System shows gradual freezing on Threadripper 3960X / TRX40. RAM slot issues are also occurring.
System shows gradual freezing on Threadripper 3960X / TRX40. RAM slot issues are also occurring.
Hello everyone, seeking advice from those familiar with fixing the Threadripper platform, particularly the TRX40. My setup includes: CPU – AMD Threadripper 3960X, Motherboard – ASUS Prime TRX40-Pro, RAM – 4×16GB DDR4, GPU – RTX 3090, SSD – Samsung 980 Pro (2TB), Cooler – Cooler Master Wraith (potential torque/torque issues with heavy towers), BIOS – latest available, OS – Windows 11 (appears OS-agnostic).
Problem: System freezes completely—no BSOD, just a full lock that needs resetting. Freezes happen even during MemTest, hinting at hardware instability. Freeze timing seems linked to which DIMM slots are active; older slots tend to freeze faster. When the system doesn’t boot, it displays Q-code 30, likely indicating a wake from S3 sleep. This code appears consistently during failed POSTs but isn’t necessarily the root cause.
Actions taken: BIOS updated, re-seated CPU multiple times, checked sockets (no visible damage), re-torqued cooler and CPU to specs, tried various RAM setups (including single stick).
Questions: Is the issue with the motherboard or CPU? How can I pinpoint which one is faulty without adding more parts? Any guidance would be greatly appreciated. Thanks ahead!
IMG_1499.HEIC
I tested each RAM stick one by one, checking for any problems. Once an issue was found, I swapped it out and tried the next stick in its place. I followed this process for every slot, making sure everything works smoothly. OG TRX 40 is quite particular about RAM, so I compared it to a new high-performance model. I even brought extra RAM (8 or 16GB) when it was in poor condition, tested them thoroughly, and kept a backup if the main RAM failed. That shows how strict OG TRX can be about RAM quality. Also, it’s a good idea to double-check everything at night to confirm it runs correctly.
The image doesn't show the error code you mentioned. The displayed value is 0C, which isn't a recognized error code. It seems the entry is marked as "Reserved for future AMI SEC error codes," meaning it's not currently in use or documented.