Shirish A. Altekar

Error-event characterization on partial-response channels

Two algorithms for characterization of input error events producing specified distance at the output of certain binary-input partial-response (PR) channels are presented. Lists of error events are tabulated for PR channels of interest in digital recording.

Improved Detection For Magnetic Recording Systems With Media Noise

In this paper, data-dependent partial local feedback noise prediction schemes will be considered for magnetic recording applications where media noise is significant. These schemes are shown to provide improvement over standard partial response maximum likelihood detection as well as non-data-dependent noise prediction.

Improvements in detectors based upon colored noise

Two techniques which improve the performance of partial response maximum likelihood systems (PRML) without increasing the number of states in the Viterbi algorithm (VA) are studied. Both techniques rely on the fact that the noise samples going into the VA are correlated. One technique, known as partial local feedback noise prediction, uses estimates of the past noise samples taken from the path memory of the VA to form a prediction of the current noise sample. The predicted noise is subtracted from the current equalizer output prior to updating the path metrics. In the other technique, the VA operates in a block mode simultaneously on X consecutive equaliser outputs. A branch metric appropriate for colored noise is used within each block of K samples. Simulation results for a Lorentzian channel equalized to a PR4 target show that block processing with 4 samples gives almost the same error probability as noise prediction using a 2-tap predictor. Results also show that the feedback in the noise prediction technique increases the length of error events.

Distance Spectra For PRML Channels

Distance spectra for partial response maximum likelihood (PRML) channels are computed by finding generating functions for their error state diagrams. A matrix oriented technique for calculating the generating function is discussed that is reasonable for systems with a relatively small number of states. For those cases where the number of states is larger, a computer search algorithm is described which yields the first few terms of the generating function. Results are presented for channels that are of a form suitable for digital magnetic and digital optical recording. The effect of preceding is also considered.

A 700 Mb/s BiCMOS read channel integrated circuit

A read channel IC achieves >1.5 dB SNR improvement over a 32/34 rate EPRML read channel at 2.8 user bit density. The 0.18 /spl mu/m BiCMOS chip operates up to 700 Mb/s with 1.8 W read mode power using 3.3 V analog and 1.8 V digital power supplies. The die area is 9.64 mm/sup 2/.

Finite truncation depth trellis codes for the dicode channel

A technique for the construction of trellis codes that double the squared free Euclidean distance on the dicode (1-D) channel is presented. This technique uses periodic pruning and permutation of the trellis to ensure that a small detector path memory length is sufficient. Some high rate codes constructed using this technique are given.

Decision-feedback equalization via separating hyperplanes

The design of finite-length decision-feedback equalization (DFE) forward and feedback filters under the assumption of genie-aided feedback and independent and equally likely transmitted symbols is considered. It is shown that the problem of determining DFE filters that minimize the probability of symbol error at high signal-to-noise ratio (SNR) is equivalent to finding the hyperplane that maximally separates two given finite groups of points in a finite-dimensional Euclidean space. The latter task can be formulated as a quadratic program which is readily solved numerically. It is also shown that the problem of finding finite-length DFE filters that minimize the probability of symbol error at any SNR subject to a certain separation condition is a convex optimization problem. The case where the transmitted data is coded using a runlength-limited code is also investigated. Examples show that this criterion yields a performance that is better than zero-forcing DFE on severely distorted channels at high SNR.