BISR Register Grouping for Memory Repair Collision Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current high-density semiconductor memory repair techniques are inefficient due to time-consuming and storage-intensive processes for storing and retrieving repair information in built-in self-repair (BISR) chains, which require a large amount of storage and increase circuit complexity.
Innovation Solution
The method involves organizing BISR registers into groups of identical length, performing collision detection, merging repair information, and storing and broadcasting the merged information in parallel, which can be compressed and decompressed as needed, to minimize repair time and storage requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If repair information is stored in BISR chains using conventional serial transfer methods, then the repair information can be permanently stored, but the process is time-consuming and inefficient
Solution Approach 1:
The patent divides the BISR chain into multiple groups, where each group contains multiple chains. Repair information is processed and merged within each group independently, allowing parallel operations across groups. This segmentation enables the repair process to be broken down into manageable units that can be handled simultaneously, reducing overall processing time while maintaining storage reliability.
Solution Approach 2:
The patent merges repair information from multiple chains within each group before storing it in the BISR chain. By combining repair data from parallel chains into a single merged result, the system reduces the total number of storage operations required and enables parallel processing of multiple repair streams, significantly decreasing the time needed to complete the repair process.
2Reliability
If conventional BISR chain methods are used to store repair information, then repair can be performed, but a large amount of storage is required
Solution Approach 1:
By organizing BISR chains into multiple groups with fewer chains per group, the patent reduces the storage burden on each individual chain. The segmentation allows repair information to be distributed across groups and merged efficiently, requiring less total storage capacity per chain while maintaining overall repair capability.
Solution Approach 2:
The merging of repair information from multiple chains within each group consolidates redundant data, reducing the total storage requirements. By combining repair streams before final storage, the system eliminates duplicate information and achieves the same repair reliability with reduced storage capacity per chain.
3Reliability
If repair information is serially transferred through BISR chains, then the information can be stored, but the circuit complexity increases
Solution Approach 1:
The patent segments the BISR chain into multiple independent groups that can be processed in parallel. This segmentation simplifies the circuit design within each group by reducing the number of chains that must be managed individually, while the parallel group structure maintains overall repair reliability without requiring complex inter-chain coordination circuits.
Solution Approach 2:
By merging repair information within each group before output, the patent reduces the complexity of information delivery. The merge operation consolidates multiple data streams into a single output per group, simplifying the circuit architecture compared to managing multiple independent serial transfer paths while preserving complete repair information delivery.
Data Source
AI summary
Systems and methods for repairing a memory. A method includes performing a repair analysis of the embedded memories to produce repair information. The method includes storing the repair information in the registers, where the registers are organized into groups having chains of identical length. The method includes performing collision detection between the repair information in each of the groups. The method includes merging the repair information in each of the groups. The method includes repairing the embedded memories using the merged repair information.