MODELING AND PERFORMANCE CHARACTERIZATION OF VISION-ASSISTED ROBOTIC CELL FOR PRINTED-CIRCUIT BOARD ASSEMBLY

Vision systems are widely used in electronics manufacturing for placement of surface mount components. One of the main objectives that drives the application of such a system is the need to acquire an on-line capability to detect and automatically compensate for errors during placement process. Although this process control does generally yield better quality products, if poorly designed, the gain in system precision due to the use of a vision system can be overridden by inaccuracies introduced by other subsystems of the placement system. The result will be a product marred by an intolerable error or defect. In this paper, the sources of error in a placement operation are characterized and analysed with the objective of identifying remedial actions to control the errors introduced.