I believe your PC might have an M.2 issue...
I believe your PC might have an M.2 issue...
Hey there, friends! I'm back sharing yet another unusual and obscure problem I haven't encountered before. A few months ago, just before Starfield was released, I upgraded my PC around 2019. I installed 32GB of Corsair RAM and a brand new 2TB M.2 drive. After roughly 70 hours of playing Starfield, I started experiencing random BSOD errors (refer to the thread I began there). It turned out the issue was linked to the SSD. I don’t know exactly what was wrong, but it seems someone—either here or on Reddit—helped me figure it out. It looked like the Windows Xbox (or GameBar, I’m not sure which) services were misusing it or expecting something specific from it. It was odd, but I’m still trying to understand.
I eventually swapped the SSD for a Samsung 970 EVO Plus, which still caused the same BSOD until I installed the Samsung NVMe drivers from their site. Once that was done, the problem disappeared—great! But then, after an update to cod: mw3, I began encountering a bizarre crash. Game elements would freeze, then my whole PC would shut down, forcing me to press the power button to restart. It was strange, limited to mw3, so I assumed it was an issue with that update.
A few people on Reddit mentioned similar experiences, but I couldn’t find much overlap. The crash would halt everything, making screen recordings nearly impossible since they would stop just before or around the crash. I managed to capture a recording with my phone (see the link at about 14:56). Fast forward to tonight—I was showering and trying to think through the problem. Maybe I could improvise a fix?
I recalled the earlier BSODs and wondered if this was similar. It seemed like random game elements might be failing to load, possibly because the SSD wasn’t sending data properly. I had an old 2.5-inch 500GB Samsung SATA SSD lying around. I installed it, removed the M.2 drive, reinstalled Windows, updated cod, Discord, Vivaldi, and other essentials. Suddenly, the PC stopped crashing during games—about 30 minutes of play without issues!
Now, I’m not sure if this was luck or a real fix. But I’m hopeful it’s resolved. My main question is: What could be causing this weird behavior? My PC currently has two unused 2TB M.2 drives and only a 500GB SSD, which isn’t ideal. What steps can I take to further troubleshoot? Any insights from others might help me figure this out.
I notice you're concerned about memory issues. Could you share more details about your symptoms and the device you're using? This will help provide a clearer understanding.
For all issues discussed, ensure your working SATA drive is removed from Samsung Magician (installed on NVMe Windows) and reports normal operation. Confirm there was any connection to the SATA 0 port while your NVMe was in the upper M.2 slot during the M.2 problems. Disconnect it from that port. Reinsert the NVMe M.2 drive into the upper slot and test again. If this doesn’t resolve it, skip steps 2 and 3 as per the guide to rule out known fixes. If none work, leave the drive connected and gather dump files for further analysis. The guide suggests checking multiple slots to avoid accidental disabling.
I began by checking their drive verification methods—three or four options confirmed everything was okay. I had a 2TB HDD in the SATA0 port and an 8TB HDD in the SATA1 slot. I installed them a while back and mainly use them for backups and cold storage. For verification, I’ll recheck the second drive but it remained stable no matter which M.2 slot I used. Initially, the TEAMGROUP SSD was in the top slot. After replacing it (which showed good performance via crystalDiskInfo), I swapped it to the bottom slot, but it didn’t help. When it was in the top slot, the SATA0 port was active. Tomorrow I’ll clear all other drives and test just the Samsung M.2 in the top slot, then share results. Right now, when my PC freezes or crashes, no dump files appear and there are no relevant logs in Event Viewer. I’ll follow your advice tomorrow and update you with what I find.
I shared a PC part picker link with complete specifications, available at the provided URL. The setup includes an i7-8086k processor with a Gigabyte Z370 AORUS GAMING WIFI motherboard, an RTX 3080 graphics card, 32GB Corsair Vengeance RGB Pro 3200mhz RAM (not XMP enabled), an rm750x power supply, and a mix of storage: two M.2 drives, a 500GB Samsung 850 EVO SSD, and two Seagate HDDs (2TB and 8TB).
The top M.2 slot uses the PCIe lanes with the SATA 0 port, which is why I suggested disconnecting anything connected to that port. This prevents any issues from NVMe drives in the top slot and SATA 0 being active. However, it shouldn<|pad|> to crash the system—only a SATA M.2 would impact the port’s bandwidth, and both drivers are NVMe M.2 devices. It’s still important to check if problems persist, as they share physical traces on the motherboard. Wait... what? The BSOD collection tool mentioned in the guide didn’t detect any dump files. Sorry, I’m surprised and will look into what to do next.
I didn't follow the BSOD guide you shared. After switching to this older 500GB hard drive, it hasn't crashed again, so using it now wouldn't be helpful. What I meant by no dump files being created was that there are never any MEMORY.DMP or minidumps in the Windows folder for me to check or analyze.
Previously handled tasks and occurrences:
- Disconnected all SATA drives
- Replaced the top M.2 Samsung NVMe SSD with a new one
- Reinstalled the Samsung Magician software
- Executed Samsung Magician with short S.M.A.R.T and full scan
- Result: Crash occurred around 9 minutes into the first game
- Moved the new SSD to the bottom M.2 slot
- Ran Samsung Magician again; crash happened ~2 minutes into the first game
- Clock displayed 10:42am on 23/12/23
- System froze at the time screen
- Used Synapse tool post-reboot; attached report .zip
- PerfMon/Report opened for over 10 minutes
- Resource Monitor stayed open despite freezing
- Windows folder showed no .DMP files or Minidumps
- Event Viewer indicated no events between 10:40 and 10:47, contradicting the frozen clock
- Screenshot attached for reference; unsure of the issue and uncertain on next steps
Have you seen what I mentioned earlier? Also, make sure to adhere to the housekeeping part of this guide and omit steps 2 and 3 just in case it’s a resolved problem from a previous update. When you execute the PowerShell script (step 4), update your original post, add a Spoiler, and include a Code block with Python settings. The Event Logs clearly highlight several problems, and it seems Windows isn’t generating any dumps. Please return once you’ve completed the House Keeping section.
Failing to generate dump files raises concerns about storage health (combining the M.2 into overall storage). Various factors might be responsible. Insufficient space prevents creation of dumps, which can reach sizes matching your RAM plus 3-4GB. I've encountered 67GB files on a system with 64GB RAM. These are later reduced to minidumps if enabled. The page file is too limited. If set to System Managed, it should grow automatically; otherwise, manual sizes require larger files. The page file holds all data. Drive encryption means the process runs outside Windows, avoiding encrypted drives. If none of these apply, is there a SATA drive connected (not M.2) you could relocate the page file to? Be aware, external drives aren't recommended as they may trigger unrelated crashes.