[SOLVED] pc hard lock ups, looking for feedback on steps taken

braddollar

Distinguished
Jan 19, 2014
4
0
18,510
ive been having some pc trouble and i am looking for some affirmation and suggestions. i have an idea for what to replace next, but i dont want to influence opinion so i will refrain from saying. any help would greatly appreciated.

system specs:
  • ryzen 9 5900x
  • asus tuf x570 plus wifi
  • 4x corsair 8gb ddr4 3200mhz vengeance lpx
  • asus tuf rx6700xt 12gb oc
  • samsung 980 evo nvme 500gb
  • samsung 980 evo nvme 1tb
  • samsung 940 840 128gb
  • thermaltake toughpower gf3 1200w
  • bunch of fans
  • some other disconnected drives
  • windows 10 ltsc 2021 and ubuntu 25.04

roughly two weeks ago i started having intermittent hard lock ups. everything would freeze, displays would show whatever was last up, no visual problems on displays, mouse unresponsive, could not toggle caps lock. this always occurred while not actively being used, though the system was still doing crap, python scripts running (discord bot), discord, browsers, the usual type of thing. i could play a game just fine for hours with no problems (last epoch). i would have to hold down power button to restart. i could get many hours of use before a problem.

no stop errors. nothing in event viewer aside from system recovered from a crash.

i tried reinstalling windows, full format of os drive. it locked up once during install. it locked up again during the 'do not turn off computer' part, but after it still booted into windows. tried installing drivers, got some lockups during the process. did get drivers installed, but still lock ups after 1 to 5 minutes. at this point i disconnected all of my hdds, the 840 sata ssd, usb stuff, sound card, extra monitors. only nvme drives still in (because pain to remove). reseated gpu. still getting lock ups. ran samsung magician simple tests on both nvme drives and they passed.

reconnected the 840 sata ssd and tried installing windows on that. did not get any lock ups during install, but once i got to desktop it started locking up after 1 to 5 minutes.

tried to make to an ubuntu live usb, but would get lock ups writing to usb drive. made live usb on another computer. once booting into linux, still got lock ups after 1 to 5 minutes.

ran memtest86+ for 20 hours with no lock ups. no errors. 12 full runs, all passed.

the psu was 12 years old from a previous system, a 1050w seasonic, so i replaced that with the thermaltake gf3 1200w listed above. reconnected the same minimal hardware. still got lock ups on both nvme, sata ssd, and linux live usb. had usb drive not show up in boot once, so switched to a different port, same results.

i would wager somebody would say try with the nvme drives disconnected, but thats a pain to do so i have not yet, and because the problem occurs in the linux live usb it does not seem likely to be the drives.

like i said, i know what my experience says to replace now, but i wanted to solicit outside opinions for a sanity check. see if i missed something obvious. im terribly annoyed and frustrated, and so im worried im missing something.

so, what would you look at replacing next? thanks in advance!
 
Solution
can you give me any links to read about that? i searched around, but could find nothing. just inundated with links about failure rates of 5000s out of the box.

im plagued indecision at this point. totally blinkered by whats going on. cant find a known good cpu to swap. would you replace the cpu?

the very frustrated part of me wants to dumpster the whole thing, but affording something totally new is troublesome.

Nothing I can directly point to. Either posts here, troubleshooting videos, and other various places.

At this point, it would be worth a shot to at least test a replacement CPU. But you could also invest in a new platform, though to match a 5900X would be somewhat pricey.
samsung 980 evo nvme 500gb
samsung 980 evo nvme 1tb
samsung 940 840 128gb

With all these drives, did you disconenct all drives except for the one you wished to install the OS onto, deleting all partitions on said OS drive so you had no remnants from the older OS install? Did you recreate your bootable USB installer to rule out a corruption in your OS installer?

asus tuf x570 plus wifi
BIOS version for your motherboard?

4x corsair 8gb ddr4 3200mhz vengeance lpx
Did you try with 2 sticks of ram. I'm assuming you got two kits, try and work with one kit of ram. You can do this by identifying the PCB revision number on the side of the memory sticks.

i would wager somebody would say try with the nvme drives disconnected, but thats a pain to do so i have not yet,
You're the one living with the system, all we can do is advise on steps to take but you're advised to rule out everything in a step by step manner. Probable the issue was staring at you all along while you didn't notice it. I would try and drop the NVMe drive onto another known working platform(a friend's or neighbor's) and see if the issue persists.

Replacing isn't the course of action, as that will only mean you're paying out of your pocket.
 
With all these drives, did you disconenct all drives except for the one you wished to install the OS onto, deleting all partitions on said OS drive so you had no remnants from the older OS install? Did you recreate your bootable USB installer to rule out a corruption in your OS installer?
i removed all the drives but the two nvme ones for the first reinstall. the nvme are a pita to get at, have to take the gpu out, which means taking apart a piece of the case due to its siie so ive (perhaps foolishly) been avoiding that. my justification to myself for avoiding that is the linux live failing, as that should not be using the drives at all. though i guess they could still cause an issue.

i did delete the old partitions when i did the reinstalls, and i fully formatted the os drive.

the usb installer was freshly made just before.

asus tuf x570 plus wifi
BIOS version for your motherboard?

