Blind signal separation (BSS) algorithm applied in EMI noise diagnosis

A new approach to conducted EMI noise separation is presented for EMI noise diagnosis based on blind signal separation algorithm in cooperation with the use of current probes in time domain measurement. In this paper, firstly the principle of BSS is introduced and then, the implementation of BSS for conducted EMI noise separation is described in detail. Experiment results show that by using BSS algorithm, the common mode and differential mode components can be separated effectively from total conducted EMI noise as well as using common mode /differential mode discrimination network. Compared with reference signal with hardware-based noise separation, the new software-based technique is more simple, stable and also cost-saved.