Reply to thread

Looks a bit noisy.. The limitation of Bode plots is that it is limited to small or large signal but does not reflect the dynamic range of step currents.  Since phase margin and overshoot are reasonably well correlated for 2nd order systems, I would inject active step pulses such as 50~100% , 10~90%,, 0 ~100% using BJTs on a suitable heatsink like an old CPU one.

 

Capture the voltage errors for various step loads to determine the capabilities and weakest pattern.

I just pulled a random discrete Buck regulator and added an active load with variable amplitude , DC offset and f from 100 Hz to 15 kHz square wave load currents 3.33V up to 100W.  https://tinyurl.com/237d5tt6