KR20010022679A - Disc drive with robust track following servo controller - Google Patents

Abstract

The servo system in the disk drive 110 provides robust servo control during track following operation. The disc drive 110 includes a rotatable disc 112 having a disc surface 180, 182 with information disposed in tracks on the disc surface 180, 182. The disc drive 110 also includes a converter 144 for accessing information on the disc surfaces 180, 182. An actuator arm 126 is operatively connected to the converter 144. An actuator 173 is operatively connected to the actuator arm 126 to move the converter 144 relative to the disk surfaces 180 and 182 to follow one of the tracks during track following. The model-based control 190 is operatively connected to the actuator 173 to control the movement of the actuator 173. The model-based servo control 190 provides robust track following control over the desired frequency range.

Description

DISC DRIVE WITH ROBUST TRACK FOLLOWING SERVO CONTROLLER}

A typical disk drive includes one or more magnetic disks installed for rotation on a hub or spindle. Typical disk drives also include one or more converters supported by hydrodynamic air bearings that slide over each magnetic disk. The converter and hydrodynamic air bearings are collectively called the data head. The drive control unit is generally used to control the disc drive system in accordance with commands received from the host system. The drive controller controls the disc drive unit to retrieve information from the magnetic disks or to store the information in the magnetic disks.

Electromagnetic actuators operate in a feedback, closed-loop servo system. The actuator generally includes an actuator arm that supports the flexure of the flexible assembly that supports the data head. The actuator serves to move the data head concentrically with respect to the disk surface for track seek operations, and to follow the converter on tracks on the disk surface for track following operations.

The information is generally stored on the magnetic disk by providing a recording signal to the data head to encode flux reversal representing the data to be recorded on the surface of the magnetic disk. Upon retrieving data from the disc, the drive control unit controls the electromagnetic actuator so that the data head slides over the magnetic disc to sense flux inversions on the magnetic disc and generate a read signal in accordance with the flux inversions. The read signal is then decoded by the drive control to recover the data represented by the flux inversions stored on the magnetic disk, which is in turn represented by the read signal provided by the data head.

Positioning the data head accurately on the track of the disc is very important for reading data from or writing data to the disc.

In conventional systems, servo operations are performed by dedicated servo heads. In the dedicated servo type of the system, the servo information is all recorded on one dedicated surface of the disk in the disk drive portion. All the heads in the disc drive are mechanically connected to the servo head used to access the servo information. Thus, all the heads in the dedicated servo disk drive are positioned according to the servo information read from the servo surface. This type of system allows the disk drive to easily perform parallel read and write operations. In other words, due to the proper circuitry in the drive control section, write and read operations can be performed in parallel using multiple data heads mounted on the actuator, the data heads being simultaneously positioned according to the servo information read from the dedicated servo surface. .

However, track densities on magnetic disks have increased over the years. Increased track densities on magnetic disks require more sophisticated and higher resolution positioning. The mechanical offset between the heads in a dedicated servo system can exceed one track width. Thus, there has been a change in the art to embedded servo information in certain applications.

In an embedded servo system, servo information is embedded on each track on each surface of each disk. Thus, each data head returns a position signal independently of the other data heads. Therefore, while a particular data head has access to information on the disk surface, a servo actuator is used to locate each individual data head. The positioning is done using embedded servo data for the track on which the data head then slides.

While this increases accuracy and resolution in positioning (since the data head is positioned independently of other data heads), disadvantages arise due to increased track density and mechanical structures. One of the drawbacks is that in a typical embedded servo system, the ability to perform parallel read and write operations using multiple heads is lost. This is because the servo system is positioned according to the information received by one individual data head, and mechanical errors are inadequate for accurately positioning other data heads in a high track density system. Of course, current actuators cannot locate data heads independently. Thus, the built-in servo system has not been able to perform parallel reading and writing, such as writing or reading all cylinders in the disk drive at the same time.

Due to the above differences between dedicated and embedded servo systems, there is a trade off between servo sample rate and efficient user data storage.

In addition, many problems exist in both systems that affect the positioning accuracy and precision of the servo system. The most important problems are the following.

1. Servo sample rate. The sample rate is limited by the spindle speed and the number of servo sectors per track of the embedded servo system.

2. Structural Modes of Arm and Head Suspension

3. External shocks and vibrations that can be linear, rotational or both.

4. Write position error caused by tracking errors that occur when servo tracks are written. This causes a repeatable runout. Runout refers to a total positioning error during the track following operation. Runout generally refers to static deviation during tracking over long periods of time. Since the write position error is synchronized with the spindle speed, it is referred to as repeatable run out.

5. Bearing nonlinearity

Such nonlinearities generate rotational drag and hysteresis, particularly when the actuator rotates at low speed.

6. Nonlinearity due to the bias force of the flux circuit on the actuator

In other words, the actuator is connected to the disc drive control unit via the flux circuit. When the actuator positions the change in other radial positions of the disk, the flux circuit bias force on the actuator can change.

7. Disc flutter causing non-repeatable runout

The amount of disk swing depends on the speed of the spindle and the stiffness of the disk substrate.

8. Different gain variation in nonlinear cross-track of magnetic transducer

9. Position error sample noise due to medium magnetovariability, electrical noise, etc.

Conventional servo control units include a PID (proportional-intergral-derivative) control consisting of two components, an observer and a regulator. The observer receives input position information and estimates position and velocity each time it traverses the servo sector. The adjuster then provides feedback in accordance with the observed signals. In the seek mode, the adjuster zeros the error between the reference velocity trajectory and the observed velocity. In the track following mode, the adjuster zeros the error between the desired track position and the observed track position. The adjuster controls according to the PID control technique.

However, the PID control unit may not be desirable for all disk drive applications. For example, in an actuator arm or in a suspension or flexible assembly, it may be desirable to provide a microactuator between the flexible assembly and the converter or slider assembly. Where a microactuator is provided, the servo actuator system receives a plurality of inputs at the microactuators and receives a plurality of inputs at each of the actuators and at the single input single output (SISO) system where the input is an error signal and the output is a voice coil current signal. It has evolved into a multiple input multiple output (MIMO) system that provides an output signal to a voice coil motor. Although the system can be controlled by decentralizing the PID control, this presents a problem. For example, if multiple heads are positioned at the same time, the positioning of one data head may be affected by the simultaneous positioning of other adjacent or adjacent data heads.

Moreover, high bandwidth positioning may stimulate structural modes of the drive and may cause vibration, ringing, or other interference that may tend to interfere with positioning of adjacent data heads.

In addition, many problems arise when trying to implement a discrete time system in a fixed-point digital signal processor in a disk drive. For example, the computational power provided to digital signal processing units that can be used in the disc drive unit is generally very limited. This can cause a lot of important problems. The size and number of registers of the digital signal processor may be very limited. In the matrix calculation, the number of bits required to store the results of the intermediate calculation may exceed the capacity of the registers in the digital signal processing section. Thus, overflow can cause serious problems. In addition, the computational speed, structure, and memory capacity of the digital signal processor may render some matrix computations impossible due to the number of computations to be performed. All digital signal processing units are fixed-point processing units. Thus, a linear discrete time system in a digital signal processor may not be possible. Moreover, the quantization errors in conventional DSPs become serious, making the control accuracy of the DSP of the servo system in the disk drive unrealistic.

