Too bad I couldn't see the VE/FA of all your models, mind posting them / post a wireframe as well (what slugheady requested).
This way we could you give some pointers for yourself to improve on.