CNC milling machine at my disposal
 

Hey guys, i'm a new member, just joined today. The thing is i've been planning to create my own waterblock for awhile now. It's for a future system, out in nov. Anyways, i've seen how you guys are amazing with autoCAD and whatnot, but are still using manual milling machines. My dad is a high school teacher, and has 4 CNC milling machines that can machine on 4 axis at his disposal. I'm open to suggestions about it. The machines at 1/38th a millimeter precision, so you guys can try out designs. I'd be willing to machine a few cooling blocks. Unfortunately, i have no resevoir or water cooling setup. I'll try to find something to test them on. Maybe just waterflow alone. Also material i machine with will PROBABLY be steel. Copper is pretty expensive for so many water blocks just to experiment with 0.o Anyways, if you can design it, i can probably make it. Please keep in mind its a CNC, so i need a format compatible with quickCAM. I don't seem keen on redrawing all those designs =p Also i don't think i can ship them out to you guys. Its a school CNC machine, profit or anything would probably get someone arrested :D

Re: CNC milling machine at my disposal
 

steel?

Alum maybe? its only about 75 cents a pound.

Re: CNC milling machine at my disposal
 

What if copper is provided?

Re: CNC milling machine at my disposal
 

alum is only 75 cents a pound? 0.o are you sure about that? i thought it was 4 dollars a pound... well depending on quality i guess. Yeah i'm pretty sure i can't do copper, even if its provided. I'm in toronto canada lol, plus i can't ship it back out, cause that'd mean i'm using the machines for my personal benefit, probably get in crap for that or something. anyways, i'll probably use steel just as a benchmark, then engineer the best waterblock with copper? i'll see what happens, then post a result. if you're going to submit a waterblock design, please remember its only a 4 axis machine. means i can't do those crater designs. well atleast not in ways that i know of.

Re: CNC milling machine at my disposal
 

really... use aluminum... its cheap and good

i am not too familiar with CNC but i heard if you are only making one piece, CNC actually takes longer. i dont know. if i were you, i would only use the CNC machine if the piece is impossible to machine on a mill or lathe.

Re: CNC milling machine at my disposal
 

yes its true CNC machines do take longer to program, especially if they are on single run items. However, the milling machine i'm using has 1/38th a milimeter accuracy, and while i have experience with a manual milling machine, well i'm not anywhere close to proficient. I also have alot of time at home with ProDesktop and QuickCAM, but only a few hours at my dad's school with the CNC. From past experience, a complex design usually takes me 4 or 5 hours to mill out, at minimum. However, the biggest advantage of using the CNC is my ability to test out designs in software. I have complex wind tunnel software, ($15000 0.o) and while its probably not the same as fluid resistance software, both air and water are fluids =p. I havn't fully tried modifying the viscosity of the air, but i'm sure i can tinker to get it to somewhat mimick the properties of water.

Re: CNC milling machine at my disposal
 

Perhaps you can find out more about this. I don't believe anyone here is really interested in any for-profit endeavors - I think most are just looking for cheaper ways to learn. It might be interesting to be able to submit designs, have you machine them and simply have them sent to hobbyist testers (of which there are a few here) who could then publish and relay their data back to the block designer. The block could then head back to the school for exhibition purposes. This may be something the school would be comfortable with.

The end result would be that the block designer would have a better understanding of the real performance data of his conceptual blocks and could then be more comfortable about spending the money on professional machining.

In any case, be sure to seek out the advice of the members here on any designs that you may have in mind. Regardless of how little a grasp you have on the concept of waterblock design, everyone here is excessively helpful, often times providing information and equations that most people simply could never understand (myself included). Also check out the block reviews here and Lee's reviews at SystemCooling, as he often provides some valuable dimensional data on the parts he reviews.

Re: CNC milling machine at my disposal
 

thanks for the advice phide. I wasn't planning to use it for any profit purposes lol. But the school would probably be dismayed that blocks are being created for other than educational purposes at the school. One thing i COULD do is convince my father to start a new "project" at the school, design a cpu waterblock XD. I doubt the school would want to have any waterblocks other than ones created by students. However, if you guys can design and test them, and make sure you send them back, my dad can use them as 'teaching aids'. I can't pay shipping, especially for the heavier 1lb blocks. I'm a high school student about to goto university 0.o. So basically, i can't machine any blocks for you, but i can test out any design you want. within constraints of the CNC of course. Guys one thing i'd like to know though, is how well my virtual windtunnel program predicts turbulence in real life, and also, is it better to have high velocity flow or to have high turbulence flow? That way, i can put out designs and test them without having to really set them up. I'll post all data recieved and also any correlations with my wind tunnel software. Hopefully we can find a ideal balance between turbulence and flow rate. I'd also have to design it in aluminum, copper is much too expensive for testing purposes solely. *edit* i might end up using copper anyways. I think my heatcore for my testing system is copper. Galvanic corrosion doesn't seem appealing.

