Jump to content
SAU Community

Water Cooled (chiller) Setup

T-Rex

By T-Rex,
in PC & Games Discussion

 Share

Recommended Posts

T-Rex Enthusiast

T-Rex

Posted
Posted

Hey Guys,

I have been playing around for a few weeks with my first water cooling setup. I had no clue what I was doing when I ordered all the parts but everything turned out pretty good! I like to do things a bit differently so I went with an aquarium chiller in place of a radiator. These are not for the faint hearted, there is a decent sized fan to keep the compressor and condenser coil cooled. The overall sound level is less than a single GPU's fans at full speed though. Its also bigger than my Bitfenix Prodigy case.. Just!

Ok so the basics of my setup are as follows;

EK Supremacy EVO Elite cooling an i7 4790K

EK FC780 cooling a single Gigabyte 780Ti Windforce OC

XSPC D5 Photon 270 reservoir/pump combo

TECO TK500 Aquarium chiller

You can use a much cheaper chiller than the one I opted for, but I figured a quality one will last longer and handle the "different" usage of trying to cool an overclocked gaming PC!

In terms of overclocks I'm pretty happy with the results. My CPU is a poor overclocker and I couldn't get a stable clock above 4400MHz using a Corsair H100i but I managed a stable 4800MHz clock using the aquarium chiller. This took 1.45V core voltage and temperatures were higher than I wanted, 70-80 degrees under sustained full load. But 1.45V with the H100i hit 100 degrees in literally 1 second and crashed the PC! I tried higher than 4800MHz clock but it crashed and I think it was because the maximum multiplier is 48. After that you have to adjust frequency? Not too sure. I have settled on a 4600MHz overclock at 1.35V to keep the temperature lower.

On to the GPU, Gigabyte factory overclocked the 780Ti to 1020MHz. I have pushed that to 1160MHz and I see a permanent boost clock of 1315MHz due to running so cool! That's almost 100MHz more than I saw on air while I was regularly reaching 80-85 degrees where the boost clock would drop.

I normally run the coolant temps at around 25-30 degrees but can go as low as 10-15 degrees while gaming but there is no benefit with my overclocks.

I will get some pictures up as soon as I can, they won't be pretty as it's still in test setup but should give you some idea of what's been done. I do plan on running SLI graphics and I think that will be the limit of my small chiller, I calculate its capacity at roughly 450-500W.

Throw some questions at me or call me stupid, whatever you desire. I had very little information on doing this so maybe someone on there could benefit if they ever wanted to try this out.

Power consumption is about 100W. The unit uses 220W but only runs for 25 minutes of the hour. More than a normal radiator and fan setup but not enough to be worried about.

cheers,

Shaun

Link to comment
https://www.sau.com.au/forums/topic/455095-water-cooled-chiller-setup/
Share on other sites
Cowboy1600 Grand Master

Cowboy1600

Posted
Posted

On a side note, my Corsair (h80i I think) leaked while the cheap Coolermaster 120V is still going strong.

So disappointed in Corsair.

I've got a H100 and it's powering along. Had it for probably 15 months or so.

I didn't bother measuring pre and post temps but gut feeling is it runs probably 10-15 lower than on air and a lot mroe stable, ie. no big spikes when pushing it really hard.

T-Rex Enthusiast

T-Rex

Posted
  • Author
Posted

I found with CPU cooling that unless you de-cap it you will never get the low temps you can achieve at the GPU. I have not de-capped a CPU before and I don't think its really worth it as some CPU's just wont overclock well even if temps are good! Years ago I had a CPU at minus 30 degrees under idle and under full load was at 10 degrees. Luckily that CPU was a good overclocker (i7 920 stock 2.66GHz, 4.8GHz overclocked and stable) I then used a i7 930 which is 2.8GHz stock but couldn't get a stable overclock above 4.5GHz.

As for the H100i, I had no issues with it. I know they had an issue with them leaking when they were in sub zero temps in the airplanes that were delivering them.

  • 3 weeks later...

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share
Go to topic listing


  • Similar Content

  • Latest Posts

    • blind_elk
      How old is too old for tyres?

      By blind_elk · Posted

      My understanding is that UV tends to accelerate the aging process. If the car has been garaged, then you could probably get away with extending beyond 10 years. FWIW, in 2015, I had tyres on my 180B SSS that had a 3-digit code (2 for week, 1 for decade), ending in 0, so could have been more than 30 years old, but still worked fine. I did replaced them very quickly, though, once I discovered what the code meant!
    • GTSBoy
      Help Understand GTT Boost Setup

      By GTSBoy · Posted

      But we haven't even gotten to the point of talking about stateless controllers or any of the good stuff yet!
    • Duncan
      Help Understand GTT Boost Setup

      By Duncan · Posted

      You guys need to take this discussion to another thread if you want to continue it, most of the last 2 pages has nothing to do with OP's questions and situation
    • MBS206
      Help Understand GTT Boost Setup

      By MBS206 · Posted

      And this, is just ONE major issue for closed loop control, particularly using PID. One such issue that is created right here, is integrator wind up. But you know GTSBoy, "it's just a simple PID controller"...  
    • GTSBoy
      Help Understand GTT Boost Setup

      By GTSBoy · Posted

      Nah. For something like boost control I wouldn't start my design with PID. I'd go with something that originates in the fuzzy logic world and use an emergency function or similar concept. PID can and does work, but at its fundamental level it is not suited to quick action. I'd be reasonably sure that the Profecs et al all transitioned to a fuzzy algorithm back in the 90s. Keep in mind also that where and when I have previously talked about using a Profec, I'm usually talking about only doing an open loop system anyway. All this talk of PID and other algorithms only comes into play when you're talking closed loop boost control, and in the context of what the OP needs and wants, we're probably actually in the realm of open loop anyway. Closed loop boost control has always bothered me, because if you sense the process value (ie the boost measurement that you want to control) in the plenum (after the throttle), then boost control to achieve a target is only desirable at WOT. When you are not WOT, you do not want the the boost to be as high as it can be (ie 100% of target). That's why you do not have the throttle at WO. You're attempting to not go as fast as you can. If the process variable is measured upstream of the throttle (ie in an RB26 plenum, or the cold side pipework in others) then yeah, sure, run the boost controller closed loop to hit a target boost there, and then the throttle does what it is supposed to do. Just for utter clarity.... an old Profec B Spec II (or whatever it is called, and I've got one, and I never look at it, so I can't remember!) and similar might have a MAP sensor, and it might show you the actual boost in the plenum (when the MAP sensor is connected to the plenum) but it does not use that value to decide what it is doing to control the boost, except to control the gating effect (where it stops holding the gate closed on the boost ramp). It's not closed loop at all. Once the gate is released, it's just the solenoid flailing away at whatever duty cycle was configured when it was set up. I'm sure that there are many people who do not understand the above points and wonder wtf is going on.  
×
×
  • Create New...