Sign in with
Sign up | Sign in
Your question
Solved

CPU water cooling loop

Last response: in Overclocking
Share
April 9, 2010 11:02:07 AM

After searching a bunch of different forums over the past few weeks I finally decided to post here because there seems to be a lot of good info bouncing around. I am building my first CPU water cooling loop and have a few quick questions on components/configuration.

(1) I'd like to mount the radiator internally if at all possible. My Lian Li case (which I don't thinks is optimum for water cooling, but I'd rather not replace) will constrain me to using a 2X120 rad after moding the top panel. According to what I've read, An i7 930 can put out approx. 250 watts of heat energy. There seems to be a consensus that a 3x120 rad is needed to deal with this temp load. furthermore as my rad will be internal (therefore surrounded by a higher ambient temp) I will possibly be looking at diminished heat dissipation. However I found this chart (I believe originally created/posted by Skinnee Labs and copied to Sidewinder computers) which seems to show this particular rad (Swiftech MCR220-QP Quiet Power 2X120mm) dealing with over 350 watts effectively at 2.0 gpm (fan dependent). The pump I'm looking at maxes out at well over 2.0 gpm (theoretically over 5.0 gpm though I realize restrictions in the water block and rad will make achieving this rate impossible). Can this radiator handle the i7?

Here's the link that has the heat dissipation chart on it: http://www.sidewindercomputers.com/swmcqupo2x.html

(2) If I do move ahead with the system I have planned here is my parts list (I have done quite a bit of price comparison and surprisingly all the parts I wanted were least expensive through this one company :??:  )

Swiftech MCR220-QP Quiet Power 2X120mm > http://www.sidewindercomputers.com/swmcqupo2x.html

Swiftech MCP655 12v DC Pump - with Speed Controller > http://www.sidewindercomputers.com/swmc12vdcpu.html

Swiftech Apogee XT CPU Waterblock > http://www.sidewindercomputers.com/swapxtcpuwa.html

Tygon R3603 1/2in. ID 3/4in. OD Laboratory Tubing # AAC00038 > http://www.sidewindercomputers.com/ty1id3odlatu.html

Bitspower 1/2in. Compression Fittings for 3/4in. OD Tubing (6x) > http://www.sidewindercomputers.com/bi1cofifor3o.html

Bitspower 3.5in. Bay Res - Black Acetal (POM) - BP-WT350P-BK > http://www.sidewindercomputers.com/bi3bayresbla.html

Scythe SFF21F 63 CFM S-FLEX 120mm Quiet FAN by Sony (2x) > http://www.sidewindercomputers.com/scsf63cfms12.html

Petra'sTech PT_Nuke Concentrated Biocide - 10mL (to be used with DI) > http://www.sidewindercomputers.com/peptcobi1.html

As you can see the system is pretty basic. The reservoir is one thing I'm not completely sold on as it only has an 80ml capacity. I chose it because my case has 2 unused 3.5 in bays and I'd like to utilize them if possible mostly to keep the inside of the case uncluttered. From what I can tell the res. mostly exists to allow the system to be filled/bled more easily and to reduce the amount of back pressure on the pump's intake. If these are the only functions of the res. then I can't see any need for a larger volume. Maybe a larger res. might make "bumping the pump" a little easier?

(3) It seems that the preferred order of water cooling systems is reservoir>pump>radiator>waterblock>reservoir. I'm having a hard time understanding why this is better than reservoir>pump>waterblock>radiator>reservoir. This is potentially a stupid question, but to me the latter configuration seems to have the advantage of running cool water through the pump. The only drawback I can think of is that the pump could potentially heat the water slightly before sending it to the water block which is not ideal.

This post turned epic :sleep:  . Any suggestions? Thanks

More about : cpu water cooling loop

a b à CPUs
a c 86 K Overclocking
April 9, 2010 2:55:37 PM

Great work. We see this amount of homework and effort rarely here.

Yout right, the res is just a place to hold water and bleed the loop.

What your trying to say about the heatload can be expressed in Delta T. The difference between the ambient air temps and the normalized water temps. Under load you can be at 250 watts but only testing loads I think. I'd say 225 is more like it.

