Two-dimensional Magnetic Recording Systems Research Articles

ASIC Implementation of Nonlinear CNN-Based Data Detector for TDMR System in 28 nm CMOS at 200 Mbits/s Throughput

In this work, we present the first application-specific integrated circuit (ASIC) implementation of a convolutional neural network (CNN)-based data detection channel for the two-dimensional magnetic recording (TDMR) system in hard disk drive (HDD). This chip is fabricated in TSMC 28 nm complementary metal–oxide–semiconductor (CMOS). TDMR signals captured from a real HDD are used to train and test the network, and the chip is designed to make inferences only. This work presents a systematic pipeline to codesign the neural network structure and the corresponding hardware implementation. To meet the high-throughput requirement, the CNN is fully unrolled with dedicated silicon for each convolution layer using systolic arrays. Quantization-aware training algorithms are used to help improve power efficiency without comprising accuracy. Overall, our chip achieves 200 Mbits/s throughput at 0.9 V supply voltage. In addition, it achieves improved detection accuracy compared to the physical-model-based conventional channel. Our chip consumes 173–219 mW power at 200 M3 clock rate for supply voltage 0.8–0.9 V, respectively.

CNN-Based Machine Learning Channel on TDMR Drive Data

In this article, we demonstrate a convolutional neural network (CNN)-based data detection channel on real data from a commercial hard disk drive (HDD) employed with two-dimensional magnetic recording (TDMR) dual-reader technology. The neural network has two sets of input neurons and is trained using 485 4-kB sectors from the drive. Data detection is performed using the trained network and a one-to-one comparison of bit-error-rate (BER) performance with the HDD channel is systematically conducted. The trained neural network demonstrates comparable or slightly better performance than the state-of-the-art HDD channel. In addition, comparing to detection with a single reader, CNN-machine learning (ML) channel autonomously enables ~6% recording density gain when the second reader is available, and there is no physical modeling needed to manually combine or “fuse” signals from the multiple readers. We believe that the understanding elucidated in this article is crucial for developing more capable, robust, and effective ML-based data detection channels with the potential to realize even greater areal density gain in TDMR systems and higher capacity HDD.

Non-Binary Constrained Codes for Two-Dimensional Magnetic Recording

The two-dimensional magnetic recording (TDMR) technology promises storage densities of $10$ terabits per square inch. However, when tracks are squeezed together, a bit stored in the two-dimensional (TD) grid suffers inter-symbol interference (ISI) from adjacent bits in the same track, and inter-track interference (ITI) from nearby bits in the adjacent tracks. A bit is highly likely to be read incorrectly if it is isolated in the middle of a $3 \times3$ square; surrounded by its complements, horizontally and vertically. We improve the reliability of TDMR systems by designing two-dimensional constrained codes that prevent these square isolation patterns. We exploit the way TD read heads operate to design our codes, and we focus on TD read heads that collect signals from three adjacent tracks. We represent the two-dimensional square isolation constraint as a one-dimensional constraint on an alphabet of eight non-binary symbols. We use this new representation to construct a non-binary lexicographically-ordered constrained code where one third of the information bits are unconstrained. Our TD constraint codes are capacity-achieving, and the data protection is achieved with redundancy less than $3\%$ and at modest complexity.

Performance Analysis and Optimization of Spatially Coupled Protograph-Based Low-Density Parity-Check Codes for Two-Dimensional Magnetic Recording Systems

As a class of convolutional-based low-density parity-check (LDPC) codes, spatially coupled protograph (SC-P) codes not only can exhibit outstanding error performance but also can realize low-complexity implementation. Motivated by the prominent superiorities, SC-P codes have been intensively investigated in various communication environments. However, no research works regarding SC-P codes in the magnetic recording (MR) scenarios have been done so far. To explore such an issue mentioned previously, this article carries out a comprehensive study on the SC-P codes in the two-dimensional magnetic recording (TDMR) systems. Specifically, a turbo-like extrinsic information transfer (EXIT) algorithm is conceived for the SC-P-coded TDMR systems. With the help of such an algorithm, we evaluate the convergence performance (i.e., decoding thresholds) for conventional SC-P codes over TDMR channels with turbo equalization and illustrate that such codes should be re-constructed to attain better performance. Therefore, we develop an optimization scheme to generate a novel class of SC-P codes, called enhanced SC-P codes, which can exhibit promising decoding thresholds without losing the linear-minimum-distance superiority. The convergence-performance analyses and simulated results reveal that such enhanced SC-P codes can obtain more desirable coding gains in contrast to the conventional SC-P codes and the state-of-the-art protograph codes for the whole signal-to-noise-ratio (SNR) range under study. The excellent performance and simple implementation make the enhanced SC-P codes a very competitive error-correction solution for the next-generation TDMR devices.

