Real-Time Scan Chain Blockage Detection Using Annotated Test Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing complexity of high-density semiconductor devices leads to more defects in scan chains, making it costly and challenging to identify and locate blockages in scan chains during testing, especially in smaller nanometer geometries, where conventional test methods are inefficient and require extensive offline analysis.
Innovation Solution
A software architecture that uses standard automated test pattern generators to analyze stuck-at patterns, categorize scan cells based on failure types, and compute the position of blockages in scan chains using a look-up table of 'Load-Capture-Unload' operations, allowing for real-time detection and localization of blockages without specialized test vectors or extensive data logging.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional test methods are used to identify blockages in scan chains, then test coverage is achieved, but extensive offline analysis and large volumes of data logging are required, reducing test efficiency and increasing costs
Solution Approach 1:
The patent applies preliminary action by pre-processing test patterns to annotate them with expected scan cell values before actual testing. This annotation includes marking which scan cells should be high or low for each pattern, enabling real-time comparison during testing without requiring post-test analysis. The look-up table is prepared in advance to map pattern indices to expected scan cell states, allowing immediate blockage detection when discrepancies are found.
Solution Approach 2:
The patent introduces an intermediary mechanism in the form of a look-up table that stores the relationship between test pattern indices and expected scan cell values. This intermediary structure acts as a reference database that the testing system queries in real-time to determine whether a blockage has occurred, eliminating the need for complex offline analysis while maintaining high identification accuracy.
2Productivity
If standard automated test pattern generators are used, then test pattern generation is efficient, but real-time blockage detection capability is lost without specialized test vectors
Solution Approach 1:
The patent applies universality by making standard automated test pattern generators multi-functional. The system takes standard test patterns as input and automatically annotates them with scan cell expected values, enabling both standard testing and real-time blockage detection using the same test patterns. This eliminates the need for specialized test vectors while maintaining real-time detection capability through the annotated pattern data and look-up table mechanism.
3Reliability
If extensive data logging is performed to capture all scan chain outputs, then complete defect information is obtained, but data volume becomes unmanageable and analysis becomes prohibitively complex
Solution Approach 1:
The patent applies the taking out principle by extracting only the critical information needed for blockage detection. Instead of logging all scan chain outputs, the system only logs the comparison result between actual and expected scan cell values for each test pattern. The look-up table provides the expected values, and only discrepancies (potential blockages) are recorded, dramatically reducing data volume while maintaining complete defect information for analysis.
4Reliability
If scan chains are made longer to increase test coverage in high-density devices, then more defects can be detected, but the probability of blockages increases and test complexity increases
Solution Approach 1:
The patent applies self-service by enabling the test system to automatically detect and locate blockages in real-time during testing. The annotated test patterns and look-up table allow the system to self-diagnose blockage locations without external intervention or complex post-test analysis. This automation handles the increased complexity of long scan chains, maintaining high test coverage while reducing the burden on test engineers.
Data Source
AI summary
A process for identifying the location of a break in a scan chain in real time as fail data is collected from a tester. Processing a test pattern before applying it on a tester provides a signature enabling a method for a tester to identify a scan cell which is stuck during the time the tester is operating on a device under test rather than accumulating voluminous test data sets for delayed offline analysis.