I'm designing a PS FB converter. I have problems with transformer saturation which might be related to a strange behaviour of the CS-pin on the UCC28950
UCC28950 EVM board (HPA471A) mounted piggy back to the gate driver board.
Current transformer in series with main transformer. Output from CT rectified with full bridge schottky diodes and then loaded with 5Ohm mounted on the EVM board. CS filtered with 1k and 330pF.
CS not connected to ADEL or other pis or circuits.
CH2 – blue: Current through transformer. AC-measured with rogowski coil.
CH3 – cyan: Voltage over CT sense resistor.
CH4 – green. CS after filtering, i.e. at the UCC28950 input
First picture shows operation close to full duty cycle but the peak current controller turns off just before max duty cycle. COMP is supplied with a DC-voltage divided from Vref.
Second picture shows operation with a lower duty cycle.
Q1) The filtered CS is pulled towards zero after turn off, even though the current is freewheeling in the transformer. Does the CS-pin sink current in this period? That is not mentioned in the data sheet.
Q2) The filtered value of CS does not follow the unfiltered value equally in the two half periods. What could be the reason for this? Note that the amplitude of the filtered CS value is not constant as it should be in peak current mode.
I have replaced the UCC28950 ith the same result.