Question MSI B550 Gaming Edge Wifi - Help

[Bjg13190]

PC Build: https://pcpartpicker.com/list/qkmbnL
---
TL;DR: Purchased and ran above system fine for 3-4 months before getting random crashes (not overheating related). Temporarily resolved by increasing CPU core voltage, but now system will not even boot to BIOS screen. Details below.
---
The Long, Torturous story of this PC build:

I built this PC back in July (2021), and it ran like a dream for about 3-4 months before the issues began. Gamed every day, did college courses, watched videos, multitasked every day, no issues.
Some time in October the PC started giving me an issue where it would reboot itself, randomly, while gaming OR performing routine tasks such as opening a browser window.

The error was a WHEA log error, similar to the following:

"A fatal hardware error has occurred.
Reported by component: Processor Core Error Source: Machine Check Exception Error Type: Cache Hierarchy Error"

I tried the following to diagnose/resolve the issue:
-Ensured no overheating of GPU/CPU was occurring
-run a check on both installed drives (m.2 nvmE, and HDD) (both fine)
-flash to current BIOS (no change)
-Clean install Windows 10 (no change)
-every combination of RAM to rule them out (no change)
-uninstall and reinstall latest GPU driver (no change)
-increase CPU Core Voltage from BIOS (Oh?)

Increasing the CPU core voltage to 1.3V seemed to have resolve my issue for quite some time. Then 1 day about a month later as my wife was playing The Sims 4, the issue reared its ugly head once again. Okay, no problem, I'll just give a slight bump to CPU Core Voltage again to see if it stabilizes, and of course, it did.
This seeming constant creep in needing to increase CPU core voltage seemed (naturally) very odd to me, but the system was running right? So I was fat, dumb, and happy once again.

Flash forward to yesterday, I was enjoying some Valheim (most of the day, probably 8+ hours) when around 11PM the crash happened once again. This time the crash continuously happened about 3-4 times until I manually shut the PC down. When I boot the PC back up it failed to even get to the BIOS screen.

Now, at power on the PC turns on. All RGB turns on, GPU turns on, fans turn on, but other than that. NOTHING!

Sometimes when it starts up the monitor will turn on, I can see the white "_" cursor for a few seconds and then the screen stays on, nothing else happens, but the screen recognizes it is connected to a PC. The other times, I will power it on and I get nothing from the monitor. Again, everything is on inside the case. Lights, GPU fans POST run, case fans are running, RAM RGB is lit up, but no one's home.

Of the MOBO error lights, well, it depends on the power on which light comes on. Sometimes, the CPU light comes on and stays lit. Sometimes the VGA light comes on and stays lit.
Only thing I've tried is flashing the BIOS using thumb drive and the flash button. It seemed to be successful as it ran for about 6 minutes and then the PC reboot, but right back to the same issue.

---

Not sure where to go from here. Obviously a hardware issue, and I'm leaning towards it being the MOBO, but it could honestly be the CPU or GPU as well.
Anyone that has lasted through that thriller of a story have any steps to take or recommendations? TIA!

 

[mancsoulja]

From what you've described, sounds like it's the CPU with the issue not the motherboard. Do you remember what voltage you were at? Have you tried resetting the CMOS?

 

[drea.drechsler]

Agree...reset CMOS first and run full stock settings for a while and see if issues resume. Only in full stock settings can you make any judgments about whether CPU or motherboard are defective.

Setting a fixed voltage, especially a low one, on a Ryzen 5000 processor can make it unstable in light threaded workloads. It has to be able to hit high voltages when it boosts as it's designed for that.

Also: be sure to install AMD's chipset drivers from the AMD web site.

 

[Bjg13190]

I have reset the CMOS once today (pulled the battery and allowed system to sit for 20 minutes). Someone in another thread suggested that I pull the battery and let it sit, unplugged overnight. Not sure if that is something I should try or not.

At the moment, I am unable to even reach BIOS, and am lucky if the monitor establishes a connection as sometimes it will, other times it doesn't even recognize the PC being turned on.

 

[Bjg13190]

From what you've described, sounds like it's the CPU with the issue not the motherboard. Do you remember what voltage you were at? Have you tried resetting the CMOS?

I think I was around 1.37V when it stopped functioning and now cant even boot toBIOS

 

[drea.drechsler]

I have reset the CMOS once today (pulled the battery and allowed system to sit for 20 minutes). Someone in another thread suggested that I pull the battery and let it sit, unplugged overnight. Not sure if that is something I should try or not.

At the moment, I am unable to even reach BIOS, and am lucky if the monitor establishes a connection as sometimes it will, other times it doesn't even recognize the PC being turned on.

Battery disconnect might help so worth trying.

Not reaching BIOS suggests something with memory...are you using the right slots? 2nd and 4th from CPU, labeled A2 an B2? fully inserted...usually a click on each end of the stick?

When it fails to boot even to BIOS are you getting trouble LED's lighting?

 

[Bjg13190]

Battery disconnect might help so worth trying.

Not reaching BIOS suggests something with memory...are you using the right slots? 2nd and 4th from CPU, labeled A2 an B2? fully inserted...usually a click on each end of the stick?

When it fails to boot even to BIOS are you getting trouble LED's lighting?

yes, so the RAM has all 4 slots being used and worked for some time. Before when it was only 2x8GB sticks I was using the proper slots and the system also ran fine. All sticks were properly seated.

It does get the trouble lights indicated in the original post. It usually stays on the CPU debug light, but on some of the start ups it goes past that to the VGA debug light. So it's random at different boot ups

 

[natcha12]

After a look around, it seems Ryzen 3 frequently boosts over 1.4v, even up to 1.5 in some cases.
I suggest, seeing as your temps are not the cause, you try to set the higher limit voltage to even 1.45. I don't use ryzen, but i hope there is separate settings for max and specific voltage setups. DO NOT set base voltage at 1.45, only maximum/boost if theres an option for it.

