Hi everyone. My PC seems to be having a bad couple of days with multiple things breaking and I'm out of ideas. So I'm listing the events chronologically in the hopes that I can get some help here.
TL;DR: Chipset provided NVME slot doesn't work, ethernet doesn't work. Windows keep rebooting. And I managed to physically damage one of the SSDs while troubleshooting. All in a span of 6 hours. Looking for ideas other than taking the motherboard to Gigabyte for service.
Ryzen 5900X
Noctua D15
Gigabyte x570 Aorus Elite
Corsair 3000MHz 8GB x4 (not a kit, I bought 2 pieces In Nov'20 and another 2 a couple months back. same model number though)
Asus TUF GTX 1650 Super
Samsung 980 Pro 1TB M.2 SSD (installed on the Processor's slot)
Crucial P1 500Gb M.2 SSD (installed on the chipset slot, with Windows 10)
Seasonic S12ii 620W power supply (this thing is about 7 years old)
5 120mm fans, 1 140mm fan, (not including the 2 on the D15). I have another 120mm fan with only the RGB header plugged in and not the fan header (because I ran out of them). I doubt any of the fans are relevant to my problem though.
Samsung 870 EVO 2.5" Sata SSD (bought yesterday. not installed when the problems started)
It all started with the PC suddenly restarted and landed on the no bootable media found screen within the 5 mins I was away from the desk. I was just doing normal JS programming before that.
Unplugged PSU, rebooted, reinstalled the Crucial SSD in the same slot. No luck. I wanted to check if it was the SSD or the mobo and while I was removing the Crucial SSD again to swap the slots, I seem to have dislodged something on the SSD (transistor/resistor/something else?). Anyway, I installed the 980 Pro on my chipset slot, installed the Crucial on the CPU slot and booted via an Xubuntu linux live USB and couldn't see both the drives. I moved the 980 pro back to the CPU NVME slot and it showed up fine in linux. So, inference, both the chipset NVMe slot and the Crucial SSD are dead.
Also, while I'm booted into linux I notice that my ethernet adapter isn't detected, but an orange light keeps blinking. From the manual I see that this is the connection speed LED. The activity LED remains Off. I dismiss this as a linux quirk, though I have never seen any linux distro not detect wired ethernet.
Next day I buy and install new Samsung 870 EVO 2.5" SATA SSD. I backup the data from 980 Pro to 870 EVO via Linux live USB. Then during windows 10 install I could only see the 980 pro. NOT the 870 EVO. Regardless, I install windows on the 980 pro. Note that Crucial SSD is still installed in the chipset slot. Once I'm booted into Windows, Ethernet's not detected here either. The fans were bothering me now (the case was open and the 100% ramp ups during each reboot annoyed me). So I download, install and open HWInfo to check the temps and before it could open the PC restarted. I tried again, same result. I install CPUID HWMonitor and even that triggers a reboot before it even shows any data. The temps in the BIOS show 38C for the CPU. I install Xubuntu to the 980 pro and check the temps there and it's fine as well. But still no ethernet. Checked lspci, lshw, there's no trace of the ethernet adapter. USB tethering via my iPhone works and a USB Archer T3U wifi dongle works.
To take a stab at windows again, i disconnected the 870 EVO, front USB cable, reseated the graphics card. Updated the BIOS from f31b to f36e. Now it won't even boot into the windows installation USB. Those spinning dots show up for a while and then it reboots. I ran memtest86 overnight and it showed no errors. I tried a different windows 10 ISO, windows 11 iso with GPT and MBR based bootable USB and all of them still reboot at the spinning dots. Linux continues to work fine (except the ethernet). I haven't done CPU intesive work but I have been browsing on it and installing a bunch of stuff.
So I am at the conclusion that something went wrong in the motherboard. Maybe some traces went bad? It may also be that some pins on the CPU went bad, but then why are there no reboots on linux and why is the SSD working only in linux? So I guess there's something wrong with the chipset and how the chipset driver gracefully handles it.
I am planning to take the motherboard to Gigabyte service. I would be super thankful if you folks have any ideas that could save me a trip to them especially that they are giving a 10-day SLA.
