Memory Die Family Classification Using Ordered SCL Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory die classification methods based on slow charge loss (SCL) measurements lead to uneven and inaccurate distribution of memory dies into die families, resulting in under-utilization and over-utilization of some families, and are difficult to accurately classify due to measurement variations and limited sampling.
Innovation Solution
A method to classify memory dies into respective die families based on a threshold percentage of SCL values, using an ordered index of SCL values and a die family threshold level to ensure equal distribution of memory dies into balanced die families, facilitating improved Block Family Error Avoidance (BFEA) bin assignment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If memory dies are classified into die families based on SCL measurements, then die family assignment is achieved, but the distribution becomes uneven and inaccurate due to measurement variations and limited sampling
Solution Approach 1:
The patent implements feedback by measuring SCL values at multiple time points (T0, T1, T2) and using the change in SCL over time to adjust and refine die family assignments. This multi-point measurement approach provides feedback on measurement variations, allowing the system to compensate for limited sampling and achieve more accurate classification despite individual measurement uncertainties.
Solution Approach 2:
The patent performs preliminary SCL measurements at multiple time points before final die family assignment. By conducting preliminary actions (measurements at T0, T1, T2) and analyzing the trend of SCL changes, the system prepares more accurate classification data in advance, reducing the impact of measurement variations on the final assignment accuracy.
2Productivity
If memory dies are classified into multiple die families, then BFEA processing is improved, but some die families become under-utilized or over-utilized due to uneven distribution
Solution Approach 1:
The patent changes the parameter used for classification from single-point SCL values to the rate of change of SCL over time (ΔSCL/Δt). This parameter transformation allows for more balanced distribution across die families because it captures the temporal behavior of charge loss rather than just instantaneous values, thereby improving both BFEA processing efficiency and utilization balance simultaneously.
Solution Approach 2:
The patent introduces dynamic measurement at multiple time points (T0, T1, T2) to capture the temporal evolution of SCL characteristics. This dynamic approach allows the classification system to adapt to variations in charge loss behavior, resulting in more balanced die family assignments that improve both processing efficiency and operational balance.
3Loss of time
If limited sampling is used for SCL measurements, then measurement time is reduced, but classification accuracy deteriorates due to measurement variations
Solution Approach 1:
The patent transforms the measurement approach by focusing on the rate of change (ΔSCL/Δt) rather than absolute SCL values. This parameter change allows accurate classification with limited sampling because the rate of change is more stable and less susceptible to measurement noise, achieving good precision without requiring extensive sampling time.
Solution Approach 2:
The patent uses a moderate number of time points (T0, T1, T2) that provides sufficient information for accurate classification without excessive measurement time. This partial action approach achieves the necessary measurement precision for reliable die family assignment without the diminishing returns of continuous or excessive sampling.
Data Source
AI summary
Processing logic of a memory sub-system to measure a set of slow charge loss (SCL) values, wherein each SCL value of the set of slow charge loss values corresponds to a memory die of a set of memory dies of a memory sub-system. An ordered SCL index based on the set of SCL values is generated. Using a die family threshold level, a set of ranges of the ordered SCL index is identified. Each range of the set of ranges is assigned to a corresponding die family of a set of die families.