Catastrophic Freeze on New 9950X3D System & Persistent POST Failure Loop on TRX40 System

[jonneymendoza]

TL;DR: My new 9950X3D system froze while gaming and no longer boots (now out for RMA). Trying to use my old Threadripper system (with the same PSU/GPU) has resulted in a complex POST failure loop that I can't solve, leaving me with zero working PCs.

Part 1: The Initial Catastrophe (New System - Out for RMA)​

This system's failure initiated the entire problem.

• Components: ASRock X870E Nova WiFi + AMD Ryzen 9 9950X3D + Corsair Vengeance RGB DDR5 6200MHz 96GB RAM + Seasonic PRIME TX-1000 PSU.
• The Failure: System froze instantly during gameplay. After a forced restart, it fails to POST entirely.
• Debug Code: Consistently failed at Code "00" (Fatal initialization failure).

Exhaustive Troubleshooting Completed on the 9950X3D System:​

• Tested both RAM sticks individually in all four DIMM slots (A1, A2, B1, B2).
• Cleared CMOS multiple times using the jumper method (including 15-minute power drain).
• Updated BIOS from version 3.20 to 3.50 using USB BIOS Flashback.
• Removed and reseated CPU—confirmed no bent pins in the socket.
• Tested with integrated graphics (GPU removed) and removed all M.2 drives.
• Visual inspection of all RAM slots, CPU socket, and contact pads revealed no physical damage.

Part 2: The Troubleshooting Nightmare (Old Threadripper System)​

My old system (ROG Zenith II Extreme, over 5 years old) was working perfectly when I decommissioned it in March/April to migrate the PSU (TX-1000) and GPU (5090) to the new build. After moving those parts back, the system failed.

• Motherboard: ROG Zenith II Extreme (TRX40)
• Initial Fix: Had to update the BIOS via FlashBack to fix a "Load VGA BIOS" error.
• The Loop: The system now enters a rotation of "POST Error" $\leftrightarrow$ "CMOS Error" $\leftrightarrow$ "Code 02 (HDD detect)" with no display output.
• Tested GPU: I also tested my old 3090 FE on this system, and the POST Error persists.

Comprehensive Troubleshooting Performed on the TRX40 System:​

Action Taken	Result & Implication
New CR2032 Battery	CMOS Error persists. (CMOS retention circuit is likely faulty).
Relieved CPU Cooler Tension	Loosened cooler screws. Failure persists. (Not a socket pressure issue).
Removed All Storage	Removed ALL NVMe drives & unplugged all SATA cables. Failure persists. (Not a drive conflict issue).
Minimal RAM Test	Tested with only one stick in DIMM_A1. The failure persists. (Not a simple RAM module fault).
Tested with Different PSU (430W)	Substituted the TX-1000 with a different, lower-wattage PSU. POST Error persists (This test is inconclusive due to the power constraint/missing GPU).

The Final Dilemma (Help Needed):​

1. PSU Uncertainty: The final test on the TRX40 was inconclusive. Given the catastrophic failure of the new 9950X3D system, I still cannot definitively rule out the Seasonic TX-1000 as the primary unstable factor in both failures. Should I wait for the TX-1000 RMA/replacement before concluding the Threadripper board is dead?
2. Zenith II Extreme Diagnosis: Given the TRX40 board's age and the consistent failure to POST even after every user-level fix, is the diagnosis of an unrecoverable motherboard hardware failure (faulty logic circuit) now confirmed?
3. Advice: What is the most critical next step: Wait for the PSU RMA, or try to find a cheap, low-power GPU to perform the final diagnostic test on the Zenith II Extreme?

Any advice on the highest probability component failure and the best path to get one machine running is appreciated!

 

[rare]

Is it worth trying to flash the BIOS on the TRX40 if it has that feature externally?

On the AM5 system it could well be another unfortunate case of an Asrock board killing a CPU..

 

[Tetras]

The Loop: The system now enters a rotation of "POST Error" $\leftrightarrow$ "CMOS Error" $\leftrightarrow$ "Code 02 (HDD detect)" with no display output.

Initial Fix: Had to update the BIOS via FlashBack to fix a "Load VGA BIOS" error.

One of the fixes online to a code 02 was to reflash the old BIOS. Since you recently flashed it, that one seems a strong possibility?

It was also mentioned that setting the PCI-E gen manually to a lower one and enabling above 4G decoding helped, but if you can't get in the BIOS I guess that's no use.