WCooling 290DC2 OC

I finally managed to install water block on my Asus DCII 290.
Been putting it off for couple months since I wasn't sure if my card was defective.
I don't know what it was that caused instability but it has been rock solid for past month or so

EK-FC R9-290X DCII - Acetal+Nickel
With CL liquid PRO

Just a 15 minute run of prime and furmark but temps were stable.
After that 5 minutes of idling to record idle temps(current).
I closed hw monitor by accident while having the GPU in loop so that's why minimum temps are higher.
Ambient@30C



Kind of surprised me how much that GFX card heated up other components in the case and that it affected CPU temps that much.
Actually have lower CPU stress temp with GPU in the loop.

CPU's are more sensative than GPUs in general. Having the GPU dump on the CPU will increase the CPU's temps by a few degrees. Nothing extreme.

Is this a before on air and after on water bench run between the two images?

How much radiator do you have in your loop?

A full list of your loop would help a lot here.

ATMINSIDE said:

How much radiator do you have in your loop?

A full list of your loop would help a lot here.

Click to expand...

Signature

GTXJackBauer said:

Signature

Click to expand...

There are a lot of sizes of the ST30

It also doesn't tell airflow direction in the case, or the case.

ATMINSIDE said:

There are a lot of sizes of the ST30

It also doesn't tell airflow direction in the case, or the case.

Click to expand...

Copy that. A little more detail will help.

6x Ultra Kaze Fans I use to use Ultra Kaze x2 on a Prolimatech Megahalems when benching, lol their screamers

Running lower temps on your GPUs is really a thrill Congrads

I run the kazes around 1300rpm and only ramp them up while benching.
And the ST30 is 360mm radiator. I should add that to sig.

First image is when GPU is still aircooled and dumping its heat inside the case.
Second is after adding the GPU to the loop eliminating a lot of radiating heat and hot air from the case, best indicated by the motherboard sensor reading.
The ST30 radiator is top mounted inside the case exhausting from top and GTX360 is mounted outside in the back of the case.

GPU added in the loop actually resulted in lower water temp than running just CPU.

It does makes kind of sense but kind of does not.
GPU dumping heat in the case the air got so warm it was not effective cooling the water passing trough the ST30 radiator and maybe even was heating it up. I did have exhaust temp well over 40C
But still adding 200W of heat to the water I would expect higher deltaT.

And I am sure this is not ambient temp change since I ran the air cooled test in the morning and watercooled test past midday. Ambient was actually around 1C warmer when I ran watercooled test.

To me this was kind of interesting

E: Oh and I have Superflower SF-2000 case which I picked up for 50euros as temp case but it has actually proven decent.
Airflow is traditional side, bottom and front in and back and top out with the exception that I exhaust from behind socket area.

"6x Ultra Kaze Fans I use to use Ultra Kaze x2 on a Prolimatech Megahalems when benching, lol their screamers

Running lower temps on your GPUs is really a thrill Congrads "


6x is not enough....... 10x Delta 5krpm pwm BI GTX Radiators here. cool 2x 290x here. Hopefully I will find the times to install the Ek water block next weekend.

Give the fans a bump up to ~1800RPM and run the same test as your second image, I bet that GTX has too high of an FPI for the fans to push air well at ~1300RPM.

May not even take that much, ~1500RPM might be enough.

Just do a quick test to see at what point your CPU temp doesn't creep up like it was

The ultra kazes are all on the GTX radiator in push/pull and they move plenty of air at 1300rpm but I was curious if raising speed makes now bigger difference since GPU is on the loop too so I ran same test @2000rpm
This time I took 3 screens, one at full load around 15 minute mark one less than minute after stopping stress and one after 5 minutes of idling. I left screens uncropped so you can see the temps that Aida gives.

15min mark

shortly after stress

5 minutes idling


Conclusion: No difference really. Ambient reads bit over 30 so we are at the same ambient range.

The GTX Radiators need around 1800-2200 RPM since they are a higher FPI rad. I think they are 20 FPI but not sure. Martin did some tests and if I recall correctly he said that 2200 RPM was the magic number for them.

The ST30 is a lower FPI rad and can have the fans run around 1300 RPM no problem.

As for the tests, you need to run them for 30mins min if you want to get real time temps. A few mins to 15 won't really cut it. You need the loop to reach an equilibrium when on full load.

I reach max temps at 9 minute mark and for the 6 remaining minutes they occasionally drop down 1C and go back up again.
I agree that running an hour or two I might see few degree of difference but most of that would be because the ambient room temp rises.

I could run it for hours and hours but I am sure I would just waste electricity heating up our living room :O

GTXJackBauer said:

The GTX Radiators need around 1800-2200 RPM since they are a higher FPI rad. I think they are 20 FPI but not sure. Martin did some tests and if I recall correctly he said that 2200 RPM was the magic number for them.

The ST30 is a lower FPI rad and can have the fans run around 1300 RPM no problem.

As for the tests, you need to run them for 30mins min if you want to get real time temps. A few mins to 15 won't really cut it. You need the loop to reach an equilibrium when on full load.

Click to expand...

Worth noting: they're a 54mm thick 20fpi.
Definitely need serious fannage for that.

ATMINSIDE said:

Worth noting: they're a 54mm thick 20fpi.
Definitely need serious fannage for that.

Click to expand...

+1

Why don't you let ultra Kaze rip it at 3000rpm and prime it for 30 minutes see what result you get. GTX radiator works best at 3Krpm and above to benefit the maximum effect.

BruceUSA said:

Why don't you let ultra Kaze rip it at 3000rpm and prime it for 30 minutes see what result you get.

Click to expand...

Because it's 3 a clock in the morning and my GF is sleeping
I have tested running them full speed before but at stock CPU speed the difference is too small to justify a degree or two of temp difference in expense of quality of life(noise)

When I bench [email protected] I need every little bit of cooling I can pull out and that's when I let the kazes rip.

I should run some OC tests but it's been so damn hot that I keep putting it off.