First time water loop complete! Help with Temps please!

[reptexas]

Hello All !

I have built a few gaming rigs but this is my first custom water cooling loop. After a few days of installing everything I am concerned with the temps I am seeing. They are not by any means "bad" but I expected them to be much better. Maybe my expectations were too high so I am hopping someone with a equal amount of hardware can share their temps or tips.

The rig: http://i42.tinypic.com/9q9v6q.jpg

CPU: FX - 8350 (STOCK No over clock)
Motherboard: Gigabyte fx-ud7
Memory: Kingston hyper x blu 12gb
GPU: 2 x 7970 in crossfire
PSU: InWin comander 1200w
Case: Thermaltake Mk-1 Chaser
3x1080p 23" monitors @ 5760x1080

Water cooling Loop:
Res: Swiftech Maelstrom
Pump: MCP35X Pump Connected via PWM set @ 4500rpm
Rad #1: 400mm Phobya with 2x cooler master 200mm fans pulling out top
Rad:#2: XSPC 120mm rad with 1x 120mm fan pulling out back
Extra fans: 200mm fan puling in from front of case, 120mmfan pulling up from case (case is elevated)
CPU block : XSPC Raystorm
GPU Blocks: XSPC New Design Raystorms
Tubing: 1/2 in OD 3/8 ID UV blu
Liquid: SIlver coil and distilled water.
Layout: Single loop - RES>Pump>400mm>CPU>120mm>GPU1>GPU2(in series via XSPC bridge)>RES

No Air. Bleed for 24hours moving case around everything runs whisper quite.

Temps: Ambient room temp 26 c Hot in Texas right now

Water Loop temps (via res):
Idle: 29.5 c
Prime (20 min): 35.0 c
Furmark (20 min): 38.0 c
Battlefield 3 multi player (20 min): 41.0

HWMonitor:
Idle: CPU 19 c / GPU 40 c (Probably needs 10c offset)
Prime: CPU 38c
Furrmark: GPU 51c
BAttleifled 3: cpu 45 c - GPU 57 c

Coretemp (w/ +10c offset added)
Idle: 30 c
Prime: 48 c
Battlefield 3: 55c !!!!

MSI after Burninger:
Idle: GPU1 40 c / GPU2 38 c
Furrmark: GPU1 49 c / GPU2 51
Battlefield 3: GPU1 57 c / GPU2 59

The conclusion I have came to with the little experience and tons of research I have come across has led me to believe two completely different things.

1: Some people with fx 8350 and 2x 7970s have way better temps even overclocked which leads me to believe something is wrong.

Or

2: Some people I see with a Fx850 and 2x7970s have almost the same temps as me which leads me to believe maybe the cards plus chips in the same loop mean that while they are all being stressed I will have higher temps even with 520mm worth of rad space.

Thanks in advance to anyone who can help me out, basically I don't know if i should add more rads, fans, re-mount heatsinks etc.. if it's not going to get much better than this.. i'd hate to rip everything apart (again) and make things worst than they are. I'd like to OC only my gpu's in the future but it seems that a $800 WC system should give me that freedom, or am I just expecting more than science will allow.

 

[Spiffy_1]

I'm sure our resident watercooling guru will reply right away but how is the static pressure on those 200mm fans? And I don't mean cfm. The bigger the fan the slower it turns lowering the static pressure. That size you almost need push pull to increase static pressure without increasing fan speeds. So do your research on your fans and see what your static pressure is and see if there are better fans in that size. More rads will lower your temps to a certain degree at which point you run into the this isn't cost effective anymore. Better fans with higher static pressure will also improve your temps to a certain extent.. Don't be throwing on Delta fans unless you like your ears bleeding.

 

[Conumdrum]

Not bad temps. All below max temps and seems okay by me. About what I would expect from your current raddage and high room temps. Actually my room is 26C or 79F, very nice. It's not your room temps causing an issue.

When you run your load tests are you running Prime 95 small ffts AND Heaven benchmark at max settings?


Both at the same time? For 30 minutes to let the loop stabilize? Try that and get back to me. After the test record room temps too. Tell me temps again after you do.

Your giving good data, just need to verify.

 

[Conumdrum]

200 mm times two fans on a 400mm rad means not much CFM making it through the rad if any.

