FSP Narrow Body Oscillation

[Guest]

Hi guys,A serious question and some observations, following eighteen months of ai, ttools and afcad "addiction."Unlike the totally reliable and consistant Aardvark 757 and 737 series ai aircraft, I am experiencing horrible oscillation problems from FSP 320 family, FSP MD80 family, Urakawa FSP 737-700 and Urakawa FSP 737-800 narrow body aircraft on approach. much like the porpoise, dive and crash behaviors that have required the use of default 737 models for other gmax-modeled aircraft being used for ai traffic. I am not experiencing any oscillation problem whatsoever with FSP wide-body aircraft such as the FSP 747-700, FSP 777-200, FSP 767-300, or FSP A-306, either their original models or their updated models. The wide bodies are as reliable and consistant as the Aardvarks.I have tried to substitute the default 737-400 air and cfg files, alone or together, using the standard contact points edits that work for other aircraft, to no avail. Likewise, any combination of tweaking the scalars has no impact. I have been surprised not to find any specific posts about this, as the aircraft involved are excellent visually, work fine taxiing, taking off, climbing and cruising. I am at a complete loss about a solution. Very frustrating. Is anyone else experiencing this and, if so, have you found an answer?Thank you for any clues or assistance.Meanwhile, some observations.For what it is worth, I have found some rules of thumb to use after modifying every major airport with afcad and building ttools schedules for almost every major airline. What an obsession! I've noticed that so long as I keep arrivals and departures on separate runways, and carefully align overlays using radius gates, if I maintain a 2 minute arrival separation (30 flights per hour) on each active arrival runway, the system tends to work, more or less. If I can keep 4 minute separations, all the better. I have also kept myself to 12 active world carriers for time's saving sake and limited my USA airports using ai traffic to 50, although the master schedules include everything. This keeps it smooth and a nice, smooth background activity while flying around on routes with the Phoenix panels and Mel Jet aircraft. But I have learned to realize that FS2002 AI is very limited, temperamental and strange. And there is noyhing we can do about it.For example, American Airlines flight 57 from EGLL to KMIA. It is scheduled in my system to depart EGLL about 11:00 GMT every day on a weekly schedule and arrive at KMIA around 20:30 GMT. However, it consistently appears on approach to KMIA daily at 15:00 GMT no matter what just like clockwork. The aircraft involved is the FSP AA 777-200. Go figure.DMS

[Guest]

1. Maybe you should try using the FDEs of PAI aircraft. They work very well as the designers built it solely for AI2. That is a bug with the AI machine - no fix for it. :(

[Guest]

I am pleased to report the successful modification of the FSP narrow body aircraft to curb oscillation on approach. The basis of the solution is the work around recommended by David Carter at AI Aardvark (see Aardvark's forum). He posted his aircraft config file format and said to use that config file with the default 737-400 air file. There was another suggestion, that I try substituting PAI aircraft instead. I have used the several generations of PAI aircraft. As a matter of individual taste, over time I evolved my library to FSP wherever possible. I found that with every early generation PAI aircraft, I had to use the three default air files (734, 773, 744) and the hybrid default config file fixes, i.e., custom contact point, lighting and identification modifications. With experience, I just like the Japanese models better and wanted to make them work so I did not follow up with this approach. So I tested David's approach as a starting point. The aircraft involved were the FSP A-319, 320 and 321, FSP MD-81/82 and Urakawa B-737-800. All the air files are authored by Hiro. I was experiencing severe oscillation on approach with all of these aircraft when using them for AI traffic. All of the aircraft would gyrate, crash and vanish once they reached approximately 2,000 feet and/or turned for a final landing sequence. None of the aircraft experienced any other problem, taxiing, taking off, climbing out and cruising steadily. The same is true for all other FSP aircraft I have installed. The others have all been approaching and landing reliably. This would include all Boeing and Airbus wide bodies.I copied David's config format, making the necessary ID and contact point changes and activated the default air file. I flew each aircraft in three test environments for 1/2 hour each, to allow the AI traffic to "settle." In each case, every aircraft continued to oscillate on approach and resulted in a combination of crashes or hard landings. it was a slight improvement but the end result remained unacceptable.Next, I re-activated Hiro's original air files. In the SIM line of the configuration, I dropped the Boeing737-400 designator for each specific Hiro designator, such as fsp_a320, fsp_md81, Boeing 737-800, etc. I then went back to my three test scenarios. In the first model, an Airbus 320 family test, EDDF at 0800 GMT, I recorded twenty consecutive error-free approaches and landings of Lufthansa 321, 320 and 319 aircraft on Runway 7R. In the second test, KDFW at 1700 GMT, there were twelve error-free approaches and landings of American MD-82 aircraft on Runway 35C. In the third, KMIA at 2300 GMT, there were fifteen error-free approaches and landings of American 737-800, MD-82, United 320 and 319 aircraft on Runway 9R.I also noticed that it is absolutely essential to tweak the original Hiro contact points to avoid hard, nose-down landings. The rear gear in all the aircraft comes out too high, especially the MD-82, if you do nothing when you marry the Hiro air file to David's modified config setup. Also, lighting must be re-aligned.However, Hiro air files are completely fine, and, most importantly, produce smooth and reliable AI approaches just like the wide-body Hiro aircraft. The landings do not display the abrupt, steep STOL-like descent, and then over-threshold float of default aircraft. The result is more realistic looking landings akin to Aardvark aircraft and wide-body FSP aircraft. Combined with tweaking scalers to .55, etc., you get probably as good a result as possible with FS2002. This is about as much as we can do until FS2004 comes out and then we will see what happens. Thank you again to Hiro for his original work and to David for the modification input.DMS