Sector error rate estimation of concatenated codes in magnetic recording

Abstract

A statistical simulation method is introduced to estimate the sector error rate (SER) of proposed triply concatenated coding systems for magnetic hard disk drives. The outer code is a standard interleaved Reed-Solomon code. The 'middle' codes are run length limiting codes used to demonstrate the effect of error propagation. The inner code(s) are block codes with 1 or 3 parity check bits. First, a baud rate simulation is used to estimate sufficient statistics to extrapolate ECC performance. Second, a Monte Carlo error event simulator generates error instances which are simply checked for correctability. In all code combinations the apparent SNR gains when considering error (or bit) event rates are considerably lessened or completely lost when considering SER.