Interesting report, clr1024! On page 15, the upper two pictures show the air velocity over the rear of a Mitsubishi Lancer with and without vortex generators. In the “without” picture, the airflow separates from the rear window and never reattaches to the bodywork. The air passing over the trunk stays green (relatively slow) all the way to the trailing edge. The separation bubble over the rear window merges with the low pressure zone behind the car.
In the “with” picture, the area of separation is smaller. The airflow still separates from the rear window, but now it is able to reattach itself to the bodywork. At the trailing edge of the trunk, the air is yellow again (relatively fast). The separation bubble over the rear window is kept apart from the low pressure zone behind the car. Also, by having the air travel fast beneath the rear wing, the wing is able to generate more downforce.
On an NSX, we would also want the airflow to reattach itself to the bodywork before it reaches the rear wing. If you put wool tufts over the rear window and on the trunk of an NSX, you will be able to see whether the airflow separates, whether it reattaches itself again and if so, where it reattaches.
I put some wool tufts on the rear window and trunk of my NSX a while ago. At some point down the rear window (not at the top), the wool tufts started flopping around, indicating that the airflow had separated and become turbulent. However, the airflow reattached itself to the trunk again relatively soon, well before the rear wing. Since the airflow reattaches itself to the bodywork, I don’t expect vortex generators would do much good. If you want to put vortex generators on an NSX anyhow, according to that Mitsubishi report, you should put them on just upstream of the point where the flow separates. So probably across the middle of the rear window.
