The "Cascade" - mini-cup directed jet-impingement block design (56K warn)

[MadDogMe]

How about using a ruler for us non americans Cathar? or do you want me to post you a penny ...

[leejsmith]

how well would this work on a p4 system ?

from the pictures it looks like the tubes are the same length as the polycarb is thick or is the o-ring lifting the polycarb of the desktop and the tubes are longer but you just cant see it.

would it be possible for you to post some pics of the jets in action without the base plate.

Lee

[Cathar]

Originally posted by MadDogMe
How about using a ruler for us non americans Cathar? or do you want me to post you a penny ...

A "dime" is 18mm in diameter.

[Cathar]

Originally posted by leejsmith
how well would this work on a p4 system ?

from the pictures it looks like the tubes are the same length as the polycarb is thick or is the o-ring lifting the polycarb of the desktop and the tubes are longer but you just cant see it.

would it be possible for you to post some pics of the jets in action without the base plate.

Should work well on a P4 system in the same way that the White Water works well on a P4. It's not necessary to focus on cooling the entire 30x30mm area. The jet cups cool the hottest section of the CPU. The oval cavity is 30mm wide and not very deep. The out-flow from the cups practically races across the surface of the copper acting somewhat like what we've come to expect from standard maze-designs in terms of cooling in this area. This means that the block is capable of offering focussed cooling where it's most needed, and still offers good cooling for residual heat. The Cascade block would make a half decent TEC block, unlike the White Water.

[Edit: Oops - forgot to answer the rest of the question]

The tubes are slightly longer than the polycarb is thick. The tubes actually extend into the holes in the copper base somewhat. The O-ring when the plate is by itself, causing the tubes to not touch the desktop so it's a little hard to see.

I'll try to get some pictures of the top two plates in action after I clear the nozzles.

[Balinju]

it's an amazing block

[satanicoo]

its perfect! it has the looks and the performance!

[Cathar]

Well picked up a 0.8mm drill bit and did the jet tubes again properly. Spoke with the machinists about what I did to fix it up and they'll program it in. Basically need to drill the holes once, and then go back over them again to clean them out once they cool down.

As a result, the block is now much more free flowing than before, and has picked up some performance as a result.

Straight away on the first remount I'm seeing a full 2.0C clear lead over the best result I had previously gotten with the White Water over 12 mounts of that block. This is using a T'Bred B at 2400MHz/1.85v running BurnK7. This is significant because it was my design goal to gain a full 2C, and secondly, it's just the first time I've mounted the block. History tells me that I rarely get it right first time, so there may be a little bit more hiding there as I remount to test again.

I snapped some pictures of the jet tubes in action without the base-plate using an Eheim 1048, and a close-up of the assembled block which is intended to highlight that the jet tubes actually do go some way into the cup holes, as some people have mentioned that it wasn't that clear what was going on there.

[Since87]

Very impressive. I really didn't think that the White Water design was going to be beat by anything that would be practical to machine. The fact that you've got better performance and (potentially) better manufacturability is truly outstanding.

Have you talked to any injection molders yet?

I did some googling since I last mentioned injection molding, and not only can polycarbonate be injection molded, it is especially good for molding components with fine features. I have no idea of the relative expense of molding polycarbonate, as compared to other plastics, but (considering the frenzy over WW availablity) I think this could definitely sell at volumes that would justify molded parts.

[Cathar]

I chatted with the machinists about it since they make plastic injection molds quite often. We're talking >$10K to get an injection mold made up for something this fine. For the top-plate, it's a snap.

I'll have to see what the final cost is to get the polycarb plates laser cut to shape and machined, as opposed to injection molding (cost of mold + casting costs). If the cost return is occurring at ~250 blocks, it may be worthwhile. If it occurs at ~1000 blocks, then it's a little "out there".

I suppose if it's molded, the tubes could be circular, rather than hexagons. I'll look into it for sure.

[Since87]

Originally posted by Cathar

I suppose if it's molded, the tubes could be circular, rather than hexagons.

Yes, lots of smooth curves and tapers become available when you go to molded parts.

[nicozeg]

Cathar, once you have a final prototype, it's relatively easy to replicate in small numbers in a DIY way.

The first step is to make a silicone mould from the original part, and then cast copies in resin, curing it inside a pressure pot. This way, any bubble that may be trapped inside is shrinked to a microscopic size. A decent pressure pot cost less than U$ 100, the materials cost is very low; the only other thing needed is experience. You could ask at someone good at model making to make them for you.

[Cathar]

Been running some tests with a single outlet blocked off, so the block is basically running as a traditional one-in/one-out block.

Performance is basically the same within what I can tell. Restriction is slightly increased, but I can fix this easy by milling the middle plate's oval cavity slightly deeper which gives more room for the water to flow. Basically the block is looking good to act as a one-in/one-out block, which makes me happy.

