Title :
Performance of maximum likelihood detectors in partial erasure
Author :
Bliss, William G.
Author_Institution :
Cirrus Logic Inc., Broomfield, CO, USA
fDate :
9/1/1996 12:00:00 AM
Abstract :
Partial Erasure is a well known phenomenon whereby closely spaced magnetic transitions in digital saturation recording partially erase each other. The resulting playback signals exhibit nonlinear pattern dependent amplitude losses. This paper defines and analyzes the performance of sequence detectors matched to a partial erasure model, matched to the RLL code d=0 or d=1 constraint, and also matched to the popular inter-symbol interference targets: PR4, EPR4, EEPR4, or EEEPR4. All of the d=0 detectors show losses on the order of -20*log10(1-e) dB, where e is the fraction of the magnetic transition amplitudes erased in a di-bit. All of the d=1 detectors exhibit much less loss than the d=0 detectors. The d=1 EEPR4 and EEEPR4 detectors have no distance losses if e<0.2 and 0.37 respectively, even though significant signal energy is lost
Keywords :
digital magnetic recording; estimation theory; intersymbol interference; maximum likelihood detection; partial response channels; runlength codes; EEEPR4; EEPR4; EPR4; PR4; RLL code; closely spaced magnetic transitions; digital saturation recording; inter-symbol interference targets; maximum likelihood detectors; nonlinear pattern dependent amplitude losses; partial erasure; playback signals; sequence detectors; signal energy; Additive white noise; Detectors; Digital magnetic recording; Logic; Magnetic analysis; Maximum likelihood detection; Noise shaping; Performance analysis; Pulse modulation; Saturation magnetization;
Journal_Title :
Magnetics, IEEE Transactions on