JPH1116308A - Magnetic disk drive - Google Patents

Abstract

(57) [Summary] In a magnetic disk drive, when external force estimation is performed using external force estimation means including a disturbance observer, gain fluctuation, model error, and position signal erroneous detection increase the transient response of the external force estimated value. Fluctuation, and deteriorates the transient response at the time of moving the head. As a means for reducing the deterioration of the transient characteristic of a disturbance estimation value due to a model error, a limiter for limiting a change amount of an observer estimation value between one sample is applied. The estimate limited by the limiter can suppress excessive fluctuation of the estimate and improve the transient response of the head position. When an error in the position error signal is detected, the position error signal in the sample is estimated using the past position error signal and the VCM control signal, thereby improving the accuracy of external force estimation in the subsequent samples.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic disk drive, and more particularly to an operation for causing a magnetic head to follow a target track. Further, X used in a semiconductor exposure apparatus and the like
The present invention relates to a magnetic disk device provided with a high-precision positioning device such as a Y stage and a magneto-optical disk device.

[0002]

2. Description of the Related Art As a positioning mechanism of a head actuator of a magnetic disk device or the like, there are a linear actuator called a direct-acting type and a rotary actuator called an oscillating type, both of which are guided by rolling bearings.

[0003] The frictional force at the bearing portion exhibits characteristics such that it is supported by a viscoelastic body such as a spring force and a viscous frictional force. Further, the gradient of the frictional force with respect to the displacement changes every time the moving direction of the rolling bearing is reversed. And the gradient of the frictional force becomes steeper as the displacement becomes smaller. As a result, a change in the frictional force with respect to the head displacement exhibits a hysteresis characteristic.

In Japanese Patent Laid-Open Publication No. Hei 3-30156, as a method of compensating for an external force applied to a positioning control motor including a frictional force, an estimated position signal calculated by integrating a motor drive signal and a signal recorded on a disk surface are used. A method of correcting a head position shift due to an external force by having a configuration in which a signal obtained by multiplying a difference e ₀ of the head position signal obtained from the servo information by a certain value is added to the operation amount, Performs measurements at several points on the disk surface, records the measurement results, creates a table of external force correction values, and compensates for external forces by feedforward by referring to the table values according to the track position on the disk A method for doing so is disclosed.

[0005]

The demand for precise positioning of the head is becoming more severe with miniaturization and higher recording density. Non-linear friction generated in a rolling bearing portion caused by minute position fluctuation of an actuator is required for precise positioning. It has been an obstacle to In the external force correction table reference method described in the prior art described above, a correction error occurs, and the influence on the positioning characteristics becomes more conspicuous as the size and weight are reduced. Therefore, it is important to estimate and compensate for the frictional force in real time by an external force estimating means including a disturbance observer.

There are two problems when external force compensation is performed by a disturbance observer. The first problem is to improve the accuracy of disturbance estimation with respect to a model error. For example, a head positioning actuator of a magnetic disk drive exhibits a characteristic of a secondary system having a natural frequency in a low frequency region due to viscoelastic friction of a carriage bearing, and its natural frequency and attenuation fluctuate depending on head displacement. Then, an error occurs between the control target and the model. Also, the MR head for signal reproduction can detect a change in a magnetic field as a change in resistance value with high sensitivity due to a magnetoresistive effect, which is effective for improving the recording density, but as a drawback, the reproduction output in the microtrack direction. The head position deviation amount from the track center and the detected position error output are not proportional, and the position detection sensitivity is high at the track center and decreases as the position approaches the periphery of the track. Therefore, the loop gain as a control system varies within the track width.

Further, the motor thrust constant changes over time due to a temperature change or the like, so that the loop gain of the control system is changed. Further, there is a model error of a resonance phenomenon in a high frequency region due to bending, twisting, and the like of the actuator. When trying to estimate the external force acting on the actuator using an observer under such conditions, the gain fluctuation and the model error greatly change the transient response of the external force estimated value.

