|
|
General Liquid/Water Cooling Discussion For discussion about Full Cooling System kits, or general cooling topics. Keep specific cooling items like pumps, radiators, etc... in their specific forums. |
Thread Tools |
02-10-2005, 01:44 PM | #1 |
Cooling Neophyte
Join Date: Feb 2005
Location: Canada
Posts: 6
|
Cooling the water with a peltier?
Hi
This is going to be my first attempt at w/cing my current system and I'd appreciate your help guys on developing the following idea I had (which I think has never been attempted). Take a look at the following GIF diagram: Basically, my search for a satisfying radiator turned out to be rather disappointing (even tried a Topaz GL 88 radiator at some point). I just need some way to cool down the water silently and I'm thinking I could maybe take advantage of the cooling function of a peltier. Now, please keep in mind the following points: - I am not overclocking. - I do not want super cool temperatures for my CPU / GPU (my CPU currently runs at around 40 C and I'd be happy to keep it that way. Although, if it can be lower, I'm not going to complain) - My ultimate objective with water cooling is total silence. I'll be making some more changes to my case once this works to further dampen the noise. My current system is an old 900 Mhz Athlon (Socket A) processor and a GF2 GTS video card. I most probably do not want to w/c the GPU now, but I'd like the w/c system I'm building to possibly be able to take up the load from a newer system in a future upgrade. So here's the idea: Attach a "medium" (copper?) heatsink to the cold side of the peltier, submerge most of it into the water (contained in what would act as horizontal reservoir) and attach a large natural covection heatsink on the hot side of the peltier to hopefully (you tell me! ) pump enough heat off the peltier. We could have two such assemblies in parallel as illustrated. The water would flow from side to side through the submerged heatsinks into and out of the fittings on the sides of the reservoir. So, what is the feasability of what I am suggesting here? Remember that the idea is to get the water back to ambiant temperature or close. That would be good enough. Which means we may not even have to run the peltiers at their max potential, thus making passive convection possible (or so am I naively thinking...) I tried making some calculations, but hell if you dont need a PhD in thermodynamics at that point. But I'll continue searching. I declare the colloquium to be officially open. Feel free to trow in what you think/know. Last edited by Skytied; 02-10-2005 at 02:32 PM. |
02-10-2005, 02:03 PM | #2 |
Thermophile
Join Date: Jun 2001
Location: The deserts of Tucson, Az
Posts: 1,264
|
Passive radiators work fine. If you want one, get one.
Passive pelts are a bad idea, because now you'll have 2 or 3 times as much heat, but less surface area. Whats the idea of that? To go from cooling a (potientially huge) radiator to cooling a small die? Not going to help you. Edit: spelling Last edited by redleader; 02-10-2005 at 02:08 PM. |
02-10-2005, 02:43 PM | #3 |
Cooling Neophyte
Join Date: Feb 2005
Location: Canada
Posts: 6
|
Well, yeah, maybe you're right. But depending on how realistic such a setup can be, if the advantage is getting water temperatures that are well below ambiant, then I see a nice advantage (considering overclocking applications, for example).
|
02-10-2005, 03:45 PM | #4 |
Cooling Neophyte
Join Date: Jan 2005
Location: UK
Posts: 96
|
But in your first post, you said you weren't overclocking or bothered about getting really cool water?
A large passive radiator, or underground pipes or elaborate arrangement of copper piping to dissipate the heat will be better. A simple example of why passive pelts won't work is this: Say your CPU and gfx put out 100 watts of heat. To keep the water temperature at ambient, you'll need more than 100 watts of peltiers to cool the water. Maybe even 200 watts worth of pelt. Now you have to passively dissipate the 200 watts of pelt rather than the 100 watts of CPU and gfx in the first place. Peltier water-chilling has been done, but I believe it needed one water loop to cool the peltiers and one to circulate the chilled water to the CPU and gfx card. Not cheap and not efficient, but probably fun! |
02-10-2005, 05:37 PM | #5 | |
Cooling Savant
Join Date: May 2004
Location: Northern VA
Posts: 383
|
Quote:
|
|
02-10-2005, 06:44 PM | #6 | |
Cooling Savant
Join Date: Sep 2002
Location: ohio
Posts: 140
|
Quote:
If you want cold temps, you need two water loops. I suppose it could be done with air-cooled pelts, but i bet you would need at least 5 or 6, and then you have to have fans to cool the hot side of you pelts anyway, so you might as well just use a radiator (much more economic). |
|
02-10-2005, 07:46 PM | #7 |
Cooling Neophyte
Join Date: Nov 2004
Location: Mendoza , Argentina
Posts: 1
|
i´m bulding a system tha uses a peltier but before it there´s a reservoir made of cooper (works as a pasive radiator) and a radiator with 2 120mm fans @ 5v. So most of the thermal energy is removed before the peltier.
