Question System that worked fine for years suddenly refuses to boot with RAM set to XMP speeds ?

[Cyber_Akuma]

Here is my build: https://pcpartpicker.com/user/Cyber_Akuma/saved/vJypBm

If you just want the important parts:
CPU: 11700K
Motherboard: Gigabyte Z490 Aorus Pro AX
RAM: G.Skill Ripjaws V 64 GB (F4-3600C18D-64GVK)

I have been basically building this system since early 2021, budget constraints have prevented me from fully finishing it but it's at least been in a functional state for a while now.

Recently after it had been running fine for months I tried to finish some more work on it, which involved re-wiring the drives, installing a sound card, and re-installing a RAID card that had been disconnected for a while but was originally on it. I also updated the BIOS.

I then started getting errors from the BIOS that a boot error was detected and that it would not let me continue trying to boot without either entering the BIOS and fixing it or just completely resetting my BIOS settings.

I tried checking my wires, removing the sound card, and other tests and eventually have narrowed it down to the RAM, specifically, when it's operating past the JEDEC standard of 2666.

Sometimes it will boot fine with XMP enabled, but about 95% of the time it will not. When it does boot it can run for literally months just fine, but every reboot after is a gamble if it will complain about a boot error or not.

I tried setting the CAS latency from 18 to 19 to see if that was it, didn't work. I tried manually adjusting the voltage from the XMP default of 1.35 in increments of 0.01 up to 1.4 (I didn't want to risk going higher) and nothing. I even tried manually adjusting the memory speed but ANY speed above 2666, even at just a mere 2700, it refuses to boot.

This utter strict adherence to the memory speed being at 2666 and not even being able to handle 2700, while other times it can randomly do the XMP profile of 3600 and stay rock solid for months even with all the stability tests I throw at it makes me worry that it's the motherboard rather than the RAM. Very odd that it can randomly boot fine at 3600 and stay fine, but then claim anything over 2666 is causing it to fail to boot. I really really hope it's not the motherboard, this thing would be a mess to completely take apart and rebuild and I doubt I can find another of the same board for any reasonable price.

Does anyone have any idea what this could be? Or what I can try? I really want to get the RAM running at XMP speeds again, it was fine from 2021 until just a few weeks ago. At this rate this thing has me pulling my hair out so much I wish I could just disable whatever blasted boot error detection it has and spend days running my own stability tests just to make sure the motherboard's boot error detection isn't just garbage.

 

[Lutfij]

You might want to replace the CMOS battery and see if that helps.

Perhaps your motherboard is the culprit all along. You might want to drop your processor and ram kit onto another motherboard with the latest BIOS version and see if the same hiccups come up. In which case try another ram kit with the processor, to see if the integrated memory controller(on your processor) is your issue.

I also updated the BIOS.
This is where you state the version of the BIOS.

 

[Cyber_Akuma]

I do not have another motherboard that I can put the CPU in, and the only other DDR4 desktop I have I am pretty sure only accepts EEC Memory, and even if it didn't only supports up to a speed of 2133.

I was on BIOS version F20, the system had been on that version since early 2021. I updated to F23. Then I used the flashback to force a downgrade to F20 because I thought that was the problem, but I am back on F23 now since the same boot errors were happening on F20 too.

 

[PiersPlowman]

Degraded memory controller? Perhaps XMP bumped the voltages up that little bit too much.

 

[Vic 40]

I tried checking my wires, removing the sound card, and other tests and eventually have narrowed it down to the RAM, specifically, when it's operating past the JEDEC standard of 2666.

You haven't tried without the raid card? Since that is another thing you added.

 

[SkyNetRising]

I then started getting errors from the BIOS that a boot error was detected and that it would not let me continue trying to boot without either entering the BIOS and fixing it or just completely resetting my BIOS settings.

Probably not related to ram at all.
OS drive is more likely suspect.

Can you show a photo of error, you're getting?
(upload to imgur.com and post link)

Also show screenshot from Disk Management.

 

[Cyber_Akuma]

You haven't tried without the raid card? Since that is another thing you added.

It was re-added, it was in the system before and worked fine. The only new addition was the sound card, which I tried removing to no effect.

Probably not related to ram at all.
OS drive is more likely suspect.

I highly doubt it's the OS drive, this is the motherboard complaining before it ever tries to boot the OS. It even overrides me pressing F12 to bypass the OS drive and bring up the boot menu when that error comes up.

Can you show a photo of error, you're getting?

