老化爐為一種用於對多個測試晶片施加熱負載，以加速其壽命驗證之設備。透過此過程，可於早期篩檢潛在失效元件並加以排除，進而確保最終產品之良率與運作穩定性。測試過程中，爐內通常同時放置多顆測試晶片，並採用強制對流空冷方式進行溫度控制。然而，由於老化爐內部空間配置之限制，導致冷卻空氣流動方向與分佈不均，部分測試晶片難以維持於目標溫度範圍內，進而影響故障檢測準確性及良率評估。 本研究採用熱流模擬軟體 FloTHERM，針對老化爐內部結構進行數值模擬分析，並將流場分為多風道入口區域與多風道出口區域，分別進行導流板與擋板之幾何尺寸與配置位置最佳化設計。研究對象為含 16 層空氣托盤之老化爐，每層設有 24 個出風口。透過逐步引入不同配置之側導流板、彎折導流板與前擋板，並進行參數化調整與比較分析，最終求得最佳設計組合。 模擬結果顯示，優化後各層入風口風量成功控制在 12.3 ± 0.6 m³/min 之範圍內，且 24 個出風口風量均落於 0.513 ± 0.026 m³/min 的目標範圍。此結果證明冷卻空氣得以均勻分配至各層與各出風口，為後續溫度精確控制提供了物理基礎。研究成果有效提升爐內氣流分佈均勻性，進而改善測試品質並支撐產品良率之提升。;Burn-in ovens are devices used to apply thermal loads to multiple test chips in order to accelerate lifetime verification. Through this process， potential defective components can be screened and eliminated at an early stage， thereby ensuring product yield and operational stability. During testing， multiple chips are typically placed inside the oven simultaneously， and forced-air cooling is adopted for temperature control. However， due to spatial constraints within the oven， airflow direction and distribution are often non-uniform， making it difficult for certain chips to remain within the target temperature range. This non-uniformity affects the accuracy of fault detection and yield evaluation. In this study， the thermal simulation software FloTHERM was employed to conduct numerical analysis of airflow distribution inside the burn-in oven. The flow field was divided into a multi-duct inlet region and a multi-duct outlet region， with targeted optimization of the geometric dimensions and placement of guide vanes and baffles. The oven under investigation consists of 16 layers of air-trays， each equipped with 24 outlets. By progressively introducing different configurations of side guide vanes， bent guide vanes， and front baffles， and conducting parametric adjustments and comparative analyses， the optimal design combination was determined. Simulation results indicate that， after optimization， the airflow at each tray inlet was successfully controlled within 12.3 ± 0.6 m³/min， and all 24 outlets achieved the target range of 0.513 ± 0.026 m³/min. These findings confirm that cooling air can be evenly distributed across all trays and outlets， thereby providing a physical basis for precise temperature control. The proposed optimization significantly improves airflow uniformity inside the oven， which in turn enhances testing quality and supports improved product yield.
