Time optimal seeks using linear velocity scheduling

Details Time optimal seeks using linear velocity scheduling Time optimal seeks using linear velocity scheduling

1999-12-29

2003-07-15

Hudspeth, David (Department: 2651)

Dynamic magnetic information storage or retrieval

Automatic control of a recorder mechanism

Controlling the head

Reexamination Certificate

active

06594105

ABSTRACT:

FIELD OF THE INVENTION
This application relates to hard disc drives and more particularly to an apparatus and method for optimizing seeks using linear velocity scheduling.
BACKGROUND OF THE INVENTION
In a disc drive data is recorded on a disc in concentric, circular paths known as tracks. During operation the disc continually rotates and a read/write head a given radius from the center of the disc would read or write data in a given track. An actuator arm swings the head in an arc across the disc surface to allow the head to read or write data in different tracks.
The read/write head is mounted upon the distal end of the actuator arm, and the arm is moved by a servo control system. Accordingly, the track position of the head is controlled by the servo system. When the head needs to access a different track, the actuator arm swings the head to the desired track location. The motion of the head from one track to another includes an acceleration and a deceleration phase, and the period during which head movement occurs is known as the seek time. For drive performance, it is desirable to minimize the seek time.
In a conventional disc drive, the movement of the actuator arm is controlled by feedforward and feedback control systems. The control process typically works as follows. A ROM (read only memory) look-up table possesses a velocity profile that indicates the target velocity of the head, given the head's distance from the desired track. Such a table assumes a nominal rate of deceleration. Typically, the table yields a target velocity for a given distance parameter based upon the relationship v(x)=[2ax]
½
, where v represents the target velocity, a represents the worst case acceleration, and x represents the distance that the head must travel, along an arc centered about the arm's pivot point, to reach its desired track position. The table is necessary because computing the velocity in real time is too processor intensive since the calculation is not linear. The target velocity is typically limited to some maximum value, v
max
.
Referring to
FIG. 8
, a prior art control system
900
is illustrated. A target velocity process
902
produces a target velocity
934
by finding from the displacement signal
932
the distance remaining to the target track. This distance is looked-up in a velocity profile stored in ROM
936
to find the target velocity value
934
. The velocity profile has been pre-determined according to the square root equation. The target velocity value
934
is fed to summation process
904
along with the actual velocity
928
that has been measured. The actual velocity is subtracted from the target velocity to produce an error quantity
906
. The error quantity
906
is amplified by scaling process
908
to produce an error value
910
. Summation process
912
combines the error value
910
with a feedforward signal
938
to produce an error current value
914
. The error current value is amplified by scaling process
916
to produce a driving current
920
. The driving current
920
is fed into the servomechanism where it is converted to acceleration
924
by conversion process
922
. The acceleration
924
is converted to velocity
928
by integration process
926
, and the velocity
928
is converted to displacement
932
by integration process
930
.
When movement begins, the arm is accelerated with the maximum torque possible. At intervals, the control system
900
gathers information regarding the actual velocity
928
of the head, and the head's distance
932
from the desired track position. Using the distance measurement, the ROM table
936
is accessed to retrieve a target velocity for the arm and thus the head. Once the target velocity
934
has been found in the table, the difference
906
between the target velocity
934
and the actual velocity
928
of the head is found. Acceleration continues until the actual velocity
928
of the head nears the target velocity, or v
max
, whichever is lower. As the distance
932
to the desired track decreases, the target velocity
934
will in turn decrease based on the square root equation. Deceleration begins when the target velocity
934
is lower than the actual velocity
928
.
During deceleration, the control system
900
once again periodically gathers information regarding the actual velocity
928
of the head, and the head's distance
932
(again, measured along an arc centered about the arm's pivot) from the desired track position. Using the distance measurement
932
, the ROM table
936
is accessed to retrieve the target velocity
934
of the head. As in the case of acceleration, calculating the velocity
934
in real time is too processor intensive and requires the table
936
to be used instead. Once the target velocity
934
has been found in the table
936
, the difference
906
between the target velocity
934
and the actual velocity
928
of the head is found. If the velocity
928
of the head exceeds the target velocity
934
, the servo system is fed with a current
920
that is proportional to the difference
906
between the head's actual
928
and target velocity
934
, and a resulting torque will be applied to the actuator arm, decelerating the arm. Deceleration continues until the head comes to rest at the desired track position.
This conventional scheme requires referencing the look-up table stored in ROM
936
because calculating the target velocity
934
in terms of distance is a non-linear, processor intensive task when constant acceleration is being applied. If the control system was able to calculate a target velocity
934
in real time, then the expensive ROM space required for the look-up table would be considerably reduced in size.
SUMMARY OF THE INVENTION
The method and apparatus in accordance with the present invention solves the aforementioned problem and other problems of producing a disc drive with an optimal seek operation. The seek operation method begins by accelerating the actuator arm of the disc drive with maximum torque. Once the acceleration has begun, a distance from the current actuator position to the desired position is determined. This distance may be determined by comparing the current position with the desired position as indicated by the command received by the disc drive from the host computer. The optimum time required to seek from the current track over the obtained distance to the desired track is acquired. This time is determined by detecting that a servo sample period has elapsed and adding the servo sample period to an initial optimum time that if stored as a negative value or subtracted from the initial optimum time if stored as a positive value. The target velocity is found from the distance to the desired track and the optimum time to seek there. The target velocity can be generated from the optimum time to reach the target for the distance to the target by finding a first target velocity component. This component is computed by scaling a zero velocity acceleration by the optimum time. A second target velocity component is obtained by scaling the distance to the target track by the mechanical motor time constant. The target velocity is then found by comparing the second component to the first component. The velocity of the head may be obtained, and then compared to the target velocity to produce an error quantity. The error quantity is multiplied by a constant to produce an error value. The error value is then combined with a feedforward quantity and a proportional error current is produced which is fed into the voice coil motor attached to the actuator arm.
The seek operation apparatus includes a voice coil motor, which is used to apply torque to an actuator arm. A transducer is coupled to the actuator arm so that it produces a signal representative of the position of the head. A microprocessor is operably connected to the transducer and to the ROM possessing acceleration and motor time constants. The microprocessor generates the actual velocity of the actuator arm from the position signal, a

Affiliated with

Also associated with

Rating

Time optimal seeks using linear velocity scheduling does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Time optimal seeks using linear velocity scheduling, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Time optimal seeks using linear velocity scheduling will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-3067715