KR20020041011A - Process for designing an optimal vibration isolation mount for a disc drive - Google Patents

Abstract

Optimal vibration mounts for use in the disk drive 100 are designed by calculating external and internal disturbance models Ξ and Θ for the disc drive and defining the inertia matrix M for the disc drive. State variable estimators, such as Kalman filters, are defined based on the inertia matrix and internal and external disturbance models, and the covariance matrix Σ is derived based on the filter logarithmetic equation. State variable estimator gain (H) is And the optimal mount damping (B) and stiffness (K) parameters are derived from the state variable estimator gain and inertia matrix H = (a).

Description

PROCESS FOR DESIGNING AN OPTIMAL VIBRATION ISOLATION MOUNT FOR A DISC DRIVE

There are two types of disturbances that affect the operation of disk drives: internal disturbances and external disturbances. Examples of internal disturbances include air pressure encountered by the head / arm assembly due to the rotating disk, noise in the position error signal (PES), structural resonance of the disk drive, actuators (including voice coil motors, actuator arms and suspensions) Movement of the assembly, instability of the disk pack assembly (including the disk pack spindle and drive motor), recording runout in the servo pattern due to interference during recording of the servo pattern, unmodeled motive forces and nonlinearities such as pivot bearing friction, etc. There is this. External disturbances include environmental vibrations (earthquake vibrations) and external reactions to internal disturbances. Moreover, environmental vibrations during servo recording adversely affect the servo pattern arrangement, resulting in write runout in the servo pattern recorded on the disc.

If the disk drive is disturbed, it can affect the radial position of the read / write head relative to the disk. In conclusion, this disturbance adversely affects head / disk tracking. For this reason, the head and data tracks on the disc have a finite width large enough to ensure that the expected movement due to interference does not substantially affect the performance of the disc drive. In conclusion, the effect of interference is a limited factor on the data density of the largest area of the disk drive. Passive damping mounts are used to damp these disturbances, but they are not entirely successful due to improper compromise if there is a conflict between conflicting demands for internal and external disturbance attenuation.

The present invention addresses the above problems and the like, and provides other advantages over the prior art by providing a process for defining the optimal damping characteristics of passive damping mounts used in disk drives to minimize the adverse effects of internal and external disturbances. do.

The present invention relates to disk drives for storing information, and more particularly to a process for identifying the characteristics of an optimal passive damping mount for a disk drive in order to minimize interference. The invention is also applicable to identifying mount parameters for a servo track recorder to minimize written-in-runout.

1 is a perspective view of a disk drive capable of implementing the features of the present invention.

2 is a diagram of a spring damping model illustrating the principles of the present invention.

3 is a graph showing the basic tradeoff of passive vibration prevention.

4 is a flow chart showing the steps used to describe an optimal manual mount system for use with a disk drive.

5 and 6 are graphs showing power spectral densities of first-order Markov process representations of voice coil motors and external earthquake disturbance torques useful in describing features of the present invention.

7 and 8 illustrate surface plots and control plots of a disk drive monitor, respectively, as a function of mount parameters useful for explaining certain features of the present invention.

The present invention relates to a process for designing an optimal vibration mount for use in a disk drive. Internal and external disturbance models are calculated for the disk drive and the internal matrix for the disk drive is defined. The state variable estimator is defined based on an internal matrix and external and internal disturbance models, and the state variable estimator minimizes a limited standard, such as a 2-norm of state variable estimation error. The gain of the state variable estimator is calculated as a solution to the filter algebraic Riccati equation, and the optimal mount and damping parameters are derived based on the gain of the calculated state variable estimator. In an embodiment of the invention, the gain of the state variable estimator is Calculated by finding the covariance matrix, Σ, as a uniquely symmetric, at least positive, approximate finite solution to the filter logarithmical equation (FARE), where M is the inertial matrix and Θ is the inertial disturbance Is the intensity matrix and Ξ is the external disturbance intensity matrix.

In another embodiment of the invention, the state variable estimator is Is a Kalman filter from which the gain H is calculated. Optical mount damping parameter (B) and optical mount stiffness parameter (K) Derived from.

Although the present invention will be described in connection with a magnetic disk drive, the principles of the present invention can be applied to an optical disk drive, a servo track recorder and a spin stand. Thus, while the present invention is described in terms of determining optimal damping for a magnetic disk drive, the present invention can also be applied to determine optimal damping for use in servo track recorders, spin stands and optical disk drives.

