Planning dive off the deep end:xtreme water cooled build. solidworks question

[disturbed13]

Originally Posted by siliconfanatic

yes i am planning on moving out, but i DON'T want to tear out the wall to put in piping that just increases the liabilities of a custom watercooled rig! and i have stated, several times in fact, that this rig will be WATERCOOLED. and yes i will have a rad to clean, in fact it'll be a biga** rad. and no, its ENTIRELY IRRELEVANT to this thread. does the thread title say i what make one big ol' loop that will leak like all god knows what? nope! if i had said i want to watercool several dozen rigs at once, your post might be appreciated. but the only other rigs are laptops or barebones. so don't even think about telling me that "what i mean" is ___, when i CLEARLY STATED what i mean SEVERAL DOZEN times. what i mean is even in the fricken title of the thread! if you want to talk about this idea(which is a good one, but not for me) make YOUR own THREAD about it! it will get positive ideas and will be relevent to itself then. next time, think carefully, and !!!!!!!!!!!!

matter of fact, it looks like your trying to commandeer this thread and use it as your own, you SHOULD have stopped the whole house loop when i said the long equivalent of no thanks. but no, you continued it like i had agreed to it, like you were some omniscient jerk who knows what i want when i clearly said what i want. which is a SELF contained water system(theres that word again!) and i simple way of doing this is to have a reservoir, and an extra pump alongside the main, with a simple logic gate/etc that turns it one should the main go AWOL. so i must ask you nicely to take it elsewhere.

hour long rant is over, case closed. anyone with relevent ideas may now post away.

??????
and now im confused
i hope you didnt think i was talking to you in my first post
i was replying to what someone else said, and was giving a general heads up
then you seemed to take it personally and now $h!t just got weird
but what ever
knock yourself out

[siliconfanatic]

yes, but then in your third post, you sounded as tho you were talking about my thread. ie, i said irrelevent, you said relevant, several words made it seem as tho you were talking about my thread. jus look a little more carefully next time you click the thread link so you dont post on the wrong one, okay? and in the first post, you could have posted in the users profile rather than rant on my thread and seem like you were suggesting something far from waht i was talking about. wouldve avoided any confusion. or you could have only posted about the air cooler only, and not about the whole-house setup.

[disturbed13]

and air cooler can be anything that uses air to cool something
so an Air Conditioner connected to your PC would be a air cooling system, with a super charger
a liquid loop is an air system, it uses the air to cool a liquid
- which can absorb the sudden thermal shock of going from idle to max cpu usage when starting a game or other CPU/GPU intensive tasks
so the fact that anyone says 'im going to air cool it' is like saying the sky is blue

[siliconfanatic]

the point is you get the point. you know what we mean, so you dont need to get specific like that. but i aint air cooling it technicly speaking, im air cooling water thats cooling my elctronics. there. happy? so if someone says air cooling, there gonna use this thing with big fins that thransfer heat to the air thats called an air cooler. thee vocab is self defined, not to be thought of as two separate words even if they are. hence water cooler is a water loop with a radiator. if you want a debate about it theres a forum for it.

[disturbed13]

okay
so you want to make a pump
then get a 120mm fan for the motor
use POM to make the impeller and the shroud
and lets see it
it will be a pump
how well will it work, yet to be defined

oh btw, good luck getting everything 'carved' using a dremel tool

[siliconfanatic]

correction: cnc dremel. i haven't got my dremel skills perfect just yet(though i know someone with a steady enough hand who could do it, but he is still in college.). and the areas the dremel head cant reach will be two piece if necessary . the inlet will simply have a fitting glued to it and sealed. and i wont need to buy a fan for the motor as those don't have large enough power reserves for my purposes. plus its much easier to get a fair-sized RC plane motor than to take one out of a fan. but thanks for the suggestion and the right plastic to use. just gotta light 'er up + let'r rip!