When this estimated value is added as a compensation signal to the actuator drive signal, the transient response at the time of moving the head is deteriorated, or an unnecessary high-frequency component is given to the current.
Excite higher order resonance modes.

A second problem is to improve the accuracy of external force estimation for erroneous detection of the head position. For example, when an MR head is used, servo position information may not be detected or an incorrect track number may be detected due to the occurrence of thermal noise (TA: thermal asperity). Such a phenomenon causes an erroneous estimated value of the external force for several samples thereafter during estimation using the observer.

[0010]

In order to solve the above two problems when the disturbance observer is applied in real time,
According to the present invention, for the first problem, a limiter (observer change amount limiter) for limiting a change amount of an estimated value of an observer is applied as means for reducing deterioration of a transient characteristic of a disturbance estimated value due to a model error. I do. The limiter can suppress an excessively large change in the estimated value and improve the transient response of the head position.

With respect to the second problem, a position error is determined by estimating a position error signal using a past position error signal and a VCM control signal, and a position error signal estimated value is determined in accordance with the determination result. Is used to improve the external force estimation accuracy. At this time, the fluctuation of the external force estimated value due to the estimation accuracy of the position error signal is reduced by using the above-described limiter that limits the amount of change of the observer estimated value.

[0012]

FIG. 13 is a perspective view of a magnetic disk drive. A spindle motor 8 that rotates at a predetermined speed is fixed to the base 10. Spindle motor 8
Is fixed to a magnetic disk 4 as a recording medium. A pivot bearing 3 supported by a base 10 is provided in a lateral direction of the magnetic disk 4 held by the spindle motor 8 so as to be parallel to the spindle motor axis.

A head support spring 7, a load arm 6 and a carriage 5, which are holding members of the magnetic head 1, are swingably fixed to a pivot bearing 3. Power for moving the magnetic head 1 is a voice coil motor (VCM) 2
Caused by A drive coil (not shown) which is a component of the voice coil motor 2 is fixed to the carriage 5.

A magnetic circuit (not shown) which is another component of the voice coil motor 2 includes a yoke (not shown) fixed to the base 10 and a magnet (not shown) fixed to the yoke. Become.

One end of an FPC (flexible print cable) 9 attached in a curved shape is fixed to the carriage 5 and the other end is fixed to the base 10, and a drive circuit for the spindle motor 8 and the voice coil motor 2 outside the apparatus. And the like. F
A read / write IC (not shown) is mounted on the PC 9, and a head signal line drawn from the magnetic head 1 is connected to an FP.
It is connected to the read / write IC via C9 and is electrically connected to the control circuit board.

The end of the coil wire of the voice coil motor 2 is also connected to the FPC 9. The control circuit board is fixed below the base 10. The magnetic disk 4 fixed to the spindle motor 8 and the carriage 5 supported by the pivot bearing 3, the load arm 6 fixed to the carriage 5, the magnetic head 1 held by the load arm 6, and the voice coil motor 2 The magnetic circuit and the coil, the head signal line of the magnetic head 1, the FPC 9 and the like are accommodated in a closed space fixed by a cover (not shown), the base 10, and screws.

The head actuator is an actuator for positioning the magnetic head 1, and includes a magnetic head 1, a load arm 6, a carriage 5, a pivot bearing 3,
It is composed of VCM2 and FPC9.

A first embodiment of the present invention will be described with reference to FIG. FIG. 1 is a block diagram showing the configuration of the magnetic head positioning control system.

Data tracks (not shown) are recorded concentrically on the magnetic disk 4, and sector servo signals (not shown) are recorded at regular intervals in the circumferential direction and radially as viewed from the center of the disk. The servo signal 30 read by the head 1 every time a sector is detected is amplified by the amplifier 11 and sent to the demodulation circuit 12. The demodulated signal 31 is captured by the AD converter 13 as a digital signal 32 into a microprocessor (not shown).

