Data reproduction apparatus and method with improved performance by adjusting filter coefficients of equalizer

A data reproduction apparatus, comprising:

• a data detector which detects data based on partial response maximum likelihood (PRML);
• an equalizer comprising a digital filter;
• a level error detector which extracts a level of an output value of the equalizer based on a reference level set in the data detector, and which detects a level error between the level of the output value and a predetermined reference value;
• a coefficient bank which stores a plurality of filter coefficients of the equalizer;
• and
• a coefficient selector which selects at least one of the plurality of the filter coefficients of the equalizer from the coefficient bank such that the level error is minimized, wherein the plurality of filter coefficients includes filter coefficients which are adapted according to a servo status, a degree of defocusing and a degree of tangential tilting.

A method of improving the performance of a data detector which detects data based on partial response maximum likelihood (PRML) by adjusting filter coefficients of an equalizer comprising a digital filter, connected before the data detector, the method comprising:

• extracting a respective level of at least a first sampling data output from the equalizer based on a respective reference level set in the data detector, and detecting at least one level error between the level of each sampling data level extracted and a corresponding predetermined reference value wherein the extracting comprises:
• predetermining at least one reference value corresponding to one of a zero level, a positive middle level, a negative middle level, a positive maximum level and a negative maximum level, which levels are classified depending on the partial response (PR) type of the data detector,
• detecting a level of the output data of the equalizer based on at least one of the classified levels, and
• detecting the level error corresponding to a difference between the detected level and the corresponding reference value;
• and
• adjusting the filter coefficients of the equalizer based on the detected level error such that the level error is minimized.

A method of improving the performance of a data detector which detects data based on partial response maximum likelihood (PRML) by adjusting filter coefficients of an equalizer comprising a digital filter, connected before the data detector, the method comprising:

• extracting a respective level of at least a first sampling data output from the equalizer based on a respective reference level set in the data detector, and detecting at least one level error between the level of each sampling data level extracted and a corresponding predetermined reference value, wherein the extracting comprises:
• extracting a level of a second sampling data output from the equalizer, the second sampling data being successive to the first sampling data,
• determining whether the product of the first and second sampling data levels is less than 0, determining which one of the first and second sampling data levels is greater than 0 if the product of the first and second sampling data levels is less than 0, determining the sampling data level greater than 0 as a positive middle level, and determining the sampling data level less than 0 as a negative middle level, and
• detecting first and second level errors, each corresponding to a difference between the level of the sampling data determined as one of the positive and negative middle levels and a corresponding reference value;
• and
• adjusting the filter coefficients of the equalizer based on the detected level error such that the level error is minimized.

A method of improving the performance of a data detector which detects data based on partial response maximum likelihood (PRML) by adjusting filter coefficients of an equalizer comprising a digital filter, connected before the data detector, the method comprising:

• extracting a respective level of at least a first sampling data output from the equalizer based on a respective reference level set in the data detector, and detecting at least one level error between the level of each sampling data level extracted and a corresponding predetermined reference value, wherein the extracting comprises:
• extracting respective levels of second and third sampling data output from the equalizer, the first, second and third sampling data being successive in order;
• determining whether all of the first, second and third sampling data levels are greater than a predetermined threshold value;
• determining the second sampling data as a positive maximum level if all of the first, second and third successive sampling data levels are greater than the predetermined threshold value;
• determining the second sampling data as a negative maximum level if all of the first, second and third successive sampling data levels are less than the predetermined threshold value;
• and
• detecting the at least one level error corresponding to a difference between the determined level and a corresponding reference value;
• and
• adjusting the filter coefficients of the equalizer based on the detected level error such that the level error is minimized.