JPH1166774A - Positioning method of disk device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enable high precision positioning control with high performance of suppressing disturbance. SOLUTION: In the disk device for reading and writing circumferential tracks on a disk under rotation by moving a head in the radial direction, the head is made to follow a target track by using a control input to be nonlinearly increased or decreased on a position error ep obtained by subtracting the present head position x from a target head position τ* as a control input τ* for controlling a position of the head.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a positioning method for positioning a head mounted on a carriage of a disk drive so as to follow a target track, and more particularly to a disk eccentricity, a frictional disturbance between a head and a medium, and a bearing frictional disturbance. An object of the present invention is to provide a magnetic disk drive having a higher storage accuracy with a higher positioning accuracy in a positioning mechanism of a disk drive that can receive uncertain disturbances such as vibration disturbance and shock disturbance.

[0002]

[Prior art]

Conventional Example 1 In the conventional method of positioning a disk device, when rolling bearing friction is a problem, the friction force is estimated using a bearing friction generation model, an input for canceling is added, and a disturbance observer is added. Other disturbances were offset.

FIG. 28 is a block diagram showing a control method in, for example, the 73rd Ordinary Meeting of the Japan Society of Mechanical Engineers (IV) “Method of compensating bearing friction force of HDD head actuator using friction force estimation observer”. It is. In FIG. 28, 55 is a compensator for executing a general following control law such as a PID controller, 56 is a plant simulating a carriage and a friction generating mechanism, 57 is a friction force estimation observer, and particularly a bearing rolling model. And both general disturbance observers.

In the conventional positioning method using the friction force estimation observer 57 configured as described above, the bearing friction is accurately estimated using a bearing rolling model, an input for canceling is added, and at the same time, other disturbances are detected. Has estimated using a disturbance observer, added an input to cancel, and effectively canceled the disturbance to improve the positioning performance.

Conventional Example 2 FIG. 29 is a block diagram showing a positioning method according to the sliding mode control law disclosed in, for example, Japanese Patent Application Laid-Open No. 5-80852, in which the positioning method is restricted to a sliding surface which is a target trajectory on a phase surface. Positioning is performed using a non-linear control input.

In FIG. 29, when a target command value P ^* is input, a positional deviation P between the movement command value P0 and the target command value P ^* is obtained.
ε is calculated. Next, the command speed v0 and the target speed v
^* , The speed deviation vε is calculated. In this case, since the target value v ^* of the speed is 0 at the completion of the positioning, the speed deviation v
ε = −v0. Next, when the sliding line Pε> 0, Pε = vε ² / 2α When Pε <0, the calculation of Pε = −vε ² / 2α (α is the maximum acceleration given by the system) is performed. The calculated speed deviation vε
Into the phase plane 60 having the sliding line shown in FIG. 30, and these positional deviation Pε and velocity deviation vε
It is determined whether the point indicating the relationship is above or below the calculated sliding line. If the driving force F is on the upper side, the positive maximum acceleration + α is output, and if it is on the lower side, the negative maximum acceleration -α is output to the integrator 61 of the speed command value v0. The integrator 61 of the speed command value v0 outputs the speed command value v0
= V0 + FΔt (Δt is sampling time). This new speed command value v0 is input to the integrator 62 of the moving speed command value P0, and the next moving command value P0 = P0 + v
0 Δt is calculated.

By repeating the above operation, the deviation between the position deviation Pε and the velocity deviation vε is reduced so as to be restricted by the sliding line as shown in FIG.
It converges to the origin O on the sliding line where the position deviation Pε = 0 and the velocity deviation vε = 0, which are the final objectives.

In the conventional positioning method using such a sliding mode control, the state is restricted to a trajectory with respect to a target trajectory defined on a phase plane, and positioning is performed as a highly robust state called a sliding mode. . The sliding surface passes through the origin on the phase plane so as to converge to the origin itself, and is determined as a curve existing only in the second and fourth quadrants not including the axis.

