Soft Error Generator for Non-Volatile Memory Verification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing memory systems struggle to reproduce and verify correction functions for soft errors in non-volatile memory cell arrays, where writing errors occur due to resistance variable devices under repetitive conditions, which are not addressed by existing error generators.

Innovation Solution

A memory system and operation program that utilize random numbers to generate write and read data considering probability errors, employing modulo and XOR operations to stochastically introduce errors, allowing for the reproduction of soft errors and verification of correction functions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If existing error generators are used to reproduce errors in non-volatile memory, then hard failures can be reproduced, but soft errors cannot be reproduced or verified

Engineering Contradiction:
Improveerror reproduction capabilityVSAvoiderror type coverage
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent changes the parameters of error generation by introducing probabilistic error injection mechanisms that operate differently from deterministic hard failure generators. The system varies error parameters including error probability, error location distribution, and error pattern types to reproduce both hard failures and soft errors, thereby resolving the contradiction between reliability and adaptability in error reproduction capability

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent applies dynamics by making the error generation process adaptive and variable rather than static. The error generator dynamically adjusts error characteristics based on memory state, access patterns, and probabilistic models to simulate real-world soft errors that occur under specific operational conditions, enabling verification of correction functions that were previously impossible to test

Inventive Principle:
Principle #15Dynamics

2Productivity

If writing is repeatedly performed under the same condition in non-volatile memory, then data can be stored, but establishment errors occur on very rare occasion

Engineering Contradiction:
Improvewriting throughputVSAvoidwriting accuracy
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent applies preliminary action by performing error detection and correction operations before they are actually needed. The system proactively verifies data integrity after writing operations and pre-corrects potential establishment errors using ECC codes, preventing rare writing failures from propagating. This allows high writing throughput to be maintained while ensuring reliability even under repetitive writing conditions

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback mechanisms where the results of writing operations are monitored and fed back to the control system. When establishment errors are detected through read-verify cycles or ECC correction attempts, the system adjusts subsequent writing operations accordingly, modifying write parameters or retrying with different conditions to prevent recurrence of the same errors

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS12148488B2Memory system and memory operation
Publication Date: 2024.11.19 SONY SEMICON SOLUTIONS CORP
  • US12148488B2 patent drawing
  • US12148488B2 patent drawing
  • US12148488B2 patent drawing

AI summary

A memory system according to an aspect of the present disclosure includes a soft error generator that generates write data or read data considering a probability error by using a specific number as a random number. In the memory system, the write data or the read data considering a probability error is generated by using the random number. Herein, whether or not an error occurs is randomly changed depending on the random number. Accordingly, an error stochastically occurs, which makes it possible to reproduce a soft error.