A Study on Block-Based Neural Network Equalization in TDMR System With LDPC Coding

To achieve a high track density, two-dimensional magnetic recording (TDMR) is combined with shingled magnetic recording (SMR). SMR makes it possible to record 1 bit on a few grains. However, the performance will be remarkably deteriorated by the increased media noise, the inter-track and inter-symbol interference (ITI and ISI). Therefore, the application of effective equalization and error control coding are required. In this paper, we investigate a simple block-based neural network equalizer (NNE) that mitigates the influence of ITI and ISI. We compare the equalization effects of the NNE and a conventional 2-D equalizer with low-density parity-check (LDPC) coding based on a random Voronoi grain media model. Simulation results show the proposed block-based NNE achieves better bit error rate performance than the conventional 2-D linear equalizer followed by the a posteriori probability (APP) detector and a sum-product (SP) decoder. In addition, we find the block-based NNE is sensitive to write errors.

Polar Channel Coding Schemes for Two-Dimensional Magnetic Recording Systems

This paper proposes new two-dimensional magnetic recording (TDMR) systems using polar channel coding as practical error correction coding. It is known that the time and space complexities of the encoding/decoding algorithms based on polar channel coding are $\mathcal {O}(N \log N)$ , where $N$ is the codeword (block) length. If we compare the error-correction performance of a polar code with that of a low-density parity-check (LDPC) code in the same rate, it is known that the polar code has a longer length and its decoder still has a lower implementation complexity than the LDPC decoder. Therefore, relatively low-complexity coding schemes are preferable for any TDMR systems under high rates and relatively long codeword lengths. In this paper, the proposed TDMR system serially concatenates a two-dimensional (2-D) modulation code with one-dimensional (1-D) polar codes in each down-track direction. These element polar codes are designed on the fundamentals of the channel polarization theory, which are applied for channels with memory. Actually, it evaluates the performance of the signal processing scheme with concatenated coding and generalized partial response equalization for the proposed TDMR system using bit-patterned media by computer simulations. As a result, it shows that the block error rate performance of the proposed TDMR system with the 2-D modulation and polar channel coding schemes is superior to that of the 1-D system with the conventional 1-D high rate modulation and LDPC coding schemes.

An Intertrack Interference Subtraction Scheme for a Rate-4/5 Modulation Code for Two-Dimensional Magnetic Recording

The performance of a rate-4/5 modulation code is evaluated in two-dimensional magnetic recording (TDMR) channels, where a magnetic medium is described by a discrete Voronoi model, and the two-dimensional sensitivity function of the reader is adopted to generate the TDMR readback signal. Since the read-head sensitivity function covers many tracks, it causes intertrack interference (ITI) that can deteriorate the system performance. Therefore, this letter proposes an ITI subtraction scheme in conjunction with the rate-4/5 modulation code in a coded TDMR channel to mitigate the ITI effect embedded in the readback signals before performing an iterative decoding process. We also investigate the data-dependent readback amplitude distributions and evaluate the TDMR system performance via computer simulation. Results show that the proposed scheme helps improve the TDMR system performance, especially when the areal density is high.