Re: CNC milling machine at my disposal
 

I'm thinking of copying *insert shocked expressions here >.>'* the cascade design. I have no formal measurements, so i'm planning to test it out myself, see what happens. Apparently, flow can be sacrificed for turbulence and reducing stagnant water zones. I don't know how that's mathmatically sound, but it seems to works in theory, so i guess I go with it 0.o

Re: CNC milling machine at my disposal
 

Insert the standard
What is the turbulence of which you speak? :P

Think a better design to copy and make multipul models of is the storm as there are only a small number of design variables and it is studied to death already. Also it wouldn't be copying as its been done before by 1000's of people in other industries.

Re: CNC milling machine at my disposal
 

Its true theres only a few variables in a storm design, but i think it would be a good one to test out first. I recently saw the storm G7 prototype, and i'm planning on trying that one out. Its got over 100 jets 0.o I just found an old system (PIII 550) and i'm planning to use it as my test setup. This thing is SLOWWW -_-' But its free =p. Also using it for other administrative test purposes. Good thing windows 2000 pro doesn't have activation. My one license is on like 8 of my systems 0.o I'll submit pictures of the completed setup when i finish it. Still getting heatcore and designing waterblock atm. it'll be awhile.

Re: CNC milling machine at my disposal
 

G7's a little tricky - requires custom tooling, which I imagine are probably solid carbide parts. I'm not terribly sure, but I believe the G7 requires two custom tools - one for creating the cups (with the obligatory "microturbulator") and another for the jets. The ID of the jets in the G7 are probably on the order of 0.3-0.4mm. If you also intend to do it right, you're going to need some silver, which is not an affordable material by any stretch.

I just pulled and cleaned out my G5 (took a few hours thanks to some interesting solid deposits in the cups), so I can take some measurements of that if you're interested. I don't have the tools to get anywhere near exact, but I can approximate everything fairly well. The G5 would be reasonable to try, I think, but still requires custom tooling.

I'd probably just try out a bunch of wacky designs if I were you, though.

Re: CNC milling machine at my disposal
 

That's what i'm confused about... What the heck is a microtubulator? The jets should be ok... i think. I'm not sure about the accuracy of the newer machines, but the old CNC has a 1/72mm precision, so it should be ok. Tools arn't really a problem, its a school lol, tech department's overflowing with them. Custom ones, well beats me. The G5 does sound good... but i was thinking of making a G5-G7 hybrid. Lol and i'm not touching silver, i'm not even planning to really overclock this system, I just want to try out water cooling =p. Wacky designs are going to be tried as well... just probably not machined. I'm going to use the wind tunnel program. But in the mean time... i have a bio test and a law ISU. heh i'll get around to drawing them when i can. If you could take some really high res photos of the G5, really close up that'd be greatly appreciated. I'm interested in the bottom of the cups, and the entrance of the jets especially. He never explained if round bottoms or flat bottoms were better. I want to learn from the techniques Cathar used, and improve on them if possible. BTW, for those looking into real cooling material, i think copper tellurium has a better heat transfer than silver. Silver really isn't that much better than copper. i think its less than 3%

Re: CNC milling machine at my disposal
 

In the Storm blocks, the tiny pin in the bottom of each cup is responsible for increasing turbulance and further disruption of the boundary layer. A turbulator is really an aircraft term, but the function in the Storm is pretty much the same. Cathar and Swiftech call them "jet disrupters", but I kind of like the term microturbulator a little better. Use whatever term you wish.

Your school will have no tool that can properly form the cups. If you should dig around and find a ~2mm flat-bottomed cylindrical carbide bit with a tiny <.25mm indentation, I'd be pretty surprised. It's possible you may be able to source the bit for the jets, but I'm not certain of it.