The present invention approaches some of these and other problems and provides advantages over the prior art.

The present invention relates to a servo system in a disk drive. In particular, the present invention relates to a disk drive having a model-based robust servo control system.

1 shows a disk drive according to an embodiment of the present invention.

2 shows an actuator assembly according to one embodiment of the invention.

3 illustrates a head gimbal assembly in accordance with aspects of the present invention.

4 is a block diagram of a part of a disk drive unit according to an embodiment of the present invention.

5 is a block diagram of a portion of a servo control circuit according to an embodiment of the present invention.

6 is a flowchart illustrating the structure of a servo control algorithm according to the present invention.

7 is a block diagram of an actuator model according to the present invention.

FIG. 8 is a graph showing a nominal model with measured data and average measured data taken while injecting disturbances into the system.

9A-9D illustrate examples of addition and multiplication uncertainty and weighting functions in accordance with the present invention.

10A to 10C show the estimated addition and multiplication uncertainty and phase error on the actual disk drive.

11 is a block diagram illustrating a track following control synthesis interconnect in accordance with the present invention.

FIG. 12 is a block diagram illustrating linear fractional transformation (LFT) of the track tracking control synthesis interconnection shown in FIG. 11.

13 is a graph illustrating model validation in accordance with the present invention.

14 is a flowchart illustrating a model validation technique according to the present invention.

15 is a flowchart illustrating an implementation of a servo controller and a digital signal processor according to the present invention.

The present invention relates to a system for approaching some of the above mentioned problems.

The servo system in the disc drive provides robust servo control during track following operation. The disc drive has a rotatable disc with a disc surface having information arranged in tracks on the disc surface. The disc drive also includes a converter for accessing information on the disc surface. The actuator arm is operatively coupled with the converter. The actuator is engaged with the actuator arm to move the converter relative to the disk surface so that it follows one track during track following. The model-based servo control is operatively connected to the actuator to control the movement of the actuator. The model-based servo control provides robust track following control over the desired frequency range.

1 is a plan view of a general disc drive unit 110. The disc drive unit 110 includes a disc pack 112 installed in a spindle motor (not shown) by a disc clamp 114. In a preferred embodiment, the disc pack 112 includes a plurality of discs installed for concentric-rotating about the central axis 115. Each disk surface on which data is stored includes an associated head gimbal assembly (HGA) 116 installed in an actuator assembly 118 in the disk drive 110. The actuator assembly shown in FIG. 1 is of a type known as a rotary moving coil actuator and includes a voice coil motor (VCM) shown at 120. Under the control of the electrical circuitry embedded in the disc drive [10], the voice coil motor 120 rotates the actuator assembly [118] and the HGAs [116] attached thereto about the pivot axis [121] so that the associated disc surfaces can be rotated. Place the HGAs 116 over the desired data tracks.

In particular, the actuator assembly 118 is pivoted about the axis 121 to rotate the head gimbal assembly 116 along the arc [119] so that each head gimbal assembly [116] is the surface of the disk in the disk pack [112]. Make sure it is located on one of the tracks of your choice. The HGA [116] can move from the tracks at the innermost radius to the tracks at the outermost radius of the disc. Each head gimbal assembly 116 includes a gimbal that elastically supports the slider with respect to a load beam, such that the slider can follow the topography of the disc. The slider includes a transducer that is used to encode the flux inversions and to read the flux inversions on the surfaces of the currently running disc.

2 is a perspective view of the actuator assembly 1118. Actuator assembly 118 includes a base portion 122, a plurality of actuator arms 126, a plurality of load beams 128, and a plurality of head gimbal assemblies 116. Base portion 122 has a hole provided for pivoting movement about axis 121 in a preferred embodiment. Actuator arm 126 extends from base portion 122 and is connected to a first end of one or two load beams 128, respectively. The load beams 128 each have a second end connected to the head gimbal assembly 116.

3 is an enlarged view of the head gimbal assembly [116]. The head gimbal assembly [116] includes a gimbal [130] having strut pairs [132] and [134], and a gimbal adhesive tongue [136]. Head gimbal assembly 116 also includes a slider 138 having an upper surface 140 and a lower air bearing surface 142. The transform section 144 is also preferably located at the front edge of the slider 138. Bonding between the slider 138 and the gimbal 130 may be in any manner desired. In a preferred embodiment, a flexible thin layer is bonded between the top surface 140 of the slider 138 and the bottom surface of the gimbal adhesive tongue 136 by an adhesive. The flexible thin layer is preferably a mylar film having a thickness of about 150 microns. The gimbal adhesive tongue 136 also ends with a mounting tab 146 at the rear edge of the slider 138, which provides a surface to which the slider 138 is attached to the gimbal adhesive tongue 136.

4 is a block diagram of a portion of the disk drive 110, showing the servo position control circuit according to the present invention. The disk drive 110 portion shown in FIG. 4 includes an actuator assembly 118, a disk pack 112, microactuators associated with each head gimbal assembly (collectively the microactuator 158), a full amplifier 160, Data and clock recovery circuit [162], error detection circuit [164], drive control section [166], data adjustment circuit [168], servo control processing section 170, power amplifying section [172], and optional microactuator control section [ 174].

The drive controller 166 is preferably a microprocessor or digital computer, or other appropriate microcontroller, and is connected via a bus 111 to a host system or other drive controller that controls a plurality of drives.

Disk pack 112 includes a spindle 176 that supports a plurality of coaxially disposed disks 178. Each disk 178 is installed to rotate with the spindle 176 about the axis of rotation 115. Each disk 178 includes a first surface 180 and a second surface 182. Surfaces 180 and [182] include concentric tracks for receiving and storing data encoded in the form of flux inversions within the track.

As described above in FIGS. 2 and 3, the actuator assembly 118 includes a base portion 122 that supports a plurality of actuator arms 126. Each actuator arm 126 is connected to at least one load beam 128. The load beam 128 supports each one of the head gimbal assemblies 116 (FIG. 3) on the disk surface 180 or [182], in order to access data in tracks on the disk surface. Each gimbal assembly also includes at least one microactuator 158 to position the transform portions of the head gimbal assembly in one track or a number of other tracks on the disk 178. As shown in FIG. 4, the microactuator 158 may include actuator arms 16, load beams 128, gimbals (or other bends) 130, between gimbals 130 and associated sliders, or other suitable. Locations may be provided. The microactuator 158 may be formed of a piezoelectric material, an electrostatic material, and bent by capacitive, fluidic, electromagnetic, electromagnetic, or thermal activity.

