My hard drive started making strange noises and now parts of it show bad sectors. I have important photos and work files on it, and I’m scared using the wrong recovery method could make things worse. What’s the safest way to recover data from a hard drive with bad sectors without causing more damage?
If your drive is throwing cyclic redundancy check errors, or File Explorer stalls forever when you open a folder, stop using the drive first. I learned this the hard way. I kept retrying reads on one old laptop disk and it got worse within an hour.
If it is your system drive, shut the machine down. If it is external, unplug it. Leaving a failing disk powered on gives Windows and the drive firmware more chances to hammer the same bad area again and again. Sometimes what starts as a file system issue turns into a hardware one. I have seen a drive go from ‘some files won’t copy’ to ‘not detected at all’ after too many retries.
What is usually going on falls into two buckets.
Some bad sectors are logical. The magnetic surface still reads, but the stored data no longer matches its error correction info. Power loss, forced shutdowns, or flaky USB connections tend to show up here. Those cases are less awful.
The other type is physical damage. Worn platter spots, weak heads, head crashes, stuff like that. Software will not repair physical damage. At best, it marks the damaged area so the drive stops trying to use it.
One thing I would not do at the start is run CHKDSK. People reach for it because it is built in and feels safe. It is not a data rescue tool. It is a file system repair tool. Different goal. If the directory structure is damaged, CHKDSK might clean things up by removing broken entries, and your files get harder to recover after tht.
The safer order is this:
What I would do first
- Get another healthy drive with enough free space.
- Make a full clone or image of the failing disk, byte for byte if possible.
- Put the original drive aside.
- Run recovery scans against the clone or image, not the original.
- Recover found files onto a third drive, not back onto the image source.
That one change, working from a copy, lowers the risk a lot.
DIY route that worked better for me
For this kind of job, I had better results with software built for unstable disks. I used Disk Drill on a drive with read errors scattered across the disk.
What stood out to me was the backup side, not the deleted-file scan stuff people usually talk about. On a bad drive, the important part is how the tool behaves when it hits unreadable blocks. Some programs hang, slow to a crawl, or keep hitting the same dead spots. That is rough on the drive.
With newer builds, especially version 6, the imaging process felt smarter. It skipped unreadable areas first, pulled easy data off fast, then circled back later with smaller reads on damaged regions. That matters. Fewer long stalls. Less stress on the drive motor and heads. More files copied before the disk gives up.
Basic workflow
This is the sequence I would follow:
1. Prepare storage
Use a healthy destination disk with enough capacity for an image or clone.
2. Create the image
Run the backup or imaging function on the failing drive.
3. Mount or attach the image
Scan the image file for recoverable files.
4. Export recovered data elsewhere
Save recovered files to another separate disk. Do not write them onto the original drive. Do not write them onto the same failing source path either.
When software stops being the answer
If the drive clicks, grinds, beeps, spins down, or does not spin up at all, I would stop there. Same if BIOS stops seeing it, or it disconnects over and over even with a known good cable and power source.
Those signs usually point to mechanical failure. DIY tools do not fix failed heads or seized spindle issues. If the files matter, family photos, tax docs, client work, old projects you do not have anywhere else, a lab is the safer move.
Places like Gillware or Techchef exist for this reason. They have cleanroom setups and hardware tools normal users do not. They swap heads, stabilize damaged drives, and read media from disks a home PC won’t even mount. It is expensive. I have seen quotes land anywhere from about $500 to $3,000, depending on damage and parts.
After the recovery
Once your files are copied out and you have opened them to verify they are good, then you can mess with the old drive if you still want to. For throwaway use only, I mean.
A full format is better than a quick one because it forces a broader read and write check. You can also run chkdsk /r in PowerShell or Command Prompt to mark bad sectors so the file system avoids them.
I still would not trust a drive after CRC errors and hanging reads. In my experience, once a disk starts acting sick, it tends to stay sick. Sometimes it limps along for months. Sometimes it dies by dinner.
Backup habit worth keeping
The fix for this mess is backups before the mess starts. I stick to 3-2-1 now:
- 3 copies of your data
- 2 different storage types
- 1 copy off-site or in the cloud
It takes some setup. It is still cheaper than recovery, and a lot less miserable.
Strange noises changes the math. Bad sectors alone are one thing. Noise plus bad sectors often means the drive is failing at the hardware level. I would be more cautious than @mikeappsreviewer on one point. If the files are irreplaceable, skip DIY first and price a lab before you do more reads.
My order would be:
- Stop booting from it.
- Do not run scans, repairs, defrag, or surface tests.
- Put a price on the data. Family photos and work files matter more than the drive.
- If the drive still mounts, copy only the highest value folders first. Small batches. No giant drag-and-drop job.
- If reads start slowing, freezing, or making louder noises, stop.
Why I say copy priority files first. Imaging a dying 2 TB drive can take many hours. A weak head sometimes dies before the image finishes. I have seen people save nothing because they chased a perfect clone. I’d grab Documents, Desktop, Photos, project folders first if the drive is still semi-readable.
If you want software, Disk Drill is fine for a second pass after the must-have files are off. Use it to preview and recover from what the file system no longer shows. Save output to another disk. Never back to the sick drive, obviosuly.
Also, swap the easy stuff first. New SATA or USB cable, different port, different enclosure, direct motherboard SATA if it was in a cheap USB dock. I have seen CRC errors come from a bad bridge board, not the disk itself.
This explains hard drive bad sectors in plain English, how to deal with bad sectors on a hard drive before data gets worse.
If the drive clicks, chirps, or drops out of BIOS, stop there. Lab time. Trying five more tools is how people turn a recoverable drive into e-waste.
Strange noises is the part that worries me more than the bad sectors, honestly. I mostly agree with @mikeappsreviewer and @suenodelbosque, but I’d push one extra point: don’t keep power-cycling the drive to “see if it works this time.” That can finish off a weak head real fast.
My take:
- If it clicks, chirps, or disappears from BIOS, stop DIY.
- If it still reads quietly, avoid long browsing sessions in File Explorer.
- Use a write blocker if you have one, or at least mount it read-only in another machine/Linux live USB.
One thing not mentioned enough is heat. Keep the drive cool during recovery. A fan pointed at it sounds janky, but it can help on borderline drives. Also disable sleep/hibernation so the job doesn’t get interrupted mid-read.
I slightly disagree with “copy priority folders first” as a universal rule. For a drive with scattered weak sectors, targeted copying can cause lots of random seeking. Sometimes a controlled image pass is gentler. Depends how bad it is, which is why there’s no perfect script here.
For software, Disk Drill makes sense after you stabilize the situation, especially if you need to recover files from an image instead of poking the original disk over and over. That’s the safer use case to me.
Also, if you’re trying to choose data recovery software for personal use, this guide to choosing the right data recovery software is a decent starting point.
Big thing is this: don’t “repair” first. Recover first. Repair later, maybe. And yeah, if the data is worth more than a few hundred bucks, lab time is probly the least expensive mistake.
