Sign in with
Sign up | Sign in
Your question

Water Cooling bootleneck

Last response: in Overclocking
Share
February 5, 2013 6:02:15 PM

I just reinstalled my watercooling loop but I am not completely satisfied with my temperatures.

Idle
CPU: 35-39 C
GPU: 24-28 C

Load (Prime95)
CPU: 51-58 C
GPU: Not tested but probably fine

I would like to bring the temperatures down, mainly on my CPU. But I do not know what would be the most efficient upgrade? Is it tubes, fittings, CPU block, pump, rad, fans or other components? Or if you suspect that I installed something incorrectly, please let me know :) 

Below is my computer specs. As you can see I do not overclock as silence, temperatures and fun is the main reason why I use water cooling. Please tell my if you need additional information.

Water cooling setup:

Laing DDC 1T (10W) pump with a XSPC Laing DDC V3 top
13/10mm compression fittings with corresponding clear tubes
XSPC RX240 radiator with 2 Noctua NF-P12 fans pushing air through
EK Supremacy Acetal CPU block
EK VGA Supremacy Acetal GPU block
XSPC Single 5.25" Bay Reservoir
EK Ekoolant clear fluid
Arctic Cooling MX thermal grease

My loop looks like this: res --> pump --> rad --> gpu --> cpu --> res

There are no other fans in my case.

Hardware:

Asrock x58 Extreme motherboard
i7 930 bloomfield (stock clock)
Sapphire Radeon HD 5850 (stock clock)
6GB Ram and some SSD disks
CoolerMaster 600W Silent Pro PSU
CoolerMaster Advanced 690 ii case

Pictures of setup:





a b à CPUs
a c 324 K Overclocking
February 5, 2013 6:09:25 PM

Temps seem fine to me...what is your ambient room temp?

Adding additional radiator space would be my recommendation, but weren't Bloomfield's relatively warm anyway? A single RX240 should easily be able to dissipate that, but you might get slightly better results with push/pull. Remember, your reported CPU temps are not the same as water temps in your loop and watercooling (just like air cooling) is 100% dependent upon ambient temps as your baseline delta.

GPU TDP 175w (from what I have found)
CPT TDP 130w

~305 watts, but in Prime95, your GPU isn't stressed, so that really shouldn't add to your thermal overhead much other than GPU idle TDP.
February 5, 2013 6:13:34 PM

Deleted
Related resources
February 5, 2013 6:20:04 PM

I have not measured the ambient room temp, but as you can see from the pictures my case is standing next to a radiator. But I tried turning it down and opening the window (freezing outside). Running a Prime95 stress test lowered the temperatures like 5-6 degrees celsius. To me that do not seem like a whole lot, so I thought there could be some optimization to be done inside the case.

I could try push pull or add another rad. But my existing radiator, or the air blown down through it, is not warm or anything. So could there be anything wrong elsewhere in the loop?
a b à CPUs
a c 324 K Overclocking
February 5, 2013 7:44:54 PM

5-6C drop in load temps is pretty decent change. The air isn't going to be very warm to you, but ambient temps do make a lot of difference. Have you tested the same scenario by pulling the side of the case and blowing a fan into the case on highest setting?

I still don't see a lot that is terribly wrong, but there might be some airflow issues to sort out. What temps were you thinking to see?
February 6, 2013 3:25:54 AM

When I stress tested with the window open I had taken both sides and the front of the case. I even installed 2 fans blowing air on the CPU.

I just woke up, and have had my radiator (heating the room) off throughout the night. The ambient room temperature is significantly lower than yesterday and now I am reading idle CPU temperatures from 29-33C. That is great, but it is slowly getting higher. So maybe you are right about airflow and ambient room temperature.

I would like to see temperatures below 30 on idle and below 50 on full load. I Do not know if that is naivly low with my current setup. But since I do not overclock I think it is a reasonable requirement :) 
a b à CPUs
a c 324 K Overclocking
February 6, 2013 2:56:45 PM

Much of it is going to be dependent upon your ambient room temp...you cannot run watercooling or normal air cooling lower than your ambient room temps without some form of sub-ambient cooling process. In short- as the temps in your room go up, so do your idle and load temps...there is nothing you can do to avoid it. I think this is more of a case and radiator airflow issue than anything else.

Try this:

Leave the windows closed like your normally would to maintain constant ambient room temps.

Load your CPU (Prime95, etc) with the sides on the case like it normally would function. Note temps.

Remove the sides of the case and blow a desk or box fan on high into the case and load again using the same tools...note temps.

