New Water Cooling System - Heat Issues

UrbanAvalon

Honorable
Jun 12, 2012
20
0
10,510
Hello,

I just installed a new water cooling system in my PC, but am having some trouble with my temperatures and was wondering if anyone could help.

I have 2 GTX 590s in my loop, and as soon as I start up my machine three GPUs read at 37 C and one reads at 68 C. From there the temperatures will keep climbing while the system remains idle, until eventually one of the Graphics cards shuts down.

I have been trying to bleed my system of all air for most of the day now, so I'm unsure if the issue is an air pocket in the heat sink, though that still could be a possibility. I haven't seen a major pocket of air in a while, though small bubble will still form in parts of my tubing, no matter how many times I clear them away.

My loop goes: Reservoir/Pump > Radiator > CPU > GPUs > Reservoir/Pump ect.

I've heard that the best set up has a 120 mm on the radiator for every cpu/graphics card you have in your loop, however my case will only fit a 240 radiator, unless you mod it extensively. Also while the one GPU climbs into the 80 & 90 C temperatures, I don't feel my radiator getting very warm to the touch.

I also still have all of the fans in my case along with the liquid cooling system, and the radiator has two 200mm fans above it.

I'm not sure where these cooling issues are coming from, any help or advice will be greatly appreciated!
 
Muscle up and rotate that case through all the angles you can,
its possible you have an air pocket in the block so on its side, front, back, other side, upside down, the whole gamut, get the air out of the loop, top the res to 100% and keep tilting until you are sure the air is gone
plus what Muffin said, more radspace is needed at a minimum, so check out the W/c gallery, scope some ideas and get modding :)
Moto
 
*sigh*

I should of figured, I was hoping that keeping all of the fans in would help disperse the heat quick enough... but I guess that was just wishful thinking...

Unless someone else has a magical way for my to keep the heat down, looks like it's time for me to invest in an external radiator stand, and research how to drain a water cooling loop...
 
even if it was the radiator, both gpus would heat up. at idle one 240 rad should cool the system, at load is where the issue will be.

this is an issue with a bad mount on the heatsink or the pump isn't pushing enough water through the system. what pump is it?

what cpu/gpu blocks also to help determine restriction on flow.
 
You may be able to put this in the top where the dual 200mm fans are.http://www.frozencpu.com/products/14803/ex-rad-329/Phobya_Xtreme_400mm_Radiator_Silverstone_Cooler_Master_Yate_Loon_Bitfenix.html?tl=g30c95s1586
And use the EX240 on the side where the dual 120 fans are
 

even if this issue is at idle?

he could be running 4x4 radiators and have the same problem. idle temps shouldn't be high, my fans run at 650 rpm at system idle, temps ~22c.

If the cpu isn't heating up, id suspect a bad heatsink mount on the gpu, flow should be ok if the cpu temps are fine.

 


That was meant to be for reference and there are other vendors with it in stock.
Magicool has a dual 180 rad that may fit.

The OP has a pair of Beast GPUs as far as heat is concerned and will need a large surface area to help disapate that.

Here are a few multi rad set-ups in the Phantom case

http://www.modsrigs.com/detail.aspx?BuildID=27017
http://www.modsrigs.com/detail.aspx?BuildID=29806
http://www.modsrigs.com/detail.aspx?BuildID=25743
http://www.modsrigs.com/detail.aspx?BuildID=27310
 
Update: Well I rolled up my sleeves, started spinning this thing like a baton and sure enough two a enormous bubble poked its ugly-unwanted head out of my rad. Once I got that out the best I could, re-filled my res and repeated.

Now my temperatures are not skyrocketing when I turn on my PC, but they are still slowly on the rise... or so I thought! Once I stopped ogling at my new set up and re attached the siding and fans, the temp stabilized and slowly started to fall.

Granted this is only at Idle, so I'm well aware more cooling will be needed, and have ordered an external stand and a new 480mm rad, which might be a bit overkill when included with my current 240mm, but why not.

Thanks for all the help and advice, hopefully one day I can pass these tips on someone else who is doing this for the first time.
 
Update: ***.

Gonna need some more help here, I've got my new 480 rad, hooked up 4 NZXT FZ-120mm's to it, but my GPU2 temp is STILL climbing up to 60+ C idle-ish. And I mean all I am doing is surfing the web and listening to music and it's climbing up there:

GPUTemp_zps9d5761a8.png


Here is a pic of my set up if it helps:

IMG_3420_zps3a31742a.jpg


IMG_3425_zps5658ad71.jpg


The other three GPUS hover around 31-33, only GP 2 climbs. I'm starting to think something is up with the thermal paste, but after going through draining this beast, then re-bleeding it again... I don't want that to be true...

BONUS QUESTION: Anyone know what could be causing this? The card works fine, been getting enough power too, but the bottom one looks a little... discolored... :

IMG_3424_zpsead16add.jpg

 
Do you know what sensor that is thats reporting the 60'c?
I think your memory may be the cause of it as 1728MHz is way high for the idled clockspeed, so thats kicking out some heat right there,
I'd have thought mem speed throttled down with the Gpu myself but I'm no expert on Gpu tweaking 😛
Moto
 
Sounds like the block isn't seated well where that GPU is. Can you pull the card and confirm thermal paste spreading?

Took me a minute to figure out where you were getting 4 GPUs...had to go back and see they were 590's... 2x GPUs each. :)
 
@Moto, That's how the card are by default, the "classified" evga cards a pre-over clocked. i tried lowering everything as far as it would let me, but no luck cooling it off.

@ rubix, that's going to require me to drain my entire loop and dis-assemble some if it huh... well that will have to be a project for tomorrow, I'll be sure to keep you all posted.
 
Ripping it all apart is a pita, but unless you are 768% certain you have a good spread on the chip, its the only way to find out, and while its apart you can make any small adjustments to the loop, shorten/reroute tubing as required for instance,
vacuum the rads of fluff, wonder where the hell your killcoil has got to,
that kind of thing 😛
Moto
 
Already drained the loop (pain in the *beep*) and am cleaning the old thermal paste off right now, looks like you all were right, as far as I can see the thermal paste didn't have a good enough connection with the water block... it also looks like the corner of the other GPU isn't connected either, should I be worried about that? I don't have much left over thermal paste, unsure if I can do both of them.

IMG_3427_zps05c442b9.jpg
 
If the paste is new, you can reuse it in a pinch as long as you have good coverage. Can you determine via test-fit where the issue is occurring? It's possible there are minor differences in how the card's components contact the block or that the screws are not creating enough tension.

When attaching and tightening screws, are you doing so in a criss-cross pattern to evenly apply pressure as you tighten?
 
Preliminary results: It's Fixed! right now my temps are 32 - 29 - 28 -31 and maintaining around there. Haven't done extensive tests yet, will follow up tomorrow after work.