Adjusting a read detection equalizer filter of a magnetic tape driving employing a recording format required control pattern

Abstract

Adjustment of a read detection equalizer filter of a magnetic tape drive is conducted utilizing a required control pattern signal which comprises a required portion of a normal recording format of the magnetic tape. An example comprises a data set separator signal. The sensed required control pattern signal is processed with respect to a target reference pattern signal to adapt the read detection equalizer filter to the target reference pattern signal.

Description

technical field [0001] This invention relates to the equalization of tape signals, and more particularly to adjusting the equalizer of a tape drive. Background technique [0002] The read waveform of a magnetic recording channel needs to be filtered to shape it prior to read detection. This filter is called an equalizer. [0003] Magnetic read / write heads are typically formed by thin-film processes that have acceptable tolerances such that the transfer functions vary somewhat and that no one equalization filter will match all heads. Therefore, an equalization filter is usually used to match a specific head. [0004] Typically, one or more multiple iterative algorithms are used to initialize the equalization filter, such as the least mean square (LMS) algorithm, using waveforms of special test patterns, or using data as the signal to which the algorithm is applied. The disadvantage of this method is that it requires many iterations to converge. The LMS scheme is discussed...

AI Technical Summary

Problems solved by technology

[0007] LMS algorithms rely on tentative decisions made on pseudorandom training data or random data waveforms, which often lead the algorithm to converge to the correct transfer function, but with enough noise, these decisions can be wrong, causing the algorithm Divergence where inappropriate updates are made to the equalizer taps

Images