Adaptive Semiconductor Test Limit Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The semiconductor industry faces challenges in managing and storing large amounts of test data from automated test equipment, leading to increased costs and resource requirements, often resulting in incomplete data analysis due to storage and time constraints.
Innovation Solution
A system and method for tracking production test data in real-time, dynamically adjusting 'PASS/FAIL' criteria based on statistical data distribution, marking units outside the normal distribution as defective, and dynamically updating pass/fail limits to improve product reliability and reduce defects per million (DPM) rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all test data is stored and analyzed, then product reliability improves through comprehensive defect detection, but storage costs and data management complexity increase significantly
Solution Approach 1:
The patent extracts only the essential statistical characteristics (mean, standard deviation, pass/fail counts) from the complete test data set, storing only these summarized parameters rather than all raw measurement data. This extraction principle reduces storage requirements while preserving the information needed for defect detection and product reliability assessment.
Solution Approach 2:
Instead of storing all data and then analyzing it, the patent inverts the approach by immediately computing statistical summaries during the testing process and storing only these condensed results. This inversion transforms a storage-intensive post-processing approach into an efficient real-time data reduction methodology.
2Measurement precision
If statistical analysis is performed on all test data, then defect detection accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The patent performs preliminary statistical computations (mean, standard deviation) during the testing process itself rather than after all data is collected. This preliminary action allows defect detection to occur in real-time with minimal post-processing, reducing overall processing time while maintaining detection accuracy.
Solution Approach 2:
The patent changes the form of data representation from raw individual measurements to aggregated statistical parameters (mean, standard deviation, pass/fail counts). This parameter transformation reduces computational complexity while preserving the essential information needed for accurate defect detection.
3Reliability
If dynamic pass/fail limits are adjusted based on real-time statistical data, then defect detection capability improves, but system complexity increases
Solution Approach 1:
The patent implements feedback by continuously monitoring statistical parameters during testing and using this information to dynamically adjust pass/fail limits. This feedback mechanism enables the system to adapt to process variations and improve defect detection capability while maintaining manageable complexity through automated statistical computations.
Solution Approach 2:
The patent transitions from static predetermined pass/fail limits to dynamic limits that automatically adjust based on real-time statistical analysis of test data. This dynamics principle allows the system to adapt to manufacturing process variations, improving defect detection while the automation keeps complexity manageable.
4Loss of information
If engineering teams manually analyze all test data, then comprehensive quality assessment is achieved, but labor costs and analysis time increase
Solution Approach 1:
The patent enables the testing system to perform self-service quality assessment through automated statistical analysis and defect detection. The system automatically computes statistical parameters, compares measurements against dynamic limits, and identifies defects without requiring manual engineering analysis, thereby maintaining comprehensive quality assessment while dramatically improving analysis efficiency.
Solution Approach 2:
The patent replaces manual mechanical analysis processes with automated computational systems that perform statistical computations and defect detection. This substitution eliminates the need for engineering teams to manually analyze all test data while maintaining or improving the completeness of quality assessment.
Data Source
AI summary
A method for testing a plurality of electronic devices includes performing tests of up to m devices at a time to measure device parameters on a device tester configured to test up to m devices at a time, where m is an integer. After each test, the method includes performing statistical analysis of the measured device parameters for all tested devices to determine statistical data, including updated mean and standard deviation for each parameter, and storing only the statistical data, and not the measured device parameters. The method further includes determining new pass/fail limits for each device parameter based on the updated mean and standard deviation, and determining pass or fail of each device based on the new pass/fail limits for each device parameter.