[0009]

The conventional method of positioning a disk drive is performed as described above. The positioning method using the frictional force estimating observer of the first conventional example requires an accurate model of the friction generating mechanism. However, it is actually difficult to obtain an accurate model because the occurrence of friction changes according to product variations, product attitudes, and years of use. Therefore, there is a problem that the disturbance cannot be canceled and the position error cannot be sufficiently reduced. Further, since the friction is obtained by calculation from the estimation model, there is a problem that an attempt to estimate the friction by a more accurate model increases the number of steps of the program, that is, increases the calculation time, and deteriorates the performance of the entire control method. . In addition, since disturbances other than bearing friction are also canceled by the disturbance observer, there is a problem that the calculation amount is increased also in this case. Further, the positioning performance after the disturbance is estimated and the canceling observer input is performed depends on the follow-up control law. In order to improve the performance, so-called high gain control must be performed, and there is a problem of stability. That is, since the control is performed to a high frequency by the high gain control, there is a problem that, for example, the vibration of the carriage is excited and the entire control system oscillates.

Further, the positioning method based on the sliding mode control of the conventional example 2 has a high robustness because the state is constrained to the sliding surface, and is effective for the movement operation from the initial position to the origin in a situation where disturbance such as friction occurs. It is. However, when the sliding surface is a linear function vε = −b · Pε (1) and equation (1) is solved, A and b are constants, and Pε = A · exp (−bt) (2) That is, as shown in FIG. 31, the motion along the sliding line becomes an exponential function, and converges to 0 asymptotically. However, in a disk drive, it is only necessary to enter an off-track without asymptotic convergence, so that this method is not necessarily highly responsive. In the process of asymptotic convergence, the speed of the carriage gradually decreases, so that the moving time of the carriage is long. That is, in FIG. 31, the closer to the origin, the lower the speed on the vertical axis, and the slower the movement of the carriage. Furthermore, when reading and writing data, it is necessary to control to keep the track confined to a given track.However, the sliding mode control has no clear design guideline against the influence of disturbance, etc. There is a problem that the position error increases.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and it has been found that the positioning of a disk drive can be performed in which positioning can be controlled with a high level of disturbance suppression performance while greatly reducing the amount of calculation and improving responsiveness. It provides a method. Further, the present invention provides a positioning method that enables higher positioning control by combining with a disturbance observer when the calculation amount has a margin.

[0012]

According to a first aspect of the present invention, there is provided a method of positioning a disk drive, comprising: moving a head in a radial direction with respect to a circumferential track on a rotating disk-shaped storage medium; In a disk device that reads and writes,
As the control input for controlling the position of the head, a control input that non-linearly increases or decreases with respect to a position error ep obtained by subtracting the current head position from the target head position is used, and the head follows the target track. It is.

In the disk device positioning method according to a second method of the present invention, the function indicating the relationship between the position error ep and the control input in the above-described positioning method is the same as the position error e.
When the magnitude of p is out of the predetermined range, it becomes a substantially linear function,
The magnitude of the control input when the magnitude of the position error ep is within a predetermined range is a non-linear function that is smaller than the magnitude of the control input calculated using the linear function.

According to a third aspect of the present invention, there is provided a disk drive positioning method, wherein a function indicating a relationship between a position error ep and a control input in the above positioning method is kff (δ is smaller than a track width) when ep> δ. Value, kff is a positive constant value), 0 when δ ≧ ep ≧ −δ, −when ep <−δ
It is a non-linear function obtained by adding a step function of kff to a linear function.

According to a fourth aspect of the present invention, there is provided a disk device positioning method for reading and writing a circumferential track on a rotating disk-shaped storage medium by moving a head in a radial direction. As a control input for controlling the position of the head, a control input for nonlinearly increasing or decreasing a position error ep obtained by subtracting the current head position from the target head position, and a disturbance estimation estimated from the control input and the current head position Using the value obtained by adding the value
The head follows the target track.

