The paper describes an online tool failure monitoring system based on acoustic emission (AE) signatures. The system was developed primarily to detect failure of polycrystalline cubic boron nitride (PCBN) inserts in milling high-chromium materials. In face milling of high-chromium materials, almost without exception, PCBN inserts fail due to fracture on the nose or rake face. The change in the root mean square of AE signals (AE /spl Delta/RMS) was found to be the best indicator of the tool failure process for the subject tool-work combination. From the experiments, the critical values of AE /spl Delta/RMS where the tool fractures, called AE /spl Delta/RMS/sub c/, were obtained. Regression analysis was performed to fit a model of.