Detection window and lineal density in digital recording systems

A theoretical model has been established describing the required detection window for a given bit error rate (BER) as a function of lineal transition density (FCI). The model was developed utilizing an extension of a quasi-optimum matched filter analysis for noise spectral density based on media or white noise-dominant recording channels. The relationship between the minimum achievable degradation in RMS timing error vs. the relative separation or pulse slimming factor is used to calculate the minimum detection window at a given error rate as a function of lineal transition density. In addition, calculations of required data detection window versus lineal data density, BER versus lineal data density, and BER versus normalized data detection window are presented. Comparison of the theoretical calculations with experimental data for a practical head-disc interface exhibit remarkable agreement near the region of operation. The model has been demonstrated to be applicable to various encoding schemes and is useful to evaluate detection window margins and to gain insight to system operating characteristics.