Memory Read Error Handling Using ARC-Based Soft Decoding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing memory sub-systems face inefficiencies in read operations during read error handling, particularly due to the high computational cost and latency associated with soft-decision decoding techniques, which are more complex and time-consuming.

Innovation Solution

Implementing auto read calibration (ARC) to obtain log-likelihood ratio (LLR) values for use in soft-decision decoding by adjusting read voltage levels and using XOR operations to identify bit flips, thereby reducing the need for extensive computational resources.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If soft-decision decoding techniques are used for read error handling, then error correction capability is improved, but computational complexity and latency increase

Engineering Contradiction:

Improveerror correction capabilityVSAvoidcomputational complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The patent performs auto-read calibration (ARC) in advance to establish voltage-to-LLR mapping relationships before actual read operations. This preliminary action creates lookup tables that store the correspondence between read voltage levels and log-likelihood ratio values, enabling the soft-decision decoder to quickly retrieve pre-computed LLR values without performing complex real-time calculations, thus reducing computational complexity while maintaining error correction capability

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent transforms the soft-decision decoding process by changing the parameter representation from raw voltage readings to log-likelihood ratio (LLR) values. By adjusting read voltage levels and mapping them to LLR parameters through ARC, the system converts complex voltage analysis into simplified discrete parameter lookups, reducing the computational burden on the decoder while preserving the probabilistic information needed for error correction

Inventive Principle:

Principle #35Parameter changes

2Reliability

If soft-decision decoding techniques are used for read error handling, then error correction capability is improved, but read operation latency increases

Engineering Contradiction:

Improveerror correction capabilityVSAvoidread operation latency

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent creates a simplified copy of the LLR values through lookup tables that mirror the essential probabilistic information needed for soft-decision decoding. Instead of performing full soft-decision decoding computations, the system uses pre-computed LLR copies stored in lookup tables, which can be rapidly retrieved and fed to the decoder, significantly reducing latency while preserving error correction performance

Inventive Principle:

Principle #26Copying

3Productivity

If auto read calibration is implemented to obtain LLR values, then read operation efficiency is improved, but device complexity increases

Engineering Contradiction:

Improveread operation efficiencyVSAvoiddevice complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the error handling process into distinct stages: auto-read calibration (ARC) for LLR value generation, lookup table construction for storing voltage-to-LLR mappings, and soft-decision decoding for final error correction. This segmentation allows each component to be optimized independently, with ARC focusing on accurate LLR generation and the lookup table providing rapid retrieval, thereby improving overall read efficiency while distributing complexity across modular components

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces lookup tables as an intermediary structure between the auto-read calibration process and the soft-decision decoder. These lookup tables store pre-computed LLR values that mediate between the complex voltage analysis performed by ARC and the decoding process, eliminating the need for real-time complex calculations during actual read operations and improving efficiency while containing complexity in the pre-computation phase

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20250349378A1Reducing read operations during read error handling in a memory device

Publication Date: 2025.11.13 MICRON TECHNOLOGY INC

AI summary

A plurality of read operations is performed on encoded host data stored in a memory device using the plurality of read voltage level adjustments to obtain current sensed data responsive to initiating an auto-read calibration operation comprising a plurality of read voltage level adjustments. A previous sensed data is obtained from a previous one of the plurality of read operations performed on the encoded host data using a previous read voltage level adjustment. One or more flipped bits of the current sensed data is identified based on the previous sensed data and the current sensed data. A likelihood value is assigned to the one or more flipped bits. Soft-decision decoding is performed on the encoded host data responsive to completion of the auto-read calibration, performing, using the assigned likelihood value.