NAND Read Recovery Using Artificial Codewords and Iterative Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data storage device controllers face high latency during read recovery operations due to the need for multiple NAND senses, as they do not effectively utilize information from previous stages to minimize additional sense operations, leading to inefficient error correction processes.
Innovation Solution
Implementing a method that uses first and second codewords obtained with shifted sense voltage thresholds, applying them to hard or soft bit decoders, and generating additional codewords through binning techniques to reduce the number of NAND senses required, thereby minimizing latency by iteratively refining the decoding process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple NAND senses are performed to obtain additional codewords for error correction, then the error correction capability is improved, but the read recovery latency increases
Solution Approach 1:
The patent applies preliminary action by generating artificial codewords from previously obtained sense values before performing additional NAND senses. The controller creates multiple artificial codewords (e.g., 3-7 codewords) by combining and processing the initial sense data, then attempts decoding on these artificial codewords first. This preliminary processing allows the system to potentially recover data without requiring additional time-consuming NAND sense operations, thereby reducing read recovery latency while maintaining error correction capability.
2Measurement precision
If additional NAND senses are performed to obtain more sense values for decoding, then the decoding accuracy is improved, but the number of data transfers increases
Solution Approach 1:
The patent applies copying by creating artificial codewords that are derived copies of the original sense values through mathematical processing. Instead of performing additional NAND senses to obtain more data, the controller generates multiple artificial codewords (second through seventh codewords) by combining and processing the initial sense values. These artificial codewords serve as substitutes for additional physical sense operations, reducing the number of data transfers required while maintaining decoding accuracy.
3Reliability
If the controller performs iterative decoding with multiple codeword generations, then the correction strength is improved, but the processing complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the error correction process into distinct stages: obtaining initial sense values, generating multiple artificial codewords from these values, attempting hard decoding on each artificial codeword in sequence, and only proceeding to additional NAND senses if all artificial codeword decodings fail. This segmented approach allows the controller to systematically increase correction strength through multiple decoding attempts while maintaining manageable processing complexity through clear stage separation.
Data Source
AI summary
The present disclosure, in various embodiments, describes technologies and techniques for use by a data storage controller for decoding codewords during an error correction read recovery process. In illustrative examples, an iterative procedure exploits artificial codewords generated using information obtained from a NAND or other non-volatile memory (NVM) in a previous sense operation. That is, procedures are described that use information obtained in one stage of read recovery to facilitate a subsequent stage to reduce the need to perform additional NAND senses. In one example, information obtained from a sense operation performed for an initial hard bit decode is used in subsequent soft bit decodes. Moreover, iterative decoding procedures are provided that progressively increase correction strength. The procedures may alternate between hard and soft reads while using syndrome weight to determine a failed bit code gradient for identifying the sensing voltage for a next hard sense.