Reader Width Research Articles

Design Considerations Including Rotating Head Technique for Shingled Heat-Assisted Magnetic Recording

Predictions for Signal-to-Noise Ratio (SNR) and User Density (UD) are presented for a large variety of shingled Heat Assisted Magnetic Recording (HAMR) systems. Calculations are performed using a micromagnetic simulation that treats the heated media cells using renormalized spin blocks. Varied parameters include grain pitch, heat spot Full Width Half Maximum, reader width, and shield to shield spacing. Results are optimized with respect to bit length and grain pitch. Unlike previous work, the base condition feature systems already available experimentally (to allow for comparison), extending to media and heads that allow greatly extended capacities. We also discuss the benefits of a rotated read head for these geometries.

Track-following system optimization for future magnetic tape data storage

Recently, the increase of track density has been shown to be one of the most effective methods to increase the areal bit density and thus the cartridge capacity of future magnetic tape data storage systems. Increasing the track density poses stringent positioning requirements to the tape track-following control system. In this work, we investigate four novel timing-based servo patterns, and the impact of their geometrical properties, i.e. azimuthal angle and subframe length, on the position estimation resolution, the system delay and track-following performance. First, the resolution of the patterns versus reader width is assessed by analyzing the noise floor in the power spectrum of the lateral position estimation. This methodology is used to determine the optimal reader width for each pattern. Second, the system delay for each of the new patterns is identified by measuring the frequency response of the track-following actuator. An analytical model of the delay is derived and verified with experimental data. Finally, the impact of resolution and delay on the track-following performance is investigated. The best track-following performance, characterized by a standard deviation of the position-error signal of 4.2nm, was achieved using a pattern with a 36°azimuthal angle. The performance could be further improved down to 4.0nm by time averaging the position estimates.

Analytical Estimation of Transition Jitter for the Heat-Assisted Magnetic Recording Process

In the heat-assisted magnetic recording (HAMR) process, the transition jitter is known to depend on the grain size as well as the heat spot thermal gradient. We find that the jitter also depends on the Voronoi grain size distribution, as well as exhibiting a surprising nonlinear dependence on reader width. By combining the noise due to these dependencies we used an analytical formulation to estimate the jitter in the HAMR process. The accuracy of the analytical formulation is verified by comparing the predicted jitter to micromagnetic simulations. This simple formulation provides both physical insight and conserves computational time relative to lengthy (and complex) recording simulations.

User Areal Density Optimization for Conventional and 2-D Detectors/Decoders

The magnetic recording industry is perennially trying to grow the areal density by optimizing the head/media, the coding/signal processing, and the bit length (BL) and track pitch (TP) of the data written on the medium. Writing data at smaller BL or TP on the medium lead to loss of SNR during readback, which can be corrected by the coding and signal processing subsystem at the cost of the code rate $R$ . The optimization of the BL, TP, and code rate for a given head and media is the task of the system integrator. In this paper, we make use of an advanced channel model: the grain-flipping probability model, to perform an equivalent optimization together with a software recording channel that is able to include the impact of the generalized partial response equalizer, soft-output Viterbi algorithm detector, and low density parity check decoder on the final performance. In this paper, we vary the BL, TP, and code rate $R$ in the simulations to search for the highest user areal density achievable for a given head/medium setup. We investigate the effect of having a constant reader width (RW) as well as with RW proportionate to TP in our channel model. We also investigate the effect of more advanced 2-D coding and detection schemes on the final system performance.

Composite media for high density heat assisted magnetic recording

A heat assisted magnetic recording composite media with a superparamagnetic writing layer is proposed. The recording process is initiated in the write layer that is magnetically softer than the long term storage layer. Upon cooling, the composite structure copies the information from the writing layer to the lower Curie temperature (Tc) storage layer, e.g., doped FePt. The advantages include insensitivity to Tc variance in the storage layer, and thus the opportunity to significantly lower the FePt Tc without the resulting Tc distribution adversely affecting the performance. The composite structure has a small jitter within 0.1 nm of the grain size limit owing to the sharp transition width of the optimized superparamagnetic writing layer. The user density of the composite structure can reach 4.7 Tb/in.2 for a Gaussian heat spot with a full-width-at-half-maximum of 30 nm, a 12 nm reader width, and an optimized bit length of 6 nm.

Development of advanced barium ferrite tape media

We developed an advanced particulate magnetic tape using fine barium ferrite (BaFe) particles for magnetic-tape storage systems. The new tape showed a signal-to-noise ratio (SNR) that was 3.5dB higher than that of the commercially available BaFe tape used for the Linear Tape Open generation 6 tape-storage system, at a linear density of 300kfci measured with a giant magnetoresistive head with a reader width of 0.45µm. Such significant increase in SNR was achieved by reducing the magnetic particle volume from 1950 to 1350nm3, while maintaining a sufficiently high thermal stability, improving the perpendicular squareness ratio from 0.66 to 0.83, and improving the surface roughness from 2.5 to 2.0nm when measured by atomic force microscopy and from 2.4 to 0.9nm when measured by optical interferometry. This paper describes the characteristics of the new BaFe particles and media, which are expected to be employed for future high-capacity linear-tape systems.

