LesWhen you continue to see AI planes fall short of a runway it is because1. Poor Flight Dynamics (FD's)2. No ILS approach code in the database.Many AI planes can only make a proper approach and land if they are assisted by the ILS approach code in clear weather (VMC). When no XML ILS approach code exist for a runway then FS uses a fallback approach hardcode in the AI_player.dll file that is the same for all airports.

Nobody knows more about XML coded approaches in FS than Jim Vile. However, there are a few things which still leave me wondering. Why, for example, would AI.dll choose a non-ILS runway such as 09 at EHAM for an AI approach in clear weather when alternative XML ILS coded runways are available? Could it be something to do with wind direction, or the initial direction from whence the plane arrives in the 'zone', or the length of the runway? Perhaps there are particular problems at airports with parallel runways such as EHAM or where there is a quasi-parallel runway such as EGKK? I often see the phenomenon described in the first post at both EGKK and LHBP, but then I often fly into and out of those airfields. For those interested in watching AI planes approach airports in low visibility, I made a small discovery. Open FS9, chose your plane, go to a parking spot at an airport with a runway which has a VORDME approach and for which there are AI flight plans in your setup. Set weather to visibility less than 3 nm, stratus at 1000 ft asl and 'into wind' direction at least 12 knots according to the runway heading. Start the simulation. Once up and running you can change the weather to CAVOK (clear, unlimited vis) and watch the AI planes make the curved approaches to which Jim Vile refers. They blow away the notion that AI planes cannot make complex approaches. If the plane insists on landing on an XML ILS coded runway you might have to disable those runways to force the VORDME alternative. Have fun!

Hi,For parallel runways, FS9 will route you to the closest runway to the direction you are arriving from. If in VFR weather any parallel runway is used. As the weather deteriorates, only runways with IFR approaches meeting the current weather criteria are used. For example, if you are arriving from the North in good weather, you will get the North runway. If arriving in bad weather and there is an IFR approach only on the South runway, that's the one you'll get.Hope this helps,

Thank you Jim Vile, for adding your knowledgeable expertise. I use many of the AFD/ILS approach files you have provided to the FS community, i.e. KATL. KMIA, EHAM, and many more. Your contributions are greatly appreciated.Mel

Nobody knows more about XML coded approaches in FS than Jim Vile. However, there are a few things which still leave me wondering. Why, for example, would AI.dll choose a non-ILS runway such as 09 at EHAM for an AI approach in clear weather when alternative XML ILS coded runways are available? Could it be something to do with wind direction, or the initial direction from whence the plane arrives in the 'zone', or the length of the runway? Perhaps there are particular problems at airports with parallel runways such as EHAM or where there is a quasi-parallel runway such as EGKK? I often see the phenomenon described in the first post at both EGKK and LHBP, but then I often fly into and out of those airfields. For those interested in watching AI planes approach airports in low visibility, I made a small discovery. Open FS9, chose your plane, go to a parking spot at an airport with a runway which has a VORDME approach and for which there are AI flight plans in your setup. Set weather to visibility less than 3 nm, stratus at 1000 ft asl and 'into wind' direction at least 12 knots according to the runway heading. Start the simulation. Once up and running you can change the weather to CAVOK (clear, unlimited vis) and watch the AI planes make the curved approaches to which Jim Vile refers. They blow away the notion that AI planes cannot make complex approaches. If the plane insists on landing on an XML ILS coded runway you might have to disable those runways to force the VORDME alternative. Have fun!

