I’ve been having issues with my m.2 drives disconnecting itself then reconnecting while loading games like Monster Hunter Wilds, Hell Divers 2.
Lately what i’ve been doing for the past year was refunding, Buying more m.2’s and each time it would have the same problem. I just bought another WD black ssd and kept getting the issue “disk write error” on steam. It would show a temp range of 40-55 temps on HWinfo but the drives themselves do feel hotter.
check list of things i’ve done is
-fresh install of windows
-Updated motherboard
-Updated GPU drivers
-set BIOS settings to Gen 4 instead of auto
-filled slots 1,2,4
-changed thermal pads for m.2
-turned the power setting for pcie to OFF
SSD’s I have are as listed
Predator M7 2tb x2
Orico 2tb
Samsung 980 2tb
WD black SN7100 1tb
Samsung 980 1tb
Check system event log. Run some SMART diagnostic program. I’m not sure which one works for Windows though. The drives keep a record of detected problems like weird controller shutdowns or when the controller had to stop because of heat.
You can for example show the metrics to an LLM and it might give you an idea of whats happening.
Remove your cooler, carefully remove your CPU and very closely inspect the CPU pads and especially the socket pins. If that all looks good then reinstall it all, make sure you have good cooler contact with good thermal compound application, it may be worth trying a CPU contact frame if you aren’t using one. If that doesn’t resolve the issues then I would guess it could be a bad motherboard or CPU.
I’m assuming you have a good power supply with every CPU & PCIe power port on the motherboard populated (including the one on the bottom of the board).
I tend to think so but I have been out of the game a long while. Not enough specific info. Kind of reminds me of some odd metallic flake issues I have run into in the past.
As agent9 suggested, but I’d add taking a close look at the m.2 and PCI ports ..
An intermittent fault, a bad contact, solder or slot for eg, could throw out errors across the whole PCI bridge n lanes.
After that, test all the drives individually, preferably in a different machine. If they all test fine, throw a single known good drive into each of the m.2 ports one at a time and test them .. then, finally, the sata.
I’ve done this with a known working drive from a different pc and it still acted the same way, I just found it hard to believe that everything else on the motherboard works but the nvme slots
pci-e lanes are controlled by the CPU, it is entirely possible for some of the lanes to have an issue while the rest of the cpu be fine, this could also be from some pins not making good contact with the CPU, reseating can sometimes help, but not really if the pins are bend or missing
there are some PCI-e lanes that are controlled through the chipset which could mean a problem with the PCH but also those lanes need to connect to the CPU in the as well
PCI is not very fault tolerant on desktop boards. An issue anywhere can throw problems everywhere. A fault in any single PCI device too, like a faulty drive .. it happens.
When you look at the socket, use your phone pix to zoom right in, from a couple angles. It doesn’t take much ..
If you look in hwmon, keep an eye on the board power consumption, see if there’s anything bouncing up n down..
M.2’s use 3.3v rail, your 3.3v rail gets as low as 2.6v, that stands out to me but I’m by no means a psu expert so I have no idea if thats too low, but it stands to reason a power draw spike with a substantial Vdroop sounds like the root of your issue
Check out windows event viewer and see if it leads you anywhere. It might give you an idea what general class of the things to look at. Device manager could also help narrow things if it is driver issues. Do a search on your mb and it’s bios version and see if there are any reoccurring issues that pop up for both.
