Identical twin boards die nearly at the same time

Sometimes I wonder if the phenomenon of identical twins dying shortly after each other applies to motherboards as well...


I have three AM4 systems in my nuclear magnetic resonance facility.

These are my own builds, all currently using AM4 hardware. These systems have been in my personal office space for quite a while, and have gone through several hardware changes throughout the years.

Two of them are using MSI x470 Gaming Plus boards bought at the same time at Newegg, 8 years ago, the other one using a Gigabyte Aorus board that I bought from Format C: a couple years ago (and it's working very smoothly).



Last week, one of the x470-based systems started glitching badly, but in a span of another week, the second of the AM4 systems with their 8 year old MSI x470 Gaming Plus boards started glitching as well in the same manner.

I've eliminated all possibilities other than the motherboard, since putting the M2 hard drives in remaining functional AM4 system lets it boot up and operate just fine, and swapping out components that I know are fully working, hasn't changed anything.

Basically, they'll boot up, and are indefinitely stable at the UEFI / BIOS screen, but as soon as I try to boot either to the hard drive, or even to a Windows 11 installation USB media drive, they just keep rebooting in an endless cycle, trying to start an automatic repair, but never getting anywhere.

Running a repair installation won't work, since I just get to a point where the installation tells me that the repair can't be completed, assuming that I can even make it that far into a boot.

Just for shits and giggles, I even tried to put a new hard drive on in one of them, and install Alma Linux, to no avail, with the same result of constant reboots.



I'm not complaining about the death of those boards, since I did get over 8 years of excellent use out of them, and they were simply collecting dust at home for a short while, which is why I had upgraded two of my personal office systems.

They had been working great for about 5 years in my office, but alas, it appears their time has come.

Naturally, given how uncooperative the University has been towards my getting new computers, they nixed the notion of getting me new systems, but reluctantly agreed to pay for the cost of the two replacement boards.

So, I have a couple of Asrock B550 boards arriving next week.

Might want to check for radiation leaks.

and test/replace the cmos batteries along with a good clearing.

Unabomber ..... you developing novel chem weaponry? Say no more, I'm in.

Aside from testing CMOS as mentioned, it could be dirty or inconsistent power in the room? Get brown outs? Other gear go bad in there at all? Could also just have been a bad manufacturing batch I guess or something (like soldering/tracing - or really the solder/copper to do so maybe degraded) if you got them at same time.

Edit: You checked the caps, right?

Thanks for the suggestions, folks.

CMOS battery socket looked clean, but I replaced the CR2032's. No difference.

I'll take a look at the capacitors Monday morning, for sure. I'm actually curious to see whether or not they're electrolytic or solid state polymer, but that's going to require a close look.

The last time I remember scrutinizing capacitors this closely were when a lot of brand name power supplies were using those horrible Fujyuu (sp?) ones, but this time around, I'm going to have to look up close without my glasses (started getting more farsighted in these last 20 years). Getting old is a downer, I guess...

All of my systems receive clean filtered power through some Emerson UPS units.

My facility is on "emergency power" that receives a steady stream, the same kind of line that the emergency lines in the hospital use.

The Avance III HD consoles (300K to 600K each) require steady power, and in the entire time I have had my facility in this building (since 2012), we've never had any brownouts. For that matter, they've been constantly on ever since June of 2012, with the only shutdowns occurring when I manually turn them off (had to replace a few of the signal boards over the years).



On another note, we do chemical and biomolecular research, no weaponry involved, and all of the isotopes we analyze (1H, 2H, 13C, 15N, 31P, etc) are all stable isotopes.

I'm a bit rusty at computer troubleshooting. Your symptoms sound like the UEFI boot configuration might have a configuration mismatch compared to the installed OSes on their respective drives. The UEFI often has a setting that can toggle between UEFI boot and CSM boot (Compatibility Support Module AKA Legacy Boot). They might be called something else depending on the motherboard brand. Try switching the setting, saving to UEFI, and rebooting to see if there's any difference in boot behavior with your drives. It's possible that your UEFI settings were changed or reset to default at some point, especially if you did a UEFI firmware update (Windows did it for security reasons on one of my laptops years ago but I think it was an optional update).

Unabomber said:

Sometimes I wonder if the phenomenon of identical twins dying shortly after each other applies to motherboards as well...

Click to expand...

Same part made in same factory had the same hidden flaw leading the the same problem after the same runtime in the same environment.

Consistency in manufacturing is a good thing, mostly.