So, taking into account your rad and fans your DT should be pretty good. Under 10C in normal usage.

The only thing I can suggest is pull cool room air through the rad, not exhaust case air. Every C of warmer air is warmer water temps.

The difference of the water temps anywhere in your loop will be less than 2C.

Rad before the CPU is best, but just make sure the res is before the pump. Thats the only routing rule.

Keep us posted.
m
0
l
April 9, 2010 4:22:32 PM

Hey Conumdrum,

Isnt that pretty much the same thing as the Swiftech H20-220-APEX ULTIMA CPU Liquid Cooling Kit w/ Apogee XT kit at Petra' with just a few tweaks? If so he can get that kit and add the parts he wants. And maybe save a little $$.

-ouch1
m
0
l
Related resources
a b à CPUs
a c 86 K Overclocking
April 10, 2010 2:14:38 AM

ouch1 said:
Hey Conumdrum,

Isnt that pretty much the same thing as the Swiftech H20-220-APEX ULTIMA CPU Liquid Cooling Kit w/ Apogee XT kit at Petra' with just a few tweaks? If so he can get that kit and add the parts he wants. And maybe save a little $$.

-ouch1



Pretty much, yea. Much nicer res, better tubing, better quality barbs etc. But pretty much the same basic stuff. The pump is the vario version, much better.

What impressed me he actually read and dug into charts. The guy even has visited Skinnees web page.
m
0
l
April 10, 2010 4:04:04 AM

Thanks for the replies. I took a look at that kit and it is really similar to my parts list with a few other things that I don't want/need. It would probably be a few bucks cheaper to go that route, but I'd have to compromise on a couple of things. It makes sense that someone's marketing a kit with these parts, because they seem to have a good quality/performance vs cost for noobs. I guess I'll stick with the small 3.5 bay reservoir despite it probably not being the most user friendly size and design.

I did come across delta T in my research. The only issue I have with it is that its "baseline" measurement depends soley on the ambient temp. Therefore you could have 2 systems operating in environments with drastically different ambient temps, yet the delta T of the systems could theoretically be the same. By mounting the rad internally I am essentially forcing it to exist in higher ambient temp. Drawing cool air from outside the case to address this would ultimately result in the wc system's heat being dumped back into the case won't it? Doing so would potentially improve delta T (by increasing the T gradient across the radiator), but wouldn't it do so at the expense of increasing ambient temp inside the case? Am I missing something?

Now I'm second guessing mounting the rad internally. It really seems like no matter what, an internal rad will compromise the cooling ability of the system. However I have a fan mounted in front of two of the 5.25 bays which would be drawing air in directly below the rad. When I get the parts I'll experiment with the rad inside and out. I hypothesis that Delta T will remain constant regardless of rad placement, but CPU temp will improve with rad mounted externally...

Also it appears that Petras and Sidewinder decided to go on vacation together :lol:  (or else they are in fact the same company). Either way no orders are processing through Petras or Sidewinder until April 17th.

m
0
l

Best solution

a b à CPUs
a c 86 K Overclocking
April 10, 2010 11:06:51 AM

The air through the rad is the primary consideration. Yes it waill warm the insides of the parts a bit more but if you look at the rad tests on Martins and Skinness you see the exhausing air is just a few C higher. Mobos etc don't mind much. You can try the fans either way, up to you.

Mounting the rad outside is always better. Since you decided to mount on the outside, the cost difference between the 120x2 and 120x3 is small. Go for the bigger rad.
Share
April 11, 2010 4:18:32 AM

Best answer selected by flylowlikerylo.
m
0
l
April 11, 2010 4:23:11 AM

Right on, thanks for the advice. I'll order this stuff when the distributor gets back from vacation and give my i7 some needed relief. As expected that stock cooler is really pretty terrible even at standard clocks...
m
0
l
a b à CPUs
a c 86 K Overclocking
April 11, 2010 4:25:04 AM

Fly? I like your thinking. Your not the avreage Toms guy. Please come by OCF when youget a chance.
m
0
l
!