Rig started crashing - possible hardware issue?

Danny Donkey · May 2, 2021

My old but stable and functional rig with FX 8350, 16 GB Memory and Zotac Geforce 970 have become dated and underperforming, so I started buying parts for a new rig a couple years ago which I build last month. Surely the core components have become somewhat dated, but are still an upgrade compared to my old rig.

Spoiler

Cabinet: Corsair Obsidian 1000D

CPU: Threadripper 1920X

MB: Asus ROG Zenith Extreme

Memory: Corsair Vengeance RGB PRO DDR4 2933MHz 64GB

Cooler: Noctua NH-U14S TR4-SP3

GPU: Radeon RX 590

PSU: Corsair AX 1200i (1200W)

OS SSD: Seagate FireCuda 510 500GB M.2

OS: Windows 10 64 bit

The rig had no issues to begin with, but have always booted up in a way that I found strange. Luckily the MB have a small LCD screen with text telling what it does. When clicking the start button the fans ramps up, and the system spends seconds checking memory without posting. Then suddenly the system plays dead for a couple seconds before starting up again with a new memory check, and then proceeding to CPU and the usual connected components before posting. This all takes like 15-20 seconds, but that might be how the old Threadripper systems are supposed to start.

The rig had it's first crash during a game three days ago, and I got PTSD reminding me of last time I had a faulty GPU because... well I have no money and GPU's are neither cheap nor available nowadays. Luckily it turns out that my RX 590 is not faulty as the same crash occurred when I had swapped it out with my working GTX 970. Anyway, the crashes are always sudden deaths - no blue screens or error codes, and no useful records in the Event Log, and have so far only happened when the computer have been in active use.

Right after this crash occurs I have to turn off the power switch on the PSU in order to be able to reset the system and turn it back on. And even then it often starts the procedure of checking memory, and crashes again.

I have also tested my memory sticks, and either each and single one of them are faulty, or none of them are. Sadly I only have this one MB with support for DDR4, so I haven't been able to confirm with another rig.

Does it exist any logging software that can help me diagnosing the issue my new rig is having? Right now I am going blind frightening that the MB itself or the CPU might be faulty meaning I have to buy a new core components, and I am living on unemployment benefits without spare money.

Juular · May 2, 2021

15 minutes ago, Danny Donkey said:

Right after this crash occurs I have to turn off the power switch on the PSU in order to be able to reset the system and turn it back on.

Something is tripping the PSU protections ... Have you tried any stress tests ? OCCT PSU stress test particularly. If it crashed any faster than usual then it's probably PSU overheating since there's no way your system trips it's OCP or OPP. It's possible that it trips UVP tho, check all PSU cables and connectors both on PSU and motherboard/GPU sides for damage, charring, melting etc.

Danny Donkey · May 7, 2021

On 5/2/2021 at 11:14 PM, Juular said:

Something is tripping the PSU protections ... Have you tried any stress tests ? OCCT PSU stress test particularly. If it crashed any faster than usual then it's probably PSU overheating since there's no way your system trips it's OCP or OPP. It's possible that it trips UVP tho, check all PSU cables and connectors both on PSU and motherboard/GPU sides for damage, charring, melting etc.

This build haven't been stress tested in any significant extent. I have played World of Warcraft: Shadowlands on ultra graphic settings, and it runs the game just fine - until it decides to die. After it started crashing I installed Cinebench to check if running benchmark would provoke a crash, but it ran fine during the tests I performed.

I have checked all cables and connectors that goes out from the PSU, and I couldn't find a single point of visible damage. Also made sure that the used cables actually supports the PSU in case the manufacturer were to send wrong cables without me detecting it during building, and I can confirm that all cables are indeed intended for my AX 1200i.

I will try to install a leftover functional AX 860 and see if that helps.