• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Air Conditioned Liquid Cooled Computer

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
If a TEC cold side cools 75 watts of heat to xx temp, the hot side of the TEC would need some cooling device to remove the 75 watts of heat, and another 75 watts of heat generated by the TEC. They don't work for high power cooling.

Thank you for saying that in a much more concise and focused manner.:thup:
 
I got ya, I was thinking a peltier (1) on each end of the rad where the tanks are, cold side of course against the rad, to help cool the rad along with fans mounted in a normal position. the problems with condensation then become easier to deal with as really the rad can be moved away from the things that make pops and crackles, and since your using them to cool the liquid in the rad then size and wattage is not as sensitive to the size and wattage output of the cpu. This might not result in a sub zero cooling solution, but seemed like it would result in a sub-ambient solution for cooling. My thinking was Stock heatsink & fan--->peltier---->rad with push/pull fans--->peltier----->stock heatsink & fan. As for power at that point you could pretty much devote another psu to that area and really about contain the whole thing is a add on case of its own, with just the hoses run into the cpu case.
 
IIRC that is a "cold box". I can't remember where I've seen or read about one.

Not trying to bust your bubble or rain on your parade. I really do like the idea. I don't know if it would be worth the time, energy, effort and expense. Jumping ahead a little bit - assuming you could create one and then discovered that it added 400Mhz to the overclock, would it be worth it?

I don't know if you have read about evaporative cooling towers, aka "bong cooling". They do work but they have to be big to work. Upside is that they aren't very expensive.
 
Actually since I work in a shop most of the parts are free, thats kinda guiding my thinking, another case with a power supply I think we just scrapped out 54 the other day. Peltiers are cheap, easily found for 20-40 depending on the rating, fans are free, so are stock heat sinks, LOL thats why I was asking because I could throw this together for the price of the peltiers alone. And I have to admit the sub-ambient area is all new to me, but I have no desire to set a worlds record with a cooling system, simply knock 15-20c off my temps. And right now I am proud to say I have setup my system to run from the stock 3.5ghz to a respectable 4.63ghz but I am learning that ambient air, even with 9 fans on a liquid system is always an issue, and that volts equals really high temps really fast. Also learning the problems that go with the next level of cooling, condensation huge power requirements, etc., and to me those issues along with the fact that winters in South Dakota can teach you about thermal transfer, windchill, and extreme cold leads me to believe there has to be a way to make the rad colder and that will benefit the system. so would I do it for another 400mhz? Oh yea I would to have a box running 5.1ghz 24/7 with temps in the upper 30's to low 40's. Wouldn't you?
Mostly I am thinking out loud and with the advice from you guys which I value so much because here at overclockers.com is where the best come, and if I can engineer this out then I may not be the best but at least feel I contribute.
 
Deleted previous post. I think from what I remember reading about them is that the amount of power needed to run a system with TEC makes it unconventional to do.
 
yep a stock heat sink and fan, since its not in the case the hot air can be handled better and by using better then stock fans you can still dissipate heat faster to keep the cold side very cold. for the price of them even a couple hyper 212 with push pull fans on each peltier would do the trick I would think. That way you could keep the cold side within an acceptable range. From my understanding the more you can deal with the heat side the better the cold side will act.
 
NM the way you work with the TEC, as conum said, you need Power to remove power... this mean added heat in the room for not that much of a gain using TEC.

It's much better and efficient to build something with an old watermachine or a real waterchiller.

TEC cooling have been tryed MANY times in the past. If conum ( the WC god out there ! ) tell you not to use it, dopnt use it .... I personnally tryed TEC cooling too and it was a real pita to work with and the efficiency was horrible since as told, you need power to remove power ....
 
OK I threw together a very fast and rough picture of what I am talking about for my idea, see if that helps with what I am seeing for this. 2 hyper 212 with 2 fans each in a push/pull config, a pelteir the rad with push/pull fans, another peltier another hyper 212 with push/pull fans, its own power supply and the whole thing housed in its own case with just the tubes feed in to the the cpu. Is this really that insane?
 

Attachments

  • cool-1.png
    cool-1.png
    451.1 KB · Views: 83
NM the way you work with the TEC, as conum said, you need Power to remove power... this mean added heat in the room for not that much of a gain using TEC.

It's much better and efficient to build something with an old watermachine or a real waterchiller.

TEC cooling have been tryed MANY times in the past. If conum ( the WC god out there ! ) tell you not to use it, dopnt use it .... I personnally tryed TEC cooling too and it was a real pita to work with and the efficiency was horrible since as told, you need power to remove power ....

