Question PC loses all power seemingly randomly to all peripherals (mouse, keyboard, monitor, etc.) during gaming or at idle

[Dexxed]

Hi!

I've been having some issues with my PC recently and have fixed most of them, however one is still stumping me, which I only just found a few days ago. During idle or at high-usage my PC will lose all power to peripherals( mouse, keyboard, monitor) but stay running. Unplugging and replugging peripherals back in did nothing.

Recently I had some unfixable sectors on my boot drive and took it as a change to upgrade my PC, so I don't know if it's an issue with faulty components or something else.

Parts List:

CPU - Ryzen 5 2600
Cooler - CRYORIG H7 49 CFM (new)
GPU - MSI rx 5700xt Mech
Motherboard - ROG Strix B450-f Gaming
RAM - Corsair Vengeance LPX 16 GB (2 x 8 GB) DDR4-3200 (old)
RAM - Corsair Vengeance LPX 16 GB (2 x 8 GB) DDR4-3200 (new)
PSU - EVGA Supernova 550 G2 80+ Gold
Storage -

Samsung 970 Evo 500 GB M.2 (new) SK hynix Gold P31 1 TB (new) Western Digital 1TB HDD Western Digital 260 GB HDD

I upgraded to two new storage devices, with the M.2 being the boot drive, added another 16 GB of ram (which are clocked at 2966 to meet CPU specifications) and switched out my stock cooler to a better one that could fit my case.

After checking the Event Viewer, before I had to force-shutdown my PC, I found these events under information (in chronological order):

All of these happened at the same time in Event Viewer. The only other system events before this were over 10 minutes before, and all after were startup events as well as Event 41, stating that my system had lost power due to my forcing a power-off.

Event 12- Kernel-General
-The operating system started at system time ‎2020‎-‎10‎-‎25T08:58:56.500000000Z.
Event 153- Kernal-Boot
-Virtualization-based security (policies: 0) is disabled.
Event 18- Kernel-Boot
-There are 0x1 boot options on this system.
Event 32- Kernel-Boot
-The bootmgr spent 0 ms waiting for user input.
Event 20- Kernel-Boot
-The last shutdown's success status was false. The last boot's success status was true.
Event 238, Kernel-Boot
-EFI time zone bias: 2047. Daylight flags: 0
Event 25, Kernel-Boot
-The boot menu policy was 0x1.
Event 27, Kernel-Boot
-The boot type was 0x0.
Event 30, Kernel-Boot
-The firmware reported boot metrics.
Event 20, Kernel-General
-The leap second configuration has been updated.
-Reason: Leap second data initialized from registry during boot
-Leap seconds enabled: true
-New leap second count: 0
-Old leap second count: 0
Event 16, HAL
-The iommu fault reporting has been initialized.

All of these happened late last night when I was playing VR Games at the instant my PC had this issue. Rough usage estimates at the time are CPU~50% GPU ~40%, VRAM ~84%, RAM ~40% which are quite stable most of the time.

Sudden shut off to VR Headset, mouse, keyboard, and monitor. It also happened a week ago when I wasn't in the room idling, so I don't think a particularly heavy load caused it.
One thing to note is I had previously been running 16GB of ram at 3200 by Overclocking .15V , because I didn't know that my CPU only supported up to 2966 (for some reason defaulting to 3000 in bios), when I added more ram and tried to OC it again, I got BSODs. After leaving the ram at 2966 I had no further BSOD issues.
I have done the Unigine Heaven stress-test on Extreme and had no issues, and have no issues playing titles like BeatSaber, CSGO, Pavlov VR, and others; except for the crash last night happened while playing VRChat.

I appreciate any advice you are willing to give and can run tests and give more logs if needed to diagnose this issue, sorry for the block of text.

 

[Vic 40]

The cpu should have no problems running ram at 3200mhz, but mixing ram can give issues. Would test both sets individually and see if that helps. Maybe even test overnight with Memtest86.

 

[Dexxed]

The cpu should have no problems running ram at 3200mhz, but mixing ram can give issues. Would test both sets individually and see if that helps. Maybe even test overnight with Memtest86.

