Data Storage Device Unreliable Bit Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Data storage devices face challenges in accurately identifying and correcting error bits in nonvolatile memory cells due to variations in threshold voltage distributions and interference between adjacent memory cells, leading to imperfect error correction operations.
Innovation Solution
A method involving a controller that reads data chunks from target memory cells based on read biases, obtains discrimination data using discrimination biases, and determines unreliable bits by comparing them to adjacent data patterns, with the option to use parity data for error correction, thereby minimizing the likelihood of incorrect bit identification as an error bit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If error correction operations are performed on all bits in data chunks, then the reliability of data storage is improved, but the complexity of the error correction process increases and the likelihood of incorrect bit identification increases
Solution Approach 1:
The patent segments the data storage system into multiple data chunks, each corresponding to a specific page in the nonvolatile memory. By dividing the error correction process into chunk-level operations rather than processing all bits uniformly, the system reduces overall complexity while maintaining reliability through localized error correction strategies.
Solution Approach 2:
The patent applies local quality by determining unreliable bits specifically within target data chunks based on discrimination data from corresponding memory cells. Instead of applying uniform error correction to all data, the system identifies and corrects errors locally in specific chunks where unreliable bits are detected, reducing unnecessary processing complexity.
2Measurement precision
If discrimination data is obtained for all memory cells, then the accuracy of unreliable bit identification is improved, but the time and computational resources required increase
Solution Approach 1:
The patent performs preliminary action by obtaining discrimination data for memory cells before executing the full error correction process. This preliminary discrimination data is used to identify unreliable bits in advance, allowing the system to focus computational resources only on chunks containing unreliable bits rather than processing all data uniformly.
Solution Approach 2:
The patent applies partial action by obtaining discrimination data selectively for target memory cells corresponding to specific data chunks, rather than for all memory cells in the storage device. This partial approach achieves sufficient identification accuracy for the current error correction task while reducing overall time and computational resource requirements.
3Reliability
If multiple read biases are applied to read data chunks, then the reliability of data retrieval is improved, but the number of read operations and time required increase
Solution Approach 1:
The patent segments the data retrieval process by applying multiple read biases to read multiple data chunks from different pages corresponding to the same word line. This segmentation allows parallel or sequential reading of chunks with different reliability characteristics, improving overall data retrieval reliability while managing read operation complexity.
Solution Approach 2:
The patent changes the parameter of read bias voltage to retrieve data chunks with different reliability levels. By varying the read bias parameter across multiple read operations, the system can identify and correct unreliable bits more effectively, trading some read time for improved data retrieval reliability.
Data Source
AI summary
A method for operating a data storage device includes reading a plurality of data chunks from a plurality of pages corresponding to target memory cells coupled to a target word line based on read biases; obtaining discrimination data corresponding to the target memory cells based on discrimination biases; determining an unreliable bit in a target data chunk among the plurality of data chunks based on the plurality of data chunks and the discrimination data; and determining whether the unreliable bit is an error bit.