Browse Prior Art Database

Disk Drive Servo Control in the Presence of Magnetic Defects

IP.com Disclosure Number: IPCOM000100236D
Original Publication Date: 1990-Mar-01
Included in the Prior Art Database: 2005-Mar-15
Document File: 3 page(s) / 111K

Publishing Venue

IBM

Related People

Daede, RJ: AUTHOR [+4]

Abstract

Described is an adaptive statistical measure of residue to enhance the servo performance by improving the estimator smoothness. It is particularly useful in the presence of magnetic noise. Conventional method sets residue=0, but measured residue shows significant autocorrelation for a single sample delay. Effective track misregistration (TMR) can also be statistically reduced by means of this method.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Disk Drive Servo Control in the Presence of Magnetic Defects

       Described is an adaptive statistical measure of residue
to enhance the servo performance by improving the estimator
smoothness.  It is particularly useful in the presence of magnetic
noise.  Conventional method sets residue=0, but measured residue
shows significant autocorrelation for a single sample delay.
Effective track misregistration (TMR) can also be statistically
reduced by means of this method.

      The digital servos used in magnetics employ state variable
based estimator/controller to achieve seek/tracking performance as
summarized by equations. (1)-(3):

                            (Image Omitted)

State Estimate Update: (1)
State Estimate Extrapolation: (2)
Controller Output: (3)
where and are state vectors before and after update, and are the
measurement and control sequence, and are the system matrices, and
are the estimator and controller gain vectors respectively.  The
result of the second group of terms on the right hand side of
equation (1) is called the residue which is expected to drive the
estimation process optimally.  However, when magnetic defects are
encountered, the value of measurement is not known, and, therefore,
the residue is set to zero.  This condition is equivalent to
performing only the extrapolation of the state vector as defined in
equation (2) without any update.  Such an approach is justifiable
when the residue has the properties of "white noise",  or that it has
zero correlation with adjacent samples of the residue time history.
However, measured data file shows a significant correlation for a
single sample delay.  Significant correlation is plausible because
the residue tends to carry components resulting from unmodeled
dynamics such as that of the resonant modes, disk/slider interface,
etc.

      Fig. 1a shows the normalized residue time history.  The
autocorrelation function of this s...