Memory Controller Read Voltage Prediction via Kalman Filter
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional memory systems face challenges in maintaining accurate data retrieval due to changes in threshold voltage over time, leading to bit errors and increased read latency, as they struggle to balance observation frequency with channel occupancy and followability of read voltage changes.
Innovation Solution
The memory system employs a management information update process using a Kalman filter algorithm or recursive least square algorithm to estimate and predict read voltage values, allowing for reduced observation frequency while maintaining followability to threshold voltage changes, by updating management information without relying on current observation values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If observation frequency of read voltage is increased to maintain accuracy, then data retrieval accuracy is improved, but channel occupancy increases and read latency increases
Solution Approach 1:
The system performs preliminary observation of the optimum read voltage at predetermined time points (e.g., immediately after program operation) and stores this information in advance. During normal read operations, the stored read voltage information is reused without requiring frequent new observations, thus reducing channel occupancy and read latency while maintaining data retrieval accuracy.
Solution Approach 2:
The system implements a feedback mechanism where the observed optimum read voltage at one time point is used to update the read voltage information for subsequent time points. The memory controller adjusts the read voltage based on feedback from previous observations and threshold voltage change predictions, enabling accurate data retrieval without frequent observations.
2Productivity
If observation frequency of read voltage is decreased to reduce channel occupancy, then channel occupancy is reduced, but followability to threshold voltage changes deteriorates
Solution Approach 1:
The system dynamically adjusts the read voltage based on the elapsed time since program operation and predicted threshold voltage changes. Instead of using a fixed observation schedule, the system adapts the read voltage in real-time by combining stored observation data with time-based predictions of threshold voltage drift, maintaining followability while reducing observation frequency.
Solution Approach 2:
The system performs preliminary observation at critical time points (immediately after programming) when threshold voltage changes are most significant. These preliminary observations capture the initial state, and subsequent read operations use time-based predictions to track threshold voltage changes between observations, reducing the need for frequent observations while maintaining reliability.
3Reliability
If read voltage is updated at every time point using current observation, then followability to threshold voltage changes is improved, but observation frequency must be increased
Solution Approach 1:
The system performs read voltage observation periodically at predetermined time points (e.g., immediately after program operation and at specific intervals) rather than continuously or at every read operation. Between observation points, the system uses time-based predictions to maintain followability, reducing observation frequency while preserving reliability.
Solution Approach 2:
The system introduces an intermediary mechanism (time-based prediction model) that bridges the gap between discrete observation points. This intermediary uses the elapsed time since program operation and stored observation data to predict threshold voltage changes, enabling the system to maintain followability without requiring observations at every time point.
Data Source
AI summary
According to one embodiment, in a memory system, a memory controller is configured to execute a first operation of observing an optimum value of a read voltage and updating a set value based on the observation result of the optimum value, at a predetermined time point of a plurality of time points for updating the set value of the read voltage for a plurality of memory cells, and execute a second operation of updating the set value based on the set value updated at one previous time point without executing the observation of the optimum value, at a time point after one time point of the predetermined time point.