JPH06203500A - Head-position controller of disk driving gear using adaptively gain-adjustable repeated learning control - Google Patents

Abstract

PURPOSE: To provide a head position controller of a disk driving device which simultaneously reduces periodical and non-periodical errors. CONSTITUTION: The head position controller has a repetitive learning controller 31 which repetitively executes learning on a decided track and generates position error compensation signals which can compensate periodical and non-periodical errors, a memory 32 storing the error compensation signals from the repetitive learning controller 31 and a 3σ calculator 35 providing the operation condition of the repetitive learning controller. Then, an interpolator 33 generating position error compensation data on the necessary track by using position error compensation data of the previously learnt adjacent tracks when a head follows the track which is not previously learnt is contained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a head position control device and method for a disk drive device, and more particularly, to adaptively adjust position error factors generated periodically and aperiodically in the disk drive device. The present invention relates to a head position control device and method for a disk drive device, which can improve the track following precision of a head by applying a repetitive learning control method capable of gain adjustment to reduce the same.

[0002]

2. Description of the Related Art Currently, information is recorded on or read from a disk as a large-capacity information recording medium, for example, a hard disk drive, an optical (magnetic) disk drive, an optical disk. Disk drive devices such as drive devices have been widely spread and used. Such a disk drive is usually
The head position control function is provided for performing track aligning on a track and a sector provided on the information storage disk to accurately align the head assembly by track following. In this case, the track search is defined as a head moving operation from the track where the current head is located to a desired target track, and the search operation is a so-called velocity profile method for stable optimization of the search time. (Velocity prof
ile method). Track following is defined as an operation for minimizing a deviation between a center of an arbitrary track on which data is recorded / read and a center of a head, that is, a position error, and is usually a PID control (proportion).
Ional-Integral-Differenti
al control) will be used.

According to such a head position control method, it is possible to relatively effectively control the position of the head in an ideal system in which the center of the track on the disk and the rotation axis of the spindle of the disk drive are exactly aligned. But there is a spindle runout,
In a real system where there is eccentricity between the track and the spindle, the position control precision of the head is reduced. Further, in a disk drive device in which the loading / unloading operation is repeated, such a situation becomes more serious, and the maximum overshoot at the target track becomes large at the final moment of the track search. The operation of the conventional device is slowed down.

On the other hand, the component of the position error generated from the disk drive device is a periodical error caused by factors such as eccentricity between the track and the spindle, accumulation of the spindle, or mechanical failure periodically generated. It is a well-known fact that it contains a large number of elements, and such a periodic element is generally the main component of the position error which is similarly repeated each time the disk is rotated, though it depends on the type of disk drive. Therefore, the track following precision of the head can be considerably improved by reducing the periodic error.

For this reason, various efforts have been made to reduce cyclic errors in disk drives.

For example, a method of repeatedly learning an error periodically generated from the disk drive device and canceling a similar cyclic error that occurs later is Tomizuka (To).
"Digital Control of Repeatable Error in Disk Drive System (Digital Control)
ol of repetitive Error in
"Disk Drive System"", IEEE Control Systems Magazine (IEEE Control Systems Magazine)
azine), pp. 16-20 (1990).

According to this method, in a disk drive device having a periodic position error, the periodic position error is repeatedly learned for each rotation cycle of the disk, so that the disk is rotated after several rotations. The position error compensation value synchronized with the position in the circumferential direction is used to reduce periodic errors. Therefore, when such a repetitive learning controller is applied to the track following operation of the disk drive device, the track following ability can be improved, so that improvement in track precision can be expected.

However, in relation to the actual recording / reading operation of the disk drive device, the iterative learning controller for the iterative learning time has a slight difference in the disk rotation time due to the characteristics of the disk drive device. Considering the corresponding learning time, there is a disadvantage that the waiting time until the recording / reading ready state is extended. Actually, the learning control must be performed for each head for each track, because the periodic pattern in which the position error is different for each head is seen. In addition, the same head must be used for learning over all triggers, because the reason is that in the case of an embedded servo system that employs a rotary actuator, the servo pattern is generally an arc in the disk radial direction. Since they are arranged in a shape, the position error compensation values of any one track cannot represent the position error compensation values of all the tracks. In such a situation, in the above-mentioned method, it is possible to reduce a periodic error by repeating learning in any one track, but when the head moves, the disk is rotated again many times with respect to the corresponding track, Since it is necessary to perform repetitive error compensation for periodic errors, there is a drawback that the operating speed of the disk drive device is slowed down.

