high efficient flyback for about 100W

You'd need some reasons to do so, as a flyback-based topology is seldom applied in a situation where other direct coupled topologies, such as the basic buck converter, would serve.

As the voltages are low, and the power level in excess of 50W in this case, it might be the last choice, if efficiency was of primary importance, due to it's higher peak/average current stresses.

The buck boost converter (direct-coupled flyback)was used fairly frequently in low power -48V input telecom applications, where it's inverting behavior fitted naturally.

Higher efficiencies are also more easily achieved when converter stored energy has to pass through only one series switching element. The flyback requires two switches in series with the main energy path. As all your voltages are greater than 10V, however, 90% efficiency is possible under loaded conditions.

Worst-case efficiency is most likely experienced at lighter loads, where control circuit losses dominate, and efficiency is of less practical importance. Perhaps you might want to restate the efficiency requirement to avoid confusion in this regard - it can add expensive complications.

RL

I assume you are talking of making an isolated flyback since they always increase the voltage in the non-isolated case.

The transformer / inductor part of the design is the very tricky part. You need extremely tight coupling and lowish capacitance. Any energy that goes into the leakage inductance is hard to get back.

A snubber that gets the energy back (mostly) looks like this:

MMMM Checking my font

**** =2E... T1 =3D=3D=3D=3D=3D=3D -----+---------))))))--------- ! ! ! D1 C1 ! ---!

Wouldn't a buck type system work better for you? It would be much simpler to construct in your case.

I guess how ever there is always the chance of over voltage if the circuit fails to regulate how, depending on the design at hand. One could simply clamp the output with a crow bar type device which would take out a fuse link. A TVS diode comes to mind!

"

Flyback is not how I would make a "buck converter".

I suggest web searching into buck converters through the point at which you see what a "catch diode" is.

- Don Klipstein ( snipped-for-privacy@misty.com)

ch

I think the OP is long gone. I think he must have been looking for an isolated design. 90% efficiency in a bucker for the voltages he is specifying is too easy to be remarkable.