應用DOE改善微機電系統麥克風產品測試良率

Abstract

目前微機電系統(Micro Electro Mechanical Systems，MEMS)在半導體產業所占的地位已經越來重要，在各項電子產品的終端應用上(如：智慧型手機、平板電腦等)也逐漸占有一席之地，越來越多的IC設計公司以及生產工廠增加本身的競爭力，提高產出、降低生產成本就成為各家公司進入微機電系統市場最重要的課題。然而因為各家設計公司以及生產工廠的產品製程以及終端應用都不盡相同，如何最大化產出一直都是各家設計公司及廠商占有微機電產品市場的重要決定因素。因此，本論文之主要目的是利用實驗設計法(Design of Experiments, DOE)中的隨機集區設計(Randomized Complete Block Design，RCBD)針對影響測試過程中的主要因子規劃實驗，以求出一組可提升產品良率（亦即產出最大化）的最佳機台參數設定值。最後，本論文以台灣某封裝測試公司目前進入微機電麥克風產品量產階段之實例，來驗證本論文利用RCBD手法所提出之最佳化機台參數設定值確實能有效提升微機電麥克風產品之良率。

Currently MEMS (Micro Electro Mechanical Systems, MEMS) have been increasingly important in the position occupied by the semiconductor industry, in the end application of the electronic products (such as: smart phones, tablet PCs, etc.) gradually place, the more An increasing number of IC design companies and manufacturing plants to increase their competitiveness, improve productivity, reduce production costs to become each company to enter the MEMS market is the most important issue. However, because of various design companies and products as well as terminal application process manufacturing plants are not the same, how to maximize the output has been occupied various design companies and manufacturers of MEMS products market is an important determinant. Therefore, the main purpose of this thesis is the use of experimental design (Design of Experiments, DOE) random pool design (Randomized Complete Block Design, RCBD) for the influence of the main factors during the test planning experiments to derive a set of improve product yield (ie maximize output) is the best machine parameter settings. Finally, the paper packaging and testing company in Taiwan has now entered a MEMS microphone instance mass production stages, to verify optimal machine parameter settings proposed by this thesis RCBD good practices can effectively improve the MEMS microphone products rate.