In operation, drive control 112 generally receives a command signal from a host system indicating that a portion of one or more disks 178 is to be accessed. In response to the command signal, the drive controller 166 provides the position (or reference) signal 165 to the servo control processor 170. The signal is a signal that indicates the particular cylinder in which the actuator assembly 118 is present to position the head gimbal assembly 166. The servo control processor 170 converts the position signal into an analog signal. The analog signal is amplified by the power amplifier 172 and provided to the voice coil motor in the actuator assembly 118. In response to the analog position signal, the actuator assembly 118 positions the load beams 128 and associated head gimbal assembly 116 over the desired cylinder.

The head gimbal assembly 116 generates a read signal including data from general data accessed in the selected portion of the disc being read, and embedded servo position data stored in the selected portion of each track of the disc being read. The read signal is provided to a preamplifier 160 that amplifies the read signal and provides it to the data and clock recovery circuit 162. The data and clock recovery circuit 162 recovers, in a known manner, the data encoded on the disk surface when the data is written to the disk surface. Of course, data and clock recovery circuitry 162 may be a partial response maximum likelihood (PRML) channel or other suitable type of read channel.

Once the data is recovered, it is provided to an error detection circuit 164 that provides an output 167 and detects whether an error has occurred in the data read from the disk. Errors are corrected in a known manner by the error detection circuit 14 or the drive control 166 or a combination of both.

During head positioning, the drive controller 166 provides a position signal to the servo control processor 170 to cause the actuator assembly 118 to position the head gimbal assembly 116 over the selected cylinder. In a sector servo positioning device (or built-in servo positioning device), the portion of each sector on the disk surface is encoded there, read by the data head data head and provided to the servo control processor 170 via the read channel. Has information The position information not only provides approximate position information indicating the specific track on which the data head slides, but also provides tuning feedback to the servo control processing for better positioning. The servo control processing unit 170 correspondingly positions the head gimbal assembly 116 in response to the positional information read from the disk.

In a preferred embodiment, the servo control processor 170 is used to control the actuator (voice coil motor) approximately, as well as to control the microactuator 158. In another preferred embodiment, a separate microactuator control (or a plurality of separate microactuator controls) [174] responds to the position request signal from the drive control [166] and embedded position information read from the disc. [158] is provided to control.

In order to record the information on the discs, the drive control section 166 receives the actual data to be recorded as well as the position of the information recorded on the disc pack 112. The position information is provided to the servo control processor 170 (and optionally the microactuator control (s) [174]) as a reference signal to roughly position the data heads relative to the corresponding disk surfaces. Thereafter, the drive control section 177 provides the data to be written to the data control circuit 179 and the data control circuit provides the information to outputs 169 to specific converters on the head gimbal assembly 116 so that the data can be stored. It is to be recorded on the surface of the disc in a known manner.

In a preferred embodiment, the microactuators 158 have a range of motion that exceeds the worst mechanical misalignment between all two head assemblies 116 supported by the actuator assembly 118. In a more preferred embodiment, each microactuator 158 has a range of motion that exceeds one track width, more preferably multiple track widths. In addition, in a preferred embodiment, the read channel provided in the disk drive 110 (in the embodiment shown in FIG. 4, includes the preamplifier 160, the data and clock recovery circuit 162 and the error detection circuit 164). May receive a number of simultaneous and parallel data signals, process the data signals in parallel, and provide them in parallel to the host system and / or drive control [166]. Moreover, in a preferred embodiment, the data conditioning circuit 168 is also suitable for providing a plurality of simultaneous and parallel write signals to the data head to perform simultaneous and parallel write operations. Furthermore, in a preferred embodiment, the servo control processor 170 and the optional microactuator controller 174 provide a position signal to the microactuator 158 to simultaneously or simultaneously position all or at least a plurality of microactuators (multiple heads In order to be able to be aligned simultaneously with the tracks on the multiple disc surfaces of the disc pack 112.

With this arrangement, many advantages are obtained. For example, accurate position control is achieved for each of the plurality of data heads. This enables accurate and simultaneous track tracking for multiple data heads, enabling parallel read and write operations. Moreover, the bandwidth of the microactuator operation is higher than that of the voice coil motor, thus providing the ability to significantly increase the track density on a given disk surface. The reason is that the bearing nonlinearity and other problems that limit the track density of the current technology in fine structures using voice coil motors for track following can be adjusted.

Moreover, in a preferred embodiment, each microactuator has a range of motion exceeding a plurality of data tracks, so the microactuator itself can be used to perform a short search operation (a search operation to search for a track within the operating range of the microactuator). have. This minimizes the occurrence of structural modes in the disc drive and enables higher bandwidth servo-control during short seeks.

Microactuator 158 may be controlled in a number of ways in accordance with the present invention. For example, the voice coil motor in the conventional disk drive unit is controlled by a servo control unit which is a SISO system. The input is a head position value obtained from the embedded servo data, and the output drives the voice coil motor through the power amplifier 172. However, in controlling the microactuator 158, the servo control system of the present invention may have multiple inputs and multiple outputs. The input includes head positions read from the embedded servo information on which the heads slide, and may optionally include the relative positions of one or more microactuators relative to the voice coil motor (or approximate actuator). Multiple outputs include a single coarse positioning unit (VCM) and outputs driving N microactuators.

A potential problem given to the current structure in which multiple data heads are located at the same time is that the positioning of one data head can be affected by the simultaneous positioning of adjacent or adjacent other data heads. High bandwidth positioning may excite the structural modes of the drive and cause vibration or other interference that may interfere with the positioning of adjacent data heads. Thus, in a preferred embodiment of the present invention, the servo control processor 170 or the micro actuator controller 174 regulates the movement of other data heads of the actuator assembly 118.

The preferred arrangement includes a single servo control including a MIMO servo control implemented as a digital signal processor (DSP). The DSP has inputs indicative of the head position of each head, the relative position of each microactuator, and a reference signal at either the host control or the disk drive control [166].

FIG. 5 shows an embodiment in which the servo control processor 170 and the microactuator 174 are combined into a single servo controller implemented as the DSP 190. The DSP 190 receives, as an input, a reference signal [165] from the host system or disk drive control section [166]. It also receives head position signals for each of the heads 181 (heads 0 to N) indicating the head position. The output at DSP 190 is provided to a voice coil motor (VCM) 173 that drives the actuator assembly (or D-block) 118. 5 also shows that DSP 190 provides output signals to all microactuators 158 coupled to actuator assembly 118 (microactuator 0 to microactuator N are indicated as [159] to [161]). Illustrated. The input to the DSP 190 also includes a head position signal [171] at the relevant heads, and relative position signals [179] of each head relative to the VCM [173]. Each microactuator is coupled to a head 181. Each microactuator also has an associated relative position sensor [183] for sensing the relative position of the microactuator relative to the voice coil motor (or actuator assembly [118]). Relative position sensor 183 may include a suitable sensor, such as a capacitive sensor or other suitable type of position sensor. Thus, it can be seen that the DSP 190 provides the approximate position signal to the voice coil motor 173 to position the entire actuator assembly 119. DSP [190] also provides a good position signal to microactuator [158].

