My main Windows 11 PC has developed a strange storage problem. A 1TB Seagate ST1000DM010 HDD becomes nearly unusable: games load very slowly, assets appear late, applications fail, and Steam sometimes reports installation errors. Windows often shows 100% active time while throughput is only a few MB/s, and the drive may repeatedly spin down and back up.
CrystalDiskInfo reports the drive as healthy at 29°C. I have tried different SATA data cables, different motherboard SATA ports, reseating everything, and a USB-to-SATA adapter. The drive works normally when connected to another, older PC. I previously had almost the same issue with an SSD: it was extremely slow on this computer, but worked normally when tested elsewhere. Reinstalling Windows several times did not resolve that SSD problem. The current 120GB system SSD still works normally in this PC.
The issue appeared around the time I upgraded the CPU without reinstalling Windows. CHKDSK /scan has now found serious NTFS corruption, including damaged file structures and crosslink information. Windows is asking to reboot for an offline repair, but I am concerned that repeatedly repairing the filesystem or reinstalling Windows will only mask the real cause.
The system uses an Intel i5-9500F, 16GB RAM, an H310M H 2.0 motherboard, an RTX 3060, one HDD, and two SATA SSDs. What could make several different storage devices perform badly on one particular PC while working correctly elsewhere? Could the cause be the motherboard or SATA controller, BIOS or chipset firmware, CPU installation, power delivery, drivers, or something else? I would especially appreciate diagnostic steps that distinguish a Windows problem from a hardware or power problem.
3 Answers
Use controlled tests to separate Windows from hardware. Boot a Linux live USB or a diagnostic environment and copy a large test file while monitoring transfer speed and errors. If the same drive is slow there, Windows and its services are unlikely to be the cause. You can also temporarily disconnect every nonessential drive and test one known-good drive at a time.
Try another SATA power connector or a different PSU cable, not only another data cable. A marginal power connection can cause resets, spin-up problems, and filesystem damage. If possible, test with a known-good power supply. Do not treat a 100% active-time reading by itself as proof of a failing disk; it can simply mean requests are waiting on retries or timeouts.
Inspect the CPU replacement carefully as well. On this platform, a disturbed socket pin, cooler pressure problem, or BIOS settings left in an unstable state can affect PCIe or chipset-connected devices. Load BIOS defaults, verify the memory is stable, and run a memory test. Then check whether the board has the latest BIOS available for the installed CPU.
If drives remain slow or produce corruption in a live environment after BIOS, chipset, power, and memory checks, the motherboard or its SATA controller becomes the leading suspect. A PCIe SATA controller can be a useful temporary comparison, though it will not rule out PSU or broader board problems.
The first priority is to back up anything important before running more repairs. Filesystem corruption can be a symptom of unreliable communication, not necessarily a failing disk. Check Event Viewer for Disk, Ntfs, storahci, and controller errors, and look for CRC, reset, timeout, or I/O messages. Also check each drive's complete SMART data, especially interface CRC errors and uncorrectable sectors.
Because the problem follows the computer rather than the drives, update the motherboard BIOS, Intel chipset and storage drivers, and any SSD or HDD firmware. A firmware or micro-timing problem in the board's SATA path can cause low throughput and corruption even when SMART says the drive is healthy.
The problem is intermittent and has briefly fixed itself before returning. The system is an i5-9500F on an H310M H 2.0 with Windows 11, 16GB RAM, an RTX 3060, and SATA storage. I will check the board firmware and event logs before assuming CHKDSK is the solution.