Novel current sensing method for mass production of power converters

Orthogonal least squares estimation algorithm is applied for the identification of unknown relationship between the drain current, the voltage across the drain and source of a fully on metal-oxide semiconductor field-effect transistor (MOSFET) under different temperatures. Very low loss current sensing can be derived based on the fitted model, the sensed temperature and voltage across the drain and source of the MOSFET. The model can easily be denormalized to cater for MOSFETs of the same model so that it is not necessary to test individual MOSFET in power converters. The proposed method could be useful for testing power converters during the manufacturing process. It could also be embedded into the controller to implement current control or over-current protection for power converters. Experiments and procedures to carry out the model fitting are presented.