Anyway, instead of giving up on the fin fences, I went back to the images from the previous tests. I noticed that without the fences, the airflow is pretty straight near the leading edge and only goes spanwise in a clearly delimited area:



The obvious conclusion was to try a new type of fence, one that goes from 0 width near the leading edge to full width near the rudder hinge line. I quickly removed the fences from the model, reshaped them right there on the field and put them back on for more tests.



I didn’t expect too much from this, but I was very surprised to see the results are clearly superior to the previous version of the fences and also showed a marked improvement compared to the normal configuration (without fin fences).



I tested the model with and without the camera and the rudder is clearly more powerful, especially on KE loop exits (last quadrant) I can be more relaxed, the model is smoother and comes back to horizontal KE much easier.