Further, according to the above-mentioned iterative learning control method, although a certain amount of canceling effect against a periodic error can be obtained, the fact that the aperiodic error can be increased is described by T. Inoue. "Practical repetitive control system design (practical repetitive)
control system design ”, Proceeding of 29th Eye EEE Conference on Decision Control (proc. of 29th IEEE conference)
enceon Decision control),
Pp. 1673-1678, 1990. Therefore, not only may there be the possibility of weakening the position control precision of the head due to amplification of aperiodic errors, but the maximum value of the difference is also increased.

Therefore, the present invention is made based on the above-mentioned conventional technique, and an object thereof is to repeatedly learn a position error which is periodically generated at the time of starting the disk drive or at a constant cycle time. The present invention provides a head position control device and method for a disk drive using adaptive gain adjustment iterative learning control in which position error compensation data is calculated and stored to improve the track following precision of the head.

Another object of the present invention is to provide a head position control device and method for a disk drive capable of reducing the aperiodic error, which weakens the position control precision of the head.

It is still another object of the present invention that, when the head moves to a non-learning track on a learned track, the position error compensation value calculated and stored by prior learning is used to determine the position of the corresponding track. A head position control device and method for a disk drive using adaptive gain adjustment iterative learning control that enables more rapid position control for each track by obtaining an error compensation value.

[0013]

A head position control device for a disk drive device according to the present invention comprises a data recording medium having a large number of data tracks, a drive means for rotating the data storage medium, and recording on the data recording medium. Or head for reading data from the recording medium, head driving means for moving the head, and head position for controlling the head driving means to hold the head on the center line of the selected data track. In the disk drive system including a control means, the head position control means controls a deviation between a center line of a selected data track of the data storage medium and a radial position of the head during rotation of the recording medium. Position error detection means for generating a certain position error signal, and for each of the data tracks,
The position error signals are input, the position error frequency components are used to extract adaptive position-adjusted periodic position errors, and the position errors corresponding to the selected tracks are set to the positions corresponding to the tracks. Position error compensation signal generating means for generating an error compensation signal, means for inputting the position error signal and the position error compensation signal to the selected track, and generating a compensated position error signal, In response to the compensated position error signal,
And a servo control means for generating a drive control signal for controlling the head drive means.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will now be described in detail with reference to the accompanying drawings.

FIG. 1 shows a block diagram of a disk drive apparatus of an embodiment using the iterative learning control according to the present invention.

As shown, the disk drive is
The disk assembly 10, the position error detector 20, the position error compensator 30, and the servo controller 50 are provided.

In general, the disk assembly includes magnetic recording media 11 and 12 in the form of stacked disks, such recording media 11 and 12 being rotated on a shaft 13 by a motor 14. Data is recorded on the center lines of concentric data tracks existing on both sides of each recording medium 11 and 12. The head actuator 11 moves a number of recording / reading heads 15 and 16 (corresponding heads on the lower disk surface are not shown) mounted on each arm in response to a drive control signal.

The position error detector 20 is controlled from a recording medium 11 or 12 rotated by a spindle motor 14 according to a recording or reading mode provided under the control of a main control unit (not shown) like the host computer. An operation of detecting a position error signal of a head with respect to a track and a sector is performed based on the detected servo pattern. As shown in FIG. 2A, the recording / reading operation on the recording mediums 11 and 12 is performed in units of a large number of tracks formed on the recording media 11 and 12 and a large number of sectors formed on the recording tracks. The sector is composed of a servo information field 41, a data field 42, an error correction field 43 and a sector interval 44 as shown in FIG. In addition, the servo information field 41 includes a servo synchronization signal 45, track identification information 46, and two burst signals A and B for obtaining fine position error information, as shown in FIG. 2C. There is. As the track identification information 46, a gray code is used in consideration of system stability.
The detected servo information 41 is used for calculating the head position error.

That is, the position error is calculated by the following equation (1).