In a preferred embodiment, when providing microactuator output for each microactuator associated with each head, the DSP [190] not only considers the head position and relative position associated with each head, but also considers the movement of adjacent or proximate heads. (Ie, consider cross-coupling of the heads). In this manner, when positioning each of the individual heads, the DSP 190 takes into account structural mode excitation. In addition, in a preferred embodiment, DSP 190 provides disturbance rejection to account for runout, to prevent excessive actuator current, and to provide noise protection.

The DSP [190] shown in FIG. 5 implements a model based algorithm. 6 (e.g., blocks [191] through [199]) is a flow diagram illustrating one method in accordance with the present invention for developing a servo control system implemented with a DSP [190]. Each block in FIG. 6 will be described briefly in comprehensive terms and will be discussed in detail later.

The nominal model first consists of a set of differential equations describing the structural dynamics of the disc drive. This is indicated by block [192] in FIG. In constructing a nominal model, it is desirable to know the exact knowledge of the driving dynamics. The knowledge is preferably obtained using empirical / measured data. The type of information is used to define the parameters in the system model, and all the information is used to build the model in first principles (or differentials).

Next, an uncertainty description is made for the disk drive. This is shown by block [194] in FIG. The uncertainty statement is preferably designed to capture drive movement relative to multiple drives in response to variations. The data is used to modify the model configured at block [192].

The performance target of the system is then specified. This is indicated by block [196] in FIG. The models are provided so that the model is compatible with existing commercialized tools to optimize and complete the development of control algorithms that can be implemented in various ways directly to the disk drive.

The model is then validated. This is shown by block 197 of FIG. Briefly, this instantaneous model is a robust control model that is not only a model of system dynamics but also a system for uncertainty and noise descriptions. In general, the model validation problem is constructed as a linear time invariant system with norm-bounded configured uncertainty and experimental data. Model validation is performed by considering model validation issues in the frequency domain. Techniques referred to as μ-decomposition and μ ₉ -decomposition are performed to determine whether the model matches the data and whether the control unit matches the model.

In the end, the model is optimized using commercially available optimization software. Algorithms for calculating and optimizing μ are available under the trade name MATLAB μ- Analysis and Synthesis Toolbox. This is represented by block 198 of FIG.

A. Actuator Model

7 shows a model 201 of a servo actuator according to the present invention. 7 includes blocks [200], [202], and [204]. Block [200] represents the servo actuator function G _nom . Block [204] represents the addition uncertainty statement and block [202] represents the multiplication uncertainty statement.

1.Nominal Model

G _nom is the actuator nominal model. The role of the nominal model accurately characterizes general system dynamics. This means that the nominal model should be the average expected system movement observed over the entire class of modeled disk drives. In a preferred embodiment, the nominal model is developed by obtaining a significant amount of accurate knowledge of the drive.

In a preferred embodiment, the nominal model is constructed in the frequency domain, since the worst case in system dynamics may include higher frequency structure modes that will result in highly variable momentary movement in the time domain. Since the disk drive is open loop instability, data is collected in a closed loop. This is accomplished by injecting disturbances into the loop and measuring the inputs and outputs to the drive actuators. For example, in a preferred embodiment, frequency domain input and output information is collected through empirical tasks. Also time domain responses according to other types of inputs, such as pulse or step inputs, are collected. This type of information is used to define the parameters of the system model. Furthermore, in a preferred embodiment, finite element decomposition is performed to obtain structural data for the various physical components in the system. All of this information is used to construct the model in the first principles (or differentials).

8 is a diagram of the frequency [203] for log size [205], showing the open loop transfer function obtained for a typical disk drive. Data was collected by injecting sinusoidal position disturbances into the measured head position during the track following operation and simultaneously measuring the voice coil motor current and head position. [206] represents a nominal model of the disk drive. [208] represents measured data when tracking by the PID control unit in the disk drive unit, and [210] is an average of measured data. Data was collected at the inner, middle and outer radii of the disks on heads 0, 1 and 3. The position disturbance amplitude was 2% to 20% of the track width.

The nominal model was chosen to be dependent on the given secondary system as follows.

Equation 1

In the above, K _DC is the actuator DC current gain, n is the number of resonant modes for the i-th (i = 1, ..., n) resonant mode, ξ _{z, i} and Is a damping and natural frequency of zeros. Ξ _pi and Is the damping and natural frequency of the poles. Similarly, ξ and ω _n define the low frequency movement of the drive.

Delay can also be modeled in the nominal model described by equation (1). Various delay sources exist within the disk drive. For example, pulse width modulation (PWM) filters, power amplifiers, voice coil time constants, and other sources can all inject delays in time. The standard procedure for modeling pure delay in continuous-time uses the following Pade approximation.

Equation 2

Where t _d is the time unit delay, s is the frequency parameter, and L {.} Represents the Laplace transform of the independent variable.

In order to model the time delay, low order (eg 1st or 2nd) Farde approximations are suitable. Other robust approximations are also possible.

2. Uncertainty Statement

Since the exact mathematical model of the actuator system in the disc drive is not possible, the uncertainty statement designated by blocks [202] and [20] 4 of FIG. 7 is used. An accurate model means that the system can be represented in a set of differentials and all parameters of the differential are known. Under normal circumstances, it is not possible to accurately model the disc drive. However, it is possible to define a set of models that "cover" the movement of the disc drive. Uncertainty modeling is one way of defining the set.

In a typical disc drive, the disc drive structure has a resonant mode at high frequencies. Moreover, the natural and damping frequencies in the low frequency mode are somewhat amplitude dependent. In other words, it is defined by some nonlinear dynamics that are most likely to occur in bearing hysteresis, flux circuit bias forces, and so on. The variation is most effectively placed in the uncertainty statement.

Block [204] shows an increase in the nominal model that is block [200] with addition uncertainty. Addition uncertainty is shown in the following set of systems.

Equation 3

Δ (s) is the standard-satisfied uncertainty perturbation, and W _a (s) is the frequency domain weighting function. The uncertainty fluctuation Δ (s) is standard-satisfied to be generally used to specify all frequency dependencies according to the magnitude of the uncertainty. Addition uncertainty is commonly used to characterize dynamic uncertainty. Therefore, uncertainty fluctuation Δ (s) will be generally thought of as a composite full matrix at each frequency.

Addition uncertainty is most effectively used to model high frequency dynamic uncertainties, such as the resonance mode in the disc drive. For example, little or no reliable information is available about the dynamics of disc drives available in the frequency range above about 2 kHz. This is because the natural and frequency damping of the structural modes vary from system to system.

The role of the addition uncertainty weighting W _a is to introduce a lower limit to the system gain at higher frequencies.

9A and 9B show a diagram of magnitude [211] versus frequency [209] and the use of addition uncertainty. In FIG. 9A, the nominal model is designated [212] and the actual measured values are designated [214]. The addition uncertainty used to account for unmodeled dynamics gives an envelope, indicated by [216]. 9B illustrates one embodiment of an add uncertainty weighting function [218].

Multiplication uncertainty describes a set of systems as follows.

