RAM cooling issue with C14S and RTX 2070 Super FE

chronomitch

Efficiency Noob
Original poster
New User
Mar 30, 2020
5
4
I recently put together an NCASE M1 v6.1 build using a Noctua C14S with the stock 140x25mm fan in low profile mode and an RTX 2070 Super FE. Aside from the fan in the SFX psu (Corsair SF600 platinum), there are no other cooling fans in the system. While I have no problems with CPU or GPU temperatures, I am finding that my RAM temperature is another issue. In some cases, my RAM is reaching 50 C, leading to stability problems.

Originally, I had the 140x25mm fan on the C14S as intake, pulling air from the side of the case through the heatsink and blowing it on the motherboard and RAM. This setup seems to give the best CPU temps, but resulted in higher RAM temps, which led to instability in Prime95 and several crashes while gaming. I reversed the fan to serve as exhaust, and this resulted in better RAM temps, but it still gets too hot for my tastes during gaming.

Here are some metrics for the RAM temperature.

  • CPU idle, GPU idle:
    • 30 C
  • C14S fan as intake, CPU loaded via Prime95 large FFTs, GPU idle:
    • 50 C
    • Prime95 failure in less than 10 minutes
  • C14S fan as intake, CPU & GPU loaded via Doom Eternal:
    • 51 C
    • game eventually crashes to desktop
  • C14S fan as exhaust, CPU loaded via Prime95 large FFTs, GPU idle:
    • 41 C
    • Prime95 stable
  • C14S fan as exhaust, CPU & GPU loaded via Doom Eternal:
    • 46 C
    • stable
  • C14S fan as exhaust, CPU loaded via Prime95 large FFTs, GPU loaded via Unigine Heaven loop:
    • 50 C
    • stable for 15 minutes, but needs more testing

I think the take away from this is that the RAM is cooler when the C14S fan is exhausting any hot air from inside the case, rather than blowing heated-up air onto the motherboard and RAM. However, once the GPU is loaded and starts dumping more hot air inside the case, the C14S's fan is not enough to exhaust all the hot air, leading to higher RAM temperatures and potential instability.

I'm wondering if anyone has run into a similar issue and how they fixed it. I am going to try adding a 92mm fan as exhaust to the back of the case. If that doesn't do the job, I will probably put an Arctic Accelero III on RTX 2070 Super and add two 120x25mm fans underneath it as exhaust.
 

MultiDoc

Airflow Optimizer
Feb 2, 2018
304
197
50C for RAM is not an issue at all. I’d suggest to investigate other issues that might be causing instability problems with your memory (Voltage ? Timings ?)
 
  • Like
Reactions: judokamak

chronomitch

Efficiency Noob
Original poster
New User
Mar 30, 2020
5
4
50C for RAM is not an issue at all. I’d suggest to investigate other issues that might be causing instability problems with your memory (Voltage ? Timings ?)
Maybe it's the chipset/pch then? I wasn't monitoring the temperature of that, unfortunately. I did notice that my ram voltage was a bit higher than it should have been according to the XMP profile. My motherboard was giving it about 1.38 V according to HWMonitor. I had to manually dial it down a bit. We'll see if it makes any difference.
 

tinyitx

Shrink Ray Wielder
Jan 25, 2018
2,279
2,338
What brand/model is your memory modules? I think not many modules have a built-in temp sensor? How is the temp measured?
Maybe you can confirm if this instability is temperature related. Just remove the side panel and run the stress tests with a desk fan blowing fresh air to the case. This should remove any hot air accumulation. If you still have instability, then it is unlikely caused by insufficient case ventilation.
 
