My computer randomly hard-freezes or powers off, even while idle. The most reliable way to trigger the problem is opening the Smithbox modding tool and inspecting a very large regulation.bin table. I have reinstalled Windows 11, checked the boot drive's S.M.A.R.T. data, run a CrystalDisk benchmark, and tested the CPU with Prime95. Prime95 usually passes but has crashed once in two runs. MemTest86 also caused a complete shutdown once, although it passed on another run. I initially suspected the CPU, motherboard, or power supply and planned to try another PSU. After testing the memory further, each RAM stick works by itself, and pairs work correctly in either dual-channel arrangement. However, the crashes only occur when all four sticks are installed. What does that point to, and how can I narrow down whether the issue is the RAM, motherboard, memory controller, or settings such as speed and voltage?
3 Answers
Test the system methodically with the complete four-stick configuration: run MemTest86 for several full passes, then try a lower RAM frequency and slightly more conservative timings with XMP/EXPO disabled. If it becomes stable only at a reduced speed, the system may simply be marginal with four modules, or the CPU’s integrated memory controller may not handle the current configuration reliably. If errors or shutdowns continue at stock settings, inspect the motherboard’s DIMM slots and consider testing with a known-good matched memory kit.
A shutdown during a memory test is a strong sign that the memory subsystem is unstable, even if MemTest86 passes on another run. Since every stick works individually and pairs work in either dual-channel arrangement, the sticks themselves may be fine. Running all four increases the electrical load on the memory controller and can expose a motherboard, CPU memory-controller, BIOS, or memory-speed issue. Try loading BIOS defaults, disabling XMP/EXPO, and testing all four sticks at the motherboard’s standard JEDEC speed. Also make sure the RAM is installed in the slots recommended by the motherboard manual and that the BIOS is current.
That fits what I’m seeing: the failure only appears with all four modules installed. I’ll test with default memory settings and a lower speed before replacing other parts.
Before blaming the power supply, check the motherboard’s BIOS memory settings and confirm that the installed modules are a matched kit with compatible specifications. Four DIMMs can be much harder to run than two, especially if they are mixed kits. Reset the CMOS, use conservative settings, and test stability after each change. If the machine still powers off during memory testing at default settings, the motherboard or CPU memory controller becomes more suspicious; swapping in a known-good PSU is still worthwhile because an abrupt power loss can also come from power delivery.
The fact that both pairs work separately but all four fail makes a defective individual stick less likely. It could still be a marginal slot or memory controller, so testing the modules in different slot combinations and using one matched kit would help isolate it.