Position error Pe = {(target track number-current head position track number) + G _f (AB)} (1) where A is the magnitude of the burst signal A and B is the burst signal. The magnitude of B (where | G _f (A−B) | <(50% of track width) and G _f is a gain function. The position error Pe obtained by the equation (1) is the target track. The position error signal, such as the deviation between the head and the centerline of the selected track, is provided to a position error compensator, which is designated by reference numeral 30. That is, as will be described later, the switch element 34 controlled by the position error 3σ calculator 35 is turned on and added to the iterative learning controller 31, and the iterative learning controller 31 sets the initial state of the system. Alternatively, when various initialization conditions are provided as described later due to environmental factors during operation, that is, when the switch element 34 is turned on by the position error 3σ calculator 35, head selection information from the main control unit, A compensation value for the periodic position error of the system is calculated based on the track number and sector number information and stored in the memory 32. In this case, the compensation value for the position error compensation value of the repetitive learning controller 31 is calculated.

In order to improve the precision, the automatic gain adjuster 36 determines the gain of the iterative learning controller 31 based on the frequency θ at which the maximum error exists and the current gain.

The position error compensation value obtained in this way is recorded in the memory 32 and then, during the actual track following operation,
It is provided to the adder 40 through the interpolator 33. Adder 4
At 0, the position error signal supplied from the position error detector 20 and the error compensation signal from the interpolator 33 are added to generate a newly compensated position error signal. In accordance with this newly compensated position error signal, the servo controller 50 switching element 51 has a position error β (track width 1
/ 2 to 1/4) or more, it is determined to be a track search operation, and switching connection is made to the "B" side, and the compensated position error signal is supplied to the target speed profile creation unit 53. In this case, the target speed profile creation unit 53 calculates the speed of the preset head according to the compensated error signal, and such a target speed control signal of the head is added by the adder 53 to the head speed determination unit. Combined with the actual head speed obtained from 55.

As a result, the adder 53 produces a speed control signal based on the difference between these two signals. The speed control signal is adjusted in gain by the gain controller 56, and the DC bias is compensated by the bias compensator 58 according to the target track through the adder 24. The DC bias-compensated speed control signal is supplied to the head actuator 11 through the saturation controller 59 and the digital / analog converter 61 to perform the track search operation by the speed profile method.

On the other hand, if the compensated position error signal is not more than β, the contact of the switching element 17 is converted to A, and the compensated position error signal is supplied to the PID control signal generator 52. The PID-controlled position control signal is provided to the head actuator 11 through the adder 57, the saturation controller 59 and the digital / analog converter 28 to perform the track following operation. In this case, the adder 54 is used. Through the DC bias compensator 58 based on the target track number, the position control signal is preset at the initialization stage of the system, for example, a flexible PCB.
After being compensated for the bias force due to the air flow or the bias force due to the air volume generated during disk rotation, the saturation controller 59 adjusts the saturation state, and the digital / analog converter 59 performs analog conversion, and then the head actuator 11
The head actuator 11 performs the desired track following operation.

The process of performing the gain adjustment iterative learning method adopted by the present invention with the above operation will be described later.

First, the position error information Pe determined by the above equation (1) is used as the switch element 34 and the position 3σ calculator 3
In addition to the above, the switching element 34 calculates the position error 3σ having a preset critical condition in order to prevent the extension of the waiting time until the information can be stored and read by the repeated learning time according to the present invention. Is turned on by the real-time control only when the system initialization stage or re-initialization condition is generated under the control of the controller 35, and makes the iterative learning controller 31 operate, while it is turned off otherwise. The position error compensation signal stored in the memory is output. That is, in order to control the switch element 34, the position error 3 [sigma] calculator 35 repeatedly determines whether or not to perform learning control when the learning controller 31 is moved, or controls using only the learned result. In the case of performing the calculation, the position error 3σ is calculated according to the on / off determination condition standard of the switch element 34 that determines whether relearning is possible due to a change in the operating environment due to a thermal factor.

Pe3σ = 31 / N Σ e (i) ² (2) = or its equalization formula, where N is the track and e (i) is the i-th track position error.

That is, from the position error 3σ calculator 35, when the position error 3σ is α (track width is 10%) or less, the switch element 34 is turned off, and when it is α or more, it is turned on.

