Taguchi DoE for ceramic substrate SMT defects improvement

Abstract

Electronic products are becoming lighter, thinner and more compact as the diversity of product grows and functions become more powerful. As a result, the size of Surface Mount Devices (SMDs) used for the Surface Mount Technology (SMT) become smaller and the density of elements becomes significantly greater. The improvement of manufacturing process is increasingly important in order to maintain product quality and yield. For the product characteristics, some defects were not allowed by customer, although the defects were accepted by IPC-A-610. This study started from reducing SMT defects in ceramic substrates for the improvement of product yield and cost saving. The Taguchi Methods were introduced to examine apertures on stencil, print wiping, oxygen content in reflow furnace and substrate thickness with less experiments, hoping to minimize variation in SMT process and quality for the optimized yield. An experiment program based on Taguchi methods was developed to improve product yields and efficiency of overall examination. This program was designed to investigate relevant factors, determine the optimized parameters with fewer experiments needed, and ultimately achieve lower process variations and stable design. The team started with a brainstorming session and came up with a list of possible factors that may lead to defects using the cause and effect analysis based on professional know-how and experience. A list of factors of great influence and their levels was produced. 7 factors were selected to determine the level of each of them. One of the factors was given 2 levels while the other 6 were given 3 levels. L18 orthogonal table was chosen for the combination above. This made it possible to represent 1458 (21×36) experiments with 18, and thus the number of experiments needed was reduced. To reduce the number of stencils for different aperture designs, the 18 designs selected for this experiment were made into a single stencil for reduced experiment time and cost. Taguchi methods were used to reduce variation and regulate target values. The experiment performed quality characteristics index as the production yield improved 3.94% overall; the solder ball rate dropped from 5.59% to 2.57% for an improvement of 3.02%; whereas the solder dross and Contamination decreased from 2.04% to 1.12% for an improvement of 0.92%. Through this study applies the Taguchi methods to accommodate experimental design, we got optimum process parameters combination in SMT process. The objectives of this study are to improve the yield, reduce the scrap rate, reduce the production cost and improve the quality of the product.