A dying drive doesn’t wait for you to read reviews. Every extra hour of Windows on a clicking disk is a worse clone. Copy the data off, then replace the disk. Not the other way around. I’ve watched people shop for the perfect NVMe while the only copy was still clicking.
How you know
- SMART: reallocated or pending sectors climbing. CrystalDiskInfo on Windows is enough to read the flags.
- The drive pauses for seconds at a time. Explorer windows that freeze, then catch up.
- Mechanical click or a long BIOS hang on the same disk.
Clone while it still enumerates
Use the enclosure and a clone tool that retries bad sectors. ddrescue on a Linux USB is the serious version. A commercial clone utility is fine if the disk is only starting to fail. Don’t defragment. Don’t run CHKDSK “to fix it” first. That can make a weak disk worse. Copy, then replace.
On the bench
External SSD, at least as large as the dying disk
Landing zone. A cheap spinning USB disk is another mechanical failure in the path. A 1TB external SSD is fast enough that I’m not babysitting a 12-hour copy on a disk that may not last 12 hours.
- Capacity ≥ the used space at minimum; ≥ the whole disk if you want a full clone.
- After the emergency, it stays useful as a backup target.
When the copy finishes, take the old disk out of daily service even if it “still works.” Pending sectors don’t go back down. Then follow Upgrade a laptop SSD without losing Windows and clone onto a new internal from the good copy.
Verdict
Read SMART, copy off immediately to an external SSD, then replace the internal. Shopping for the perfect NVMe while the old one is still the only copy is how data disappears.