Referring to FIG. 3, a mathematical model of the iterative learning controller 35 is shown. Switch element 1
When 1 is on, the iterative learning controller 35 repeatedly learns a periodic position error factor for several tracks for each head based on the head selection information, track number and sector number information added from the main controller. A position error compensation value is calculated based on the learning result and stored in the memory 32 based on the head selection information, track number and sector number information to be hit. When the iterative learning is completed, the switch element 34 is turned off, that is, the position error compensation value is continuously applied to the corresponding head and track sectors using the data stored in the memory 32 in the inactive state of the iterative learning controller 31. Therefore, the real time recording / reading of data is improved. Further, the position error compensation value stored in the memory 32 is learned over a large number of preset tracks. For example, when the target track is not the learned track during the information recording / reading operation, the vicinity of the target track is learned. An interpolator 33 is provided so that the position error correction data from the adjacent track subjected to the learning control can be interpolated and used. This interpolator 3
3 is a position error compensation value that is already learned based on the head selection information and track number and sector number information that are input externally, and is interpolated as follows to newly calculate a position error compensation value. To do.

That is, PECD (i, j) is the position error compensation data in the i-th sector and the j-th track, and j ₁ ≤j≤j ₂ . (J ₁ , j ₂ ) are the numbers of the adjacent tracks on which the repeated control is performed.

Then PECD (i, j) can be calculated by the following equation.

PECD (i, j) = (j ₂ −j ₁ ) (j−j ₂ ) + PECD (i, j ₁ ) ... (3) Furthermore, referring to FIG. A mathematical model of a learning controller is disclosed, the transfer function of which is:

Gr (z) = Kr · Z ^−N q (z) / [1−Z ^−N q (z) {Gs (z)} − 1] (4) Here, Gs is Gr Shows the Pereloop transfer function of the entire system in the absence of the condition, and q (z) is generally expressed as the characteristic of the FIR low pass filter (LPF) as follows.

Q (z) = 1 / L + 2 (L + Z ⁻¹ + Z) (5) Here, L = 2, 4 and 6 are selected.

At this time, Kr is the iterative learning controller 12
The gain is not stable unless 0 <Kr <2. If this Kr increases, the periodic element of the error becomes smaller and the number of non-periodic elements may also decrease.
The error cannot be reduced without proper adjustment. For this reason, an arbitrary standardized frequency is defined as follows.

The minimum value of θ = (w−w0 K) / w0 (0 ≦ θ ≦ 2π) where w0 = 2π / NT, K is a natural number, and N is a sector −
, T = sampling time For any normalized frequency θ, Kr0 (θ) =
1-cos (θ) / α (θ) (α (θ) is FIR LP
If (the size of F) exists and the current Kr is smaller than or equal to Kr0, then Kr can be made smaller, and if Kr is larger than Kr0, then Kr can be made larger to reduce the error corresponding to θ. Therefore, assuming that the error exists in the frequency domain at the frequency where the maximum error is θm, Kr0 (θm) = 1-cos (θm) / α (θm)
And the current Kr (θm) is calculated as Kr0 (θm)
Kr can be adjusted to reduce the maximum error value.

The gain adjustment algorithm can be summarized as follows.

First, in order to predict an actual disk drive system, when considering a normal linear single-input single-output discrete repetitive control system, R (Z) is the Z-transform of the reference input,
C (Z) is output Z conversion, E (Z) is difference Z conversion, Gr
(Z) is the transfer function of the repetitive controller, G0 (Z) is the transfer function of the controlled system, and Gi (Z- ¹ ) is Gs (Z) = Gs when modeling of the perfect controlled system is possible.
Select the inverse transfer function of (Z) / (1 + G0 (Z)), and
r, Z- ^N , T, and NT are gains of the iterative learning controller,
The time delay element, sampling time, and repetition period are shown, and the repetition period is equivalent to the period of the fundamental frequency of the periodic error.

A signal which is periodic but whose frequency is 2IIk / (NT) [rad / sec] with respect to a constant k is regarded as a periodic error, and all other signals are aperiodic. As the low pass filter shown in FIG. 3, the following FIR low pass filter is provided.

Q (Z) =-1 / L + (L + Z + IZ) (6) Here, L is usually selected as 2, 4, or 6.
q (Z) satisfies the following equation.

| Q (Z) | ≦ 1, and <(q (Z)) = 0 (7) Accordingly, the closed loop transfer function Gcl (Z) of the iterative control system can be expressed as follows.

Gcl (Z) = 1 + Gr (Z)) G0 (Z) / 1 + (1
+ Gr (Z)) G0 ( Z) = [{1-Z -N q (Z) + KrZ -N q (Z) Gi
(Z)} Gs (Z)] Gs (Z)] (8) Further, if the error transfer function Ge (Z) = 1-Gcl (Z) is defined, Ge (Z) is expressed as follows. it can.

