JPH05334819A - Recording head positioning device - Google Patents

Abstract

PURPOSE:To always perform a proper track following operation. CONSTITUTION:Prior to track following control, a command signal of a prescribed frequency is given to a step motor 22, and the step motor 22 is driven. A displacement amt. of the step motor 22 is detected by an encoder 23, and hence a gain of the above command signal is detected. Fuzzy inference is performed by this gain and a rule determined previously in a fuzzy inference device 30, so that a control parameter of a PID compensator 21 is derived. PID operation of the PID compensator 21 is performed by the derived optimum control parameter, and feedback control for following a track by a magnetic head 13 is performed.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a recording head positioning device.

[0002]

2. Description of the Related Art In a disk-shaped recording medium, it is necessary to position a recording head on a track of the recording medium. The positioning of the recording head is performed by using, for example, a step motor, but the recording tracks are densified with the increase in capacity of the recording medium, and more precise positioning is required.

In order to meet this demand, a device for positioning the recording head by feedback control has appeared (for example, Yoshiura et al. Tracking Servo for High Density FDD Magnetic Recording Research Society Material MR90 of the Institute of Electronics, Information and Communication Engineers).
-18). FIG. 6 shows an example of the configuration of an apparatus for positioning the recording head by such feedback control.

The circuit shown in FIG. 6 is composed of a first servo circuit 10 and a second servo circuit 20.

An eccentricity amount of the recording medium is detected, and a signal * x representing the eccentricity amount is given to the subtractor 11 of the first servo circuit 10. A signal representing the magnetic head position from the magnetic head 13 is also input to the subtractor 11. The subtractor 11 obtains a signal e representing the amount of positional deviation of the magnetic head 13 due to the eccentricity of the recording medium, and supplies it to the deviation compensator 12.

Magnetic head input by deviation compensator 12
The magnetic head 13 based on the signal indicating the position deviation amount of 13
Command signal r to follow the track to be positioned
Is output.

The first servo circuit 10 positions the magnetic head 13 on the target track so as to prevent displacement of the magnetic head 13 due to eccentricity of the recording medium, and feedback control is performed so that the magnetic head 13 does not deviate from the target track.

The command signal r output from the deviation compensator 12 is given to the subtractor 24 of the second servo circuit 20. The magnetic head positioning device includes an encoder 23 for detecting the rotational position of a step motor 22 that drives the magnetic head 13. The position signal of the step motor 22 represented by the output signal of the encoder 23 is also given to the subtractor 24. A signal representing the amount of displacement of the magnetic head 13 due to the vibration of the recording medium is obtained by the subtractor 24 and the PID is obtained.
It is given to the compensator 21.

A command signal is outputted by the PID compensator 21 so as to correct the positional deviation of the magnetic head 13 due to the vibration of the recording medium, and the command signal is given to the step motor 22.

The second feedback circuit 20 positions the magnetic head 13 on the target track so as to prevent displacement of the magnetic head 13 due to vibration of the recording medium.

A detailed block diagram of the second servo circuit 20 included in the magnetic head positioning apparatus shown in FIG. 6 is shown in FIG.

In FIG. 7, K _I is an integral gain, K _d is a differential gain, K _c is a loop gain, K _r0 is a command value correction coefficient, K _r1 is a command value displacement gain, and K _f is a stiffness (step). The thrust constant of the motor 22), M is the inertia (mass of the mover of the magnetic head), and D is the viscosity coefficient.

The PID compensator 21 applies a gain of K _c (1 + K _I / s + K _d s) to the signal input to the PID compensator 21, and corrects the gain to give it to the step motor 22 as a command signal.

In the magnetic head positioning apparatus shown in FIGS. 6 and 7, the gain value of each circuit in FIG. 7 is set to a constant value when the magnetic head positioning apparatus is manufactured. However, there is a problem in that the gain value changes due to the accuracy of assembly and changes over time, which affects the positioning accuracy of the magnetic head, and the stability of positioning cannot be ensured.

[0015]

DISCLOSURE OF THE INVENTION An object of the present invention is to improve the positioning accuracy of a recording head and ensure the stability of the positioning of the recording head.

The recording head positioning device of the present invention comprises:
A drive motor for moving the recording head by feedback control so that the recording head follows the recording track of the recording medium, a detection means for applying a predetermined drive signal to the drive motor, and detecting the displacement amount of the drive motor, the detection Fuzzy inference means for performing fuzzy inference according to the displacement amount of the drive motor detected by the means and predetermined rules set in advance to derive control parameters used for feedback control, and feedback control using the fuzzy inference control parameters A control means for controlling the recording track following control of the recording medium of the recording head is provided.

According to the present invention, the drive signal is applied to the drive motor, and the displacement amount of the drive motor caused by the drive signal is detected. Fuzzy inference is performed according to the detected displacement and preset rules to derive the optimum control parameters.

