I ran Memory Diagnostic Tool 2 times, as far as I understand it should show any discrepancies and should fix them automatically.
No, if it finds errors it means that the RAM is having errors which usually means it’s faulty. It can’t fix anything. With the two crash errors you’ve had, I would suspect the CPU.
Please provide the dump files as instructed by the bot.
Need the dump files.
Let’s see what CDI says.
?cdi (Bot command for instructions)
One was corrupted so of the four working ones: One looks like RAM, doesn’t really blame anything it just found corrupted data. One was the filter manager driver which makes sure that files are handled correctly, often storage related. The final two were wof.sys which is storage related.
Third gen Ryzen has a quirk where a faulty CPU usually looks like RAM and on all CPUs, if the memory controller fails it will look like RAM. The memory controller is in the CPU.
These dumps don’t really look like RAM though. RAM and storage can look a lot like each other because of the page file, but this actually looks like storage.
Were both drives SATA 2.5"/3.5"? Did you use the same cable? Was the SSD removed when you tested the HDD for Windows? Did you check where the page file was put on both installs?
Can confirm cwsink’s conclusion. Install the BSOD patch tool from Lenovo. This might have to be reinstalled after every Nvidia driver update.
It just shows the WiFi card. Did you try this driver?
Need the dump files. And how did you test the RAM and storage?
Cloning isn’t guaranteed to work. Yes, the board works with NVMes.
Need the dump files.
A BSOD is a different name for a bugcheck, a specific type of crash. A BSOD will leave logs and can be caused by all manner of different things. That error screen you have means that a file is corrupted. Pretty much no wiggle room.
That’s not a BSOD. It’s telling you that the OS is corrupt.
Freezes are really annoying to diagnose as there are no logs. You just have to go down the list of most common culprits. I would say GPU and PSU share the first place here. If you can test the GPU in a different machine and a different PSU in that machine, I would do so.
If you test a different PSU, DO NOT re-use the cables. Use the cable that belong to each PSU. PSU outputs and therefore the cables are not standardized. Even a different version of the same model PSU could have different outputs and cables. Using the wrong cables can kill components.
Microsoft support said that there is no kernal-41 error in my event log so they are saying it’s not a power issue, but they stopped helping me.
Kernel ID 41 is not power related, so you spoke to someone who has no idea what they are doing. Kernel ID 41 means unexpected shutdown, so any crash, power loss or holding down the power button. Anything that causes the PC to restart or shut down without using the buttons for it in Windows.
That being said, you should have had those events.
With shutdowns, the main suspects are temps and the PSU. With the monitor going black a second before it shuts down, the GPU as well. If it’s not any of those, it could be anything. You just have to try one part at a time or take it to repair shop and have them do it.
From the dump files it looks like a faulty storage drive. I have no idea why you would want to look at Windows events when you have the dump files, they have a ton more information than events.
I can’t see which drive, but it’s almost certainly the drive with the page file. You run this command in Powershell to see which drive has the page file:
Get-WmiObject -Query "SELECT * FROM Win32_PageFileUsage" | Select-Object Name
If you used the MBR2GPT tool and immediately started crashing, chances are that it corrupted the OS. This happens some of the time.
This computer is more than capable of running Windows 11, you just need secure boot and fTPM enabled in the BIOS. As you just reset I would do a clean install of Windows. Our Windows 11 install guide has instructions for converting the drive to GPT if that was changed with the factory reset. You will get an error during the install telling you that you need GPT. The change is done from within the installation.
Note that this does delete everything on the OS drive. When installing, we recommend that you only have the OS drive connected as Windows will put the EFI partition which has the boot files on whichever drive the motherboard designated as Disk 0. So if the drive that is Disk 0 is not your OS drive and that drive dies or needs to be replaced, you might have to reinstall the OS. With only one drive connected, we guarantee that it’s on the OS drive.
Microsoft made a tool that converts it live, no format required. Of course it does corrupt the OS 1-2% of the time requiring a format and reinstall.
You didn’t have any minidump files? This error is a bit iffy to debug with just one dump file and you can only have one kernel dump at a time.
It looks like RAM from the dump files. First try setting the RAM to 5600MT/s or slower as that is the highest officially supported by the CPU and higher speeds overclocks the memory controller in the CPU.
If it still crashes use the machine normally with one stick at a time and see if only one of them cause crashes.
As this is intermittent, you could try a Memtest86 as well, but be aware that memory testers are not super reliable with DDR4 and newer so a clean scan doesn’t mean anything. No false positives.