Ge (Z) = Gre (Z) Geo (Z) (9) Here, Geo (Z) = (1-Gs (Z)) is error transmission when the repetitive controller 12 is not applied. In the function, Gre (Z) = 1 / (1 + Gr (ZGs (Z)) is expressed as a ratio of Ge (Z) and Geo (Z).
Let e (Z) be a relative difference transfer function, and let Eo (Z) be an error in the state where the iterative learning controller is not applied to Eo (Z) = Geo (Z) R (Z). ) E (Z) = Ge (Z) R (Z) and Gre (Z) = G
re (Z) / Geo (Z), so E (Z) and Eo
The function of (Z) is expressed as follows.

E (Z) = Gre (Z) Eo (Z) (11) If Z is exp (jwT) and θ is wNT, Gr
e (Z) is a function of Kr and θ and is expressed as follows.

E (Kr, θ) = Gre (Kr, θ) Eo (θ) (12) where θ is 0 ≦ θ ≦ 2π E (θ) maximum value | E | m max | E (θ) |
The frequency θm where E | m exists is arg (max) | E
(Θ) | Then, it can be seen that | E | m can be reduced by adjusting Kr as follows.

In the iterative learning control system of equation 8, 0 <K
r <2, | Gi (Z) Gs (Z) | = 1, <Gi (Z)
On the assumption that Gs (Z) = 0 and Gs (Z) stabilizes in an approaching manner, ▽ Kr (| Gre (Kr, θm) | <
If 0, increase Kr or ▽ Kr (| Gre (K
When r, θm) |> 0, | E | m decreases as Kr decreases.

That is, from Equation 12, when θ = θm, E
The magnitude of (Kr, θ) is expressed by the following equation.

| E (Kr, θm) || Eo (θm) | = | E | ... (13) Here, the differential value of | E (Kr, θm) | with respect to Kr is obtained by the following equation 14. To be

Σ (| E (Kr, θm) |) = σ (| Gre (Kr, θm) | / σ (Kr) σ (Kr) | Eo (θm) (14) From Expression 14, σ ( Kr)> 0, Kr (| Gre (Kr, θ
m) | <0, or σ (Kr <0),
If Kr (| Gre (Kr, θm) |> 0, then σ
It can be seen that (| E (Kr, θm) | <0.

Therefore, θm and Kr (| Gre
It can be seen that if only the sign of (Kr, θm) |) is given, whether or not | E | m can be increased or decreased by increasing or decreasing Kr.

Θm is a short time fast Fourier transform (short time F
ast Fourier Transform) or filter banks. As a result, Kr (| Gre (Kr, θ
The sign of m) |) is also | Gre (Kr, θm) |, which can be easily obtained as follows based on the quantitative analysis.

In order to easily observe the behavior of the aperiodic error, | Gi (Z) Gs (Z) | = 1, <(Gi) (Z)
If Gs (Z) = 0, Gre (Kr, θ) is obtained by the following equation.

Gre (Kr, θ) = 1−exp (−jθ) q (θ) /
1-exp (-jθ) q (θ) + Kr exp (-j
θ) q (θ) ... (15) Further, | Gre (Kr, θ) | is obtained as follows.

| Gre (Kr, θ) | = [(β / q (θ) ² Kr ² −rKr + β)] ^1/2 (16) where β = 1 + q (θ) ² −2q (θ ) Cos (θ)
And r = 2q (θ) ² −2q (θ) cos (θ).

In this case, | Gre (θ) | is a periodic function whose fundamental frequency w0 is w0 = 2π / (NT), and therefore Bk = {wq | w0k} ≤w0 (k
+1) is mapped to the frequency domain normalized by 0 ≦ θ ≦ 2π. Therefore, the frequencies where θ is 0 or 2π correspond to the periodic frequencies, and the frequencies where θ is other than 0 and 2π correspond to the non-periodic frequencies. Also, solve the following equation.

▽ Kr | Gre (Kr, θ) | = (γ-
2q) (θ) ² Kr) β ^1/2 / [2 (q (θ) ² Kr ²
−γKr + β) ^3/2 ] = 0 ... (17) Here, Kr is 0 <Kr <2.

