All aircraft have different control input characteristics, much the same as cars handle differently.  Each flight sim modeler is trying to duplicate those individual characteristics when he produces a plane.  To reduce this to some common demoninator would be false and extremely boring.  I've been a pilot going on fifty years and I can assure you that, if they all did the same push/pull, left/right, left foot here right foot there, I'd have found something else to do.  What all this boils down to is, adjust your behavior rather than the aircraft - it's part of being a pilot .

FS flight dynamics is a highly specialised subject & not many people know much about it. Your best bet on this forum might be Brett Henderson. You could try sending him a PM. http://205.252.250.26/cgi-bin/yabb2/YaBB.pl?action=viewprofile;username=Brett_He...

I'm here ..   

It would take a whole forum section; and many contributions from many people, to do this subject justice..  And it would be an on-going thing.. everyone learning as we go along 

In quick response.. best I can do is relate the theories.

First.. If you start chasing an aircraft handling goal in the air-file.. you're fighting a battle that can't be won. Most of what's in there is over-ridden by the cfg file... and the parts that aren't, are pretty generic relationships that are tied closely to the MSFS flight-model. Any tweaking of consequence done in the air-file, would be for a VERY unconventional model.

-The biggest factors regarding control surface function/use; is where the CG is located, and where the wing-apex is located. The CG is the mathematcal focal point for gravity.. the wing-apex is the mathematical focal point for lift. The first indicator that those are problematic in the cgf file will show up on the loading screen (where the loaded (not empty) CG is shown in realtionship to the center of lift.)

-Next is the 3D location  of the control-surfaces (of course assuming the wing/stabilzer areas in square-feet are accurate). Think of an elevator(or rudder) as a hand on a lever... and the longitudinal location, as the length of the lever. A too long lever has a stronger effect, but is less responsive.. and vice-versa. This is where I've found the need to cheat a bit. Real-world dimensions for vertical/horizontal stailizer location gives you a "sloppy" sort of stability.. especially at low airspeed (which consequently works out nicely for something like my V-tail Bonanza.. notoriously "wobbly" at approach speed). This is where lengthy experimentation comes in.. mixing stabilzer location, with control-surface area/deflection.

-The turn-coordinator only knows how the "mass" of the airplane is moving, relative to its loaded CG and center of lift. Now, this is mostly realistic, as the amount of rudder needed to coordinate a turn, is indeed a function of the loaded CG.. but that is  NOT  where you start the tuning process. This is where olderndirt's suggestion comes in.. After you've set the numbers up realistically (and are getting good, in-flight, non-turning perfromance).. you adapt your flying, to fit the aircraft (and its loading)

Now.. as you're doing all of this tuning.. the climb/cruise performance aspect of the model are getting bounced around.. so at several points during the tuning, you might have to adjust thrust and wing efficiency, too.

As you can see.. it's a very tedious process. The best thing to do; is to start from scratch .. because picking just one aspect of in-flight performance, and trying to tune just THAT aspect, will have you chasing your tail while opening cans of worms 

"From scratch" means.. Make the reference-datum, the same as the model-origin (without access to the source-files, you'll have to estimate the origin by noting the focal point in 'spot-mode').. then  sit down with 3D drawings, re-do all of the 3D coordinates, lights, contact points, engine locations, etc.

THEN, load the model.. empty it completely (no pilots/passengers/fuel/cargo), and just go look at the loading screen and see where the EMPTY CG lies..

Keep in mind that many light GA, SHOULD have their empty CG well forward.. and then with light fuel and just a pilot, it should be reasonably near center (on the loading screen).. If your weight stations are accurate, you'll confirm it by trying different loading scenarios.. and start to get an idea of how to load an airplane (and why its important). I've seen too many 3rd party models where ALL the loading stations are very close to the empty CG.. making everything short of an over-load of no consequence 


OOPs.. got carried away 

ANYway.. a quick-fix for too much rudder, could be handled three ways (stay out of the air-file until you're ready to tune things like gear-drag, and gauge response)

1) Increase the rudder area
2) Increase rudder deflection
3) Last (and most cheat-like), is to increase rudder effectivenss in the 'Flight_Tuning' paragraph

Just keep in mind that too much rudder needed, is likely a function of a bad CG, or center of lift, and/or poor rudder placement.. 

Wow Brett, now I've also learned a lot! Thank you very much!!

THe MOIs are as tricky as everything else. As you'd assume, they're the inertia for rotation around a CG.

We can get so deep into fighting within the MSFS limitations, that we lose track of the area between easy, and realistic.. trying to avoid outright erroneous.

I use MOIs for in-flight "feel" only (bigger models feel heavier)... and then try to get control responsiveness to fill in the gaps. If you get a yaw MOI to feel correct during control use.. it's likely to be "off" when representing the "heaviness" of the model in general.. and can really mess up ground-handling. Like, on some larger models; enough yaw MOI to make in-flight turning realistic, could result in ridiculously sluggish ground handling.

When you get in-flight MOIs "feeling" proper.. then you can sneak into the cheat-zone of 'Flight_Tuning', and tweak the pitch/bank/yaw stability.

It's truly an endless fight, because of the limitaions of a desktop simulator trying to realistically mimmick what  should  require a warehouse full of super-computers 

You could end up cheating so much that a third party reader of you cfg dile, will think you're a nut 

TRy downloading my Bonanza... Load it with two, 200 pound, front-seat passengers (pilot and right-seat).. give it 60% fuel, add 100 pounds to the rear cargo bay.. set trim to neutral (by the indicator), and go fly it. 

You designer guys are going to laugh at such a small feat, but I DID IT, and it makes me happy!

I increased Yaw Moment (Weathervane Stab) by 100%
Decreased Yaw Moment -Roll rate Adverse by 50%
Increased Dihedral Effect by 50%

The result is that the plane now requires about a quarter throw of rudder during the turn, and leans into the roll without it decaying into a spiral when rudder is applied steadily.  It certainly "feels" better, if not completely correct.  I still suspect that once I establish the initial coordination, and apply correct back pressure, that I then should be able to ease off of the rudder somewhat more.  That isn't happening, but since I'm not a real pilot, I really don't know if this steady, constant rudder input is appropriate for this particular airplane.  The A2A Stratocruiser certainly requires it, but the Beech Duke model allows me to ease off to an extent and the ball stays centered while back pressure is maintained.

In the abscence of actually flying the Twin Comanche, at least I can say that the turning now feels believable.  (it could have been right to begin with too, and the real plane is just completely disagreeable )

It feels good to tame the flight simulator beast from time to time, so thanks for your help guys!

olderndirt wrote on May 14^th, 2010 at 8:52am:


Opposite aileron?  I actually increased the dihedral effect a bit to tone down this spiraling effect.  It is still present, but I'm only becoming aware of it at steep bank angles.  Right now it suffices to simply center the ailerons, and let the dihedral counteract the roll moment of the rudder. How common is this, for instance on planes that you have flown.  I'll bet you don't need to do this on a high wing Cessna?  But maybe an older Twin engined Piper?

(watch me wind up reverting back to original flight model, just like you said above )