Correlation Study on Voiding in Underfill of Large Quantity Ball Grid Array Chip Using Machine Learning

This paper investigates voiding issues in the underfilling process of ball grid array (BGA) chip packages under various parameter settings such as chip conveyor speed, valve pressure, temperature, and dispense pattern complicate. The study identifies valve pressure as the primary cause of voiding in large quantity BGA chips, achieving 88.9% in accuracy, supported with the deformation of the valve nozzle. Additionally, the findings reveal that racing effects occurs due to asymmetry of the solder ball array arrangement with percentage difference between the TSAM BGA chips experiments and its simulation counterparts in the range of 0.089–3.65%.