Kr0 = arg (max | Gre (Kr,
θ) |) is obtained as follows. Kr0 = 1-cos θ / q (θ) (18) From the fixed θ, if Kr ≦ Kro, then Kr (| Gre
(Kr, θ) |) ≧ 0, and if Kr> Kr0, then ▽
It can be seen that Kr (| Gre (Kr, θ) |) <.

If θm is known, Kr0 can be found by calculating equation 27.

Therefore, the currently applied Kr and Kr0
By comparing Kr (| Gre (Kr, θ
The sign of m) |) becomes easy to be judged.

The above process can be summarized as an algorithm as follows.

Algorithm 1: Tolerance b for the iterative learning control system expressed by Equation 3
appropriately select sound ε> 0 and gain adjustment width ΔKr> 0. i represents the number of repetitions, and | E | m, Kr, θ
m, and the i-th value of Kro.

For initialization i = 1, | E | m (0) =
Set to 0. Kr (1) is set to 1 because of the rapid convergence speed.

That is, (| E | m (i)-| E | m
In the case of (i-1) / (Em (i))> ε, | E | m (i) and θm (i) are obtained by spectrum analysis of E (i), and θm (i) is given by Equation (13). Kro after substituting
(I) is calculated, and Kr (i + 1) = Kr (i) + ΔKr, for (Kr
(I) −Kro (i))> 0 Kr (i) + ΔKr, for (Kr (i) −Kro
(I)) Kr (i + 1) is updated so that ≦ 0, and the number of repetitions is updated to i = i + 1. It can be seen that the application of the gain of the iterative learning controller 12 obtained by the above algorithm reduces the periodic and aperiodic errors.

The head position controller for the disk drive which operates as described above will be described below with reference to the flow charts shown in FIGS.

FIG. 4 shows the system operation when the iterative learning control is activated, and FIG. 5 shows the system operation after the iterative learning control shown in FIG. 4 is performed.

In step S61 of FIG. 4, initialization is performed after power is applied to the system, or new learning control is performed due to environmental factors when the cartridge is mounted or while the system operation is in progress. If initialization is required, reinitialization will occur.

If such a condition is satisfied, the iterative learning controller 3 such as the position error compensation data is selected in step S62.
After initializing the data required for the operation of No. 1, the track follow-up operation is performed in steps S63 and S64. In step S66, the learning controller 31 is repeatedly operated for learning control with respect to the periodic and aperiodic position errors determined during the track following operation, and the position error compensating device is operated as described above. To calculate. After this, the iterative learning controller 3
In step S67 during the operation of 1, each disk rotation,
The position error 3σ is calculated based on the position error generated each time the head passes each sector, and if it is equal to or smaller than the preset σ, it is determined that the desired track following precision is ensured and the repeated learning control is performed. Then, in step S68, the position error compensation data calculated by the learning control operation is stored in the memory 13.

On the other hand, in the step S67, the position error 3σ
If is greater than or equal to α, it is determined that the learning is necessary again, and after returning to step S66, the position error compensation data by the subsequent step, that is, the iterative learning is calculated. Here, the process of steps S64 to S68 is repeated for each head, so that each head can deal with the position error with a different periodic pattern.

Further, step S6 is performed for each selected track.
By repeating the steps 5 to S68, the position error compensation value of only one track is used to complement the position error pattern that may occur when the device is used as the position error compensator of all tracks. This way,
High-accuracy attenuation is possible for position error in each track for each head.

Referring to the flow chart shown in FIG. 5, in step S71, the target head number and target track number information for the recording / reading operation are input and determined, and in step S72, the target head number and the target track number or their vicinity are input. The position error compensation data learned by the process of the flowchart appearing in FIG. 4 is read from the track. Then, in step S73, if the target track number and the track number subjected to the learning control do not match, the interpolation position error compensation data is set as new position error compensation data based on the position error compensation data learned near the target track. To use.

In this case, the interpolation method may be a linear interpolation method (that is, a linear interpolation method) or a high-order interpolation method if necessary.

In step S74, the new position error data obtained by adding the interpolated position error compensation data to the actual position error data for each sector by 1: 1 synchronization is added for each sector, that is, every sampling. In step S75, the new position error data is used as a position error during track following and track searching, and track following and searching operations are performed.
Next, in step S76, it is determined that the track precision is sufficient, as in the step S67 of FIG. 4, for each rotation of the disk during the track following operation or for every fixed time (for example, 30 minutes), and it is determined to be sufficient. If it can be determined, step S7
7 and apply the corresponding position error compensation value to reduce the cyclic position error, but if it is judged to be insufficient, proceed to step S78 to re-initialize the learning control of step S61 of FIG. The stability of the system is secured by performing the process of FIG.