1 is a perspective view of a disk drive 100 useful in the present invention. Disk drive 100 includes a housing having a base 102 and a top cover (not shown). The disk drive 100 further includes a disk pack 106 mounted on a spindle motor (not shown) by the disk clamp 108. The disc pack 106 includes a plurality of individual discs mounted on the spindle 109 mounted to rotate simultaneously about the spindle center axis. Each disk surface has an associated disk head-slider 110 mounted to the disk drive 100 to communicate with the disk surface that is facing. The head slider 110 is a conversion head arranged to record data to and read data from a slider structure arranged to fly over an associated disk surface of an individual disk of the disk pack 106 and concentric tracks on the opposing disk surface. It includes. In the example shown in FIG. 1, the head slider 110 is supported by a suspension 112 that is in turn attached to a track that accesses the arm 114 of the actuator 116. The actuator 116 is driven by a voice coil motor (VCM) 118 for rotating the actuator and the head 100 attached to the actuator about the pivot axis 120. Rotation of the actuator 116 moves the head along the motion path 122 to position the head over a predetermined data track between the disc inner diameter 124 and the disc outer diameter 126. Voice coil motor (VCM) 118 is driven by a servo device included on circuit board 130 based on signals generated by head-slider 110 and a host computer (not shown). The reading and writing apparatus also provides a signal to the host computer based on the data read from the disk pack 106 by the head-slider 110 and writes to the recording head of the head-slider 110 for writing data to the disk. It is included in the circuit board 130 to provide a write signal.

The passive vibration mount 132 supports the base 102 of the disk drive 100. Mount 132 isolates the drive from seismic vibrations, indicated by arrow 134. As the stiffness of the mount 132 decreases, a better reduction in seismic vibrations can be obtained. For example, the servo track recorder and spin stand may be blocked by a pneumatic mount with a vibration free natural frequency as low as 2 Hz. This soft mount prevents external vibrations due to seismic disturbances at significantly lower frequencies. However, the soft preventive mount may be subjected to internally generated disturbances such as reaction force 136 (VCM torque) generated by voice coil motor 118 and reaction force 138 generated by imbalance of spindle 109. Reduces error movement caused by Typically the error movement due to this disturbance is reduced by the hard mount 132. Thus, the external and internal vibration sources have a conflicting effect on the error movement of the disc drive such that attenuation of one source by manual mounting results in amplification of the other source.

2 is a spring damper model showing the principle of blocking a mass M such as a disk drive 100. External disturbances 134, such as earthquake movements collectively represented by x ₀ , are applied to the platform P supporting the mass M, and internal disturbances 136, 138 collectively represented by f _d are the masses M Acts directly). The six-dimensional disturbance vector f _d models internal disturbances such as disk pack imbalance 138 and VCM torque 136. The six-dimensional external disturbance vector (x ₀ ) models the seismic vibration 134 which is susceptible to the disk drive. The equation of motion for a rigid body part supported by a decentralized prevention device Where x is a six-dimensional vector containing six rigid body degrees of freedom, M is a mass / inertia matrix, B is a damping matrix, K is a stiffness matrix, x ₀ is a six-dimensional vector of external disturbances, and f _d Is a six-dimensional vector of internal disturbances. Through the Laplace transform and other mathematical calculations, the transfer function ( ) In harmony with feedback control terms, P (s) is a sensitivity function and Q (s) is a complementary sensitivity function. These transfer function matrices add up to the identity matrix (I), i.e. Becomes This relationship explains the basic tradeoffs, and full attenuation of internal and external disturbances is not possible. 3 is a board plot of P (s) and Q (s) for signal degrees of freedom, indicating that enhanced attenuation of external disturbances occurs using attenuation of internal disturbances and vice versa. This is a basic tradeoff due to passive vibration protection. The present invention is directed to a process for optimizing tradeoff and damping passive damping mounts to identify optimal stiffness.

The present invention uses a state variable estimator, such as a Kalman filter, which is an optimal state variable estimator used to minimize the two-standard of state variable estimation error for a system susceptible to process disturbances and sensor noise. When these disturbances and noise sources act as white noise processors, the Kalman filter provides an optimal solution to the state variable estimator problem. The present invention addresses the design of passive mount parameters (eg, stiffness and damping) as a Kalman filtering problem where internal and external disturbances can be modeled as white noise processes. Optimal manual mount integration for disk drives It is cast in the form of general Kalman filtering by representing the plant (disk drive 100) as a double integrator of the form, where M is the inertia matrix of the disk drive mass and f _d (t) and x ₀ (t) are independent Gaussian Is assumed to be a zero-averaged white noise process. If equation (4) is expressed as a general Kalman filtering problem, , Where A is the matrix Where L is the matrix Where C is the matrix to be. Thus, the condition of the Kalman filter where [A, L] can be stabilized (controllable) and [A, C] is detectable (observable) is satisfied.

