Error correction code processor employing adjustable correction power for miscorrection minimization
AssigneeCirrus Logic, Inc.
InventorChristopher P. Zook
Life of this patent
Estimating the term…
Loading patent…
AssigneeCirrus Logic, Inc.
InventorChristopher P. Zook
Estimating the term…
An error correction code (ECC) processor which minimizes the probability of miscorrecting an ECC codeword in a multi-dimensional code, such as a product code, thereby avoiding the added latency in recovering from the miscorrection. Initially, the ECC processor sets the error correction power of the code to a minimum value so that the probability of making a miscorrection is low even though the probability of detecting an uncorrectable codeword is high. The ECC processor then performs iterative error correction passes over the multi-dimensional code and incrementally increases the error correction power if no corrections are made during a current pass. Increasing the correction power may render a previously uncorrectable codeword correctable, and after correcting the codeword, it may render a corresponding intersecting codeword in the other dimension correctable. If a codeword is corrected during a current pass, the correction power is reset to the minimum value for the next pass and the correction process continues. In this manner, the probability of making a miscorrection is minimized since the number of remnant redundancy symbols used to check for a miscorrection is maximized.
An error correction code (ECC) processor for correcting errors in binary data of an ECC codeword, including a plurality of ECC redundancy symbols for correcting a maximum number of correctable errors in the ECC codeword, read from a disc storage medium, the ECC codeword comprising a first and a second set of intersecting ECC codewords of a multi-dimensional code wherein the error correction code processor processes the first and second sets of ECC codewords in iterative error correction passes, the ECC processor comprising:
An error correction code (ECC) method for correcting errors in binary data of an ECC codeword, including a plurality of ECC redundancy symbols for correcting a maximum number of correctable errors in the ECC codeword, read from a disc storage medium, the ECC codeword comprising a first and a second set of intersecting ECC codewords of a multi-dimensional code, the method comprising the steps of:
An error correction code (ECC) processor for correcting errors in binary data of a ECC product codeword read from a disc storage medium, the ECC product codeword comprising a first and second set of intersecting ECC codewords, the error correction code processor processing the first and second set of intersecting ECC codewords in iterative error correction passes and comprising:
Dependencies are inferred from the claim text.