[0073]

As described above, according to the present invention, stability and track are improved by applying the iterative learning control method to the position error generated periodically and aperiodically in the head position control system of the disk drive. The density is improved to improve the recording density of the disc.

Further, by dividing the real-time execution process of the learning control and the process of using the position error compensation value obtained according to the execution, and performing the learning control according to the initialization stage or situation of the system, the learning control is required. Depending on the time, it is possible to eliminate the influence on the waiting time due to the actual recording and reading operations of the system, and it is possible to accurately recognize the target track position during the track search by using the position error compensation data during the track search. As a result, the fixing time can be reduced, and the track search operation can be switched to the track following operation more quickly.

Further, while the learning control method is being executed, learning is performed on a plurality of tracks selected for each head, and after the repetition learning control is completed, when the position error compensation value is used, the target head and target track or By interpolating using the track-learned position error compensation value in the vicinity thereof to newly form a position error compensation value, the position error compensation value is stabilized, and a stable track following operation can be secured.

The head position control apparatus and method according to the present invention described above provides a head position control apparatus and method for a hard disk drive or a disk drive that uses a replaceable hard disk repetitive learning control.

[Brief description of drawings]

FIG. 1 is a block diagram of a head position control device of a disk drive device using gain adjustment iterative learning control according to the present invention.

FIG. 2 is a diagram schematically showing servo control recorded on an information storage disk.

FIG. 3 is a block diagram showing a mathematical model of a gain adjustment iterative learning control unit that performs the iterative learning control shown in FIG.

FIG. 4 is a flowchart showing an operation of updating error compensation data according to the present invention.

5 is a flowchart illustrating a head position operation using a position error compensation value obtained by the iterative learning control method in FIG.

[Explanation of symbols]

 10 Disk Assembly 20 Position Error Detector 30 Position Error Compensator 31 Iterative Learning Controller 32 Memory 33 Interpolator 35 3σ Calculator 36 Automatic Gain Adjuster 50 Servo Controller

Claims (4)

[Claims] 1. A data recording medium having a large number of data tracks, a driving means for rotating the data storage medium, a head for recording data on the data recording medium or reading data from the recording medium, and moving the head. In the disk drive system, the head position control means comprises: a head drive means for controlling the head drive means, and a head position control means for controlling the head drive means to hold the head on the center line of the selected data track. Position error detection means for generating a position error signal which is a deviation between a center line of a selected data track of the data storage medium and a position of the head in the radial direction while the recording medium is rotating; , Each position error signal is input, and the position error frequency component is used to adaptively use the position error signal. Position error compensation signal generating means for extracting an adjusted periodic position error and generating a position error compensation signal for each track corresponding to each selected track, and for the selected track Means for inputting the position error signal and the position error compensation signal to generate a compensated position error signal, and drive control for controlling the head drive means according to the compensated position error signal A head position control device for a disk drive device, which comprises a servo control means for generating a signal. 2. The position error compensation signal generating means extracts a periodic position error for each track according to a position error signal for each track, and position error compensation corresponding to the extracted periodic error. 2. An iterative learning control means for generating a value, and an adaptive gain adjusting means for adaptively adjusting a gain of the iterative learning control means according to a frequency component of a position error signal for each corresponding track. Position control device for disk drive device of. 3. The position error compensation signal generating means selectively supplies a position error signal to each track by the iterative learning control means, and the position error signal is generated according to the position error signal. 3. A head position controller for a disk drive according to claim 2, further comprising a switch driver for driving the switch so that the position error signal is repeatedly provided to the learning controller only when the set reference value is exceeded. . 4. The recording medium has N tracks,
The iterative learning control means generates the position error compensation signals for a preset number of tracks, and the position error compensation signal generation means assigns each position error compensation signal of the track to the corresponding track. Position error compensation signal storage means for storing and outputting the position error compensation signal of the corresponding track stored according to the selection of the track, and when a track other than the preset track is selected, the selected track 4. The head position of the disk drive device according to claim 3, further comprising: a reinforcing unit that interpolates a position error compensation signal of a preset track that is closest to the selected track to generate a position error compensation signal of the selected track. Control device.