Since fuzzy inference is performed to obtain optimum control parameters necessary for feedback control in recording head positioning, optimum control parameters according to the situation such as the recording track to follow can be obtained and the optimum recording head is always available. Position control is performed. Therefore, the track deviation of the recording head can be minimized, and the recording head can be positioned relatively stably.

Further, since the recording head can always be positioned stably, the positioning accuracy is improved even if the assembly accuracy of the apparatus is lowered, and the manufacturing cost of the apparatus can be reduced.

The predetermined drive signal preferably uses a signal having a frequency lower than the natural frequency of the drive motor and a signal having a frequency higher than the natural frequency of the drive motor.

[0021]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows an embodiment of the present invention and is a block diagram of a magnetic head positioning device. In FIG. 1, the same parts as those shown in FIG. 6 are designated by the same reference numerals and the description thereof will be omitted.

Prior to the description of FIG. 1, a method of calculating the command-gain coefficient K _r and the stiffness K _f of the step motor 22 will be described.

The transfer function of the step motor can be generally approximated by the second-order delay system of the following equation.

[0024]

[Number 1] _{G (s) = K f ·} K r / (Ms 2 + Ds + K f) ... formula 1

The Bode plot of Equation 1 is presented in FIG. In FIG. 2, ω _n is the natural frequency of the step motor.

Command-displacement gain G [dB] and command-
The relationship with the displacement gain coefficient K _r is expressed by the following equation.

[0027]

[Equation 2]

There is a relationship between the natural frequency ω _n and the inertia M and stiffness K _f of the step motor expressed by the following equation.

[0029]

[Equation 3]

Further, as shown in FIG. 2, the Bode diagram shows that in the frequency region above the natural frequency ω _n , the natural frequency ω
Inclination passing through the intersection A between _n and the command-displacement gain G-40 [d
B / dec].

First frequency ω _{equal to} or lower than natural frequency ω _n
1 (for example, if the natural frequency ω _n is 150 [Hz], 50
If the command-displacement gain in [Hz]) is known, the command-displacement gain coefficient K _r can be obtained from Equation 2. The second frequency above the natural frequency omega _n omega ₂ (e.g. 30
If the command-displacement gain at 0 [Hz]) is known, the natural frequency is determined by the intersection of a straight line passing through this gain value and having a slope of −40 [dB / dec] and a straight line passing through the gain value of frequency ω ₂ and having a slope of ω _n is obtained. By knowing the natural frequency ω _n and the inertia M, the stiffness K _f can be obtained from the equation 3.

Next, a method for calculating the integral gain K _I , the differential gain K _d , the loop gain K _c , and the command-displacement gain correction coefficient K _r0 from these command-displacement gain value K _r and stiffness K _f will be described. State.

Referring to FIG. 7, it is preferable that the following equation is established between the command-displacement gain K _r1 and the command-displacement gain correction coefficient K _r0 .

[0034]

[Equation 4] K _r1 · K _r0 = 1 Equation 4

From equation 4, the command-displacement gain coefficient K _r0 is obtained from the following equation.

[0036]

[Equation 5] K _r0 = 1 / K _r1 Equation 5

K is the loop gain of the second servo circuit 20.
_{If L} , this loop gain K _L is constant. Assuming that Equation 4 holds, the loop gain K _L is obtained by the following equation.

[0038]

[Equation 6] K _L = K _c · K _f Equation 6

From Equation 6, the loop gain K _c is obtained by the following equation.

[0040]

## EQU7 ## K _c = K _L / K _f Equation 7

Further, the cutoff frequency of integration is f _ci [H
z], the relationship between the cutoff frequency f _ci and the integral gain K _I is given by the following equation.

[0042]

## _EQU8 ## K _I = f _ci × 2π Equation 8

Further, the cutoff frequency of differentiation is f _cd [H
z], the relationship between the cutoff frequency f _cd and the differential gain K _d is given by the following equation.

[0044]

## EQU9 ## K _d = 1 / (f _cd × 2π) Equation 9

Here, if the stiffness K _f is large, both the cutoff frequency f _{ci for} integration and the cutoff frequency f _{cd for} differentiation are large, and if the stiffness K _f is small, the cutoff for integration is small. Frequency f _ci
It is generally known that both the cutoff frequency f _cd and the differential cutoff frequency may be reduced.

Therefore, the relation between the above-mentioned stickiness K _f and the cutoff frequencies f _ci and f _cd , the relation between the integral cutoff frequency f _ci and the integral gain K _I expressed by the equations 8 and 9, and the cutoff of the derivative. From the relationship between the off frequency f _cd and the differential gain K _d , the relationship between the stiffness K _f and the integral gain K _I and the differential gain K _d is derived. That is, if the stiffness K _f is large, the integral gain K _I is increased, but the differential gain K _d is decreased. If the stiffness K _f has a small value, the integral gain K _I is reduced, but the differential gain K _d is increased.