[0016]

BEST MODE FOR CARRYING OUT THE INVENTION

Embodiment 1 FIG. 1 (a) and 1 (b) are a plan view and a front view showing a schematic configuration of a disk device according to Embodiment 1 of the present invention, and FIG. 2 is a configuration diagram showing a positioning mechanism unit and a calculation unit of the disk device. 3 is a first embodiment of the present invention.
FIG. 4 is a flowchart showing a positioning method according to the first embodiment of the present invention, FIG. 5 is a flowchart specifically showing calculation of a control input, and FIG. Function.

In FIG. 1, 1 is a disk as a recording medium, 2 is a spindle motor for rotating the disk, 3
Is a head for reading and writing information to and from the disk 1, 4 is a magnet, 5 is an iron core, 6 is a coil, 7 is a carriage, 8 is a bearing, 9 is a guide shaft, and 10 is a mounting bracket for fixing the guide shaft to the outside. It is. 33 is, for example, 8 μm on the disk 1.
This is a concentric track with a constant width of m.
A voice coil motor composed of a magnet 4, an iron core 5, and a coil 6 generates an electromagnetic force to obtain a thrust for driving the carriage 7. In FIG. 2, reference numeral 28 denotes a head position detection signal, which directly measures a position from a data track recorded on the disk 1. Reference numeral 29 denotes an A / D converter for discretizing the head position detection signal 28, and reference numeral 30 denotes a DSP for calculating a control input, which fetches a target position and calculates a control input. 31 is a D / A converter, 32 is a voice coil motor driver, and 34 is a drive current flowing through the coil 6.

A positioning method according to the first embodiment of the present invention will be described with reference to FIGS. In FIG. 3, when a head target position x ^* is input, a position error (deviation) ep between the head position detection values x and x ^* is calculated (S).
01). Differential values ep and ep of ep in the PID controller 11
Is calculated (S02), and a linear control input τ is calculated by multiplying the integral value ei by the proportional gain kp, the speed gain kv, and the integral gain ki (S03). τ = kp · ep + kv · ev + ki · ei In FIG. 3, S in the PID controller 11 is a Laplace operator, Laplace operator S is differentiation, 1 / S is integration, and Ei is obtained by integration of ev and ep by differentiation. On the other hand, in the nonlinear switching controller 12, ep
Is within the dead zone ± δ (S04), and if ep is within the dead zone, a control input (τ ^* ) is obtained with τ ^* = τ (S07). If it is outside the dead zone, the compensation amount kff is determined, and finally the control input (τ
^* ) Is obtained (S05). τ ^* = τ + kff · sgn (ep) The obtained input τ ^* is output from the D / A converter 31 (S
07). The control input τ ^* is converted into a coil input current 34 by a current-voltage conversion coefficient 14 after an operation delay 13.
The coil generates thrust according to the current thrust conversion coefficient 15,
It becomes the driving force of the carriage 16. The disturbance vibration 17 acts on the carriage 16 in a form subtracted from the thrust.

The block diagram of FIG. 3 is rewritten as shown in FIG. In FIG. 5, reference numeral 18 denotes a block for calculating a position error ep obtained by subtracting the current position from the target position, a differential value or speed error ed of the position error, and an integrated value ei of the position error.
Reference numeral 19 denotes a block for calculating a control input by using two functions 20, 21 of a step function 20 for compensating friction and a linear function 21 with respect to the position error ep. It is equivalent to Reference numeral 23 denotes a block for calculating a control input using a linear function for the velocity error ev, and reference numeral 24 denotes a block for calculating a control input using a linear function for the integral value ei of the position error.
25 is a block for calculating the sum of three control inputs, and 26 is a D / A output. FIG. 6 is a diagram showing the block 22 in detail. The function 22 is such that when ep> δ, kff (δ is smaller than the track width, kff is a positive constant value), δ ≧ ep
This is a non-linear function obtained by adding a linear function 21 to a step function 20 that is 0 when ≧ −δ and −kff when ep <−δ. The position error ep is used as the nonlinear function 22.
Is substantially a linear function when the magnitude of the control input is outside the predetermined range, and the magnitude of the control input when the magnitude of the position error ep is within the predetermined range is the magnitude of the control input calculated using the linear function. A non-linear function that becomes smaller than that, for example, a continuous curve as shown in FIG. 7 may be used. The nonlinear function in FIG. 7A is represented by the following equation. f = ep− | ep | · ep · exp (− | ep
|) This non-linear function may be a continuous but non-smooth function as shown in FIG. 7 (b). The nonlinear function in FIG. 7B is represented by the following equation. f = ep · | ep | (| ep | ≦ 1) f = ep (| ep |> 1) Both of the continuous curves shown in FIG. 7 are non-linear when ep is close to the origin, and become almost linear as ep increases. .

