High temps under water?

sdamaged99

New member
OK specs below

4960x @ 4.6Ghz (1.35v)
Asus Rampage IV Extreme (BIOS 4901)
360mm Alphacool Monsta Radiator
480mm EK Water Blocks EK-CoolStream PE Radiator
Scythe Gentle Typhoon AP-15s on all rads
Corsair 900D
Radeon 295X2 Full XSPC Waterblock
EK 4.0 Pump
250mm Tube Res
EK Supreme HF Waterblock (it was a faff to get this mounted on my board but i think it is mounted correctly)

Running Asus RealBench my CPU is hitting 75 degrees, that seems much higher than i expected for what i am assuming is a decent watercooling setup

I've read online other people at similar voltages getting temps in the mid 50s, and thats running Prime95 etc

I think the pump is adequate althought the flow into the tube res is more like a trickle (but i don't know what to expect to be honest as never had a tube res)

Could the pump be the issue? Perhaps not shifting the water quick enough?

Rads get warm, but nothing more than that

On the GPU side, the temps never go above 53 degrees, no matter what i run (Heaven, Crysis 3, Sniper Elite 3, 3dMark etc) which would suggest the issue was more with the CPU and not the loop?

Have i done anything wrong?

I've noticed that i may have some trapped air in my loop. When i turn the PC off, the tube from the processor to the reservoir appears to fill with air (as in there is no liquid)

If i turn the bleed valve on my top rad, i hear a small sss of air releasing. However, if i turn the system on, i can bleed again, and then turn off and bleed again - rinse and repeat

I'm not convinced this is doing anything though as the reservoir goes back to being half full, which i worry may be trapped air

I reset CPU to stock frequency and then ran Prime95 and furmark together for 30 mins to see what kind of temps my watercooling loop could handle, and CPU was hitting 78 degrees (stock @4Ghz) and both GPU cores were hitting 55 degrees

Radiators felt very warm, and the 295X2 was so hot you couldn't touch it
 
The radiator surface area should be more than enough, what voltage are the fans running at?

Considering the quality of the rest of the components, the EK-DCP 4.0 pump seems a bit low end, and I'd recommend investing in a D5 vario when you have the chance. I doubt this would cause your temps to suffer greatly though, flow rate isn't that big a factor.

Have you only just put the loop together? The temperatures will naturally come down a bit the longer you've had the loop running because it gradually gets rid of more air. If the res fills with a lot of air when you turn it off, you should bleed it more and fill it more. My loop is fairly complex with three blocks and two radiators, and I spend at least half an hour bleeding it, turning it all sorts of angles and shaking it about gently.

I know it's a massive hassle to alter a loop, but are you confident that you've done a good job installing the waterblocks, and have you used a decent quality thermal paste?

Edit: For some kind of comparison, my i5 25000k @4.5GHz (can't remember the voltage, sorry) gets to about 70°C and my pair of GTX580s get to about 50°C under load, that's with 60mm thick 240mm and 360mm radiators, and Corsair SP120s in push/pull at 5V.
 
Last edited:
Yea that pump won't be helping mate, they are ideally only for single block loops. I have one to cool my G3258, and there is a flow meter connected after the rad, just before the res, and the flow rate is really slow. Even before the flow meter was added, I was using the top port of the res as an inlet, and it was barely a trickle.

It might be a case of so many blocks being pumped through that, the restriction is causing the water to be very slow in moving round, not allowing the heat to be transferred adequately.

Firstly though, I'd re-seat everything again, and if that doesn't work, I'd change the pump mate.
 
I have to admit, the water flowing (trickling!) into the res did make me wonder if that was the problem

Does flow rate massively affect temperatures though? Just want to make sure i don't fork out another £100 ish for no reason!

The 295X2 block seems on great, and thermal paste was IC diamond on the GPU and MX-4 on the CPU
 
Last edited:
I'd be leaning toward a possible mount issue on the CPU block too, just because you mentioned it was a pain. :) That sounds a bit toasty on the CPU to me, and the 295x2 is a heat monster (from what I read, no experience with it), and the GPU temps aren't bad at all. A D5 is always a good idea, but not sure if it would solve your problem. Sure would help the flow though.

Also, what speed are you running the fans at? What's your ambient room temp like?
 
I'm not 100%, but I would imagine if it is major slow it won't be taking the heat away very quick. So in turn not moving the fluid through the radiator, quick enough to be cooled. At least that's how I'm thinking any way mate, I'm sure if I'm way of the mark, a pro will be along to correct me. I'm only saying this as I no the pump and with just one block, and a 240 mm thin rad it was piss poor, so I can only imagine what your flow is like.
 
