Showing results for 
Search instead for 
Did you mean: 

PC Building


NVMe disappears on restart

So I recently moved on to Phase 2 of what I've now dubbed "Project Spectre." That meant transferring to a new, larger case and droping a 5800X3D into the spot where the 5600 used to live. I also swapped over to a dual stick 32GB RAM kit.


I updated the BIOS to the most recent one for my B550 A-Pro, and once I got sorted out with all my boot settings going wonky, my boot drive in NVMe slot 1 vanished from BIOS. So I powered off the PC with a plan  of re-seating the drive (it never moved during the upgrades) and figured... just hit the button. If it does it again, we have a problem. Well, problem went away.

Update the chipset drivers... SAME DARNED THING. So I cut power, restarted, got BSOD'd. Let Windows do its thing, and so far it's been smooth sailing. Any idea what it was? Bug in the BIOS? I made sure not to use a beta BIOS. Bad chipset driver? It's been stable since the restart off the BSOD.

XMP is on, but MSI won't let me touch PBO on the X3D, so I'm locked to 3.4GHz. I could muck with stuff in Ryzen Master, but the reality is, I don't really need the boost.

Performance over Pretty.
8 Replies
Volunteer Moderator

It could be the M.2 NVMe drive starting to fail.  You should at least back up the important stuff on that drive in case it fails shortly.  It might also be that M.2 slot on the motherboard that has problems.

As Albert Einstein said, "I could have done so much more with a Big Al's Computer!".

It better not be. Just over a year from purchase on both.

Performance over Pretty.
Community Manager

I'm not 100% sure, but the new BIOS could have required a reboot after the installation. That simple? 

Your biggest fan!
CPU: AMD Ryzen 7 5800X3D GPU: AMD Radeon RX 6800XT

It did. Then it did a POST loop into BIOS every time. Once I powered it off, let it sit for a bit, it was good. It's fine off a cold start, too. I'll try a couple of restarts before I call it a night and update.

Performance over Pretty.

As for the nvme, I would either check in another board if you have the chance or get a very inexpensive external nvme enclosure to test it via USB.

Does it shows the nvme inside BIOS?

If yes, can diskpart detect it?

Good luck 



The Englishman

It's been stable since yesterday. After the BIOS and driver updates, it vanished in BIOS, but after I powered off, let it sit for a minute or two and restarted, it was there. No issues since. I even did a chain of restarts before bed last night and it came back like nothing ever happened. Might have been a gremlin.

It just scared the bejeezus out of me because it was my boot drive. I didn't cheap out either, going with a KC3000. Worst case scenario if it starts going trippy on me again, I'll clone it to the 2TB WD SN850 I mounted up as an additional drive for games.

Performance over Pretty.

UPDATE: Stil causing grief. I'm debating ordering a new MoBo off Amazon since the B550 A-pro is super cheap right now. I have an M.2 USB enclosure on the way for future use, but next step is taking the KC3000 boot drive out of the Gen4 slot and moving it to the Gen3 slot. If it runs fine, either my Gen4 slot is toast, or something else might be afoot. Like the socket getting gunked up by thermal paste squeeze-out. I thought I gave it a very thorough cleaning with lots of IPA, but you never know, right?

Performance over Pretty.

Had a cable order come in today, so I figured I'd best just tear half the machine down and rearrange everything. Then I had an "OH 💩!" moment when one of my M.2 screws FELL INTO THE PSU. Yeah. RIGHT INTO THE POWER SUPPLY. How, you might ask? Well, my case has a perforated floor plate, and the fan is facing up because my EPS and 24 pin cables are just a teensy bit too short to reach properly. So... BLOOP! Into the PSU the screw falls. I tear down, pull the PSU out, and start shaking for all I'm worth. No idea if it came out or got wedged. So... I'm kind of at my "I'm screwed no matter what I decide" point. I make sure the caps are discharged, and I start peeling labels off the PSU to get at the screws. There was no way it would get RMA'd because some ham-fisted ape dropped a teeny little screw in there, and if I didn't make darned sure it was gone, I could face a very catastrophic failure. And at worst: I get zapped because the caps weren't fully drained and this post wouldn't be happening. So half an hour of fiddling, and half a can of compressed air later... nothing comes out. So I start checking the floor... and there's the screw, 3/4 buried in my carpet. So just to make 100% sure nothing was gonna go KABOOM on me, I dig out my old AM3+ mobo, plug the 24 pin in to it and spin it up. PSU worked as intended. I figured a couple hours of idle power would be about as good as 30 seconds of high draw if I was shorting somewhere due to an electrically conductive foreign object in there. Even if the OCP saves the rest of my rig, I'm out $250 on a new power supply because I'm clumsy.

Now for the meat and taters: the KC3000 booted immediately in the Gen3 slot once I got everything reassembled. I've got a funny feeling that the M.2 controller contacts in the CPU socket (not sure of the right wording there, but I hope everyone gets it) might be gunked up from some thermal paste that didn't get fully cleaned out. I opted to err on the side of caution and pull the trigger on the new MoBo. $130CAD with a Prime Day and "Welcome back to Prime" discount is a small price to pay for peace of mind. When I can afford a day of down time, I'll take the old board in and have the guys at MemEx test it before I send it off for RMA. Worst case, it's a PEBKAC again. Best case, they send me a new board and I sell it locally, or do another AM4 build.

I'm planning on doing some more thorough testing on my nephew's mobo when I build his rig next week. In the mean time, if anyone can recommend some good SSD testing software that's free, I'll be owing you a beer or something cold to drink.

Performance over Pretty.