U.S. patent application number 11/789784 was filed with the patent office on 2008-03-13 for position control method, position control device, and medium storage device having disturbance suppression function.
This patent application is currently assigned to Fujitsu Limited. Invention is credited to Kazuhiko Takaishi.
Application Number | 20080065240 11/789784 |
Document ID | / |
Family ID | 38786896 |
Filed Date | 2008-03-13 |
United States Patent
Application |
20080065240 |
Kind Code |
A1 |
Takaishi; Kazuhiko |
March 13, 2008 |
Position control method, position control device, and medium
storage device having disturbance suppression function
Abstract
In a position control device having a disturbance suppression
function, loop gain is calibrated without stopping disturbance
suppression control. The position control device has, for
disturbance suppression control, a feedback controller for changing
the loop characteristic, a table for storing a target gain
according to a disturbance frequency, and a gain calibration
section for calibrating open loop gain. The gain is calibrated
using a target gain in the table according to the change of the
loop characteristic of the feedback controller. Open loop gain can
be calibrated without interrupting the disturbance suppression
control, so the open loop gain can be accurately calibrated without
being affected by disturbance, and accurate position control is
possible.
Inventors: |
Takaishi; Kazuhiko;
(Kawasaki, JP) |
Correspondence
Address: |
GREER, BURNS & CRAIN
300 S WACKER DR, 25TH FLOOR
CHICAGO
IL
60606
US
|
Assignee: |
Fujitsu Limited
Kawasaki-shi
JP
|
Family ID: |
38786896 |
Appl. No.: |
11/789784 |
Filed: |
April 25, 2007 |
Current U.S.
Class: |
700/37 ;
G9B/5.198; G9B/5.216; G9B/5.228 |
Current CPC
Class: |
G11B 5/5582 20130101;
G11B 5/59688 20130101; G11B 5/596 20130101 |
Class at
Publication: |
700/37 |
International
Class: |
G05B 13/02 20060101
G05B013/02 |
Foreign Application Data
Date |
Code |
Application Number |
Sep 12, 2006 |
JP |
2006-246451 |
Claims
1. A position control method for controlling a position of an
object to a predetermined position by an actuator, comprising: a
step of computing a position error based on a target position of
the object and a current position of the object; a step of
computing a control value in which a disturbance frequency
component is suppressed, using a predetermined feedback loop based
on the position error and computing a drive value of the actuator
by multiplying the control value by a loop gain; a step of fetching
a target loop gain according to the disturbance frequency from a
table; a step of adding disturbance of a measurement frequency to
the feedback loop and measuring a loop gain of the feedback loop;
and a step of calibrating the loop gain of the control value
computing step based on the measured loop gain and the target loop
gain.
2. The position control method according to claim 1, wherein the
step of fetching the target loop gain comprises a step of fetching
a measurement frequency according to the disturbance frequency, and
wherein the measurement step comprises a step of adding disturbance
of the fetched measurement frequency to the feedback loop, and
measuring a loop gain of the feedback loop.
3. The position control method according to claim 1, wherein the
step of fetching the measurement frequency comprises a step of
fetching a measurement frequency which does not overlap the
disturbance frequency.
4. The position control method according to claim 1, wherein the
drive value computing step comprises: a step of estimating the
disturbance frequency according to the position error based on an
adaptive control; a step of computing a control value in which the
disturbance frequency component is suppressed, according to the
estimated disturbance frequency; and a step of computing a drive
value of the actuator by multiplying the control value by a loop
gain.
5. The position control method according to claim 4, further
comprising a step of interrupting the estimation of the disturbance
frequency according to the position error during calibration of the
loop gain.
6. The position control method according to claim 4, wherein the
drive value computing step comprises: a step of estimating the
disturbance frequency according to the position error based on an
adaptive control, and changing a constant of a controller according
to the estimated disturbance frequency; and a step of computing a
control value in which the disturbance frequency component is
suppressed, using the changed controller according to the position
error, and multiplying the result by a loop gain, so as to compute
a drive value of the actuator.
7. The position control method according to claim 4, wherein the
drive value computing step comprises: a step of estimating the
disturbance frequency according to the position error based on an
adaptive control; a step of changing a constant of a controller
constructed with an observer according to the estimated disturbance
frequency; and a computing step of computing a control value in
which the disturbance frequency component is suppressed according
to the position error, using the changed observer, and multiplying
the result by a loop gain, so as to compute a drive value of the
actuator.
8. A position control device for controlling a position of an
object to a predetermined position by an actuator, comprising: a
control unit for computing a position error based on a target
position of the object and a current position of the object,
computing a control value in which a disturbance frequency
component is suppressed, using a predetermined feedback loop based
on the position error, and multiplying the result by a loop gain,
so as to compute a drive value of the actuator; and a table for
storing a target loop gain according to the disturbance frequency,
wherein the control unit fetches the target loop gain according to
the disturbance frequency from the table, adds a disturbance of a
measurement frequency to the feedback loop, measures a loop gain of
the feedback loop, and calibrates the loop gain in the control
value computing step based on the measured loop gain and the target
loop gain.
9. The position control device according to claim 8, wherein the
control unit fetches a measurement frequency according to the
disturbance frequency from the table, adds a disturbance of the
fetched measurement frequency to the feedback loop, and measures a
loop gain of the feedback loop.
10. The position control device according to claim 8, wherein the
control unit fetches a measurement frequency which does not overlap
with the disturbance frequency from the table.
11. The position control device according to claim 8, wherein the
control unit estimates the disturbance frequency according to the
position error based on an adaptive control, computes a control
value in which the disturbance frequency component is suppressed
according to the estimated disturbance frequency, and multiplies
the result by a loop gain, so as to compute a drive value of the
actuator.
12. The position control device according to claim 11, wherein the
control unit interrupts the estimation of the disturbance frequency
according to the position error during calibration of the loop
gain.
13. The position control device according to claim 11, wherein the
control unit estimates the disturbance frequency according to the
position error based on adaptive control, changes a constant of a
controller according to the estimated disturbance frequency,
computes a control value in which the disturbance frequency
component is suppressed, using the changed controller, according to
the position error, and multiplies the result by a loop gain, so as
to compute a drive value of the actuator.
14. The position control device according to claim 11, wherein the
control unit estimates the disturbance frequency according to the
position error based on adaptive control, changes a constant of a
controller constructed with an observer according to the estimated
disturbance frequency, computes a control value in which the
disturbance frequency component is suppressed according to the
position error, using the changed observer, and multiplies the
result by a loop gain, so as to compute a drive value of the
actuator.
15. A medium storage device, comprising: a head for at least
reading data on a storage medium; an actuator for positioning the
head to a predetermined position on the storage medium; a control
section for computing a position error based on a target position
of the head and a current position acquired from the head,
computing a control value in which a disturbance frequency
component is suppressed, using a predetermined feedback loop based
on the position error, and multiplying the result by a loop gain,
so as to compute a drive value of the actuator; and a table for
storing a target loop gain according to the disturbance frequency,
wherein the control section fetches the target loop gain according
to the disturbance frequency from the table, adds a disturbance of
a measurement frequency to the feedback loop, measures loop gain of
the feedback loop, and calibrates the loop gain in the control
value computing step based on the measured loop gain and the target
loop gain.
16. The medium storage device according to claim 15, wherein the
control section fetches a measurement frequency according to the
disturbance frequency from the table, adds a disturbance of the
fetched measurement frequency to the feedback loop, and measures a
loop gain of the feedback loop.
17. The medium storage device according to claim 15, wherein the
control section fetches a measurement frequency which does not
overlap with the disturbance frequency from the table.
18. The medium storage device according to claim 15, wherein the
control section estimates the disturbance frequency according to
the position error based on adaptive control, computes a control
value in which the disturbance frequency component is suppressed,
according to the estimated disturbance frequency, and multiplies
the control value by a loop gain, so as to compute a drive value of
the actuator.
19. The medium storage device according to claim 18, wherein the
control section interrupts the estimation of the disturbance
frequency according to the position error during calibration of the
loop gain.
20. The medium storage device according to claim 18, wherein the
control section estimates the disturbance frequency according to
the position error based on adaptive control, changes a constant of
a controller according to the estimated disturbance frequency,
computes a control value in which the disturbance frequency
component is suppressed, using the changed controller, according to
the position error, and multiplies the result by a loop gain, so as
to compute a drive value of the actuator.