[disturbed13]

so are you going to use the less effective brush motor
or a brushless?
an how are you going to deliver power to it?
are you going to have a whole different 12v loop from a wall wart?
or are you going to try and suck enough power from the PC PSU?
and how big of a PSU will you use/do you have

[siliconfanatic]

dont call them LESS effective. this thing ishuge. this isnt from those tiny little park flyers that have somehow come to symbolize rcs. but i also happen to have rc motors, brushed and brushless, of many sizes. and as for the psu, that is unofficial. ill total up the power requirements, THEN make a decision. i would like to use brushed, but i dont need that much torque. ill be going for brusless. this will be moving at really high speeds, for the main pump at least.
plus, if my gradfather isnt getting rusty about that knowledge, brushless doesnt quite need as much power. i havent yet heard what they sound like without a propeller, but dang these things can wail like a banshee in the air. and they can SCOOT! ive had many experiences so i would know. been taken flying since a iwas a month old. anyways, which would you suggest? its looking like ima need a roswell psu for the peltiers alone. of course the peltiers will wait a while yet. once my gradfather gets home ill check just how much power this thing gulps.
edit: i wont be able to post very much for a while, just got horrid grades for first quarter. i will try and update wen i can

[disturbed13]

i would go with all of the torque that you can get
especially for a pump
and as for PSU ionno
there needs to be specs for the PC

[siliconfanatic]

Thats just it; i cant specify what the power requirements for a pc will be in 3 years, for one with intels tick-tock cycle, and with the power revolution going on as of late. For all that i know we could be seeing graphene cpus eating the same amount of power, belching out anywhere from one third to ten times the heat and nailing windows with ten hundred times the number-crunching power. Or we could still be seeing silicon cpus eating three times the power, ten times the heat and six times the power. So the powr requirements will wait for last. But i would like to use no more than 3 psus. Merely for wanting to plug the darn thing into a wall socket without ripping the wall out to beef up the power lines and sockets.

[disturbed13]

the graphite nano-tube cpus could be possible
but not for the consumer marker
maybe for the business and data servers
cpu's will continue on the same trend that they are on now, always getting smaller and faster
gpu's have been on a bit of an opposite trend
they keep demanding more power each year
so its a double sided coin

[SuicideNeil]

Originally Posted by siliconfanatic

Merely for wanting to plug the darn thing into a wall socket without ripping the wall out to beef up the power lines and sockets.

A standard water cooling loop would be so much simpler, with your party piece 'turbine'* as a reservoir ; no need for over complicated wiring or extra power supplies; running r/c electronics in a pc is one thing, but building your own pump unit ( not just the inlet/outlet cover, but motor and impeller too ) is something I've never seen anyone do & probably for good reason...

*http://forums.bit-tech.net/showthrea...=145276&page=8 & http://forums.bit-tech.net/showthrea...=76374&page=18

[KayinBlack]

I've seen it done once, and it was in a place where normal hardware was impossible to get. Its performance wasn't stellar, either.

It takes a lot more than an idea to make something like that. CAD software? Check. Flow mapping software (like CATIA) as well. Now let's be ready to make a bunch of prototypes to find out what thousands of hours of CPU simulations couldn't tell us. How do I know this? Seven patents thus far, a few more pending. None are even profitable. Cost of entry into the market is too high.

There is nothing wrong with having ambition, but I think you're running into areas that people have doctorates in for a reason. I have a few degrees, a ton of certifications, a couple of licenses and many years of experience, and every other day I see something that makes me scratch my head and wonder how it works. Many of the ideas thrown around in this thread have been tried and discarded before by very large companies with very big R&D budgets. There is a reason we're all using off-the-shelf pumps, for example. What if you built your pump, but it decided to throw off enough EMI to nuke board and CPU or something? Or your impeller decides to disintegrate (seen it happen one with an off the shelf pump-boy was I glad for that warranty then) or a potted coil goes kaput? With each part that is assembled from off-the-shelf systems, we can generally eliminate an issue by controlling it as best possible. Knight's Rest has a backbone of familiar parts (standard PC innards, Laing DDC pump, Swiftech blocks, Black Ice radiator) surrounded by enough bespoke electronics to run it to that I had to draw myself diagrams. Sure, there's some really wild stuff, like the reactive armor panels. But if I give myself freedom there, I need to trade another area (like IB and Z77 instead of something more fun like Itanium2 or something) in order to keep the project manageable. It's one thing to reinvent the wheel. It's another to try to recreate the car with a set of hand tools and a photograph of one.

Keep thinking big, but if you change all the variables, it's not a proper experiment. Dial it back a bit and stick your nose in some books on thermodynamics and fluid physics. What you learn will help you immeasurably when you get ready to make something new. I still spend part of every day studying something just to keep up with how fast the industry moves. Buckle down, study, and one day we might be buying your parts-but one thing at a time. That CATIA rig will cost you, too-mine's a four CPU 16 core Opteron setup, though I have a pair of hexacores here to migrate to when I get the opportunity.

