NAND Flash Threshold Estimation via Gaussian Intersection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
NAND flash devices experience increased read errors due to incorrect decision thresholds, leading to reading failures and data loss, and existing methods for estimating optimal thresholds are inefficient and time-consuming, causing significant latency and quality of service drops.
Innovation Solution
A method that reads a page with a default threshold, attempts hard decoding, and if failed, reads twice with an offset voltage for soft decoding, approximates the empirical distribution of successfully decoded bits using an iteratively reweighted least squares (IRLS) method to find the intersection of Gaussian distributions as the new threshold, and adjusts the read operation accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional threshold estimation methods (Soft Decoding, LRE) are used to improve threshold accuracy, then measurement precision improves, but loss of time increases significantly due to multiple additional read operations
Solution Approach 1:
The patent performs preliminary actions by using the default threshold read results and soft decoding outcomes to gather necessary information about bit distributions before attempting to determine the optimal threshold. This preliminary data collection avoids the need for multiple additional read operations required by conventional methods like LRE, thereby reducing time loss while maintaining measurement precision.
Solution Approach 2:
The patent extracts only the essential information needed for threshold estimation from the soft decoding results - specifically, the distributions of successfully decoded bits at different voltage levels. By extracting only this critical data rather than performing complete additional read operations, the method achieves accurate threshold estimation without the significant time penalty of conventional approaches.
2Measurement precision
If multiple additional read operations are performed to estimate optimal threshold, then measurement precision improves, but productivity decreases due to increased read latency
Solution Approach 1:
The patent utilizes preliminary read operations with default threshold and soft decoding to gather distribution information before threshold optimization. This preliminary action framework allows the system to achieve accurate threshold estimation without requiring the multiple additional read operations that would degrade productivity and quality of service.
Solution Approach 2:
The patent performs partial action by using only the information from default threshold reads and soft decoding results that is necessary for threshold estimation. It avoids the excessive action of performing multiple additional read operations at different thresholds, thereby maintaining productivity and quality of service while still achieving the needed measurement precision.
3Device complexity
If default threshold is used for all read operations, then device complexity is reduced, but measurement precision deteriorates leading to increased read errors
Solution Approach 1:
The patent implements self-service by using the default threshold read results and soft decoding outputs to automatically determine the optimal threshold for the current read conditions. This self-determined threshold adapts to varying conditions (block, wordline, wear level) without requiring complex external intervention or multiple predetermined threshold tables, thus maintaining low device complexity while improving measurement precision.
Solution Approach 2:
The patent introduces dynamics by making the read threshold adaptive rather than static. Instead of using a fixed default threshold for all operations, the system dynamically determines the optimal threshold based on real-time analysis of soft decoding results and bit distributions, thereby improving read accuracy without significantly increasing device complexity.
Data Source
AI summary
A method for determining an optimal threshold of a nonvolatile memory device, the method including: reading a page from a nonvolatile memory device with a default threshold and attempting to hard decode the page using the default threshold; reading the page two more times with a predetermined offset voltage when the hard decoding fails and attempting to soft decode the page using the default threshold; approximating an empirical distribution of successfully decoded bits with a Gaussian distribution for each level; finding an intersection of the Gaussian distributions; and setting the intersection as a new reading threshold and reading the page again with the new reading threshold.