21. The medium storage device according to claim 18, wherein the
control section estimates the disturbance frequency according to
the position error based on adaptive control, changes a constant of
a controller constructed with an observer according to the
estimated disturbance frequency, computes a control value in which
the disturbance frequency component is suppressed, according to the
position error, using the changed observer, and multiplies the
result by a loop gain, so as to compute a drive value of the
actuator.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is based upon and claims the benefit of
priority from the prior Japanese Patent Application No.
2006-246451, filed on Sep. 12, 2006, the entire contents of which
are incorporated herein by reference.
BACKGROUND OF THE INVENTION
[0002] 1. Field of the Invention
[0003] The present invention relates to a position control method,
position control device and medium storage device for controlling
the position of an object by suppressing disturbance, and more
particularly to a position control method, position control device,
and medium storage device for suppressing the position deviation of
an object due to disturbance.
[0004] 2. Description of the Related Art
[0005] A position control device is for controlling the position of
an object to a specified position and is widely used in various
fields. For example, in a medium storage device, such as a magnetic
disk device and an optical disk device, it is extremely critical to
accurately position the head to a target track to improve recording
density.
[0006] In this position control, it is known that disturbance
affects positioning accuracy. In order to suppress this disturbance
by a control system, a method of installing a filter for
suppressing specific frequencies (e.g. see U.S. Pat. No. 6,487,028
B1, R. J. Bickel and M. Tomizuka: "Disturbance observer based
hybrid impedance control", (Proceedings of the American Control
Conference, 1995, pp. 729-733)), and a method of suppressing such
disturbance as eccentricity by observer control (e.g. see Japanese
Patent Application Laid-Open No. H7-50075 and Japanese Patent
Application Laid-Open No. 2000-21104) have been proposed.
[0007] On the other hand, in order for the positioning control
system to operate accurately, calibration of gain of open loop
characteristics (open loop gain) is indispensable. The open loop
gain changes if the characteristics of elements constituting a
feedback loop change (e.g. temperature, age-based deterioration).
Always maintaining this open loop gain at an optimum contributes to
the performance of the feedback loop. This calibration is performed
when the power of a device is turned ON, when temperature is
changed, or when a predetermined time elapsed, for example, so as
to calibrate the open loop gain at an optimum.
[0008] As a method for calibrating a gain, applying a sine wave
disturbance to a position or current and acquiring and comparing
waveforms before and after applying the sine wave disturbance so as
to measure the gain of the open loop characteristic is known (e.g.
see Japanese Patent Application Laid-Open No. H11-328891, Japanese
Patent Application Laid-Open No. H8-167160, and Japanese Patent
Application Laid-Open No. H11-96704).
[0009] In the case of this conventional gain calibration method, a
target gain of the loop characteristic for adjusting the gain and a
sine wave disturbance frequency for adjusting the gain are fixed
values. In other words, in the case of the conventional gain
calibration method, gain is calibrated assuming that the position
control system has only one type of characteristic during gain
adjustment.
[0010] However in the case of a position control system with a
disturbance suppression function, the characteristics of the
position control system change according to the disturbance
frequency to be suppressed. For example, in a control system using
adaptive control, the loop characteristic of the control system is
different from that before disturbance is applied, when the
position control system is following up disturbance based on the
adaptive control in a state of applied disturbance vibration.
[0011] In order to execute gain adjustment while this adaptive
control is following up disturbance, it is necessary to stop the
adaptive control and switch to the control system for gain
adjustment. In this case, the external vibrations cannot be
sufficiently suppressed and positioning accuracy drops since the
adaptive control is stopped. As a result, accuracy of gain
adjustment drops.
[0012] In other words, in the case of such a position control
system as an adaptive control system, disturbance is suppressed by
changing the loop characteristic according to disturbance, so it is
difficult to accurately calibrate the open loop gain of the control
system using prior art. Particularly if suppression width is taken
wide or if disturbance in a high frequency area is suppressed to
meet the recent demand for adapting a wide range of disturbance
frequencies, the original characteristics of the controller are
influenced, so it is more difficult to calibrate gain
accurately.
SUMMARY OF THE INVENTION
[0013] With the foregoing in view, it is an object of the present
invention to provide a position control method, position control
device, and medium storage device for accurately calibrating the
open loop gain in a position control system using the disturbance
adaptive control.
[0014] It is another object of the present invention to provide a
position control method, position control device, and medium
storage device for accurately calibrating the open loop gain even
while disturbance is being applied.
[0015] It is still another object of the present invention to
provide a position control method, position control device, and
medium storage device for accurately calibrating the open loop
gain, even if the disturbance suppression frequency is set from the
outside.
[0016] It is still another object of the present invention to
provide a position control method, position control device, and
medium storage device for accurately calibrating the open loop
gain, even if the position control system adapts to disturbance
suppression frequencies in a wide range.
[0017] A position control method according to the present invention
is a position control method for controlling a position of an
object to a predetermined position by an actuator, having: a step
of computing a position error based on a target position of the
object and a current position of the object; a step of computing a
control value in which a disturbance frequency component is
suppressed, using a predetermined feedback loop based on the
position error, and multiplying the result by a loop gain, so as to
compute a drive value of the actuator; a step of fetching a target
loop gain according to the disturbance frequency from a table; a
step of adding the disturbance of a measurement frequency to the
feedback loop and measuring a loop gain of the feedback loop; and a
step of calibrating the loop gain of the control value computing
step based on the measured loop gain and the target loop gain.
[0018] A position control device for controlling a position of an
object to a predetermined position by an actuator according to the
present invention, has: a control section for computing a position
error based on a target position of the object and a current
position of the object, computing a control value in which a
disturbance frequency component is suppressed, using a
predetermined feedback loop based on the position error, and
multiplying the result by a loop gain, so as to compute a drive
value of the actuator; and a table for storing a target loop gain
according to the disturbance frequency, wherein the control section
fetches the target loop gain according to the disturbance frequency
from the table, adds the disturbance of a measurement frequency to
the feedback loop, measures a loop gain of the feedback loop, and
calibrates the loop gain in the control value computing step based
on the measured loop gain and the target loop gain.
[0019] A medium storage device according to the present invention,
has: a head for at least reading data on a storage medium; an
actuator for positioning the head to a predetermined position on
the storage medium; a control section for computing a position
error based on a target position of the head and a current position
acquired from the head, computing a control value in which a
disturbance frequency component is suppressed, using a
predetermined feedback loop based on the position error, and
multiplying the result by a loop gain, so as to compute a drive
value of the actuator; and a table for storing a target loop gain
according to the disturbance frequency, wherein the control section
fetches the target loop gain according to the disturbance frequency
from the table, adds the disturbance of a measurement frequency to
the feedback loop, measures a loop gain of the feedback loop, and
calibrates the loop gain in the control value computing step based
on the measured loop gain and the target loop gain.
[0020] In the present invention, it is preferable that the step of
fetching the target loop gain further has a step of fetching a
measurement frequency according to the disturbance frequency, and
the measurement step further has a step of adding the disturbance
of the fetched measurement frequency to the feedback loop and
measuring a loop gain of the feedback loop.
[0021] Also in the present invention, it is preferable that the
step of fetching the measurement frequency further has a step of
fetching a measurement frequency which does not overlap with the
disturbance frequency.
[0022] Also in the present invention, it is preferable that the
drive value computing step further has a step of estimating the
disturbance frequency according to the position error based on an
adaptive control, computing a control value in which the
disturbance frequency component is suppressed, according to the
estimated disturbance frequency, and multiplying the result by a
loop gain, so as to compute a drive value of the actuator.
[0023] It is also preferable that the present invention further has
a step of interrupting the estimation of the disturbance frequency
according to the position error during calibration of the loop
gain.
[0024] Also in the present invention, it is preferable that the
drive value computing step further has a step of estimating the
disturbance frequency according to the position error based on an
adaptive control, and changing a constant of a controller according
to the estimated disturbance frequency, and a step of computing a
control value in which the disturbance frequency component is
suppressed, using the changed controller according to the position
error, and multiplying the result by a loop gain so as to compute a
drive value of the actuator.
