Codeword Detection in Multi-Level Memory Under Signal Drift
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing read-detection techniques in solid-state storage devices face challenges with drift in physical quantities measured during cell readout, such as electrical resistance in PCM devices, which affects the accuracy of detecting codewords stored in multi-level cells, particularly due to stochastic drift over time and varying cell levels.
Innovation Solution
A method and apparatus that involve obtaining read signals from memory cells storing N-symbol codewords, ordering and averaging these signals to determine reference signal levels based on predefined probabilities of symbol values, allowing for accurate detection of codewords by correlating the average read signal with predefined probabilities to estimate current reference signal levels, thereby adapting to drift without requiring additional reference cells.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If reference-cell techniques are used to compensate for drift, then detection accuracy is improved, but device complexity and memory capacity loss increase
Solution Approach 1:
The system uses the user cells themselves to generate reference signal levels through statistical processing of read signals, eliminating the need for separate reference cells. The controller performs self-calibration by computing average signal levels from multiple reads of the same data, thereby serving its own reference needs without external assistance.
Solution Approach 2:
User cells are made to serve dual purposes: storing user data and simultaneously providing reference signal levels for drift compensation. The same memory cells that hold information also generate the reference levels needed for accurate detection, eliminating the need for dedicated reference cells and reducing overall system complexity.
2Measurement precision
If reference-cell techniques are used to compensate for drift, then detection accuracy is improved, but memory capacity is reduced
Solution Approach 1:
User cells are made to serve dual purposes: storing user data and simultaneously providing reference signal levels for drift compensation. The same memory cells that hold information also generate the reference levels needed for accurate detection, eliminating the need for dedicated reference cells and reducing overall system complexity.
Solution Approach 2:
The system uses the user cells themselves to generate reference signal levels through statistical processing of read signals, eliminating the need for separate reference cells. The controller performs self-calibration by computing average signal levels from multiple reads of the same data, thereby serving its own reference needs without external assistance.
3Measurement precision
If reference-cell techniques are used to compensate for drift, then detection accuracy is improved, but readout latency increases
Solution Approach 1:
The system performs preliminary statistical analysis of read signals to establish reference levels that can be reused for subsequent detections. By computing average signal levels from multiple reads and storing these reference values, the system reduces the need for repeated full calibration sequences, thereby lowering latency in ongoing operations.
4Measurement precision
If model-based drift cancellation is used, then detection accuracy is improved, but implementation difficulty increases
Solution Approach 1:
The system uses the user cells themselves to generate reference signal levels through statistical processing of read signals, eliminating the need for separate reference cells. The controller performs self-calibration by computing average signal levels from multiple reads of the same data, thereby serving its own reference needs without external assistance.
Solution Approach 2:
The system changes the approach from using physical reference cells to using statistical parameters (average signal levels) derived from user cell reads. This parameter transformation simplifies the system architecture while maintaining drift compensation effectiveness.
Data Source
AI summary
A method for detecting codewords in solid-state storage devices. The method includes the steps of: obtaining respective read signals by reading memory cells that stores a group of codewords, where each of the read signals includes N signal components corresponding to respective symbols of the codeword; producing an ordered read signal by ordering the components of each of the read signals according to a signal level; producing an average read signal by averaging corresponding components of the ordered read signals; determining a reference signal level that corresponds to each of q levels of the memory cells in relation to the average read signal with predefined probabilities of each symbol value occurring at each symbol position in the codeword, where the symbols of the codeword are ordered according to the symbol value; and detecting the codeword corresponding to each of the read signal in relation to the reference signal levels.