Lecture 8.7: The Flyback Converter 4 (RCD Snubber) 

Cool question, I haven't really discussed transformer measurement yet. Just a warning that this explanation is coming off the top off my head. I don't know what your measurement set up is, so I can't diagnose all your issues, but maybe I can give you some theoretical backing. First of all, leakage inductance of 7 mH seems quite large. Perhaps you mean 7 uH? It usually helps to consider the "T" or "pi" models of a transformer to understand how to measure the magnetizing and leakage inductances. For multi-winding transformers it gets a little more complex to manipulate. We can do an open circuit test to determine the leakage inductance and magnetizing inductance (L_lk,1 + L_mag) seem from (let's say) the primary winding. Then you can do short circuit tests for each winding to determine the leakage of that winding along with the reflected leakages of the other windings in parallel with the magnetizing inductance seen from that winding (L_lk,k + (n_1/n_k)^2L_mag || (n_2/n_k)^2L_lk,1 || ... || (n_k-1/n_k)^2L_lk,k-1 ). Shorting the secondary and tertiary windings will make the apparent inductance seen from the primary decrease because you're effectively removing the magnetizing inductance from the measurement. Measure the other windings and see what happens. The amount of leakage will definitely depend on core geometry and winding/turn placement. I think the fact that gapless cores will have a much larger magnetizing inductance compared to gapped cores (typically) is what increases that percentage. The leakage inductance does not change much if a small gap is introduced, but the magnetizing inductance tends decrease significantly due to the total reluctance of the transformer being dominated by the gap reluctance. Good luck!

CCB is capacitor charge balance. I will get to an active clamp flyback eventually. Secondary side RCD snubber is pretty much the same as primary. The best design is dependent on the switch. When you're balancing Vsn and snubber loss, effectively it boils down to minimizing the sum of snubber, conduction, and switching loss for that switch. Switches with higher blocking voltages may have slower transition times. There is no "best" design that I can give you to you without context.

@@timmcrae3831 Thanks for your answer, I have tow more question, 1.) when Dsn turn on, the current through Dsn will divide two current path right? one is flow through Csn, one is flow through Rsn? 2.) Do you have video talking about QR flyback, I am very curious about how to design it.

This is one way of thinking about it. Either you use a very high voltage switch, or you use a lower voltage switch in conjunction with a snubber. Passive snubbers introduce some loss, but there's a tradeoff between this additional loss and the loss associated with a larger switch.

The snubber diode needs to block Vg + Vsn when the main switch is on. You probably also want to include some extra on that for safety.

Awesome. Thanks for your reply.

It doesn't go away, this was more of a convenience to demonstrate the current paths which effect the ringing phenomenon. When the voltage across the voltage across the transistor reaches Vg+Vout/n, the voltage across the primary side of the transformer reaches Vout/n. Assuming the diode on the secondary side is ideal, it begins conducting. When this happens, the current in the magnetizing inductance can now flow to the output and by small ripple approximation, the voltage across the primary side of the transformer remains relatively constant at Vout/n. Effectively this creates two loops of interest. One loop is the magnetizing inductance discharging into the output, and the other is the leakage inductance flowing into the switch. The magnetizing inductance current tends only to flow to the output and not into the switching, allowing us to ignore it in the leakage inductance loop.

@@timmcrae3831Thanks. I understand it now.