4021
the latest is 5021. my fear is that with the lock ups, it would lock up while trying to do a bios update.

4x corsair 8gb ddr4 3200mhz vengeance lpx
Did you try with 2 sticks of ram. I'm assuming you got two kits, try and work with one kit of ram. You can do this by identifying the PCB revision number on the side of the memory sticks.

i did not consider that. yes you are correct, it is two matched sets. i had not tried it since it passed the long memtest. ive now tried each set in each of the two ddr dimms, and it locked up all four times.

i would wager somebody would say try with the nvme drives disconnected, but thats a pain to do so i
You're the one living with the system, all we can do is advise on steps to take but you're advised to rule out everything in a step by step manner. Probable the issue was staring at you all along while you didn't notice it. I would try and drop the NVMe drive onto another known working platform(a friend's or neighbor's) and see if the issue persists.

sadly, no known good systems to work with or pull from. you are right of course.
Replacing isn't the course of action, as that will only mean you're paying out of your pocket.

unfortunately this is all slightly older than three years, so most things are out of warranty. aside from the ram, fans, and now the psu.

thank you for your time.
 
CPU re-seat? Not usually a thing with AM4, but worth a shot.

Otherwise it is either the motherboard or the CPU I would wager. Ryzen 3000 and 5000 are somewhat failure prone when it comes to the I/O die, they just fail. Sometimes you lose a memory channel, and it could be that is what you are seeing. Early signs of a memory channel dying. That you passed memtest may be an ideal condition, full load will keep the I/O and CPU dies at full blast, not bouncing around from idle to load conditions.

Buggy motherboard voltages could also be the problem and once in a while you get a voltage sag and the CPU/Memory lose stability.

Also, don't rule out silly things. I once troubleshot a computer for an hour only to find a bad SATA power cable. System would get into the BIOS, even start booting, but as soon as the drivers started loading crash.

Unplug all unnecessary fans too, never know when a little internal short inside of something can bring cause a voltage fault in the PSU and trigger a shut down. Though the symptom there would be the PSU failing to restart without being disconnected from power.
 
Otherwise it is either the motherboard or the CPU I would wager. Ryzen 3000 and 5000 are somewhat failure prone when it comes to the I/O die, they just fail. Sometimes you lose a memory channel, and it could be that is what you are seeing. Early signs of a memory channel dying. That you passed memtest may be an ideal condition, full load will keep the I/O and CPU dies at full blast, not bouncing around from idle to load conditions.
can you give me any links to read about that? i searched around, but could find nothing. just inundated with links about failure rates of 5000s out of the box.


so i stripped it down to cpu, ram, gpu, cpu fan, kb and mouse. still locked up.

replaced the mobo with an asus b550, ran with the same hardware in previous sentence, still locked up. managed to source an old r9 290 and swapped gpu, still locked up.

at this point i assumed it was cpu, but figured thered be no harm trying without cpu fan if it was being replaced anyways. ran for 9 hours, idling, that way. temps reached 92c, but it ran.

plugged cpu fan in again, locked up in minutes. unplugged cpu fan, plugged chassis fan into cpu 4pin, ran for an hour or so no lock up. swapped back to cpu fan, lock up in minutes. tried cpu fan on optcpu 4pin, same lock up. tried cpu fan in chassis 4pin, same lock up. unplugged cpu fan, and plugged in two fans on a y splitter to cpu 4pin, no lock up. tried a second set of 2 fans on y split, again ran fine. plugged in all chasis fans, ran fine for several hours. worth noting the cpu fan never stopped spinning when it was plugged in, no whining. its at about 28k hours and rated for 40k hours.

thinking this cpu fan was the source, i swapped back to x570 board. plugged all fans in, two m2 ssds, and swapped back to rx6700. did not plug in cpu fan, and it locked up after several hours. removed all fans again, locked up after an hour. removed one of the m2s, ran fine for five hours. plugged the m2 drive in again, ran fine for five hours.

now ive added back in one of the y split chassis fan sets, and waiting.

im plagued indecision at this point. totally blinkered by whats going on. cant find a known good cpu to swap. would you replace the cpu?

the very frustrated part of me wants to dumpster the whole thing, but affording something totally new is troublesome.
 
can you give me any links to read about that? i searched around, but could find nothing. just inundated with links about failure rates of 5000s out of the box.

im plagued indecision at this point. totally blinkered by whats going on. cant find a known good cpu to swap. would you replace the cpu?

the very frustrated part of me wants to dumpster the whole thing, but affording something totally new is troublesome.

Nothing I can directly point to. Either posts here, troubleshooting videos, and other various places.

At this point, it would be worth a shot to at least test a replacement CPU. But you could also invest in a new platform, though to match a 5900X would be somewhat pricey.
 
  • Like
Reactions: braddollar
Solution
At this point, it would be worth a shot to at least test a replacement CPU. But you could also invest in a new platform, though to match a 5900X would be somewhat pricey.
yeah, id had the same thought. i dont like the idea of spending on a replacement that is now over 4 years old. while its gonna hurt my pocket a lot i ordered a 9900x, x870, and ram.

thank you for your assistance.