Read Retry Table Generation for Non-Volatile Memory Cells
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory devices face challenges in efficiently generating or changing read voltages for non-volatile memory cells, leading to issues with data retention, power consumption, and manufacturing costs.
Innovation Solution
An apparatus and method are developed to generate a read retry table using machine learning algorithms, which collects data from multiple memory dies, selects cluster candidates based on pass rates, performs clustering on word lines, and updates the read retry table to optimize read voltage levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional methods are used to generate read voltages for non-volatile memory cells, then manufacturing costs and development time are reduced, but data retention reliability and read accuracy deteriorate
Solution Approach 1:
The system performs preliminary characterization of memory cell threshold voltage distributions during manufacturing and stores this data in a lookup table. This preliminary action enables the read voltage generator to quickly retrieve and apply pre-determined optimal read voltages without performing complex real-time analysis, thereby improving data retention reliability while avoiding the high computational costs that would otherwise be required.
Solution Approach 2:
The read voltage generator dynamically adjusts read voltages based on the specific memory cell block being accessed and its characteristics stored in the lookup table. Instead of using fixed read voltages, the system adapts the read voltage levels to match the actual threshold voltage distributions of different memory blocks, improving read accuracy and data retention reliability while maintaining manufacturing efficiency.
2Measurement precision
If complex algorithms are used to optimize read voltage generation, then read accuracy improves, but computational resources and processing time increase
Solution Approach 1:
Complex threshold voltage distribution analysis and read voltage optimization are performed in advance during manufacturing or initialization, with results stored in a lookup table. During actual read operations, the system simply queries the lookup table and applies the pre-determined voltages, achieving high read accuracy without requiring complex real-time computation.
Solution Approach 2:
Instead of performing complex computational algorithms during read operations, the system creates a simplified copy of the optimal read voltage settings in a lookup table structure. This copy contains pre-calculated voltage values that replicate the results of complex algorithms, enabling fast and accurate read operations without the computational overhead.
3Measurement precision
If fixed read voltages are used for all memory blocks, then device complexity is reduced, but read accuracy and data retention vary across different memory blocks
Solution Approach 1:
The system implements local quality by storing and applying specific read voltage settings for each memory block based on its unique threshold voltage characteristics. Instead of using a single fixed voltage for all blocks, the read voltage generator retrieves block-specific voltages from the lookup table, ensuring optimal read accuracy for each local memory region while maintaining manageable system complexity through the tabular storage approach.
Data Source
AI summary
A read retry table generator is coupled to a plurality of memory dies via a data path. The read retry table generator is configured to: collect data from a plurality of memory cells coupled to a plurality of word lines in the plurality of memory dies; determine a pass rate of collected data appertaining to a plurality of clusters; select a cluster candidate among the plurality of clusters, based on a pass rate of collected data; and cluster the collected data into the cluster candidate to generate a read retry table.