4-D shock-sensing for hard-disk drives

Details 4-D shock-sensing for hard-disk drives 4-D shock-sensing for hard-disk drives

1999-11-19

2003-05-20

Hudspeth, David (Department: 2651)

Dynamic magnetic information storage or retrieval

Automatic control of a recorder mechanism

Controlling the head

C360S060000, C360S061000, C360S077080

Reexamination Certificate

active

06567233

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates to the field of hard disk drives for use in computer systems. In particular, the present invention relates to a technique for improving tracking of a write head and improving shock sensing in a hard disk drive system.
BACKGROUND OF THE INVENTION
Areal densities for magnetic hard disk-drives continue to grow aggressively in recent years, with no sign of abating in the foreseeable future. As areal density grows beyond 10 Gb/in
2
towards 100 Gb/in
2
, there is a need to shrink the bit aspect ratio from the present 15-20:1 down to about 4:1, in order to achieve sufficient head/media signal-to-noise ratio. Consequently, track densities are required to grow substantially more aggressively than linear densities as the storage industry progresses on the 60% (or higher) areal density growth curve.
As tracks-per-inch (TPI) grow from 25K TPI (10 Gb/in
2
) to 200K TPI (100 Gb/in
2
), issues related to data integrity in the presence of operating shock and the ability to servo precisely/quickly on tracks 0.12 &mgr;m wide will become substantially more severe. These issues are described, for example, in Chew K.K., “Control Systems Challenges to High Track Density Magnetic Disk Storage”,
IEEE Trans on Magnetics,
Vol. 32, No. 3, May 1996, pp. 1799-1804, and Chew K.K., “Efficient Adaptive Fast Short Moves for Magnetic Disk Drives Utilizing Magneto-resistive Heads”,
Proceedings of the Second World Congress on Intelligent Control and Intelligent Automation,
Jun. 23-27, 1997, pp. 1053-1057.
State-of-art servo technology uses so-called embedded servo position information to derive position information to position the read/write head to the desired read/write position. Position information is embedded within the data track, usually taking up less than 10% of the overall capacity. The servo sampling time ranges (depending on high or low performance) from 100 &mgr;s to 200 &mgr;s in current technology (10K TPI for high-end and 17K TPI for mobile drives) and is projected to be reduced by square root of increase in TPI.
To position/move the read/write head more accurately/quickly, a higher bandwidth servo system is required. This requires a higher sampling rate and more servo information to be embedded in the data tracks. Since position information is derived from finite discrete servo fields embedded within a data track, the servo control system is essentially operating open-loop in-between servo information. Any shock of sufficient magnitude may force the read/write head to wander significantly off track-center without the controller's knowledge.
There are two aspects of data integrity in the presence of operating shock. If a shock happens during a read operation, the head will start to read adjacent track information. The error-correcting code (ECC) capability in the drive will indicate invalid information in the data stream, causing the erroneous data sector to be read again. This is a soft-error and affects the performance and throughput in the drive, but does not affect data integrity, as the ECC will detect the error and cause the drive to re-read the erroneously read sector or sectors.
If a shock happens during a write operation with the write-head continuing to write while the head traverses off track-center over adjacent tracks, previous information in adjacent tracks will be written over. This causes a hard-error and inability to correctly retrieve data on current and adjacent tracks. Once data is written onto adjacent tracks, adjacent track data may be altered in a manner which is uncorrectable. Moreover the tracks for which data was intended to be written to may contain old data or an uncorrectable combination old and new data.
Present embedded servo methodology compromises position accuracy (through bandwidth limitations) and data integrity (due to lack of servo information while reading or writing data). Moreover, random access times are typically sacrificed (because of inadequate seeking bandwidth) with higher acoustic noise induced during seeking (due to large step changes in control signals).
In prior art hard drives, when the read/write head is between servo fields, the drive controller does not know whether the read/write head on- or off-track. If an external shock is inflicted on the drive during this time, the read/write head may wander substantially off-track. For example, if the drive sustains a shock of 100 g (typical of opening and closing a desk drawer), it may move more than 100&mgr;″ off-track in between servo fields, equivalent to 1-track in a 10K TPI drive. If the head is in a write mode, such an event will lead to catastrophic overwriting and corruption of an adjacent track.
One present solution is to include a shock sensor in the drive that senses external shocks of sufficient magnitude and immediately shuts off writing. See, for example, Levy L., “Smart Shock Sensors Preserve Data Integrity in Hard Drives”,
Data Storage Magazine,
May/June 1996, pp. 33-39.
The shock sensor approach has at least two disadvantages. First, the shock sensor is itself a mechanical device, having its own mechanical characteristics with regard to shock and vibration. During normal drive seeking operations, the drive may generate internal shocks in excess of 50 g. If a sensor is set to trip above 20 g, the sensor may falsely trigger even in normal drive operations. Moreover, the sensor may take a longer time than the seek settling time for its own vibrations to die away, taking a much longer time than is necessary to remove write inhibit signals. Second, the cost of the shock sensor is an additional cost to the drive. The external shock sensor cost depends on the number of axes required to detect transient shocks.
Another solution is to write more servo information per track, so that off-track position measurements may be made more frequently, enabling the drive to issue a write inhibit command before the read/write head wanders too far off-track to cause catastrophic data corruption during writing. With more servo information per track, a higher servo bandwidth may be achieved, leading to better seeking, settling and tracking performances as well as reduction of acoustic noise during seeking from a higher sampling rate.
Using additional servo information is also an unattractive solution, as it increases the servo overhead and reduces drive format efficiency (i.e., ratio of data sectors to servo sectors). In addition, using additional servo information reduces overall drive capacity and leads to increased cost per megabyte. Consequently, more aggressive areal density design points are required to meet drive capacity requirements, which ultimately adversely affecting final manufacturing yields.
In the prior art, early hard drives were known which provided one or more disk surfaces comprising only servo sectors (a so-called “dedicated” servo surface). Such hard disk drive might comprise a number of disks arranged along a common spindle with a number of read/write heads arranged along a common control arm. One side of one disk might comprise only servo sectors, and no data. That servo disk may be used to track data reads and write to all other surfaces of all other disks in the drive. Such techniques, however, a rarely used today, due to the high overhead associated with using an entire disk surface for servo sectors.
Tung, U.S. Pat. No. 4,924,160, issued May 8, 1980, discloses the use of staggered servo sectors on alternate sides of a disk (FIG.
4
). Tung is primarily directed toward a technique for accelerating and decelerating a drive head. Since Tung accelerates the drive head so rapidly, the chances of missing a servo sector while crossing the disk are increased (Col. 2, lines 47-51). Thus, Tung uses a “staggered seek” to reduce the likelihood of missing a servo sector during such a rapid seek.
Tung teaches that all read heads are available during a seek operation as “there is no write operation taking place during seek mode in a typical disk drive” (Col. 3, lines 4-6). Thus, other drive heads may be used for seeking othe

Affiliated with

Also associated with

Rating

4-D shock-sensing for hard-disk drives does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with 4-D shock-sensing for hard-disk drives, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and 4-D shock-sensing for hard-disk drives will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3030720