The covariance of the external disturbance is the product of the external disturbance intensity matrix (Ξ) and the dirac delta function (δ (t-τ), and the covariance of the internal disturbance is the internal disturbance intensity matrix (Θ) and the dirac delta function (δ (t-τ) Is the enemy of)):

In the Kalman filtering problem, the cost function J which is the sum of the state variable estimation error changes is minimized.

Applying the Kalman filtering algorithm to the problem of determining the optimal mount parameters for the disk drive minimizes the error moved state variable vector (position and speed of the mass of the disk drive 100).

If Σ represents a fixed covariance of the state variable motion error, i.e. In this case, the cost function J from equation (7) can be expressed as J = tr [Σ].

Kalmaen filter state variables for the motion error affects the external disturbance (x ₀₎ internal interference (f _d) Easy system ( Estimate the minimum 2-standard of Karlman filter equation Where H is the gain of the Kalman filter to be.

The covariance matrix Σ is obtained as a uniquely symmetric, at least positive value, approximately defined solution to the filter logarithmetic Carty equation (FARE).

For the passive mount integration problem, Calman Gain (H) Where B and K are the optimal mount damping and stiffness matrices.

The process of the present invention is executed in accordance with the flowchart of FIG. The inertia matrix M of the disk drive 100 is obtained in step 200. In step 202, the internal disturbance x ₀ is modeled as white noise with an intensity matrix Θ, and the external disturbance f _d is modeled as independent white noise with an intensity matrix Θ, where Ξ And Θ is Is, to be.

In step 204, the Calman filtering problem Is defined by the in-filter equation (4). The covariance matrix (Σ) for the Kalman filter Is calculated in step 206. The gain (H) of the Kalman filter is expressed by equation (9) Is calculated at step 208 using the values of Σ and Θ defined at steps 202 and 206 in the form of. In step 210, equation (1) is solved for the values of mount damping matrix (B) and stiffness matrix (K), thereby defining an optimal anti-vibration mount system for the disk drive.

Yes

The process of the present invention is applied to the x15 Cheetah ^® 9LP disk drive model from Seagate Technologies. The rotational inertia of the disk drive is measured at 2.47 gm-in ² . Internal disturbances caused by movement of voice coil motor and spindle imbalance are modeled using a white noise process. External disturbances are estimated using the International Organization for Standardization (ISO) for seismic vibrations and other movements of computer devices.

Internal disturbances due to the imbalance of the spindle 109 (FIG. 1) correspond to a single frequency with high harmonics and can be modeled directly from drive design and performance. While the movement of the voice coil motor 118 excites a wider range of frequencies, for the purposes of the present invention only excitations resulting from the voice coil movement are considered. The profile of the torque disturbance generated by the voice coil motor 118 is It is modeled to be a first-order Markov process given by. Moreover, the power spectral density (PSD) of the voice coil torque profile is determined during a particular search and a first-order Markov process is obtained that closely matches the PSD profile. 5 illustrates this. It has been found that the Markov parameters β = 200 Hz and σ = 10 nm are quite suitable for the PSD of the search profile.

External disturbances (i.e. seismic vibrations and movement of other components in the chassis) are modeled by proximity to the ISO specification for computer devices. The power spectral density of the first-order Markov process corresponds to the power spectral density plot of the external disturbance as shown in FIG. The Markov parameters β = 10Hz and σ = 1500e-9rads are chosen to be fairly compatible with ISO standards.

For this example, the frequency form of the power spectral density of the voice coil motor torque disturbance and external vibrations are ignored. Indeed, frequency shaping can be considered or ignored using a shaping filter for the white noise process in the derivation of the Kalman filter equation. Ignoring the frequency shaping of the power spectral density representation for this example does not result in a noticeable loss in precision because the case of not neglecting and ignoring frequency shaping has a similar frequency profile (band-limited white noise).

Kalmaen gain is calculated (step 208, Fig. 4) for a disk drive, the results are derived by the damping of the disk drive in the mass standard form, i.e. the natural frequency and the rotation θ _z direction. Fig. 7 shows a change surface plot of the movement of the disc drive in the θ _z direction and shows the convexity of the optimization problem. 8 shows a contour plot in which the best mount parameters can be identified. From Fig. 8, the center of the slope is the optimum natural frequency of 325 KHz and 0.707 (1 / It can be seen that it limits the optimal damping. Optimal damping is particularly interesting because it is common for the Kalman filtering process that the pole is present in the Butterworth pattern in the complex S-plane.