[Since87]

Just wondering...

Do you think it would be beneficial to chamfer the tops of the cups to allow water to exit easier?

Just shallow drilling with a larger drill bit at the top of each cup...

[Zymrgy]

Not sure what I think about the whole single inlet/outlet change. Ture, it is more traditional. However, one of the main reasons I borrowed off of your design was it is a great way to have your block serve as a manifold...one side supplying the HD or whatever & the other cooling the vid card. Still, it is interesting that just plugging one ot the outlets shows no performance decrease. I woulda though with the water not circulating on the side that you plugged it woulda shown a hit.

[Cathar]

Originally posted by Since87
Just wondering...

Do you think it would be beneficial to chamfer the tops of the cups to allow water to exit easier?

Just shallow drilling with a larger drill bit at the top of each cup...

I had the chamfering on the earlier Rev 1 block. Made no discernable difference. The Rev 2 block actually has more clearance between the cup walls and the jet tube walls than the Rev 1 block did.

Still, I'll do it on this prototype again and check the flow rate before and after, since the design is finer. It may make a difference in terms of lowering overall block flow restriction.

[Cathar]

Originally posted by Zymrgy
Still, it is interesting that just plugging one ot the outlets shows no performance decrease. I woulda though with the water not circulating on the side that you plugged it woulda shown a hit.

When the block is first installed and the pump is turned on, you can watch the air-bubbles rush through the block. The side that's blocked actually does still get a lot of flowing water. It shoots out towards where the outlet was, and then splits into two arcs and loops back around the outer edge of the elliptical cavity towards the single outlet. Actually it looks pretty cool to watch it in action.

Still, there is no heat making it that far. No significant heat is making it further than about 1.5mm away from the CPU in any direction, laterally or vertically. This is why the block's performance is unaffected. I was really only concerned if the block would become more restrictive and slow down flow rates, but it seems that this hasn't happened to any large degree, and I can even fix the degree to which it is happening presently.

[bigben2k]

I wish there was a way to do an inlet/outlet scheme similar to Radius, where one is incorporated within the other. I might give it a shot, when I have tools and time...:shrug:

[Blackeagle]

Man, great gains Cathar!

That very fine pattern of cups is impressive indeed. And a full 2C gain over the WW along with the good looks this block has will insure it is a top seller for a long time. And the injection molding, with round micro tube/injectors, might gian a fraction more in performance.

[Cathar]

Okay, got some final pictures. I'll be getting a batch of these made up in about two weeks.

The silver plate is a P4 mounting bracket which fits over the top of the block and is held on with longer clamping screws (just like what was done for the White Water). The final P4 bracket will have a chamfer on the inner oval edge which isn't there on this prototype bracket. This gives a much cleaner look.

[utabintarbo]

I prefer the 2 outlet design, but purely on esthetic grounds.

Has any further thought gone into the mold idea for the "jet plate"? I still think it could be cost-effective.

Bob

[Cathar]

Originally posted by utabintarbo
I prefer the 2 outlet design, but purely on esthetic grounds.

Has any further thought gone into the mold idea for the "jet plate"? I still think it could be cost-effective.

Bob

Had a talk to a number of injection molding people. The jet tubes were the stumbling block every time. My machinists were happy to make a mold for me for a very reasonable price, but the practicalities of getting the fluid polycarb to fill out the jet tubes was something that we couldn't seem to solve, even with air drafting (I think that was the term) to help the flow. Trying to fill 52x3mm long tubes that are 0.3mm wide on the walls and hollow in the middle, and positioned sideways to the injection point (as they would need to be) was something that all agreed would just end up in a high percentage of malformed tubes.

[bigben2k]

Given the velocity of the water through the small tubes, I'm guessing that the inside surface/wall would have to be smooth (as molding would provide), otherwise it could become an additional flow restriction (as laser etching would do).

Is anyone following my train of thought here?

[hara]

Maybe you should find a different kind of plastic. IMO, what's the use of polycarb if you can't see the fine details?

[Cathar]

Originally posted by hara
Maybe you should find a different kind of plastic. IMO, what's the use of polycarb if you can't see the fine details?

Oh, but you can see the details, especially when water is flowing.

An important aspect of the clear tops, not just for appearance, is that when servicing and/or assembling the block, it allows you to see that the tubes are aligned properly into the holes. The clearances are very tight but there's always some small movement available (~0.2mm) when clamping the block together and that's enough to get the jets mis-aligned, so being able to see the inside of the block when assembling really is a good thing.

[hara]

Maybe it's the photos then. What I meant also is that there could be a plastic that would be "moldable" enough, so that a high persentage of the tubes are perfect.