NAND Memory Read Voltage Offsets for Accurate Page Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The correctness of data reading in NAND memories is affected by the variation of stored charges over time and with repeated read operations, leading to inaccuracies in data retrieval.
Innovation Solution
A memory device and system that utilize a peripheral circuit to determine a target valley voltage by adjusting read voltages based on the position of word lines and blank physical pages, using offset values and adjustments to ensure accurate data reading.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If read voltage is kept constant for simplicity, then device complexity is reduced, but data reading accuracy deteriorates due to charge variation over time and repeated reads
Solution Approach 1:
The patent applies preliminary action by pre-calculating and storing offset values in an offset table before actual read operations. The offset values, which compensate for charge loss, are determined in advance based on physical page positions and programming states, eliminating the need for complex real-time calculations during read operations and maintaining high reading accuracy.
Solution Approach 2:
The system implements self-service by automatically selecting and applying the appropriate offset value from the offset table based on the current physical page position and programming state. This automatic selection process eliminates manual intervention and ensures the correct compensation is applied without adding operational complexity to the read process.
2Measurement precision
If offset values are calculated in real-time based on word line position and blank page location, then data reading accuracy is improved, but processing time increases
Solution Approach 1:
The offset values are pre-calculated and stored in an offset table during manufacturing or initialization, rather than being calculated in real-time during read operations. This preliminary action eliminates time-consuming calculations during actual reads while maintaining accurate compensation for charge loss based on physical page positions and programming states.
Solution Approach 2:
The patent uses an offset table that stores pre-computed offset values as a reference copy of the compensation data. Instead of recalculating offsets during each read operation, the system copies and applies the appropriate pre-stored offset value from the table, significantly reducing processing time while maintaining accuracy.
3Measurement precision
If multiple offset values are used based on programming state and position, then reading accuracy is improved, but device complexity increases
Solution Approach 1:
The offset table is segmented into multiple entries based on different programming states (S1, S2, S3) and physical page positions. Each segment contains pre-calculated offset values specific to that condition, allowing the system to select the most accurate offset without requiring complex real-time analysis of the current state.
Solution Approach 2:
The offset table acts as an intermediary structure that pre-organizes all possible offset values according to programming states and physical positions. This intermediary table simplifies the complexity by providing a ready-reference lookup mechanism, eliminating the need for complex decision logic during read operations.
Data Source
AI summary
Examples of the present disclosure include a memory device including: a memory cell array including a plurality of blocks. The blocks include a plurality of word lines, and a plurality of memory cells coupled to the plurality of word lines. The plurality of memory cells coupled to a same word line form a physical page. A physical page includes one or more code words; and a peripheral circuit coupled to the memory cell array and configured to: acquire a predicted initial read voltage of the one or more code words according to a position of the word line coupled to the one or more code words in an open block and a position of a first blank physical page in the open block; and obtain a target valley voltage of the code words according to a first result corresponding to the code words at the predicted initial read voltage.