WHAT the...?

i think the fish people either use evaporative cooling which works by free flowing water through air, or some form of thermo electric, which is what im proposing. now there is a lot of physics involved in tec cooling, and i am by no means any where near to being an expert. so lets see if it works as well as I want it to. but i think with 13 on there I could certainly ice up the water.
 
now i really don't want that though, because as soon as the water temp goes below the dew point, it will form condensation on the outside of tubing or plates or wherever water is flowing that is below the dew point. my hypno wheel shouldn't have any problem cooling 2 gpu 2 cpu and the chipset blocks to a few degrees below ambient temps in the case. now to keep the ambient in the case the same as out of the case we need to take the water from the hot side and take it to radiators outside the case. and then control the power to the tecs based on sensors for water, air and ambient temps.
 
so to answer your question, this thing should be enough ridiculous overkill by itself. but who am i to stop you, you tell me what you want to do, and I can engineer something to fit.
 
now as far as milling it. I use nate for milling the copper parts at this time. i can mill any acrylic on my equipment, but I am not set up for copper at the moment. I might get my bridgeport working some day, but i just haven't had the time. so once i get all the bugs out of this design, i will mill it completely from acrylic, then do some flow tests to see how the water moves through the system, then I will have nate mill the copper if he wants to from code that I can give him. now this stuff is pretty big, so he will need to do it on his new setup when he gets it ready.
 
I will be happy to design just about anything imaginable. now keep in mind, I think most people don't go crazy like this because if the sensors mess up or if something breaks, you have the potential of soaking thousands of dollars of electronics. I def recommend you spend a sizable amount of time researching and understanding the condensation issues before you decide to go this route. now the way I am doing this will allow normally 3 rad cooling when the tecs are off, so i can run without worrying about it till i 'prove' reliability. 
 
by the way, do a search on bong cooling. that is essentially what the fish guys do.

from what i have seen so far, the tecs can be controlled via pwm which is really just pulsed on and off power to maintain the corect block temps. easily achieved with some of the products available. So I am confidant i can maintain a constant temp curve in relation to ambient. the block and water should theoretically be very close to the same temp once equalized. and yes a 'drip' tray could be used and would be relativly unnoticable when intergrated. it will be fun to play with. I just dont want to conformat and waterproof the chip sockets and such, but if the failsafes break for some reason, this setup I am sure is large enough delta to ice the lines. using all 13 at 100% would consume 2860 watts and move 2000 watts of heat. like i said though, vastly inefitient. but when run at 10-20% they actually are very efitient, taking 300-520 watts or so and moving around 500-1000 watts of heat.
 
yah, lol i thought about hanging the mb upside down as well, decided against it because of visual reasons, and the fact that heat rises. I would love to see a build that way though.

Exact control. but basically it works via a t -controler or a custom control which I will make that is similar to the t-controller. the problem would be if the blocks get colder than the water then need to do a 2 stage to initialy cool to make the blocks at dew point then after system equalizes gradually cool down more till it water reaches dew point. thats where i need to experiment with how to make the custom control multi staged and utilizing algorithm from multiple temp sensors at diferent points in the loop.
 
http://www.t-balancer.com/english/bng.htm
 
edit
if i build my own it will be sorta like the 3rd scematic here
http://www.bit-tech.net/modding/2001/12/03/pwm_fan_controller/1
  
 

np, it will take me a wile to get it all working but we got time, right? btw, seeing as you have come up with the product name for me...hypnosis...you qualify for a 1% discount

Going on Active Duty for the next 3 weeks.  I will finish your CPU blocks as SOON as i get back. 
 
I got yours and Forgottons SR2 block sitting right next to each other (and the scrap piece of copper).  Just need to finish cutting the CPU blocks. 
 
I will be on-line daily and will still answer PM's but won't be on here 24-7 like I usually am. 

nateman_doo

Going on Active Duty for the next 3 weeks.  I will finish your CPU blocks as SOON as i get back. 

I got yours and Forgottons SR2 block sitting right next to each other (and the scrap piece of copper).  Just need to finish cutting the CPU blocks. 

I will be on-line daily and will still answer PM's but won't be on here 24-7 like I usually am. 

Humm. So you've actually made the air/water block? because i hadn't paid you for that piece yet but will be quite willing to ! :)
 

nateman_doo

I am making CPU blocks for gratuitous.

ahh yes kk just makin sure lol well hope your service goes easy on ya and let me know when yer ready for 2 more orders ;)

ajohnes2

70lbs!!!!  I can only imagine the madness that will ensue if it is made.

