For the most part, everything is fine, but I noticed that one of my 990pro drives was running hot. I am assuming since it’s where drive 0 was indicated in the Windows install that it’s the drive on the MB which is on the same side as the CPU.
I thought I installed everything correctly and I think there wasn’t a peel on the Framework heatsink?
Any ideas? I have another Noctua fan coming that I am going to install this weekend, but ideas in the meantime would be helpful!
My concern is with the only intake fan blowing on the CPU, and the way the CPU fins are oriented it will blow some hot air off the cpu heat sink right onto the SSD. And if there is high disk activity but low CPU, the fan won’t run at all.
If it is the second case, then try setting the fan controls in the firmware so that the CPU fan comes on (low) at 35, so that it is always running a bit. That way even if there isn’t a lot of processor activity, you will still get air blowing across the ssd. And for my first concern regarding if there is high ssd and cpu activity, the cpu fan should be blowing fast enough that the air temp going across the ssd may not be too much.
Those temps are kind of insane. If you look up the published materials from Samsung, the drive is only rated to operate up to 70C. It should throttle and start shutting itself down when getting much hotter than that.
I’d install the biggest heat sink that fits within the case and like others said, install an 80mm fan as intake to get more continuous airflow into the case.
There was an option in the BIOS to turn on the fan all the time at low power. The DRAM temps are now the same more or less between the two drives, but the boot drive - the one next to the CPU still has a controller temp that is almost 20C higher than the one in the back…
@SE_M I forgot to say, enabling ASPM (i.e. PCIe power saving) will also make a huge difference. I don’t know whether Framework’s BIOS has that option though.
I got the noctua 80mm and installed it - I set the fans to be alway on in the BIOS and now the temps are under control and really, the system is no louder than before.
This is after 12 minutes of stressing the system with AIDA64 AND Crystal Disk Mark
OK, I guess I’m just not sure what each of the temperatures for your C/E drive are, since it’s reporting 3 temperatures for that one physical drive. If you use Crystal Disk Info, which temperature it shows corresponds to the HWINFO64 info shown above?
I’ve been super focused on “Temperature 3” but if the controller (which is generally what Crystal Disk Info reports) is actually “Temperature 1” then that’s the one that shows a peak of 64C in your testing. That’s under the 70C critical temp and thus would be fine I guess?
