3D Nonvolatile Memory Sampling for Post-Write Read Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Three-dimensional nonvolatile memory arrays face challenges in efficiently and accurately storing data due to higher error rates in higher density formats, necessitating a method to quickly determine data integrity without extensive read and compare operations.
Innovation Solution
A 3-D nonvolatile memory array operates by writing data in both high-density (MLC) and low-density (SLC) formats, sampling data from all word lines and a sample word line of each string to determine if the high-density data meets a standard, allowing for the discarding of lower error rate data if the sampled data is acceptable, and using the lower error rate data as a backup to rewrite the high-density data if it does not meet the standard.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is written in high-density format (MLC) to increase storage capacity, then storage density is improved, but error rate increases
Solution Approach 1:
The patent creates a backup copy of data written in high-density MLC format by writing the same data in low-density SLC format to a separate block. This backup copy serves as a reference for verification, allowing the system to detect and correct errors in the primary MLC data without sacrificing the high storage density benefit.
Solution Approach 2:
The system uses a temporary backup block in SLC format that can be discarded after verification. If the MLC data passes verification, the SLC backup is erased and the MLC block is marked as valid. This disposable backup mechanism provides error checking capability without permanently occupying additional high-density storage space.
2Measurement precision
If all data in a block is read and compared to verify integrity, then measurement precision is improved, but read time increases
Solution Approach 1:
Instead of reading and comparing all data in a block simultaneously, the patent divides the verification process into segments by reading only specific sample data points from the MLC block and comparing them against the corresponding SLC backup. This segmentation dramatically reduces the amount of data that must be read while still providing sufficient verification confidence.
Solution Approach 2:
The system performs partial verification by reading only a sample of data points rather than the complete block. This partial action is sufficient to detect likely failure modes, providing adequate measurement precision for verification purposes without the time cost of reading every single data point in the block.
3Reliability
If sampling pattern reads all word lines of a string and at least one word line of each string, then reliability is improved, but read time increases
Solution Approach 1:
The sampling pattern segments the block into strings and word lines, reading only specific representative samples from each segment rather than the entire block. By reading all word lines of a sample string and at least one word line from each string, the system achieves reliable failure mode detection while reducing read time through strategic sampling.
Solution Approach 2:
The verification process applies local quality by focusing sampling efforts on specific critical locations (all word lines of one string and at least one word line from each string) rather than uniformly reading the entire block. This targeted approach ensures reliable detection of likely failure modes while minimizing overall read time.
Data Source
AI summary
Data that is stored in a higher error rate format in a 3-D nonvolatile memory is backed up in a lower error rate format. Later, the higher error rate copy is sampled to determine if it is acceptable. A sampling pattern samples all word lines of a string and at least one word line of each string of the block.