PC shuts down and then displays a black screen.
PC shuts down and then displays a black screen.
I have encountered an issue where the PC freezes abruptly and then displays a black screen, even though it remains powered on. The CPU diagnostic light remains red, and the system does not stop unless a hard reboot occurs.
System details:
- Model: MOBO: MSI B450 Gaming Plus.
- CPU: AMD Ryzen 5 2600.
- GPU: MSI RX 580 Armor OC 8GB.
- RAM: Corsair Vengeance LPX 2666MHz 16GB (2x8) @ A-XMP.
- SSD: Samsung 970 EVO 500GB.
- HDD#1: Seagate Barracuda 5400RPM 1TB.
- HDD#2: Seagate Barracuda 5400RPM 4TB.
- PSU#1: Corsair CV750 80+ Bronze.
- PSU#2: Cooler Master MWE 230V 450 80+ White.
- CASE: Armageddon Kagami K-5 (front panel removed).
- FAN: Arctic P12 x5 (3 front intakes, 2 exhausts).
During disassembly, I accidentally altered the power settings multiple times due to concerns about video output not functioning. After reassembly, the PC operated normally until a sudden freeze and black screen occurred even during idle browsing.
My actions at that time included:
- Monitoring temperatures; they stayed below 80°C, with shutdowns only above 100°C.
- Re-seating and benchmarking CPU with Cinebench.
- Re-seating and testing GPU with Furmark.
- Reseating RAM twice, including testing on another system.
- Reinstalling Windows 10.
- Performing CHKDSK, DISM, and SFC checks.
All connections were secure; the cable was firmly connected.
Swapping the Corsair PSU for a Cooler Master resolved the problem (as expected). Later, I installed LMDE 6, which functioned well for several months until I accidentally damaged my boot and EFI. After a live setup, moving files to a laptop via SSFTP initially worked, but after a day of changes and an update in the live environment, the cursor froze and the screen went black again.
I then tested CPU and GPU performance using Sysbench and Furmark for 5-20 minutes. The results were recorded here. Based on the data, it seems the issue might not be related to the SSD, as the GNOME disk reported healthy S.M.A.R.T metrics.
I attempted to install Mint but encountered problems; the installer stopped working entirely. It had been functional previously, possibly due to a bug. I switched to OpenSUSE Tumbleweed, which worked temporarily for a couple of days, but the problem reappeared during internet browsing, requiring a hard reboot.
My planned steps:
- Update BIOS (last version 2022).
- Purchase a Prolink Line Inverter UPS (simulated sinewave).
- If unsuccessful, consider a Fractal ION+ 2 560W 80+ Platinum.
- As a backup, a B450M Mortar Max.
My questions:
- Is it typical to experience two PSU failures in such a short time? The first might have been accidental, but the second?
- Should I replace the PSU immediately rather than invest in a UPS?
- How can I confirm if MOBO is the root cause?
- Can I rule out unstable power supply as the problem source? My laptop remains operational without batteries.
- Is it safe to use the MSI B450 Gaming Plus MAX BIOS with my non-MAX model? The MOBO BIOS has been a beta version for the past two years.
Welcome to the forums, newcomer!
The PC remains active despite a red CPU diagnostic light, and it stops only when a hard reboot is performed. This might suggest the CPU is failing, or the memory controller is defective, or that your motherboard's BIOS has become corrupted.
PSU#2: Cooler Master MWE 230V 450 80+ White.
This isn't a particularly dependable unit, and it's not suitable as a power supply for this situation. From a wattage perspective, you were satisfied with the 750W model. Opting for a lower-wattage PSU as a temporary fix doesn't eliminate any potential causes of your problem.
Tracking the temperatures is generally okay—mostly within the normal range at around 80°C. Thermal shutdown happens only when reaching 100°C.
You didn't specify the case model or the cooler you're using to manage the processor heat.
Reinstalling Windows 10
It seems you obtained the OS installer from where? Did you recreate it to check for corruption? If this doesn't resolve the issue, consider upgrading to a new Fractal ION+ 2 560W 80+ Platinum unit.
Since your RX580 processor is in use, a stable 650W PSU would be more appropriate for the entire system.
Determining if MOBO is the problem? Try replacing all your components with a known working motherboard that matches the correct socket and BIOS version. If the issue disappears, the problem likely lies with your motherboard or its BIOS.