In this way, if the command-displacement gain K _r and the stickiness K _f are known from the equations 5 and 7 and the relationship between the stiffness K _f and the integral gain K _I and the differential gain K _d , the second value is obtained. Other control parameters of the servo circuit 20 (command-displacement gain correction coefficient K _r0 , loop gain K _L , integral gain K _I and differential gain K _d ) can be obtained.

The gain G ₁ of the frequency ω ₁ of less than or equal to the natural frequency ω _n From the above it and the natural frequency ω _n or more of the frequency ω ₂
If the gain G ₂ is known, the second servo circuit 20 shown in FIG.
All control parameters of can be known.

The gain G ₁ of the frequency ω _{1 equal to} or lower than the natural frequency ω _n , the gain G ₂ of the frequency _{equal to} or higher than the natural frequency ω _n , and the control parameters K _c , K _I , and K _{D of} the second servo circuit 20.
The relationship between K _r0 and K _r0 is shown in FIGS. 3 (A)-(C). FIG. 3 (A) shows a gain G ₁ of frequency ω _{1 which is equal to} or lower than the natural frequency ω _n.
And the command-displacement gain correction coefficient K _r0 are shown in FIG.
(B) the gain G ₂ of gain G ₁ and the natural frequency omega _n frequencies above the frequencies below the natural frequency omega _n, loop
The relationship between the gain K _c and the differential gain K _D is shown in FIG.
(C) shows the relationship between the gain G ₁ of the gain G ₁ and the natural frequency omega _n below the frequency of the higher natural frequency omega _n and the integral gain K _I.

As shown in FIG. 3 (A), the gain G ₁ and the command-displacement gain correction coefficient K _{r0 at} a frequency _{equal to} or lower than the natural frequency ω _n are obtained.
Is inversely proportional to, and if the gain G _{1 at} a frequency _{equal to} or lower than the natural frequency ω _n increases, the command-displacement gain correction coefficient K
_The command-displacement gain correction coefficient K _r0 increases when _r0 decreases and the gain G _{1 at} a frequency _{equal to} or lower than the natural frequency ω _n decreases.

Further, as shown in FIG. 3B, the natural frequency ω _n
The gain G ₂ of the above frequency is proportional to the loop gain K _c and the differential gain K _d with a negative proportional coefficient, and when the gain G ₂ of the frequency above the natural frequency ω _n increases, the loop gain K _c and the differential gain K _d decreases and the natural frequency ω _n
If the gain of the following frequencies decreases, the loop gain K _c
And the differential gain K _d increase. Furthermore, the natural frequency ω _n
When the gain G ₁ of the following frequencies becomes smaller, each value becomes smaller.

Further, as shown in FIG. 3 (C), the natural frequency ω
The gain G _{1 at} a frequency of _n or more is directly proportional to the integral gain K _I, and when the gain G _{2 at} a frequency of the natural frequency ω _n or more increases, the integral gain K _I also increases, and the gain of the frequency of the natural frequency ω _n or more becomes When it decreases, the integral gain K _I also decreases. Further, when the gain at frequencies below the natural frequency ω _n becomes smaller, the integral gain K _I increases.

The rules shown in FIGS. 3A to 3C are as follows. If G ₁ = NL then K _r0 = PL If G ₁ = PL, G ₂ = NL then K _c = PL, K
_r0 = NL, K _I = NL, K _d = PL If G ₁ = ZR, G ₂ = ZR then K _c = ZR, K
_r0 = ZR, K _I = ZR, K _d = ZR If G ₁ = NL, G ₂ = PL then K _c = NL, K
_r0 = PL, K _I = PL, K _d = NL If G ₁ = NM, G ₂ = NM then K _c = PM, K
_r0 = PM, K _I = NM, K _d = NL If G ₁ = NM, G ₂ = PM then K _c = NM, K
_r0 = PM, K _I = PM, K _d = PL

Here, NL is Negative Large (negative and large), NM is Negative Middle (negative negative), NS is Negative Small (negative and small), ZR is Zero (almost zero), and PS is Positive Small (positive). Is small), PM is Positive Middle, and PL is Positive L
Represents arge (positive and large).

FIG. 4 shows an example of the membership function relating to the variables of the antecedent part representing the states of G ₁ and G ₂ described above.

FIG. 5 shows an example of the membership function relating to the variables of the consequent part representing the states of K _c , K _r0 , K _I and K _d described above.