A Study of TDMR Signal Processing Opportunities Based on Quasi-Micromagnetic Simulations

This paper presents the results of a comprehensive study of the shingled writing process and various signal processing and data detection approaches applied to the readback waveforms. The recording simulations include realistic head fields, a random granular media, magnetostatic and exchange interactions, and a READ head sensitivity function. Readback waveforms are examined in both one and two dimensions in terms of signal characteristics (linear and nonlinear), noise behavior (stationary and signal-dependent), and intertrack interference. Different equalization and detection approaches are compared and about a 10% density gain is reported for such 2-D magnetic recording compared with traditional single-track recording. These gains depend strongly on the number of readers, the reader positioning, and the reader width.

Additional Reader Noise in TDMR Reader for ITI Cancelation

Shingled writing eliminates the writer width limitation and enables further bit aspect ratio reduction to extend the areal density growth of current perpendicular magnetic recording. The achievable track density of shingled magnetic recording (SMR) is determined by the reader width rather than the writer width. However, continuously scaling of reader width increases the reader resistance and results in higher reader noise. It would be an issue to maintain enough signal-to-noise ratio of reader. The 2-D magnetic recording (TDMR) was proposed to extend the density growth by advanced signal processing technique. In TDMR configuration, there are multiple readers to read multiple tracks simultaneously. Recently, the TDMR reader was proposed as a solution of intertrack interference cancelation, which allows to use wider reader to reproduce signals from narrow tracks. In this paper, an additional reader noise is observed when the reader was positioned to read two tracks at the same time. The formation of the reader sensor appears to be closely related to the current reader structure with hard bias magnet. This issue highlights that a solution is required to enable TDMR, probably a reader structure without hard-biased.

Side-Reading Effects in High-Track-Density Tape Recording

Aggressive track-density scaling in linear magnetic tape recording has been identified as a key means of sustaining capacity scaling. As the reader width decreases, the relative signal contribution of the written track region outside the lateral extent of the reader increases. We model, analyze and quantify side-reading effects of shielded magneto-resistive read heads reading narrow tracks on perpendicular-oriented tape media. We compare results from extended analytical models with experimental data. An understanding of these effects will be important in designing track misregistration budgets and SNR margins for future operating points of parallel-channel tape recording systems.

Ideas for Detection in Two-Dimensional Magnetic Recording Systems

In two-dimensional magnetic recording, detection for a perfect writer and perfect reader based on an iterative threshold technique are compared to detection employing single threshold and adaptive threshold techniques. The simulation indicates that the algorithm can reach a BER of 9.3% compared to 11.2% and 10.5% for the latter cases when the bit size is 8 by 8 nm with a grain of 8.7 nm diameter. This increases data capacity by 11% to 6.6 TBits/inch2. Additionally, detection with an array of three conventional heads has been studied: the simulation shows that for bits with dimension of 12 × 12 nm, grains of 8.7 nm diameter and reader width of 18 nm, the BER drops from 11.3% for a single conventional head to 5.4% for the head array with a two-dimensional generalized partial response equalizer at the optimum offset of 10 nm. The density is increased by 42%, reaching 4.0 Tbits/Inch2 .

Characterization of the Reader Width Using the Micro-Track Test in Perpendicular Recording

Micro-track test is a valuable tool used to characterize the electrical reader width in perpendicular recording. The test is based on trimming both sides of a track written with a single tone and followed by a cross-track scan of the remaining track. The reader cross-track profile or read sensitivity function is characterized by a narrow band measurement of the read-back signal as a function of cross-track position. With increasing track density, the signal amplitude decreases due to incomplete media saturation and to an increase in the off-track noise due to percolation and opposite polarity media saturation. These effects cause a lump at the edge of the micro-track profile that changes the width of the profile (MT10 and MT50). Micro-magnetic analysis indicates that the off-track demagnetization field creates a track-edge section in the media with opposite polarity relative to the originally written track. This section is coherent with the main track and decreases the read-back signal during the scan of the micro-track. Consequently, the shape of the cross-track read sensitivity function becomes distorted, causing a reduction in MT10 and a side lump. The solution to this problem is to increase the amplitude of the micro-track read-back signal by reduced trimming and by the use of a higher density (KFCI) in the main track that minimizes the off-track demagnetization field. These test improvements are shown to preserve the integrity of the micro-track profile and to enable an accurate measurement of MT10 and MT50. Optimizing reader and writer widths in perpendicular recording requires an accurate measurement of these parameters.

Transition Curvature and Transition Broadening Analysis Using the Spin-Stand Imaging Technique

This paper presents an improved methodology of estimating transition curvature and transition broadening in perpendicular magnetic recording, using a spin-stand imaging technique. Due to the finite read resolution widths in the cross-track direction, the curvature and broadening of the measured transition will deviate from that in the magnetization. It is demonstrated that the reader width effects can be substantially eliminated by performing a specially designed trimming process which allows the reader to only see the written transitions from a small fraction of the track. Simulations illustrate that the trimming process effectively produces a narrow boxcar reader. This paper compares the results of this method for two heads with similar reader width but different writer width. It is shown that the trimming process has a stronger impact on the head with a larger ratio of reader width to writer width.