 

[reptexas]

You Guys are FREAKIN BRILLIANT!

*SOLVED*
I had a spare 200mm fan and a 120mm basic case fan just sittin around so i Squezeed each under the 400mm rad just to test out the fan theory. Loaded up battlefield 3 again as that is the highest temps i was getting. After 20 minutes of gaming the results landed at,

CPU: 45 C
GPU1: 47 c
GPU2: 51 c
water cooling temp: 35 c

with an ambient temp of 26 c

The idle temps stayed exactly the same but the load temps all fell by 10 c so the two 200mm on the 400mm rad just aren't enough. I will look into what fan combo i could use on this rad and place an order to give it an even push/pull but i can run it like this in the meanwhile.

Thanks guys ! so glad I joined up as you saved me from dumping the water for a 5th time and re-mounting the blocks for a 3rd time.

 

[Conumdrum]

Wait.

YOU STILL HAVEN"T RUN THE TEST I NEED TO SEE.

So goodbye, good luck, no help to the future folks that ask questions like this because you wouldn't run a test.

'Click'....................

 

[S_I_N]

wow ur giving us Texans a bad name by this type of behavior. Should always follow up on whats offered and not **** off those who are here to help.

 

[reptexas]

wow ur giving us Texans a bad name by this type of behavior. Should always follow up on whats offered and not **** off those who are here to help.

Woah SIN slow down there bro.

I just stepped away to watch my San Antonio Spurs in Game two of the finals. Jeesh.

I said thank you to the advice and i'm pretty sure that it was correct. I didn't even think about running the test for other folks who have the same types of questions.

My bad Conundrum, I will run the tests after the game here and post the numbers.

Again, thanks for helping it is greatly appreciated!

 

[Conumdrum]

Go Spurs! Is basketball played in Amerika? I know golf and football is. Jus' kidding!

Go Bronco's!

 

[reptexas]

Yes, Go American Professional Basketball Sports team the Spurs from San Antonio!!

Okay during half time i ran Prime using small ffts settings while running Furrmark (dont have 3dmark) at full 5706x1080 aax2 settings for 20 minutes.

At 20 minutes the results are as follows:
Coretemp: 56c (w/10c+ for offeset)
MsiAfterburner: GPU1 54c / GPU2 57c
Water loop temp per res therm: 40.1c
Ambient air was at 26c but is now at 27c after the test LOL

This is with adding the extra 200mm and 120mm fan to below the 400mm rad to help push some air up through. Better temps then before for sure, but now i have no idea what to go with?

For the 400mm rad should i step away from 200mm fans all together, or would it be safe to add 4 120mm typhoon under the rad in push and keep the 2 200mm cooler master fans on top in pull? Or should i keep 200mm fans and just add two below? Hmm.

 

[Spiffy_1]

Well you know you need push pull on that 200mm rad. Test your 200mm fans and compare them against 120mm to see what works for temps.. See where that leads you. If you're going to go with the typhoons you might want to look into a shroud if you have the space so you use all of your rad.

 

[Conumdrum]

Your Dt (which is talked about in the stickies) is pretty high. 13 diff between air temps and water temps. But you were under xtreme load, not usually that high. P95 and Furmark type concurrent loads aren't normal.

I'd say your fine. While gaming you'll be under 10C Dt which is fine.

 

[JeremyCT]

Okay during half time i ran Prime using small ffts settings while running Furrmark (dont have 3dmark) at full 5706x1080 aax2 settings for 20 minutes.

He asked you to run the Heaven benchmark, it's a better indicator than Furmark. It's a little less demanding than Furmark, but more representative of the load an actual graphically intensive game would put on the graphics system, and nVidia/AMD don't automatically downclock their cards when they sense it running so it's a better stress test for the modern stuff.

It's free, just Google it. Easy to find, install, and use.

Be sure to come back and update this when you find and install any new fans. We like people who provide good data on changes and the results they get from those changes.

 

[Conumdrum]

Unengine has a new one called Valley. It's free to DL. Bit updated from heaven, but Heaven is a fine realistic load proggy. Furmark burn in mode is a bit xtreme but is a good stress tester.

 

[reptexas]

More awesome info guys thanks, I will download Valley and Heaven.

I always thought Furrmark was the norm to test systems but it makes sense to have a more realistic every day scenario to test a water loop by.

