Issue link: https://iconnect007.uberflip.com/i/1529411
42 PCB007 MAGAZINE I NOVEMBER 2024 dimensionless numbers. For example, you might remember that the Reynolds Number (a measure of inertia force divided by viscous force) is used in fluid flow calculations 4 . How Do Semiconductor Fabs Model Yields? e semiconductor industry has spent a lot of time modeling its yields based on die size, defect density ( AD), and wafer size. They have established four types of defect models (Figure 1): The Poisson Model: (The defect density is constant across the wafer) The Murphy Model: (The defect density increases toward the edge of the wafer) The SEEDS Model: (The defects tend to cluster across the wafer) The Moore Model: (The defect cluster (SEEDs) and higher density at wafer edge) (Murphy) 3. I read an article titled "Statistical Analy- sis for Multivariable Systems" in Chemical Engineering in 1970, written by a 3M engi- neer who introduced me to Plackett and Burman's 1 factorial design. 4. e only statistics book I enjoyed reading (and learning) was "Conversational Sta- tistics" by Harry Roberts of the University of Chicago 2 . But the free NIST "Engineer- ing Statistics Handbook" will have all the soware and training you will ever need and has one of the most thorough statis- tics manuals (including examples for self- paced learning). 3 e critical parameter in the calculation of first-pass yield (FPY ) was the metric called complexity index. It is a dimensionless param- eter in what I call a pseudo-independent vari- able: It replaces several important independent variables that would be complicated if taken individually. is variable is invented, but it follows several pseudo-independent variables that are essential for modern engineering: the Figure 1: Four ASIC yield models.