I designed a waterblock

[AndrewZorn]

Go on to page 2 to see the redesign, the one that is hopefully feasible!

Am I worth my own thread?
I've been working on this for a while (~3days, many hours though). Learning experience, fun, and I might actually get to make it someday.
It's Fuzionlike because the Fuzion is good and because I own it.




Need to add threading and get exact sizes for fittings.


Notice the different sized return holes in the midplate, to account for the distance to outlet. Fancy directional fillets.


Is this impossible to CNC? Everything is 1mm+. All cuts are either horizontal or circular. Need to align the circular pattern to vertical.


Nozzling


Exploded. O-Rings are rectangular to show that they will be compressed to the rectangular channels once assembled.


To show detail of top plate. Observe generous rounding and opposite return hole diverge...


Cross-section of midplate, to show reverse-funneling effect of the under-loft. Water velocity is highest at the lower edge/circle, and it is forced down and outward.


Need to make a more universal plate. This is S775. I'll probably make the plate thicker and fuller to avoid it bending (like I did to mine).


Plate fits over/around block detail and screws. It's free-moving. Need to add tolerances to the whole thing (the crosshatch looking stuff is basically where two surfaces meet, that plate won't fit on there with things EXACT).


PERFECT LAPPING ZOMG Can't figure out if I should conceal the screws. I want to, and I can...

Is this good? Is it worthy? Will it be possible to make?

While I worked really hard to balance the flow with COSMOSFloXpress, it crashes when I have the base's pins/detail turned on. Not sure if it's machine stability yet, but the base is symmetric, so it shouldn't change from a flat inner base.
EDIT ok so it keeps crashing now. I wanted to get a picture, because it looked really cool. Basically I made it account for the fact that different corners of the block have a higher travel distance to the outlet. The outlet also has a 'suction' effect on it, or a lower pressure than the inlet. Doing this, I balanced the holes so that the liquid's travel was evenly spaced and the speed vectors were approximately equal (even cooling). It was awesome! I don't know what I changed in the design that crashes SW when starting CFX, it used to run fine.

MS Paint scales/compresses horribly. SolidWorks render's dont look very good, I need a better program.

[Conumdrum]

Need more pins for more heat transfer methinks, but neato project you have been making.

[AndrewZorn]

I was going to make it more intricate, but I'm really worried of working really hard to hear "yeah, so nothing but supernice corporation equipment can cut that tight".

(definitely interested in more pins though if that's the consensus)

[Martinm210]

Looks good, give it to one of our CNC guys like iandh to evaluate if it's constructable. I think anything is doable, but I think generally if it requires the really small endmills that are under a couple of millimeters in diameter, they break really often and that's why we typically see slotting saw type designs when you want something really micro.

Solid works looks cool...looks like a really fun tool!

[AndrewZorn]

SW is really awesome. I had a 'class' on it in college (I put 'class' in 'these' because it consisted of doing the built-in tutorials... no instruction) and just really like it. It's satisfying when you finally get to extrude a sketch!

I have so many questions...

Can't figure out how to thread/the thread dimensions...
General make-sense dimensions:
- I made the block 50mm x 50mm. Fuzion is 55x55 but that's because the screws are out to the side.
- 1mm thick base. Can I go lower? All the pins/walls collide with the midplate's loft so that should provide lots of support.
- Should I step the base?
- 1mm thick pins/walls. Can I go lower?
- M2 screws to hold it together should be OK right? I want to stick to METRIC/ANSI ONLY (to ENSURE availability) and because my block is all within a square, I need 'thinner' ones.
- Are overall dimensions normalish?

Because I practiced AWESOME engineering procedures, it's actually really easy to change any of this. I don't have everything 100% linked up... for instance, I can't just change one screwhole size and EVERYTHING changes or adjust the inlet size and the nozzle fixes accordingly... the software is still kind of cryptic sometimes, and it goes NUTS if you are referencing too much stuff and/or overdefining. However I did use mirroring and referencing whenever possible, so I'm not afraid of criticism that would make me change things. The only thing I'm pretty dead-set on is the flow-equalizing design of the midplate, which I'm waiting for someone to tell me about its wrongness. I think it's a really good idea and spent about half the entire time on the midplate and flow testing.

