B0G3N

Hi again, folks! This time, I bring interesting news. At least to me... As said, I was feeling a bit puzzled by the fact that my GPU was being so warm. After the feedback I have gotten here, I decided to do a few things: - Drain the loop, move the drain valve to the lowest point in the loop (next to the pump). - Re-apply thermal paste to both GPU core and CPU. Was out of paste, so picked up some Noctua NT-H1 paste from a close-by store. - Remove the standard Phanteks rear (intake) fan and install a Noctua NF-A14 200 0RPM fan (that I had lying around from another PC) as exhaust. - Fill the loop and redo testing Turns out, when using the overclock I have on my MSI R9 390 Gaming card (1150MHz/1575MHz, +100 Aux voltage and +100 Power limit), the GPU core stays as low as 55oC when all rad fans are running at max and the rear exhaust is running at 1000 RPM. The water got to 43oC, and stayed there during the whole testing period of one hour using Valley benchmark. I do not know whether it was the previous thermal compound not being correctly applied or if it was the change in type of compound that did the trick (possibly both?). Anyways, a major thanks to suggestions in this thread. That said, the loop still gets warm, just not as warm as before.. Anyways, here is the current build: So, next question is; Do I go for hard tubing next? And in that case, would you then keep the pump/res combo or exchange with a separate pump and reservoir? And where do I in that case put the res? Cheers!

Yes, my cooling setup is having an EK SE 240 in the front as intake, using two F3 Vardar (1850 rpm) fans in push. The top rad is an EK PE 240 rad with two corsair SP120 (quiet ed, 1450 rpm) fans in push as exhaust. The rear fan was set as intake to get some positive pressure, but I can turn it around and see how that works out. I am using a bitspower temperature sensor connected to the temperature probe input on my motherboard. It gives a temperature readout in the motherboard software (AI suite 3), and seems pretty on point. Good to know, thanks! I also have suspected the Hawaii core of my GPU to be working properly, just being a very warm GPU in general. However, to see if I had done anything wrong, I spent yesterday evening cleaning out my loop, re-applying thermal compound to both CPU and GPU, and I will have a go at doing some testing today or tomorrow. And as Necrodead says, I do not get these crazy temps when gaming; the GPU core usually stays around 65-70C when gaming, so that is a much more comfortable temperature range. I was just pushing the system to see if I still would be within the comfort zone in a "worst case" scenario, which I think I am not... Hah. Will get back to you when I have done some more testing, hopefully this may help other R9 390 users as well.. On another note; While draining my system yesterday, I totally got to experience the bad placement of my drain port. The new loop will definately have a drain port at the lowest point, located next to the pump. Will also take some more photos on the rig, so you guys get a clearer picture of the system. Thanks for the help!

Ok, thank you for your reply. However, I am still questioning the relation between high coolant temperatures and poorly applied thermal paste. Does not poorly applied thermal paste on the GPU core imply that the core itself will overheat due to lack of heat transfer into the cooling medium (i.e. giving higher core temps and lower coolant temps)? I have had the thought that having a 250W+ GPU in the loop with the power limit turned up might just add too much heat into my loop. Cooling capacity of the two rads should (from my research) be well above the needed for a single GPU+CPU loop, but I might be wrong here. Have also thought about going for a water cooled GTX 1070, as I imagine the temperatures (and thus sound levels) will go down significantly when using a card with almost half the TDP.. Again, thank you for the help debugging my system. Anyone else here with experience with a water cooled MSI R9 390?

The idle temperatures with radiator fans running at max speed: GPU is doing 33 C and CPU is doing 30-35 C. Room temps are 21 C at the moment. Thank you for your help!

Ambient temps in the room are between 20-25oC.

Thank you both for your feedback, guys. I have suspected the temperatures of the water to be a bit on the warmer side - i can elaborate on the conditions: When benchmarking in Valley, using ultra settings and 8x AA in 1440p, the GPU reaches core temps of 76oC after 40 (ish) mins. At this time, the liquid coolant is reaching 54oC with both radiator fan pairs running at their respective maximum rpms (that is; 1850 rpm for the two EK Vardars and 1450 rpm for the Corsair SP120s). The rear fan is running at 1050 rpm to let in some fresh air and to keep case pressure positive. The pump is doing 4500 rpm, and making a loud whine/hum (famous for the Laing DDC pumps, I suppose). Ambient temps in the room are between 20-25oC. I am fairly sure that the loop is drained of air, and that the thermal paste is applied in a generous but correct fashion. I speculate that a lack of thermal paste might have shown up as high GPU core temps but low water temps? Or am I assuming this link between coolant temperature and GPU core heat transfer wrongly? Also, the radiator placement might have some impact on the cooling efficiency, no? From what I have seen, having a front radiator as intake and a top radiator should work out quite alright, or have I misunderstood? Regarding the drain; I know it looks silly, but I figured that was an easy access point in the case, especially considering that I might change the loop soonish to redo the tubing. It is in the middle of the loop when the case is oriented this way, but when tilted, it becomes the lowest point in the loop. Will definitely try to do a drain at the lowest point (where the pump-res-combo now is) the next time I need to drain the system! Again, thank you for the feedback, it is much appreciated!

Hi there! Some months ago, I decided to try to liquid cool my computer. So far, I have had loads of fun learning and researching different aspects of the liquid cooling systems, but there is so much out there that it all seems a bit daunting at first. I have a couple of questions, but first; the system. Case: Phanteks Enthoo Evolv mATX Mobo: Asus ROG Maximus VIII Gene CPU: Intel i5-6600K RAM: 2x8GB Crucial Ballistix Sport (2400 MHz) GPU: MSI R9 390 Gaming 8G PSU: Corsair RM750x Storage: 512GB Samsung 850 Evo + 1TB Seagate Laptop SSHD Pump+Res: EK-XRES 100 DDC MX 3.2 PWM Front Rad+fans: EK CoolStream SE 240 + 2x Vardar 1850RPM fans Top Rad+fans: EK CoolStream PE 240 + 2x Corsair SP120 Quiet Ed. fans Had some difficulties figuring out where to rout the tubing in this build, but am quite happy with the way it turned out. I am tempted to redo the tubing in hard acrylic or PETG, but I am a bit confused on the options and ranges of tube sizes available. From what I gather tube size is mostly a matter of visual preference? Are there any other things to look our for when choosing tube and fittings size? I am thinking of incorporating more white into the build, maybe using white fittings and some lighting. Also considdering using white pastel coolant. However, my GPU core temperatures seem to go a bit warm (the GPU core is touching on 80-82oC after benchmarking in Unigine Valley for 30+ mins). Are these temperatures normal? The cooling liquid is getting close to 60oC (upper 50-ies). I am aware that the R9 390 is notorious for high operating temps, so I am not sure if these temps are normal or not. Might turn this thread into a sort of build log if I decide to go for hard tubing. ^^ Comments and suggestions are much welcome - exciting to be joining such a knowledgable community!