Ryzen 9 7950X failed after installing thermal paste (Need help from PLS).
Ryzen 9 7950X failed after installing thermal paste (Need help from PLS).
I upgraded my computer to a R9 7950x with Gigabyte X670E Aorus Master and Corsair 32gb (2x8) DDR5 during the initial setup (about two weeks ago). During the process, I realized the thermal paste was depleted. There was a small amount left, but it didn’t seem sufficient, so I used what remained and purchased fresh paste later. The system functioned properly; temperatures stayed normal for that CPU—around 95°C at full load, dropping to 5.35 GHz when using AIO. I encountered some challenges with AMD’s EXPO mode, which required the RAM to hit 5600MHz. After enabling it, the boot process displayed only 24GB instead of the expected 32GB. Once I rebooted and re-enabled EXPO, everything worked correctly at the higher frequency.
A few days prior, I ordered a Noctua NT-H2 thermal paste. It arrived yesterday, and upon waking, I powered on the PC for roughly ten minutes before shutting it down again. After waiting about half an hour for cooling, I removed the AIO cooler, cleared the old paste, applied the new one, reassembled everything, restarted, and immediately noticed a problem: I hadn’t connected the power and data cables for the AIO pump. I quickly powered off the system (it ran briefly for under five seconds), verified all connections were secure, and then turned it back on. Everything functioned normally afterward—fans spun, display appeared correct, no error codes showed up. The numeric panel didn’t list any POST codes, and the Q-LED transitioned smoothly from DRAM to CPU every 10–20 seconds.
After several attempts, I resolved the issue by ensuring proper power connections and confirming all components were correctly seated. The board was stable, and no warning lights appeared. I also double-checked BIOS updates, cleaned all RAM slots, used a clear CMOS reset, and even disassembled the backplate to inspect the CMOS battery. Despite thorough checks, the system remained unresponsive without any visible faults.
I’ve tried various fixes—tightening the AIO, using different RAM configurations, flashing BIOS, removing the GPU, and even booting without RAM. Still, I suspect a hardware fault, possibly related to the RAM or thermal management. The low-pitched noise during operation was unusual but might have been from the voltage regulator on the DDR5 sticks.
I’m unsure if I’ve made any critical mistakes, but given how far this goes beyond my first upgrade, I’m concerned about the reliability of the build. If I’m stuck, should I consider returning it for a repair or replacement?
The postcode field remains empty and doesn't display valid codes. It might be stuck at 00 due to socket issues or power delivery problems. A dirty VRM on a P8z68 Deluxe could cause the board to show blank postcode, leading to CPU coldness when touched—unless it's a low-power chip or you quickly shut off the PSU. After water exposure, the CPU may be completely dead (rare), usually due to 00 indicating CPU ded. Posture: check for common causes before extreme measures.
It seems the display remains blank and unlit. I suspect the VRMs might be dirty, but the board is only a few weeks old and it worked before taking the cooler apart. I’m surprised I messed up the socket—wouldn’t a damaged socket stop Q-Flashing? If the CPU is fully removed, the board behaves differently.
Capture an image of the socket. Occasionally boards fail unexpectedly, which isn't surprising.
I can remove it from the socket too, though I made sure everything was in good shape a few hours back.
Remove the CPU to inspect the pins, and capture a photo of the bottom surface including the golden pads.