EDIT and SolidWorks just crashes now trying to do CFX. It's not system stability. I don't know whether to be glad or not. It worked earlier today!

[NaeKuh]

that base almost looks fusion and enzoish.

[bluphysted]

Do you use any simulation on the waterflow? Will Cosmos simulate that?

[ThrustSSC]

Just thinking, your pins dont look quite efficient.

[AndrewZorn]

Do you use any simulation on the waterflow? Will Cosmos simulate that?

That's what it does, that's what I meant...

Just thinking, your pins dont look quite efficient.

Again, should I make them smaller and more dense? I have to keep manufacturability in mind.

[WaterFlex]

How much will it cost ?

[iandh]

Looks good, give it to one of our CNC guys like iandh to evaluate if it's constructable. I think anything is doable, but I think generally if it requires the really small endmills that are under a couple of millimeters in diameter, they break really often and that's why we typically see slotting saw type designs when you want something really micro.

Solid works looks cool...looks like a really fun tool!

This pretty much sums it up.


This is possible but very irritating and slow. Lots of tooling cost.

[AndrewZorn]

Which is the difficult part?
The pins are hopefully easy enough, as I'm about to start reducing them in size/distance apart...
The 'walls', with the exception of the outermost few, are simply circles with some sections cut out.

I'll fix it if I can figure out what to fix.

I tried to be very conservative... everything in the base is 1mm or greater (the gaps and the solid) and I'm hearing that they aren't small/dense enough. Is a cross-cut diamond design pretty much the best thing to do when trying to make something on a non-business scale like this?

My goal was to make it easier than this (didn't I?):



EDIT I have 7 center pins in diameter, the Fuzion has 11. So it's easier than the Fuzion. This says two things: that I can improve the efficiency, but I'm nearing the manufacturing process of a Fuzion. Is that bad?

[ShArKo]

very nice men........I liked...

[MpG]

The Fuzion is drop-forged, if I'm not mistaken, which basically works around a lot of machining limitations.

[AndrewZorn]

The Fuzion is drop-forged, if I'm not mistaken, which basically works around a lot of machining limitations.

Worth noting is the method in which the base plate has been made. Rather than the traditionally more expensive CNC-machining used in most designs, D-TEK has chosen to create their base plate with the help of a copper cast mold. Undoubtedly this allows D-TEK to offer this waterblock at a much more competitive price point than would be possible with the machining that would be otherwise required.

Well, that sucks. Wish I could have imagined that as being a possibility...

I guess I'll start the base over. It didn't take near as long as the other stuff, like the midplate and top, that are the prouder points of my design anyway.

EDIT

and the injection-molded, 2-part shell is extremely durable.

Wow. Great. So I've basically designed something all-around that requires astronomical amounts of money to make?

[Waterlogged]

The Fuzion is drop-forged, if I'm not mistaken, which basically works around a lot of machining limitations.

Actually, it's a blend of casting and forging. A few of us got into a pretty serious "debate" a while ago about that very thing as I asserted my belief that it was forged as well and "others" stated that it was cast. I can't seem to fined the post ATM and I need to leave for work shortly. If no one finds it by time I get back, I'll start looking for it again.

[Martinm210]

Yeah, I believe the fuzion and MC-TDX amongst others is a special casting process. You can make that fairly complex yet constructable for reasonable costs at high volumes.

I think pretty much all of the machined bases in commercial blocks these days use jewlers slotting saws to make either micro channels or variations of square or diamond shaped pins.

iandh is the only one here with cnc experience in the forums. You could always bring some pictures to your local machine shop and ask them though to give you a quote.

Someone did that with a different design a while back on OCnet, and they gave hime a $1000 quote so I know it's very expensive.

