| | | | I was wondering how successful industry has been in quantifying relationships between defects rates of particular solder problems (like bridging, opens,..) and PCA design parameters (like spacing, part orientation, pad dimensions, ...). For wave soldering and reflow soldering, I would think defects can be attributed partly to design layout problems, not just to uncontrolled process variability. I am familiar with the logistic regression model which relates the probability of a successful binary outcome (the occurrence of a particular solder defect) to a set of independent variables. I have had some success modeling solder bridges on wave soldered SOICs, but I believe a few explanatory variables are missing from my model. | Hi Ted, | Just curious to know what type of solder defect modeling it is? Is it a multi-variated hybrid model? | rgs | chiakl Hi Chiakl, I am looking at relating defect rates to multiple variables. A univariate model would not capture the multiple-variable dependencies of the solder defect problems.

reply »