I put in an order for two additional coolermaster 200mm fans since they are onsale for $8 on newegg and I would like to start the testing out on the more silent options.

I will post the results when they come in and hopefully that is the missing link.

 

[dealmaster]

I don't know for certain, but I've read in several places that Heaven is more GPU intensive and Valley is more CPU intensive. Can anyone confirm or deny?

 

[JeremyCT]

Can anyone confirm or deny?

Quick testing on my rig. *not strictly scientific, may or may not be broadly applicable*

1080p, Ultra, 0xAA, Windowed, Normal tessellation (Heaven, Valley doesn't have tessellation settings)

Both tests had the same GPU temps, but Valley had marginally higher CPU temps (2 deg C). Total power consumption was also 5-10 watts higher with Valley.

Personally, I like Heaven better. Valley has a lot of texture pop-in and draw distance problems that make me not really like it. Since the load is similar (not exactly the same, but very close), I'll continue to use Heaven.

Plus, y'know, dragon ... it's hard to argue against dragon content.

 

[GTXJackBauer]

I've read in several places that Heaven is more GPU intensive and Valley is more CPU intensive. Can anyone confirm or deny?

I have heard this as well.

I put in an order for two additional coolermaster 200mm fans since they are onsale for $8 on newegg and I would like to start the testing out on the more silent options.

I am not sure how silent your systems going to be since they are 200mm and they need to spin up to build and keep that pressure. Its easier for smaller fans to sustain pressure than it is for bigger fans at a low RPM.

Also after reading all this info you do realize the rule of thumb per chip is 120.2 (avg.) of rad space. I did some math and you seem to have just over 170mm of rad space per chip. If you buy some really good fans as I am not familiar with the 400 mm rad and or 200mm fans, you should see a real improvement.

(Heaven, Valley doesn't have tessellation settings)

I swear I thought they do.



To the OP again. Testing the rig under load with Prime95+Furmark at the same time gives you an indication of your MAX in DeltaTemps, Water Temp, CPU and GPU temps, etc, all under the H20 system. You can see how much room you have to play with (OC). Heaven benchmarks are more realistic in real world situations. Games will be moving at a higher eye candy pace soon since the new consoles will be out later this year. Things are about to change very fast.

Nice setup btw. I hope the 780 will handle my 5760x1080 as well. My 480s SC are keeping up just fine in high settings in BF3.

 

[JeremyCT]

I stated it poorly. Heaven has tessellation options, I set it to "Normal". Valley has no specific tessellation settings on the setup screen, but I know it uses tessellation in its rendering. Free/Basic versions in both cases, perhaps the pay versions have more settings.

 

[reptexas]

Also after reading all this info you do realize the rule of thumb per chip is 120.2 (avg.) of rad space. I did some math and you seem to have just over 170mm of rad space per chip. If you buy some really good fans as I am not familiar with the 400 mm rad and or 200mm fans, you should see a real improvement.

To the OP again. Testing the rig under load with Prime95+Furmark at the same time gives you an indication of your MAX in DeltaTemps, Water Temp, CPU and GPU temps, etc, all under the H20 system. You can see how much room you have to play with (OC). Heaven benchmarks are more realistic in real world situations. Games will be moving at a higher eye candy pace soon since the new consoles will be out later this year. Things are about to change very fast.

Nice setup btw. I hope the 780 will handle my 5760x1080 as well. My 480s SC are keeping up just fine in high settings in BF3.[/QUOTE]

Thanks!

Well, the 200 mm fans came in so i set them up and ran some test. Basically, the temps were worst with 4x 200mm fans in push pull, then with 2x200mm on top and one 120mm on the bottom when I was testing the fan theory.

Guess i will put in an order for 6x 120mm XSPC 1600rpm fans and some fan splitters. I'm interested to see the temps vs noise between the two.

With the 4x200mm CPU gets to 56c and GPUs to 52 when under load with Prime + Heaven. The cooling loop shot up to 40c.

Thanks again everyone for the input and I will be sure to post the numbers next week sometime with 2x200mm on top and 6x120mm fans on the bottom of a 400mm rad. hopefully it doesn't sound like a bee hive in here.

BTW: ave FPS with Bf3 maxed @ 2aa, 50 fps @ 5760x1080 with the 2x7970s .