Memory Error Correction Using Physical Bit Reliability Groups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing error correction methods for memory storage face challenges in accurately identifying and correcting errors due to physical property drift and variations in memory cells, leading to unreliable data retrieval.
Innovation Solution
An apparatus and method that utilize an array of memory cells and a controller to read encoded bits, divide them into reliability groups based on persistent physical characteristics, and provide these estimates to a soft decision decoder for improved error correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional error correction methods are used without considering physical characteristics, then the decoding process is simpler, but the accuracy of error identification and correction deteriorates
Solution Approach 1:
The system performs preliminary assessment of bit reliability based on physical characteristics (such as resistance values in resistive memory) before the decoding process. This preliminary action groups bits into reliability categories, providing the decoder with advance information about which bits are more likely to be erroneous, thereby improving error identification accuracy without significantly increasing decoding complexity
Solution Approach 2:
The patent introduces an intermediary reliability assessment mechanism that bridges the physical memory characteristics and the logical decoding process. This intermediary layer translates physical properties (like resistance measurements) into reliability estimates that guide the decoder, improving accuracy while maintaining a clear separation between physical measurement and logical correction functions
2Reliability
If memory cells are monitored continuously for physical property drift, then data reliability improves, but the time and resources required for monitoring and recalibration increase
Solution Approach 1:
The system implements periodic monitoring of physical characteristics at predetermined intervals or after specific numbers of write cycles, rather than continuous monitoring. This periodic approach maintains data reliability by detecting drift at appropriate intervals while minimizing the time and resources spent on monitoring, as the memory can operate normally between monitoring events
Solution Approach 2:
The patent monitors changes in physical parameters (such as resistance values) over time and triggers recalibration only when these parameters exceed predetermined thresholds. This approach improves reliability by detecting actual drift conditions while avoiding unnecessary recalibrations, thereby reducing the time and resources wasted on frequent recalibration operations
Data Source
AI summary
Apparatuses, systems, and methods are presented for error correction based on physical characteristics for memory. A controller may be configured to read a set of encoded bits from a set of cells of a memory array. The controller may be configured to divide the encoded bits into reliability groups based on one or more persistent physical characteristics associated with cells of the set of cells. The controller may be configured to provide reliability estimates based on the reliability groups to a soft decision decoder for decoding the encoded bits.