NAND Read Threshold Calibration Using Tree-Based Hardware Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The challenge of maintaining process uniformity in NAND process shrinking and three-dimensional stacking, coupled with the need to support a wide range of operational conditions, leads to significant variability in read thresholds for memory pages, resulting in increased bit error rates and performance degradation due to inaccurate reading.
Innovation Solution
A modular, tree-based hardware module for read threshold calibration that allows flexible operation of tree groups based on user requirements and system conditions, optimizing accuracy, latency, and power consumption through adaptive usage of tree-based inferencing models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If read threshold calibration is performed multiple times to improve accuracy, then reading accuracy improves, but latency and power consumption increase
Solution Approach 1:
The patent applies partial action by performing read threshold calibration a determined number of times rather than always performing full calibration. The system selectively executes calibration operations based on whether the read threshold falls within an expected range, performing calibration multiple times only when necessary to achieve accuracy while avoiding unnecessary calibration operations to reduce latency
2Measurement precision
If read threshold calibration is performed multiple times to improve accuracy, then reading accuracy improves, but power consumption increases
Solution Approach 1:
The system performs calibration operations a determined number of times based on whether accuracy requirements are met, avoiding excessive calibration operations that would waste energy. The selective execution of calibration based on expected range checks ensures power is consumed only when necessary to achieve the required read threshold accuracy
3Ease of operation
If fixed read thresholds are used to simplify operation, then device complexity reduces, but bit error rate increases due to variability under different operational conditions
Solution Approach 1:
The patent implements dynamic read threshold adjustment by performing calibration operations a determined number of times based on operational conditions. Instead of using fixed read thresholds, the system adaptively calibrates thresholds dynamically to account for variability under different program/erase cycles, retention times, and temperatures, thereby maintaining reliability while managing complexity through controlled calibration operations
Data Source
AI summary
A predictive read threshold calibration method is provided in which a dedicated hardware module is used for read threshold calibration. The hardware module uses a tree-based inferencing model to generate trees that produce respective inference results based on a plurality of read thresholds. An inferred read threshold is obtained by summing the inference results of the trees, and the memory is read using the inferred read threshold. Other embodiments are provided.