Non-adaptive Scan Chain Pattern Reordering for Diagnostic Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic test equipment (ATE) has limited fail buffer capacity, leading to restricted failure data collection during testing, which negatively impacts scan chain diagnostic resolution in volume production environments, where extensive failure logging is not feasible due to time constraints.
Innovation Solution
A non-adaptive pattern reordering method that computes penalty scores for test patterns based on subchain lengths without sensitive bits, reordering them to maximize diagnostic resolution without requiring adaptive testing flows or extensive ATE modifications.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If adaptive pattern reordering algorithms are used to improve scan chain diagnostic resolution, then diagnostic resolution is improved, but test complexity and ATE modification requirements increase
Solution Approach 1:
The patent pre-calculates and stores penalty scores for all test patterns before actual testing. This preliminary action eliminates the need for adaptive decision-making during testing, as the reordering is predetermined based on pre-computed metrics. The penalty score for each pattern is calculated based on the sum of squares of subchain lengths without sensitive bits, and this information is stored for direct application during test execution.
Solution Approach 2:
The patent creates a simplified copy of the diagnostic problem by using penalty scores as a proxy metric for diagnostic quality. Instead of implementing complex adaptive algorithms that evaluate multiple factors in real-time, the system uses a single pre-computed penalty score to represent the diagnostic value of each pattern, making the reordering process straightforward and compatible with standard ATE operations.
2Measurement precision
If more failure cycles are collected to improve diagnostic resolution, then diagnostic accuracy is improved, but test time increases
Solution Approach 1:
The patent replaces the mechanical approach of collecting more failure cycles with a computational approach using penalty scores. Instead of increasing the volume of failure data collection, the system uses pre-computed penalty scores to identify and prioritize the most diagnostic patterns, achieving high diagnostic accuracy with fewer patterns applied.
Solution Approach 2:
The patent changes the parameter being optimized from the quantity of failure cycles to the quality of patterns selected. By using penalty scores that measure the diagnostic information content of each pattern (based on subchain lengths without sensitive bits), the system prioritizes pattern quality over quantity, achieving better diagnostic resolution with fewer test cycles.
3Ease of operation
If non-adaptive pattern reordering is implemented to reduce test complexity, then ease of operation is improved, but diagnostic resolution may be compromised
Solution Approach 1:
The patent makes the testing system self-optimizing by pre-calculating penalty scores that automatically identify the most diagnostic patterns. The system serves itself by using the penalty score metric to determine the optimal pattern sequence without requiring external adaptive control or complex decision-making during testing. This self-service approach maintains high diagnostic resolution while simplifying the testing flow.
Data Source
AI summary
Systems and methods for re-ordering test patterns for circuit design or testing. A method includes receiving a set of scan chains and associated test patterns, and computing a penalty score for each test pattern in the set of test patterns. The method includes selecting a first pattern of the set of test patterns that has a lowest computed penalty score in the set of test patterns, and removing the first pattern from the set of test patterns and adding the first pattern to a set of ordered patterns. The method includes, for each remaining test pattern, computing an accumulated penalty score for each remaining pattern, selecting a next pattern of the set of test patterns that has a lowest accumulated penalty score in the set of test patterns, removing the next pattern from the set of test patterns, and adding the next pattern to the set of ordered patterns.