Assessing whether unstable electricity is the cause: I have a battery-free laptop that functions normally.
It would be wise to consult a certified electrician and inspect your wiring. Also, using a UPS could help during unexpected power outages. A well-built PSU will improve power quality from the wall.
Can I use MSI B450 Gaming Plus MAX BIOS for my MSI B450 Gaming Plus model that isn't MAX? Given that the BIOS for my MOBO is a "beta version" over the past two years, always follow the BIOS recommended for your specific motherboard, even if it shares the same chipset.
Can a failing CPU handle a full workload? It seems unlikely, but a stress test could trigger the faulty component right away.
True, the Cooler Master PSU I purchased back in 2017/2018 is the original one. It lasted about eight years before this issue occurred.
Yes, sorry. The model is Armageddon Kagami K5 with an AMD Wraith Stock cooler. It has five Arctic 120mm fans, three front intakes, one back and one bottom exhaust, plus Cooler Master Mastergel thermal paste.
Came directly from Microsoft’s site. I’m unsure if reinstalling the original installer fixes it, but a fresh setup worked the same. I’ve confirmed the ISO.
I’ll consider this. For my UPS of choice, any advice or notes? I can’t afford a full Pure Sinewave Online UPS since it’s nearly as expensive as the whole system.
Even if I have to go with the beta version?
PSU#1: Corsair CX750 80+ Bronze (not the green CX).
That PSU was actually the original one. It lasted eight years without any issues, until now.
The Corsair model you referred to, CX750, has a five-year warranty. You used it for three years past that period. A power supply will naturally degrade over time, and this degradation speeds up when it faces constant high loads or high room temperatures.
It's a lesser-known brand named Armageddon Kagami K5
View: https://youtu.be/H2Tc8q9ocb8
To be honest, with minimal ventilation for the fans to bring in cooler air, this unit becomes a hot box—ready to damage internal components. All your internal cooling efforts are pointless if you can't supply them with cooler ambient air from outside.
But the 450W model remains reliable for eight years, wouldn’t you think it would be okay with 560W?
Wait. I’m now confused. Which PSU is eight years old? Even if it means adopting the beta version?
If the newest version is still in beta while I’m seeing instability in the older one, I try to flash it to the beta in hope it fixes the problem. That’s what I do—I don’t recommend others follow the same approach, as they might not know how to spot the issue if it arises.
I realize now what I meant. The Cooler Master 230V MWE 80+ White was the one I bought in 2017/2018 with the RX 580. It wasn't until late 2023 that I upgraded to the Corsair CV750 80+ Bronze (I mixed up the CV with the CX). You just reminded me I still have warranty coverage. Still wondering if the CV series are worth it. I should have mentioned this earlier. I usually take off the front panel or sometimes the side panel. It's fair that I could downgrade since there were no issues with its original BIOS. If it doesn't work, I'll keep going through the beta.
It wasn't until late 2023 that I finally upgraded to Corsair CV750 80+ Bronze (I mixed up the CV with the CX).
The CV was a rebadged VS unit (the one with the yellow label) which is actually a step down from the CX unit—this indicates poor quality PSU. CX's are only acceptable if they're just below that threshold; anything else is a waste of money.
Should've mentioned this earlier. I do take the front panel off, or sometimes the side panel.
This defeats the purpose of having a case.
Fair enough, I could downgrade since there was never an issue with its original BIOS. And if it doesn't work then continue progressively through the beta.
Please don't downgrade the BIOS as there are instances where the process ended up making the board an expensive paperweight. Either move up or leave the BIOS alone.
There seems to be a BIOS Flashback+ button on the back I/O panel. Is it still considered risky?
Isn't that what you're suggesting? I'm not sure.
Unfortunately, I can't help with that.
The PSU came in just yesterday, so I reassembled everything carefully.
That night it started up but only recognized 8GB of RAM out of the 2x8 CMK16GX4M2Z2666C16 model.
I then executed the command to check memory details.
The BIOS confirmed two RAM modules were installed at 2666MHz.
This morning I powered the PC back on and it booted into Mint, though performance was slow.
After a brief pause, I turned off the system, unplugged the PC, and reinserted the RAM in the same slots (DIMMA2 and DIMMB2).
Now the issue persists—DRAM red debug light appears, and I can't access BIOS.