TUF GAMING A620M-PLUS with Wi-Fi and a 7600 processor, flashing the power button quickly when starting up.
TUF GAMING A620M-PLUS with Wi-Fi and a 7600 processor, flashing the power button quickly when starting up.
Hi, when I assembled my PC it started up properly with all default configurations (RAM at 4800 MHz, no BIOS updates). After performing some OCCT and memory tests, I encountered numerous errors that eventually led to a crash. Once I configured an Expo profile for 5600 MHz, everything functioned smoothly without any issues. Following the BIOS update, tests passed and performance remained stable regardless of whether Expo was enabled. The only persistent concern is the unusually long boot time—around fifty seconds—and the rapid flashing of the power button logo before the system loads, which I found unusual. According to the manual, this rapid flashing could signal a RAM or CPU problem (referred to as Q-LED CORE). Interestingly, despite this issue, the PC operates correctly otherwise. Would you think there might still be a RAM or CPU fault that needs addressing? Should I consider returning it for repair, or is this behavior typical for an AM5 platform? Thanks, David.
Slow boot on DDR5 AM5 systems are very well known and is extremely common. It is very likely that the flashing is just due to the memory training it has to go through every time it boots. If everything works fine while in the desktop, playing games or running stresstests I wouldn't worry.
It seems you're asking about running RAM at 4800MHz with auto settings. The advice suggests it should work, but if you notice odd behavior with Expo, consider disabling Expo and adjusting frequency, timings, and voltage manually. Setting primary timings directly may produce similar results while offering different performance.
I wasn't able to run it with the auto settings (4800, all auto) on the old BIOS that came with the PC. When I tried the OCCT CPU test and memtest, there were a lot of errors and it crashed eventually. Once I updated the BIOS to the latest version, no errors appeared at all—tested one hour of OCCT and think two memtest runs worked, but I’ll try again today.