Problem occurs when the system begins or restarts, causing a freeze.
Problem occurs when the system begins or restarts, causing a freeze.
Hello, I recently assembled a workstation PC for our print shop. Recently, the system would freeze during various startup phases. It functioned normally for a few days. After building, I performed MemTest with XMP and disabled settings—both passed without issues. I ran OCCT extensively for an hour, and it operated smoothly. I also executed a System Platinum stability test lasting 12 hours, which completed successfully. The certificate link is available if needed. We used the machine for three days before testing AMD AI bundles, specifically Amuse, which proved useful for certain tasks. After installing a few more models, we began with the most demanding one (estimated VRAM requirement: 24-26GB). Since our card supports 24GB, shared memory was activated, which resolved the problem. Around 30 seconds in, the PC froze; the GPU temperature reached about 50°C, and shortly after, the screen went black—displaying a full-screen black rectangle without shutdown. Five minutes later, after rebooting, it started normally but then locked up again, showing a black screen. We tried restarting multiple times, but the issue persisted. Eventually, we suspected Amuse was handling resource-intensive tasks and opened Task Manager to check for activity. However, no anomalies appeared. Drivers were suspected, so I downloaded DDU and booted into safe mode to uninstall them. In safe mode, the PC froze during startup, then briefly after a few moments before restarting. After several reboots, it returned to normal operation. During those restarts, the PC displayed A0 on the POST screen, which is not an error but a status code. Once we exited safe mode, it booted normally. At some point, the PC froze during the "Welcome" screen transition, showing code 06—a potential error. We attempted to clear CMOS, but it didn’t help. The issue persisted even after clearing. After replacing the GPU with a working RX560, the problem continued. We then tried booting into Memtest and confirmed no freezing during that test. Since the AI relies heavily on GPU performance, we considered the possibility of GPU failure, though this seemed unlikely. We transferred the GPU to another system, which worked, eliminating GPU as the primary cause. The PC would freeze again when transitioning from the welcome screen to desktop, suggesting a driver or configuration issue rather than hardware. BIOS adjustments—such as disabling PWM fan control, changing USB4 settings, and modifying BIOS settings—did not resolve the problem. We also reset CMOS and disabled iGPU in MSConfig. The issue remained consistent across reboots. After several attempts, we reached a point where the PC would freeze at startup, specifically during Windows loading. This occurred before the "Welcome" screen appeared. The error code 06 was observed. We tried clearing CMOS again but saw no improvement. Eventually, we returned the GPU to our system and continued experiencing freezes. Now, to investigate further, I considered whether Windows itself might be causing instability, especially since GPU usage was critical.
There should be a small LED above the display when the system starts up. Consider checking the connection under the CPU, as some pins could be damaged or melted from overheating. You might need to take out the CPU and examine the CPU pads and the motherboard slots closely. Look for any signs of burn marks, discoloration, or misaligned pins. To confirm, try installing a single stick of RAM to rule out that issue. You’re likely familiar with clearing CMOS—remember, when doing so, disconnect the PC from power, either by shorting the relevant pins or removing the battery and briefly pressing the power button.
The display lacked LED illumination when the PC froze, only a red and yellow indicator appeared before cycling through codes—this is typical. I’ll check the CPU next and test with a single stick of RAM. Regarding CMOS, I cleared it using the back button marked "clear CMOS," but the settings reset afterward. On this board, I can’t find the battery; it might be hidden under the chipset or within the VRM/IO heatsink.
It seems the cooling component is located under the chipset heatsink. However, since the board includes a specific button for it, it might not be necessary to touch it. I plan to reset the CMOS once more while the computer is off, as this step is clearly outlined in the user manual for this model. I expect it will function properly after resetting, but I’ll double-check before proceeding with any further complicated or costly actions.
I updated the CMOS by holding the "clear CMOS" button for 20 seconds. The PC had been unplugged overnight, and now it displayed a message confirming the CMOS was cleared. However, the settings reset unexpectedly. The system didn’t start properly, and even basic auto-repair froze. The freezing seemed to occur more frequently after each restart. I tested with only one RAM stick, which finally allowed the desktop to load after several restarts. Then I swapped RAM sticks, tried different slots, but nothing resolved it. Using just a single stick helped me reach the desktop after many restarts. I suspect the memory controller might be struggling or there’s an issue with its stability. I checked the BIOS and adjusted the NB/SOC voltage between 1.12 and 1.24, but increasing or decreasing it didn’t improve performance. Lowering it slightly helped for about 15 minutes before freezing again. After some web browsing, setting the voltage to 1.14 and then 1.12 allowed me to work for around an hour. Now even at 1.00, the system stalls at the Windows loading screen, and booting from a USB also caused freezing.
I have another machine running an AM5 platform. I tested the components you mentioned: a Ryzen 5 7600x from MSI B650 Tomahawk Corsair RGB with 32GB of RAM (2x16 at 6000Mhz). I swapped the Gskill RAM for Corsair and managed to boot Windows. The PC ran smoothly for a few hours, stayed on overnight without problems, and started again in the morning. However, when I tried more demanding applications like Amuse, it froze completely. On another machine with the same X870E motherboard, running Gskill RAM worked mostly well except for occasional issues with that specific setup. This suggests the problem might be related to the CPU rather than the RAM itself. I replaced the 9950X CPU, but it still experienced random freezes during startup and would not even boot at all, only showing red and yellow lights. The other X870E unit worked fine with both RAM types and the 7600X processor. So I’m leaning toward the CPU or possibly the NB/SOC module being faulty, and we might need to return it for repair. Regarding Corsair RAM on the B650 system, I haven’t encountered any issues so far—maybe 32GB is less taxing than 96GB for the IMC unit. The problem appeared only after loading tasks.
It looks like the system might be reaching the end of its life. Double-check that the B650 model has BIOS version 7D75v1H or higher for this chip. I couldn't locate a clear CPU compatibility guide, and the MSI website doesn’t seem reliable. The reference link provided seems to be useful.
Yes, the most recent BIOS updates are available for the B650 and X870E models. There are typically 2 to 3 versions, though it can be a bit unclear which one is the exact latest at any given time. The system was even newer than that, being nearly half a year older when it came out.