inklap

SMT Head on Pillow Defect Detection Using AI Machine Learning

SMTA International · 2021

ABSTRACT As the electronics assembly industry has moved to environment friendly manufacturing and thin BGA (Ball Grid Array) packages with finer ball pitches, the incidence of HoP (Head on Pillow) defects during manufacturing have increased. HoP defects occur when the BGA ball sits on top of the solidified solder paste post reflow, without forming a continuous joint. These HoP defects are very difficult to detect during inspections in the printed circuit board (PCB) assembly process and often escape detection. They can result in electrical opens in the field at time of usage. The goal of this study was to come up with a reliable method of detecting HoP defects during the PCB assembly process. As the HoP defects consist of one metallurgical mass in contact with another, the joint can still pass structural electrical tests like Flying Probe or In Circuit Test, especially when pressure is applied to the solder joints during these tests. The nature of the joint also prevents easy detection by manual review of grayscale x-ray inspection images. A reliable detection method still eludes us. In this research work, we have applied Artifi

📖 افتح في inklap 🔗 DOI 📮 اطلب بحثاً