NVMe, SATA, USB — same habits, different blind spots
Hard Disk Sentinel reads through whatever link sits between the controller and your OS. Weak links masquerade as sick disks until you isolate them.
NVMe & M.2
Pads, screws, and thermal pads matter as much as SMART rows. Reseat after shipping trauma; note whether throttling correlates with sequential writes or tiny random bursts — each implies different fixes.
Driver stacks after major Windows upgrades can change error surfaces — capture exports before and after.
SATA & internal cabling
CRC and interface errors often precede realloc drama. Budget spare SATA cables for towers you open often — tension and dust behave like “slow disk disease.”
Label cables near hot GPUs or radiators; heat-softened clips loosen over years.
USB & docks
Translation layers reinterpret SMART — sometimes bluntly. Treat USB readings as “disk + bridge contract”: record chipset model, firmware feeling, and cable length beside serial numbers.
Same disk on motherboard SATA vs cheap enclosure can tell completely different stories.
Arrays and pools add another lens — correlate physical bay IDs with serials in Sentinel before you blame the wrong spindle in a NAS or shelf setup.