Servo pattern enhancements for high areal density hard disk drives
Date of Issue2017-09-05
School of Electrical and Electronic Engineering
The growth of areal density (AD) in hard disk drive (HDD), which measures the data bits per unit area, has slowed down considerably in recent years from the average of 30% annual increase rate achieved in past decades. The main reason is that the current perpendicular magnetic recording (PMR) technology has reached its bottleneck of the so-called superparamagnetism. In the meantime, HDD is facing a strong competition and high market share challenge from solid state drive (SSD). As a result, HDD industry is striving to improve the AD with some new technologies. Heat-assisted magnetic recording (HAMR) is the most viable and promising technology to bring the AD beyond a few terabyte per square inch in the coming decade. When the AD is bigger than one terabyte per square inch, the track width is expected to be smaller than a few tens of nanometers. Being a mechanical device, ever decreasing track width has imposed more stringent requirements on the servo system for the better track-following performance of read and write operations. This thesis first focuses on writing good quality servo patterns such as burst field and repeatable run-out (RRO) compensation field under HAMR writing conditions. The HAMR writing mechanism is fundamentally different from the PMR. For PMR, the written-in signal quality is mainly determined by the strength and the gradient of the writing magnetic field. However, for HAMR, although the writing magnetic field is important, the dominating factor for recording quality is determined by the media thermal profile. The media thermal profile is in turn determined by the thermal properties of the media stack and the thermal power delivered to the media. Given the write head design, and especially the near field transducer (NFT), the thermal power delivered to the media can be controlled through input laser power and the spacing between the NFT and the media. Unfortunately, the NFT-to-media spacing (NMS) keeps changing due to the laser-induced heat sources, which makes the written-in signal quality control a much more complex task compared with that of PMR. In order to predict and control the written-in signal quality, the dynamic model for the HAMR writing process is constructed and identified. Using the HAMR writing process model, the NMS and input laser current can be optimized to achieve more consistent written-in signal quality. The focus of this thesis is to write high-quality RRO field and servo burst field from the first bit to the last bit, from the first servo sector to the last sector or even from the first track to the last track. However, this methodology applies to data writing as well. Next, the servo track-following performance is enhanced by improving the burst pattern non-linearity and RRO compensation scheme. The position error signal (PES), which is the only feedback signal for the closed-loop servo control to tell the head position in the radial direction, is demodulated from the burst pattern. The analysis will be performed on how to reduce the non-linearity of the written-in burst pattern to minimize the induced PES demodulation error and improve the head position accuracy. In addition, the multi-track blending (MTB) is invented to improve the RRO compensation performance at the off-track locations, which is very critical for the much narrower track width of future high areal density HDDs.