[0025] Also in the present invention, it is preferable that the
drive value computing step further has a step of estimating the
disturbance frequency according to the position error based on an
adaptive control, and changing a constant of a controller
constructed with an observer according to the estimated disturbance
frequency, and a computing step of computing a control value in
which the disturbance frequency component is suppressed, according
to the position error using the changed observer, and multiplying
the result by a loop gain, so as to compute a drive value of the
actuator.
[0026] Because of disturbance suppression control, open loop gain
is calibrated using a target gain according to a disturbance
frequency, even if a loop characteristic of a feedback controller
changes, so the open loop gain can be calibrated without
interrupting the disturbance suppression control. Therefore the
open loop gain can be accurately calibrated without being affected
by disturbance, and accurate position control is possible.
BRIEF DESCRIPTION OF THE DRAWINGS
[0027] FIG. 1 is a block diagram depicting a medium storage device
according to an embodiment of the present invention;
[0028] FIG. 2 is a diagram depicting the position signals of the
medium storage device in FIG. 1;
[0029] FIG. 3 is a diagram depicting details of the position
signals in FIG. 2;
[0030] FIG. 4 is a block diagram depicting a position control
system according to the first embodiment of the present
invention;
[0031] FIG. 5 shows a target gain table in FIG. 4;
[0032] FIG. 6 is a characteristic diagram of the sensitivity
function of the feedback loop in FIG. 4;
[0033] FIG. 7 is an open loop characteristic diagram of the
feedback loop in FIG. 4;
[0034] FIG. 8 is a flow chart depicting the gain calibration
processing of the gain calibration block in FIG. 4;
[0035] FIG. 9 is a diagram depicting the sine wave disturbance of
the gain calibration processing in FIG. 8;
[0036] FIG. 10 is a block diagram depicting a position control
system according to the second embodiment of the present
invention;
[0037] FIG. 11 shows a target gain table of the embodiment in FIG.
10;
[0038] FIG. 12 is a block diagram depicting a position control
system according to the third embodiment of the present
invention;
[0039] FIG. 13 is a block diagram depicting an embodiment where the
controller in FIG. 12 is constructed with a current observer;
[0040] FIG. 14 shows a parameter table of the embodiment in FIG.
13;
[0041] FIG. 15 shows another parameter table of the embodiment in
FIG. 13;
[0042] FIG. 16 is a diagram depicting a measurement frequency in
the parameter table in FIG. 15;
[0043] FIG. 17 shows still another parameter table of the
embodiment in FIG. 13;
[0044] FIG. 18 is a diagram depicting a measurement frequency of
the parameter table in FIG. 17;
[0045] FIG. 19 is a block diagram depicting a position control
system according to the fourth embodiment of the present invention;
and
[0046] FIG. 20 is a block diagram depicting a position control
system according to the fifth embodiment of the present
invention.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0047] Embodiments of the present invention will now be described
in the sequence of the position control device, first embodiment of
the position control device, loop gain calibration processing,
second embodiment, third embodiment, design of disturbance
observer, fourth embodiment, fifth embodiment, and other
embodiments, but the present invention is not limited to these
embodiments.
Position Control Device
[0048] FIG. 1 is a block diagram depicting the position control
device according to an embodiment of the present invention, FIG. 2
is a diagram depicting the arrangement of the position signals of
the magnetic disk in FIG. 1, and FIG. 3 is a diagram depicting the
position signals of the magnetic disk in FIG. 1 and FIG. 2.
[0049] FIG. 1 shows a magnetic disk device, which is a type of disk
device, as a position control device. As FIG. 1 shows, in a
magnetic disk device, a magnetic disk 4, which is a magnetic
storage medium, is installed at a rotation axis 2 of a spindle
motor 5. The spindle motor 5 rotates the magnetic disk 4. An
actuator (VCM) 1 has a magnetic head 3 at the tip, and moves the
magnetic head 3 in the radius direction of the magnetic disk 4.
[0050] The actuator 1 is comprised of a voice coil motor (VCM)
which rotates with the rotation axis as the center. In FIG. 1, two
magnetic disks 4 are mounted on the magnetic disk device, and four
magnetic heads 3 are simultaneously driven by the same actuator
1.
[0051] The magnetic head 3 is a separation type magnetic head which
has a read element and a write element. The magnetic head 3 is
comprised of a read element, including a magneto-resistance (MR)
element, stacked on a slider, and a write element, including a
write coil, stacked thereon.
[0052] A position detection circuit 7 converts the position signals
(analog signals) read by the magnetic head 3 into digital signals.
A read/write (R/W) circuit 10 controls the reading and writing of
the magnetic head 3. A spindle motor (SPM) device circuit 8 drives
the spindle motor 5. A voice coil motor (VCM) drive circuit 6
supplies drive current to the voice coil motor (VCM) 1, and drives
the VCM 1.
[0053] A microcontroller (MCU) 14 detects (demodulates) the current
position from the digital position signals from the position
detection circuit 7, and computes a VCM drive command value
according to an error between the detected current position and a
target position. In other words, the microcontroller 14 performs
position demodulation and servo control, including the disturbance
suppression described in FIG. 4 and later. A read only memory (ROM)
13 stores a control program of the MCU 14. A random access memory
(RAM) 12 stores data for processing of the MCU 14.
[0054] A hard disk controller (HDC) 11 judges a position in one
track based on a sector number of a servo signal, and
records/reproduces the data. A random access memory (RAM) for
buffer 15 temporarily stores the read data or write data. The HDC
11 communicates with a host via an interface IF, such as USB, ATA
and SCSI. A bus 9 connects these composing elements.
[0055] As FIG. 2 shows, on the magnetic disk 4, servo signals
(position signals) 16 are arrayed in each track in the
circumferential direction from the outer circumference to the inner
circumference with an equal interval. Each track has a plurality of
sectors, and the solid lines in FIG. 2 indicate positions where the
servo signals 16 are recorded. As FIG. 3 shows, the position signal
is comprised of a servo mark ServoMark, track number GrayCode,
index Index, and offset information (servo burst) PosA, PosB, PosC
and PosD. The dotted line in FIG. 3 shows the track center.
[0056] The position signals in FIG. 3 are read by the head 3, and
the position of the magnetic head in the radius direction is
detected using the track number GrayCode and offset information
PosA, PosB, PosC and PosD. Also the position of the magnetic head
in the circumference direction is acquired based on the index
signal Index.
[0057] For example, the sector number when the index signal is
detected is set to No. 0, which is counted up every time the servo
signal is detected, so as to acquire the sector number of each
sector of the track. The sector number of the servo signals is used
as a reference when data is recorded/reproduced. There is one index
signal in one track. The sector number may be set instead of the
index signal.
[0058] The MCU 14 in FIG. 1 confirms the position of the actuator 1
through the position detection circuit 7, performs servo
computation, and supplies appropriate current to the VCM 1. In
other words, in seek control, the head is moved to the target
position through the transition from coarse control, settling
control and following control. For all these controls, the current
position of the head must be detected.
[0059] To confirm the position like this, the servo signals are
recorded on the magnetic disk in advance, as mentioned in FIG. 2.
In other words, as FIG. 3 shows, servo marks which indicate the
start position of the servo signal, gray code which indicates the
track number, index signal, and signals PosA to PosD which indicate
the offset are recorded on the magnetic disk in advance. These
signals are read by the magnetic head, and these servo signals are
converted into digital values by the position detection circuit
7.
First Embodiment of Position Control System
[0060] FIG. 4 is a block diagram depicting a first embodiment of
the position control system of the present invention, and is a
block diagram of a position control system for suppressing
disturbance which the MCU 14 in FIG. 1 executes. FIG. 5 shows the
target gain table in FIG. 4, and FIG. 6 and FIG. 7 are
characteristic diagrams of the disturbance adaptive control in FIG.
4.
[0061] The position control system in FIG. 4 controls a disturbance
suppression compensation function of a controller which is set from
the outside, or according to a detected disturbance frequency
Fdist. A gain adjustment function is added to this position control
system. An error computing block 24 subtracts an observation
position (current position) `y` from a target position `r` to
compute a position error `e`.
[0062] According to the position error `e`, a controller 20
computes a drive instruction value `u` of a plant 22 (1, 3) to make
the position error `e`. The controller 20 computes the drive
instruction value `u` by a known PID control, PI control+LeadLag,
and observer control, for example.
