Flash Memory Read Decoding Using A Priori Probability Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The performance and reliability of multi-bit flash memory cells decrease due to overlapping threshold voltage distributions, leading to read errors, especially as the number of threshold voltages increases, and existing memory systems do not effectively utilize prior probability information to improve data reading accuracy.
Innovation Solution
A memory system that uses a priori probability information to read stored data by calculating a posteriori probabilities based on state information stored as meta-data, performing a soft-decision operation to detect logical states, and employing a memory controller with an a priori probability computation unit, an a posteriori probability computation unit, and a soft-decision decoder to determine bit values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multiple threshold voltage distributions are used to store multiple bits of data, then storage capacity increases, but threshold voltage distributions overlap causing read errors and reliability decreases
Solution Approach 1:
The patent applies preliminary action by calculating and storing a priori probability information about the distribution of data across different threshold voltage distributions before the read operation. This pre-computed probability information is then used during reading to weight the likelihood of each threshold voltage distribution, allowing the system to compensate for overlapping distributions and improve reading accuracy without changing the multi-bit storage structure itself.
2Productivity
If the number of threshold voltages increases to store more bits, then storage density improves, but threshold voltage distributions become overlapped leading to performance degradation
Solution Approach 1:
The patent implements feedback by using the stored a priori probability information to adjust the interpretation of read values. When reading data, the system references the pre-stored probability distribution information to determine the most likely original data value, creating a feedback loop that compensates for the degradation caused by overlapping threshold voltage distributions and maintains performance despite increased storage density.
3Measurement precision
If a priori probability information is calculated and used in the reading process, then data reading accuracy improves, but computational complexity increases
Solution Approach 1:
The patent reduces computational complexity during reading by performing the complex probability calculations in advance during write operations or initialization. The a priori probability information is pre-computed and stored with the data, so that during reading, the system only needs to perform simpler comparisons and lookups based on this pre-prepared information, rather than performing complex calculations in real-time.
Data Source
AI summary
A memory system comprises a non-volatile memory device that stores user data and state information regarding the user data. In a read operation of the non-volatile memory device, a memory controller calculates a priori probabilities for the user data based on the state information, calculates a posteriori probabilities based on the a priori probabilities, and performs a soft-decision operation to determine values of the user data based on the a posteriori probabilities.