The head 3 is of a contact type, and reads and writes information by contacting the medium 1. The carriage 7 is supported by bearings 8. Therefore, since the friction due to the medium and the friction due to the bearing are generated, there is a possibility that the positioning performance is largely impaired as compared with the case where there is no friction. In block 19, the function 20 is a step-like function that gives a certain value to the sign of the position error, and the compensation value uses the nominal value of 0.0784N of the frictional force. If the amount of compensation is set to be larger than the value of friction measured in an experiment or the like, the problem that the control input is smaller than the frictional force and the state is stopped can be avoided. In the present embodiment, a dead band of 0.39 μm is provided near the origin to suppress chattering, and the compensation value is set to zero. The P feedback configured in this manner is a non-linear function equivalent to the function 22. When the target position is not reached by adding the estimated friction force to the error, the coil thrust is applied in such a direction as to cancel the friction force, and in the section that goes too far from the target value, the coil thrust is applied in the same direction as the friction force. Brake to reduce the amount of overshoot.

While the disturbance observer works to cancel the frictional force at any time, in the present embodiment, the friction observer cancels the frictional force according to the situation, and the friction observer increases the frictional force in order to apply a brake. Switching is performed according to the position error code.

When the calculation of the nonlinear function is divided into a step function part represented by the function 20 and a linear part represented by the function 21, the calculation is simple in the case of the step function, so that the number of calculation steps hardly increases. .

In the following, the effect of this embodiment by an experiment will be described. The function shown in FIG. 6 was used as the nonlinear function for the position error. FIG. 8 and FIG. 9 show the experimental results when the above disk drive has a medium eccentricity of 12 Hz and 10 μm. FIG. 8 shows a case where only the conventional PID control is performed.
9 shows the position error and the time change of the drive current when the control according to the first embodiment of the present invention is performed. In FIG. 8, while the position error of ± 0.836 μm is shown, FIG.
In this case, the position error is ± 0.58 μm, and the position error can be reduced to about 70% by the control according to the first embodiment of the present invention. In a disk drive, the medium is always eccentric to some extent, and the eccentricity tracking error can be reduced by improving the responsiveness. 10 and 11 show sinusoidal vibration disturbance, ± 0.25
G, Experimental results when 100 Hz is applied. FIG. 10 shows the change over time of the position error when only the conventional PID control is performed, and FIG. 11 shows the time change of the position error when the control according to the first embodiment of the present invention is performed. In contrast, in FIG. 11, the position error is ± 0.7 μm, and the position error is reduced by about 20% in the control according to the first embodiment of the present invention.
Can be 12 and 13 show random vibration disturbance, ± 0.75 Gr. m. s. It is an experimental result when adding. FIG. 12 shows a case where only the conventional PID control is performed.
3 shows a time change of the position error when the control according to the first embodiment of the present invention is performed.
While a position error of 6 μm occurs, FIG.
4.6 μm, and the position error can be reduced to about 60% in the control according to the first embodiment of the present invention.

In the sliding mode control, the sliding surface passes through the origin on the phase surface so as to be stable itself, is given as a curve existing only in the second and fourth quadrants not including the axis, and the control law is switched. In the present embodiment, since the switching is performed with respect to the sign of the position error, that is, the speed axis, the sliding mode control is not performed. The sliding surface in the sliding mode control is defined as a function of the position error and the speed error. In the present embodiment, since the calculation is performed using a nonlinear function with reference to only the position error, the position error, the speed error, and the position error integrated value are calculated. The functions for the control input calculation corresponding to each are independent. In the dead zone to prevent chattering, only PI
The evaluation may be performed as D control.

