MTS在測試流程改善的應用--以筆記型電腦為例

Abstract

筆記型電腦的生產過程中，最重要的製程就是功能測試，除了測試主機板與各基板間零件的功能外，並藉由燒機測試，提早將基板上不良的零件，能在工廠內篩選出來。為了確保成品的出貨品質，通常會在燒機前與燒機後，作部分相同的功能測試，但是，重覆的測試並不能保證零組件的良率，亦造成了工廠製造成本高居不下。 本論文應用MTS（Mahalanobis-Taguchi System）法來了解筆記型電腦成品的測試項目是否皆為必需的檢測項目。 首先利用馬式距離對各測試項目的觀測值進行綜合評價，再依據田口的望大特性，決定出重要的測試變數。本論文經由實例分析，研究結果顯示，MTS法可成功的界定必要測試項目，經最後進行確認數據後，驗證了此方法在保持原有良率水準下確實可行。

The most important process during laptop manufacturing is the functional test which can not only test the functions of parts among the mother board and other substrates but also can screen out the failed parts by burn-in tests. Consequently, in order to ensure the quality of finished goods, some identical functional tests are necessary prior to and followed by the burn-in tests. However, the repeated tests can’t guarantee the parts’ yields and will keep the manufacturing cost high as well. This thesis applied the MTS, Mahalanobis-Taguchi System, which will verify if all the test items are necessary items of the Note Book Computer. At first, to determine the value of observation from each test items by using Ma- halanobis’s distance, and defined the major testing items by Taguchi’s LTB（Large-The-Better）. According to the case study in this thesis, MTS model could successfully determine the necessary testing items and proved that it’s feasible to remain the original yields under this method by evidencing the numbers at the end.