This paper considers the modeling and compensator design for Self-ServoWriting (SSW) process in disk drives. An Iterative Learning Control (ILC) based scheme is established to deal with radial error propagation and improve the quality of written tracks. In the proposed scheme, a feedback controller for track following is first designed to achieve good disturbance attenuation. Then, an ILC structure is applied to generate an external signal, which is injected into the feedback loop in order to compensate for the written-in errors in the previous track while the next track is written. As a result, the error propagation can be contained. The learning controller is synthesized by solving Linear Matrix Inequality (LMI) equations to ensure the stability and monotonic convergence of the control algorithm. Simulation results show the effectiveness of the proposed scheme on the error containment which results in good quality written tracks.
In the sector servo system for Hard Disk Drives (HDDs), a circular disk is divided into equally sized angular pieces, called servo sectors. Laid on the boundaries of the servo sectors are servo fields on which special servo patterns are embedded. Whenever the Read/Write head (R/W head) passes over the servo patterns, a waveform is read back and decoded to generate a position error signal (PES) which indicates the deviation of the R/W head from the center of the track. PES is then utilized for the control of the head position. With the continuously increasing track density, the required accuracy on the placement of servo patterns is proportionally increased and becomes a crucial issue.
Servowriting is the process to write servo patterns onto the disk. The goal is to place the servo patterns in centric circles with minimum variation of track spacing for all tracks. Conventionally, the process is accomplished by a high precision device called a servowriter, which uses external positional references to accurately position the heads in the disk drive for writing the servo patterns. During the process, in order for the device to access the heads, the drive cover has to be removed. Hence, a clean-room environment is required to servowrite a disk drive. With nowadays high density drives, the servowriting process is extremely time consuming. Consequently, long hours spent on expensive servowriters in clean-room space result in substantial manufacturing cost. Recently, a technique called Self-ServoWriting (SSW) has been developed in order to cut down the cost.
For one type of SSW process [1], the basic concept is to propagate the servo patterns from a few seed tracks by the drives itself. Only the seed tracks are written by an external device. Thereafter, the majority portion of the servowriting process is done after the drive is assembled. The time spent in the clean-room and thus the manufacturing cost can be dramatically reduced. However, in SSW, an absolute position reference is not available. Instead, a previously written track is used as a relative position reference. As a result, the written-in error may be re-produced and even amplified when successive tracks are written. The phenomenon is called radial error propagation, the avoidance of which is the major challenge in SSW. Therefore, the motivation of this research is to prevent radial error propagation and improve the quality of tracks in terms of written-in errors.
In control theories, Iterative Learning Control (ILC) and Repetitive Control (RC) both have the capability in learning through iterations to reject periodic disturbances. The major difference between them is the setting of the initial conditions. ILC is intended for discontinues operation, where the initial conditions are reset in each iteration. On the other hand, RC aims for continuous operation, where the initial conditions are set to the final conditions of the previous iteration [5]. In SSW process, after one track is written, the read head seeks to the written track for the write head to start servowriting the next track. Therefore, ILC is suitable for the application because of the reset natural of the process. It can be an effective approach to mitigate radial error propagation. Some researchers (see [2] and [3]) have formulated the problem into a similar manner and provided heuristic design methods. In this paper, we propose an ILC based scheme and establish a systematic framework for the compensator design. The learning controller in the ILC structure is synthesized by solving Linear Matrix Inequalities (LMIs) to ensure the stability and monotonic convergence of the control algorithm. The disturbance and noise rejection is also taken into consideration. The remainder of this paper is organized as follows. In Section 2, the modeling of the SSW process and the root cause of radial error propagation are briefly reviewed. Then Section 3 describes the proposed ILC scheme. LMI conditions are derived for controller synthesis. In Section 4, the proposed method is applied to a HDD benchmark problem. Finally, conclusions are given in Section 5.
This paper discussed the Self-ServoWriting process in HDDs. The major challenge is to prevent the written-in errors from propagating across the radial direction in the process. An ILC structure was applied to mitigate the error propagation effect. The disturbance and noise rejection was also considered in the design framework. The problem was transformed to the lifted domain and formulated as a set of LMI equations. The proposed LMI-based synthesis scheme provides a systematic design approach. Simulation results have demonstrated that the radial error propagation was contained which results in good quality written tracks with small track distortion and track squeeze.