 

[Bjg13190]

After a look around, it seems Ryzen 3 frequently boosts over 1.4v, even up to 1.5 in some cases.
I suggest, seeing as your temps are not the cause, you try to set the higher limit voltage to even 1.45. I don't use ryzen, but i hope there is separate settings for max and specific voltage setups. DO NOT set base voltage at 1.45, only maximum/boost if theres an option for it.

That's solid info, I appreciate it. If I'm able to get back in the system I will give it a try. Currently I'm unable to even reach BIOS any more to tinker with settings.

 

[drea.drechsler]

After a look around, it seems Ryzen 3 frequently boosts over 1.4v, even up to 1.5 in some cases.
I suggest, seeing as your temps are not the cause, you try to set the higher limit voltage to even 1.45....

Don't set any fixed voltage with Ryzen and setting a fixed limit that high is almost a certain early death. Set it in AUTO, especially when troubleshooting. Just expect to see spikes as high as 1.5V when boosting. The boost algorithm will pull voltage back as far as 1.2V in heavy processing.

@Bjg13190 : pull the two extra sticks, leave only the two in A2 and B2. THEN reset CMOS and see if it will start.

If it boots into BIOS great, but make no changes yet. Shut down and put in the next two sticks of memory and RESET CMOS AGAIN. Try now to start up.

 

[natcha12]

Don't set any fixed voltage with Ryzen and setting a fixed limit that high is almost a certain early death. Set it in AUTO, especially when troubleshooting. Just expect to see spikes as high as 1.5V when boosting

Please read the full post. Including the very obvious DO NOT

 

[drea.drechsler]

Please read the full post. Including the very obvious DO NOT

I guess I missed the meaning of your post as it's terminology threw me a bit. There is no option to set a 'limit' for VCore beyond a fixed voltage setting with Ryzen. It's possible to set offsets but the algorithm still manages the limits based on temperature and each processor's fuzed-in FIT parameters.

 

[natcha12]

I guess I missed the meaning of your post as it's terminology threw me a bit. There is no option to set a 'limit' for VCore beyond a fixed voltage setting with Ryzen. It's possible to set offsets but the algorithm still manages the limits based on temperature and each processor's fuzed-in FIT parameters.

I guess it looked clearer to me at the time. I fully agree there should be no changing the settings to hit 1.45 or 1.5 as a baseline, but would it be possible, if all else fails, to turn off the boost mode and only then increase the voltage offset? It does seem like his pc is unstable because it is too low not too high

 

[Bjg13190]

Attempted alternating RAM and resetting CMOS, no joy there. Still no BIOS.

I also removed the CMOS nattery overnight and reinstalled it today. No joy there.

 

[natcha12]

It doesn't matter toooooo much, but you don't need to leave battery out overnight. The full CMOS reset just need no power supply and empty board capacitors. That is done with battery and wall socket removed, then pressing your pc on button to discharge caps.

Is it posting, but without bios? Or absolutely nothing?

 

[drea.drechsler]

Attempted alternating RAM and resetting CMOS, no joy there. Still no BIOS.

I also removed the CMOS nattery overnight and reinstalled it today. No joy there.

At this point, with CMOS reset's not working to at least get you into BIOS I'd agree something is wrong. Isolating board or cpu or PSU or even memory means swapping components.

It could easily be PSU if it's not holding a stable voltage for it to boot up and stay up. What PSU do you have in there now? Do you have another one you could try out?

I'd not try flashing BIOS at all, not until you are absolutely confident it's a stable system. Doing it previously could have bricked the board but you did get it to boot up for a few minutes at least.

 

[Bjg13190]

It doesn't matter toooooo much, but you don't need to leave battery out overnight. The full CMOS reset just need no power supply and empty board capacitors. That is done with battery and wall socket removed, then pressing your pc on button to discharge caps.

Is it posting, but without bios? Or absolutely nothing?

It POSTs up to either the CPU debug light remaining lit or the VGA light remaining lit. It alternates between the two at different boot ups

 

[Bjg13190]

I just tested a separate PSU (motherboard 24 pin and CPU 8 pin and now it boots with the DRAM debug light lit and when I attempt to power it off with the case button (which I used to turn it on) it doesnt work and I have to toggle the I/O switch on the PSU itself

 

[natcha12]

I'm sorry but I need to explain terminology I think.

POSTing is before the BIOS loads, then the pc boots Windows to finish up. POST<BIOS<Boot.

When you say it boots with the dram light on, do you actually mean that, or it simply POSTs?

 

[Bjg13190]

I'm sorry but I need to explain terminology I think.

POSTing is before the BIOS loads, then the pc boots Windows to finish up. POST<BIOS<Boot.

When you say it boots with the dram light on, do you actually mean that, or it simply POSTs?

Fully with you.

I can observe the board POST, the DRAM light flashes on and off as it is checked. BOOT light flashes on and off indicating a bootable drive is present. The CPU light will either come on and stay on, or come on for 10ish seconds and then go to the VGA debug light and that will remain lit.

So either the CPU or VGA light remains lit after POST, but we never get BIOS.

 

[natcha12]

Try your 3070 in the other pcie, and while it's out double check the gold contacts for damage

 

[Bjg13190]

Try your 3070 in the other pcie, and while it's out double check the gold contacts for damage

Sorry for delay. Was able to test the 4x slot today, same indications as before including no BIOS.

Contacts on the GPU look proper to the naked eye.