Marvell International Limited

Marvell Asia Pte., Ltd.

Marvellworld Trade Ltd.

Nitin Nangare

Gregory Burd

Zining Wu

Nedeljko Varnica

Michael Madden

Panu Chaichanavong

Hongxin Song

Shaohua Yang

Mats Oberg

Ke Han

Yifei Zhang

William J Mitchem

Vincent M Wong

Noam Mizrahi

Hongying Sheng

Ruwan Ratnayake

Konstantin Kudryavtsev

Leo Jiang

Igor Polivanyi

Wen Lung Chang

Jinlu Shen

Quynh Chau

Han-Ting Chiang

Ahmed Medhat Ahmed Fahmi Eid Aboutaleb

A data channel on an integrated circuit device includes a non-linear equalizer having as inputs digitized samples of signals on the data channel, decoding circuitry configured to determine from outputs of the non-linear equalizer a respective value of each of the signals, and adaptation circuitry configured to adapt parameters of the non-linear equalizer based on respective ones of the value. The non-linear equalizer includes a non-linear filter portion, and a front-end filter portion configured to reduce numbers of the inputs from the digitized samples. The non-linear equalizer may be a neural network equalizer, such as a multi-layer perceptron neural network equalizer, a reduced complexity multi-layer perceptron neural network equalizer, or a radial-basis function neural network equalizer. Alternatively, the non-linear equalizer may include a linear filter and a non-linear activation function, which may be a hyperbolic tangent function.

Non-linear neural network equalizer for high-speed data channel

Data read from a storage medium is first processed through a first data path including a first decoder configured to decode data output from at least one first finite impulse response (FIR) filter and first FIR adaptation circuitry configured to adjust a first FIR coefficient for the at least one first FIR filter. The data is then processed through a second data path, which includes at least one second FIR filter and second FIR adaptation circuitry configured to adjust a second FIR coefficient to reach an FIR coefficient that achieves a target minimum number of errors. The second FIR adaptation circuitry is configured to reach the FIR coefficient that achieves the target minimum number of errors faster than the first FIR adaptation circuitry. A second decoder in the second data path is configured to decode data output by the at least one second FIR filter.

Method and apparatus for multi-stage equalization for reading data from storage media

A method of reading data from a rotating magnetic storage medium, having at least one read head, includes storing respective digitized data samples from each respective read head of the at least one read head in a respective timing buffer, determining a zero-phase start phase angle from a preamble of the digitized data samples, feeding forward the zero-phase start phase angle to an interpolator, selecting an interpolation filter based on the fed-forward zero-phase start phase angle, releasing the respective digitized data from the respective timing buffer after a duration sufficient for completion of the determining, the feeding forward and the selecting, and interpolating samples of the digitized data released from the respective timing buffer.

Digital timing recovery in hard disk drive read channel for preamble reduction

A method for digital timing recovery from oversampled analog signals includes computing filter coefficients for digitized samples of the oversampled analog signals based on an oversampling factor of the oversampled analog signals, using the filter coefficients in a rotation filter to compensate for the oversampling factor in the digitized samples of the oversampled analog signals, deriving a starting phase and magnitude from the compensated digitized samples of the oversampled analog signals, and using the starting phase and magnitude in a timing recovery loop to recover a clock from the compensated digitized samples of the oversampled analog signals. The rotation filter may include a plurality of taps, and the circuitry may be configured to compute respective sets of coefficients for respective taps. Each set of coefficients may be dependent on another set of coefficients, or the coefficients may be approximate with each set of approximate coefficients being independent.

Rotation filter for digital timing recovery in hard disk drive read channel

The present disclosure describes aspects of health management for magnetic storage media. In some aspects, a media health manager determines, with a read channel, read metrics for a sector of magnetic storage media that resides in a zone of magnetic storage media. The media health manager accesses read metrics of the zone and updates the read metrics of the zone based on the read metrics determined for the sector to provide updated read metrics for the zone of magnetic storage media. A health score for the zone of magnetic storage media is then determined with a neural network based on the updated read metrics of the zone of magnetic storage media. By so doing, gradual wear of the magnetic storage media may be predicted using the health score, enabling replacement of a magnetic storage media device before failure to improve reliability or availability of data stored to the device.