In this embodiment, for example, a non-linear function as shown in FIG. 6 is used. However, when the position error exceeds the dead zone, the gain is increased in a stepwise manner to enhance the position control performance. Since the increase in the gain is small, it is more stable than simple high gain control.
When the carriage is defined as an inertial mass m and the PD control law in which the integral compensation ki / S of the third term in the PID controller 11 is omitted is used, the equation of motion is as follows: When the position is x, v = dx / dt, m · dv / Dt = −kp · x−kv · v, the phase plane trajectory is as shown in FIG. Assuming that λ is a solution of the characteristic equation m · s ² + kp · s + kv = 0, the damping coefficient of the PD control law is often given by about 0.7, so the origin is called a stable spiral point, and α
> 0, ω> 0, and j are given as imaginary units as λ = −α ± j · ω. The state trajectory is inclined with respect to the axis by the function v = −α · x. If there is kinetic friction in the system subjected to the PD control, the equation of motion m · dv / dt = −kp · x−kv · v−d as the kinetic friction force d acting in the direction opposite to the speed.
Since sgn (v) is obtained, the phase plane trajectory is as shown in FIG. A point (± d / kp, 0) on the x-axis is called a singular point. On a line segment connecting these two points, the state stops due to friction, and a steady-state deviation occurs to cause deterioration of control performance. The position of the singular point is determined by the magnitudes of the dynamic frictional force d and kp, and it is necessary to increase kp in order to meet the required specifications. However, if the gain is set to a high gain, the stability is impaired. Alternatively, the steady-state error can be reduced to zero by performing integral compensation.
I can do it, but I cannot expect much responsiveness. FIG. 16 shows the phase plane trajectory of the present embodiment. The equation of motion is m · dv / dt = −kp · x−kv · v−d
Sgn (v) + kff · sgn (x) When the singularity is x> 0 and v> 0, ((−d−k
ff) / kp, 0), when x> 0 and v <0, ((d−
kff) / kp, 0), when x <0 and v <0, ((d
+ Kff) / kp, 0), when x <0, v> 0,
((−d + kff) / kp, 0). The position x corresponds to the position error ep, and the speed v corresponds to the speed error ev. A sharp change in the speed error with respect to the position error indicates that the effect of suppressing the position error is great. In other words, it is possible to quickly move to the inside of the off-track and improve responsiveness. Although there is a dead zone near the origin, the state stops, but if the dead zone is set smaller than the required specification, there is no problem in the operation of the product even if the state stops. Note that the state trajectory when there is no dead zone is as shown in FIG. 17, and the origin is an unstable point, so that a limit cycle is performed.

Embodiment 2 FIG. FIG. 18 is a block diagram of a control law showing a positioning method according to Embodiment 2 of the present invention,
FIG. 19 is a flowchart showing a positioning method according to the second embodiment of the present invention. FIG. 18 shows a configuration in which a disturbance observer is added to the first embodiment shown in FIG.

