Fault residuals for compact disc players based on redundant and non-linear sensors

In the process of achieving good performance of compact disc players it is important to handle surface defects as well as possible. The first prerequisite for handling these defects is to detect their beginnings and ends. Two servo loops are formed to keep the optical pick-up focused on, and radially tracked at the information track on the compact disc. The pick-up feeds the controllers with sensor signals, and some signals for defect detection. However, due to optical cross couplings detection based on these signals can at times give false or no detections. A method to estimate fault residuals is designed in such a way that the cross couplings are removed. This is done based on a model of the optical system, from the physical focus and radial distances to the four detector signals. Combining this model with a fault model, enable us to solve an inverse function to find the distances, as it would have been if no faults had occurred. Two different approaches are pursued in this paper. Both are based on the Newton-Raphson method. Both methods solve the respective inverse problems.