PAT SAY 制程平均测试与统计.ppt

PAT SYAProcess Average Testing and Statistical Yield Analysis Developed by MARK YASIN Member of the Technical Staff Motorola Corporate - Science Advisory Board Associate Process Average Testing and Statistical Yield Analysis Down to Earth Concepts, Global Impact Tools That... Flag Process Shifts Allow You to Detect Problems Early Indicate the Source of a Problem Allow You to Contain and Correct a Problem Before Shipping the Product Lead to Total Customer Satisfaction Process Average Testing and Statistical Yield Analysis UNDERSTANDING PAT and SYA Definitions Advantages of using PAT / SYA Applying PAT / SYA Success of Suppliers using PAT / SYA PAT / SYA - Definitions Normal Distribution - probability density function from which statistical inferences can be made - A graph of the function forms a bell shaped curve Maverick - Nonconforming Part that meets the specification, but falls outside the normal distribution or performance of the product Pareto Diagram - frequency histogram with defect categories ranked in descending order Advantages of Using PAT / SYA Allows for Differing Manufacturing Processes by Different Suppliers Uses the Characteristics of Each Part Type to Establish Limits Allows for Static and Dynamic Application Screens for Hidden Anomalies Advantages of Using PAT / SYA Identifies Subtle Process Perturbations Provides Immediate Feedback of Manufacturing Problems Contains Maverick Parts Before Shipping them to the Customer Reduces Failure Analysis Costs Guarantees a Quality Product Process Average Testing - PAT PAT is the use of statistical limits for part acceptance, instead of specification limits, for significant part parameters. Applying PAT to a One Sided Distribution Steps for Applying Process Average Testing Use PAT at Any Point where 100% Tests Are Performed First Measure and Record Data on Selected Significant Parameters from a Statistically Representative Sample Calculate the Mean and Standard