Machine Vision In Robotic Guidance For Pga Burn-in Board Loading

Eng Chin Peng Penang Package Technology Development Intel Technology Sdn.Bhd. Bayan Lepas Free Trade Zone 1 1900 Penang, Malaysia The prototype Auto Burn-in Board Loading machine is designed to automate the manual process of loading Pin Grid Array (PGA) package to burn-in board. The challenge of this project is the capability of auto-inserting high pin count PGA for example, a 168 lead PGA package into a 168 hole socket precisely onto the burn-in board socket. Pioneering effort had been made in computer vision. By integrating a low cost PC based vision system to guide a Cartesian robot, the integrated system proved to perform accurate insertion of PGA package to the burn-in board socket. This paper describes the system approach of a PC based vision system with a Cartesian robot integration. It also covers image processing algorithms and techniques in positioning recognition and compensation of PGA package and burn-in board socket.