Dynamic Fault Clustering for Semiconductor Memory BIRA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fault detection and clustering methods in semiconductor memory cells are inefficient, leading to increased test costs and time consumption due to the high integration of semiconductor memory products, particularly in determining appropriate redundancy for fault repair.
Innovation Solution
A dynamic fault clustering method and apparatus that allows for simultaneous fault detection and clustering by checking address matches and determining redundancy requirements on the fly, using a redundancy storage device and address storage device to manage fault mapping and repair.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If static fault clustering is used (detect all faults first, then cluster), then clustering accuracy is improved, but test time and productivity are worsened
Solution Approach 1:
The patent performs preliminary clustering actions during the fault detection process itself, rather than waiting until all faults are detected. The clustering algorithm continuously updates cluster assignments as faults are detected, making use of partial information to guide redundancy allocation earlier in the test process, thus reducing overall test time while maintaining reasonable clustering accuracy
Solution Approach 2:
The patent implements dynamic fault clustering where cluster assignments are continuously updated and refined as new faults are detected during testing. The clustering algorithm adapts to newly detected faults in real-time, adjusting cluster boundaries and redundancy allocations dynamically rather than requiring a static, complete fault map beforehand
2Productivity
If dynamic fault clustering is used (cluster during detection), then test time is reduced, but clustering accuracy and repair rate are worsened
Solution Approach 1:
The patent incorporates feedback mechanisms where the clustering algorithm continuously receives information about detected faults and their characteristics, then updates cluster assignments and redundancy allocations based on this feedback. The system learns from each detected fault and refines its clustering decisions, improving accuracy over time even as testing progresses
Solution Approach 2:
The patent performs preliminary clustering actions during the fault detection process itself, using partial fault information to make initial clustering decisions that guide redundancy allocation. These preliminary clustering actions are later refined as more fault information becomes available, allowing the system to achieve reasonable clustering accuracy without waiting for complete fault detection
3Adaptability or versatility
If redundancy allocation is delayed until all faults are detected, then redundancy utilization is improved, but repair time is worsened
Solution Approach 1:
The patent performs preliminary redundancy allocation during the fault detection process rather than delaying until all faults are detected. It uses the information from detected faults to make preliminary redundancy assignments that can be executed immediately, while continuing to refine these assignments as more faults are detected, thus reducing repair time without completely sacrificing redundancy utilization efficiency
Solution Approach 2:
The patent implements dynamic redundancy allocation where redundancy assignments are continuously updated and refined as new faults are detected during testing. The system transitions from static, post-detection allocation to a dynamic process that adapts redundancy assignments in real-time, balancing the trade-off between early repair capability and optimal redundancy utilization
Data Source
AI summary
A dynamic fault clustering method and apparatus for efficiently managing redundancy in semiconductor memories performs a collection operation of searching for and detecting a fault and an operation of appropriately clustering the fault at the same time, which reduces an amount of time spent performing Built-In Redundancy Analysis (BIRA).