Automated solder inspection technique for BGA-Mounted substrates by means of oblique computed tomography

High-density large-scale integration (LSI) packages such as ball grid arrays (BGAs) are being utilized in car electronics and communication infrastructure products. These products require a high speed and reliable inspection technique for their solder joints. Oblique computed tomography (OCT) was proposed as a novel imaging technique for BGA-mounted substrates, and it is being introduced in many manufacturing factories. Although operators examine OCT images manually, the establishment of an automated inspection technique is required from the viewpoint of an operator's load and the fluctuation of the inspection results. In this paper, a novel automated solder inspection technique by means of OCT is proposed. This technique consists of position adjustment, bump extraction, character extraction and judgement. Moreover, by combining five characteristic features, the condition of a solder bump was determined in computer algorithm. In this paper, linear discriminate analysis and an artificial neural network technique are introduced as the determination methods. In the experiments, these techniques are evaluated by using actual BGA-mounted substrates. The correct rate of inspection reached 99.7% in both the determination methods, which clearly indicates that proposed method may be useful in practice.