In the position signal generating means 14, the demodulated signal 32
From the main position error PESN (not shown) and the sub position error signal PESQ (not shown) having a phase difference of 90 degrees from the main position error PESN (not shown), and a position signal 33 obtained by adding a track number generated from the demodulated signal 32 to these signals. create. By subtracting the position signal 33 from the target position signal 35, the position error signal 34
Create

If the position error is large based on the position error signal 34, that is, if the head is located far from the target track, a seek control system (speed control system) is used. When the head approaches the target track, the switches 15 and 23 are switched to use the following control system (position control system).

At the time of following, the position error signal is taken into the positioning compensator 20a. At the time of seeking, the speed signal generating means 16 calculates the speed based on the position error signal 34 by using a method such as detecting a difference between the position signals for each sample, and outputs a speed signal 39.

In the speed command generating means 17, a speed target value 40 corresponding to the distance to the target value 19 is created based on the position error signal 34, and a difference signal 41 between the speed target value 40 and the speed signal 39 is calculated by the positioning compensation. Is taken into the vessel 20b.

The acceleration compensation generating means 18 responds to a speed command generated based on the position error signal 34 by
The acceleration compensating signal 37 for providing the operation amount required for acceleration and deceleration in feedforward is generated, and the positioning compensator 2
0b is added to the output 38. The positioning compensator 20a uses phase lead / lag compensation for securing stability and improving servo characteristics in a low frequency region, and the compensator 20b sets a servo system band only by gain adjustment.

The external force correction table 22 includes an external force compensation signal 43 created by measuring an external force due to an FPC (flexible print cable), an external force caused by a frictional force at a bearing, and the like for each track position. In the table, the output of the external force correction table 22 is
To generate a VCM operation signal 46.

In the external force compensating means 21, the position error signal 34
And the operation signal 45 are used to calculate the external force estimated value 42 to compensate for a time-varying external force that cannot be completely compensated by the external force correction table 22.

The external force compensating means 21 incorporates means for limiting the output of the external force compensation value of the present invention. The VCM operation signal 46 is converted into an analog operation amount 47 by the DA converter 24, a VCM drive current 48 is generated by the drive circuit 25, and the head is moved by supplying a current to the VCM 2. By the above means, the head can be positioned on the target track.

Next, the calculation performed by the external force compensating means and the details of the observer change amount limiter will be described.

FIG. 2 is a functional block diagram including the external force compensating means of the present invention. The derivation of a variable block appearing in the functional block diagram of FIG. 2 will be described. The state equation of the control target in the continuous time system is expressed by the following equation.

[0030]

(Equation 1)

X is a state vector, u is an operation amount, d ₁ is an external force, y is a position error signal, x is position, v is velocity, m is the equivalent mass of the head actuator, and K ₁ is the dac output to driving force. gain, K ₂ is the gain until the position error signal from the position.

[0032] In Equation 1, but chose to 2 × 2 matrix as A _p, since the formula is the case of n × n and more generalized.
B _p and B _dp are n × 1, and C _p is 1 × n. An observer for estimating the external force applied to the carriage based on this model is expressed by the following equation.

[0033]

(Equation 2)

Here, d ₁ is the estimated value of the external force applied to the carriage, L is the feedback gain of the observer, and H is the disturbance estimation gain. Here, L is the characteristic equation for determining the observer poles, p _1, p ₂ poles, ..., if you specify p _n,

[0035]

| Z · I _n −A _d + L · C _p | = (z−p ₁ ) (z−p ₂ )... (Z−p _n ) (Expression 3) Is determined as a simple vector.
Derivation of H finds a transfer function G _dy (z) from the external force d ₁ to the position error signal y,

[0036]

(Equation 4)

H is determined so that the DC component of the transfer function G _dy (z) becomes zero. That is, H satisfying G _dy (0) = 0
Ask for. Here, the state equation and output equation of the compensator are

[0038]

E (k) = r (k) −y (k) X _c (k + 1) = A _c X _c (k) + B _c e (k) (Equation 5) u _e (k) = C _c · X _c (k) + D _c · e (k).

