This problem has been popping up for months, but until last night only popped up every few weeks to months. Basically, the PC will blackscreen, turn off peripherals (keyboards backlights go out) and just sort of "idle", still running but not actually doing or displaying anything. Initially I thought my GPU was dying, I ran a Vega 56 before, bought it used and had it for 3 years, so not unreasonable. But now it's also happening with the RX 7600. It happened 3 times last night, and I was able to replicate it once this morning (all four times while playing Snowrunner. Nothing happened yesterday while playing Red Dead Online for hours before that). Event Viewer shows nothing for the time of the crash. I've been trying to replicate it again with GPU-Z open on the other monitor, but no "luck" so far. There's not even any readings that could indicate where the problem is. Temperature for both CPU and GPU remains in the 50s (around 70 for GPU hotspot), fans function normally, board power draw doesn't go above 160W, nothing spikes randomly or anything of the sort and the new GPU should also be using significantly less power than the old one. I've also checked that the GPU is plugged in properly and drivers are updated.
The only things I can think of are either faulty RAM, a faulty motherboard or a faulty PSU. The PSU is one of the oldest parts of the PC, probably coming up on 10 years now, so if any part is failing I would expect it to be that. However, I still wanted to explore other possibilities before spending 100 bucks or more on a new, high quality PSU.
EDIT: Would also like to add that I've only found two threads on reddit that describe a problem that matches mine closely (there's a ton of different black screen crashes, so a bit hard to find), but neither had a solution.
The solution was to disable "Core Performance Boost" in the BIOS, but I'm not sure if that's even an option for a 2700X (I only owned a 3700X, 5800X and 5800X3D so far). The guy also had a B450 board.
Besides that.. the classic steps I'd take:
Update your BIOS to the newest version (And set XMP to on again)
Run DDU and reinstall the newest GPU drivers (I owned a 5700 XT before switching to a 3080 and had plenty of issues with AMD drivers at times)
Run in a terminal as admin: sfc /scannow, if Windows finds any errors and fixes them, run the command again (maybe after a restart) till no more errors are found
Run a scan on your disks if they are faulty, you can also do CPU, GPU and RAM testing. Though if you did steps 1 to 3 it might already be fixed
If nothing helps then it might just be your PSU, but that would be weird to be honest. The PSU either can take the load or it can't, what happens if you run stress tests? Furmark + a CPU heavy test at the same time, does the PSU crap out?
GPU drivers have been clean installed not even two weeks ago.
I could try doing that, the main problem is that I have no reliable way of replicating the crash, even when closely recreating the scenario of the previous crash. Even crashed under different loads, I think it might've even crashed without a heavy load, just using a browser, in the past, but I'm not exactly sure about that. There's also some differences between that thread and my issue, like mine persisting even with a different GPU. Also neither the CPU or GPU were under a particularly heavy load (70% or so?) nor showing high temps. So even if the fix would work, I really have no way of knowing, could just happen again next week or in two months.
I have to admit, I've never updated my BIOS, and with the fear that it might be the PSU, it sounds kinda risky, but I'm not really qualified to assess that.
If it is the PSU, my guess would be that there's some very random power spikes it can't handle and that causes the shutdown. Otherwise it would be weird, yes, since the new GPU has a lower power draw. I've had a failing PSU before and that just shut the PC down completely when it couldn't handle the draw.
It might seem a bit silly, but I am going to try a new PSU first before I stress test it again. If it is a power issue and my PSU is on the way out, I'd rather not mess with it. I'm in a financial situation now where I can stomach potentially wasting 150€ for a new PSU I may or may not need, but not a thousand to replace my PC, you know?
Anyway, thank you, still. If it happens again with the new PSU, I'll go through these steps. And if it still happens after that... I don't know, I'll probably have a little cry.
Definitely try a BIOS update first, Ryzen did a massive amount of bug fixes and optimizations with each version. After the BIOS update also make sure to get the current chipset drivers from AMD!
But be very careful when selecting your BIOS update, there might be later versions where they removed support for older Ryzen CPUs (as space is limited). So check the notes if they say anything like "removed support for Ryzen 2000 series", but I think B450 should be safe for your CPU.
If your PSU is broken your computer would go 100% dark (or straight up restart).
From what you've described, my suspicion would be the power supply. It sounds like something is happening where it doesn't completely shut down, but stops delivering sufficient power to the system. The fact that it's a black screen and not a blue screen points toward it not being a ram problem, and unlikely to be your motherboard.
If you have a multimeter, you could check the voltages on some of your cables during normal operation and compare them to the readings you get the next time this happens.
Yeah, I guess I should just buy new PSU at this point. Worst case scenario, I have a new PSU. I've still been unable to replicate it again, so it seems pretty random and not purely dependend on load.
maybe some driver crashes somwere or an kernel level anticheat ? google tells me windows is a monolithic kernel wich means if a driver of anything on kernel level crashes it takes the whole system down with it. sorry i have absolutely no idea how you would check that on windows because i don't use it. the only way to check that i can think of reinstalling or trying another os alltogether
I would expect a bluescreen in that case, I think. Probably something in event viewer, too. I'm pretty sure it's hardware related, but thank you still.