This tests your current airflow against a simulated best case scenario of very high airflow in your case to determine if there is an issue with how much air is made available to your components and more specifically, your cooler...the radiator.
February 6, 2013 8:25:48 PM

Are your fan pulling outside air or are you pushing case air out through your raid?
a c 103 à CPUs
a c 190 K Overclocking
February 6, 2013 9:02:02 PM

He's pushing case air out through the rad, I'll input more when I get to my Pc in the Am
Moto
a b à CPUs
a b K Overclocking
February 6, 2013 9:35:25 PM

I would like to add that you really should have more rad space for that cpu and gpu

I am running an i7 930 @4.4 and 5870 at 925 and my idle temps are ~32C ish with load temps at 63C ( must say i have ALOT more rad space than you so your temps seem in line with the rad space you have ( 6x120) ambient is ) 18C +/-
February 6, 2013 10:51:46 PM

Motopsychojdn said:
He's pushing case air out through the rad, I'll input more when I get to my Pc in the Am
Moto



While I am FAR FAR FAR from an expert if he was pulling cooler air from outside he might see the temp drop a bit more.
a b à CPUs
a b K Overclocking
February 6, 2013 11:06:05 PM

belezeebub said:
While I am FAR FAR FAR from an expert if he was pulling cooler air from outside he might see the temp drop a bit more.



You are correct. he would see a drop of 2-3C
a b à CPUs
a c 324 K Overclocking
February 7, 2013 1:02:39 AM

Stock CPU + GPU in that loop should be fine with a single RX240 with good airflow. But, you can never really have too much radiator space...
February 7, 2013 5:59:50 AM

Now I have been running some tests. Ran Prime95 for 10-15 minutes on each test. Listing the temps from each core.

Test 1
Closed case. 2 rad fans on top pushing air down.
56 53 55 52

Test 2
Closed case. 2 rad fans on top pushing air up
54 51 53 49

Test 3
Open case. 2 rad fans on top pushing air up. 2 case fans pushing air on CPU.
53 51 52 48

Test 4
Open case. 2 rad fans on top pushing air up. 2 rad fans on bottom pushing air up (push pull configuration).
52 49 50 47

I think that pushing air up instead of down, will be slightly better. Furthermore I believe that another rad on top of the case, a push pull configuration and some low RPM case fans would lower my CPU temps the most.

What do you conclude from my tests?
a c 113 à CPUs
a c 168 K Overclocking
February 7, 2013 7:49:46 AM

To me at least, it that you may have case airflow issues. Though your rad is intaking air, so it seems that there aren't enough exhaust fans to get rid of the heat being dumped into the case by the rad.
From the looks of it, your back 120mm is intaking air, might want to flip that around. Cant see them, but you also want any fans at the top exhausting. Could even try flipping the PSU, though I don't think it will do much.

See if that makes a difference.
a c 103 à CPUs
a c 190 K Overclocking
February 7, 2013 11:44:51 AM

You've been busy I see, good man
As you have seen, pushpull on the rad has given you best results whilst pulling cooler outside air into the case,
As mentioned that rearfan is 'backwards', you want that blowing out of the case, coupled with some intake fans at the front, If you check out the W/c gallery there are a few 690 builds that may give you inspiriration and yes, another rad with more fans is a good move imo
I'm a great advocate of having more radiatorspace than strictly necessary, its where the silence part of W/c really comes into play
Moto
a b à CPUs
a c 324 K Overclocking
February 7, 2013 12:58:37 PM

With the open case tests, did you also use the large fan for additional airflow? This would help determine if it's an airflow issue or not. It would seem by your numbers above that airflow is an issue if you are seeing 4C (~7 F) difference on load temps. That's a decent number just by opening the side of your case to allow more free-flowing air.

February 8, 2013 4:42:33 AM

When you say "the large fan", do you mean a particular fan? I only did two tests with an open case. One where my two yate loons pushed air at my CPU and another with a push pull configuration on my rad, with no fans cooling the CPU.

I do not think I will be making more tests for now. But i appreciate you replies. I will be upgrading, with fans and maybe another rad, when I need to redo the loop in 6 months. I can follow up on this thread if I detect significant changes in temperatures :) 

I am just relieved that there does not seem to be any issues with my tubing, blocks, pump or other components :) 
a b à CPUs
a c 324 K Overclocking
February 8, 2013 5:48:13 PM

Or this:



This will help eliminate if it's a case airflow issue or not by providing more airflow than you'd ever see in any computer case.
a c 103 à CPUs
a c 190 K Overclocking
February 9, 2013 6:57:57 AM

Hmm, got a link for those Rubix? I have an idea....
:p 
Moto
a b à CPUs
a c 324 K Overclocking
February 10, 2013 3:00:42 AM

It was just a normal Home Depot average box fan...found at any home improvement or DIY store.
a c 103 à CPUs
a c 190 K Overclocking
February 10, 2013 7:46:07 AM

Looks a nice size for a 1080 is all :p 
Moto
!