[0063] A gain multiplication block 26 multiplies the drive
instruction value `u` from the controller 20 by a gain which is set
(open loop gain), and outputs the result. A power amplifier, which
is not illustrated, converts this output to a drive current I of
the plant 22 (1, 3), and drives the plant 22 (1, 3).
[0064] A disturbance suppression compensation block 30 changes a
characteristic (e.g. constant) of the controller 20 according to a
disturbance suppression frequency which is set from the outside, or
an estimated disturbance frequency Fdist, and adds a disturbance
frequency suppression characteristic by the controller 20. This
suppression characteristic will be described using the sensitivity
function (1/(1+CP)) and open loop characteristic (CP). C is a
characteristic of a controller, and P is a characteristic of a
plant.
[0065] FIG. 6 shows an example of the frequency characteristic of a
sensitivity function of the position control system, where a
frequency (Hz) vs. gain characteristic (dB) is shown on the top,
and a frequency (Hz) vs. phase characteristic (deg) is shown at the
bottom of FIG. 6. As FIG. 6 shows, the sensitivity function is
changed as shown in the thin line from the original characteristic
of the controller which is shown in the thick line, according to
the frequency to be suppressed. Here the sensitivity function in
the case of suppressing a 500 Hz disturbance frequency is
shown.
[0066] FIG. 7 shows an example of a frequency characteristic of the
open loop characteristic of the position control system
corresponding to FIG. 6, where a frequency (Hz) vs. gain
characteristic (dB) is shown on the top, and a frequency (Hz) vs.
phase characteristic (deg) is shown at the bottom of FIG. 7. As
FIG. 7 shows, the open loop characteristic is changed as shown in
the thin line from the original characteristic of the controller
which is shown in the thick line, according to the frequency to be
suppressed. Here the open loop characteristic in the case of
suppressing a 500 Hz disturbance frequency is shown.
[0067] A controller 20 which implements this sensitivity function
or open loop characteristic for suppressing a specific disturbance
frequency will be concretely described in FIG. 12 or later.
[0068] Back in FIG. 4, a gain calibration block 34 applies a sine
wave disturbance SD having a predetermined frequency according to a
gain calibration instruction, detects signals in a loop before and
after applying the sine wave disturbance, and calibrates a gain in
the gain multiplication block 26. In FIG. 4, a sine wave
disturbance SD for measurement is applied from an addition block 28
to a feedback loop according to a position (position error) which
is an input of the controller 20, and position errors before and
after applying the sine wave disturbance are observed. A target
gain table 32 stores a target gain G corresponding to a disturbance
frequency Fdist, supplies the target gain G corresponding to the
disturbance frequency Fdist to the gain calibration block 34, and
uses it as a reference of the gain calibration in the gain
calibration block 34.
[0069] This target gain table 34 stores target gains TG1, TG2, . .
. , TGnwo for each disturbance frequency Fdist, as shown in FIG. 5.
This target gain is decided corresponding to the loop
characteristic, which changes by the disturbance suppression
control, as described in FIG. 6 and FIG. 7.
Gain Calibration Processing
[0070] Now the calibration processing of the gain calibration block
34 will be described. FIG. 8 is a flow chart depicting the gain
calibration processing which the gain calibration block 34
executes, and FIG. 9 is a diagram depicting the sine wave
disturbance thereof.
[0071] (S10) The gain calibration block 34 acquires a target gain
TG corresponding to a disturbance frequency from the target gain
table 34.
[0072] (S12) The gain calibration block 34 applies a sine wave
disturbance SD to the addition block 28. FIG. 9 shows an example of
a waveform of a sine wave disturbance SD to be applied, and a sine
wave at 800 Hz, for example, is used.
[0073] (S14) The gain calibration block 34 observes signals S1 and
S2 before and after the disturbance is applied. In FIG. 4, the
input stage of the addition block 28 is observed (acquired) as a
signal (position error) S1 before the disturbance is applied, and
the output stage of the addition block 28 as a signal (position
error) S2 after the disturbance is applied.
[0074] (S16) The gain calibration block 34 performs DFT (Digital
Fourier Transfer) operation on each of the observed signals S1 and
S2, so as to determine the level (amplitude) of each signal S1 and
S2. This is repeated for N number of turns of the disk (e.g. 10
turns), and the sums thereof .SIGMA.S1 and .SIGMA.S2 are
calculated.
[0075] (S18) The gain calibration block 34 divides the sum of the
amplitudes of the signals S1 before the disturbance is applied,
that is .SIGMA.S1, by the sum of the amplitudes of the signals S2
after the disturbance is applied, that is .SIGMA.S2, so as to
calculate an open loop gain Tm.
[0076] (S20) Then the gain calibration block 34 calculates a
correction gain Tc by dividing the target gain TG corresponding to
the disturbance frequency, which is acquired in step S10, by the
measured open loop gain Tm.
[0077] (S22) The gain calibration block 34 sets this correction
gain Tc in the gain calibration block 26, and ends the
calibration.
[0078] It is desirable to set this target gain TG for each
disturbance frequency, but this increases the memory capacity of
the target gain memory 32. Therefore as FIG. 5 shows, the table 32
may store the target gain TG at a multiple of Fr (e.g. rotation
frequency), so that for disturbance frequencies between these
multiple frequencies, a target gain corresponding to the
disturbance frequency Fr is calculated by interpolation.
[0079] By suppressing disturbance like this, the open loop gain is
calibrated using a target gain according to the disturbance
frequency, so the open loop gain can be calibrated without
interrupting disturbance suppression control, even if the loop
characteristic of the controller 20 is changed. Therefore the open
loop gain can be accurately calibrated without being affected by
the disturbance.
Second Embodiment of Position Control System
[0080] FIG. 10 is a block diagram depicting a second embodiment of
the position control system of the present invention. In FIG. 10,
composing elements the same as those in FIG. 4 are denoted with the
same reference symbols.
[0081] Just like FIG. 4, an error computing block 24 subtracts an
observation position (current position) y from the target position
r to compute a position error e. A controller 20 computes,
according to position the error e, a drive instruction value u of a
plant 22 (1, 3) to make the position error e zero. The controller
20 computes the drive instruction value u by a known PID control,
PI control+LeadLag, and observer control, for example.
[0082] A gain multiplication block 26 multiplies the drive
instruction value u from the controller 20 by a gain which is set
(open loop gain), and outputs the result. A power amplifier, which
is not illustrated, converts this output into a drive current I of
the plant 22 (1, 3), and drives the plant 22 (1, 3).
[0083] A disturbance suppression compensation block 30 changes the
characteristic (e.g. constant) of the controller 20 according to a
disturbance suppression frequency which is set from the outside, or
an estimated disturbance frequency Fdist, and adds a disturbance
frequency suppression characteristic by the controller 20.
[0084] A gain calibration block 34 applies a sine wave disturbance
SD having a predetermined frequency according to a gain calibration
instruction, detects signals in a loop before and after applying
the sine wave disturbance, and calibrates a gain in the gain
multiplication block 26. In FIG. 11, a sine wave disturbance SD for
measurement is applied from an addition block 28-1 to a feedback
loop according to a current level (drive instruction value) which
is an output of the controller, and currents before and after
applying the sine wave disturbance are observed. A target gain
table 32 stores a target gain G corresponding to a disturbance
frequency Fdist, supplies the target gain G corresponding to the
disturbance frequency Fdist to the gain calibration block 34, and
uses it as a reference of the gain calibration in the gain
calibration block 34.
[0085] This target gain table 34 stores target gains TG1, TG2, . .
. , TGn for each disturbance frequency Fdist, as shown in FIG. 5 or
FIG. 10. This target gain is decided corresponding to the loop
characteristic, which changes by the disturbance suppression
control, as described in FIG. 6 and FIG. 7.
[0086] This gain calibration processing is the same as FIG. 8,
except that the observation target is the current level of the
output stage. In this way, gain can be calibrated also by observing
the current level.
[0087] FIG. 11 shows another target gain table. Compared with the
table 32 in FIG. 5, this table 32-1 has a measurement frequency
column in addition to the target gain, for each disturbance
frequency Fdist. In other words, the measurement frequency can also
be changed according to the disturbance frequency Fdist.
