A few days ago I asked everyone here if turning the fans from exhaust to intake would work better and everyone uniformly responded with "Intake!"
https://www.overclock.net/forum/61-w...l#post28419716
Well, I did my own testing and found that both rads as intake is 4C hotter GPU core, 5C hotter on the GDDR6 and 6C hotter on VRM / MOSFET within only 20 minutes (stability not reached).
I recorded the testing.
YouTube (Short URL)
Testing methodology:
I performed two 20 minute runs back to back starting with 25C on the GPU and 66-67F ambient (21C and fan speed @ 40% RPM, unchanged, for both runs, I induce full load on the GPU (in this case 95%, not sure if it's a CPU bottleneck or RTX but this game doesn't go up to 99% with 8700k @ 5.0 GHz) and I run around the Caspian Sea in Metro Exodus all settings maxed (RT: High) @ 3440x1440. Temps start out good, around 40C but slowly ramp up over time. I alt+tab around the 10 minute mark both runs and pull up Hwinfo64 (still recording) to check the temps on what's not shown in RTSS (memory, hard-drives, Sabrent Rocket 2TB M.2 PCI-E that is mounted directly to them motherboard sitting below CPU socket, above GPU).
The only thing that changes between the runs is that on the 2nd run I revert all of the fans (6 on the PE 360, 3 on the SE 420) to push-pull OUT of the case / exhaust and the rear 140mm fan reverted to intake.
2nd session has the same ambient and same starting temp on GPU.
Data:
All rad fans as intake, rear 140mm fan as exhaust
10 min mark:
Sabrent Rocket 2TB: 52-55C
2x16 G.Skill Trident Z Royal Memory: 51 and 52C
CPU VRM MOS: 43C
GPU: 46C
20 min mark:
Above mostly unchanged but now GPU core is 49C, one of the banks of memory is at 52C and GPU VRM / MOSFET is at 49C!
All rad fans as exhaust, rear 140mm fan as intake
10 min mark:
Sabrent Rocket 2TB: 30C!
2x16 G.Skill Trident Z Royal Memory: 35 and 41C!
CPU VRM MOS: 36C
GPU: 43C
20 min mark:
Above mostly unchanged, GPU is now 46C on core, 47 on that same bank of MEM1 that was 52C, and 43C on GPU VRM MOSFET!
Other observations, back-plate gets really hot without that rear 140mm fan as intake! That combined with the fact that hot air is blowing directly over distribution plate all of the 14/10 acrylic tubing, motherboard, GPU back-plate etc. All of that heats up with the radiator heat, it's basic common sense (and DDC pump, no-one even bothered to look at my system before immediately suggesting all rad fans as intake, if they saw that the DDC mounted to the distro plate was literally 1.5 inches in front of PE 360 that would otherwise be pulling air across it and out of the case as exhaust would no be dumping heat directly onto it would they still recommend all rad fans as intake considering DDC is air-cooled. Unexepected pump failure in what 1-2 years with that heat? When your pump goes and takes your $1200 GPU out with it, that's always fun. Luckily I did my own testing and found that rads as intake is definitely not better than exhaust despite what conventional wisdom dictates AND I'm replacing single DDC with 2x D5 in serial, just waiting on my larger rads to come in)
If you don't want to watch me run around the Caspian Sea in Metro Exodus I alt+tab out in the all fans as intake testing session at the 20 min mark in the video and I also go over the data at the end and before starting second testing session with fans reverted to intake @ 36 min mark and you can see the ending temps @ the end of the second 20 minute run around 1:02:30 mark.
Needless to say I'm glad I just didn't listen to all of the sage advice here and run my rads as intake! I've been doing this for a while and I've always questioned why you would want to run your radiators as intake, pumping all of that HOT air into your case and over your components (and hard tubing + liquid), the very same components youre trying to cool with said radiators. I understand the concept, the effectiveness of a radiator depends on the temp of the air passing through it, but you can provide ample air to said radiators from inside the case. It may be a few degrees celsius warmer, but at least in a View 71 (and Air 540) with 40% fan RPM it's definitely better than pumping all of that hot air into your case. I've always run my rads as exhaust and until today I've never done back to back testing like this to test my hypothesis, to me it's just common sense.
Well now I know for a FACT that multiple radiators are much better exhausting that heat out of a case, at least if it's a View 71 and youre only running the fans at 40% RPM (inaudible, I don't like noise, hence why I'm replacing the DDC on the distro plate with 2x D5 which I will run at 70% RPM).
As I state in the video, maybe rads as intake is better if you run your rad fans at 70-100% RPM and the case doesn't breathe as well as Thermaltake View 71. I don't mean to offend anyone with my findings nor the way that I present them here, each case is different, I'm sure many of you have done your own testing and may have found that the opposite is true. My question is, did you actually look at the temps of your memory and your mobo mounted M.2 PCI-E drives? +10-15C on the memory and +20C on the M.2 PCI-E boot drive is MASSIVE!
Anyhow, thanks for reading and watching, hopefully this experience and data will help others to buck conventional wisdom and do their own research when it comes to radiator fan orientation.
System in question:
YouTube (Short URL)
https://imgur.com/a/i7zCusE