I'd run a long memtest and/or clean/reset ram and disk and maybe cpu.

brinstar117 said:

I'm a bit rusty at computer troubleshooting. Your symptoms sound like the UEFI boot configuration might have a configuration mismatch compared to the installed OSes on their respective drives. The UEFI often has a setting that can toggle between UEFI boot and CSM boot (Compatibility Support Module AKA Legacy Boot). They might be called something else depending on the motherboard brand. Try switching the setting, saving to UEFI, and rebooting to see if there's any difference in boot behavior with your drives. It's possible that your UEFI settings were changed or reset to default at some point, especially if you did a UEFI firmware update (Windows did it for security reasons on one of my laptops years ago but I think it was an optional update).

Click to expand...

Thanks. I'll definitely try that this Monday.

It had originally be setup for CSM mode back in 2018, but after I got rid of the old DVD-RW drive that required that mode, I converted it to UEFI, and changed the UEFI setting. I'll take a look at the settings.

Yeah, if I can keep the old warriors going, I'd like to, since they've been excellent boards throughout the time. That, plus I won't have to bother the IT guy and register new MAC addresses.

Just to be safe, I took out the old warrior power supply (a 19 year old Antec Truepower Trio 650 watt unit) on one of them, and put in a new Corsair RM750e power supply.

Memtest86 showed no memory errors, even after 4 cycles.

Capacitors all look good (solid state), although I honestly don't know how to spot a defective solid state capacitor...

I checked the UEFI settings.

Surprisingly, it had been set to Legacy + UEFI, even though I know that I had set it to UEFI-only several years ago, and hadn't done any BIOS updates since then.

I changed it back to UEFI only, and tried to boot from the hard drive. It got to the login screen, and then it promptly rebooted again, starting the endless cycle once more.

I also tried putting a different fresh hard drive in there, and did a Windows 11 Pro installation on it. The installation actually finished the first round, but when it rebooted, it got to the initial welcome screen (white screen with Windows logo), playing the Windows sound, and then rebooting even before getting to the login screen.

Flashing the BIOS to the latest version (and making sure it's UEFI-only) results in the same behavior (continual reboot cycle).



The M2 hard drive works fine in the third system that uses the B550 board. I also transplanted the memory, GPU, and CPU as well, just to rule out any other possibilities.

OCCT didn't show any errors with that Frankenstein system.

Asrock B550 Gaming 4 / AC boards came in today. Both are working fine, but I had to find out the hard way that these boards do not support M.2 SATA drives (NVMe-only).

No big deal since one system was using a NVMe drive, and I had a 1 TB SATA 2.5" SSD (Teamgroup) sitting around for the other system, and Windows 11 Education installed without a hitch on both.

I'm actually impressed with the Asrock boards. They're solidly built, and use solid state capacitors.

I don't intend on using these boards for another 8 years, since I'm hoping that:

1) The university powers that be stop being cheapskates, and either gets me new systems or lets me upgrade to something more substantial

and

2) The price of RAM will one day come down, to the point where I don't have to stick with a DDR4 platform.

In the meantime, these Ryzen 5 systems are plenty powerful, so I should be good for several more years.

Unabomber said:

550 Gaming 4 / AC boards came in today. Both are working fine, but I had to find out the hard way that these boards do not support M.2 SATA drives (NVMe-only).

No big deal since one system was using a NVMe drive, and I had a 1 TB SATA 2.5" SSD (Teamgroup) sitting around for the other system, and Windows 11 Education installed without a hitch on both.

I'm actually impressed with the Asrock boards. They're solidly built, and use solid state capacitors.

I don't intend on using these boards for another 8 years, since I'm hoping that:

1) The university powers that be stop being cheapskates, and either gets me new systems or lets me upgrade to something more substantial

and

2) The price of RAM will one day come down, to the point where I don't have to stick with a DDR4 platform.

In the meantime, these Ryzen 5 systems are plenty powerful,

Click to expand...

Before I got my Synology in 2018 I did DIY media servers/WMC DVR servers and used exclusively consumer low power ASRock embedded CPU boards for that (also what I'm going back to after I'm done with the Synology) - also they do a lot of industrial versions of those embedded (but also socketed as well) boards and would be my pick there too - they make good stuff for that IMO

Maybe they both got hit by cosmic particle bit flips that broke them. Joking but really makes me ponder such possibilities.

CrimsonKnight13 said:

Maybe they both got hit by cosmic particle bit flips that broke them. Joking but really makes me ponder such possibilities.

Click to expand...

Neutrons are serious business