Wear measurement of ultrathin grinding wheel using fiber optical sensor for high-precision wafer dicing.

When the wafer dicing saw processes hard and brittle materials, the wear rate of the grinding wheel blade accelerates. To detect blade wear in time, a grinding wheel blade wear detection method based on a fiber optic sensor was proposed. This paper studied the principle of grinding wheel blade wear detection method, detection device, hardware system composition, and detection signal triggering process. To improve the accuracy of blade wear detection, the proportional–integral–derivative + feedforward + notch filter control algorithm was used to improve the dynamic and static response characteristics of the servo control system. This study aimed to analyze the influencing factors of the detection accuracy and efficiency of blade wear by the changing optical fiber power response mode, brightness setpoint, and signal capture detection speed. According to the experimental results, fine mode, 50% brightness setpoint, and signal capture detection speed of 0.4 mm/s were selected as the optimal parameters for blade wear detection. The performance verification experiment of the blade wear detection method was carried out under the optimized parameters. Results revealed that the maximum measurement error of the blade wear detection method was 3.8 μm. Because the measurement error could meet the industrial requirement that the measurement error of grinding wheel blade wear detection of wafer dicing saw was less than 5 μm, this method was feasible. [ABSTRACT FROM AUTHOR]