Adaptive LLR Lookup for NAND Error Recovery Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Non-volatile storage devices often lack the capability to generate optimal log-likelihood ratio (LLR) values for soft decision decoding, which is essential for error recovery, especially in systems where soft decision information is not readily available.
Innovation Solution
A system and method for generating LLR values through multiple reads of non-volatile storage devices with different reference voltage settings, using a set of decoding factors to select the number of reads and aggregation mode, and employing a mapping engine to generate and index LLR values for soft decision decoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If soft decision decoding is implemented to improve error recovery capability, then error recovery reliability is improved, but the system requires LLR values which are not available from storage devices that do not provide soft decision information
Solution Approach 1:
The patent introduces an intermediary component (LLR generation circuit or mapping engine) that converts hard decision read results into soft decision information (LLR values). This intermediary bridges the gap between storage devices lacking native soft decision capability and the soft decision decoder, enabling error recovery without requiring the storage device to provide LLR values directly.
Solution Approach 2:
The patent creates a copy or representation of soft decision information by generating LLR values from hard decision read results. Instead of requiring the storage device to output native soft decision data, the system creates equivalent soft decision information through LLR generation circuits or lookup tables, allowing the soft decision decoder to function with devices that only provide hard decision outputs.
2Measurement precision
If multiple reads are performed with different reference voltage settings to generate accurate LLR values, then error recovery accuracy is improved, but the read operation time and complexity increase
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and storing mapping relationships between hard decision read results and corresponding LLR values in lookup tables. This pre-computation allows the system to quickly generate accurate LLR values during error recovery operations without performing multiple time-consuming reads, thus reducing latency while maintaining precision.
Solution Approach 2:
The patent changes parameters by using different reference voltage settings for multiple reads to gather information about bit reliability. By varying the reference voltage and observing how read results change, the system can infer LLR values that reflect the reliability of each bit, improving accuracy without requiring fundamentally new read mechanisms.
3Productivity
If an LLR lookup table is used to quickly retrieve LLR values during error recovery, then decoding speed is improved, but the memory resources and table size requirements increase
Solution Approach 1:
The patent segments the LLR lookup table into multiple smaller tables or partitions, each handling specific cases or subsets of read results. This segmentation reduces the size of individual tables that need to be maintained in memory while still providing comprehensive coverage for all possible error conditions, thus lowering memory resource requirements while maintaining fast decoding performance.
Solution Approach 2:
The patent designs the LLR lookup table to serve multiple functions: it not only provides LLR values for error recovery but also can be used for other decoding operations and can accommodate different error patterns and read conditions. This multi-functionality reduces the need for separate tables for different purposes, optimizing memory resource utilization while maintaining high decoding speed.
Data Source
AI summary
Systems and methods are provided for performing error recovery using LLRs generated from multi-read operations. A method may comprise selecting a set of decoding factors for a multi-read operation to read a non-volatile storage device multiple times. The set of decoding factors may include a total number of reads, an aggregation mode for aggregating read results of multiple reads, and whether the read results include soft data. The method may further comprise issuing a command to the non-volatile storage device to read user data according to the set of decoding factors, generating a plurality of Log-Likelihood Ratio (LLR) values using a mapping engine from a pre-selected set of LLR value magnitudes based on the set of decoding factors, obtaining an aggregated read result in accordance with the aggregation mode and obtaining an LLR value from the plurality of LLR values using the aggregated read result as an index.