Two-Dimensional Magnetic Recording With a Novel Write Precompensation Scheme for 2-D Nonlinear Transition Shift

This paper proposes a novel write precompensation scheme that adjusts the current magnitude, not the timing, to mitigate the serious 2-D nonlinear transition shift (NLTS) during the shingled writing process in a two-dimensional magnetic recording (TDMR) system. This scheme calculates the needed optimum write field for the current transition being written by subtracting the demagnetizing field produced by the previous written bits from a fixed write field and determines the optimum writing current based on the minimum mean-squared-error criterion between the needed field and the available fields within an established field library. Such a write precompensation scheme is tested with a micromagnetic writer and exchange coupled composite media, and readback with a rotated reader array combined with a read channel. Simulation indicates that the proposed write precompensation technique can decrease bit error rate compared with the case without write precompensation and approach the performance of a system with negligible NLTS. Correspondingly, the maximum user areal density is increased from 3.2 to 4.1 Tbits/in $^{\mathrm {\mathbf {2}}}$ for a TDMR system with 8 nm Voronoi grains due to the proposed write precompensation.

Spinstand demonstration of areal density enhancement using two-dimensional magnetic recording (invited)

Exponential growth of the areal density has driven the magnetic recording industry for almost sixty years. But now areal density growth is slowing down, suggesting that current technologies are reaching their fundamental limit. The next generation of recording technologies, namely, energy-assisted writing and bit-patterned media, remains just over the horizon. Two-Dimensional Magnetic Recording (TDMR) is a promising new approach, enabling continued areal density growth with only modest changes to the heads and recording electronics. We demonstrate a first generation implementation of TDMR by using a dual-element read sensor to improve the recovery of data encoded by a conventional low-density parity-check (LDPC) channel. The signals are combined with a 2D equalizer into a single modified waveform that is decoded by a standard LDPC channel. Our detection hardware can perform simultaneous measurement of the pre- and post-combined error rate information, allowing one set of measurements to assess the absolute areal density capability of the TDMR system as well as the gain over a conventional shingled magnetic recording system with identical components. We discuss areal density measurements using this hardware and demonstrate gains exceeding five percent based on experimental dual reader components.

Study of Two-Dimensional Magnetic Recording System Including Micromagnetic Writer

Study of Two-Dimensional Magnetic Recording System Including Micromagnetic Writer

An iterative inter-track interference mitigation method for two-dimensional magnetic recording systems

At high recording density, the readback signal of two-dimensional magnetic recording is inevitably corrupted by the two-dimensional (2D) interference consisting of inter-symbol interference and inter-track interference (ITI), which can significantly degrade the overall system performance. This paper proposes an iterative ITI mitigation method using three modified 2D soft-output Viterbi algorithm (2D-SOVA) detectors in conjunction with an iterative processing technique to combat the 2D interference. The codeword of the outer code is divided and then written on three separate tracks. For every iteration, all 2D-SOVA detectors exchange the soft information to improve the reliability of the a priori information and use it in the branch metric calculation, before feeding the refined soft information to the outer decoder. Simulation results show that the proposed method outperforms the conventional receiver and the existing partial ITI mitigation method.

Detection for Two-Dimensional Magnetic Recording Systems

Detection for Two-Dimensional Magnetic 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 .

Read Channel Modeling for Detection in Two-Dimensional Magnetic Recording Systems

In this paper, we describe a read channel model for detector design for two-dimensional magnetic recording (TDMR) system, a novel strategy for recording at upto 10 Tb/in <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> . We describe a scheme for (1) modeling of the recording medium, (2) modeling of the write/readback process, and (3) an experimental method for the characterization of noise in the TDMR channel, occurring due to irregularities in the bit-boundaries in the recording medium, that can be used for detection purposes.

2-D Magnetic Recording: Read Channel Modeling and Detection

Two-dimensional magnetic recording (TDMR) is a novel storage architecture that, in theory, can achieve a density of up to 10 Tb/in <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> . It uniquely differs from other proposed next-generation architectures because of its reliance on sophisticated 2-D signal-processing algorithms. Recently, a number of contributions have been made in the development of read-channel models and detectors for TDMR systems. In this paper, we provide a detailed review on all important read-channel models under consideration. Our discussion focuses on the purpose of each model, placing a special emphasis on the suitability of the Voronoi model for the purpose of designing detectors. We also propose several detection schemes for TDMR based on the Voronoi model and present some numerical results.