The signal flow of the functional block diagram of FIG. 2 will be described based on the parameters described above. The VCM operation signal u (k) 51 is converted into an analog signal via the DA converter 24, and is input to the system represented by Expression 1 as an operation amount u52. As a result, the system outputs a position error signal y53. The disturbance estimating means is configured as an observer represented by Expression 2, and an operation amount u (k) 51 input to the head actuator is input to the observer.

The observer system calculates the estimated position error signal y
_e (k) 55 is output, and a difference signal e ₀ (k) 56 from the detected position error signal y (k) 54 from the actual plant is calculated. e ₀ (k) is converted to the signal e ₁ (k−1) one sample before.
The signal e ₁ (k) 57 is generated by limiting the width of the change with a limiter 64 in comparison with the signal e ₁ (k) 58. E found
₁ (k) 57 is multiplied by the observer feedback gain vector L obtained by Expression 3 and fed back to the observer. Further, a disturbance compensation signal w (k) 59 for canceling the external force d ₁ 60 input to the actual system is generated by multiplying e ₁ (k) 57 by the gain H obtained from Expression 4.

In order to position the head at the target position, the detected position error signal y (k) 54 and the target position r (k)
The operation amount u is calculated by calculating a difference signal e (k) 62 from the signal 63 and feeding it back to the compensator represented by Expression 5.
_e (k) 61 is calculated. The operation signal u (k) 5 is obtained by subtracting the disturbance compensation signal w (k) 59 from u _e (k).
Find 1 The above is the details of the operation of the external force compensating means of the present invention.

Next, FIG. 3 shows the transient response of the detected position error signal y (k) 54 and the external force compensation value w (k) 59 when the external force compensating means is operated. An external force of 1 mN is applied to the actuator, and as a model error, the plant model of the observer has a loop gain of 1 with respect to the actual plant.
Based on the output sensitivity of the MR head, the position error signal detection sensitivity is set to be different depending on the position within the track based on the output sensitivity of the MR head, and a plant model having a different loop gain depending on the position on the disk is assumed. When there is a limiter, the fluctuation of the estimated external force caused by the fluctuation of the gain of the plant is small, and the setting of the position error signal is fast. On the other hand, when there is no limiter, the fluctuation of the disturbance estimation value appears greatly,
As a result, the position error signal oscillates and settling is slow.

When the limiter applies a larger external force to a control system in a stepwise manner, the external force estimated value changes only with a constant width with respect to the sample, so that the external force estimation is delayed, The response of the position error signal y is delayed.

Therefore, when setting the width of the limiter, it is necessary to grasp the range of the absolute value of the external force acting on the actual plant and to give the limit value of the limiter which is not too small. Conversely, by applying an external force expected to exceed the range of the limiter, it is possible to determine whether the limiter is used.

Here, consider the time response of the VCM operation signal u47 to the input disturbance 60 when the configuration of the present invention is adopted. The input disturbance is a pulse-like voltage command, and the voltage level takes two steps of 0 to VA and 0 to VB. Voltage level V
There is a threshold value VS between A and VB that defines the limit value of the observer output limiter. The disturbance is added to the amplifier at the preceding stage of the VCM current amplifier, or the adder formed by the operational amplifier, the resistor R, and the power supply shown in FIG. 4 is connected in series between the drive circuit 25 and the D / A converter 24. Insert and add via a resistor.

FIG. 5 shows a signal waveform of the input disturbance 60. First, an input disturbance pulse is applied at a level from 0 to VA. However, VA> VS. VCM operation signal 4 at this time
7 is shown in FIG. During d = 0, the operation signal indicates a constant value Voff. When the input disturbance 60 indicates VA, the operation signal 47 first changes by several samples in steps of the threshold value VS, and thereafter the width of the change gradually decreases, and an operation signal of a level Voff-VA that cancels the level VA of the input disturbance is generated. I do.