The cpu is listed with it's highest speed being 2933, so I'm just running it there. Had BSOD errors when I tried to run it at 3200, and intermittent crashes at 3000.
I ran memtest86 overnight yesterday but didn't figure out how to set it for multiple passes and will do it again today.

Last night the same issue happened again, this time with the screen flashing green then same issue, no power to peripherals but with the computer still running, also in vr, but just in the steam vr home which is non-intensive.
I will update this thread after the next test, though I'm thinking it may be a PSU issue since mine is four years old at this point, hoping it's not the motherboard.

I also may have found something that causes similar issues-apparently running multiple hardware monitoring services (specifically temperature monitoring services) causes a similar issue on the ROG Strix B450-f, but with the screen typically freezing. However, the only hardware monitoring service I'm running most of the time is Radeon Settings.

 

[Vic 40]

Only would run one hardware monitor at a time, they can interfere.

The psu might be a cause, not the quality and it has a 7 year warranty if i remember correctly, but it might be that with that card the powerdraw is in such a way it gets into trouble. I do doubt it, but it might be possible. Looked again and since idling it also can happen do i suspect it's not the psu. Maybe more likely the motherboard or even that gpu. There are more threads here where the 5700XT series has been a problem.

 

[Dexxed]

Only would run one hardware monitor at a time, they can interfere.

The psu might be a cause, not the quality and it has a 7 year warranty if i remember correctly, but it might be that with that card the powerdraw is in such a way it gets into trouble. I do doubt it, but it might be possible. Looked again and since idling it also can happen do i suspect it's not the psu. Maybe more likely the motherboard or even that gpu. There are more threads here where the 5700XT series has been a problem.

Hasn't happened at idle in the past week, only happened once as well while I was out of the room, and my oculus may have been plugged in then. I've looked at the recommended wattage for PSUs regarding my setup, and it should take ~ 450W, and there are mixed opinions on whether someone should buy a 550W or a 600+W psu for a rx5700xt. I didn't have this problem with the gpu before installing my new parts, so I don't think its that, however there is a similar problem that does seem somewhat prevalent among rx5700xt and rx5700 owners after looking again. I have posted on the AMD forum with related information in case they have something there, since this seems to be an oddly common problem..

 

[Dexxed]

UPDATE: Ran OCCT's PSU, GPU, and CPU tests. PC only crashed on the PSU test, and I'm ordering a new (750W) one to see if it fixes the issue. Will post update when it arrives.

 

[Dexxed]

@Vic 40 I got the new PSU and installed it. Ran the OCCT PSU test for an hour with no issue, where as previous it would crash at 15-20 mins in. Sad thing is that, just a minute ago it crashed while semi-idle again. All I was doing at the time was listening to spotify and installing a steam game while browsing the web. Long green-screen (5-15 seconds) and then no power to peripherals again.

 

[Vic 40]

Hard to say what is the problem, but as said earlier my more likely suspects are motherboard or gpu. Since power to peripherals are lost might it well be motherboard.

 

[Dexxed]

Hard to say what is the problem, but as said earlier my more likely suspects are motherboard or gpu. Since power to peripherals are lost might it well be motherboard.

Thought I might've fixed it after updating the chipset and reinstalling the drivers with DDU, since I was able to play games and use VR for over 2.5 hours a few nights ago. A day after that though, it shut down doing some light blender work after playing two matches of cs:go. Was able to replicate it with 2 hours of light blender work today. Today I also ran aida64 and got errors when doing the stress cache or stress system memory stability tests. Took out my new ram and it ran without issue for 20 minutes, set them back in again and it ran without issue for the 15 minutes I tested it (Would crash within 2 minutes otherwise). Now it has random black screens, which appear similar to when restarting the video driver with windows + shift + ctrl + b. Every now and again, with seemingly random timings. Will time when it happens again.

I've also heard of power to peripherals being lost when it was a gpu issue.

Edit: just remembered I forgot to post some new errors I got here:


Error Source: Machine Check Exception
Error Type: Bus/Interconnect Error

Processor APIC ID: 0



Error Source: Machine Check Exception
Error Type: Cache Hierarchy Error
Processor APIC ID: 2



Error Source: Machine Check Exception
Error Type: Cache Hierarchy Error
Processor APIC ID: 13

These errors were given once, and the time it gave the errors it rebooted instead of needing a hard-reset.