印刷電路板組裝（PCBA，Printed Circuit Board Assembly）作業流程分為表面黏著 ( Surface Mount Assembly )、 人工插件 ( Dual In-line Package；DIP )、功能測試 ( Function Test ) 及組合與包裝 ( Assembly & Package ) 等四大流程。而每個製程都有其不同的製作方法與差異。 如果將製造流程加以細分，可以將這些流程區分為一般製程與關鍵製程。一般製程產生的問題，大部分可以使用例行的工作手法或工作經驗去解決；可是對於關鍵製程，則必須運用一些統計的手段，使問題變得更清析明朗，更容易解決。因此，在面對客戶的稽核與詢問：「這個製程，你為什麼這麼做？」，「這麼多參數，你為什麼選這一個？」，實驗設計（DOE）之因子分析，確實是一項真實而有力的工具。 實驗設計之因子分析是在可控制的情況下，使用不同的配置，對某一系統進行實驗與觀察。再經由實驗數據的統計與分析，以獲得變異的控制。因此，對於製程參數的變化，具有直接而有效的說明。當然，如果對製造流程愈了解，對於變異的控制與製程的改善就更愈具效果。所以，本文即以實驗設計之因子分析的方法，將製造關鍵之錫膏印刷與溫度控制等因素，予量化、模組化、公式化處理，以提供客戶更準確、更具說服力的參考。 Printed circuit boards Assembly are manufactured by the following four processes: surface mount assembly, dual in-line package, function test, assembly and package. Each of these processes can be implemented differently to serve the same purpose. The manufacturing process can be further categorized into Normal Processes and Critical Processes. Most of the problems which arise from Normal Processes can be solved by experience. However, it is necessary to use techniques such as statistical analysis to facilitate understanding of the problems that arise from Critical Processes. When we faced with questions raised by customers such as “why do you implement this manufacturing technique?”, or “why do you choose this variable among all the variables?” Design of Experiments (DOE) is indeed a powerful tool that will be able to address these concerns. DOE involves the design of experiments in a controlled environment and collect data while changing variables. Then by analyzing experimental data, the effect of the variable can be observed. DOE experiments can directly reflect changes caused by variables in manufacturing processes. Furthermore, the better one understands the manufacturing process, the more effective one can implement changes that improve the process.
