In the fast paced PCB industry of today windows of opportunity are short. One needs to be able to find quickly the key variables to a new process or an existing process so that the capability requirement can be met. Guesswork and traditional experimental design approaches will not work. Control charts by themselves will not work. Very large experimental studies will not work. What will work are simple, well defined, statistically designed experiments. This paper is an example of how simple, statistically based experiments were used to reduce haloing on Teflon® circuit boards. The drilling and plating processes using different laminate materials were investigated using Box‐Wilson and Plackett‐Burman experimental designs. Different data analysis techniques such as confidence intervals, scree plots, analysis of variance tables and analysis of means charts were used to determine the true significant (key) variables and the insignificant (non key) variables. An effective experimental study must have a capable measurement system. A method of determining the measurement system capability and improving it for haloing using isoplots is presented. In the end, two economically feasible paths to reduce the haloing problem were found.
CitationDownload as .RIS
MCB UP Ltd
Copyright © 1990, MCB UP Limited