Referring to FIG. 18 and FIG. 27 is a disturbance observer. In FIG. 19, when the head target position x ^* is input, a deviation ep between the head position detection value x and x ^* is calculated (S10). The differential value ev of ep and the integral value ei of ep are calculated in the PID controller 11 (S1
1) Calculate a linear input τ by multiplying each by a proportional gain kp, a speed gain kv, and an integral gain ki (S1).
2). τ = kp · ep + kv · ev + ki · ei The nonlinear switching controller 12 determines whether the acquired ep is within the dead band ± δ (S13), and if ep is within the dead band, τ ^* = τ Obtain control input (τ ^* ) (S1
6). If it is outside the dead zone, the compensation amount kff is determined, and a control input (τ ^* ) is finally obtained as the sum of the respective outputs (S14). τ ^* = τ + kff · sgn (ep) The obtained τ ^* is first used as a control input (S15).
In the disturbance observer 27, the control input τ one step before
^{** A} disturbance estimation value -Fd is calculated from (-1) and the head position detection value (S17). An input Fd for canceling the disturbance is added to the control input τ ^* to obtain a new control input τ ^** (S18). τ ^** = τ ^* + Fd τ ^** is stored for the next operation by the disturbance observer (S19). Finally, the control input τ ^** is output from the D / A (S20). The control input τ ^* is converted into an input current 34 to the coil by a current-voltage conversion coefficient 14 via an operation delay 13. The coil generates a thrust according to the current thrust conversion coefficient 15 and becomes a driving force of the carriage 16. Disturbance vibration 1
7 acts on the carriage in such a way that it is subtracted from the thrust.

In the positioning method according to the first embodiment,
Positioning was performed by using the estimated values of the bearing friction and the medium friction. In the second embodiment, the bearing and the medium friction that are easy to estimate are canceled by a nonlinear function with respect to the position, and the disturbance received from the outside of the drive such as random vibration. This is a positioning method that is dealt with by a disturbance observer.

FIGS. 20, 21, and 22 show the eccentricity follow-up response, the sinusoidal vibration disturbance response,
This is a random vibration disturbance response. In the second embodiment, eccentricity tracking ± 0.438 μm, sinusoidal vibration disturbance response ±
0.51 μm and random vibration disturbance response ± 4.4 μm, which are compared with the case of only the conventional PID control shown in FIGS. 8 to 13 or the case of performing the control according to the first embodiment.
It can be confirmed that the performance is further improved.

FIGS. 23 to 26 show the case where only the conventional PID controller is used, the case where the control according to the first embodiment is performed, and the case where the control is performed using a conventional PID controller with a disturbance observer added thereto. 9 shows eccentricity follow-up simulation results when the control according to the second embodiment is performed. PID
In the case of the control alone, the tracking error is ± 2.8 μm, whereas in the first embodiment, it is ± 0.9 to 1.2 μm, which is about ３. On the other hand, in the combination of the PID control and the disturbance observer, it is ± 1.5 μm, which is about １ ／ as compared with the PID controller alone. However, in the second embodiment, ± 0.5 μm
Approximately 1/6 of the PID controller. Further, the current input is also smooth. It can be seen that in the second embodiment in which a disturbance observer is added to the first embodiment, the performance is improved by the complementary effect.

The complementary effect can be explained as follows. The disturbance observer has frequency characteristics, and a cutoff frequency of the disturbance that can be estimated and canceled is designed by adjusting the disturbance observer gain. On the other hand, in the control device used in the first embodiment, the controller is designed in the form of setting a dead zone and a compensation amount that explicitly reflect the specification for positioning. That is, in the second embodiment, the controller can be designed from two aspects of the frequency and the magnitude of the error with respect to the disturbance and the off-track amount. If the disturbance observer gain is too high, high-frequency vibration is excited and the system becomes unstable, so there is a limit. However, the method of using the non-linear function shown in the present invention for the position error has the effect of improving the positioning accuracy without high gain, so the disturbance is suppressed in the frequency domain by the disturbance observer, and the specification is not satisfied. The second embodiment, in which the compensation is made with a nonlinear function relating to the position error in this case, is an effective method.

FIG. 27 shows (a) an eccentricity follow-up response, (b) a sinusoidal vibration disturbance response, and (c) a random vibration disturbance response.
3 shows the result of comparison with the case where positioning is performed. In the case of Embodiment 1, the conventionally used PI
It can be seen that the position error can be greatly reduced as compared with the case of only the D control. Also, in the case of the second embodiment, it can be seen that the position error can be further reduced.

