Question B650M Pro RS. Sudden problems with POST that are only resolved by doing CMOS reset.

[ichik]

tl;dr
Built the PC in late 2023, it worked without any major issues with the only exception being RAM was not able to run with EXPO profile, so I just ran it on 5200. Since BIOS version 3.20 I was able to use it with EXPO.
Since a week ago I started getting problems with POST on reboots (in particular after installing OS updates usually, I run Fedora 42). It shows white cursor in the top left and only clearing CMOS helps after which it works for a few days without any issue.
It could be related to EXPO, but so far lowering to 5800 did not fix the issue, neither did flashing back to 3.25.

Config:
MB AsRock B650M Pro RS (BIOS versions 3.30 and 3.25)
CPU 7600X (with Arctic Liquid Freezer II 240 AIO)
GPU Sapphire Nitro+ 7900 XTX
RAM G.Skill DDR5-6000 32GB (F5-6000J3636F16GX2-FX5)
Storage WD SN850X 2TB + a few SATA SSDs
PSU Corsair RM850x
OS Nobara Linux 42 (Fedora-based)
Case Fractal Design Pop Air

---

Longer description.
It first occured while playing videos with MPV, suddenly the next one in the playlist couldn't be played and neither the previous one or any at all. On reboot: symptoms described above. I tried resitting GPU and checking slots and cables: everything is in order. Resetting CMOS helped, but after a couple of days problem came back.
So far, I tried rolling back to 3.25, lowering memory speed to 5800, but none had long-term effect.
Checking with MemTest86 shows no errors with EXPO turned on.
Normally I am able to sleep/wake/reboot/shutdown/powerup the PC. Turning off the mains, unplugging the cable, etc. It comes back up… until it doesn't. At which point only CMOS clear helps.
BIOS settings are usually as follows: EXPO profile turned on, ECO mode turned on for the CPU, custom curves on CPU and case fans, dGPU Only mode turned on.

---

What could be the culprit? Is there anything else I could try doing before trying to RMA the MB? (It's under warranty until October).

 

[Lutfij]

Welcome to the forums, newcomer!

You could source(borrow, not buy) another ram kit with AMD's E.X.P.O advertised on it to see if your ram kit is at fault and requiring an RMA.

Just curious, was your chipset drivers updated prior to you flashing the BIOS to the latest? That's what you need to do if you were on a Windows OS install.

You might want to take the processor and ram kit over onto another known working motherboard(with the right BIOS version) at your friend's or neighbors and see if the issue persists. The probable issue might be your ram kit doesn't like going to DDR5-6000MHz or the E.X.P.O profile or the integrated memory controller(on your processor) is having issues. If the problem is non existent on the donor platform, then the issue is with the motherboard, namely the BIOS.

Just an observation;
GPU Sapphire Nitro+ 7900 XTX
+
PSU Corsair RM850x
If I were you, I'd have looked at a higher wattage PSU for headroom.

 

[ichik]

I think like I've overfocused on EXPO in my post. I've just ran 2 cycles with Memtest86 without errors (EXPO on, ECO mode off this time), now running OCCT, but I doubt anything will be found.

AsRock support got back to me and suggested I try checking the storage next time it happens and interestingly enough checking the SMART status I see that one of the drives has 2 "Data Address Mark Errors" which is still interpreted as "Good" by partition manager, but is not 0. And that's the same drive on which the malfunctioning videos were located (although I did later play them correctly after the first CMOS reset).

 

[CountMike]

I think like I've overfocused on EXPO in my post. I've just ran 2 cycles with Memtest86 without errors (EXPO on, ECO mode off this time), now running OCCT, but I doubt anything will be found.

AsRock support got back to me and suggested I try checking the storage next time it happens and interestingly enough checking the SMART status I see that one of the drives has 2 "Data Address Mark Errors" which is still interpreted as "Good" by partition manager, but is not 0. And that's the same drive on which the malfunctioning videos were located (although I did later play them correctly after the first CMOS reset).

I would suggest you backtrack BIOS version to 2.10, Newer ones are focused on 9000 series and newest ones on 9000x3d CPUs with different requirements. It also allows for better performance because of less constraints with 7000 series.2.10 is what I run with my 7900x and tried all other ones but 2.10 is still best. Best RAM for Ryzen are single rank

 

[ichik]

I would suggest you backtrack BIOS version to 2.10, Newer ones are focused on 9000 series and newest ones on 9000x3d CPUs with different requirements. It also allows for better performance because of less constraints with 7000 series.2.10 is what I run with my 7900x and tried all other ones but 2.10 is still best. Best RAM for Ryzen are single rank

Gonna try your advice next. In the meantime chasing faulty drive issue did not lead to anything.
It happened again during reboot after installing some OS kernel updates. It is beyond me though how OS updates could cause such behavior in a motherboard. One by one I disconnected all the drives, but it had no effect. I disconnected 1 SATA SSDs, then another one, then the GPU, and then the M.2 drive. System was getting stuck on POST with VGA LED on. Putting back GPU without any drives in did not change anything except allowed me to see the cursor (notably without GPU, with the DP cable connected to the motherboard port I was not getting a cursor, although display was getting signal).
Resetting CMOS brought system back just as before. I'm currently running the 3.30 with ECO mode, 5200 speed for RAM (not through EXPO, everything else at default) and custom fan curves.