Did you ever wonder how your hard disk drive works?

Without it, your computer would feel empty inside

Magnetic hard-disk drives are used to store most of the data accessible by personal computers and workstations, as well as much of the data being processed by large enterprise servers. The data is stored digitally as tiny magnetized regions, called bits, on the disk. A magnetic orientation in one direction on the disk could represent a "1", an orientation in the opposite direction could represent a "0". Data is arranged in sectors along a number of concentric tracks. These tracks are arranged from the inner diameter of the disk to near its outer edge. Disk drives may contain more than one disk in a stacked assembly. Data is written onto each disk surface (top and bottom) by a separate recording head. So a disk drive with three disks will usually have six separate recording heads.

IBM and the
history of the
hard drive

Visualize MR and
GMR Heads in

Observe the
physics of GMR
in motion

You're not the only one who can read and write

In modern disk drives using the IBM innovations of magnetoresistive (MR) or giant magnetoresistive (GMR) heads, the bits are written and read by separate elements in a recording head as it flies over the spinning disk. A writing element writes bits onto the disk and a reading element read the bits by detecting the presence of their faint but tell-tale magnetic fields. The head itself is attached to a slider, an aerodynamically shaped block that allows the head to maintain a consistent flying height above the disk. In turn, the slider is connected to a suspension arm that is controlled by an actuator which can move the head to any track of bits on the disk, from the inner to the outer diameter. Special electronic circuits encode data from the computer's processor prior to writing, and decode the bit pattern after reading. Additional vital electronic circuits keep track of where data is stored on the disk so that it can be readily retrieved when needed, and monitor the motion of the disks and heads so their positions over the disks is always known precisely.

Keep reading -- this is the juicy part

When a command is made to store some data on a disk, the following chain of events occurs:
  • The data flows into a cache where it is encoded using special mathematically derived formulae, ensuring that any subsequent errors caused by noise can be detected and corrected.
  • Free sectors on the disk are selected and the actuator moves the heads over those sectors just prior to writing. (The time it takes the actuator to move to the selected data track is called the "seek" time.)
  • Once over the data track, the heads must not write the data until the selected free sectors on that track pass beneath the head. This time is related to the rotation speed of the disk: the faster the speed, the shorter this "latency" period.
  • When it's time to write, a pattern of electrical pulses representing the data pass through a coil in the writing element of the recording head, producing a related pattern of magnetic fields at a gap in the head nearest the disk. These magnetic fields alter the magnetic orientations of bit regions on the disk itself, so the bits now represent the data.

When a command is made to read some data on a disk, a similar process occurs in reverse. After consulting the table of stored data locations in the drive's electronics, the actuator moves the head over the track where the chosen data is located. When the correct sectors pass beneath the head, the magnetic fields from the bits induce resistivity changes in the sensitive MR or GMR materials located in the reading elements within the head. These elements are connected to electronic circuits, and the current flowing through those circuits change with the resistivity changes. The current variations are then detected and decoded to reveal the data that had been stored on the disk.

Faster, smaller, cheaper, better -- we're up to the challenge

As manufacturers such as IBM improve the capacity and performance of disk drives, each of the elements involved must be improved. To name just a few: magnetic materials on the disk must be made finer so smaller bits can be written and still be read; read elements must be made more sensitive so they can read the smaller bits; smooth operating motors and lubricants must permit the disk to be spun faster to reduce latency; the actuator's electrical and mechanical systems must be improved to position the head accurately over narrower tracks in less time (to reduce the seek time); the disk drive's electronics must be upgraded to manage -- without errors -- the increasing torrent of data flowing in and out of the heads when smaller bits pass by at faster speeds.

[GMR Lead Story | IBM Research home page]

[IBM home page | Order | Search | Contact IBM | Legal]