Cooling the water in a wc loop?

[6_5x47lapua]

water cooling and air cooling are very effective, but the lowest temp you can hope to achieve is room temperature.
I understand that if you can have a system run at room temperature than you are doing damn good, but unless you have a massive, massive cooling setup your temps will run higher, becoming less and less effective the hotter the environment.
refrigerated systems have problems with condensation and such, most people stay clear of them for alot of reasons.

but just a theory, what is stopping from having a water loop with the water running through blocks on a few peltier coolers/cold plate or through a mini fridge?

i know there is disagreement if a fridge can handle a heat source, (that will take a bit of R & D testing) but is the theory good?

what else could cool the water? A small ac unit blowing through a large inline heat sink or radiator?

i like to look at all options, even crazy ones, it would be cool (pun intended) if it was efficient and preformed well though!

 

[rubix_1011]

Fridge or freezer are not designed to continually remove heat- this will cause the compressor to burn up. They are only designed to remove heat to a specific temperature and then maintain that temperature. You could potentially have an A/C unit blow through a large radiator, but you'd have better results using the A/C or fridge compressors modded into phase change or water chillers. Many people have done this, but it's a fairly advanced cooling method...you need to know what you are doing.

 

[6_5x47lapua]

sounds interesting, i will have to read into that.

i know this would not be effective to cool a entire system, but how effective would would it be to take a cpu block (one with a back plate) and clamp a cpu sized peltier cooler into it, just to take the edge off the heat, say between gpu and cpu in the loop?

 

[rubix_1011]

I guess in theory it would work, but in most instances where peltiers are used, the hot side is cooled by watercooling...how would you go about cooling the hot side?

 

[4Ryan6]

Here's a simple below ambient water cooling solution.

http://www.tomshardware.com/forum/267412-29-sandy-bridge-water-cooling-project

 

[6_5x47lapua]

i guess it would be kind of pointless since in essence, you would be back where you started. what actually happens if you don't cool the hot side? there are no sensitive components to fry... unless the high heat harms the plate itself, or if the heat transferred through the back plate and back into the block, is it totally necessary to cool? anything more than an old cpu sink and a fan on the hot side is starting to be a bit bulky, and straight up stupid for this idea.

+1 ryan, good article, but i could see it being a pain to be swapping out ice constantly, a little more than i want to mess with all the time.

 

[rubix_1011]

i guess it would be kind of pointless since in essence, you would be back where you started. what actually happens if you don't cool the hot side? there are no sensitive components to fry... unless the high heat harms the plate itself, or if the heat transferred through the back plate and back into the block, is it totally necessary to cool?

The TEC gets damaged and no longer works- you need to cool the 'hot' side.

anything more than an old cpu sink and a fan on the hot side is starting to be a bit bulky, and straight up stupid for this idea.

This is why there are integrated waterblocks on the hot side of most watercooling TEC's...the watercooling loop cools the TEC, while the peltier cold side super cools the CPU/GPU component it's attached to. Applying the cold side of a TEC to a watercooling component (like a radiator or block) will only cool the water in the specific area for a very brief time...much like water in a waterblock only gets heated by a CPU or GPU under load for a very brief time.

You have to remember that for every action, there is an equal and opposite reaction. Water responds to heat sources similarly to how it responds to cold sources- you are simply suggesting to swap some of them, but not address the opposite effect needed to be taken into account; this is leaving unaccounted holes in your cooling idea.

 

[6_5x47lapua]

thanks. i don't think i want to get into too much of this then. I like phase change cooing except, i would like to cool cpu and graphics card(s). and it wouldn't really make sense to have phase change AND a water loop. i don't know much about phase change yet- are there two tubes going to the cooling plate, or just one? two would be cool, as you could have a sequence of cooled components. i know they have dual evaporators, but over 1k is a little out of my range.

 

[rubix_1011]

You can build your own phase cooler...there are a lot of DIY's out there on that...just not on this site.

http://www.overclock.net/f/133/phase-change
http://www.xtremesystems.org/forums/forumdisplay.php?80-Vapor-Phase-Change-Cooling

 

[6_5x47lapua]

yesh. just read into it a little bit.. im not scared to mod, but getting condensation and drips is a bit too scary for me. i understand that if its done right, you are relatively safe, but that almost negligable tolerance for error is putting me off, especially something of my design.
I am fairly confident in my abilities, but i'm no HVAC tech. plus this doesnt solve the fact that i would still have to cool the GPUs as well. (i can get a really good deal on watercooled ones from a friend who isnt using them, so just going with watercooling for them is a necessity.

something else i ran into is cooling the water loop with phase change.

the obvious problem is freezing the water. that can be solved easily by using something along the lines of anti freeze instead. i may still have to do some insulating on the watertubes, but so be it.

a submerged evaporator is the most simple method i can see, plus no added restriction of a block.

i don't need crazy low numbers like -55 that some people are getting, even only sub-ambient i would be very happy with. A very small, low power phase changer that could maintain 40F would be perfect (assuming the theory is good, and even if phase changers can run that hot.)