Thanks for the replies guys

Fans are set to half way, using Nesteq fan max (but when stress testing i put to full speed) and the Gentle Typhoons are decent fans aren't they. PC is in my conservatory so quite cool now

I might try remounting the waterblock on the CPU, but i have mounted it around 4 times now (had a couple of faulty boards) and the temps were always much the same (even on my 3930K) so i wonder if the EK mounting system simply isn't very good

I've just taken a quick video of my flow rate into my reservoir, it's a few seconds long (14MB) but should give you a good indication of how good / bad it is

I would have a hard time persuading my wife to let me now splash out on both a new pump and a CPU block, so ideally need to prove one or the other!

https://www.dropbox.com/s/0kbv7rnsoec5fip/2014-09-27%2020.38.50.mp4?dl=0
 
Those temperatures sound very close to what I get if I rage my CPU volts past 1.3v and have my D5 vario set pretty slow. I have noticed that the pump speed effects temps just as significantly as fan speed but after a point increasing them both is not beneficial. There is definitely a point when the pump is too slow and it holds the fans back. When I have got the cables to make my aquaero's work properly I am going to graph fan speed vs. temps at varying pump speeds to find an optimal pump setting.

JR
 
Try just the cpu block on it's own with the rads, see if that helps, at least you will no if it's the mounting or something else. If the temps drop at least you'll no it's mounted OK. It might be a lot of faffing around removing this and that, but it will be worth the time it takes.

I'd drain the loop plug up the gpu, and unplug pcie cables, then just re route the loop bypassing the gpu.
 
Picture quality isn't wonderful as the conservatory is quite dark and i don't have any LED lights in the case yet

Basically it goes like this :-


Reservoir>Pump>480mm Rad>GPU>360mm Rad>CPU>Reservoir

Also, if you're wondering why the PCI-E cables come from the top hole, i've pulled them tight and it stops the 295X2 sagging, as it weighs almost 2 kilos, and i think looks crap when it sags

Couldn't think of any other way to do it!

Pump and 480 rad at the bottom of the case of course

2014-09-27225834_zpsf4669723.jpg
[/URL][/IMG]
 
OK, i've just taken the CPU block off, cleaned, re timmed and put back on. Its definitely on right

Ran RealTemp and temps are exactly the same. I guess this proved the CPU block isn't at fault

The only other thing could be the pump speed
 
I've found i have a lot of trapped air inside the loop, the GPU is being cooled fine as the block is less restrictive, but the CPU is being affected

I'm going to change the intel on the reservoir to the bottom, as i think the water going in the top is creating air

I'll let you know how i get on!
 
that gpu is going to be introducing a lot of heat into the loop which is the exasperated by the cpu putting out a lot of heat due to volts I think that could be your main problem. Ive had air trapped in my loop rocking the case back and forth so it bumps about a bit normally helps clear the majority of it the rest normally clears over time but Ive never seen a change in temps before and after clearing the air from my loops.
 
OK Laing D5 fitted and tubing redone. Spent nearly an hour moving case around to get trapped air out, and i am confident all air is now gone

Checked temps within windows at stock idle and cpu was around 22 degrees

Under OCCT linpack hitting 68 degrees at 1.24v (4Ghz)

Still seems a tad high for me, i will test at 4.5Ghz later 1.35v and see what the temps go up to

Are these just very hot chips? My 3930k was much cooler

GPU temps have dropped under load to 49 degrees though, so the trapped air was making a good 5 degrees

difference

Look at the temps this guy is getting and this is on air with 1.30 volts?

http://www.overclockers.com/forums/showthread.php/749698-Man-why-such-high-temps-on-a-4960X
 
Last edited:
Well i changed my waterblock from a Supreme HF to a Supremacy EVO and its dropped my temps by nearly 15 degrees

Either other waterblock is faulty or it just wouldn't mount properly

Good news though!
 
as others have said the diference in temps for +0.05v vs the extra speed you can use stable may be something you want to contemplate.
But if you have it down to 60c at 1.35v now i would imagine it is fine. however for me if it cant stay at or preferably under 70c running IBT for at least 4 max tests in a row then the voltage/speed is to high. and i prefer it did that with the fans staying at 20-50% but if they do have to get to 100% to keep it at 70c i would call that a pass too. (i would just try to add more rad space after)
 
Or the EVO is far less restrictive and it's improved your flow rate perhaps? Sounds like a result either way.

JR
 
Back
Top