All I know is I've made nearly $1/hr on the half a dozen job I manually milled and simply refuse to do much more except for the occasional friend project. It's incredibly time consuming on a manual machine and just not worth the time it takes.

But one offs are alot of fun especially if you have some means like a friend with the machinery to tinker yourself...

[AndrewZorn]

I'm not sure I understand. If someone has a CNC mill, what's the worry about how long it might take, or how irritating it might be? As iandh said, the tooling cost + time, right?

I gave up trying to use the hole wizard, ignoring threading, etc... I'm making my own! And wow, it's hard... I looked up the BSP standards, made a profile sketch of one thread, and accurately extruded along a helical path to make my own threading! It's a little difficult to find out what most manufacturers use for fitting dimensions. Here's what I've got as a "safely-long" fitting:



Here's my main sketch of my fitting. It's a little long, but again, I want to make sure almost any fitting will go all the way in without tapping my inlet/outlet an absurd amount. Here is Swiftech's fitting, which is the 'deepest' I could find. The threading I've matched, all that's important is that theirs is ~6.85mm from the bottom to the underside of the hula-hoop part where the o-ring resides... ~6.85mm deep. All the other ones, based on the thread count, don't seem near that.

To anyone curious, here are the websites I found that gave me enough information to construct the barb and holes:
http://www.btr.nl/pages/gb/produkte/...aee4625b175f59 - Inner/outer diameter of threading
http://en.wikipedia.org/wiki/ISO_228 - Pitch and threadform

[iandh]

The problem that you run into here is that to do these bases even semi-cost effectively you need a small, very high speed, high precision machine. With a cutter at .040", you have to be running screaming fast spindle speeds (10k rpm +) to be able to get any machining done without breaking a cutter every three seconds.

Most shops have heavy machines with 4k or 6k spindles, and they really hum once they get spinning at full speed, on all but the highest end machines. You can't have that when working on this small of a scale.

Any shops that did have a suitable machine for production use would probably have it assigned to a certain part type, so it would be tough finding time on one, or finding one at a privately owned shop that works for the pubilc.



That's the main reason slots are used on non cast blocks. Blocks like the supreme and GTZ push the edge of profitability and manufacturability for that matter, at least using conventional machining methods.

[AndrewZorn]

Would a slot-cut base as detailed as one of those (that has no outer walls like mine, so you could cut straight across) cost a lot to have machined?
EDIT the EK Supreme has walls, was it still slot-cut?

Guess I'm not doing pins/radial like I expected...

Are the midplate and top doable? Does the fact that they are acetal/delrin make it 'easier' to cut them, and thus be able to have the level of detail in those models?

[Martinm210]

Yeah, I would go forth and have fun in the software, just be aware of the fabrication limitations if your intent is to really make some of these. Most folks on these forums that have an inkling for block design sooner or later end up making one of their own. It's alot of fun and a great learning experience. Unfortunately along with that fun comes the reality that profit margins on commercial blocks is pretty small and only pencils out in high volumes.

I did it more for the personal satisfaction that I could, here was my block:




It now sit quietly in my junk drawer, but it was still alot of fun..

[RickCain]

The inlet and outlet appear to be so close compression fittings will not work. I would take this into consideration!

[AndrewZorn]

Does anyone use compression fittings?
Surely not many blocks allow for them then, because it's in the traditional diagonal design. I know mine are very close together (Fuzion).
I'm working on that spacing anyway...

I'm not trying to make money off it, I just think it'd be neat to do and have. If the top/midplate are machineable, that is... that's really what I'm proud of, how I've equalized the flow without adding two outlet ports.

[DeathWalking]

Only if he cares about compression fittings. From the barb he created, it appears that he wants to go with regular fittings.

[CyberDruid]

I think you should work on the spacing of the inlet outlet as well. It's my one gripe about the FuZion: it's almost impossible to use 1/2ID-3/4OD tube and get clamps on. BTW Bitspower is reviving interest in compression fittings. I never liked them until I saw the Bitspower set ups in the Water Cooled Case Gallery here. Take a look at the recent builds: lots of compression fittings being used now.