FIG. 1 shows a magnetic head positioning device.
A fuzzy reasoning device 30 is included as shown in FIG. The fuzzy inference device 30 applies the gains G ₁ and G ₂ of the frequencies above and below the natural frequency ω _n to the membership function shown in FIG. 4 in accordance with the rules described above, and performs the MIN or MAX calculation of their goodness of fit. Fuzzy inference is performed by further applying the result to the membership function shown in FIG. 5, and the inference result is defuzzified (defuzzified) to obtain each value K _c , K _r0 , K _I and K _d.
Ask for. Positioning control of the magnetic head is performed using the obtained values.

Next, the procedure for obtaining each control parameter of the second servo circuit 20 and performing the positioning control of the magnetic head will be generally described with reference to FIG.

The magnetic head positioning device shown in FIG. 1 includes a sine wave generator 31, a fuzzy inference device 30, a peak hold circuit 32, and a memory 33, as compared with the magnetic head positioning device shown in FIG.

Firstly a sine wave generator 31 from the natural frequency omega _n following the first frequency (e.g. 50 Hz) natural frequency omega _n
Two kinds of sinusoidal command signals having the above second frequency (for example, 300 Hz) are output and given to the step motor 22 and the fuzzy inference device 30, respectively. This drives the step motor 22.

The displacement amount of the step motor 22 according to the command signal is detected by the encoder 23. The detected displacement amount is given to the peak hold circuit 32 and 50 [Hz]
The amplitude value when the command signal is input to the step motor 22 and the amplitude value when the command signal of 300 [Hz] is input to the step motor 22 are detected. Each detected amplitude value is given to the fuzzy reasoning device 30,
The amplitudes of the command signals of the respective frequencies input from the sine wave generator 31 are compared. By this comparison, the natural frequency ω _n
Following the first gain G ₁ and the gain G ₂ of the natural frequency omega _n or more second frequency of the frequency it is calculated.

Natural frequency ω in fuzzy reasoning apparatus 30
The gain G ₁ of the first frequency _{equal to} or less than _n and the natural frequency ω _n
Fuzzy inference according to the above rule is performed from the gain G ₂ of the second frequency. By this fuzzy inference, each control parameter of the second servo circuit 20 can be obtained. The calculated control parameters are given and stored in the memory 33.

Each control parameter is detected at a plurality of positions on the track of the recording medium, and the optimum control parameter for the track is stored in the memory 33. The control parameters are detected for all the tracks of the recording medium, and the optimum control parameters for each track of the recording medium are stored in the memory 33.

When a recording or reproducing command is given to the magnetic head positioning device, the track number to be recorded or reproduced is read, and the control parameter corresponding to the read track number is read from the memory 33 and given to the PID compensator 21. Be done. P determined by the given control parameters
The ID operation is performed by the PID compensator 21, and the second servo circuit 20 prevents the track off due to the vibration of the recording medium.

The first servo circuit 10 prevents the off-track due to the eccentricity of the recording medium, as described in the background art section.

In the above-described embodiment, the sine wave generator 31 is used in the magnetic head positioning device to give a sine wave to the step motor 22 to determine the control parameters for the PID compensator 21. Instead of this, a step-shaped or ramp-shaped command signal may be given to the step motor 22.

Further, the sine wave generator 31, the fuzzy inference device
30 and peak hold circuit 32 can be implemented using software as well as hardware.

Further, the step motor can be applied regardless of whether it is a rotary type or a linear type.

[Brief description of drawings]

FIG. 1 is a block diagram of an electric configuration of a magnetic head positioning device according to an embodiment of the present invention.

FIG. 2 is a Bode diagram of a step motor.

[Fig. 3] Gain and P for frequency of step motor
It shows the relationship with the control parameters of the ID compensator.
(A) shows the relationship between the step motor frequency and the command-displacement gain correction coefficient, and (B) shows the relationship between the step motor frequency and the loop gain and differential gain.
(C) shows the relationship between the step motor frequency and the integral gain.

FIG. 4 shows an example of a membership function of the antecedent part of a fuzzy rule.

FIG. 5 shows an example of a membership function of the consequent part of a fuzzy rule.

FIG. 6 is a block diagram showing an example of a magnetic head positioning device.

FIG. 7 is a block diagram of a second servo circuit included in the magnetic head positioning device.

[Explanation of symbols]

 20 Second servo circuit 21 PID compensator 22 Step motor 30 Fuzzy inference device 31 Sine wave generator 33 Memory

Claims (1)

[Claims] 1. A drive motor for moving the recording head by feedback control so that the recording head follows a recording track of a recording medium, a predetermined driving signal is given to the driving motor, and a displacement amount of the driving motor is detected. The detection means, the displacement amount of the drive motor detected by the detection means, and fuzzy inference means for performing fuzzy inference according to a predetermined rule set in advance to derive control parameters used for feedback control, and the fuzzy inference control parameter are used. A recording head positioning device comprising a control means for performing feedback control to perform recording track following control of the recording medium of the recording head.