NOR Flash Memory Degradation Detection via Time-to-Completion

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

NOR Flash memory devices suffer from degradation mechanisms that lead to performance errors over time, necessitating an early warning system to detect degradation before data corruption occurs, which existing technologies have not adequately addressed.

Innovation Solution

The implementation of a system that estimates the dispersion of threshold voltages in NOR Flash memory cells during read operations by measuring time-to-completion values, providing an early warning signal when significant degradation is detected, allowing for data relocation and extending the operational life of the device.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If error correcting codes are used to address bit errors, then reliability is improved, but device complexity increases

Engineering Contradiction:
Improveerror correction capabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent measures time-to-completion values during normal read operations to detect degradation before it causes errors. By performing this measurement as a preliminary action during routine reads, the system can identify at-risk data and relocate it before corruption occurs, reducing reliance on complex error correction codes while maintaining reliability

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses the existing read operation infrastructure to simultaneously perform data retrieval and degradation detection. The time-to-completion measurement leverages the same read circuitry and timing mechanisms already present in the memory system, allowing self-diagnosis without adding separate complex detection hardware

Inventive Principle:
Principle #25Self-service

2Reliability

If data is relocated to extend operational life, then reliability is improved, but loss of time occurs due to data movement

Engineering Contradiction:
Improveoperational life extensionVSAvoiddata reconstitution time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system detects degradation early using time-to-completion measurements during normal reads, allowing data to be relocated proactively before corruption occurs. This preliminary detection enables planned data movement during low-utilization periods rather than emergency relocation when errors are detected, minimizing disruption and time loss

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system continuously monitors time-to-completion values and compares them against thresholds to provide feedback on degradation status. This feedback mechanism enables dynamic decision-making about when to relocate data, optimizing the balance between extending operational life and minimizing data movement overhead by acting only when necessary

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS8369143B2Early detection of degradation in NOR flash memory
Publication Date: 2013.02.05 SANDISK TECHNOLOGIES LLC
  • US8369143B2 patent drawing
  • US8369143B2 patent drawing
  • US8369143B2 patent drawing

AI summary

The embodiments of the invention in this disclosure describe techniques for early warning of degradation in NOR Flash memories by estimating the dispersion of the threshold voltages (VT's), of a set of NOR Flash memory cells during read operations. In an embodiment invention the time-to-completion (TTC) values for the read operation for the memory cells are used as a proxy for dispersion of the threshold voltages (VT's). If the measured TTC dispersion differs by more than a selected amount from the reference dispersion value, a warning signal is provided to indicate that the page of memory has degraded significantly. Higher level components in the system can use the warning signal to take appropriate action. Since every cell's VT position in an ideal distribution can be estimated, the data from each cell can have a confidence level assigned based on deviation from the mean of an ideal distribution.