Equation 4

Δ (s) is the standard-satisfied uncertainty fluctuation and W _m (s) is the frequency domain weighting function. Along with addition uncertainty, Δ (s) is generally thought of as a composite full matrix at each frequency.

Multiplication uncertainty is effective in modeling low frequency gain variations. 9C illustrates using multiplication uncertainty to account for gain variation. FIG. 9C is similar to FIG. 9A, and like elements are numbered with like numbers. However, [216] shows the envelope (envelope) for a set of systems with multiplication uncertainty at lower frequencies. The role of the multiplication uncertainty weighting W _m is to provide an envelope around the nominal transfer function. In many cases, the envelope is + or-about 50% up to about 100 Hertz, and then gradually decreases below 1%. FIG. 9D shows the multiplication uncertainty weighting function [218] for the uncertainty [216] shown in FIG. 9C.

In the given model shown in FIG. 7, the designer's role is to select the uncertainty weights W _m and W _s that adequately cover the data collected at the disk drive, not too conservatively. Several simple estimates are provided through FIGS. 10A-10C showing the frequency [215] for magnitude [217] and phase error [219].

Examples of addition and multiplication uncertainty levels in a typical disk drive are shown in FIGS. 10A, 10B, and 10-3. 10A shows estimated multiplication uncertainty [220]. FIG. 10B shows the estimated addition uncertainty [222], and FIG. 10C shows the phase error [224] as a function of frequency (used for delay estimation). Note that addition and multiplication uncertainty can be used for both low, medium and high frequency uncertainty modes. The design selects an uncertainty model for simple weighting. Moreover, multiple addition and multiplication uncertainties may be used.

B. ㅗㅗ H _∞ and μ-synthetic control design

Once the actuator model 201 shown in FIG. 7 is obtained, the next step is to design a control that is optimized for specifying performance goals and controlling the modeled actuator. 12 shows a block diagram of a track follower control synthesis interconnect 230 representing a disk drive. Note that system 230 includes actuator model 201. The control is designed to optimize the weighted interconnect 230.

In FIG. 11, the system block is preferably for control block [232], actuator model [201], ideal actuator model [234], current and tracking error signals [235] (u) and [237] (y), respectively. Positioned composite weights [231] (W _u ) and [233] (W _p ), external inputs including center offset position [239] (r), position sensor noise [241] (n), signal [269] (y _m ), weighting [267] (W _n ), current disturbance [243] (d _i ), input weighting [259] (w _di ), output weighting [261] (w _d0 ) and position disturbance [245] ( d ₀ ), and controlled outputs [247] (z1) and [249] (z2) corresponding to current anomalies and tracking errors, respectively.

In a preferred embodiment, the control block [232] is provided in a two degree of freedom control arrangement comprising a compensator K ₂ and a precompensator K ₁ . Complement compensator K ₁ is included to improve the instantaneous response of the closed loop system.

The interconnection shown in FIG. 11 is preferably designed to satisfy the following equations describing actuator current anomalies and tracking errors.

Equation 5

Equation 6

The interconnect shown in FIG. 11 is simplified to the equivalent linear fractional transformation shown in FIG. 12 that includes the controller 261, the actuator 253, and the uncertainty 265. Linear fractional transformation (LFT) is often used in modern control theory to simplify the representation of interconnections in linear systems, and is also considered as a mathematical representation of a block diagram consisting of linear blocks with feedback.

In the LFT shown in FIG. 12, the resulting _H∞ optimization criteria are as follows.

Equation 7

In the above, F ₁ (F _u (Δ, G), K) shows the feedback interconnect w → z.

In the above, ∥I _∞ represents the derived ∞-standard of independent variables, and C is a set of complex numbers.

For robust performance, the closed loop system includes a control K to stabilize the set of plants.

Equation 8

{P: P = F _u (△, G), ∀ △ εβ △},

In the above,

β △ ₁ = {△ ε △ ₁ : ∥ △ ∥ _∞ ≤1}

It satisfies the following performance target.

Equation 9

∥F _l (F _u (△, G), K) ∥ _∞ <1∀ △ εβ △

Therefore, in equations 5, 6 and 7, the following inequality can be derived.

Equation 10

∥ω _j → z _{i ∞} ≤ ∥ω → z∥ _∞ , ∀i, j

And in the _{∥ω → z∥ ∞ = ∥F l (} F u (△, G), K) ∥ ∞, ω j → z i z _i is the transfer function to at ω _j.

Therefore, when the control unit K is designed to satisfy ω ω z _∞ <γ, in the H _∞ optimization criterion from Equation 7 and the mean inequality from Equation 10, the following performance inequality is valid at each frequency ω.

Equation 11

Equation 12

In addition, the following actuator inequality (actuator inequality is also valid):

It can be seen that the performance inequalities from Equations 11-14 serve as a set of performance limits on the system. Equation 11 defines the tracking limit, Equation 12 defines the input disturbance rejection (rotational vibration) limit, Equation 13 rejects the current disturbance, write error rejection, repeatable runout rejection, nonrepeatable runout rejection, and bias current error. Defines a disturbance rejection, including, and Equation 14 defines a noise rejection constraint.

The tracking constraints in Equation 11 are located on the system such that by selecting weights [233] (w _p ) and [251] (w _r ), the designer allows the controller to possibly meet the desired response. For example, the designer would like to have a step response of the actuator to have a certain appearance in the frequency domain (eg, when moving a track or performing a read-write offset). In general, the weights will be chosen such that the tracking constraints are severely damped. The ringing can excite the structural modes of the drive, and severe damping tends to avoid the problem. It should be noted that providing the controller K as a degree of freedom 2 control provides additional degrees of freedom for the system to implement tracking constraints. The transfer function G _ideal in equation 11 is also preferably designed to provide an ideal response to the command offsets. Then, the performance goal given by Equation 11 ensures that the disk drive response matches the G _ideal response. This type of tracking goal is referred to as explicit model following. There is also an implicit technique that does not require embedding the model in the composite interconnect, avoiding additional states.

However, obvious techniques are preferred because they are simpler to design.

The input disturb rejection performance limitations in Equations 12 and 13 allow the track following servo system to reject disturbances from external sources and thus follow constant tracking performance under normal operating conditions. For example, the search appearance generates actuator position and speed errors at the time the tracking control is switched on. The user configuration also allows the disc drive to be subject to translational and rotational shock and vibration. Good tracking performance thus requires the servo system to react quickly and to attenuate continuous and stepped (or momentary) bias currents, torque, position and velocity disturbances. Moreover, very high amplitudes or long time constant ringing are undesirable for such systems. Requirements such as fraud are satisfied by standard-limits on specific transfer functions. In some defined frequency ranges, the effects of bias current and torque disturbances can be minimized through standard-limiting attenuation induced as follows.

Equation 19

More attenuation from DC to the critical frequency results in faster response. Ringing is caused by peaks in the transfer function.

Similarly, positional disturbances

Equation 20

Attenuation equations 19 and 20 can be obtained by setting the performance weights of equations 12 and 13.

