Other Parts Discussed in Thread: LM5025, UCC24610, UCC24612
I've designed a forward converter with the LM5026. It seems there is a fundamental problem with the forward converter topology when using secondary side FETs.
At light loads, IL will be able to become negative through S4. This is a huge problem, because when S4 turns off L-output will have nowhere to freewheel. All it can do is break-down S4.
Indeed, in my prototype I see that S4 runs in avalanche during light-loads where IL is forced to go negative. The converter actually overheats for zero output load, but works fine for 5A output current.
What am I missing here? How do you actually make these converters work correctly for light load? Turning off S4 at IL zero-crossing so that the inductor current never goes negative would solve it, but accomplishing that goal would be quite tricky and probably require some fairly expensive very-high-speed low-offset comparators.
