Z-Score Memory Block Screening for Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional memory arrays face challenges in detecting and preemptively handling defect-prone blocks, which can lead to data loss and increased manufacturing costs due to stricter defect per million (DPPM) requirements and the slow process of addressing NAND defects.
Innovation Solution
A metric-based approach using Z-scores is employed to track and analyze the time of memory operations such as programming and erasing, allowing for the identification and handling of defect-prone blocks before they fail, by comparing generated Z-scores to thresholds and updating statistics for future calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional memory arrays are used without metric-based screening, then manufacturing process is simpler, but defect-prone blocks cannot be detected preemptively leading to data loss
Solution Approach 1:
The patent performs preliminary screening of memory blocks by tracking metrics during programming and erasing operations before defects manifest. Z-scores are calculated and compared against thresholds to identify defect-prone blocks in advance, allowing preemptive handling before data loss occurs.
Solution Approach 2:
The system continuously monitors memory operation metrics and provides feedback through Z-score calculations. When metrics deviate from expected ranges, the feedback mechanism triggers identification of defect-prone blocks, enabling dynamic adjustment and screening based on actual performance data.
2Manufacturing precision
If stricter DPPM requirements are implemented, then manufacturing precision improves, but processing time increases due to slow defect addressing
Solution Approach 1:
The patent replaces slow traditional defect detection methods with a metric-based statistical system using Z-scores. This substitution enables rapid identification of defect-prone blocks by comparing tracked metrics against thresholds, significantly accelerating the screening process while maintaining strict DPPM requirements.
Solution Approach 2:
The system changes the approach from reactive defect handling to proactive parameter-based screening. By tracking metrics such as programming and erasing times and comparing them against statistically determined thresholds, the system rapidly identifies defect-prone blocks without slowing down manufacturing processes.
3Measurement precision
If metric tracking and Z-score calculation are implemented, then defect detection accuracy improves, but computational overhead increases
Solution Approach 1:
The patent applies partial action by calculating Z-scores and comparing metrics against thresholds only when necessary during programming and erasing operations. This selective approach provides sufficient detection accuracy without continuously processing all memory blocks, thereby reducing unnecessary computational overhead while maintaining defect detection precision.
Data Source
AI summary
A variety of applications can include apparatus and/or methods to preemptively detect defect prone memory blocks in a memory device and handle these memory blocks before they fail and trigger a data loss event. Metrics based on memory operations can be used to facilitate the examination of the memory blocks. One or more metrics associated with a memory operation on a block of memory can be tracked and a Z-score for each metric can be generated. In response to a comparison of a Z-score for a metric to a Z-score threshold for the metric, operations can be performed to control possible retirement of the memory block beginning with the comparison. Additional apparatus, systems, and methods are disclosed.