Question Stability issues, random crashes

[studiomatthewbrooks]

Hi all,

Long time lurker, first time poster.

I recently put together a workstation based around a Xeon E5-1650 v4 and Supermicro X10SRA-F. I pulled 128GB (8x16GB) of Samsung DDR4-2400 ECC RAM from an HP Z440 workstation, so I'll admit that the RAM compatibility with the board is a little iffy. The rest of the components are an EVGA 1000W P2 PSU, Quadro K2200 GPU, NZXT H700 case, Samsung 860 Pro 256GB (boot drive), Noctua NH-U14S heatsink (Thermal Grizzly Kryonaut paste), and five Noctua NF-A14 iPPC-2000 140mm case fans. The heatsink and case fans are controlled using Noctua NA-FC1 controllers to keep them at max, so cooling is not an issue.

Everything seemed to be working well initially but I've had some random crashes and instability which I'm inclined to think is either a RAM or PSU issue. I was able to perform basic benchmarks without any trouble (Passmark, Geekbench, CrystalDiskMark, etc.) but around 30min of Aida64 produced a hard crash. No BSOD, just turns off suddenly and then starts booting again. I also had a few random crashes with little to no load, so it's a little hard to tell what's going on exactly. Some random crashes were shortly after logging into W10, others were during longer sessions.

I made a bootable USB with memtest86 and was able to get it running, but had the same hard crash after ~10min (no errors reported before the crash). It also exhibited some weird behaviour after crashing where it tried to reboot, fans came on for 1-2sec, then off again, then tried to boot again, fans on for 1-2sec, then off again, etc.. maybe three or four times like that and then POST/booted as normal.

From what I've described, does this sound more like a RAM or PSU issue? Bad motherboard? I can get a replacement motherboard or PSU since they're both still under warranty, but the RAM is a little more complicated so hopefully it doesn't come to that. Any thoughts?

System specs
Mobo: Supermicro X10SRA-F
CPU: Xeon E5-1650 v4
RAM: Samsung 128GB PC4-2400T DDR4-2400 ECC (8x16GB)
GPU: Quadro K2200
PSU: EVGA 1000W P2
Heatsink: Noctua NH-U14S, Thermal Grizzly Kryonaut paste, Noctua NF-A14 iPPC-2000 (x1)
Case: NZXT H700, Noctua NF-A14 iPPC-2000 (x5), Noctua NA-FC1 fan controllers
Storage: Samsung 860 Pro 256GB (boot drive), Intel Optane 905P 380GB x2 (media/working drive)

 

[studiomatthewbrooks]

I was probably too vague describing the RAM, so here are images of the sticks for reference: imgur.com/a/0tSM5YH

 

[studiomatthewbrooks]

Sorry for posting again but my previous thread was either deleted or flagged as junk (?)

System specs
Mobo: Supermicro X10SRA-F
CPU: Xeon E5-1650 v4
RAM: Samsung 128GB (8x16GB) PC4-2400T 1Rx4 DDR4-2400 ECC
GPU: Quadro K2200
PSU: EVGA 1000W P2
Case: NZXT H700, Noctua NF-A14 iPPC-2000 140mm case fans (x5), Noctua NA-FC1 fan controller (x2)
Heatsink: Noctua NH-U14S, Thermal Grizzly Kryonaut, NF-A14 iPPC-2000 140mm fan
Storage: Samsung 860 Pro 256GB (boot drive), Intel Optane 905P 380GB (working drive)
OS: Windows 10 Pro 64-bit, version 1903

Long story short, this is a workstation build with mostly new components but a couple being reused from an HP Z440 workstation (CPU and RAM). Everything seemed to be working great and I was able to run several benchmarks (Geekbench, Passmark, CrystalDiskMark, etc.) with no issues. However, today I noticed a couple of strange crashes (suddenly turning off, no BSOD, proceeds to reboot)—both while benching/testing and little to no load—which led me to do some stress testing with Aida64. After about 25min, I got the same hard crash with no BSOD and a reboot.

My first inclination was either a RAM or PSU issue, or maybe a compatibility issue between the mobo/RAM since the latter was pulled from a Z440. I made a bootable USB with memtest86 and ran it for about 10min before getting the same hard crash with no BSOD, no errors reported before the crash. After crashing, it also exhibited some strange behaviour where the fans came on for 1-2sec, then off, restarted again, fans on for 1-2sec, then off.. maybe three or four times before POST/booting properly. To me that behaviour seemed more PSU related but I'm not sure.

From what I've described, does this sound more like a RAM or PSU issue? Or perhaps the mobo is dodgy? I can get a replacement PSU or mobo since they're both under warranty, but I'd rather not replace the RAM if possible. I've gone through every component in the system and made sure that the driver and/or firmware is up to date, including the X10SRA-F BIOS/IPMI/BMC. Any ideas?