[0033]

As described above, according to the first method of the present invention, a disk drive for reading and writing a circumferential track on a rotating disk-shaped storage medium by moving a head in a radial direction. In the above, a control input that nonlinearly increases or decreases a position error ep obtained by subtracting the current head position from the target head position is used as a control input for controlling the position of the head, so that the head follows the target track. As a result, the effects of bearing friction, the friction between the medium and the head, and the effects of vibration and shock disturbance from outside the disk drive are reduced, and the head can be positioned with high precision following the eccentricity of the rotating medium. is there. as a result,
By increasing the reliability of the disk device and narrowing the track width, a large-capacity disk device can be obtained. Further, the positioning method of the present invention devises a calculation method of the controller,
There is no increase in material cost and the amount of calculation is small, so that it can be easily mounted, and there is an effect that a high-performance disk device can be obtained at low cost.

According to the second method of the present invention, the function indicating the relationship between the position error ep and the control input in the above-described positioning method is substantially equal to the function when the magnitude of the position error ep is out of the predetermined range. It becomes a linear function, and when the magnitude of the position error ep is within a predetermined range, the magnitude of the control input is
Since this is a non-linear function that is smaller than the magnitude of the control input calculated using the linear function, it is possible to design a control system based on the position error specification, and to control with higher positioning accuracy.

According to the third method of the present invention, when the function indicating the relationship between the position error ep and the control input in the above positioning method is ep> δ, kff (δ is smaller than the track width, kff is positive Is constant when δ ≧ ep ≧ −δ and −kff when ep <−δ,
Since this is a non-linear function added to a linear function, highly accurate positioning can be performed without increasing the number of calculation steps.

According to the fourth method of the present invention, in a disk device for reading and writing a head in a radial direction with respect to a circumferential track on a rotating disk-shaped storage medium, the position of the head is determined. As a control input to control,
A control input that increases or decreases nonlinearly with respect to a position error ep obtained by subtracting the current head position from the target head position;
Since the head is made to follow the target track by using a value obtained by adding the control input and the disturbance estimation value estimated from the current head position, there is an effect that the positioning can be performed with higher accuracy.

[Brief description of the drawings]

FIG. 1 is a plan view and a front view showing a schematic configuration of a disk device according to a first embodiment of the present invention.

FIG. 2 is a configuration diagram showing a positioning mechanism unit and a calculation unit of the disk device according to the first embodiment of the present invention.

FIG. 3 is a block diagram of a control law showing a positioning method according to the first embodiment of the present invention.

FIG. 4 is a flowchart showing a positioning method according to the first embodiment of the present invention.

FIG. 5 is a flowchart specifically showing calculation of a control input according to the first embodiment of the present invention;

FIG. 6 is a diagram showing a nonlinear function according to the first embodiment of the present invention.

FIG. 7 is a diagram showing another nonlinear function according to the first embodiment of the present invention.

FIG. 8 is a diagram showing an experimental result (eccentric follow-up response) when positioning is performed only by conventional PID control.

FIG. 9 is a diagram showing an experimental result (eccentric follow-up response) when positioning is performed according to the first embodiment of the present invention.

FIG. 10 is a diagram showing an experimental result (sinusoidal vibration disturbance response) when positioning is performed only by conventional PID control.

FIG. 11 is a diagram illustrating an experimental result (sinusoidal vibration disturbance response) when positioning is performed according to the first embodiment of the present invention.

FIG. 12 is a diagram showing an experimental result (random vibration disturbance response) when positioning is performed only by conventional PID control.

FIG. 13 is a diagram showing an experimental result (random vibration disturbance response) when positioning is performed according to the first embodiment of the present invention.

FIG. 14 is a diagram showing a phase plane state trajectory of a system to which the conventional PD control is applied.

FIG. 15 is a diagram illustrating a phase plane state trajectory of a frictional system subjected to a conventional PD control.

FIG. 16 is a diagram showing a state trajectory according to the first embodiment of the present invention.

FIG. 17 is a diagram showing a state trajectory in the case where there is no dead zone according to Embodiment 1 of the present invention.

FIG. 18 is a block diagram of a control law showing a positioning method according to a second embodiment of the present invention.

FIG. 19 is a flowchart showing a positioning method according to Embodiment 2 of the present invention.

