May 3, 2012
Fault Detection and Classification (FDC)
When equipment and process parameters are controlled, process outcomes are also well controlled. Fault Detection and Classification (FDC) is based on the idea that you can detect changing conditions within equipment and use that knowledge to improve process. First, detect a changing condition within the equipment that results in an abnormal status, then classify the detected abnormality as a root cause failure.
Using near real-time and historical data collected from the fab floor, FDC analyzes and tests that data against models to both detect problems as they occur and react to emergent conditions. By identifying processes with frequent alarm occurrences, placing definitive metrics on consumable life, and enabling preventative maintenance, the FDC process lowers cost by drastically reducing scrap, rework, cycle time, Test/Qual wafers, and unscheduled downtime. FDC also curbs costs indirectly by empowering device manufacturers to measure and compare equipment health, transfer and tune recipes, compare and match tools and chambers, pre-classify faults, and capture engineering “best known methods”.