cancel
Showing results for 
Search instead for 
Did you mean: 

New Water-Cooling Setup

Dr__Zchivago
Level 12
Hey all,

I'm beginning the process of putting together a new boutique build so that my previous WC rig can be dedicated solely to math research, and I would like some feedback on my ideas. The components are set, and will not change; but, I would like to hear about possible air-flow ideas and anything that may be critical that I may have missed.

I picked up a brand new Corsair c70 Vengeance case for cheap (only missing the HD trays, which would've been removed anyway), and attached is my proposed flow diagram.

I intend to run all of the case fans at the lowest speed, and the radiator fans will be controlled via BIOS (maybe, and at first).
Initially, I will be water-cooling the CPU only - and, since the fans on the graphics cards are practically silent under heavy load - I'll be adding a second loop for graphics later.

I know, I have an unhealthy love for anything Noctua.

32295

The image doesn't show it, but the expansion slot covers are ported, but can always be removed to allow reduction in case pressure (ideas).

System Components:

PSU: Seasonic Platinum 1000W
MB: Asus Crosshair V Formula Z
CPU: AMD FX-8320
RAM: Crucial Ballistix Elite 2x8 DDR3-1866
Graphics: (x2) Asus Radeon HD 7850 GDDR5 2 GB v2 (in Crossfire)

Let me know what you guys think! And, thanks!

Dr. Z
128 Views
27 REPLIES 27

Dr__Zchivago
Level 12
Ahh, I see what you're getting at. I have made a bracket that will mount the radiator to the lower media tray, and brackets to mount the radiator to the floor are included in my order from EK.

Ultimately, the radiator will be tied on all four corners to the case - the radiator fans will be ~1.5 cm from the ends of the graphics cards' fan shrouds, and the intake side of the radiator will be ~4 cm from the intake NF-P12s. I'm hoping that the side-panel fans provide enough cool air intake to not hurt GPU cooling too much - especially since air will be flowing around the radiator as well - but I'll see once it's all running. I can always mold some plastic shrouds to isolate the GPU heat-sinks from the radiator output (and give them dedicated intake-case fans), but only if it becomes necessary - as you can probably imagine, it's a serious pain.

If all else fails, and I absolutely have to mount the radiator to the front of the case, I can always use three or four of the little brass motherboard mounts in series to put some distance between the case fans and the radiator (that's how my previous water-cooled build was set up). http://rog.asus.com/forum/attachment.php?attachmentid=31016&d=1386864591

Antronman
Level 10
Ah. Well good luck. Glad it's an AMD setup. When mantle comes out, we're going to destroy the competition.

On another note, I am laughing about the specs on the Xbone and Piece of Sh!t 4.
The majority of the games aren't even on 1080p resolution, and the guys taking them apart are talking about how the PS4's GPU is 50% more powerful. And all the fools in the comments are talking about how the Xbox one and PS4 get "better patches" than PC games, and are "as good as middle-end pcs". America is too lazy to even do something that will allow you to enjoy your laziness even more...
Say hi to the next generation.

Peace is a lie, there is only Passion

Through passion, I gain strength

Through strength, I gain victory

Through victory, my chains are broken

The Republic of Gamers shall free me

Dr__Zchivago
Level 12
1. While I am straight-up PC gamer who owns zero consoles, consoles have the benefit of only performing a few jobs at a time, while a PC has A LOT more on its plate all the time.

2. Using #$%^&* to mask inappropriate language in this forum isn't tolerated, as I understand; so, I'd knock that off.

3. This thread has nothing to do with consoles.

4. AMD FTW

Dr. Z

Antronman
Level 10
That depends on the tasks the PC is performing. If all the PC is doing is running the OS and another program, it is not doing much more than a console. A console is just a very simplified, low-end pc.
PCs will always do better, AMD will always be my favorite unless they go out of business. I just wish there was a steamroller 8-core :(.
Say hi to the next generation.

Peace is a lie, there is only Passion

Through passion, I gain strength

Through strength, I gain victory

Through victory, my chains are broken

The Republic of Gamers shall free me

Dr__Zchivago
Level 12
Okay guys, it's time to make a final decision on operating systems.

I have licenses for Win7 HP and Win8.1 Pro. However, I can't decide which to use.

Given the components in this machine (see original post), does Windows 8.1 still have issues with AMD drivers and Crossfire compatibility?

I'm willing to learn 8, I got 8.1 for free from my university, but 7 is so comfy.

Z

Antronman
Level 10
8.1 Pro 8.1 Pro 8.1 Pro...8.1 Pro.
Say hi to the next generation.

Peace is a lie, there is only Passion

Through passion, I gain strength

Through strength, I gain victory

Through victory, my chains are broken

The Republic of Gamers shall free me

kkn
Level 14
w7HP supports 16gig on 64bit and 4 on 32bit.
w8 pro supports 512gig of ram on 64bit, and 4 on 32bit.

nleksan
Level 10
? Push Pull setups have, not a huge, but certainly a discernable benefit over single-side fan setups for radiators....

