| Abstract |
In this study, a low cost current sensing approach based on module parasitics (inductances and resistances) is examined. It was found that the best results are achieved by using an active low pass filter as sensor element for the typical inductive-resistive measurement sections. This procedure, however, bears the drawback that the resulting sensor signal is highly dependent on other influencing parameters, especially module temperature and load current transient in the to-be-measured current pulse. To compensate these effects, this paper examines the origin of the sensor signal in detail and shows interesting results: It could be found that the information for current and the influencing parameters is embedded in the sensor signal waveform. This circumstance can be exploited by special sensor low pass time constants and an adapted sampling scheme of this signal. The samples can be processed to analyse the influencing parameters as well as the module current. In the following, considerations are given for different processing approaches including the application of an artificial neural network (ANN). This method is also tested in a close to reality simulative situation to proof the concept. It could be shown that the mathematical designed approach works and gives the first opportunity for a module parasitic current sensing approach without disturbance by temperature. |