PS- i just wanted to compliment you rubix, you are most helpful, and ever courteous. i wish everyone on the internet was more like you! +1 for general helpfulness!

 

[rubix_1011]

something else i ran into is cooling the water loop with phase change.

the obvious problem is freezing the water. that can be solved easily by using something along the lines of anti freeze instead. i may still have to do some insulating on the watertubes, but so be it.

So you are looking at something like a water chiller build, then?

a submerged evaporator is the most simple method i can see, plus no added restriction of a block.

I'm curious how you have this drawn up, mentally...It's interesting, so if you have a few mins, can you elaborate/provide pics/info? I've seen a handful of chillers, so I'm wondering what you had in mind, here. I'm not trying to discourage you at all...if anything, I'm one of the biggest promoters of DIY (I built my own case out of wood for crying out loud). If you are going to go this route, please document and post pics...we need some projects with additional cooling alternatives. It's on my list at some point in the future to convert a phase or chiller.

PS- i just wanted to compliment you rubix, you are most helpful, and ever courteous. i wish everyone on the internet was more like you! +1 for general helpfulness!

Thank you for the kind remarks- I just know that it's difficult to trudge your way through concepts that are new to you, but there really isn't an easy way to dive in and get acquainted with the terminology and technology without sticking your neck out. We've done a pretty good job here over the past year turning a revolving door forum into more of a hangout with some very good information and long-term forum members. In fact, we were so disorganized that we were trying to pick and choose watercooling topics out from the heatsink and air cooling questions. Now we have a culture, organization and ongoing support. It's been tough to get things turned around, but with the effort of many, it has transpired into what you see today.

I agree that Tom's isn't the world's foremost watercooling expertise forum, but many of us are on other forums and have the ability to migrate knowledge and experience (learned and personal) to members here so that the interactions we have here are much higher quality than we ever had before.

 

[silverliquicity]

I like the sound of Electro Thermic Cooler.

 

[rubix_1011]

TEC or actual phase change?

 

[6_5x47lapua]

Here's where i got the idea from, meant to post but got ahead of myself.

http://www.overclock.net/t/761977/proposed-phase-change-cooled-liquid-cooling-loop

I will draw up a blueprint when I have access my other computer, have no good sketching software on this one.

the above is not my picture or exact idea, but it is very similar, where I am basing mine from.

 

[rubix_1011]

Yeah, water chiller. This is possible, but depending on the size of the heat exchanger and how you plan to perform that piece of the project, will depend on overall performance.

 

[6_5x47lapua]

(not pretty but it works)
No heat exchanger, evaporator is in the reservoir. I figured less flow restriction, why not?

 

[rubix_1011]

Yeah, I've seen a couple of builds like this in person. It takes a decent amount of work and know-how, so make sure you do your homework.

If this transpires, let me know- I'd like to see the end result and a worklog if you decide.

Good luck.

 

[6_5x47lapua]

will do sir- thank you for all your input.

my one final concern is that if i cant get a phaser to run hot at around 40 F then the extreme low temp will carry into the water, causing the need for insulation.
no biggy- i can foresee difficulty insulating around the CPU block and GPU(s). factory evaps. have shrouds built around them, might be a bit hard to replicate.

Depending on if the sub-ambient temps carry on to the GPU blocks, that would be really hard to insulate since you would need to do the block itself.

 

[rubix_1011]

my one final concern is that if i cant get a phaser to run hot at around 40 F then the extreme low temp will carry into the water, causing the need for insulation.

40 F or 40 C? Big difference...40 F isn't warm at all...

Depending on how cold you run and ambient temps and ambient dew point, you'll definitely need to consider insulation, and you'd have to do a lot of insulation and waterproofing on more than just blocks and tubing...

 

[6_5x47lapua]

40 F. I know its not 'hot' but pretty warm for a phase changer, being as they run around -50 (not sure if F or C on this one, either way its really cold) in most cases i've seen them, I'm not even sure if they can operate that high.

the answer you want to jump to 'just use a small one' but there is a chance that the actual reaction that occurs is always that cold, and size of the device only increases the area it can cool. (just my theory, i might be talking out of my A** on this one. )

 

[rubix_1011]

This is beyond my knowledge. You might want to give some of the guys in those forums a shout and see what you get back.

 

[6_5x47lapua]

And yep, pretty much everything dealing with the changer will have to be insulated, i was just naming what could potentially present problems. thats why i would like to run near-ambient but above freezing, and avoid the issue entirely if it proves to be reliable.

With a radiator also in the loop the changer doesn't necessarily have to be run all the time, I probably only will when I game, or such.

 

[rubix_1011]

It all depends on how you plan to go. A lot of people consider having alternative cooling in place when they don't run their primary active cooling.

 

[6_5x47lapua]

haha that makes two of us. i may post my findings on this thread so anyone following it isn't left hanging

 

[rubix_1011]

That's fine, sounds good.