Resistance-Based Memory Transfer Function for Error Rate Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Resistance shifts and increased noise levels in resistance-based memory cells lead to bit errors and decreased reliability, especially when storing multiple bits, due to manufacturing variations, environmental factors, and usage-induced changes.
Innovation Solution
A transfer function is used to predict and estimate resistance shifts and noise variance based on workload indicators such as retention time, number of cycles, temperature, and neighboring data, allowing for the adjustment of threshold resistances during program and read operations to reduce error rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If resistance-based memory cells are used to store multiple bits, then storage capacity is improved, but resistance shifts and noise levels increase leading to higher bit error rates
Solution Approach 1:
The system performs preliminary characterization of memory cells during manufacturing and before use, measuring their initial resistance values and workload indicator responses. This advance information is stored and used to predict future resistance shifts, allowing the system to proactively compensate for expected changes rather than reacting to errors after they occur.
Solution Approach 2:
The system continuously monitors workload indicators (read cycles, write cycles, retention time, temperature) and uses this feedback to dynamically adjust read thresholds. The transfer function processes current workload data to predict resistance shifts, and the system compensates by adjusting thresholds in real-time, creating a closed-loop control system that maintains reliability despite changing conditions.
2Device complexity
If fixed threshold resistances are used for program and read operations, then device complexity is reduced, but resistance shifts cause increased error rates
Solution Approach 1:
Each memory cell effectively manages its own threshold requirements through the transfer function, which uses the cell's specific workload history and characteristics to predict its resistance shifts. The system autonomously determines optimal thresholds for each cell based on its unique degradation pattern, eliminating the need for complex external calibration or manual threshold adjustment.
Solution Approach 2:
The system dynamically changes the threshold parameter based on predicted resistance shifts. Instead of using fixed thresholds, the transfer function calculates optimal threshold values by considering workload indicators and predicted resistance changes, allowing the threshold parameter to adapt to changing memory cell conditions while maintaining simple operational procedures.
3Reliability
If transfer functions are used to predict resistance shifts, then error rates are reduced, but measurement and characterization requirements increase
Solution Approach 1:
The transfer function serves multiple purposes: it characterizes memory cells during manufacturing, predicts resistance shifts during operation, and determines optimal read thresholds. This multi-functional approach consolidates what could be separate complex systems into a single unified model, reducing overall measurement and characterization burden while maintaining high reliability.
Data Source
AI summary
Two or more workload indicators affecting a memory cell of a resistance-based, non-volatile memory are measured. The two or more workload indicators are applied to a transfer function that predicts a resistance shift and/or resistance noise variance in response to the two or more workload indicators. A result of the transfer function is applied to shift and/or determine a threshold resistance used for at least one of a program operation and a read operation affecting the memory cell. An error rate of the memory cell is reduced as a result.