Jitter in a Voronoi Pattern Media-Effect of Grain Size Distribution and Reader Width

We calculate jitter of the signal from transitions written on Voronoi-pattern media and played back with readers of varying widths. The expected dependence of jitter on reader width, as predicted by the microtrack model, is verified down to very narrow readers. Decrease in grain size variance results in less jitter. The original Voronoi patterns give a jitter value very close to the microtrack model. However, when the Voronoi patterns are modified such that grain size distribution is reduced, the jitter can be significantly smaller than in the microtrack model. This is due to the reduced randomness in the distribution of grain centers.

Adjacent Track Interference Due to Side Reading and BER Performance

We investigate the effect of the adjacent track interference (ATI) due to side reading on the bit-error rate (BER), using an E/sup 2/PR4 recording channel model and Monte Carlo simulation. We focus on how the reader width (RW), writer width (WW), and microtrack profile shape of the reader affect the ATI and the BER. In the Monte Carlo simulation, the RW, the WW, PW/sub 50/, and amplitude of the head are randomly selected from distributions. For each head, the microtrack profile is randomly selected from a collection of measured profiles. The ATI has strong correlation to the RW/WW ratio, but interestingly, the latter does not correlate to the BER. Further, the tail area of the microtrack profile has strong correlation to both the BER and the ATI.

Tunneling Magnetoresistive Heads Beyond 150>tex<$hboxGb/in^2$>/tex<

Tunneling magnetoresistive (TMR) readers capable of 150 Gb/in/sup 2/ of areal density magnetic recording for hard disk drive were demonstrated with bit-error-rate performance. The head design used is basically a bottom type stack of Ta/PtMn/CoFe/Ru/CoFe/oxide barrier/CoFe/NiFe/Ta cap with abutted hard bias stabilization. The electrical reader width is about 4 /spl mu/'' to reach a very high track density and shield-to-shield spacing is about 700 /spl Aring/ for high linear density. On-track bit error floor is better than 10/sup -5/ at a linear density of 900 KBPI and the recording system noise is dominated by the media. The best areal density achieved (using -4 OTC reference level) is 143 Gb/in/sup 2/ using symmetric squeeze and 152 Gb/in/sup 2/ using asymmetric squeeze method, respectively. It was found that the TMR head has several decibels more signal-to-noise ratio gain over spin valve readers at 150 Gb/in/sup 2/ and beyond. The TMR head is also suitable for perpendicular recording application.

Spin-valve heads with self-stabilized free layer by antiferromagnet

We demonstrate the feasibility of a spin-valve read head using an adjacent antiferromagnetic (AFM) layer to provide a stabilization Held, eliminating the hard bias permanent magnet layer. The head stack consists of seed/AFM/CoFe/Ru/CoFe/Cu/CoFe (and/or NiFe)/Cu/IrMn/cap where the AFM is PtMn or IrMn. The abutted junction is about 0.1 /spl mu/m wide at the free layer and no permanent magnet materials are used. Micromagnetic modeling indicates that a proper bias field is needed to have a stable head response. The bias field from the AFM layer can be modified by a nonmagnetic spacer layer between AFM and FM. The coupling field decays exponentially with the thickness of the spacer layer, consistent with a pinhole-dominated mechanism. Unshielded sensors show much higher sensitivity as,compared with a standard PM abutted junction head. Microtrack profile based on the spin-stand test shows a regular shape without apparent lumps. The electrical reader width is about 6 /spl mu/" (0.15 /spl mu/m), corresponding to a track density of 112 ktpi. Combined with a linear density of 591 kbpi the heads in this work are capable of areal density around 66 Gb/in/sup 2/.

Finite element and Monte Carlo modeling of the electrical and thermal properties of high areal density MR read-back heads

The current density distribution in magnetoresistive read-back heads with abutted permanent magnets was studied by using finite element modeling. The parasitic junction resistance was calculated as a function of reader width and stripe height. The Joule heat, which was generated by the local current density, was subsequently used as a heat source for a subsequent calculation of the local temperature rise in the active sensor area. The effect of the dispersion in reader width and stripe height on the average temperature rise was studied by Monte Carlo simulations. The thermal modeling predictions were in agreement with the experimental data from actual read-back heads. The thermal modeling was further used to study the trends in the bias current and voltage as a function of areal density.

Measurements of media jitter down to 0.2 μm track widths

The design of high areal density recording systems requires a detailed understanding of how the signal-to-noise ratio depends on the system geometry. We have made measurements using various head-media combinations of the variation of the media jitter with recorded track width. An analogy is made between the width of these recorded micro tracks and an equivalent reader width. The results are then compared with the recording theory expectation that the jitter varies as the inverse square root of the read track width. For the two head-media combinations discussed in this article, the jitter exponents are found to be −0.34 and −0.43, respectively. The deviation from a −0.5 exponent can be correlated to nonuniformities in the micro-track profiles of the heads used in the measurement.