Flexible Inter-Track Interference Cancellation for Generalized Interlaced Magnetic Recording

Interlaced magnetic recording (IMR) shows potential to achieve higher areal density capability (ADC) than conventional magnetic recording by recording tracks in an interlaced manner [1]. Top and bottom tracks can be recorded in different frequencies such that the bit error rate (BER) performance can be balanced, providing extra track or linear density pushes. For example, heat assisted IMR (HIMR) enables higher ADC than conventional heat assisted magnetic recording (HAMR) and shingled recorded HAMR by allowing broader bottom tracks with sharper bit transitions and narrow top tracks by control the heat spot [2]. On the other hand, bottom tracks in IMR can suffer from sever inter-track interference (ITI) by subsequent top track writes on both sides, especially as the track density increases [3]. In order to mitigate the ITI effect for such double-sided squeezed tracks, customized ITI cancellation scheme has been introduced, where the asynchronously written tracks are processed in the oversampled domain and overall ADC can be improved [4]. In this study, ITI cancellation scheme is extended for generalized IMR framework, where the ITI can be distributed for top and bottom tracks by applying dual write configurations. One of the major contributions of the tangible ADC gain by HIMR is the additional laser spot control to form narrow but sharp bit transitions in top tracks [2]. In this manner, the wide and sharp bottom tracks are only partially erased by the subsequent top track writes, and, thus, the ITI can be effectively managed. Likewise, the write track widths can be modulated for conventional perpendicular recording and may provide the ADC gain once the ITI is efficiently handled. Note that ITI cancellation can also be needed for top tracks as well, and the relative linear density differences of ITI can be both positive and negative. Therefore, flexible ITI cancellation scheme is proposed in this study to handle the widely varying synchronization issues. The waveforms need to be over-sampled first to effectively synchronize the neighboring tracks, and double of baud rate of lower density track can be used. For example, under 2500 and 2000 kBPI dual writer configurations, the oversampled data can be denoted as x(t)2x, x(b)1.6x respectively, relative to its baud rate. Then, the error signal is first estimated, e(t)2x, e(b)1.6x, resampled to the side track data rate, e’(t)1.25x, e’(b)0.8x, and then the ITI response is estimated for tracks mt and mb as, {h’(t)mt−1, h’(t)mt+1}=arg minh(t)mt−1,h(t)mt+1 || e’(t)1.25x − h(t)mt−1*nmt−1−h(t)mt+1*nmt+1 ||² {h’(b)mb−1, h’(b)mb+1}=arg minh(b)mb−1,h(b)mb+1 || e’(b)0.8x − h(b)mb−1*nmb−1−h(b)mb+1*nmb+1 ||² where nm is the recorded binary data of the track m, and estimated side track data will be used for actual implementation. The ITI signal is estimated from {h’}’s, resampled to baud-rate if needed, and subtracted from the baudrate readback signal. The feasibility of such generalized IMR with flexible ITI cancellation is investigated with the numerical channel simulations. For numerical evaluations of IMR, microcell channel model [1] is employed, where the media is modeled by the square micro-cell of 1.5nm sides, and the granular effect is accounted by shifting the bit and track boundaries randomly for the locally clustered cells. If the writer configuration is maintained except the write frequency for IMR, bottom tracks should be recorded at the relaxed density due to the ITI. If dual write configurations can be used for narrow top tracks, bottom tracks can be recorded with aggressive linear density by forming relatively sharp bit transitions in the remained portion. Figure 1 illustrates the averaged media polarization of repeated [1 1 –1 −1] patterns with single write configuration in (a) bottom and (b) top tracks at 2000 and 2500 kBPI with 529 kTPI. The writer width is assumed to be 60 nm, and the transition curvature is modeled by the ellipse with depth of 20 nm. As illustrated, the bottom tracks are erased by the top track writes and expected to suffer from ITI. On the other hand, same track density 2T patterns with dual write configurations is illustrated in (c) bottom and (d) top tracks at 2500 and 2000 kBPI. The writer width is changed 20% and −20%, while the depth is factored 1/2^½ and 2^½, respectively. As shown, the ITI is distributed for both tracks, and ITI cancellation is needed for both tracks. Readback waveforms are evaluated with the Gaussian read sensitivity function and BER performance is investigated. Other experimental setups are set to as [1], and head electronic signal to noise ratio (HESNR) is scanned from 12 to 30 dB under the track and linear densities illustrated in Fig. 1. Figure 2. shows BER with flexible ITI cancellations in (a) single and (b) dual write configuration IMRs. For the single mode, bottom track BER performance improved significantly by ITI cancellation, while the top track experiences negligible ITI. On the dual mode, both tracks are experiencing the ITI and are improved by the ITIC. In overall, −1.5 order target BER can be achieved at 30 dB HESNR by the dual mode IMR with the flexible ITIC.