When the input disturbance shows 0 again after a certain period of time, the operation signal 47 first changes by several samples at the step of the threshold value, and thereafter the width of the change gradually decreases and returns to the level Voff again. The same response is repeated every time a disturbance pulse arrives. After the disturbance pulse rises from d = 0 to d = VA, the time required for the output of the operation signal 47 to reach Voff−VA (for example, the time required for u> 0.9 ^* (Voff−VA)) to be tA. And

Next, an input disturbance pulse is applied at a level from 0 to VB. However, VB <VS. The input disturbance pulse at this time is shown in FIG. 5, and the time response of the operation signal 47 is shown in FIG.

During d = 0, the operation signal 47 has a constant value Voff
Is shown. When the input disturbance d indicates VB, the operation signal 47 changes in steps smaller than the threshold value VS in the first sample, and reaches the level Voff-VB with the step width smaller than that of the immediately preceding sample after the next sample.

When the input disturbance 60 shows 0 again after a certain period of time, the operation signal 47 changes in steps smaller than the threshold value in the first sample, and the level Voff becomes smaller while the step width after the next sample is made smaller than that of the immediately preceding sample. To reach.

The pulse of the input disturbance 47 is changed from d = 0 to d = V.
B, the output of the manipulated variable 47 becomes Voff−VB
(For example, u> 0.9 ^* (Voff−
Assuming that the time required to reach VB) is tB, the relationship between tA and tB is tA = tB when there is no observer output limiter, whereas in this case, tA> tB
It shows the characteristic that it becomes tB.

Here, the amount of change in the disturbance estimation value between one sample is limited as the operation method of the limiter. However, the absolute value of the disturbance estimation value is limited so that the estimation value that cannot be realized in reality is limited. You may do it. Further, the change amount of the change amount of the disturbance estimation value, that is, the second-order differential value may be limited. Further, the higher-order differential value may be limited.

The limit value of the limiter may be made variable. The limit value may be increased when the distance to the target track is large, and may be reduced when the distance to the target track is short.

In the first embodiment described above, by applying a limiter (observer change amount limiter) for limiting the amount of change of the estimated value to the disturbance observer, an excessively large change in the estimated value is suppressed, and the head position is adjusted. It is shown that the transient response can be improved.

A second embodiment of the present invention will be described with reference to FIG. The second embodiment differs from the first embodiment of the present invention in that:
It is configured by adding a block for detecting an abnormal position signal and correcting the position signal.

The detected position error signal 34 is represented by an operation amount u
The difference signal 71 of the estimated position error signal estimation value 69 is calculated by the position error signal estimation means 65 based on the position error signal _a 44 and the position error signal 68. The difference signal 71 is input to the position signal erroneous detection judging means 66, and when the erroneous detection is judged, the switch 67 is switched, and the position signal estimated value 69 is fed back as a position error signal 68 to the magnetic disk controller and the observer. Position error signal 6
8 is the position error signal 34 of the first embodiment.

Here, as an erroneous detection determination method, if the difference signal differs by x tracks or more, it is regarded as erroneous detection. The criterion x is given according to the distance to the target track. For example, it is given as a function of speed when seeking, x is large when the speed is high, x is small when the speed is low, and is constant when the vehicle reaches the target track. To give.

Next, a processing algorithm when an abnormal position signal is detected is shown by a flowchart of FIG. If the difference signal 71 between the estimated position signal 69 and the detected position signal 34 is larger than the erroneous position detection criterion set value x, the estimated position error signal 69 is output as the position error signal 68.

Based on the output position error signal 68, the output of the positioning compensator and the external force are estimated by the method of the first embodiment, and the VCM operation signal is calculated. When the estimated position error signal is used, the external force compensation signal w (k) is set to the value of the previous sample. Next, some estimation algorithms of the position error signal will be described.

For example, assuming a constant acceleration,

[0061]

Y (k) = 3 · y (k−1) −3 · y (k−2) + y (k−3) (Equation 6) When the manipulated variable _ua (k-1) 44 of the previous sample is used,

[0062]

(Equation 7)

Is obtained. Here, T _s is the sampling time. Furthermore, plant model formula

[0064]

Equation 8] _{y = c p (z · I} n -A d) Expand ^{_{-1 B d · u a (z}} ) ... ( 8), when the sampled representation,

[0065]

(Equation 9)

Is obtained. The coefficient of each term in Equation 9 is represented by Equation 8
By expanding the pulse transfer function of

Next, the effect of this embodiment will be described with reference to FIG. FIG. 9 shows a target track 20 (μ)
m) Response of a position error signal when a position signal is erroneously detected in the foreground.

