Please judge this design!

[hara]

I need comments and critique on this design. What is your opinion?

1mm channels

Number of channels and dimensions are not finalised. Is the concept worth a shot?

It is a dual inlet, dual outlet design. It relies on lateral impingement to do the job.

Thanks

[hara]

Top view. Remember, number of channels and their dimensions aren't finalised.

[Rayman2k2]

are you going to use jet impingmnet? if so, I would get rid of piece of copper in the middle, that way you dont loose alot of you impingment on something that might hinder performance...

[hara]

It is a dual inlet, dual outlet design. It relies on lateral impingement to do the job.

[bigben2k]

There is no lateral inpingement: you haven't met the conditions for such.

It's an ok design. It'll perform better than a straight lateral design, but it won't beat WW.

Similar to Radius, in some ways.

[Rayman2k2]

Originally posted by hara

hahaha :shrug: my bad...

[hara]

Originally posted by bigben2k
There is no lateral inpingement: you haven't met the conditions for such.

It's an ok design. It'll perform better than a straight lateral design, but it won't beat WW.

Similar to Radius, in some ways.

http://dictionary.reference.com/search?q=impingement

Isn't water colliding when turning the 90 degree angle?

[Arcturius]

Won't this be awfully restrictive, with four waterpaths containing a 90 Deg turn, each? (Though flow should be exceptionally well balanced, due to that...)

[hara]

I think that the turns would add the restriction of an elbow. I think that would not be a problem though. What are the conditions for impingement?

[bigben2k]

Originally posted by hara
http://dictionary.reference.com/search?q=impingement

Isn't water colliding when turning the 90 degree angle?

Yes and no: it's being diverted, which implies *some* collision, but it's not "jet inpingement".

You also have an issue with the water entering the fin area: that's the equivalent of an orifice plate, and is restrictive.

Also, the first 90 is well over the core, and so is the second, but where does the third fall?

Don't give up: it's a good design. Maybe 3 * 120 degree turns would be interesting.

[hara]

Originally posted by bigben2k
[You also have an issue with the water entering the fin area: that's the equivalent of an orifice plate, and is restrictive.

I don't understand

[8-Ball]

I think what he is saying is that the end of the fins, where the water enters is half blocked off, ie a sudden reduction in cross-sectional area.

There was a thread earlier which mentioned that various curved profiles would reduce the pressure drop associated with an orifice plate, allowing for a greater velocity.

Is this what you mean BB2k?

8-ball

[bigben2k]

Yes!

[hara]

Now I understand. But that's not a big problem. It can be rectified by making the fins pointed.

[Cathar]

The design is extremely similar to a concept that I prototyped on my toy thermal simulator.

Yes, there's impingement in the true sense of the word, the issue being that the impingement isn't exactly where you want it. "Jet impingement" means narrowing the inlet orifice to boost water velocity.

A simplistic appraisal:

The heat is coming from the bottom. Your impingement action is coming from the side. Much of that impingement action in your design therefore occurs high up the copper walls. Now since the total waterflow is being distributed over the walls, the thermal gradient through the walls pretty much means that not all of the water is being used to cool the hottest bits of the block being the base of the block and the bottom of the channel walls.

The design will perform okay. I'd predict slightly better than Swiftech MCW5000 performance.

Please note that this is just a casual obervation backed with some simulation results.

[8-Ball]

exactly, but it's by fine tuning these little things that blocks get good.

That's why forums are so great, provided you can accept constructive criticism and solve problems as they are pointed out. I'd say your pretty good at that.

8-ball

[8-Ball]

Sorry, my last post was meant to read before cathar's, and is not related to his post.

I gotta start using the quote feature more often.

8-ball

[hara]

Originally posted by Cathar
The design is extremely similar to a concept that I prototyped on my toy thermal simulator.

Yes, there's impingement in the true sense of the word, the issue being that the impingement isn't exactly where you want it. "Jet impingement" means narrowing the inlet orifice to boost water velocity.

A simplistic appraisal:

The heat is coming from the bottom. Your impingement action is coming from the side. Much of that impingement action in your design therefore occurs high up the copper walls. Now since the total waterflow is being distributed over the walls, the thermal gradient through the walls pretty much means that not all of the water is being used to cool the hottest bits of the block being the base of the block and the bottom of the channel walls.

The design will perform okay. I'd predict slightly better than Swiftech MCW5000 performance.

Please note that this is just a casual obervation backed with some simulation results.

I understand your reasoning. How much do you think there is space for optimisation with the current design or is it worth it to start over?

[Cathar]

Originally posted by hara
I understand your reasoning. How much do you think there is space for optimisation with the current design or is it worth it to start over?

Sent PM.

[Edit: oops - no I didn't!] :shrug:

Hmmm, about the only optimisation I can easily think of is making everything small, thereby becoming a true micro-channel design, but in that instance, just having parallel flows (no turns) would work better from a pressure drop perspective.

[Rayman2k2]

Originally posted by Cathar
Sent PM.

[Edit: oops - no I didn't!] :shrug:

Hmmm, about the only optimisation I can easily think of is making everything small, thereby becoming a true micro-channel design, but in that instance, just having parallel flows (no turns) would work better from a pressure drop perspective.

and then you'd have cathar's block, well, pretty much