It would be a gargantuan task to try and beat out the effectiveness of the Storm design across a broad range of flowrates as Cathar as done. It's one thing to design a practical, functional jet impingement block by referencing and calculation. It's another to make the block perform well at a variety of flowrates. That demands a fair amount of knowledge of a number of aspects of watercooling - not just block design. There are some things one can do to improve performance at one end of the spectrum or the other, but I doubt Cathar will tell you what they are even though you aren't seeking profits.

He probably came to the conclusion that flat-bottomed cups are ideal by referencing existing white papers and through his own simulator. As far as I can tell, the jet inlets are just chamfered once. Nothing terribly fancy.

I'm unfamiliar with copper tellurium, but its thermal conductivity is 205 (BTU/(hr degreesF ft^2/ft) while silver is 235 BTU and copper at 223. So its thermal conductivity, compared to pure copper, is poor. This is typical with all alloys, even silver/copper alloys. Silver also has the advantage of being somewhat antimicrobial in nature. There are a few scattered companies that produce fittings and tubing that release a slow stream of silver ions and effectively keeps bacteria from forming on their surfaces.

Re: CNC milling machine at my disposal
 

Would recommend a different test setup. A better method might be to build something which makes the block HOT(300 to 400W) and do some scaling. That way cheap testing equipment can be used.

The differences in conductivity between copper tellurium and proper copper I think are small. They are going to effect things sure but the effort probably isn’t worth it.

As this is an educational project jets in cups have another bonus point. Jet impingement is extensively studied at university. Its used as a test for CFD programs and has heavy uses industry. If you did a test on proper test samples varying different parameters then the test would be of a publishable level in a scientific journal.

The walls at the base of the cup should be at right angles. When the jet hits the base of the wall it makes a thin film of water spraying out at high velocity; a wall jet. When this hits the cup wall it creates a jet impingement effect of a fast moving stream hitting a wall. In short you want the wall as square as possible and there should be a tool around to do this anyway.

My advice would be to ditch the turbulator from your design. The performance improvements are likely to be small. In a strict sense it doesn’t really turbulate the jet. I believe it introduces non-steady un-axisymmetric features in the jet. These cause the jet to pulse a little, which increases heat transfer. They also have the effect of reducing jet maximum speed is bad and while I’m sure cathar knows what he is doing I remain sceptical about the performance trade off.

Re: CNC milling machine at my disposal
 

these are both very interesting points phide and bobo, i believe i'll end up ditching the turbulator. I have endmills that are finer than 2mm, and drill bits even finer. As for publishing, i don't know how much people would believe me. I'm not even out of high school yet lol, 2 months till graduation.

I can test out performance at different flow rates easily. The wind tunnel software allows me to set the velocity of wind, which hopefully can mimic water. I'd probably go with a design for a flow rate that i'd be using, its its really only for me. I'd still report all my findings about different flow rate performance though.

I don't know how effective the turbulators would be... although i'd be very happy to test them out in my virtual wind tunnel.

All i have to do now is find time to draw all this =p.

BTW, what kind of simulator is cathar using?

And how can i design something to emit 300-400Ws of heat? 0.o i have temperature sensors already.

Re: CNC milling machine at my disposal
 

Anyways, maybe i should make this a bit more clear. Send me your designs. I will machine for free. Its a 4 axis CNC with 1/72 mm precision. You just can't have the block. I'll tell you how well it does though, I'd also be happy to help you make any improvements and retest it.

Re: CNC milling machine at my disposal
 

well i finally sat down and drew out this thing. Baseplate atleast. It looks alot like the G7 for now... but the jets are going to be very different. I'm looking at a suction system directly above the jets, so instead of turbulence outside the cups and restricted flow by the jets towards the output, flow might go up a bit. I don't know yet. Anyways. this took forever to draw... and i might end up redrawing it -_-' If anyone knows how to use proDesktop, please tell me. I'm having quite a hard time with its more advanced features =/. anyways. its not dead center, its off by .2 a mm or so... mostly because i couldn't find a way to link all the jets together as a single object. I'll make the jets, then run it through the wind tunnel and see what results. Does anyone have a suggestion for test system construction? I'd be even more grateful if they drew it out... >.>' turns out that prodesktop is surprisingly difficult for repetitive tasks... duplicate or pattern commands don't allow copies to be later linked up. Or atleast as far as i know. Anyways. heres the baseplate. http://img511.imageshack.us/img511/5875/baserev27kx.jpg The block is also going to be circular, not ovalish like the G7.

