Multilevel Memory ECC with Concatenated Codes for Lower Parity Overhead

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The increasing raw bit error rate (BER) due to shrinking technology nodes and higher bits per cell in nonvolatile memory (NVM) devices leads to a corresponding increase in parity information bits, offsetting the gain in storage capacity, especially when using hard-decision ECC solutions like Hamming and BCH codes.

Innovation Solution

The implementation of advanced error correcting code (ECC) schemes and calculation systems that select optimal parameters, such as ECC schemes and bit per cell combinations, to increase storage capacity while maintaining desired error rates, including the use of concatenated codes with trellis code modulation (TCM) and Reed Solomon (RS) codes to reduce parity overhead and improve reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If hard-decision ECC solutions like Hamming and BCH codes are used, then error correction capability is improved, but storage capacity is reduced due to increased parity information bits

Engineering Contradiction:
Improveerror correction capabilityVSAvoidstorage capacity
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent changes the fundamental parameters of error correction by transitioning from hard-decision ECC (Hamming/BCH codes) to soft-decision ECC schemes. This parameter change in the decision mechanism allows for more efficient error correction with reduced parity overhead, thereby improving storage capacity while maintaining reliability

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent employs composite coding schemes that combine multiple error correction techniques (concatenated codes combining outer and inner codes, TCM integration) to achieve superior error correction performance with lower parity bit requirements compared to single-code hard-decision approaches

Inventive Principle:
Principle #40Composite materials

2Quantity of substance

If the number of bits per cell is increased, then storage capacity is improved, but raw bit error rate increases leading to more parity bits needed

Engineering Contradiction:
Improvestorage capacityVSAvoidraw bit error rate
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent converts the harmful effect of increased raw bit error rate (which naturally occurs with higher bits per cell) into a benefit by using soft-decision ECC that can effectively correct these errors. The soft-decision mechanism leverages analog signal information to correct errors that would be uncorrectable with hard-decision methods, thus enabling high-density storage while maintaining reliability

Inventive Principle:
Principle #22Blessing in disguise (Convert harm into benefit)

3Quantity of substance

If technology nodes are shrunk, then device density is improved, but raw bit error rate increases requiring more parity information bits

Engineering Contradiction:
Improvedevice densityVSAvoidraw bit error rate
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent applies parameter changes by adopting soft-decision ECC schemes that fundamentally alter how errors are detected and corrected. This change in the error correction parameter allows the system to handle the increased error rates from shrunk technology nodes without proportionally increasing parity bit requirements

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS8370702B2Error correcting codes for increased storage capacity in multilevel memory devices
Publication Date: 2013.02.05 MICRON TECHNOLOGY INC
  • US8370702B2 patent drawing
  • US8370702B2 patent drawing
  • US8370702B2 patent drawing

AI summary

Embodiments of the present disclosure provide methods, systems, and apparatuses related to multilevel encoding with error correction. In some embodiments, data may be programmed and/or read from a matrix of nonvolatile memory cells with concatenated encoding/decoding schemes. In some embodiments, a calculation module may determine an actual bit per cell value of a given combination of parameters of a nonvolatile memory device. Still other embodiments may be described and claimed.