You can see the madness we are planning one page back ;)
or by clicking this.
http://forums.evga.com/fb.ashx?m=1138590

Oh I do not doubt it will be made, just the idea is massive for a computer.  I saw the idea, just never thought to comment on it because I had no idea how to put thoughts to forum.

nateman_doo

I was just re-reading this thread about condensation.  Worry more about the motherboard, and any block touching it.  The tubes are easy, any type of insulation will work.  Weatherproofing a motherboard, takes some serious work ;)

If you do that, might as well scrap all this, and just build a chiller.  That will give you the most bang for the buck.

I've already begun buying silver for the project 8 bars so far, so I'm already vesting in it and planing on melting them into a workable shape. for the toobing i've come up with a simpler solution I'm planning on just using a thicker grade of tubing and perhaps using it over the normal tubing types and I'm going to make sure that I'll be able to use the inverse (or upside-down) motherboard install. Though i am not opposed to a chiller design and considering taking apart the mini refrigerator i have here at the house and incorporating into the design.

nateman_doo

silver? smelting? whaaa?

skip the fridge, use a window air conditioner. 

I would, however i already have the fridge in my possession it's been unused for about a year and well better this then unused space. lol, Do  you feel you have underestimated my willingness to use unconventional and or new ideas in the making of a unique and perhaps overwhelmingly unique system?

Gratuitous

I have been considering phase change as well. I think what i really want is to keep the tecs regulated via pwm to keep the cold plate just above the dew point. this could be done with temp sensors and humidity sensors and an algorithm in a custom pwm controller. that way I can maintain a lower than ambient, but above dew point cold plate, without any manual tweeking. this would be the way to control phase change as well, but im not sure i want to go there.

i don't want to have to insulate anything. as much as going way lower than ambient is awsome, I just don't think i need that. if it is 90F in the room ambient. which it is alot because i dont run the ac when no one is home, I want the liquid to the cooling blocks at 80F or 70F depending on the delta between ambient and dew point for this specific minute of the day, controlled by the controller. doing this will allow the computer to be on at 100% load 24/7 without me even having to glance at it. I will have to get the heat from the hot side away from the computer thogh so as to not increase dew point  close to the computer and defeat the purpose.

so yes, it might be simpler to run a window ac as a chiller and use the same type controller to do the same thing. the diferance is the size of the solution, 12 tecs can fit nicely in the case and the heat routed to standard radiators in a box away from the computer, phase change requires a much bigger external component. I havnt completely decided on that yet. the heat exchanger part that I am designing now can potentially do either solution though. (if it works)

I sure hope it does ;) Even should it encounter problems the silver I've begun collecting will stand as a good investment that at some point i would sell and hopefully receive a profit ;)

Gratuitous

your idea is also very interesting, and I have been thinking about it a lot forgotten. you know how watches have jewels in the inside and you never see them but they are there. well this is sort of the same. you have some value inside your case and it just happens to serve a purpose as well. been looking at the investment and sand casting techniques and if we were to use a investment style then the positive part is sacrificial. I might be able to mill the positive from something that will melt away when you cast the silver. I need to do a little more research on that style to see if i can mill that material and if a big block of it will actually melt properly.

I'm certain we can smelting is a very simple process.  In addition this allows me to hide expencive silver in my computer case so if someone was to break in and seek monetary gain well a safe would be obvious, and the system I'm building will be very heavy and therefore a bit bulky for someone seeking to rob the house and may be left behind. Especialy if they are totaly unable to understand how to care for or even transport of a water-cooled system. Just may dissuade someone from taking it. Or even realizing the true value built into it.
Then even beyond that if there is an issue with my system at some point I will have within it a means to re-purchase and replace damaged components with the valuable metal.

Gratuitous

so, what, from your understanding, do you think the 1st original should be milled from? I can easily mill it from plastic. if you were thinking of a sand casting tequnique that might work well. having dever done smelting or casting, i am only speculating at the outcome. I like the idea though as all you would need to do is tap the holes and lap the bottom to get the finished piece. that way there are not a million tiny shreddings to collect and remelt later.

if you are referring to making the cast out of plastic. I do not believe that would be a good idea as the melting of the silver would require it to be at a high temp. that would in all likelihood end up melting the cast. If you are referring to the shape of the component to create the cast from. I do not see any problems with that being a good first step towards creating the ultimate mold for us to melt the silver into. I do not have experience with sand blasting or other techniques used in milling so i will need to rely on your and doo's experience in that area. including the tapping i do not have the equipment for that. However smelting can be achieved with very simple means, A camp fire and a stone bowl can be enough to complete the process.