Multi-bit Bad Block Flag Classification for Non-volatile Memory Yield
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing semiconductor memory devices often discard or underutilize memory die due to defects, as they lack a comprehensive method to identify and utilize recoverable blocks, leading to reduced yield and inefficient resource allocation.
Innovation Solution
Implementing a multi-bit bad block flag system during the die sort process to classify and recover blocks with defects, allowing for flexible block access and utilization in memory systems, enabling the selection of blocks based on their recoverability during power-on sequences and wear leveling operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional binary bad block flag system is used, then manufacturing process is simple, but memory yield is reduced due to discarding recoverable blocks
Solution Approach 1:
The patent segments the binary bad block flag into multiple sub-flags (first sub-flag, second sub-flag, third sub-flag) that can be independently set based on different defect conditions. This segmentation allows for fine-grained classification of block quality, enabling the system to distinguish between permanently defective blocks and recoverable blocks, thereby improving memory yield without requiring complete system redesign
Solution Approach 2:
The patent transitions from a one-dimensional binary flag system (good/bad) to a multi-dimensional classification system using multiple sub-flags. This dimensional expansion creates a more nuanced state space for block classification, allowing the system to track different types of defects and recovery statuses simultaneously, thus resolving the contradiction between simplicity and yield improvement
2Productivity
If all defective blocks are discarded, then reliability is maintained, but resource utilization efficiency decreases
Solution Approach 1:
The patent applies local quality by treating different blocks with different flag states based on their specific defect characteristics. Instead of uniformly discarding all defective blocks, the system selectively marks blocks with appropriate sub-flag combinations, allowing recoverable blocks to be utilized while maintaining reliability by excluding truly defective blocks through the first sub-flag mechanism
Solution Approach 2:
The patent implements a selective discard and recovery strategy where blocks are not immediately discarded upon defect detection. Instead, the multi-sub-flag system enables identification of recoverable blocks (marked by second and/or third sub-flags) that can be restored through repair processes, while only blocks with critical defects (first sub-flag) are permanently discarded, thus improving resource utilization without compromising reliability
3Measurement precision
If multi-bit bad block flag system is implemented, then block classification precision is improved, but manufacturing complexity increases
Solution Approach 1:
The patent segments the classification task into distinct sub-flags, each responsible for a specific aspect of block quality assessment. This segmentation simplifies the manufacturing process by providing a modular approach where each sub-flag can be set based on specific test results, reducing the cognitive and procedural complexity compared to a monolithic multi-bit classification system
Solution Approach 2:
The patent performs preliminary classification actions during the die sort process by setting specific sub-flags based on test results before final block utilization decisions are made. This preliminary action allows for early identification of recoverable blocks, enabling repair and retesting processes to be planned and executed efficiently, thereby reducing overall manufacturing complexity despite the enhanced classification capability
Data Source
AI summary
For a non-volatile memory die formed of multiple blocks of memory cells, the memory die has a multi-bit bad block flag for each block stored on the memory die, such as in a fuse ROM. For each block, the multi-bit flag indicates if the block has few defects and is of the highest reliability category, is too defective to be used, or is in of one of multiple recoverability categories. The multi-bit bad blocks values can be determined as part a test process on fresh devices, where the test of a block can be fail stop for critical category errors, but, for recoverable categories, the test continues and tracks the test results to determine a recoverability category for the block and write this onto the die as a bad block flag for each block. These recoverability categories can be incorporated into wear leveling operations.