Stacking the blades on an arbor with spacers will keep the project more simplified. You will cut down the number of cutting operations by half or more.

Having had a few chunks thrown at me in the past, I most strongly concur!:eek:

I was thinking more of useing the dremel as a router, moving the router not the block, BB just mentioned he had the C~drill as well as the router atatchment. I don't think he planned on using it except 'maybe' to bore holes then join then by hand with the router...

I think everyone has tried drilling a hole in something while holding it with their hands once..., once only! :p . especialy Cu!!, it's a grabber...

I finished hand drawing the fin pattern, by hand, which took me a few hours. I'll scan and post it tmo.

The twin blade idea would actually save some time, but I understand that it's hard to see the necessity without this drawing.



Fixittt, I still need to know what size blade (20, 25 or 32mm) you need, to make a 5mm deep cut. The collateral damage to the other fins depends on the size of this blade. Either way, it looks like finishing this thing with a 1 mm endmill is the best approach. I can still minimize the use of this small endmill, but it looks like a requirement, for tiny, tiny little finishing touches.

In case it's not clear, I suggest cutting the 5mm channels, as much as the round saw blade can be used, without cutting into other channels, then cutting the rest of the channels with a 1 mm endmill. If you're still up to doing this, I would of course get these cutting bits (blades, endmills) to you. I am still offering to compensate you, if you want to make this block.

I still don't know what level of precicion you're capable of. So far, we have a 1.5 mm wide fin, and a 1.0 mm wide channel. The angle of the cuts is full circle, in 11.25 degree steps, with off-center offsets of 1/2 the fin width, or +/- 0.75 mm. Does this fall within your level of precision? What's your margin of error, or tolerance?



From looking closely at the design, I've observed that it's not optimized for uniform flow, but that as long as the nozzle isn't smaller than 3/8 inch, it should still function nicely. I say 3/8 because that's the point where the tip of fin set#3 is visible from a 3/8 opening, allowing the coolant to seperate into 16 channels pretty much equally. The flow should otherwise "theoretically" split in a 1.5 to 1 ratio, at the tip of fin set #4.

It looks like I'd also need a set of "drop-in" nozzles, so that I can run tests at different flow rates. I'll post specs later.



I agree that the "turntable" would be pretty rough. Definitely, the block would have had to be bolted to this table: I'm not that silly! Judging from the sheer density of the copper, and with the little (not nil) experience I have cutting metals, I wouldn't trust it to stay still without a sturdy hold-down.

just get me the drawing of the block how you want it, or how you think it can be done. I review it, and have a friend of mine review it. Together I am sure we can come up with something.

Sound good?

Sounds good, Fixittt!

Here's my hand drawn rendition of the fin pattern:
Fin width: 1.5 mm
Channel width: 1.0 mm
pattern radius: 20 mm (increased from 15)

Utabintarbo is working on the CAD drawing.

B&W but still psychadelic:D I think you can scrap the little bits at the outer edge of the circle, not much use for them and they just complicate the settup for making it.

Of course! They have no function, given the distance from the core. They are only the result of the 1.0 mm channel cut, with a pattern radius of 20 mm (diameter = 40mm). With a radius of 15mm (diameter = 30 mm), they do not exist/appear.

They don't complicate the setup, they're just there! It would be more work to cut them off.

You are making two cuts instead of one by having them there though.

No, it's two cuts no matter what. It would involve a third cut to take it out.

Refer to the colored drawing: there is one cut on each side of fin set #4 (red). Since the cut is 1.0 mm wide, the result is fin set #5 (the posts).

Here's Utabintarbo's CAD representation:

Fixittt, what now?

PS: the fin pattern has been increased to a radius of 20mm (diameter = 40 mm). This is not necessary, but it's there so that it can all be more easily understood. 30mm diameter would not include the posts at diameter 30.

Do you see what I see after some quick paint skills to the lower portion of your design? Pairs of parallel cuts with small pie wedges left at the outer edge which does not sit over the die area. What is wrong with simplifying in this way? Everyone always loves to dig out the KISS addage ya know:)

I see!

Yes, that would be fine. It's outside of the 15 mm radius, so the effect is negligeable.

I shall adjust the design for your perusal and edification! :D Stay tuned!

Bob

email me a DXF file of the drawing.
fixittt@cimtel.net

I talked to paul today (My master machinist friend) and he said he will try anything. As a personal favor to me. I feel I am gonna have to cash in a whole lot of favors for this one. Machining time is gonna be a B!t@H But hey now you have me wrapped up in this thing as much as U are.

U do know that if it is a good preformer, that I am gonna have to make at least 3 or 4 of them. One for you, one for me, and one for your fab. Cad designer. None other then the great utabintarbo !!!

I also want to try and make one in Plexy. Solid acrylic. So we can see how flow runs thure the block by adding small air bulles into the water.

Sound good?

I second the idea, Fixittt. If nothing else, acrylic is cheap and cuts fast so you can setup the machinery for a test run. :D

The only draw back is that acrylic gets gummy when hot and we are talking about a .040 in endmill here.

Geesh kinda scary! But I would love to see it etched out of acrylic!!!!

Come on guys lets put this into perspective. a 1 mm endmill has the thickness of a dime. the human hair is .004 and paper is .003 to .004 in.

Im wondering if we can get away with either a 3/64 or 1/16 if you want to know the MM equal remember the chart I posted in the machining preameters.

Fixittt: doesn't a lower rotation rate help to alleviate the melting of the acrylic?
Besides, we're all cooling geeks, I'm sure we can find a way to keep it cool enough to make chips without the melting...:D

Polycarbonate machines much better without the gumming like acrylic.;)

First of all, thanks for the compliment. I am unworthy of it, however. I am not great, just really, really, really, really good! LOL!:D ;) (God, I make myself ill!)