  • Like
Reactions: rfarmer

chronomitch

Efficiency Noob
Original poster
New User
Mar 30, 2020
5
4
I performed some more investigation and found that my PCH was hitting 70 C when the CPU & GPU were loaded before experiencing a failure. I'm using a Gigabyte Z390 I Auros Pro Wifi, which has some kind of M.2 heatsink covering the regular PCH heatsink. I think the idea is that an M.2 SSD would be sandwiched between the PCH heatsink and the M.2 heatsink. I don't know how well that would actually work, but I am not using an M.2 SSD, so the M.2 heatsink was just covering the PCH heatsink loosly and preventing airflow. I removed the M.2 heatsink and did another round of testing. This time, my PCH peaked at 61 C and averaged around 57 C.

I still have a few gremlins to sort out, but my system seems considerably more stable now.
 
  • Like
Reactions: Odd_Fox

HyperActive

Cable-Tie Ninja
Jul 22, 2019
183
75
I performed some more investigation and found that my PCH was hitting 70 C when the CPU & GPU were loaded before experiencing a failure. I'm using a Gigabyte Z390 I Auros Pro Wifi, which has some kind of M.2 heatsink covering the regular PCH heatsink. I think the idea is that an M.2 SSD would be sandwiched between the PCH heatsink and the M.2 heatsink. I don't know how well that would actually work, but I am not using an M.2 SSD, so the M.2 heatsink was just covering the PCH heatsink loosly and preventing airflow. I removed the M.2 heatsink and did another round of testing. This time, my PCH peaked at 61 C and averaged around 57 C.

I still have a few gremlins to sort out, but my system seems considerably more stable now.
Strange, mobo chips usually can hit higher before throttle or crash I believe. But good you found out! Do you have the c14s on intake or exhaust? I found that intake gave me better cpu and gpu thermals whilst also giving 10 degrees drop in mobo chip temps.
 

MultiDoc

Airflow Optimizer
Feb 2, 2018
304
197
I still think that your stability issues are NOT temp related. Neither 50C for RAM, nor 70C for chipset are high temperatures. Technically yes the chipset being at 70C is not ideal, but it’s max temp before any instability occurs is nowhere near that.

So I’d still suggest you double check the bios settings for voltages, xmp, timings, etc and also maybe check all your power connections.
 

chronomitch

Efficiency Noob
Original poster
New User
Mar 30, 2020
5
4
Strange, mobo chips usually can hit higher before throttle or crash I believe. But good you found out! Do you have the c14s on intake or exhaust? I found that intake gave me better cpu and gpu thermals whilst also giving 10 degrees drop in mobo chip temps.
All my tests so far have shown that using the C14S fan as exhaust gives better thermals for everything except the CPU, at least when the GPU is loaded and starts dumping more heat into the case. At that point, it seems to be better overall to try to get all the heat out of the case.
 

chronomitch

Efficiency Noob
Original poster
New User
Mar 30, 2020
5
4
I still think that your stability issues are NOT temp related. Neither 50C for RAM, nor 70C for chipset are high temperatures. Technically yes the chipset being at 70C is not ideal, but it’s max temp before any instability occurs is nowhere near that.

So I’d still suggest you double check the bios settings for voltages, xmp, timings, etc and also maybe check all your power connections.
Thanks for suggesting this. I did more research and tweaking. While getting the PCH cooler probably helped somewhat, I think the underlying problem was that my bios was doing something funky with the VCCIO and VCCSA voltages. I was able to leave these two settings on auto when the same CPU and RAM were in my previous ASUS ATX motherboard, but I guess my new Gigabyte ITX motherboard handles it differently. They are now manually set to 1.2v. I was able to get through two hours of stress testing with RealBench, and all my temperatures were in a safe range. I will probably do a longer stress testing run tonight, but I think the system is in good shape now.
 

HyperActive

Cable-Tie Ninja
Jul 22, 2019
183
75
All my tests so far have shown that using the C14S fan as exhaust gives better thermals for everything except the CPU, at least when the GPU is loaded and starts dumping more heat into the case. At that point, it seems to be better overall to try to get all the heat out of the case.
Not from my tests on 20 min of realbench testing. Cpu was cooler by couple of degrees and gpu only by 1 degrees. Different hardware, different settings I guess.

But while cpu and gpu didn't change that much, mobo chipset temps dropped a lot!