Signal-processing for digital magnetic recording channels

In this dissertation we explore the application of communication theory and information theory to the digital magnetic recording channel. In Chapter 1 we begin with an overview of the recording channel, including descriptions of some of the common signal processing practices used today. In Chapters 2 and 3, variations on the commonly used binary (d,k) codes are introduced. Application of these codes to magnetic recording indicate potential density gains with increased error margins over conventional schemes. The information theoretic limits of a discrete time and a continuous time model of the recording channel are explored in Chapters 4 and 5, respectively. A lower bound on the capacity of a discrete memoryless channel with an M-ary run-length constrained input is given in Chapter 4. In Chapter 5, an upper bound on the capacity of a peak power constrained Gaussian channel is given as a function of certain recording channel parameters. A conjectured lower bound for the same channel is given as well. In Chapter 6, we present a Viterbi detector for a specific modulation/equalization scheme known as compact spectrum. Gains in signal-to-noise ratio are demonstrated. Finally, ideas for future research can be found in Chapter 7.