i already read all the posts concerning this problem, but none of the hints and tipps led to a final solution.
my WOAI traffic touches down in the water in varying distances (from only a couple of hundred feets to several miles) most of the time. happens to md83s, 737. airbus. interestingly the 744 seems to always gets a perfect landing. all other traffic is not affected (i.e. GA props and business jets, dehavilland turboprops and everything thats smaller then the general airline jets).
this is no afcad problem. someone already mentioned that this has to do something with the internal workings of how FS9 handles the approach on runway 9 without ILS for AI. they do the visual approach and crash through the "hard surfarce" at ~2000ft, then are not able to hold this height with gears and flaps down.
i tried to give the AI aircrafs more lift through the flight tuning. that kinda worked, but of course had tremendous effect on their landing behavior (ever saw a 737 slowly gliding down on the runway at around 60 kias?
then i tried to remove the loading weight of passengers. lighter plane, better stability? didnt worked out.
changing all flightplans (with ttools) that go to TNCM to FL310 didnt do the trick either.
setting visibility to <3 miles (as suggested) seems to work. maybe theres a way to hack something so TNCM always thinks its <3 visibility, while having the weather set to normal ~40miles visibility (i dont want to fly around the caribbean without seeing anything
this not just happens in the first 15 minutes (where it seems normal that AI behaves stupid until everything is sorted out).
there was something mentioned about the ATC AI distance and some point that is used to get AI on the approach and that this point is too close to TNCM. so i wonder if there is any way to change any of this for TNCM in the fs9 internals. maybe some hex editing, xml modifcation. whatever.
maybe changing something so the AI use instrumental approach on tncm.
maybe its a problem with the flight dynamics all WOAI aircrafts use and other traffic is not affected, i would just love to fix this problem somehow, if only someone could give me some more insights, as all my amateurish tries lead to no good results.
EDIT: after updating all the WOAI traffic airlines that fly to TNCM with Jan Martins improved (more realistic) flight dynamics a lot of testing and waiting and planespotting i had excellent results that mostly solve the problems i was having. the initial AI spawn of course still show the well known weird behavior but after a few minutes all the arriving traffic is able to approach correctly now.
i used ai separation and the fs2004 sdk ai toolbox module (which btw is very great as it gives you extrem good control over AI with instant position changes, viewing and whatnot) to monitor the traffic.
i have yet to see how the change of flight dynamics will affect the behavior at other airports but for the moment i am happy with how things worked out for me at TNCM and now can concentrate on my fspassenger winair twotter flights again