Intel hardware malfunctioning, unable to start properly or stop functioning at all
Intel hardware malfunctioning, unable to start properly or stop functioning at all
Before diving in, keep in mind I’m familiar with the setup and just need some external perspectives on my diagnosis. Here’s a quick rundown of what I’ve checked:
- CPU: i7-11700k, no overclocking
- Cooler: Cooler Master ML240 (240mm CLC)
- RAM: Two sets of 2x8GB Corsair Vengeance RGB Pro CMW16GX4M2C3200C16 (functioned with and without XMP enabled)
- Motherboard: MEG Z590 Tomahawk Wifi
- GPU: EVGA RTX3070
- Power Supply: EVGA 850GA (also tried with Corsair RM1000x)
- SSD: 2TB Crucial P5
- Case: Corsair 4000D Airflow
The main issue started with the system freezing and crashing after a few minutes. It would then fail to even POST. My troubleshooting steps included:
- Running Memtest86 on each RAM stick (single, pairs, all four) and verifying slot compatibility.
- Checking temperatures – everything normal.
- Confirming GPU setup with three 120mm fans in a high-airflow case.
- Swapping the motherboard to ASUS TUF Z590 Plus Wifi; now it boots but only with one or two RAM sticks. Four or more cause POST failure.
- Testing RAM on another system – all four passed simultaneously on my TUF X570-Gaming Plus Wifi with R9-5900x.
It seems the problem might lie in the CPU’s memory controller. I’m considering an Intel RMA, but the system boots to Windows with the new board and two RAM sticks, then crashes into a BSOD after about 20 seconds.
I usually resort to fixing things myself. Here’s what I’ve done before: What does the BSOD indicate? Have you saved any diagnostic files? Just because Memtest succeeds doesn’t guarantee the hardware is solid. Have you attempted a Linux live USB or a clean Windows installation? That’s unlikely, but plugging it into another power source could help. A power surge would be unexpected but possible. Have you tried booting without any USB devices? Also, did you skip the GPU during startup?
Prior to the Mobo swap, minidumps showed HIDCLASS.SYS issues and the USB Human Interface Devices driver was failing. I’m using WinDbg for analysis and haven’t attempted a clean Windows install due to POST problems. I’ve also tried booting from Windows 10 PE (based on 1909) and Ubuntu 22.04 via USB, both working fine. All systems run through APC UPS, and I’ve tested with and without the GPU. The PSU has been changed as mentioned. Files provided: 050523-22453-01.dmp, 051123-17281-01.dmp
Super cool, I thought you put in a lot of work, but it’s always nice to confirm you’re reading right. You should be able to start and run these without issues, right? No unexpected crashes mean there might be a problem with the Windows setup. Checking your crash logs shows HIDCLASS appears in both, which could point to a faulty mouse or keyboard driver. It might help fixing things with SFC/DISM, and a checkdisk scan is likely needed too. Trying safemode could also resolve it. If everything checks out, you may need a replacement drive to install the fix.
I sent the data using three sticks, then opened BIOS and spotted a detail I’d overlooked. The stick in Channel 1 Slot A had a rating of 2,133 while the one in Slot B was 2,600. As expected, the kits varied—both came with 3,200 CL16 XMP profiles from Corsair, but each had distinct JEDEC clocks and timings. I put the 2,133 sticks into Slot A on both channels and the 2,666 into Slot B so the system defaults to the lower clock speed. Now I’m running another full 4x8GB Memtest 9.4 test. If it completes successfully, I’ll try again with XMP enabled. After that, I’ll attempt to prevent the BSOD in Windows. The most frustrating part was when the system worked until the crash began—RAM was in XMP mode and everything seemed normal until then.
These dump files consistently reference HIDCLASS.sys, suggesting the CPU isn't the issue. HID typically relates to USB, but it can also involve other interfaces. You might be able to turn off the WiFi card in the BIOS settings. The BT component connects to HID, making it difficult to remove without affecting functionality.
Reversing the DIMM arrangement resolved the issue. Restarted in Safe Mode and turned off Procs and Services to get back into Windows. After updating drivers, I left the system overnight during stress tests without any issues. I saved the stable configuration as a user profile in the BIOS, enabled XMP, and ran the tests again successfully. I'm really glad I didn't have to return the CPU for replacement. Thanks for your help—I just needed some outside advice to verify this.