| Abstrakt: |
Abstract In this paper, we describe a new solution for driving the voice coil motor (VCM) in hard disk drives (HDDs), in which the usual current driver for the voice coil motor has been replaced by a voltage driver. When used in conjunction with PWM power stages, this solution becomes fully digital and has proven advantages in terms of cost, power dissipation and silicon area. To achieve the same performance of the current driver, the voltage driver requires a pre-filter placed at its input, to cancel out the electrical pole of the VCM. In the paper, we describe the voice coil model-based control (VCMC), a solution for implementing the pre-filter, which replicates the behavior of a current loop by using a model of the voice coil motor and its driver, including the saturation of the power stage. It is worth noticing that VCM resistance may change by 30% during HDD operations, increasing when the disk is performing repetitive seeks or decreasing when the disk cools down during track following. Such variation may lead to an unsatisfactory performance of the VCMC, so an adaptation mechanism, capable of tracking variations of VCM coil resistance, must be set up. This paper presents a pair of on-line estimation procedure, used to get the value of the VCM coil resistance during seek and track following, respectively. The first one is based on a standard LMS approach, aimed at identifying a simplified model of the plant with the least computational effort. The second exploits the presence of repetitive disturbances (RROs) to detect variations of plant gain and, in turn, of VCM resistance. It will be shown that both procedures are quite accurate in estimating variations of VCM resistance and their output can be used to perform on-line tuning of the VCMC. Experimental results show that the servo performance with the adaptive VCMC is not affected by resistance variation and equivalent to that of the standard current driver. |
Full text from SpringerLink