TL;DR: Chipset provided NVME slot doesn't work, ethernet doesn't work. Windows keep rebooting. And I managed to physically damage one of the SSDs while troubleshooting. All in a span of 6 hours. Looking for ideas other than taking the motherboard to Gigabyte for service.
Ryzen 5900X
Noctua D15
Gigabyte x570 Aorus Elite
Corsair 3000MHz 8GB x4 (not a kit, I bought 2 pieces In Nov'20 and another 2 a couple months back. same model number though)
Asus TUF GTX 1650 Super
Samsung 980 Pro 1TB M.2 SSD (installed on the Processor's slot)
Crucial P1 500Gb M.2 SSD (installed on the chipset slot, with Windows 10)
Seasonic S12ii 620W power supply (this thing is about 7 years old)
5 120mm fans, 1 140mm fan, (not including the 2 on the D15). I have another 120mm fan with only the RGB header plugged in and not the fan header (because I ran out of them). I doubt any of the fans are relevant to my problem though.
Samsung 870 EVO 2.5" Sata SSD (bought yesterday. not installed when the problems started)
It all started with the PC suddenly restarted and landed on the no bootable media found screen within the 5 mins I was away from the desk. I was just doing normal JS programming before that.
Unplugged PSU, rebooted, reinstalled the Crucial SSD in the same slot. No luck. I wanted to check if it was the SSD or the mobo and while I was removing the Crucial SSD again to swap the slots, I seem to have dislodged something on the SSD (transistor/resistor/something else?). Anyway, I installed the 980 Pro on my chipset slot, installed the Crucial on the CPU slot and booted via an Xubuntu linux live USB and couldn't see both the drives. I moved the 980 pro back to the CPU NVME slot and it showed up fine in linux. So, inference, both the chipset NVMe slot and the Crucial SSD are dead.
Also, while I'm booted into linux I notice that my ethernet adapter isn't detected, but an orange light keeps blinking. From the manual I see that this is the connection speed LED. The activity LED remains Off. I dismiss this as a linux quirk, though I have never seen any linux distro not detect wired ethernet.
Next day I buy and install new Samsung 870 EVO 2.5" SATA SSD. I backup the data from 980 Pro to 870 EVO via Linux live USB. Then during windows 10 install I could only see the 980 pro. NOT the 870 EVO. Regardless, I install windows on the 980 pro. Note that Crucial SSD is still installed in the chipset slot. Once I'm booted into Windows, Ethernet's not detected here either. The fans were bothering me now (the case was open and the 100% ramp ups during each reboot annoyed me). So I download, install and open HWInfo to check the temps and before it could open the PC restarted. I tried again, same result. I install CPUID HWMonitor and even that triggers a reboot before it even shows any data. The temps in the BIOS show 38C for the CPU. I install Xubuntu to the 980 pro and check the temps there and it's fine as well. But still no ethernet. Checked lspci, lshw, there's no trace of the ethernet adapter. USB tethering via my iPhone works and a USB Archer T3U wifi dongle works.
To take a stab at windows again, i disconnected the 870 EVO, front USB cable, reseated the graphics card. Updated the BIOS from f31b to f36e. Now it won't even boot into the windows installation USB. Those spinning dots show up for a while and then it reboots. I ran memtest86 overnight and it showed no errors. I tried a different windows 10 ISO, windows 11 iso with GPT and MBR based bootable USB and all of them still reboot at the spinning dots. Linux continues to work fine (except the ethernet). I haven't done CPU intesive work but I have been browsing on it and installing a bunch of stuff.
So I am at the conclusion that something went wrong in the motherboard. Maybe some traces went bad? It may also be that some pins on the CPU went bad, but then why are there no reboots on linux and why is the SSD working only in linux? So I guess there's something wrong with the chipset and how the chipset driver gracefully handles it.
I am planning to take the motherboard to Gigabyte service. I would be super thankful if you folks have any ideas that could save me a trip to them especially that they are giving a 10-day SLA.
Facebook
Twitter
Instagram