yet another new idea, sorry
 

http://www.digital-explosion.co.uk/u.../PaulMK3-2.gif

Basically, thats the base plate, the design in the centre is a group of 4 concentric circles, 2mm wide, 2mm deep into a 3mm base, and 5mm apart, cut in with a lathe.

Circles @ 20mm diameter, 15, 10 and 5

The same set of circles is then cut in again, offset 2.5 mm to the right and left.

Single inlet over the core, then 2 outlets right and left, ala WW.

Any thoughts?

Your design is an interesting way to use alternative tools cutting a WB. Considering your limitations, it's difficult to comment about what you could do :(

The intake jet design will be tricky (just having a "normal" gush in intake would be pointless) since it will also have to help support the centeral area that's been weakened by the circle cuts.

I think a thin cross shaped jet would be ideal, but would be difficult to construct (given it's required structural element)

What sort of inlet design do you reckon Volenti? pics?

this is what I had in mind, the vertical jet may not be needed, but would reduce pressure drop slightly.

They really should be structural, actually making contact with the base.

Actually, thinking about it some more I believe that a simple set of concentric circles(with the cross jet) would perform slightly better (you could use a finer lathe tool and make thinner rings/channels)

May as well try for 1mm rings/channels, also makes it far simpler to machine, since you only set it up on the lathe once.

I'll have a think about how to set up the jet's structurally.

The idea with the circles is that there would be turbulence at every intersection, This pic might show the layout of the channels slightly better

http://www.gta45.pwp.blueyonder.co.uk/sblockbottom2.jpg

I thought about another design, although lack of a central inlet might hinder it slightly.

This would be the base, channels 1mm apart, 2mm wide, 2mm deep into a 3mm base.

http://www.gta45.pwp.blueyonder.co.u...blockbase1.jpg

The inlet and outlet would be mounted on top like this

http://www.gta45.pwp.blueyonder.co.u...blockbase2.jpg

As well as that, another inlet and outlet could be created perpendicular to that, at 3 and 9 o'clock respectivly.

Thoughts on how good a block this would be?

That has potential, with a 3 and 9 o'clock narrow slitted inlet it should give pretty good performance, it also eliminates the structural problems, also don't limit yourself to square channels, you could try triangular cross section rings, U shaped channels ect since you'll have a grind a specific tool for this anyway.

TOP

http://www.gta45.pwp.blueyonder.co.u...etblocktop.jpg

MIDDLE

http://www.gta45.pwp.blueyonder.co.u...lockmiddle.jpg

BASE

http://www.gta45.pwp.blueyonder.co.u...tblockbase.jpg

( co-ordinates starting at bottom left of block, 1mm increments )

Top is 50x50x2mm

- Holes are 15mm diameter
- Hole centres at (25,39) and (25,11)

Middle is 50x50x10mm

- Rectangular slits are 8x23mm
- Corners at

[ 1st set : (21,0) (29,0) (21,23) (29,23) ]
[ 2nd set : (21,50) (29,50) (21,27) (29,27) ]

Base is 50x50x3mm

- Groves are 2mm deep

- Grove placement described (outside,inside) as measured from the centres of the block.

(50,48 )
(46,44)
(42,40)
(38,36)
(34,32)
(30,28 )
(26,24)
(22,20)
(18,16)
(14,12)
(10,8 )
(6,4)

Then a 4mm diameter hole 2 mm deep in the centre of the block

Any suggested improvments?

Looks like this when rendered properly

http://www.gta45.pwp.blueyonder.co.uk/1.gif
http://www.gta45.pwp.blueyonder.co.uk/2.gif
http://www.gta45.pwp.blueyonder.co.uk/3.gif

Puzzling... I can't figure out the flow dynamics:shrug: I think you lost me in the revised idea.:confused:

I think I get it....Water goes in on one half of the circle(s), and follows it around to the other half, where it leaves

If that's the case, you only have half a jet inpingement, because the outlet has the flow reversed. It would also be unecessarily restrictive.

Well, it took a while, but I finally got the first prototype for this.

I've made a few changes to the design very recently, basically increasing the size of the outlet groove to create a larger pressure drop.