Plus, you have more options in the seemingly endless war between noisy and cool versus quiet and hot. P-P is worth between 20-50% increase depending on fin density and proper fan selection (and that's assuming you're using fans properly suited to the radiator, general rule I use is no less than 1.5mmH2O but 2.25 or more is ideal, once it gets to between 12-17fpi or so then a minimum of 3mmH2O; 18-24 and 3.5-4mmH2O minimum; 25-30fpi and you want at least 5mmH2O).
What push-pull does is the following:
- INCREASE *Effective* static pressure
- Disturb boundary layer airflow, increasing first-molecule heat transfer efficiency
- Increase Total Volume of air moving through radiator over any given period of time (say we go from 30cfm to 60cfm; that's 2x as much air to move the same amount of heat)
- Increased overall efficiency

Heat transfer depends on the difference, or DELTA, between what is radiating and what is absorbing heat; the greater the dT, the more heat can be transferred in x amount of time.
Since air is a poor substance to use for absorbption of heat, we have to use A LOT of it (hence the many thousands of percent increase in surface area via high fin count radiators) otherwise stagnation occurs and the radiator would simply heat up with nothing to cool it.


Another thing....

When using push-pull, you can achieve the same effective static pressure/airflow that you get from JUST Push OR Pull, but at significantly lower fan speeds while still retaining the majority of the other benefits P-P brings with it.
Since "noise" increases as a logarithmic function, 2x as many fans operating all at the same speed does not mean 2x as much noise; it's actually an increase of around 1.8dB on a 480 radiator. For reference, an increase of 10dB (technically it's about 8.93dB, but 10 is so much simpler) is an approximate DOUBLING of perceived loudness; furthermore, the generally accepted sound pressure increase required for the average person to NOTICE an increase in loudness is +3dB, much below that and it goes from "you might hear the difference listening really closely in a perfect environment" to "only in an anechoic chamber while holding your breath" to "physically, and anatomically, impossible".


Consider that just driving in the highway, with windows up and no sources of direct sound (radio, etc), in a brand new BMW 760Li or Bentley Continental GT Speed (premium luxury cars, especially European, and in particular German, have tens of millions of dollars put into keeping the cabin as removed from outside noise as possible), the AMBIENT SOUND PRESSURE is OVER 65dB!!! For a 5 to 10 year old, significantly less expensive economical vehicle like a Civic, you can expect it to be TWICE as loud: aka up to 75dB!

Have you ever ridden in a higher end or entirely-luxury-focused car and thought "my God, the sounds are everywhere, so unbearable...the noise, THE NOISE!!!!!"?
I know that is something that has never once crossed my mind, and I eat and breath BMW, Porsche, etc... My (incredibly) extensively modified 2003.5 BMW M3 track car (weight stripped down to 2841lbs from 3481lbs; power from 272rwhp/239rwtq to 369rwhp/279rwtq), which has a suspension that would turn you into an unwitting kidney donor should you try to cross a speed bump anywhere in excess of, oh, about 2 miles per hour, has zero sound deadening, a welded chromoly steel 14-point cage, no catalytic converters/mufflers (SuperSprint V1 Stepped headers into custom Burns X-Pipe into Akrapovic Race Evolution Titanium exhaust... so basicallly, really expensive straight pipes).... and which rides on 18x10f 11r wheels wrapped either in only-technically-street-legal tires like Pilot Sport Cup 2's or not-legal-anywhere-but-closed-course-track tires (R-Compound; the rubber is soft enough that I usually get 2 weekends out of a set, and if I'm competing then a fresh set is used every event)....

It's interior sound pressure level, as measured?
Idle: 73dB
70mph in 6th Gear: 77dB
2nd Gear Redline (8750rpm engine speed and 64.4mph): 84dB

Or, how about a firearm?

Shooting the lovely .300 Blackout round, 220gr Hornady subsonic handloads (1105fps avg), through my Noveske 10.3"-barreled custom SBR AR with a lovely GemTech titanium-billet .30cal suppressor on the end?
98-111dB depending on conditions.

Shooting the lovely (but in a very different way) .338 Lapua Magnum round, 300gr Hornady BTHP accurized handloads (2925fps) from my Accuracy International AWSM, with the 20" Surgeon barrel and just a crowned muzzle?
181dB Peak
Throw on the AAC Titan-Ti supressor with the 24" fluted and threaded accurized barrel and with the same supersonic loads?
150dB Peak
(goes down to 122dB with subsonic rounds through the can)



Point is that fan sound pressure levels are so HILARIOUSLY and blatantly lies that I wish they'd save the ink and not even print them. I promise you, any two fans at the same speed will not be more than ~5dB apart and that mostly comes down to the motor (about 75:25 motor:blade design). For any REAL change in LOUDNESS, you need to have a change in AIR SPEED which is only affected in any meaninful way by increasing the speed of the fan!




Anyway, sorry, for the rant.


I'd add a second 240 or 280 rad up top, too, if you're cooling more than just a CPU.

Dr__Zchivago
Level 12
The radiator is 80 mm thick (10 fins per inch), and is rated for fans rated at 1.5 mmH_2O, and the NF-F12s that will be used in pull are rated at 2.6 mmH_2O.

I don't have the room for a push/pull configuration in this case, and the radiator isn't going to change.

Plus, I have adjusted the case so that I get outstanding case air-flow and all fans are running at ~850 RPM (and those NF-P12s are almost silent at 1200).

Currently, the only fan I can actually hear is the CPU fan, and I am definitely a stickler for noise.

The rig that has been dubbed "SRT8 - Challenge This!" after a custom 2010 Dodge Challenger (white body, black striping, red accents, tan interior), and it's officially finished!

Initial performance has been phenomenal - I've managed to pull the CPU up from 3.5 GHz to 4.74 GHz (21.5x220, see signature). I'm still working on memory - that appears to be a wholly different monster.

Overall, however, my CPU idles in Windows at 27 degrees C (19-20 ambient), and under current overclock, tops out at 54.

To those of you with the AMD experience, I'd like some input on where I should go next!

Thanks all!

33114331153311633117