While the system is unable to filter or reject sensor noise, it is desirable to be designed so that the noise does not become more severe. This may be possible if there is peaking in the transfer function from equation (14). The peaking represents a gain of 0 dB or more in some frequency ranges. Peaking is generally used to refer to unwanted amplification in a closed loop transfer function. Where high gains exist, the transfer function is nearly constant. Therefore, when there is excessive noise at frequencies where the control unit is expected to have high performance, it is desirable to specify a noise filter using priority information on the noise spectrum. In the disc drive, this is achieved by using a repeatable runout compensator which fixes in accordance with the repeatable noise caused by the recording error in accordance with the servo track recording system.

Actuator limits from equations 15-18 include tracking limiting (equation 15), input disturbance limiting (equation 16), output disturbance limiting (equation 17), and noise limiting (equation 18).

The designer's role is to select the frequency weights w _p , w _u , w _r , w _di , w _do , and w _n corresponding to the performance requirements for the disk drive. If the performance requirements are easily expressed in the frequency domain, the weighted selection is simple. In general, there will also be time domain requirements. For example, such requirements include overshoot, rise time, settling time, and ringing, which are equivalent characteristics (e.g., bandwidth, Attenuation, picking, etc.).

Actuator limits (or targets) from Equations 15-18 penalize the actuator signal. This plays several roles. For systems with little or no uncertainty, the goals allow the control to operate to limit bandwidth. However, these goals are used to limit the size of the control signal, minimizing the possibility of saturation of the control signal of the servo actuator.

Adjusting related weights is an iterative process using data from simulations and tests to determine appropriate weights.

C. Robustness and Model Validation and Optimization

Once the model is developed and the appropriate weights are selected, the robustness of the model is validated. In a preferred embodiment, this is achieved using H _∞ and μ synthetic techniques. The matrix function μ can be calculated and used to analyze the stability and performance characteristics of the interconnect structure shown in FIG. 12, including the controller [261], actuator [163], and uncertainty [265]. μ analysis is used to analyze the stability and performance characteristics of interconnected systems that are subject to standard-satisfied uncertainties. 12 provides a general interconnection consisting of uncertainty suitable for μ analysis and μ synthesis.

The matrix function μ is calculated for the system, and then the function μ _g is calculated. μ _g is an extension of the μ framework in which the shake block is divided into two sets (one that satisfies the maximum standard limit (similar to mu) and another that satisfies the minimum standard limit). This formulation will solve the model validation problem. Computable lower and upper boundaries for μ _g are known.

In essence, after synthesizing the control, μ defines the maximum size of disturbances, noises, and model uncertainties that the closed loop system preserves stability and performance. For a given set of data, model validation μ _g -analysis defines the minimum size of disturbance, noise and model uncertainty needed to fit the model to the data. Closed loop systems are robust if the control is capable of preserving stability and performance for the actual system. Simply assign, and if μ (ω) is less than μ _g (ω) ∀ωεΩ, the model matches the data and the control unit matches the model. Therefore, the closed loop system is robust over the frequency set Ω.

13 is a graph of frequency [271] versus magnitude [273] showing stability and performance robustness for a closed loop tracking control. The μ _g function is represented by [242]. Μ (ω) for the performance target is indicated by [244] and μ (ω) for stability is indicated by [246]. At each frequency, the closed loop system is stable when μ is less than μ _g . The critical frequency range within the disc drive occurs where structural modes are uncertain (ie about 1 to 2 kHz). Accordingly, it can be seen from FIG. 13 that the closed loop system may lose stability within the frequency range of 2 kHz to 4 kHz. There are some frequencies in the range where μ _g falls below μ (ω) for both stability and performance analysis.

By iteration over uncertainty and performance weighting, the robustness of the stability can be improved. This requires either or both of increasing the level of uncertainty or decreasing the desired level of performance. Each iteration requires that the μ consistency analysis is recalculated and the uncertainty and performance weighting adjusted accordingly.

Adjusting the control unit according to the invention is preferably carried out in two steps. The first step is simulation and the second step is implementation. In both stages, open and closed loop data are used to evaluate the effectiveness of the control design. According to the data, the nominal model, uncertainty weighting, or control performance weightings, or all of the above, are adjusted.

One simple technique to tune is to adjust the frequency weighting w _p with the performance limitations shown in FIGS. This directly affects all relevant performance goals. In this way, the bandwidth of the controller can be easily adjusted by multiplying w _p by a scalar. This eliminates the need for more complex processes of each of the other frequency weightings in the performance limits shown in Equations 11-14.

This type of adjustment can be preferably applied when installing a new servo control on a new disc drive product. Weighting in previous generation designs can be used with a scalar of w _p to reduce the bandwidth of the new disk drive product until a specific design weighting is completed. This creates the potential to start new products faster.

In addition, one parameter (loop gain of the servo control) can be adjusted on-line. In fact, the loop gain can be adjusted per drive principle. For example, by scalaring the servo control, the loop gain can be increased or decreased resulting in an increase or decrease in the closed loop bandwidth.

FIG. 14 is a flowchart including blocks [308-330] showing the above-described process in which the disk drive model is validated and optimized. First, all weights for uncertainty descriptions and performance constraints are selected by the designer to design the model. Of course, the items can be intuitively selected through experimentation. This is indicated by blocks [310] and [312].

Next, matrix function [mu] ([omega]) is calculated to perform [mu] analysis on stability and performance. This is indicated by block 320. The function μ (ω) can be calculated using the technique described above or other known technique (or the upper and lower boundaries of the function can be calculated).

μ _g (ω) is then calculated using the weights selected at block [310]. This is indicated by block [322]. As mentioned above, once μ (ω) and μ _g (ω) are calculated, two things are known. First, you need a minimum of uncertainty in the models for fit to the data (or data matching) model known in μ _g. It is also known (from μ) that the maximum amount of uncertainty the control can allow and follow stability and meet performance limitations.

Thus, the next step is to compare μ and μ _g to ensure that μ is below μ _g at all desired frequencies. If this is not true, there are some frequencies in the dependent frequency range for which the control does not achieve robust performance. In other words, there are some frequencies that become unstable, fail to meet performance limitations, or both. In that case, the process returns to block 310 where new weighting functions are selected. This is indicated by block [324]. Just as μ _g is calculated with the new weighting functions, μ (ω) for performance and stability analysis is then recalculated. This is indicated by blocks [310-322].

However, at block 324, it is determined that μ is less than or equal to μ _g at all desired frequencies, and robust performance may be achieved by the control obtained when using the weighting function selected at block 310. This is indicated by block [326].

Next, it is determined whether further optimization of the model is necessary. If so, the process returns to block 310 where a new weighting function is selected again. To optimize the system, weighting functions can be adjusted to provide better performance or less uncertainty or both. μ and μ _g can be calculated using the new values to determine whether the control continues to provide robust performance. This can be repeated until the optimization reaches the desired level. In other words, while following a control that provides robust performance in the dependent frequency range, the amount of conservative trends in the model can be repeated until it is reduced to achieve the desired performance level. Once generated, optimization is not desired, the model has already been validated and the control has already been optimized to the desired part and the design is complete so that the control can be realized. This is indicated by blocks [328] and [330].