Re: CNC milling machine at my disposal
 

What's the diameter of each cup? I'd speculate that the diameter of the G7's cups are on the order of ~1.5mm. The tight density should allow for an even thinner baseplate, so I imagine the cups are somewhat deeper than 3mm, with the baseplate being 4mm at its thickest point.

You'll probably have to learn ProDesktop through trial and error. If it's anything like SolidWorks or AutoCAD, there should be some simple tutorials that may be able to assist you. I've gained quite a considerable grasp of SolidWorks just by fiddling around.

For a diagram of a very robust and capable test system, head over to SystemCooling and take a look at Lee's/Robotech's configuration. Do keep in mind that building a proper testbench is pretty cost prohibitive, which is why I suggested contacting one of the testers here or at the Waterblock Testing Alliance (wbta.us, I believe). There aren't many left willing to do testing, but BigBen might be interested in doing some work for you, though I have no idea if he'd be interesting in testing a Storm clone.

I'm not sure I understand what you mean by suction over the cups - care to elaborate? There are already areas of suction at various important parts of the Storm.

Re: CNC milling machine at my disposal
 

Talked to my physics teacher at school, hes got a PhD 0.o Anyways, basically said that the diameter of the tubing and jets is directly proportional to the pressure of the jets. Should known that beforehand >.>' Anyways, so i'm now redesigning it for the smallest jet possible. The cup diameter will be the diameter of the jet plus the area of the water, converted to diameter. calc is a bit rusty, so i havn't done the math yet. Anyways, basically the cups should be very small, just like the jets. I'm going to see what kind of material gives me the option of having extremely fine details without breaking or stuff like that. The depth of the cups and the length of the jets have yet to be tested. I'm also considering a tapered jet design down the tube. Anyways, the entrance of the jets will be a parabolic arch, supposedly gives the greatest velocity with the least turbulence. No point in losing flow and speed before cooling the actual block 0.o.

I plan to place suction in 3 strips surrounding the jets, that way the water flows out vertically instead of being moved off to one side and sucked like the cascade or storms.

Anyways i hope to have performance close to cathar's design, without using his proprietary stuff. I asked him for information about the jets, and hopefully he'll point me to some jet impingement materials.

On a side note, did you know copper and vinegar is toxic? I found out after my chem teacher heard what i did yesterday >.>' Cleaned out a radiator and at the same time, washed my hands with the stuff for 3 hours. MSDS doesn't say any serious side effects, so i'm lucky about that... but heavy metal poisoning doesn't sound fun either way. Anyways, now i have 6L of copper acetate that i can't pour down the drain... have to get it treated by a professional 0.o. I think the city pays for it =p. Oh well... just a heads up for the rest of you guys. Don't try cleaning your copper rads or waterblocks with anything other than steel wool. You'll most likely end up creating copper ions. And unless you have very good lab skills, you're probably going to end up touching some too. Plus you have to find a way to treat it yourself, or get someone else to.

And i shall check out the testers... i might end up setting up my own block anyways... I live in toronto.. might cost more to ship these blocks than it would to build my own bench >.>'

Re: CNC milling machine at my disposal
 

Does your school have a toolroom that can make custom endmills?
The problem you might run into is that if the spacing between the jets is going to be as close as the drawing shows, then finding tiny endmills (less than 2mm dia.) that can cut deeper than a few mm's will be difficult.
The G5 design will allow larger cups and jets and make the tooling part a lot easier.
Delrin is the material of choice for the jet plate.

Re: CNC milling machine at my disposal
 

Hmm... i don't think i can make custom endmills... atleast not as small as you mention. the smallest end mill i have is 1/16"... can probably aquire smaller ones... Basically i only need an end mill that can drill like 4mm... the spacing is .27mm in the drawing. However, i'm going to change up the drawing anyways, but yeah i'll consider increaing the spacing between cups.

Delrin? hmm alrite... is it easy to machine?

Re: CNC milling machine at my disposal
 

There are different varieties of acetal resins, so you need to be a little mindful when selecting one for your purposes. You can either get an acetal homopolymer (sold under the Du Point brand Delrin) or acetal copolymer (majority of which is made by BASF). There are also different grades, most of which you'll find are designed for injection molding. You obviously want machining grade acetal, which is ideally going to be lubricated.

Machining-grade Delrin homopolymer should be fine for your needs, but I'd probably go with a copolymer. It seems slightly more ideal for the demands of a water block, and you'll probably find that most copolymers are cheaper than Delrin.

Cathar may have actually used a copolymer and just called it Delrin so people would know what he was talking about.