Multi-bit Error Correction in Multi-level Memory Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-density semiconductor memory systems face challenges in maintaining data integrity due to environmental fluctuations, such as temperature and power variations, which affect the characteristic parameters used for storing multiple bits in a single memory cell, leading to data distortion and errors.
Innovation Solution
A method is introduced where memory cells are divided into characteristic parameter bands, with binary multi-bit values assigned to each band. The system tracks and compares stored and retrieved counts of these values, using statistical methods and probability distributions to identify and correct errors by adjusting the characteristic parameter values within adjacent bands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If multiple bits are stored in a single memory cell using multi-level cell technology, then storage density is improved, but data integrity deteriorates due to environmental fluctuations affecting characteristic parameters
Solution Approach 1:
The patent segments the characteristic parameter spectrum into multiple non-overlapping bands, with each band representing a specific multi-bit value. This segmentation allows the system to distinguish between different stored values even when environmental fluctuations cause parameter drift, as long as the drift does not cause crossing between bands. The segmentation approach resolves the contradiction by creating distinct regions that maintain reliability while enabling high storage density.
Solution Approach 2:
The patent performs preliminary error detection and correction by comparing retrieved characteristic parameter values against expected band ranges before final data interpretation. Count values are pre-calculated and stored for each band, and during retrieval, the system checks whether retrieved parameters fall within expected bands and adjusts counts accordingly. This preliminary action prevents corrupted data from being misinterpreted, maintaining data integrity despite environmental fluctuations.
2Quantity of substance
If the spectrum of characteristic parameter variation is divided into more bands to store more bits per cell, then storage capacity is improved, but measurement precision deteriorates due to narrower band widths
Solution Approach 1:
The patent creates a virtual copy of the characteristic parameter distribution by calculating and storing count values for each band based on initial measurements. These count copies are then used during retrieval to determine the most likely stored value without requiring precise direct measurement. The system compares retrieved parameters against all band ranges and uses the pre-calculated counts to resolve ambiguities, effectively copying the reference distribution for comparison purposes.
Solution Approach 2:
The patent changes the approach from direct parameter measurement to parameter range comparison. Instead of measuring the exact characteristic parameter value and mapping it directly to a bit value, the system compares the retrieved parameter against multiple band ranges and uses statistical count information to determine the most probable original value. This parameter change approach allows narrower bands to be used while maintaining discrimination accuracy through comparative analysis.
3Quantity of substance
If characteristic parameter bands are made narrower to increase the number of storable bits, then storage density is improved, but device complexity increases due to more frequent errors from environmental fluctuations
Solution Approach 1:
The patent implements a self-service error correction mechanism where the system uses its own stored count information to detect and correct errors. By comparing retrieved characteristic parameter values against expected band ranges and adjusting counts based on retrieval results, the system automatically compensates for errors caused by environmental fluctuations. This self-service approach reduces the need for external correction mechanisms while maintaining reliability with narrow bands.
Solution Approach 2:
The patent establishes a feedback loop where retrieval results are used to update and refine the count information for each band. The system compares retrieved values against stored counts, identifies discrepancies caused by environmental fluctuations, and adjusts the interpretation accordingly. This feedback mechanism allows the system to adapt to drift and maintain accuracy even with narrow bands, reducing the complexity burden of error correction.
Data Source
AI summary
A method, system, and computer software product for operating a memory cell collection. Memory cells in the collection store binary multi-bit values delimited by characteristic parameter bands of a characteristic parameter. In one embodiment, a comparing unit compares a retrieved count and a stored count for each binary multi-bit value. The retrieved count, equal to the number of occurrences the binary multi-bit value, is retrieved from the memory cell collection. The stored count, equal to the number of occurrences the binary multi-bit value, is stored in the memory cell collection. An error correction unit then assigns the error memory cell(s) a corrected binary multi-bit value with the characteristic parameter value within the characteristic parameter band adjacent to the characteristic parameter band associated with the retrieved binary multi-bit value such that the retrieved count of each binary multi-bit value is equal to the stored count of each binary multi-bit value.