[0088] Here if the disturbance frequency Fdist is a predetermined
measurement frequency fsd, the measurement frequency is changed to
fsd+.alpha. (.alpha..noteq.0). In other words, applying a
measurement frequency the same as a disturbance frequency to the
control system as a disturbance while the position control system
is controlling disturbance suppression means that the disturbance
frequency is supplied into the loop while suppressing the same
disturbance frequency.
[0089] Therefore the sine wave disturbance of the measurement
frequency is also suppressed by the disturbance suppression
function, and accurate open loop gain cannot be measured. So the
measurement frequency is shifted so as not to overlap with the
disturbance frequency, then the open loop gain can be accurately
measured.
[0090] The table 32-1 in FIG. 11 can also be applied to the first
embodiment in FIG. 4.
Third Embodiment
[0091] FIG. 12 is a block diagram depicting a third embodiment of
the position control system of the present invention, FIG. 13 is a
block diagram when the control system in FIG. 12 is constructed
with a current observer, and FIG. 14 is the parameter table in FIG.
13.
[0092] FIG. 12 shows a position control system for detecting a
disturbance frequency and suppressing the disturbance by adaptive
control, and in FIG. 12, composing elements the same as those in
FIG. 4 are denoted with the same reference symbols.
[0093] In other words, an error computing block 24 subtracts an
observation position (current position) `y` from a target position
`r` to compute a position error `e`. According to the position
error `e`, a controller 20 computes a drive instruction value `u`
of a plant 22 (1, 3) to make the position error e zero. The
controller 20 computes the drive instruction value `u` by an
observer control to be described later in FIG. 13, for example.
[0094] A gain multiplication block 26 multiplies the drive
instruction value `u` from the controller 20 by a gain which is set
(open loop gain), and outputs the result. A power amplifier, which
is not illustrated, converts this output into a drive current I of
the plant 22 (1, 3), and drives the plant 22 (1, 3).
[0095] A disturbance suppression compensation block 30 is comprised
of an external suppression adaptive control system. This adaptive
control system has a .omega. estimation section 30-1 for estimating
a disturbance frequency Fdist (.omega.) according to an adaptive
rule based on the position error of the controller 20, and a table
30-2 for storing estimated gains L and A of the controller 20
according to an estimated frequency (angular frequency .omega. in
this case).
[0096] A gain calibration block 34 applies a sine wave disturbance
SD having a predetermined frequency according to a gain calibration
instruction, detects signals in a loop before and after applying
the sine wave disturbance, and calibrates a gain in the gain
multiplication block 26. In FIG. 12, a sine wave disturbance SD for
measurement is applied from the addition block 28 to a feedback
loop according to a position which is an input of the controller,
and positions before and after applying the sine wave disturbance
are observed.
[0097] A target gain table 32 stores a target gain G corresponding
to a disturbance frequency Fdist, supplies the target gain G
corresponding to the disturbance frequency Fdist from the .omega.
estimation section 30-1 to the gain calibration block 34, and uses
it as a reference of the gain calibration in the gain calibration
block 34.
[0098] The present embodiment also has a switch SW for stopping the
input of a position error of the controller 20 to the .omega.
estimation section 30-1 during gain calibration. By this, the
.omega. estimation section 30-1 maintains the estimated disturbance
frequency before starting gain calibration during gain calibration.
Therefore during gain calibration, the disturbance suppression
control is executed, but adaptive control is interrupted so that
unnecessary adaptive control is not performed by a sine wave
disturbance for measuring the gain.
[0099] In the present embodiment, the characteristic (e.g.
constant) of the controller 20 is changed according to an estimated
disturbance frequency Fdist, and the disturbance frequency
suppression characteristic is added by the controller 20.
[0100] The present embodiment will be described in more detail
using the controller 20 based on a current observer in FIG. 13. In
FIG. 13, composing elements the same as those in FIG. 12 are shown
with the same reference symbols. The current observer shown in FIG.
13 is a current observer which includes bias compensation shown in
the following Expressions (1), (2), (3), (4) and (5).
( x ( k ) v ( k ) b ( k ) z 1 ( k ) z 2 ( k ) ) = ( x ( k ) v ( k )
b ( k ) z 1 ( k ) z 2 ( k ) ) + ( L 1 L 2 L 3 L 4 L 5 ) ( y ( k ) -
x ( k ) ) ( 1 ) u ( k ) = - ( F 1 F 2 ) ( x ( k ) v ( k ) ) ( 2 )
uout ( k ) = u ( k ) - ( F 3 F 4 F 5 ) ( b ( k ) z 1 ( k ) z 2 ( k
) ) ( 3 ) ( x ( k + 1 ) v ( k + 1 ) ) = ( 1 1 0 1 ) ( x ( k ) v ( k
) ) + Bl m 1 Lp T 2 ( 1 / 2 1 ) u ( k ) ( 4 ) b ( k + 1 ) = b ( k )
( z 1 ( k + 1 ) z 2 ( k + 1 ) ) = ( a 11 a 12 a 21 a 22 ) ( z 1 ( k
) z 2 ( k ) ) } ( 5 ) ##EQU00001##
[0101] In other words, this embodiment is an example of an adaptive
control system where the controller 20 is separated into a model of
the controller and a disturbance model.
[0102] In FIG. 13, a first computing block 24 computes an actual
position error er[k] by subtracting a target position `r` from an
observation position y[k] which is acquired by demodulating the
servo information read by the head 3. A second computing block 40
computes an estimated position error e[k] by subtracting an
estimated position x[k] of the observer from the actual position
error er[k].
[0103] In the controller model, this estimated position error e[k]
is input to a state estimation block 42, and an estimated
correction value (right hand side of Expression (1)) is computed
using an estimated gain La (L1, L2) of the controller. And in the
addition block 44, the result is added with state quantities (left
hand side of Expression (1)) x[k] and v[k] from the delay block 46,
and estimated position x[k] and estimated velocity v[k] are
acquired, as shown in Expression (1). In Expression (1), the
estimated position error e[k] is indicated by (y[k]-x[k]).
[0104] The estimated values x[k] and v[k] are multiplied by a state
feedback gain (-Fa=F1, F2) in a fourth computing block 48, and a
first drive value u[k] for the actuator 1 is acquired, as shown in
Expression (2). On the other hand, the estimated values x[k] and
v[k] of Expression (1) from an addition block 44 are multiplied by
an estimated gain Aa (2.times.2 matrix (1, 0) in Expression (4)) in
a fifth computing block 52.
[0105] The drive value u[k] in the fourth computing block 48 is
multiplied by an estimated gain Ba (a value by which u[k] in
Expression (4) is multiplied) in a sixth computing block 50. Both
of the multiplication results are added in an addition block 54,
and estimated state quantities x[k+1] and v[k+1] of the next sample
in Expression (4) are acquired.
[0106] The estimated state quantity of the next sample is input to
a delay block 46, and is corrected with the estimated correction
value in the state estimation block 42. And for the estimated value
of Expression (1) from the addition block 44, the estimated
position x[k] is acquired in a seventh computing block 56, and is
input to the second computing block 40.
[0107] In the disturbance model, on the other hand, the estimated
position error e[k] is input to a state estimation block 60 of the
disturbance, and an estimated correction value (right hand side of
Expression (1)) is computed using estimated gains Ld1 (L3, L4, L5).
And the result is added with a state quantity (left hand side of
Expression (1)) from a delay block 62 in an addition block 66, and
an estimated bias value b[k] and estimated disturbance suppression
values z1[k] and z2[k] are acquired, as shown in Expression
(1).
[0108] The estimated values b[k], z1[k] and z2[k] are multiplied by
a state feedback gain (Fd1=F3, F4, F5) in an eighth computing block
68, and the disturbance suppression drive value of the actuator 1
is acquired, as shown in Expression (3).
[0109] The estimated values b[k], z1[k] and z2[k] of Expression (1)
from the addition block 66, on the other hand, are multiplied by an
estimated gain Ad1 (gain of b[k] of Expression (5) and gain of
2.times.2 matrix A) in a ninth computing block 64, the result is
input to the delay block 62, and estimated values b[k+1], z1[k+1]
and z2[k+1] of the next sample are acquired.
[0110] And in an addition block 70, the disturbance suppression
drive value is subtracted from the drive value u[k], and an output
drive value uout[k] of Expression (3) is acquired.