JohnThe reason why ATC will choose rwy 09 when the airport is VMC over the ILS 18/36 runways is based on winds. There are a lot of "IF", "AND", "OR" statements in the runway score. Wind carry's the highest score during VMC conditions where as a ILS XML code carry's the highest score in IMC conditions (based on winds). Each airport runway confiquration has to be put through the runway scoring process by evaluation. KLAX does not score the runways the same as EHAM which does not score the runways the same as EGLL, etc.FS has to factor all the runway score coding and use what the confiquration of the runways dictate.What parts effect the runway score selection arewinds ---- carry a very high scoreapproach code ---- ILS carry's the highest score and NDB carry's the lowest score which is just above a visual approachbase vs recip end ----- base end of runway scores 100 points higher then the recip (secondary) end parallel runways vs non-parallel ------ parallel rules see Tom's post above. Non-parallel use a different set of rulesCW runway technique in use to simulate a parallel group within 7.9 degrees ---- Throw out most of the parallel rules and combine the non-parallel rulesweather visibility ------ see your post aboveStart Location ----- is also the Flightplan ending location for all AI/User airplanes based on winds or override by ATCHold short node at takeoff end ----- 90 degree taxiway entrance vs a taxiway that enters the runway at 60, 45, 30 or 0 degree entranceMain Terminal placement between parallel runways or to the side of parallel runways ---- FS9/X uses the old FS2002 runway overlay system based on terminal placement. Ground freq is bridged across the entire airport rather then using 2 ARP's parking spot Lat/Lon vs runway center line ---- runway center line circles the entire world and is the fundamental code used for all curved approaches runway length based on type aircraft (6 default type aircraft in the .air vs EW in the .cfg based on Tom Gibson's research) ----- we could write a book on this one for runway score selection. The runway selection score for weight restriction runway usage is embedded in this code. FS did not activate this score but we activate when using my Xwind runway technique. Empty Weight restriction is also the way I activate all parallel runways (5 at KATL) at the same time. By default FS only uses the outer 2 for arrival. I activated all 5 runways at ImaginSim's FS2004 KATL if you purchase that scenery. I also added the weight restriction code to the most outer 2 runways. Only B737 or smaller size aircraft are vectored to land on those 2 outer runways as per ATC. Much heavier planes (above 74,000 EW lbs.) are vectored by ATC to use the 3 inner runways for arrivals. Many users are of the impression that runways should be closed. In some cases that works but for every runway that is closed that doubles the usage of another runway. The objective at large airports with multiple runways (parallel and non-parallel) is to open all runways which allows more planes to land per hour then what FS has calculated (Reggie has the spec for planes landing per hour).One of my favorite airport/runways to watch AI arrivals is the stock PHNL (if you don't use my FS9 PHNL). Runway 26L is a LDA Approach which is FS9's stock curved approach. Using your technique to set weather to IMC and then back out to VMC all the AI Planes assigned to runway 26L will bank left on short final over the bay and land on runway 26L. If you have any payware/freeware version of Kai Tak I wrote all those checkerboard curved approaches for both IMC and VMC. FS9/X missed the mark on that one since all their exotic approaches are IMC only. That is the reason many users do not know they exist and should use your IMC/VMC find to watch the more complex AI plane approaches.@MelThankyou for the kind words. I am working on updates for all the more complex airports now that ADE9 is up and running. I need to redo/update all the approach code and ADE9 allows everything to be in 1 bgl unlike AFCAD. My newer airports will be under the banner of AIG which has quality control simalar to what we had with the old PAI AFCAD Airports.jim

To Jim, you are most welcome. I'm looking forward to your updates and future work through AIG. BTW, I downloaded ADE9 beta version, installed, and it looks like a great piece of work. I'm sure it will be a worthy replacement of the venerable AFCAD.

JimThanks so much for all that information. I realise that all of that took long hours to unscramble - it's a superb contribution.I've been away from FS for a while but am picking up again and will be taking a look at ADE9 - it's great that you've done an FS9 version. I didn't know about your payware KATL and will be tracking this down, and I will be taking a look at PHNL just as soon as I get back home this week. You kindly gave me some help a couple of years back with PAAQ and PADU. I managed to code NDB approaches to both runways at PADU although the northeast approach was very sporting with just a few hundred feet clearance of the mountains to the south. It was not good enough to upload anywhere but perhaps ADE9 will allow me to improve on it. Thanks again.For those who mailed me about how to know if a particular airport has VORDME or NDB approaches, and any others interested, use your GPS unit - the approaches available will show as white dotted lines in FS9 - I do not know about FSX.