What if we made the base in copper and fab'd up a housing in acrylic? Would that fill the bill? We could epoxy/jbweld it to the base. Whaddya think?

Bob

ohh I know a way to help keep it cool when machining. I put ice in my coolant bucket. Then add a little dish soap to the mix.

Hey utabintarbo You and I need to have a solid works show off.

I need some web space to host pics!!!

Would'nt it be easier to make if the channels were where the fins are and the fins were where the channels are?. if you know what I mean?,lot better for flow I'd of thought, get rid of the solid 'cross' in the middlebetter still have it like a pie sliced into very thin slices. you coul;d always rework the optimisation out to suit it could'nt you?...

Thanks to everyone for your help, this couldn't be done without you!
CAD drawings and renders: Utabintarbo
Milling: Fixittt
Cube res drawing: Morphling1
cutting blade link: gone_fishin

and everyone else who participated: Puzzdre, MeltMan, myv65, Cathar, MadDogMe, fresno12, Can O' Beans, Dix Dogfight, Arcturius, N8 and #Rotor.

Fixittt, yeah, I guess that we could run some flow test, in an all acrylic prototype block. I already have a fair idea of how it's going to turn out, but I could be off: I'm not briefed on fluid dynamics beyond what plumbers should be. I was thinking about throwing some sparkly stuff in the water (inspired by my stepdaughter:D )!

I've cut acrylic panels before, with a jigsaw, and I found that there is a specific range of feed rate that will keep the blade clear. Trial and error?

3/64 is 1.19mm. 1/16 is 1.587 mm. I've already established that 1/16 is too much. At 3/64, the 4th set of fins (psychadelic red)turns into "pointy thing", plus it will throw off the nozzle. I'd rather stick to 1 mm, as much as possible.

A dime, huh? That would make a neat saw blade... (NOT!)

Some progress: I picked up a foot of braided 1 1/2 ID (2" OD) tubing. It came on a roll so I have to straighten it out, but otherwise, that's the flow outlet.

MadDogMe: I thought about that, but the amount of copper left from all the cross cuts, would leave the core bare. There's no point in having any fins at all, if they don't cover the core.

Another update: Utabintarbo and I have been discussing some minor items.

1-I haven't specified a mount. I intend to go with a bolt mount. However, I believe that if the block is kept to the size of the socket, that an Innovatech type hold-down could be used, but that's not where I'm headed.

2-See http://forums.procooling.com/vbb/sho...&threadid=4324 . Paul Vodrazka has some interesting thoughts about the O-ring. We'll go with a 1/16 o-ring. I have to admit, the Swiftech gasket looks appealing...

3-(ref #2) I've asked Utabintarbo to make the top secured with 8 bolts, instead of 4.

4-Minor mods to the top (outlets). I figured that the outer channel would be "turned", where Utabintarbo saw it milled with a ball end endmill. I figured that the outside channel would have a radius of 5 mm. The outlets should be at the top of that curve. Since the outlet tube is now 1 1/2 ID, the top will be slightly modified. No effect on flow, but the fin pattern is going back to a 30 mm diameter.

We both agree about one thing: the top is part of the structural support, in addition to the fins themselves. I wouldn't want the baseplate to start caving in, even by a fraction of a mm, but neither of us is a mechanical engineer, so we just don't know for sure if an acrylic top would be sufficient, or if a copper one would be required, and in either case, at what thickness. Anyone?

Here's the latest render:

It may just live after all:) BTW nice show of class giving honorable mention to the help others provided;)

If you continue on with the KISS concept wherever you can, consider foregoing the ever popular o-ring channel and opt for a thin clear silicone gasket sheeting. I believe I posted a source here at one time. One less, and really unnecessary, milling operation. This also insures no cross channel leakage like can happen with an o-ring if it is not fully compressed.

DO you mean KISS or KISS A**;)

Another simplification I've been thinking about: there is a cut on each side of the first set of fins (set#1, psychadelic orange). These channels could be cut with the cutting wheels, leaving a curved pattern in the center, instead of going back with a 1 mm endmill, to cut it out.

It would make some weird kind of pyramid, about 5 mm high, and would optimize flow at that point.

The 1 mm endmill will still be needed, to finish the other channel junctions.

Searching for "silicone gasket"...

I found 3 references:

http://forums.procooling.com/vbb/sho...ilicone+gasket

http://forums.procooling.com/vbb/sho...ilicone+gasket

http://forums.procooling.com/vbb/sho...ilicone+gasket

but they're all from JayDee116!

This turned out almost the same I had in mind including piramid in the centre which I believe it's not a bad thing.
Now how the hell am I supose to come out with another original idea that would work, cathar came up with his and now you got the whole team behind you will beat me with this one :shrug:
Oh well I guess I'll focus on graphic card blocks...
As my man Arnie said I'll be back :D :cool:

I posted about it here, sorry no link but it is widely available at any hardware outlet.
In discussion here

A quote from myself,
""Transparent Silicone
Durometer 40A
Tensile Strength, psi 800
Elongation % 250

Get a clear view of your materials through this transparent rubber. Sheets are covered with a protective clear film (remove film before using). Rubber has a temperature range of -80° to +425° F. "

That was straight from the spec sheet of a supplier. It is not impossible to find and goes for about $12.00 per square foot at 1mm thickness. That will get you 24 gaskets at 2" X 3" so I don't see it being a huge cost but rather a cost saver. If you get it adhesive backed, it's a couple dollars more per square foot." end quote

From that discussion you should be able to know what to look for.

Morphling1, don't despair, give up the ship, hang it up, throw in the towel..... there is always room for improvement.:D