As shown in FIG. 9, even if an erroneous detection is made, the subsequent position error signal is not significantly affected. That is, it indicates that the voice coil motor operation signal 47 calculated based on the position error signal is not adversely affected.

FIG. 10 shows the response of the position error signal, the disturbance estimation value, and the current when Expression 9 is used. In the figure, the external force estimation value at the time of occurrence of the erroneous detection is circled. However, the estimation of the external force after the occurrence of the erroneous detection is not disturbed, and the transient response of the position error signal is also good.

In order to actually confirm this effect, an erroneous position error signal is generated by, for example, externally applying pulse noise to the head signal. At this time, if the configuration of the present invention is adopted, a track seek error does not occur by a pulse application during a seek, and a seek time does not become long.

In particular, during the following operation, the voltage V with respect to the input disturbance to the head signal when the configuration of the present invention is adopted.
6 shows a time response of the CM drive current 80. The input disturbance is a pulse-like voltage signal having a duration corresponding to the sampling time, and is applied to the head signal line or the detection circuit directly or via an addition circuit, and is added to the head position signal.

First, the level of the input disturbance pulse 80 is set to V _h
Head position signal 33 when given as
FIG. 11 shows the time response of the CM drive current 48. Level V
_It is assumed that _h is sufficiently larger than the erroneous position detection criterion x. Therefore, the pulse _Vh is determined as an erroneous detection. In this case the observed head position signal 33, applied pulse V _h varies greatly added time points, but shows the values of the position signal 33 before again pulse application when the pulse is extinguished. The VCM drive current 48 at that time does not change at the time when the pulse is applied.

Next, assuming that the level of the input disturbance pulse is VL, the observation head position signal 33 at this time and the VCM
FIG. 12 shows the time response of the drive current 48. It is assumed that the level VL is smaller than the erroneous position detection determination reference x. Therefore, the pulse VL is not determined to be erroneously detected.

At this time, the observation head position signal 33 generates a signal that cancels the pulse at the time when the applied pulse VL is applied. This phenomenon occurs because the control system recognizes PESL as a true displacement.

In the second embodiment described above, an error in position detection is determined by estimating a position error signal using a past position error signal and a VCM control signal, and the estimated position is determined in accordance with the determination result. A method for improving the estimation accuracy of the disturbance observer using the error signal will be described. By combining the method with the method described in the first embodiment, the estimation accuracy of the external force is improved, which is effective for improving the transient response of the head position. That was shown.

The contents described in the above embodiments are as follows.
The present invention is not limited to the magnetic disk device, and may be applied to an XY stage used for a semiconductor exposure device or the like, or a high-precision positioning device such as a magneto-optical disk device.

[0077]

As described above, in a magnetic disk drive in which miniaturization and high recording density are progressing, the occurrence of friction in the bearing portion causes a decrease in positioning accuracy at the time of minute positioning. When an external force estimating means including a disturbance observer is used to compensate for a disturbance including friction, a gain fluctuation and a model error greatly change the transient response of the estimated external force value and deteriorate the transient response when the head is moved. Further, when the servo position information cannot be detected or an incorrect track number is detected, an erroneous estimated external force is generated for several samples thereafter.

According to the present invention, by applying a limiter for limiting the amount of change in the estimated value of the observer, it is possible to suppress a fluctuation in the estimated value that is too large and to improve the transient response of the head position.

Further, the position error signal estimated by using the past position error signal and the VCM control signal is compared with the detected position error signal to determine whether the position error signal is abnormal.
By using the estimated position error signal as needed, the accuracy of external force estimation can be improved, and the transient response during head movement can be improved.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a first embodiment of the present invention.

