If you don't achieve turbulent flow (Reynolds 4'000), then the surface friction is irrelevant: laminar flow at those (low) flow rates doesn't care about a grainy surface. I think...

That's a nice link! I just might use it!

But the texture can be a problem to uncast the part...

If you're trying to make a mold from it, yes. (sneaky devil! :D )

If you need assistance getting this service, PM me, maybe we can arrange something.

EDIT, oh a mold?

I experimented with different base size objects, and up to a 8" x 10" base the only factor in price is the height. $45 is the minimum they charge, and That's up to about 6mm. But it need to be only one part; so it can be modeled several variations of the base, joined by an easy to break small bar.

Great for flow test different variations! :D

Jaydee: Yes, I plan to cast it in silver using jewelers tools.

More exactly, a Girl I know that do jewel stuff, She's..... MMMM :drool: :dome:

Ben, thanks for the offer!

Prolly this is going to take some time, I'm going to do it only when I have a final design. Hope it don't take as much time as radius. :D

Get out of my mind!
 

:eek:

This is similar to an idea I had a while ago while working on Radius.... It would use a modified Radius type top. It would also have a concave central pin thing to "break" the flow.

Bob

He he, I hear you:p

The problem with the spiral is that it leaves no fins in the center, and more importantly, no continous fins (the pin is a substitute, but with partial performance). That's a limitation of the radial design.

It's a critical area of the block, because it's over the core. Without it, you won't even beat Radius:p

What are you going to use for the molding material? What temp does this plastic stuff melt at? What temp does silver melt at? Are you going to have it pressure casted so it gets all the bubbles out and maintains a level of density? How much does casting cost?

Hehe, sorry lots of questions but I am very interested in this. :)

The performance has to be seen. I think this design has less hard direction changes in the flow, and an easier entrance path. That leads to increase overall flowrate. But you pointed the main problem with this designs: As more channels are used, the center empty space is bigger. If they are too few, the pressure drop increase. I found that a good tradeoff was 12 channels, but testing could say something different.


I dont have the answers yet, but this material is going to be for the original, from wich the mold is going to be made, it don't need to handle temperatures. :)

As I told, she is the expert. Now where I left her phone number? :confused:

Ok, if you've got "12" as the optimal # of channels, maybe you could apply what pHaestus stated in his article, about the most efficient configuration of an elbow (i.e. length of sweep).

pHaestus' article

Also, I think you would do well to look into the jet inpingement effect.

I don't quite understand that. It will have to be able NOT to melt when the molding material is applied. The ony way to make a mold is to have something in liquid form that will cool off and harden. This will also not be able to melt when the silver is applied. So in other words it seems to me this plastic prototype needs to have a higher melting temp than silver, because the stuff that the mold will be made of needs to have a higher melting point than silver (obviously). And in order to make the mold the stuff that the mold will be made of will have to be molten, which will have to be at higher temp then what silver will turn to liquid from, or the mold will melt when the silver is cast into it. So in other words this protoype needs to have a very high melting point.

Finally found it. Melting point of pure silver, 1763.2 °F. This means the mold needs to be able to withstand this and a few hundered degrees more as getting silver into a fluid state has to be hotter than melting point. So the mold needs to be able to handle that temp, which means this plastic prototype needs to be able to handle around 2000F.

Of course this is from my very limited knowlegde of casting, the above process could be totally wrong, but I would look into it before you try it.

No no no... You're not thinking 4 dimensionally!

The mold can be used to make a plaster cast, which in turn can be used to make a sand cast, ready for silver:D

You loose some precision because of the multiple steps, but it's cheap and easy, so who cares:)

As I know, plaster can handle silver temps. And plaster cures with water, at ambient. (It is a close relative to concrete) The main deal of most casting materials is being able to replicate a “weak” original with a high strength material. Traditionally most cast iron originals were made from carved wood.

Bruno Facca in his thread explained the characteristics of several casting sands.

Ben; the main factor of flow resistance in an elbow is the ratio between turn radius and pipe diameter. In this case it's not so bad.

The sharpest turn is at the beginning of the channel, close to the core; where it has a useful side effect: Centrifugal force brings the flow close to the wall, reducing the boundary layer.

See? The variable radius flow channel is not casual, there’s lot of things I've thought in this design.

But the opposite is also true: you'll increase the boundary layer on the other side of the channel. Effect cancelled.

Also, in centrifugal flow, the heavier water molecules, which are colder, will be pushed to the outside. That's good if that's where your hot spot is, but that's not the case here: it's on both sides.

I will maintain that the circular motion only does one thing: add to the flow restriction. Further, because your channels are longer, you're adding more flow restriction.

In Radius, the flow restriction is progressively lowered by channel splitting, for an overall minimized impact, so that the flow can reach its highest velocity within the critical area, and within an area that is not too big for the pump to handle (which I have yet to demonstrate, I know...).

Anytime the water has to turn, it's restricted.

I can't find it, but someone mentionned something about capillary channels: as myv65 pointed out, in flow restriction, there comes a point where the channel is so narrow, that the height doesn't matter, the restriction is the same.

You should make that central post a pin, otherwise the incoming coolant will spray all around it, ignoring the pin's hot base.

This is maybe a stupid bet, but I hope that the benefit in the outer side of the channel is more than drawback on the inner side.

About capillars: remember that my channel widht is the same as in radius.

That's on the go, now that I'm free from geometric limitations of milling, I'm going to use full NURBS surfaces modelling software. :cool:
Consider the images I posted as a first rough sketch :D

Ahhhh! I knew there was another way to do it. :D Hummmmm, this is very interesting. I wonder what they would charge for something like this and in a microchannel design like Cathars. Silver WW anyone? :evilaugh:

He he you got me there!

I was really surprised by the flow graphs: I didn't think it was possible that the top of the channels would be stagnant.:shrug:

I investigated sand-casting. The channels are too fine for anything other than a one-off cast as the fins of the mould would snap as they are too deep. At least this is what about 20 different sand-casting places told me before I got the hint...

Man that sucks. I imagine the same would hold true with this aswell.

How about using a nozzel like this:

http://www.customcooledpc.com/nozzel.jpg

and having that center pin just slightly below the bottom of it? Wouldn't that add a lot of volocity right at the center. or would this mess up the flow through the channels? :shrug:

That's because the heavy texture of the sand, And copper can't be casted in plaster. It's melting point is over the point of calcination, where the chemical bond between plaster and water is broken. The water will be released, and the cast become soft powder. This process also absorbs a lot of energy, that's why plaster is used as fire retardant in steel buildings.

Could be good, but may produce a huge pressure drop. :shrug:

A while back Jon Fettig approached me with a very similar idea. He had designed a multi-channel spiral block, but at the time we decided it would be too expensive to manufacture the small channels that he designed. Here's one of his pics.
http://www.logiccooling.com/ocf/jfettig.jpg