During dead time, the device Vds has oscillations which create a much longer than expected dead-time, deteriorating the efficiency. The transformer is an ML91S EI core set, with approximately 600nH/60uH magnetizing inductance, with good coupling (slightly over 99%). Since you have a few planar transformer designs, I was wondering if you have some experience in combatting this effect?
Below are the waveforms, I’ve included my attempt to fix this, by adding a 1nF, and 2nF capacitance on the LV of the transformer, to try to reduce dV/dt, but it seems to make the phenomenon worse.
Hello Greg,
usually this is due to lack of energy in the leakage inductance of the transformer. The small energy cannot properly flip the node, which occurs by charging the Coss of the devices.
To increase this energy you might want to reduce the magnetizing inductance of your transformer, in order to increase the magnetizing current. You can do a quick test by spacing a little the two halves of the core set and see what happens. Also you might also try to increase the leakage inductance a little. Likely adding a small value inductor in series. Adding capacitance should not help. Did you try to simulate the circuit with the transformer parameters you have?
Indeed I have simulated the situation (LTSpice), and I get the same oscillations. However in simulation the issue does get fixed as you said with a smaller magnetizing inductance. In practice I found that the ringing disappears but only with an even smaller magnetizing inductance, with an “extreme” air gap (which could make sense in a way as leakage also decreases). I’ll see if I can increase the series inductance to patch this issue.
