I know what your talking about. I am actually considering redoing my exhausts. I have yet to check and see what the area is in between the start and end shape I used. So I may still go that route since you bring up a very good point. If I use thrust vectoring, which I plan to, the exhaust should be more intensively designed than the intakes were. I will also most likely do the same thing I did on the intakes.
My goal was to have a constant area all the way to the fan blades. The problem is that the spinner cone on the front of the fan takes up a lot of area. And most intakes do not take that into account. So what I did with the new ones, is account for the area that is taken up by the spinner. So the intakes continue shrinking, from 31.65 sq. cm down to 29.10 sq. cm. at the start of the spinner. Then the intake follows the contour of the spinner to keep a constant 29.10 sq. cm. area all the way to the fan blades.
So for the exhaust I will have to utilize the same concept. I'll have to put tail cones on the fans (already have them) and make the exhaust follow the contour of the tail cones to keep a constant area to the end of the duct. The only question that really remains for me is what to end the duct at. My thought is that if I do thrust vectoring, I should keep the exhaust @ 100% FSA, and use the nozzles for getting it down to the desired 70 or 80%. The problem with that is that the nozzles are very short. And this would create a pretty abrupt change in the airflow. So should I make the exhaust duct for from 100% FSA down to 85 or 90% FSA, and then have the nozzles do the rest.
In tests that I have done with a long exhaust tube, cutting the exhaust down to 90% did not produce much change in either air velocity or static thrust. But this may be different up in the air. So who knows?