FDC (Fault Detection and Classification) monitors process tool sensor data in real-time to detect abnormal conditions and classify the type of fault for rapid response. Principle: During each process run, dozens to hundreds of tool sensors (pressure, temperature, gas flows, RF power, current, voltage, endpoint signals) record time-series data. FDC analyzes this data against expected signatures. Detection: Statistical comparison of sensor traces against golden reference traces or control limits. Deviations flagged as faults. Classification: After detecting an anomaly, FDC categorizes the fault type (gas leak, plasma instability, heater failure, particle event, recipe error). Enables targeted corrective action. Univariate vs multivariate: Simple FDC checks individual parameters against limits. Advanced FDC uses multivariate statistical methods (PCA, PLS) to detect complex interaction effects. Real-time: FDC operates during the process run. Can trigger alarms or automatic tool shutdown if critical fault detected. Post-process: Trace data also analyzed after run for quality decision (lot hold/release). Integration with APC: FDC detects tool problems while APC adjusts for normal process drift. Complementary systems. Data volume: Massive data streams from modern tools (sensors sampling at kHz rates). Requires efficient data infrastructure. Benefits: Reduce scrap by catching problems immediately. Improve tool uptime with predictive fault detection. Enable faster root cause analysis. Equipment intelligence: Modern tools have built-in FDC capabilities. Additional fab-level FDC overlays provide cross-tool monitoring.