電子連接器之失效分析與可靠度評估

Failure Analysis and Reliability Assessment of Electronic Connectors

劉明偉(Ming-Wei Liu)

指導教授 ： 吳文方

國立臺灣大學/工學院/機械工程學研究所/碩士(2009年)

https://doi.org/10.6342/NTU.2009.03173

摘要

針對電子產品的可靠度預估，產業界所採用的標準甚多，其中通訊用電子產品常以美商所發展的Telcordia SR-332為之，其係藉由內建多種依電子零組件型式分類之失效率套表，快速預估產品的可靠度指標。但這些套表是由經驗累積與零件廠商提供彙整而成，且零組件型式的分類也不盡齊全，特別是像電子連接器這種型式多樣的產品，藉由Telcordia SR-332所提供套表預估的結果不一定符合該產品之實質特性。因此，本研究目的即針對某廠商所自行設計生產的一種電子連接器(Electronic Connector)，規劃測試與分析流程，進行連接器系統之可靠度預估，以瞭解產品的品質特性及與可靠度水準。
本研究首先藉由「失效模式及影響性分析」(Failure Model Effects Analysis)與「失效樹分析」(Fault Tree Analysis)找出影響該產品壽命的關鍵潛在失效因子為溫度與濕度，而後就不同的溫、濕環境實施多種加速試驗(Accelerated Test)，再將測試結果輔以迴歸分析(Regression Analysis)，以取得產品的失效數據，最後利用適合度檢定(Goodness-of-Fit Test)與串聯系統可靠度模式處理失效數據，並依據艾林模型(Eyring Model)估算此產品在40 ℃/30% RH一般環境連續使用下之平均失效時間(Mean Time To Failure)為990.86小時。同樣的試驗結果若依Telcordia SR-332可靠度預估模型計算，其失效率為5,497.26 FITs (Failures Per 109 Device Hours)，換算成平均失效時間為181,908.81小時。此結果顯示，兩模型預估所得呈現顯著差異，因此在解讀Telcordia模型的分析結果時，需格外留意與謹慎。
本研究也發現試件之PCB量測孔於試驗後有明顯腐蝕現象，而會導致量測誤差，因此建議相關產品欲進行溫濕加速試驗時，量測孔應做好抗腐蝕處理。所幸，此誤差造成我們所預估的平均失效時間落在保守的一端，亦即產品實際的平均時效時間應比990.86小時還高。在物理現象方面，本研究也建立接觸正向力下降量與環境溫度的關係曲線，可提供爾後產品設計的參考。

關鍵字

可靠度預估 ； 失效模式與影響性分析 ； 加速試驗 ； 串聯系統可靠度模式 ； Telcordia SR-332 ； 適合度檢定

並列摘要

There are many standards set up and methods proposed for the reliability prediction of electronic products. Among them, Telcordia SR-332 Reliability Prediction Model is frequently used in telecommunication industry. The approach provides us an easy way to estimate a product’s reliability through examining failure-rate tables and environmental factors for a variety of electronic devices. Those tables are constructed based on data collected by and experiences gained from several suppliers. The approach is too rough for devices such as electronic connectors which are extremely multitudinous and, therefore, results in inaccurate reliability prediction. The major purpose of the present study is to design a test program and analyze its result to find the reliability of a particular type of connector. First of all, Failure Mode and Effect Analysis (FMEA) and Fault Tree Analysis (FTA) are used to identify the most relevant failure modes of the electronic connector. Based on results of FMEA and FTA, a series of Accelerated Test (AT) is performed to find the failure trend of the connector. After obtaining failure data of tested connectors by regression analysis, the study uses goodness-of-fit test and Series System Reliability Model to quantify failure data. The result shows that the connector’s Mean Time To Failure (MTTF) is 990.86 hours under the reference condition of 40 ℃/30% RH if Eyring model is adopted. It is quite different from the value of 181,908.81 hours calculated from the failure rate of 5,497.26 FITs (Failures Per 109 Device Hours) which is, in turn, obtained based on Telcordia SR-332 Reliability Prediction Model. After the tests, the PCB gauging holes of the connectors are found to be corroded. It may result in measuring error. Therefore, it is suggested that gauging holes should be strengthened to resist corrosion for accelerated tests to be performed in the future. For the studied connectors, this measuring error causes the reliability predicted above falls on the conservative side, i.e., the real MTTF should be longer than 990.86 hours. Aside from the above results, a relation between the contact normal force and temperature is found for the studied connectors. It can be used to improve the design of the studied or similar types of connectors.

並列關鍵字

Reliability Prediction ； Failure Mode and Effect Analysis (FMEA) ； Accelerated Test (AT) ； Series System Reliability Model ； Telcordia SR-332 ； Goodness-of-Fit Test

參考文獻

11. 闕延洲，自行車內變速系統之可靠度研究，國立台灣大學機械工程研究所碩士論文，民國93年。

26. 陳耀鍾，被動元件之可靠度預估－以通訊用連接器為例，國立台灣大學機械工程研究所碩士論文，民國98年。

6. Ebeling, C. E., “An Introduction to Reliability and Maintainability Engineering,” McGraw-Hill, 1997.

12. Kececioglu, D. and Jacks, T. Q., “The Arrhenius, Eyring, Inverse Power Law and Combination Models in Accelerated Testing,”Reliability Engineering, 1983.

13. Caruso, H. and Dasgupta, A., “A Fundamental Overview of Accelerated-Testing Analytical Models,” Reliability and Maintainability Symposium, pp. 389-393, 1998.