(A) study on the optimum design of a pick-up actuator using a time optimal technique

Abstract

The most important requirement in a design of optical pick-up actuator is minimization of seek time. Thus, an actuator is needed to make mass and size small. Also, an actuator with a large drive force and high resonant frequency is required for a shorter seek time. Researches to satisfy these short seek times and the characteristic values of a 2-axis actuator were carried out. Also, there was an optimum design of 2-axis actuator using Taguchi method. However, these design methods merely increased the drive acceleration of the actuator and only considered the performances for 2-axis actuator. In this thesis, we propose an optimal design method for a dual-stage actuator of VCM type and introduce the performances of the actuator. The performances of a dual-stage actuator are defined as the seek time, natural frequency, 5Hz sensitivity, 200Hz sensitivity and Q-value. Here, the seek time is related to the dual-stage actuator and the others are related to the 2-axis actuator. Because the performances within the ranges of design parameters are broadly varied, an optimal design method is needed in order to obtain the design parameters for the required performances. Here, SQP is used to obtain the optimal parameters. Because SQP is proper to the constrained optimal problem, the method is selected in our constrained optimal problems. We determine the cost functions and constraints of the optimum design. The cost function must be the minimum seek time for the dual-stage actuator. Thus, we use time optimal technique to achieve the minimum seek time for the actuator. According to this method, the design parameters must be coincided with the real values. However, because the real actuator cannot be made as the optimal design parameters, a robust design insensitive to the variations of parameters is needed. On the other hand, in the case of a 2-axis actuator, the performances become important factors for the dynamic properties and the properties of control. Thus, in order to...