FIG. 20 is a diagram showing an experimental result (eccentricity following response) when positioning is performed according to the second embodiment of the present invention.

FIG. 21 is a diagram showing an experimental result (sinusoidal vibration disturbance response) when positioning is performed according to the second embodiment of the present invention.

FIG. 22 is a diagram showing an experimental result (random vibration disturbance response) when positioning is performed according to the second embodiment of the present invention.

FIG. 23 is a diagram showing an eccentricity follow-up time waveform by the conventional PID control.

FIG. 24 is a diagram showing an eccentricity following time waveform according to the first embodiment of the present invention.

FIG. 25 is a diagram showing an eccentricity follow-up time waveform obtained by a positioning method in which a disturbance observer is added to conventional PID control.

FIG. 26 is a diagram showing an eccentric tracking time waveform according to the second embodiment of the present invention.

FIG. 27 is a graph summarizing experimental results according to the first and second embodiments of the present invention.

FIG. 28 is a block diagram showing a positioning method using a conventional frictional force estimation observer.

FIG. 29 is a block diagram showing a positioning method based on a conventional sliding mode control law.

FIG. 30 is a diagram showing a sliding line in conventional sliding mode control.

FIG. 31 is a diagram showing a state trajectory by conventional sliding mode control.

[Explanation of symbols]

1 disk, 2 spindle motors, 3 heads, 4
Magnet, 5 iron core, 6 coil, 7 carriage, 8 bearing, 9 guide shaft, 10 mounting bracket, 11 PID controller, 12 nonlinear switching controller, 13 calculation delay, 14
Current-voltage conversion coefficient, 15 Current thrust conversion coefficient, 16
Carriage, 17 disturbance vibration, 18 position error ep, differential value or velocity error ed of position error, block for calculating integral value ei of position error, 19 block for calculating control input from position error ep, 20 step function, 21 linear Function, 22 nonlinear function, 23 block for calculating control input from speed error ev, 24 integral value ei of position error
Block for calculating more control inputs, 25 block for calculating the sum of three control inputs, 26 D / A output, 27
Disturbance observer, 28 Head position detection value, 29 A /
D converter, 30 DSP, 31 D / A converter, 32
Voice coil motor driver, 33 tracks, 34
Drive current, 55 compensator, 56 plant, 57 frictional force estimation observer, 60 sliding surface, 61, 6
2 Integrator.

 ────────────────────────────────────────────────── ─── Continued on front page (72) Inventor Takeshi Ouchi 2-3-2 Marunouchi, Chiyoda-ku, Tokyo Mitsubishi Electric Corporation

Claims (4)

[Claims] 1. A disk drive for reading and writing a circumferential track on a rotating disk-shaped storage medium by moving the head in a radial direction, wherein a target head position is used as a control input for controlling the position of the head. And a control input for nonlinearly increasing or decreasing a position error ep obtained by subtracting a current head position from the head, and causing the head to follow a target track. 2. A function indicating the relationship between the position error ep and the control input is a substantially linear function when the magnitude of the position error ep is outside a predetermined range, and the magnitude of the position error ep is within a predetermined range. 2. The method according to claim 1, wherein the magnitude of the control input is a non-linear function that is smaller than the magnitude of the control input calculated using the linear function. 3. A function indicating the relationship between the position error ep and the control input is kff (δ is smaller than the track width, kff is a positive constant value) when ep> δ, and when δ ≧ ep ≧ −δ. 2. The method according to claim 1, wherein the step function is a non-linear function obtained by adding a step function which becomes -kff when ep <-. Delta. To a linear function. 4. A disk drive for reading and writing a circumferential track on a rotating disk-shaped storage medium by moving the head in a radial direction, wherein a target head position is used as a control input for controlling the position of the head. The head is moved to the target track using a value obtained by adding a control input that non-linearly increases / decreases a position error ep obtained by subtracting the current head position from the current head position, and a disturbance estimation value estimated from the current head position. A method for positioning a disk drive, characterized in that the disk drive is made to follow.