybe it'll give you some ideas..

Yeah, similar to that, but with lots of small pins, and a 90o sweep inlet, and the outlets given 45o sweeps to be level

BB2K here is what I meant by smaller fins.

Or would the larger fins be better in this case?

[8-Ball]

If you have access to CNC, then you could make the fins slightly wavy. I would assume that this would help to break down laminar flow going across straight fins.

8-ball

[bigben2k]

Nice, but hard to make waves in that thin of fins.

More fins is always better. Just keep the ratio of fin-to-channel between 0.75 and 1.5; you're otherwise probably limited by your capabilities.

[1398342003]

Sadly I have no access to a CNC or a mill in general. I have to use a drill press. The fins will be slightly wavy because of that.

So the gaps should be the same size as the fins? Like 2mm to 2mm?

[bigben2k]

Not bad, but you might also consider this fine tool that Volenti uses.

[1398342003]

I had planned to make the walls a part of the base, and have the top as a seperate piece. If I used that tool the base would have the fins, and the top would have the walls. Wouldn't that require twice the copper, and therefore cost about twice as much?

[bigben2k]

In Volenti's case, he prefers to use a clear material for the top, in more than one layer: much simpler to build.

The outside wall of a block does not cool the CPU anyways, especially with a bp that thin.

[1398342003]

I fear cracked tops, and they generally have to be thicker than their metal counterparts.