[siliconfanatic]

the party piece turbine is already taken, being used for radiators. part of the thing in this was a turbine takes in cool air, compresses it then super heats it and spits it out. rad does a similair thing. several o fthem in tandem... the loop wouldn be too complicated as its not one single loop, rather several loops... however i am open to any ideas as to how ram enough air into it to feed all those hungry fans. i will post pics of exact layout later. any ideas as to what else i can do to integrate the res are welcome.
and like i said, could have this turning real slow as the display pump, while the one thats actually pumping liquid would be hidden. the custom one would be on a "fake" loop branching off the real one. although i will admit wiring that many fans is difficult, but...

[jay2]

umm... i think i lost the topic...
some pics please..

[siliconfanatic]

will get those up asap

[siliconfanatic]

Originally Posted by KayinBlack

I've seen it done once, and it was in a place where normal hardware was impossible to get. Its performance wasn't stellar, either.

It takes a lot more than an idea to make something like that. CAD software? Check. Flow mapping software (like CATIA) as well. Now let's be ready to make a bunch of prototypes to find out what thousands of hours of CPU simulations couldn't tell us. How do I know this? Seven patents thus far, a few more pending. None are even profitable. Cost of entry into the market is too high.

There is nothing wrong with having ambition, but I think you're running into areas that people have doctorates in for a reason. I have a few degrees, a ton of certifications, a couple of licenses and many years of experience, and every other day I see something that makes me scratch my head and wonder how it works. Many of the ideas thrown around in this thread have been tried and discarded before by very large companies with very big R&D budgets. There is a reason we're all using off-the-shelf pumps, for example. What if you built your pump, but it decided to throw off enough EMI to nuke board and CPU or something? Or your impeller decides to disintegrate (seen it happen one with an off the shelf pump-boy was I glad for that warranty then) or a potted coil goes kaput? With each part that is assembled from off-the-shelf systems, we can generally eliminate an issue by controlling it as best possible. Knight's Rest has a backbone of familiar parts (standard PC innards, Laing DDC pump, Swiftech blocks, Black Ice radiator) surrounded by enough bespoke electronics to run it to that I had to draw myself diagrams. Sure, there's some really wild stuff, like the reactive armor panels. But if I give myself freedom there, I need to trade another area (like IB and Z77 instead of something more fun like Itanium2 or something) in order to keep the project manageable. It's one thing to reinvent the wheel. It's another to try to recreate the car with a set of hand tools and a photograph of one.

Keep thinking big, but if you change all the variables, it's not a proper experiment. Dial it back a bit and stick your nose in some books on thermodynamics and fluid physics. What you learn will help you immeasurably when you get ready to make something new. I still spend part of every day studying something just to keep up with how fast the industry moves. Buckle down, study, and one day we might be buying your parts-but one thing at a time. That CATIA rig will cost you, too-mine's a four CPU 16 core Opteron setup, though I have a pair of hexacores here to migrate to when I get the opportunity.

thanks for your truly well based wisdom. but ive already nuked my account. ill try and find some books to help me comprehend what i need to read. and i am going for a goal very similair to this. masters in fluid dynamics, and @least a minor in thermal. i have heeded warnings, and will use this as an aesthetic thing. actual pumping will be done with at least a mid-range pump. which will be in the real loop. sorry for yet another repeat

gotta go; cornbread and coffee dont exactly mix very well(current engineering program is simultaneously fun while nontrivialy whooping your mental a#$

[siliconfanatic]

and speaking of the CATIA/solidworks rig; which i have finally felt the power of not just sw, but autodesk INventor(introduction to engineering and design class, the rig/program didnt cost me; for me its free aside from the fa,ily favors now owed.
anyways, ill skip to the point: now that i know ill need to do an assembley, does anybody know of a decent(resource rich) sw component/part site? for things like radiators, model fans, mobo's etc? i do most of it my self but i aint quite confident enough with my math and such of it. still gotta get that darned tubing tutorial done but it keeps glitching small things in the base model(reduce router ithink). maybe in afew days ovr thnksgivin's break...

[SuicideNeil]

Google sketch up?

[siliconfanatic]

not sketchup, unless solidworks(which i now love so much that i find sketchup repulsive) can recognize the model as a solid part.... which i am doubtful about. alls i need is fittings, radiators, fans, waterblocks and bigger fans( all sizes, from 40 mm all the way up to 240 mm. ) most other parts i can model myself. just need to finish that darned routing tutorial. in fact, what i learned from skp actually inhibited my learning sw. i was so used to coping with rather innacurate 3d sketching that 2d plane sketching was a bit of a leaning curve for me.
edit... there are plenty of sites for downloading things like these. my primary concern is that they are trustworthy