I built a new PC with an Intel Ultra 7 270K Plus, an RTX 5080, 32GB of Corsair DDR5-6000 RGB RAM, and an ASUS Z890-P WiFi motherboard. I'm getting recurring blue screens and application crashes, and MemTest86 points to memory instability.
With both RAM sticks installed in the recommended slots, Test 3 quickly reaches 10,000 errors, all associated with bit 4000. Each stick passes when tested alone, and testing individual sticks in the available slots also appears clean. However, using both sticks in slots 1 and 3 produces the same errors.
Reseating the CPU temporarily made the errors and crashes disappear, but after about three days the same MemTest86 failures returned. This has happened repeatedly, so I'm wondering whether the problem is a defective memory kit, a motherboard slot or trace issue, bent CPU-socket pins, excessive socket or cooler pressure, or a problem with the CPU's integrated memory controller.
I have several BSOD dump files available for analysis. Before I start replacing expensive parts, what would be the best way to isolate the faulty component?
3 Answers
Run the tests methodically with one stick at a time in every slot, then repeat with the other stick. After that, test both sticks at stock settings rather than DDR5-6000/XMP. If both sticks pass alone and fail together even at default settings, the likely suspects are the motherboard’s dual-channel path or the CPU memory controller, not simply one defective RAM module.
Check that the kit is on the board’s memory compatibility list and that the modules are installed in the recommended paired slots. Load BIOS defaults, disable XMP and any memory overclocking, and retest for several complete passes. If it still fails, the most useful next step is swapping one component at a time—ideally testing the RAM in another compatible system or trying a known-good two-stick kit in this board. Replacing parts without isolating the failure could get expensive quickly.
Since each stick passes by itself but the pair fails, the issue may not be a bad DIMM. It could be the motherboard’s memory channel, the CPU’s integrated memory controller, socket contact, or cooler pressure on the CPU. Inspect the LGA socket carefully for bent pins, make sure the CPU and cooler are seated evenly, and avoid overtightening the cooler. Also update the motherboard BIOS and test with XMP disabled at default JEDEC settings before changing hardware.
I have tested the sticks individually in every slot, and the errors only appear when both are installed. Reseating the CPU fixes it temporarily, which makes me suspicious of the socket or contact pressure.
The kit is running at 6000 MT/s, so I’ll repeat the testing with XMP disabled and verify the BIOS version before deciding whether to RMA the board or CPU.