Click of death is a term that became common in the late 1990s referring to the clicking sound in disk storage systems that signals a disk drive has failed, often catastrophically.
The clicking sound itself arises from the unexpected movement of the disk's read-write actuator. At startup, and during use, the disk head must move correctly and be able to confirm it is correctly tracking data on the disk. If the head fails to move as expected or upon moving cannot track the disk surface correctly, the disk controller may attempt to recover from the error by returning the head to its home position and then retrying, at times causing an audible "click". In some devices, the process automatically retries, causing a repeated or rhythmic clicking sound.
The phrase "click of death" originated to describe a failure mode of the Iomega ZIP drives, appearing in print as early as January 30, 1998. In his podcast of September 18, 2008, Mac journalist Tim Robertson claimed to have coined the phrase in the early '90s. The phrase was then applied to other drives exhibiting a similar unusual clicking sound usually associated with failure.
Iomega Zip drives were prone to developing misaligned heads. Dust inside the Zip disks or dirty heads caused by oxide build-up could misalign the heads, but in newer devices it was due to poor quality control and manufacturing defects in the drive itself. Magnetic fields could also cause the drive heads to become misaligned, as the drives were not internally shielded from external magnetic fields. The heads caused the data on the cartridge to become misaligned, rendering it unreadable. The Zip cartridges also wore out, grew defects, or otherwise lost all four 'Z tracks'. If all four redundant 'Z tracks' were lost, the cartridge was unusable, since retail drives are unable to low-level format the disks.
With a malfunctioning drive or cartridge, the drive heads tried to read the Zip disk, but could not find a good Z track or hit a bad spot during a read operation. As part of the drive's retry program, the controller would quickly snap the head arm back into the drive and out again, producing a specific number of 'clicks'. This happens each time a data request fails: the drive parks the heads to calibrate them and presumably clean them. Parking and relaunching continued until the data had been read or a set number of attempts had been reached.