But this is what I've got right now, and is basically what the final thing will look like cosmeticly, but with a clear middle plate as well, and the top plate won't have that frosted look, it'll just be totally transparent.

Fits AMD and P4, there's a few issues with that, but as a first prototype, its not bad, after I make the changes I've got planned, should make a reasonable block. I wanna work on a socket tabs holddown for it, because without the top, it looks badass where you can see a lot of the AMD chip because the block itself is so small.

Same base and middle plate, different tops, and it fits NB and GFX cards.

Anyway, enough talking, take a look at the pics:

The base, size comparasion with a quid.

http://www.gta45.pwp.blueyonder.co.uk/sizecomp.jpg

Each component separatly.

http://www.gta45.pwp.blueyonder.co.u...wnexploded.jpg

Base with copper middle plate

http://www.gta45.pwp.blueyonder.co.uk/isowithmiddle.jpg

The assembled block

http://www.gta45.pwp.blueyonder.co.uk/isoassembled.jpg[/

i still can't understand the connection between these 2.

The only problem I see is that the center of the circles isn't getting very much water. Maybe I'm just not thinking about it right or something.

This looks like a good CPU block, and probably an EXCELLENT peltier block.

Ignore that, the one on the left is what I've done, the one on the right is a previous idea, with offset grooves 1/2 a radius each way, I abandoned it.

@ Phamt0m51 actaully, the centre will get the most water, because its the path of least resistance for the water. The smallest groove gets the most flow, then the second smallest, then the third smallest, and so on to the largest groove. The idea is that there should be more flow closest to the heatscource, although none of the grooves are exactly miles away from the core.

@Althornin : It wouldn't be that great as a TEC block, its smaller than a peltier ;) But yes, expanding the idea to more circular grooves wouldn't be a problem, and probably a decent pelt block.

I did'nt think this would be that good last time you posted about it, but seeing it in the flesh it looks better. It may not be a WW killer but it looks OK to me, impingment or not. What other machinery did you use apart from a lathe?, it all looks like it was made with TLC...

just an idea. have you ever thought of putting some dimplets for turbulance in the smallest circle?

How are you planngin to seal it?
Looking very good, btw :)

@ MadDogMe

Its definatly not a WW killer, but I'm happy with the design. Its hard to say one design is better than another, the WW fulfills its design requirements wonderfully, but my requirements were different. For a start it had to be less that 45mmx45mm to fit on GFX cards.

Base was made on a lathe, then drilled and tapped holes in each corner. Middle plate was manually milled for the slots, then drilled in each corner again. Top is drilled, tapped holes with a lathed o-ring groove ( not pictured )

@ Balinju

Don't actually need to, and probably couldn't anyway. The base is VERY thin, and the lathing process leaves a slightly round surface are the bottom of each channel anyway. ( slight ridges, about 0.1-0.2mm high and wide, perdendicular to the channel walls, the ridges are about 0.5mm appart. I nice little side effect in my opinion.

@Kwissus

Sealed with an o-ring on the base and one on the top plate. The machinist forgot to put an o-ring groove on the top plate this time, but its only a prototype, so I'm not that bothered, I'm just going to make a gasket from silicone sealant, and clamp down on that with the socket screws.

done any tests to see how it performs yet GTA? :) is this the block you were talking about taking into production? its looking interesting

Yes, this block is going into production, even if it is a limited run where I lose money on each and every block.

I've always wanted to produce and sell a block to other people, its a dream of mine, sad as it may seem, and I'm prepared to pay for that dream to become reality.

I take it you are the same king Richard from OcUK, and if you want to post this there, go for it, and if you want to buy or test this block, just ask, and I'll send it to you.

As posted before, this is not a WW killer, but at a guess, within 3 degrees is not a target too high to aim for. This means it should beat a Maze3 by a couple of degrees.

This all remains to be seem though, this prototype is NOT complete by any means, but its close, and the results should be reasonable.

let me know, gta@digital-explosion.co.uk, or find me on the forums at the same page.

Maze 3? maybe you should be more concerned with the Maze 4?