FIG. 2 is a functional block diagram of the first embodiment.

FIG. 3 is a diagram showing the effect of the first embodiment.

FIG. 4 is a diagram showing a disturbance pulse addition circuit.

FIG. 5 is a time chart of an input disturbance 60;

FIG. 6 is a time chart of an operation signal 47 when a disturbance is applied.

FIG. 7 is a configuration diagram of a second embodiment of the present invention.

FIG. 8 is a flowchart of a second embodiment of the present invention.

FIG. 9 is a diagram showing a position error signal estimation accuracy of the second embodiment.

FIG. 10 is a diagram showing the effect of the second embodiment.

FIG. 11 shows a signal 33 and a signal 4 when a position signal disturbance 80 is applied.
8 time chart (1).

FIG. 12 shows a signal 33 and a signal 4 when a position signal disturbance 80 is applied.
8 time chart (part 2).

FIG. 13 is a perspective view of a magnetic disk drive.

[Description of References] 1 ... magnetic head, 2 ... voice coil motor, 3 ... pivot bearing, 4 ... magnetic disk, 5 ... carry, 6 ... load arm, 7 ... head support spring, 8 ... spindle motor, 9 ... FPC, 10 ... bass, 11 ... amplifier, 12 ...
Demodulation circuit, 13 AD converter, 14 position signal generating means, 15, 23 switch, 16 speed signal generating means,
17 speed command generating means, 20 positioning compensator, 21
... External force compensation means, 22 ... External force compensation table, 24 ... DA
Converter, 25 drive circuit, 32 demodulation signal, 33 position signal, 34 position error signal, 37 acceleration compensation signal, 3
9 speed signal, 40 target speed value, 41 difference signal, 4
2 ... External force estimated value, 45 ... Operation signal, 46 ... VCM operation signal, 47 ... Operation amount, 48 ... Drive current.

Claims (4)

[Claims] At least one rotating disk having concentric data tracks, a head for reading servo information defining a center line of a data track, means for deriving a head position error signal from the servo information, A voice coil motor for positioning the head with respect to a track, a processor for generating a digital control signal from the head position error signal, means for converting the digital control signal into an analog control signal, and amplifying the analog signal; In a magnetic disk drive having an amplifier for generating an input current used in the voice coil motor, a disturbance compensating means for compensating disturbance applied to a control target by a digital control signal generated from the head position error signal; Variable limits for the amount of change at each signal output time Or a means for limiting the range of the disturbance compensation signal to a variable limit value, or limiting a differential value of second or higher order of the disturbance compensation signal to a variable limit value. Characteristic magnetic disk drive. 2. The method according to claim 1, wherein when adding a pulse signal as a disturbance to the voice coil motor driving circuit directly or via an adding circuit, when an amplitude value of the disturbance pulse signal is larger than a threshold value. The operation signal, which is the command value of the voice coil motor input current, changes by a value limited to the threshold for a certain period and then converges, and the time required for the operation signal to settle depends on the amplitude value of the disturbance pulse. 2. The magnetic disk drive according to claim 1, wherein the length of the magnetic disk drive becomes longer. 3. A method for estimating a head position from the head position error signal assuming a constant speed or a constant acceleration;
An estimated head position error calculated by using a method of estimating a head position based on a dynamic model of a control object from the head position error signal and the voice coil motor input current or a history of the operation signal according to claim 2. A signal is compared with the head position error signal to determine whether the position error signal is normal or abnormal. If it is determined that the position error signal is abnormal, a digital control signal is calculated using the estimated head position error signal. The magnetic disk drive according to claim 1. 4. When a pulse signal is added to a signal output from the head directly or via an adding circuit, when a pulse amplitude value is larger than a certain threshold value, the voice coil motor drive current fluctuates when a pulse is applied. Without the pulse, the position error signal indicates the value before the pulse is applied after the pulse is removed. If the pulse amplitude value is smaller than a certain threshold, the drive current fluctuates at the time of pulse application, and the position error signal fluctuates after the pulse is removed. The magnetic disk drive according to claim 3.