Browse Prior Art Database

Hybrid Equalization for FTDM Coding on Magnetic Disks

IP.com Disclosure Number: IPCOM000037221D
Original Publication Date: 1989-Dec-01
Included in the Prior Art Database: 2005-Jan-29
Document File: 3 page(s) / 18K

Publishing Venue

IBM

Related People

Feig, E: AUTHOR

Abstract

Fourier transform division multiplexing (FTDM) schemes 1,2 transmit data in contiguous blocks. They augment redundant data to these blocks in order to equalize the effects of the channel memory. This redundancy may be a significant fraction of the data. We disclose here how to drastically reduce this redundancy without loss of channel integrity (at the expense of extra computation in the decoder).

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 3

Hybrid Equalization for FTDM Coding on Magnetic Disks

Fourier transform division multiplexing (FTDM) schemes 1,2 transmit data in contiguous blocks. They augment redundant data to these blocks in order to equalize the effects of the channel memory. This redundancy may be a significant fraction of the data. We disclose here how to drastically reduce this redundancy without loss of channel integrity (at the expense of extra computation in the decoder).

The channel we are interested in is a linearized magnetic recording channel. Its impulse response function h(t) is significant over some time interval. The channel response to the sampled input function fk is given by the convolution. The response function is significantly non-zero for some finite values of k around
0. We call this set of k the support of the channel memory.

The equalization suggested in 1, 2 is rather costly. There they suggest repeating the part of the initial data sequence at the end of the block, the length of the part being the order of the support of the channel memory. As an example, the channel in [2] had the order of its memory support 32, and these were appended to blocks of length 158. This introduced a redundancy over twenty percent.

We propose here to equalize the channel output so that the new (equalized) channel has very short memory. Attempt should be made to equalize the channel to any of the standard short memory partial response channels, such as PR-II, PR-IV, etc. An algorithm such as in [3] may be used to find an appropriate equaliz...