A steady click, once or twice a second
That's what most people call "click of death." When heads can't read their servo data they keep retracting & attempting to read the data indefinitely. It is a mechanical fault and won't fix itself.
Heads
The click is the head assembly retracting to its parking ramp; recalibrating and moving back out to try again – over and over. This is a loop not a one-off failure and each cycle costs you something. Turn it off before reading any further.
The drive cannot get a lock on its servo information so it retracts the heads and tries again. When the heads are damaged, each of those passes has the heads crossing the surface holding your data.
All content below is worth knowing. Nothing in the following compensates for the minutes the drive spends running while you are reading it.
A clicking drive is deteriorating at a rate that will be measurable, and the amount ultimately recoverable tracks closely with how long you run the drive after the noise begins. That is the only variable entirely under your control, and no equipment afterwards can make up for spending that time.
The best course of action here is to turn it off, unplug it, and let it sit. Don't try one more time to get your file. And don't even think about trying to connect it to another machine as a way to "see if that works." They're the same actions done repeatedly. And they'll each cost you.
That's mechanically precise. Which explains why you hear such an exact timing in the sound.
Inside every hard drive there are read/write heads floating just a few nanometres over spinning disks. The heads use servo data—positioning info placed on the surface of the disk by the manufacturer—to determine their location.
When a head can't read that servo data, the drive doesn't just stop working. Instead it pulls the entire head assembly back onto its parking ramp at the edge of the disk. Once safely parked, it resets itself using previously known positions. After resetting, it swings back out to attempt reading again. That act of pulling back and returning is what you hear as a single "click". Because the process is a continuous loop—seek, fail, reset, seek—it sounds like a constant repetition.
So that's why the noise is so distinct & repeatable. It's not a repeated breakage of the drive. It's just the drive repeating the process as many times as possible, using its best judgment.
You don't need all those words. Those are the descriptions we hear on the phone when someone calls with this problem.
That's what most people call "click of death." When heads can't read their servo data they keep retracting & attempting to read the data indefinitely. It is a mechanical fault and won't fix itself.
Drive attempted to access data, failed, and intentionally parked. This type of failure is often firmware/service area related, not head damage. So it has a better prognosis for recovery without disassembling the drive.
Platter head contact. Removing magnetic coating while running. Everything that was present in the damaged band is destroyed forever. The time when literally minutes matter.
Not the heads at all. Either a seized spindle bearing or heads are stuck to the platter because they have been sitting powered off for so long. Same instruction. Different failure.
Each has been suggested by some individual online, and each will make an expert recovery more difficult.
Some historical basis exists here. Cooling sometimes shrank a failed component enough on older drives to get one more spin. Anything modern produces condensation on the platters upon returning to room temp. Moisture where the heads operate is detrimental.
The idea is to unstick a stuck head. On a drive with bad heads, tapping/dropping drives those bad heads into the platter surface, converting a head fault into permanent scoring.
No matter how long your recovery software runs, it will never fix a physical problem. Running a deep scan is just hours of continuous reading from failing hardware which has a limited number of reads left. Drives regularly fail partway through a scan intended to save them.
The commonest and the most costly. No matter how much value that single file has to you, there's no way that one file can justify damaging every other item on your entire hard disk.
You'll see exactly what you're buying; and why it is not a desk job.
Before a drive is powered, its specifications are examined — family, firmware revision, head configuration — because a mismatched donor will make things worse instead of making them better. Locating a matched donor — same family, compatible firmware, correct number of heads — this can be the slowest process. This is also the reason an older or uncommon drive will take longer without being any less recoverable.
The head stack on the drive is swapped out with the internal filter changed to catch any debris which was generated by the drive's failure. The drive is then imaged read-only using short, controlled reads while setting timeouts for each individual head. The top priority of the read operation is the directory you requested — because a drive which has been repaired might provide one readable pass; therefore, every attempt at reading the drive should be made deliberate rather than exploratory.
All other operations occur against the disk image of the failed drive. At no point during these operations is there ever a write to the original disk. There is little expectation that the original drive will function again once all operations have completed. In Manchester this service will cost from £300 + VAT, after a free 48-hour diagnostic, with a 50% deposit because donor parts and bench time are committed before anyone knows what the platters look like.
The reason for the head clicking is due to the inability of the heads to receive servo information, which allows the drive to determine their position. Therefore, the drive has to retract the heads into the park position, perform a calibration cycle, then try to find the correct track. This retraction is what you hear as the "click," and it repeats because the drive is looping through the same failed seek.
It does not restore it to usability. The term recovery implies accessing the contents by removing the damaged head from the drive, replacing it with a matching donor head assembly, and accessing the contents once. After recovery, the drive is typically unusable. The primary focus during recovery is obtaining access to your data, not a working disk.
Do not use it. Every time the platter rotates, damage occurs. How long the drive operated after the first sign of failure can be the major determining factor of how much recoverable data exists. One area of influence over the amount of recoverable data is entirely under your control.
No. Condensation forms on the platters as the drive warms back up. Additionally, moisture located near where the heads move is destructive. In some cases on older drives manufactured decades ago, moisture may have been a concern. However, today it is an active hazard to your recovery efforts.
Generally better news. This type of behaviour can indicate that the drive has identified an issue and will park itself, pointing to either the firmware or the service area as opposed to damage to the mechanical heads, both of which can be recovered from without having to open the drive.
From £300 + VAT for a single drive. The cost is quoted in writing, after a free 48-hour diagnostic. A 50% deposit is required for donor part work, with the balance due only on success.