Search

We evaluate a testing specification proposed by the National Storage Industry Consortium's (NSIC) Extremely High Density Recording (EHDR) group for evaluating head and media compatibility for servo performance in magnetic disk drives. These tests use average amplitude and average noise profile measurements across isolated tracks to predict the shape, linearity, noise, and long-term stability of position error signal (PES) patterns. We compare the predictions from these tests to measurements from null and amplitude PES patterns written on a spin-stand. Results show average PES-profile prediction errors of 1%-2 % track width and noise level prediction within a factor of 2. We present data from tests for long-term stability of the magnetoresistive (MR) read element after repeated write cycles by the inductive write head. In the set of heads we tested, the MR head's center and effective width changed only slightly. Although we evaluated the NSIC specification for MR read elements, the specification should be equally valid for other read head types also, as long as the PES patterns are similar. Index Terms--Disk drive, position signal.

We present improved re-read strategies that try to recover data that would otherwise be lost. These strategies store the block samples at the detector input and average them with the samples obtained from succeeding re-reads. We evaluate the data reliability improvements of such schemes as a function of a weighting function that weights different reads with different importances, for both single and double retrials. Finally, we compare various schemes. Index Terms--Automatic repeat request, disk drives, error analysis, error correction coding, error detection coding, system recovery.

The permanent-magnet motor is often the most important element in hard disk drive (HDD) spindles and also a frequent source of vibration and acoustic noise. The eccentricity between stator and rotor is inevitably introduced during the manufacturing process, and includes mass unbalance, shaft bow, and bearing tolerances. This paper analytically discusses the effects of rotor eccentricity on motor performance for symmetric and asymmetric motors, such as local traction, unbalanced force, cogging torque, back electromotive force (EMF), phase current, and torque ripple. An asymmetric motor, mostly chosen to reduce the cogging torque, shows a worse effect on cogging torque and unbalanced force when the eccentricity exists. It also adversely affects the flux linkage and introduces variation of the phase back EMF depending on winding and rotating eccentricity, thus yielding increased mutual torque ripple. Index Terms--Cogging force, hard disk drive spindle motor, rotor eccentricity, torque ripple, unbalanced radial forces.

The time dependence of static friction is an important aspect of tribological studies whereby the static friction increases with an increase in the time of stationary contact (rest static friction or rest stiction). This effect is commonly observed in computer disk drives especially at high humidities, in which the static friction increases rapidly with an increase in rest time between some head slider and the disk surfaces. For the first time, a comprehensive kinetic meniscus model is proposed to explain this phenomenon, both for a single asperity and multiple asperity contacts at a liquid mediated interface. It is found that the static friction increases up to a certain equilibrium time with a power law relationship after which it remains constant. The equilibrium time is dependent on the liquid film thickness, the liquid viscosity and the contact geometry. The model developed is applied to a rough textured disk, a polished disk and a super smooth disk. To study the effect of the scan size on the meniscus force, the static friction is calculated by using roughness parameters obtained from measurements at different scan sizes.

Acoustic emission is used to study the contact behavior of subambient pressure tripad sliders during start/stop and constant speed operation. The contact force at the slider/disk interface is determined as a function of velocity and the dependence of contact force on flying height is investigated. The results indicate that contact forces for typical subambient pressure tri-pad sliders are on the order of a few mN.

This series of three papers presents an experimental and numerical analysis of the friction between a head slider and a thin-film rigid disk with circumferential surface texture, under various humidities. In Part I, we develop a model of a cylindrical contact of a slider rail with strong anisotropic asperities under dry conditions. This model predicts that the real contact pressure is primarily determined by the height distribution and geometry of asperities on the disk, whereas it is very weakly dependent on the surface curvature of the rail. The result of the latter calculation suggests that the frictional coefficient is insensitive to variation in the head contact angle due to head pitching motion.