I don't believe I can take a screenshot of it so here's a photo:

https://imgur.com/BemjuXn

View: https://i.imgur.com/BemjuXn.jpg

Also show screenshot from Disk Management.

The drives are a bit of a mess since it's a temporary setup, but I tried to make it as readable as possible. Normally I have two other (non-bootable) USB flash-drives connected:

https://imgur.com/PkhTG1h

View: https://i.imgur.com/PkhTG1h.png

 

[kerberos_20]

you did mention updating bios, its possible that current bios version doesnt like your ram, you can try roll back to previous bios version to rule out hardware defect

 

[Cyber_Akuma]

you did mention updating bios, its possible that current bios version doesnt like your ram, you can try roll back to previous bios version to rule out hardware defect

I did try that, but the exact same problem re-appeared in the old BIOS which was working fine for two years before.

 

[SkyNetRising]

I highly doubt it's the OS drive, this is the motherboard complaining before it ever tries to boot the OS. It even overrides me pressing F12 to bypass the OS drive and bring up the boot menu when that error comes up.

The drives are a bit of a mess since it's a temporary setup, but I tried to make it as readable as possible. Normally I have two other (non-bootable) USB flash-drives connected:

Your storage configuration doesn't match pcpartpicker list.
What is model name of 120GB OS drive?

Also there's something weird going on with OS drive.
Active "System Reserved" partition should be identified as "System". But you have OS partition as "System".
Did you photoshop the screenshot?
Or did you set "System Reserved" partition as active after booting up your pc?

I'd suggest, you recreate the bootloader.
Execute from elevated command prompt. Regular command prompt will give error on last step.
If you get any errors, then stop immediately and show screenshot.

diskpart​

list disk​

select disk 4​

(select 120GB disk)​

list partition​

select partition 3​

(select 499MB recovery partition)​

delete partition override​

create partition primary​

format fs=ntfs quick​

active​

assign letter=H​

exit​

bcdboot C:\windows /s H:​

And reboot your pc.
C: - partition should be identified as "Boot".
New 499MB H: partition (new bootloader) should be identified as "Active, System".

After you have successfully completed this, you can delete 50MB "System Reserved" partition.

diskpart​

list disk​

select disk 4​

(select 120GB disk)​

list partition​

select partition 1​

(select 50MB System Reserved partition)​

delete partition​

 

[Cyber_Akuma]

It doesn't match the list because drives C, S, and Q are all temporary drives that I will disconnect later.

If it was the bootloader then I would be getting an error that there is no OS detected instead of a hardware boot failure, again, it will give me this error even if I attempt to boot from a USB drive instead of the SSD. It's my motherboard reporting this error before it attempts to boot from ANY drive.

 

[Cyber_Akuma]

Ok, so just to make sure I tried disconnecting the drive while the RAM is still set to JEDEC speeds, none of the other drives in the system have even a bootloader on them, much less an OS.

After the motherboard logo instead of seeing the spinning circles from when Windows loads (Or the error message if it had a boot error) I saw a cursor keep jumping up and down, pretty sure that's the board searching each drive for a bootloader, and then loading me into the CMOS settings. Expected behavior, that's generally what happens when the hardware is working fine but the board did not find anything to boot. I didn't get an error message.

So I tried setting it back to XMP on expecting to get the error message, but like I said before sometimes it will rarely decide to work, and it did (still dumped me into the CMOS, but again with no error) so before it flakes out again I quickly booted up Memtest86+ 7.0 to see if I can catch anything wrong with the RAM.

The test is still going but I did notice something odd. Unless this is a bug in Memtest86+ itself, it's claiming my IMC is reporting an insane speed and laughably high timings:

EDIT: Also, I just noticed. It's saying it's testing "66GB - 67GB" .... out of 64GB?

https://imgur.com/JHM9lku

View: https://i.imgur.com/JHM9lku.jpg

 

[Cyber_Akuma]

This problem is still happening. The system was behaving for a few days, but it started happening again, RAM tests if the system manages to POST don't show any errors. I also tried a completely different boot drive and install of Windows and it still happened, so definitely not because of the install.

 

[artk2219]

This problem is still happening. The system was behaving for a few days, but it started happening again, RAM tests if the system manages to POST don't show any errors. I also tried a completely different boot drive and install of Windows and it still happened, so definitely not because of the install.

It sounds like it's hardware at this point, either the memory controller in the CPU, or the motherboard. What if you set the xmp settings manually? Will the ram run at the rated xmp speeds without the xmp timings?