With a 75w peltier cell you can chill water almost to ambient temp sorry about the bad english. |
02-10-2005, 08:45 PM | #8 | |
Thermophile
Join Date: Jun 2001
Location: The deserts of Tucson, Az
Posts: 1,264
|
Quote:
|
|
02-11-2005, 10:54 AM | #9 | |
Cooling Savant
Join Date: May 2004
Location: Northern VA
Posts: 383
|
Quote:
Yeah I know I was jokin Should have put sarcasm tags sorry. There have been some really large lab tec coolers of the same design on ebay though, can't find any right now =\ |
|
02-12-2005, 01:59 AM | #10 | |
Cooling Neophyte
Join Date: Mar 2004
Location: Canada
Posts: 18
|
Quote:
This system uses 4 - 72W TEC's wired together in series (so, they are making 1/4 the output per TEC?) mounted on a 92x92mm heatsink and has a high-speed (3000-4000RPM) 92mm fan sucking air from inside the computer through the heatsink to the back of the computer for one "chiller". This computer has a second "chiller" mounted on the top as well. This unit can cool the CPU and GPU to roughly the same temperatures as an air-cooled CPU/GPU under-load and to around ambient while idle. These readings come from the software readings in the motherboard software. I do not know why there is such a big difference in temperature except to speculate that the inefficiency of the TEC's aren't capable of dissipating a full-load and have a "breaking point" where they start to act more "inefficient" than efficient. |
|
02-12-2005, 09:24 AM | #11 |
Cooling Neophyte
Join Date: Feb 2005
Location: Canada
Posts: 6
|
molak: yeah, that could be a good idea. I'll consider it as an option if I want to build something more "extreme" in the future.
Thanks for your input, guys. I'll go with a passive solution for now and I'll leave trying exotic stuff for later. I'll actually make my rad from copper tubing. Which brings me to this question then: Is there any kind of special copper that is flexible enough to be bent into circles without causing weird stuff to happen to it? I think I've seen such tubing at a Home Depot style store here in Canada and it seemed flexible enough. It was about 7$CAD for 10' if I remember well (there was 25' long tubing as well). How long should be enough? What diameter is ideal for, say, a 1/2" system? I'm thinking maybe I can buy two 25' long "rollers" and use them in parallel like a setup I've seen on the internet somewhere. I'm ordering the pump and water block today, too. |
02-12-2005, 10:42 AM | #12 |
Cooling Savant
Join Date: Sep 2002
Location: ohio
Posts: 140
|
I've made a radiator out of coiled copper tubing before, although it was such a long time ago all i cooled with it was a 750mhz duron or something.
The tubing you are thinking of should be flexible enough, but do yourself a favor and buy a 'pipe/tubing bender' or whatever it is actually called. It only costs a few bucks and looks like a coiled up spring, but you slide it over the tubing to bend it. It works better with this tool because the tubing won't crease or fold if you use it. |
02-12-2005, 12:06 PM | #13 |
Cooling Savant
Join Date: Dec 2003
Location: France
Posts: 291
|
|
02-12-2005, 04:52 PM | #14 |
Cooling Neophyte
Join Date: Feb 2005
Location: Canada
Posts: 6
|
Albigger: Nice bit of advice, thanks.
lolito_fr: lol. I'll have a couple of bandages handy. And, oh my. Look what I found: Swiftech MCW-CHILL 452 |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|