A matrix of specially formulated greases composed of perfluorinated polyalkylether (PFPE) oil and telomers of polytetrafluoroethylene (PTFE) particulate thickener were tested. The PFPE greases were chosen for this evaluation because their oils have a very low vapor pressure and they will not volatilize in the disk drive. Base oil viscosity, PTFE particle size and percentage of oil content were varied. The rheological properties of complex viscosity, storage and loss moduli, and loss factor were measured. Percent oil was determined using thermal gravimetric analysis. The thickener's PTFE particle sizes were measured and their shapes were imaged using scanning electron microscopy. Bearing low speed torque and bearing noise tests were performed to evaluate the effect of the grease on bearing performance. The head settle track misregistration (TMR) was measured on disk drives to measure the effect of bearing lubrication on the servo performance. Generally, greases with high base oil viscosity had the lowest complex viscosity. Greases with highest viscosity PFPE oils had the highest torque and lowest noise. Greases with large PTFE particles had high loss factors and exhibited high torque and noise levels. High PTFE thickener to PFPE oil ratio (thicker grease) causes the bearings to have fluctuations in torque and noise levels. Actuator bearings lubricated with greases having higher viscosity oil had lower head settle TMR.

This series of three papers presents an experimental and numerical analysis of friction in a thin-film rigid disk with a circumferential surface texture, under various humidities. In Part II, we develop a theoretical formulation of the meniscus force of elliptical contact at isolated liquid bridges. This model predicts that the mean real pressure due to the meniscus force apparently decreases the net real contact pressure [p.sub.r] and increases the frictional coefficient. This effect becomes significant when the relative surface energy of the liquid 2[Gamma]/[sigma] is comparable to [p.sub.r], where [sigma] is the standard deviation of asperity height distribution. When the disk is covered with lubricant and the liquid bridges are connected with each other, the solid noncontacting asperities play an important role in the formation of the meniscus force, and the relative lubricant thickness t/[sigma] is the critical factor determining the friction.

The performance of a simple servo-derived clocking scheme for magnetic disk recording channels is presented. Using a theoretical model, the bit clock jitter is analyzed by computer simulation, and it is shown that in the bit clock jitter the mechanical component (velocity jitter) is more influential than the jitter component which originates from the position noise in the detected clock mark readback pulse. Cancellation error of an eccentricity jitter component between writing and reading moments due to PLL (Phase-Locked Loop) gain variation can be reduced by a feed-forward eccentricity suppression technique. Part of simulation results are verified by jitter measurements using an experimental recording channel, a patterned magnetic disk and an MR/Inductive head. Bit clock jitter of less than 2.5 ns(rms) is obtained for a spindle with air-bearing, supporting the analytical results. The simple servo-derived clocking scheme can be used in magnetic disk drives with a data rate of 20 Mbit/s or more.

Recently, equalization and detection techniques have been applied to the magnetic recording channel. For medium to high normalized recording densities (S [greater than or equal to] 2), decision feedback equalization (DFE) represents the only performant equalization method. Despite many proposed schemes, no specific analytical performance has been evaluated yet when applying DFE to the recording channel. In this paper, we derive both the zero forcing (ZF) and the mean square error (MSE) performance measures for DFE to the Lorentzian channel. Simple expressions have been found and allow us to determine from which recording density the MMSE-DFE significantly outperforms the ZF-DFE. These expressions have also revealed to be useful to evaluate the loss induced when combining coding and DFE for the Lorentzian channel.

Conner Peripherals' Conner Disk-Stor CP30254 hard disk drive outperforms products in the same price and feature range made by other manufacturers. It is very easy to install, comes with clear instructions and has excellent technical support from the company. It is fast at 14ms average access time with a 32K buffer and consistently performs as expected. Its manufacturer's list price is $329.

Disk scheduling algorithms are used to analyze the performance of a disk drive. Elementary calculus and probability techniques are applied to determine the mean response time for the scan algorithm and the first-come-first-served algorithm. A comparison of mean response times shows that the scan algorithm generally performs better than thefirst-come-first-served algorithm, while the latter performs better at low rates of request for disk access.