[0111] In other words, the estimated gain L is separated between
the controller model and disturbance model, and the feedback gain F
is separated between the controller model and disturbance model, so
as to design the controller model and disturbance model separately.
The design of the disturbance observer will be described in detail
later.
[0112] Now as an input to an adaptive control system 30-1, the
estimated position error e[k] of the observer is supplied. The
estimated position error e[k] of the observer is a difference value
between an actual position error (r-y[k]) of a computing block 40
and the estimated position x[k] of the observer.
[0113] The disturbance suppression adaptive control system has a
.omega. estimation section 30-1 for estimating a disturbance
frequency according to an adaptive rule, and a table 30-2 (32) for
storing estimated gains L and A and a target gain according to an
estimated frequency (angular frequency .omega. in this case). A
.omega. estimation section 24 calculates an estimated angular
frequency .omega.1[k] from an estimated error e[k] using the
following adaptive Expression (6).
.omega. 1 [ k ] = .omega. 1 [ k - 1 ] + Ka L 5 z 1 [ k ] - L 4 z 2
[ k ] z 1 [ k ] 2 + z 2 [ k ] 2 e [ k ] . ( 6 ) ##EQU00002##
[0114] This adaptive expression has an integration form for
adaptively correcting an estimated angular frequency .omega.1[k-1]
one sample before using estimated disturbance gains L4 and L5,
estimated disturbance values z1[k] and z2[k] and estimation
position error e[k]. Ka is a predetermined gain.
[0115] Based on the estimated value of the addition block 66, the
estimated disturbance values z1[k] and z2[k] are acquired, and
output to the .omega. estimation section 30-1. The .omega.
estimation section 30-1 has a computing section for computing the
second term (Ka - - - e[k] of the .omega. adaptive expression in
Expression (6), a delay section for delaying the estimated
.omega.[k] by one sample, and an addition section for adding the
delayed .omega.(.omega.[k-1]) and the computing result of the
second term in the computing section. In other words, the adaptive
expression of Expression (6) is computed.
[0116] A table 30-2, on the other hand, stores L1, L2, L3, L4 and
L5 according to the value of the estimated disturbance frequency
Fdist (estimated angular frequency .omega.), a11, a12, a21 and a22,
and values of the target gain, as shown in FIG. 14. Based on L1,
L2, L3, L4 and L5 of this table 30-2, L1, L2, L3, L4 and L5 of the
state estimation blocks 42 and 60 are changed according to the
estimated angular frequency. Also based on a11, a12, a21 and a22 of
this table 30-2, a11, a12, a21 and a22 of the ninth computing block
64 (see Expression (5)) are changed according to the estimated
angular frequency.
[0117] In other words, according to the disturbance (angular)
frequency .omega., the disturbance model and the estimated gain are
changed, without changing the state feedback gain F. Here all the
estimated gains of the observer are influenced, not only the
disturbance model for shaping the frequency characteristic in a
notch filter form. In other words, if the disturbance frequency
.omega. or the disturbance model changes, not only the disturbance
estimated gains L4 and L5 of Expression (1), but all of the gains
L1, L2 and L3 of position, velocity and bias are influenced.
Particularly this influence is major if the .zeta.2 value is high
in the pole assignment when the disturbance model is designed in
the form of a shaping filter, that is if the width of the
suppression area in the notch filter form is wide in the frequency
characteristic. Therefore all the estimated gains from L1 to L5
must be changed according to the disturbance frequency. The values
of the estimated gains are calculated by the pole assignment method
and stored in the table 30-2 in advance.
[0118] On the other hand, just like FIG. 12, the gain calibration
block 34 applies a sine wave disturbance SD having a predetermined
frequency according to the gain calibration instruction, detects
signals in the loop before and after applying the sine wave
disturbance, and calculates a gain of the gain multiplication block
26. The sine wave disturbance SD for measurement is supplied from
the addition block 28 to the feedback loop for a position which is
an input of the controller 20 constructed with the disturbance
observer, and positions before and after applying the sine wave
disturbance are observed.
[0119] Here the target gain table is integrated with the parameter
table 30-2, as shown in FIG. 14. In other words, the parameter
table 30-2 stores a target gain G corresponding to a disturbance
frequency Fdist, supplies the target gain G corresponding to the
disturbance frequency Fdist from the .omega. estimation section
30-1 to the gain calibration block 34, and uses it as a reference
of gain calibration in the gain calibration block 34.
[0120] The present embodiment also has a switch SW for stopping the
input of a position error of the controller 20 to the .omega.
estimation section 30-1 during gain calibration. By this, the
.omega. estimation section 30-1 maintains the estimated disturbance
frequency before starting gain calibration during gain calibration.
Therefore during gain calibration, the disturbance suppression
control is executed, but adaptive control is interrupted so that
unnecessary adaptive control is not performed by a sine wave
disturbance for measuring the gain.
[0121] In this way, a desired disturbance suppression function can
be easily added to the controller by constructing the controller
with the disturbance observer. Also the disturbance suppression
adaptive control can be easily implemented by changing the
estimated gain.
[0122] FIG. 15 shows another parameter table, and FIG. 16 is a
diagrams depicting the operation thereof. Compared with the table
30-2 in FIG. 14, a column of measurement frequency Fcal is created
for each disturbance frequency Fdist in this table 30-2, in
addition to the estimated gain and target gain. In other words, the
measurement frequency can also be changed according to the
disturbance frequency Fdist.
[0123] Here, as FIG. 16 shows, a frequency deviated from the
disturbance frequency Fdist by .alpha. is set for the measurement
frequency. Therefore the measurement frequency is shifted from the
disturbance frequency without overlapping, and open loop gain can
be accurately measured.
[0124] FIG. 17 shows another parameter table, and FIG. 18 is a
diagram depicting the operation thereof. Compared with the table
30-2 in FIG. 14, a column of measurement frequency Fcal is created
for each disturbance frequency Fdist in this table 30-2, in
addition to the estimated gain and target gain.
[0125] In FIG. 17, just like FIG. 11, if the disturbance frequency
Fdist is the measurement frequency fsd, the measurement frequency
is changed to fsd+.alpha. (.alpha..noteq.0) or fsd-.alpha.. Also,
as FIG. 18 shows, the measurement frequency is set to be higher
than the disturbance frequency Fdist (e.g. fsd+.alpha.) if the
disturbance frequency is lower than this measurement frequency fsd,
and if the disturbance frequency is higher than this measurement
frequency fsd, the measurement frequency is set to be lower than
the disturbance frequency Fdist (e.g. fsd-.alpha.).
[0126] In other words, the measurement frequency is shifted from
the disturbance frequency, but if the shifting range is
inappropriate, the gain calibration becomes difficult. For example,
if the measurement frequency is set low, the measurement frequency
extends beyond the lower limit of the servo band, and if the
measurement frequency is set high, the measurement frequency
extends beyond the upper limit of the servo band, so measurement
itself becomes difficult.
[0127] Therefore in an area where the disturbance frequency is low,
the measurement frequency is set to higher than the disturbance
frequency Fdist (e.g. fsd+.alpha.), and in an area where the
disturbance frequency is high, the measurement frequency is set to
lower than the disturbance frequency Fdist (e.g. fsd-.alpha.).
[0128] And in the column of the disturbance frequency Fr*K(=fsd),
two measurement frequencies, (fsd+.alpha.) and (fsd-.alpha.), are
set, and the corresponding target gains GK1 and GK2 are set. By
this, the interpolation of the gain can be switched at the
boundary.
Designing Disturbance Observer
[0129] Now the design procedure of the disturbance observer will be
described. The observer control system when the actuator 1 is a
double integral model is given by the following analog expression,
Expression (7).
s ( x v ) = ( 0 1 0 0 ) ( x v ) + Bl m ( 0 1 ) u + ( L 1 L 2 ) ( y
- x ) y = ( 1 0 ) ( x v ) u = - ( Fx Fv ) ( x v ) } ( 7 )
##EQU00003##
[0130] In Expression (7), `s` is a Laplace operator, `x` is an
estimated position, `v` is an estimated velocity, `y` is a current
position, `r` is a target position, L1 and L2 are estimated gains
of position and velocity respectively, `u` is a drive current, and
B1/m is a force constant of the actuator 1.