Ohh I am not disagreeing with anyone, I just not sure I am explaining my side of it well. I trust the experience and knowledge here or simply I would not be here. Just maybe my old fashioned thinking, if my car runs hot I make the rad cooler which in turns makes the engine cooler. I don't make the engine cooler to make the rad cooler.
 
Yes. You look at the area the TEC would cover vs the total water flow area. You will ONLY cool a very very small amount of the water passing through the rad. The area of the TEC is I'm sure under 1% of what the fin area is on the rad.

I'll just let you know honestly, it's a pipe dream and it will not make any measurable difference in the water temps of the rad. I can assure you that worldwide this has for sure been tried and tried again. And never ONCE have I seen a completed project like this that actually worked.

Otherwise, :bang head:bang head, it's your stuff.
 
OK I can respect that as I said I value your expertise and thoughts in this area if not then I would not be on these boards.
 
And never ONCE have I seen a completed project like this that actually worked.

Otherwise, :bang head:bang head, it's your stuff.

I seen one single TEC project that actually seemed to work great and it was a real peice of engineering ... The guy ran a dual ( if not triple or quadruple ) loop system with the main loop cooling the CPU and the other loop cooling the main loop with blocks and TEC ... He had like 2000w of TEC and 6 or more 360mm radiator ,.... the project was insane and possibly cost 1500+++ dollars ....

Do what you want but this will not work. Trust me, and Conum.....

I'm out.
 
If you wanna use a TEC, you need really strong ones and a water loop to cool them. The proper way to use them is directly over the cpu, and cool the hot side with the loop.

Problem is they don't cool all that well, going below 0 will be impossible, you ca however run at ambient temps or slightly below for giggles, but I'm not sure it's worth the work.

You need a custom waterblock and mounting mechanism, and a controller for the TEC.
It's a lot of work for little gain.

Try the water chiller idea I tossed if you want something more extreme, if not build a regular loop or use air cooling and call it a day.

Look at this guy's tec setup: http://www.overclock.net/t/1406591/tec-block-on-4770k
 
OK been studying a little bit, and see some idea that thanks to where I work I have the ability to try. I am going to try a couple peltier coolers stacked with voltage control to try and keep them right above the dew point, then the top hot side cooled with a Coolermaster hyper 212 with push pull fans, if what I am reading is correct a colder/hot one (more wattage) cooled by a smaller less rated wattage, then cooled by either liquid or a good air system should work as long as you don't run it past the dew point and start the condensation issues.... Got some stuff ordered for this, and a couple play boxes at work to try them in just to see in they will function and to set up some control for them, sound like fun and if not I am sure I could find some thing to do with a couple peltier coolers.
 
OK I got a 40mm peltier today so here is what my study so far has reveled.

More voltage in, more cold out also more heat out.

Most people I have seen are putting these on the 12v leg of the power supply while there is a 3v and a 5v leg that would stop the peltier from cooling to sub zero temps, reducing the need for exotic cooling, and reducing the amount of condensation to deal with.
one of the first "test" that they suggest involves hooking a battery to the peltier to determine which side gets cold/hot with the leads hook red to positive and black to negative, if you can get cooling with a battery I am confused why they need for 12v is even attempted unless you just want sub zero on the cold side.

So now I have the peltier, a Coolermaster 212 configured with 2 fans in push/pull. Thankfully I also have several older desktops at the shop to experiment with.

Hoping to drop the cpu temps to -10 below ambient, which right now would be around 14c.
 
Problem is you get sub zero easily without an active heat load on it...that's what the peltier rated power means, how much heat it can drain. For -10c you'd need an absurd amount of cooling for the hot side and a really powerful peltier, or a few of them in paralell (water cooling is a must).
That for modern cpus of course...you can easily freeze a pentium 3.
 
I think it's way doable, its only about 14c drop which isn't that difficult, but so far it's cost me $5 the computers and cooler, etc are free so worst comes to worst I blow up a old computer and a $5 peltier. I remember a day when any water cooling was considered insane, so it gives me something to try for.
Right now I am running about 27C with the liquid cool system alone, and thats at the overclock listed in my siggy, not to far from ambient right now, all I am looking for is a simple 10c drop, not run at 100% prime95 at 14c, just a simple 14c at no load idle on a stock cpu to start with. The Hyper 212 would cool this 6 core at 4.4 2 40c so if it will do a stock processor at less then the 212 would do it without major water then thats simply all I am after.
 
Back