Health management for magnetic storage media

A physical layer transceiver for connecting a host device to a wireline channel medium includes a host interface for coupling to the host device, a line interface for coupling to the channel medium, a transmit path operatively coupled to the host interface and the line interface, a receive path operatively coupled to the line interface and the host interface, and adaptive filter circuitry operatively coupled to at least one of the transmit path and the receive path for filtering signals on the at least one of the transmit path and the receive path, the adaptive filter circuitry including a non-linear equalizer. The non-linear equalizer may be a neural network equalizer based on a multi-layer perceptron or a radial-basis function, or may be a linear equalizer with a non-linear activation function. The non-linear equalizer also may have a front-end filter to reduce input complexity.

Ethernet physical layer transceiver with non-linear neural network equalizers

A receiver for use in a data channel on an integrated circuit device includes a non-linear equalizer having as inputs digitized samples of signals on the data channel, decision circuitry configured to determine from outputs of the non-linear equalizer a respective value of each of the signals, and adaptation circuitry configured to adapt parameters of the non-linear equalizer based on respective ones of the value. The non-linear equalizer may be a neural network equalizer, such as a multi-layer perceptron neural network equalizer, or a reduced complexity multi-layer perceptron neural network equalizer. A method for detecting data on a data channel on an integrated circuit device includes performing non-linear equalization of digitized samples of input signals on the data channel, determining from output signals of the non-linear equalization a respective value of each of the output signals, and adapting parameters of the non-linear equalization based on respective ones of the value.

Systems and methods for adaptation of a two-dimensional magnetic recording (TDMR) channel are provided. Read-back signals from respective read sensors of a TDMR channel are received at an equalizer, the read-back signals corresponding to a digital signal value. A log-likelihood ratio (LLR) signal is generated based at least in part on the read-back signals. A cross-entropy value is generated indicative of a mismatch between a probability of detected bit and a probability of the true recorded bit. The equalizer is adapted by setting an equalizer parameter to a value that corresponds to a minimum cross-entropy value from among the computed cross-entropy value and one or more previously computed cross-entropy values, to decrease a read-back bit error rate for the TDMR channel.

Systems and methods for adaptation of a two-dimensional magnetic recording channel

A storage control device coupled to a storage device and located remotely from a host device receives media object data from the host device. The storage control device identifies a type of the media object data and select, based on the identified type, a computational model from among a plurality of computational models for use by a computational engine of the storage control device. The computational engine uses the selected computational model to generate metadata describing the media object data. The metadata is stored in the storage device so as to be selectively retrievable from the storage device separately from the media object data.

Systems and methods for generating metadata describing unstructured data objects at the storage edge

In a data storage device having a storage medium, wherein data is written to tracks on the storage medium, data for each track including a preamble, and wherein the preamble in any current track is orthogonal to the preamble in any track adjacent to the current track, and wherein data accumulated in a multi-dimensional space is representative of a relationship between signals from the current track and signals from at least one adjacent track, the relationship between the signals from the current track and the signals from the at least one adjacent track is calibrated by, for each respective position out of a plurality of positions in the multi-dimensional space, accumulating a plurality of data values for the respective position, and determining, from the plurality of data values for the respective position, an inverted covariance matrix without performing a matrix inversion operation, either during run-time or prior to run-time.

Method and apparatus for calibrating read-head off-track detector

A storage device includes read circuitry having a read head having a detector that outputs signals representing data from a first track and an adjacent track. The read head is subject to off-track excursions during which the read head detects signals from both the first track and an adjacent track. Data on each track includes a preamble including a repeating pattern. The repeating pattern in any first track is orthogonal to the repeating pattern in any track adjacent to the first track. The read circuitry also includes respective Discrete Fourier Transform circuits to identify components in the signals corresponding to respective frequencies characteristic of the repeating pattern on the first track and the repeating pattern on the second track, and computation circuitry to determine from the components a ratio by which the read head is off-track. Corresponding methods are provided for operating such a storage device and for reading data.

Method and apparatus for determining read-head deviation using orthogonal preambles

In a storage device, data is written to tracks on a storage medium. Data for each track includes a preamble. The preamble in any current track is orthogonal to the preamble in any track adjacent to the current track. A first read head is positioned over the current track and off-center toward a first adjacent track on a first side of the current track to detect first signals from among the current track, the first adjacent track, and a second adjacent track on a second side of the current track. A second read head is positioned over the current track and off-center relative to the current track toward the second adjacent track to detect second signals from among the current track, the first adjacent track, and the second adjacent track. Analyzing the first and second signals determines an amount by which the read heads are off-track from the current track.

Growing community of inventors

Santa Clara, CA, United States of America

Nitin Nangare