[0131] This control system has a sensitivity function 1/(1+CP), and
the disturbance suppression for this sensitivity function is
defined by the second degree filter in the following Expression
(8).
s 2 + 2 1 .omega. 1 s + .omega. 1 2 s 2 + 2 2 .omega. 2 s + .omega.
2 2 ( 8 ) ##EQU00004##
[0132] The disturbance model, of which denominator is a numerator
of the shaping filter, is given by the following Expression
(9).
1 s 2 + 2 1 .omega. 1 s + .omega. 1 2 ( 9 ) ##EQU00005##
[0133] There are three possible methods to set up this disturbance
model in the observer of the original controller (Expression
(7)).
[0134] The first method is setting up the disturbance model in
Expression (9) as is. In other words, this is a second degree
filter, so if the estimated state quantities of a disturbance are
z1 and z2, and the estimated gains of disturbance are L3 and L4,
the observer control system is expression by Expression (10).
s ( x v z 1 z 2 ) = ( 0 1 0 0 0 0 1 0 0 0 0 1 0 0 - .omega. 1 2 - 2
1 .omega. 1 ) ( x v z 1 z 2 ) + Bl m ( 0 1 0 0 ) u + ( L 1 L 2 L 3
L 4 ) ( y - x ) y = ( 1 0 0 0 ) ( x v z 1 z 2 ) u = - ( Fx Fv K 0 )
( x v z 1 z 2 ) K = m Bl } ( 10 ) ##EQU00006##
[0135] The second method is dispersing the term of the square of
.omega.1, and Expression (11) is acquired by transforming
Expression (10).
s ( x v z 1 z 2 ) = ( 0 1 0 0 0 0 1 0 0 0 0 .omega. 1 0 0 - .omega.
1 - 2 1 .omega. 1 ) ( x v z 1 z 2 ) + Bl m ( 0 1 0 0 ) u + ( L 1 L
2 L 3 L 4 ) ( y - x ) y = ( 1 0 0 0 ) ( x v z 1 z 2 ) u = - ( Fx Fv
K 0 ) ( x v z 1 z 2 ) K = m Bl } ( 11 ) ##EQU00007##
[0136] The third method is inverting the sign of .omega.1 in
Expression (11), and is given by Expression (12).
s ( x v z 1 z 2 ) = ( 0 1 0 0 0 0 1 0 0 0 0 - .omega. 1 0 0 .omega.
1 - 2 1 .omega. 1 ) ( x v z 1 z 2 ) + Bl m ( 0 1 0 0 ) u + ( L 1 L
2 L 3 L 4 ) ( y - x ) y = ( 1 0 0 0 ) ( x v z 1 z 2 ) u = - ( Fx Fv
K 0 ) ( x v z 1 z 2 ) K = m Bl } ( 12 ) ##EQU00008##
[0137] Design is possible by any of these methods. The second and
third methods are effective particularly when the model is
transformed into a digital control system. In other words, the two
state variables z1 and z2 are balanced, and the values of the
estimated gains L3 and L4 of the observer for the two state
variables are not very far apart.
[0138] At this time, the values of the estimated gains L1, L2, L3
and L4 are designed by specifying the poles combining the pole of
the shaping filter of Expression (8) (derived from denominator=0 in
Expression (8)) and the poles used for designing the original
observer control system.
[0139] The observer control system combining the second degree
filter shaping and the conventional steady state bias estimation is
given by the following Expression (13).
s ( x v b z 1 z 2 ) = ( 0 1 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0
.omega. 1 0 0 0 - .omega. 1 - 2 .omega. 1 ) ( x v b z 1 z 2 ) + Bl
m ( 0 1 0 0 0 ) u + ( L 1 L 2 L 3 L 4 L 5 ) ( y - x ) y = ( 1 0 0 0
0 ) ( x v b z 1 z 2 ) u = - ( Fx Fv K K 0 ) ( x v b z 1 z 2 ) K = m
Bl } ( 13 ) ##EQU00009##
[0140] In this way, the filter form for shaping is considered
first, then the disturbance model is added to the observer in
designing. Therefore forms can be freely shaped without being
limited by the physical response characteristic of the original
disturbance model.
[0141] Thus far description dealt with analog design. But in order
to design a digital control system, on the other hand, the
disturbance model is created in an analog space and an enlarged
model is constructed, then after the enlarged model is transformed
(digitized) in the digital space, the pole assignment is specified
in the digital space.
[0142] When the disturbance model has the characteristic of the
second degree filter, if the enlarged model is transformed into a
discrete system, both of the two variables z1 and z2 of the
disturbance model in matrix A for designing the estimated gain of
the observer influence the actuator 1.
[0143] Therefore correction is made so that only one of the
variables of the disturbance model influences the actuator 1, more
specifically, so that only a variable, the same as that of analog
design, influences the actuator 1. In other words, after
digitizing, the enlarged model is corrected.
[0144] Specifically, when the analog model in the form of
Expression (11), using the second degree filter, is digitized (that
is z-transformation is performed and the result is converted into
SI units), the following Expression (14) is established.
z ( x [ k ] v [ k ] z 1 [ k ] z 2 [ k ] ) = ( 1 T A 13 A 14 0 1 A
23 A 24 0 0 A 33 A 34 0 0 A 43 A 44 ) ( x [ k ] v [ k ] z 1 [ k ] z
2 [ k ] ) + Bl m ( T 2 / 2 T 0 0 ) u [ k ] y = ( 1 0 0 0 ) ( x [ k
] v [ k ] z 1 [ k ] z 2 [ k ] ) } ( 14 ) ##EQU00010##
[0145] In Expression (14), z is a Z transformer and T is a sampling
cycle. Here matrix A, that is A13, A14, A23 and A24, is focused on.
Neither A14 nor A24 become "0" merely by digitizing. In other
words, both of the two variables z1 and z2 of the disturbance model
in matrix A for designing the estimated gain of the observer
influence the actuator 1.
[0146] Therefore after digitizing the analog model, coefficients
with which the state variables z1 and z2 of the disturbance model
in matrix A influence the actuator 1, are replaced.
[0147] In the case of the examples of Expression (14), matrix A is
corrected as the following Expression (15).
A 14 = A 24 = 0 A 13 = T 2 / 2 A 23 = T } ( 15 ) ##EQU00011##
[0148] In the digital control system, the unit of distance is a
track, the current value is normalized with the maximum current as
"1", and the velocity and acceleration are not in second units, but
must be normalized by a sampling frequency.
[0149] In the same way, if the observer in analog format in
Expression (13) is transformed into the format of the current
observer, Expression (16) is established.
( x ( k ) v ( k ) b ( k ) z 1 ( k ) z 2 ( k ) ) = ( x ( k ) v ( k )
b ( k ) z 1 ( k ) z 2 ( k ) ) + ( L 1 L 2 L 3 L 4 L 5 ) ( y ( k ) -
x ( k ) ) u ( k ) = - ( F 1 F 2 F 3 F 4 F 5 ) ( x ( k ) v ( k ) b (
k ) z 1 ( k ) z 2 ( k ) ) ( x ( k + 1 ) v ( k + 1 ) b ( k + 1 ) z 1
( k + 1 ) z 2 ( k + 1 ) ) = ( 1 1 1 / 2 1 / 2 0 0 1 1 1 0 0 0 1 0 0
0 0 0 a 11 a 12 0 0 0 a 21 a 22 ) ( x ( k ) v ( k ) b ( k ) z 1 ( k
) z 2 ( k ) ) + Bl m 1 Lp T 2 ( 1 / 2 1 0 0 0 ) u ( k ) } ( 16 )
##EQU00012##
[0150] In this way, if the disturbance model is designed to be a
separate configuration, Expression (16) can be set up with the
disturbance model separately, as shown in FIG. 12.
[0151] In other words, in the comparison of Expression (16) with
Expression (1) to Expression (5), Expressions (2) and (4) are
Expression (16), wherein the model of the controller is
independent, and Expression (3) and (5) are Expression (16),
wherein the disturbance model 50 is separated.
Fourth Embodiment
[0152] FIG. 19 is a block diagram depicting the fourth embodiment
of the position control system for suppressing disturbance which
the MCU 14 in FIG. 1 executes. This position control system is a
control system for detecting a disturbance frequency and
suppressing the sine wave disturbance having a predetermined
frequency. In FIG. 19, composing elements the same as those in FIG.