D. Implementation of Control Unit in Digital Signal Processing Unit (DSP)

After obtaining the transfer function for the control unit as described above, the frequency domain transfer function has a discrete-time state-space realization as follows.

Equation 21

x (k + 1) = Ax (k) + Bu (k)

y (k) = Cx (k) + Du (k)

In the above,

x ∈ R ⁿ is the control state,

u ∈ R ^m is a control input (e.g., derived from a position error signal (PES)),

y ∈ R ^p is the control output (e.g., output to a voice coil motor and / or microactuator),

A, B, C, and D are matrices of appropriate dimensions

R ⁿ , R ^m , and R ^p are the n-, m-, and p-dimensional vectors of real.

However, the implementation of the discrete-time system into the digital signal processor can be problematic. As discussed above in the context of the application, the above problems include intermediate calculations exceeding register capacity, general computational power and memory capacity of conventional signal digital processing units, conversion from discrete-time systems to fixed-point processing units, and within the control unit. Problems related to the reduction and elimination of quantization errors.

Fig. 15 is a flowchart including blocks [339 to 347] showing one preferred technique for implementing the control unit in the digital signal processing unit according to the present invention. Each block in FIG. 15 will be briefly described, and then each block will be described later in more detail in the applications.

First, as described above, matrices for the discrete-time state-space realization from equation 21 are obtained. This is indicated by [340].

The states of the control are then scaled. In one preferred embodiment, under the worst-case prediction, the boundaries of the controller states are determined and state transitions are made to scale the boundaries to the desired level. This significantly reduces the possibility of overflow during intermediate control calculations. This is indicated by block [342].

The states of the control are then converted into a more desirable structure. In one preferred embodiment, the states of the control section are transformed into a two-orthogonal structure that significantly reduces the computation time and memory requirements of the digital signal processing section. This is indicated by block [344].

Finally, the control parameters are converted to shift from the floating point format to the fixed point format. In carrying out the above, it was necessary to be around to significantly reduce the quantization errors that resulted in acceptable levels. In one preferred embodiment, the matrices are shifted to the fractional binary format to position the decimal point in the fractional binary representation at the desired location. This is indicated by block [346].

1. Scaling of controller states to reduce overflow

The state calculation for discrete-time realization from equation 21 is

Equation 22

Where x (0) is the initial condition.

The present invention minimizes or at least reduces the likelihood of encountering an overflow in the intermediate control calculation. In a preferred embodiment, this is done by automatically scaling the states of the control. Thus, for each sample k, for each input u _i

Equation 23

Can be assumed (generality is maintained)

In the above, m is defined as the number of inputs to the controller. Under the worst-case assumption, the k th sample of the state at the controller satisfies the following boundaries.

Equation 24

Where m and M _j are defined as in equation 23.

For a stable system (i.e., p (A) <1), there is a predetermined constant N where A ^N becomes very small. Then, in equation 24, the states described in equation 24 satisfy the following boundaries.

Equation 25

The boundary on the control states given by equation 25 (the worst case for the samples) can be used to form a state transition.

First, T is defined as follows.

Equation 26

In the above, diag (.) Is defined as an orthogonal matrix whose main diagonal is a vector separated by an independent variable (.).

Then, state transition z = T (x) will satisfy the boundaries.

Equation 27

Therefore, the new controller realization is as follows.

Equation 28

Note that the controller that implements the pure integrator cannot be stable because ρ (A) = i. In this case, there is no integer N where A ^N approaches zero. In many cases (and especially in H and μ-synthesis), the control will not include pure integrators, but will have a spectral radius close to a single. Thus, there will be an integer N where A ^N approaches zero. The integer N will be very large. This means that the state transition from equation 26 will result in very conservative boundaries for everything except integral states. Therefore, in the case where the spectral radius of the controller is very constant, an appropriate anti-integrator windup technique is implemented to reduce the likelihood of integrator overflow. The remaining control state is limited as described above.

2. Conversion of scaled states to reduce the need for computational power and memory capacity in the DSP

The discrete-time system from equation 21 may have arbitrary elements in matrices A, B, C, and D. Thus, in each sample, this may require up to (n + p) x (m + m) multiplication and (n + p) x (n + m−1) addition. This is very undesirable because it requires a very unrealistically large number of mathematical operations (multiplication and addition) to be performed by the controller. However, using state transitions, the number of multiplications and additions may be greatly reduced.

In embodiments where the control is implemented as a single input single output (SISO) control, the canonical forms may be applied to be very computationally simplified. In this embodiment, the preferred form of control norm for realizing the control from Equation 21 is as follows.

Equation 29

In the above, c _c and d are arbitrary matrices, and A _c and b _c are defined as follows.

Equation 30

Where n is the number of states of the control and a _i is the coefficient of the s ⁿⁱ term in the characteristic polynomial for A.

In each sample, it can be seen in equations 29-31 that there are only 2n + 1 multiplications and 2n additions for realizing the control normative form. This is reduced over (n + 1) x (n + 1) multiplication and n x (n + 1) addition to the realization from equation 21. While state transitions are highly desirable for SISO implementations, the normative form will not be well generalized for multiple input multiple output (MIMO) systems. Thus, in a MIMO implementation, two-orthogonal realization can be applied to obtain significant computational simplification.

For the purposes of current applications, the two-orthogonal realization is that the matrix of equation 21 is separated into a block orthogonal matrix of real eigenvalues and complex conjugate pairs. This is a conversion of the realization of Equation 21, resulting in the following realization.

Equation 32

In the above, B _bd , C _bd , and D are arbitrary matrices,

Equation 33

Where r is the number of real eigen values of the controller, C is the number of complex conjugate eigen value pairs (n = r + 2C), and each λ _i is the first of A in equation 21 Corresponds to the Eigen value, and each Is a 2 x 2 matrix corresponding to the j th complex conjugate eigen value pair of A.

Thus, in equations 32 and 33, there are nx (m + 1) + px (n + m) + 2c multiplication and nxm + px (n + m-1) + 2c addition for each sample for a 2-orthogonal realization. You will see that. In other words, for each sample, the two-orthogonal realization from equations 32 and 33 has n ² -n-2c smaller multiplication and n ² -n-2c smaller addition. Moreover, as the number of states increases (for the number of inputs and outputs to the controller), the calculation for the 2-orthogonal realization is linear in the number of states for the quadratic equation for the standard realization from equation 21. to be. Note that a two-orthogonal transformation is used to scale the states of the control and can be applied before the transformation bounding the states. This allows the integration of states to be independent prior to scaling in order to minimize or prevent overflow in intermediate computations.

3. Conversion of matrices to fixed-point format

Since the digital signal processing unit is a fixed-point control unit, the matrix defining the discrete-time control unit must be converted to the fixed-point format so that the control unit is realized on the DSP. Floating point numbers can be converted to fixed-point equivalents using fractional binary digits. Fractional binary numbers are represented as binary numbers followed by binary decimal points followed by binary fractions. For example, fractional binary numbers can be expressed as follows.