The contour diagram of FIG. 8 explains that damping plays a prominent role in minimizing error movement due to rotational vibration. Damping can be ignored, but a disk drive that is rigidly mounted to the chassis generates high natural frequencies that cause problems with track registration. The operating point for this disc is at the upper left corner of Figure 8, where the slope is quite high, as indicated by the adjacent lines in the contour diagram. An efficient damping mechanism for obtaining the high damping rate defined by the present invention is likely to include nonlinear damping principles such as Coulomb friction.

Thus, the present invention can provide a process for designing an optimal vibration mount 132 for the disk drive 100. In step 202, the model is calculated as the external disturbance 134 of the disk drive, and the model Σ is calculated as the internal disturbances 136 and 138 in the disk drive. In step 200, the internal matrix M is defined for the disk drive. In step 204, the state variable estimator is defined based on the inner matrix M and the external and internal disturbance matrix models Ξ and Θ. The state variable estimator is selected to minimize the two-standard state variable estimator error. In step 208, the gain of the state variable estimator is calculated (step 206) based on the zero matrix error. In step 210, the optimal mount damping (B) and stiffness (K) parameters are derived from the calculated gain and inertia matrix.

In an embodiment, the gain H of the state variable estimator is a filter logarithmic Riccati equation (FARE) equation. Is derived first by calculating the covariance matrix (Σ) at step 206, based on the solution, and at step 208 to the covariance matrix (Σ), the external disturbance matrix model (Ξ) and the internal disturbance matrix model (Θ). Calculate the estimated state variable estimator gain. The optimal mount damping (B) and stiffness (K) parameters are related in step 210. Derived from. In another embodiment, the state variable estimator is a Kalman filter.

Although the present invention has been described in terms of designing a passive antivibration mount system for a magnetic disk drive, those skilled in the art will appreciate that a passive antivibration mount system may be implemented in other environments where such a requirement is required. More specifically, the present invention can be used to design optimal damping and stiffness parameters of an optical disc drive. Moreover, although the present invention has been described with reference to a disk drive having a plurality of disk surfaces, the present invention can be implemented where a single disk causes the internal disturbance.

While the various features and advantages of the embodiments have been described above in conjunction with the detailed description of the functionality of the various embodiments of the invention, these are for illustration only and modifications of the construction and arrangement of the invention are possible within the scope of the appended claims. For example, the filter algebraic Riccati equation (FARE) described above may have additional terms for minimizing different standards. Other estimators other than Kalman filters may be used for the determination of the optimal mount based on the particular application without departing from the scope and spirit of the present invention. Other changes may be made without departing from the scope and spirit of the invention.

Claims (8)

The process of designing the optimal vibration mount for a disk drive, a) calculating an external disturbance model for the disk drive; b) calculating an internal disturbance model for the disk drive; c) defining an inertia matrix for the disk drive; d) defining a state variable estimator based on the inertia matrix and the external and internal disturbance models to minimize a defined standard of state variable estimation error; e) calculating gain of said state variable estimator as a solution to a filter logarithmetic equation; And f) defining an optimal mount damping and stiffness parameter based on the calculated state variable estimator gain. The method of claim 1, wherein step (e) e1) calculating a covariance matrix based on said solution to said filter logarithmetic equation; And e2) calculating the state variable estimator gain based on the covariance matrix. The method of claim 1, wherein step (e) (e1) Calculate a covariance matrix (Σ) from a solution of a filter logarithmic equation of the form where M is the inertial matrix, Θ is the internal disturbance matrix and Ð the external disturbance matrix, and (e2) And calculating the state variable estimator gain (H). The method of claim 3, wherein step (f) comprises: f1) for B and K , And f2) a process characterized by setting the optimal mount damping matrix to B and setting the optimal mount stiffness matrix to K. The process of claim 1 wherein the state variable estimator is a Kalman filter. The method of claim 5, wherein step (e) e1) calculating a covariance matrix based on the solution to the filter logarithmic Riccati equation; And e2) calculating said Kalman filter based on said covariance matrix and said inertia matrix. The method of claim 6, wherein step (e) (e1) Calculate a covariance matrix (Σ) from a solution of a filter logarithmic equation of the form where M is the inertial matrix, Θ is the internal disturbance matrix and Ð the external disturbance matrix, and (e2) And calculating the Kalman filter gain H from. The method of claim 7, wherein the step (f), f1) for B and K , And f2) a process characterized by setting the optimal mount damping matrix to B and setting the optimal mount stiffness matrix to K.