4 are denoted with the same reference symbols.
[0153] A position error e between a target position r supplied from
an interface circuit 11-1 (in HDC 11) and an observation position y
is computed in a computing unit 24. This position error e is input
to a controller 20 (Cn) which performs feedback control. The
controller 20 outputs a control current value Un by a known PID
control, PI control+LeadLag, and observer control.
[0154] A frequency estimation unit (.omega. estimation) 30 for
estimating the frequency of disturbance, and a compensator (Cd)
20-1 for suppressing the disturbance having a specific frequency by
adaptive control, are added to this controller 20.
[0155] A sum U of an output Un of the controller 20 (Cn) and an
output Ud of the compensator 20-1 (Cd) is determined in an addition
block 20-2, and is supplied to a plant 22 (1, 3) via a gain
multiplication block 26. By this, the position of the head 3 driven
by the actuator 1, which is the control target 22, is controlled so
as to follow up the disturbance. In other words, the device is
vibrated by the distance, so the position of the head 3, with
respect to the magnetic disk 4, is also controlled to follow up the
disturbance, so the position relationship of the head 3 and the
magnetic disk 4 does not change.
[0156] This frequency estimation unit 30 estimates an angular
frequency .omega. (=2.pi.f) of the disturbance based on a position
error e, as shown in FIG. 12, and supplies it to the transfer
function of the disturbance frequency suppression of the
compensator 20-1. The compensator 20-1 calculates the recurrence
formula (adaptive control expression) of the sine wave based on the
position error e and this estimated angular frequency .omega., and
a compensating current output Ud is calculated.
[0157] In this way, in order to handle disturbance with an unknown
frequency in a certain range, the frequency of disturbance is
detected and the unknown frequency is suppressed. As a method for
estimating an unknown frequency and suppressing disturbance of the
unknown frequency, assuming the recurrent formula of a sine wave,
or the above mentioned method of supplying adaptive rule based on
the error signal e and correcting the drive quantity of the control
target, can be used. Also a method of estimating an unknown
frequency based on an error signal e, generating a disturbance
suppression signal on the position level, correcting the error
signal, and inputting it into the controller, can also be used.
[0158] Here the interface circuit 11-1 receives a disturbance
suppression frequency from the outside, and sets it in the
frequency estimation unit 30 as an initial value of the frequency
estimation unit 30 (initial value of the angular frequency of
disturbance). Therefore the compensator 20-1 performs adaptive
control from this initial value.
[0159] In other words, the initial value of the frequency
estimation unit 30, which is normally based on the assumption that
the disturbance frequency is unknown, is set at the center of the
follow up range, with a position error e, and gradually reaches the
disturbance frequency, but in the present embodiment, a known
disturbance frequency is set as the initial value, so position
control starts directly with the known disturbance frequency, and
even if the frequency changes thereafter, the estimated frequency
follows up from there.
[0160] In this position control system where disturbance frequency
is set from the outside, a gain calibration block 34 and a target
gain table 32 are installed. The gain calibration block 34 applies
a sine wave disturbance SD having a predetermined frequency
according to a gain calibration instruction, detects signals in a
loop before and after applying in the sine wave disturbance, and
calibrates the gain of the gain multiplication block 26. In FIG.
19, the sine wave disturbance SD for measurement is supplied from
the addition block 28 to the feedback loop according to a position
which is an input of the controller, and positions before and after
applying the sine wave disturbance are observed.
[0161] The target gain table 32 stores a target gain G
corresponding to a disturbance frequency Fdist, supplies the target
gain G corresponding to the disturbance frequency Fdist from the
.omega. estimation section 30 to the gain calibration block 34, and
uses it as a reference of gain calibration in the gain calibration
block 34.
[0162] The present embodiment also has a switch SW for stopping the
input of a position error to the .omega. estimation section 30
during gain calibration. By this, the .omega. estimation section 30
maintains the estimated disturbance frequency before starting gain
calibration during gain calibration. Therefore during gain
calibration, the distance suppression control is executed, but
adaptive control is interrupted so that unnecessary adaptive
control is not performed by a sine wave disturbance for measuring
the gain.
[0163] In this way, the present invention can also be applied to
the position control system where disturbance frequency is set from
the outside.
Fifth Embodiment
[0164] FIG. 20 is a block diagram depicting the fifth embodiment of
a position control system for suppressing disturbance which the MCU
14 in FIG. 1 executes. This position control system is a control
system for detecting a disturbance frequency and suppressing sine
wave disturbance having a predetermined frequency. In FIG. 20,
composing elements the same as those in FIG. 4, FIG. 12 and FIG. 19
are denoted with the same reference symbols.
[0165] In FIG. 20, a position error e between a target position r
supplied from an interface circuit 11-1 (in HDC 11) and an
observation position y is computed in a computing unit 24. This
position error e is input to a controller 20 (Cn) which performs
feedback control. The controller 20 outputs a control current value
Un by a known PID control, PI control+LeadLag, and observer
control.
[0166] A frequency converter 30 for converting a frequency of
disturbance to a corresponding angular frequency, and a compensator
(Cd) 20-1 for suppressing disturbance having a specific frequency
by adaptive control are added to this controller 20.
[0167] A sum U of an output Un of the controller 20 (Cn) and an
output Ud of the compensator 20-1 (Cd) is supplied to a plant 22
(1, 3) via an addition block 20-2 and a gain multiplication block
26. By this, position of the head 3 driven by the actuator 1, which
is the control target 22, is controlled so as to follow up the
disturbance. In other words, the device is vibrated by the
disturbance, so the position of the head 3 with respect to the
magnetic disk 4 is also controlled to follow up the disturbance,
and the position relationship of the head 3 and the magnetic disk 4
does not change.
[0168] This frequency converter 30 supplies an angular frequency
.omega. (=2.pi.f) to a transfer function of the disturbance
frequency suppression of the compensator 20-1. The compensator 20-1
calculates the recurrence formula (adaptive control expression) of
the sine wave based on the position error e and this estimated
angular frequency .omega., and a compensating current output Ud is
calculated.
[0169] In this way, in order to handle disturbance with a frequency
in a certain range, disturbance frequency which changes is
suppressed according to the frequency of disturbance. As this
method, assuming a recurrence formula of a sine wave, or the above
mentioned method of introducing an adaptive rule based on the error
signal e and correcting the drive quantity of the control target,
can be used. Also a method of estimating an unknown frequency based
on an error signal `e`, generating a disturbance suppression signal
on the position level, correcting the error signal, and inputting
it into the controller, can also be used.
[0170] In the present embodiment, the interface circuit 11-1
receives a disturbance suppression frequency from the outside, and
sets it in the frequency converter 30. Therefore the compensator
20-1 performs adaptive control from this initial value (angular
frequency).
[0171] Since a known disturbance frequency is set as an initial
value, position control starts directly with the known disturbance
frequency, and even if the frequency changes thereafter, the
compensating current Ud of the compensator 20-1 follows up from
there.
[0172] In this way, this position control system has a means of
changing an internal constant (angular frequency in the case of
FIG. 5 and FIG. 6), or the configuration according to the set value
of the disturbance frequency to be selectively suppressed, and the
disturbance frequency can be referred to or set from the outside
via the interface 11-1.
Other Embodiment
[0173] In the above embodiments, the position control device was
described using an example of a head positioning device of a
magnetic disk device, but the present invention can also be applied
to other medium storage devices, such as an optical disk device or
other devices to control the position of an object. The number of
disturbance frequencies can be arbitrary according to necessity,
and the number of disturbance models to be used can be arbitrary
accordingly. The embodiments were described using a second degree
filter, but a first degree filter or a combination of a first
degree filter and a second degree filter may be used according to
the frequency which need be suppressed.
[0174] The present invention was described using embodiments, but
the present invention can be modified in various ways within the
scope of the essential character thereof, and these variant forms
shall not be excluded from the scope of the present invention.
[0175] Because of disturbance suppression control, open loop gain
is calibrated using a target gain according to a disturbance
frequency even if a loop characteristics of a feedback controller
changes, so the open loop gain can be calibrated without
interrupting the disturbance suppression control. Therefore the
open loop gain can be accurately calibrated without being affected
by disturbance, and accurate position control is possible.
* * * * *