Equation 34

Where N is the word length (or number of bits in the binary word), Q is the position of the binary decimal point, and (b _n-1 , b _n-2 , ... b ₁ , b ₀ ) is binary Bits in the word.

For the purposes of this application, fractional binary digits are represented by Qn, where n is the position of the binary decimal point. For example, the Q15 fractional binary number represents a binary number with 15 bits after the binary decimal point.

The elements of the matrix from the two-orthogonal realization of equation 32 must be converted to fractional binary digits. In a preferred embodiment, this is achieved by right-shifting the elements of each matrix until the matrix can be represented in the desired fractional binary format. Thus, the fixed-point two-orthogonal realization from equation 32 is realized as follows.

Equation 35

S _A , s _B , s _C , and s _D are positive numbers equal to the number of right shifts required to represent each matrix in fractional binary (Qn) format. And,

silver Corresponds to the Qn expression of.

In one preferred embodiment, the discrete-time control realization from Equation 21 is a fixed-point, with a standard arithmetic logic unit (ALU) that provides coded and uncoded multiplication and addition instructions for 16-bit independent variables. It is implemented on a 16-bit digital signal processor. The result and sum are stored internally in 32-bit form. In a preferred embodiment, the processor is designated as the same as the TMS320C2xLP DSP core from Texas Instruments. Experimental studies have shown that using 16 bits (ie, Q15 format) for the control matrix, measurements and outputs provides acceptable operation. However, 32 bits (Q31 format) were desirable for state calculation.

Accordingly, the present invention provides a preferred disk drive in which the model-based servo control is implemented. The servo control controls the movement of the actuator so that the converter moves relative to the surface of the rotatable disk in the disk drive. The model-based servo control provides robust track following control of the servo system over the desired frequency range.

In a preferred embodiment, the servo control 190 is a track located in the nominal performance [200] of the disc drive 110, the uncertainty description [202], [204] characterizing the disc drive, and the servo control model [201]. It is synthesized in the servo control model [201] considering tracking performance limitations. In another preferred embodiment, multiple converters 181 are provided for multiple disk surfaces 180, 182. Actuator arm [118] includes arm portion [126] and a plurality of suspension assemblies [128], [116], where each suspension assembly [128], [116] has a corresponding disk surface [180], [182]. The conversion unit 144 is supported by the &quot; In this embodiment, the actuator 173 comprises a coarse actuator connected to the arm portion and a plurality of microactuators 158, wherein at least one of the plurality of microactuators 158 is the arm portion 126 and the corresponding one. It is operatively connected to the converter 144 and controls the movement of the corresponding converter 144 relative to the arm portion 126.

In another preferred embodiment, the servo controller 190 has a plurality of error inputs configured to receive error signals 171 at the converter 144, and a command signal [165] and error signals [171]. Accordingly comprises a plurality of control outputs configured to provide control signals to the plurality of microactuators 158 and the approximate actuator 173. Moreover, in another preferred embodiment, the servo controller 190 is configured to receive a plurality of relative position signals 179 indicating the relative position of the plurality of microactuators 158 relative to the approximate actuator 173. Relative position inputs of. In this preferred embodiment, the plurality of relative position sensors 183 are operatively connected to the plurality of microactuators 158 and provide a plurality of relative position signals 179.

In another preferred embodiment, the servo control 190 is a servo control signal to compensate for the structural resonance of the disk drive 110, position input disturbances, and / or simultaneous movement of the plurality of actuators 118, 158. It is configured to adjust.

The features and advantages of embodiments of the present invention have been described together with the detailed description of the structure and function of the various embodiments of the present invention. The technique is for illustrative purposes only and various changes are possible within the broadest scope defined by the claims. For example, certain elements may be used depending on the particular drive or drive type without departing from the spirit of the invention.

Claims (10)

A servo system for controlling the positioning of a converter relative to a disk surface having tracks containing information of a rotatable disk in a disk drive, An actuator arm operatively connected to the converter; An actuator operatively connected to the actuator arm to move the converter with respect to the disc surface to follow one of the tracks during track following; And And a model-based servo control operatively coupled to the actuator to provide a servo control signal for controlling the movement of the actuator, the model-based servo control having a desired performance level over a desired frequency range. And provide a servo control signal to take into account variations in operating parameters of the servo system from nominal operating parameters to achieve stable track following control of the servo actuator. The method of claim 1, wherein the model-based servo control unit comprises: And a servo control synthesized from a servo control model, wherein the servo control model includes a nominal performance of the disc drive, an uncertainty description characterizing the disc drive, and a track following performance limit located in the servo control model. Servo system which takes into account track following performance constraints. The method of claim 2, wherein the conversion unit provides an error signal indicating a tracking position error, the servo control unit, An error input unit configured to receive the error signal; A command input configured to receive a command signal indicative of a track to be followed; And And a control output connected to the actuator, wherein the servo controller is configured to provide a servo control signal in accordance with the error signal and the command signal. 4. The disk drive of claim 3, wherein the disk drive further comprises a plurality of disk surfaces and a plurality of converters, wherein the actuator arm comprises: Cancer part; And Each suspension assembly includes a plurality of suspension assemblies, supporting one transition against a corresponding disk surface, The actuator includes an approximate actuator coupled to the arm portion and a plurality of microactuators, wherein at least one of the plurality of microactuators is operatively connected to the arm portion and a corresponding converter to the arm portion relative to the arm portion. Move the corresponding converter, The servo control unit, A plurality of error inputs configured to receive error signals at the converters; And A plurality of control outputs configured to provide control signals to the plurality of micro actuators and the coarse actuator in accordance with the command signal and the error signals. 5. The servo of claim 4 wherein the servo control includes a plurality of relative position inputs configured to receive a plurality of relative position signals indicative of the relative position of the plurality of microactuators with respect to the coarse actuator. system. 6. The servo system of claim 5, further comprising a plurality of relative position sensors operatively coupled to the plurality of microactuators and providing the plurality of relative position signals. The servo system of claim 1, wherein the model-based servo control unit comprises a control unit synthesized in a model using H-∞ and μ control unit synthesis. 5. The apparatus of claim 4, wherein the operating parameters change according to variations caused by structural resonances in the disk drive, and the servo controller compensates for variations in the structural resonances of the disk drive and operating parameters of the disk drive. To adjust the servo control signal, and the operating parameters are: A servo system selected from the group consisting of positional input disturbances to the servo control and simultaneous movements of the plurality of microactuators. A rotatable disk having the disk surface having information disposed in tracks on the disk surface; A conversion unit; An actuator arm operatively connected to the converter; An actuator operatively connected to the actuator arm to move the converter with respect to the disc surface to follow one of the tracks during track following; And And a model-based servo control operatively connected to the actuator for controlling the movement of the actuator, wherein the model-based servo control provides robust track following control over a desired frequency range. Drive part. 10. The disk drive unit of claim 9, wherein the model-based servo control unit comprises a control unit synthesized in a model using H-∞ and μ control unit synthesis.