Nonvolatile Memory Read Voltage Search via Regression Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Nonvolatile memory devices face challenges in accurately determining optimal read voltages due to shifting threshold voltage distributions caused by charge leakage, program disturbances, and environmental changes, leading to overlapping states and potential read errors.
Innovation Solution
A method using regression analysis to quickly and accurately determine the optimal read voltage level by selecting multiple voltage levels, performing data read operations, and generating cell count values, which are then used to calculate the optimal voltage based on a first-order polynomial, reducing calculation complexity and improving data integrity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional methods are used to determine read voltage, then the process is simple, but the accuracy and precision of optimal read voltage determination deteriorates due to threshold voltage distribution shifts
Solution Approach 1:
The patent applies parameter changes by using a first-order polynomial model to represent the relationship between read voltage and cell count, transforming a complex determination problem into a simplified linear regression task. This allows accurate optimal read voltage determination while maintaining computational efficiency through parameter transformation and mathematical modeling.
2Measurement precision
If multiple read voltage levels are tested to improve accuracy, then the precision of optimal voltage determination improves, but the time required for voltage search increases
Solution Approach 1:
The patent applies preliminary action by pre-selecting multiple candidate read voltage levels before the actual determination process. By preparing a set of candidate voltages in advance and using polynomial fitting to evaluate them, the system achieves precise optimal voltage determination without exhaustive testing, thus reducing the time loss while maintaining high precision.
3Reliability
If threshold voltage distributions are allowed to shift naturally, then the device operation is simple, but the reliability of data reading deteriorates due to overlapping states
Solution Approach 1:
The patent applies feedback by using cell count measurements at different read voltage levels to inform the polynomial fitting process. The cell count data serves as feedback that reveals the actual threshold voltage distribution state, allowing the system to dynamically adjust and determine the optimal read voltage that accounts for distribution shifts, thereby maintaining high data read reliability.
Data Source
AI summary
A nonvolatile memory device includes a plurality of memory cells that have a first state and a second state different from each other. A method of searching a read voltage of the nonvolatile memory device includes determining a number n that represents a number of times a data read operation is performed, selecting n read voltage levels of the read voltage such that a number of read voltage levels is equal to the number of times the data read operation, where the n read voltage levels differ from each other, generating n cell count values by performing n data read operations on the plurality of memory cells using all of the n read voltage levels, and generating an optimal read voltage level of the read voltage by performing a regression analysis based on a first-order polynomial using the n read voltage levels and the n cell count values.