That post proves the point I was making. Incorrect assumptions that don't take everything into consideration are pretty useless.
First and foremost, all references to a given value must be removed from the airfile before removal from the config file, or the results will not be as expected. All variables in the airfile are backwards compatible! FSX reads every table that ever existed from FS2002 through P3D.
Second, one must understand that MSFS flight models are modular, and your assumptions about what something should do should keep this in mind before determining something as ineffective. Modular means that if elevator lift is zero, but elevator moment is 1000, you will still have elevator response and your assumptions about any test results will be wrong. If wing_apex is moved, which changes wing shape and rightfully, CG% MAC, that doesn't automatically mean the wing pitch moment will change. The relationship between CG position (which can only be changed through weight) and Wing Center of Lift need to change for aerodynamic effects to be seen.
We must understand the inter-dependencies before assuming a valid test!!!!!!!!
For example:
mgh said:
I carried out a test to try to understand which values in aircraft.cfg are used and their defaults. I took the default FSX C172 and commented out the entrries in [geometry] from wing_dihedral to vtail_pos_vert inclusive.
I compared times and speed to take-off and to 2000 ft with the original .cfg file and the modified one. There were no significant differences.
See above. Zero out all control moments, then set CL_xx values as appropriate (default FDEs are no where near appropriate. You will have to do the math to find the lift pounds that will create the desired moment at FULL deflection, then set this coefficient accordingly. I believe there is a simplified formula floating around for that...). Now set the geometry correctly in one test case and incorrectly in another. You will see a difference. Be sure you have removed all control surface, geometry, and moments from the airfile; however do not zero out the deflection tables such as 517-519 or equivalent. They affect the deflection of the surface itself, and without deflection you will only produce a constant (uncontrollable) lift force from each surface.
As for the wing_xxxx variables, they mostly affect computational things, not actual wing dynamics.
Those variables, (wing: chord, area, sweep, and apex) are used to find aspect ratio that is applied to find induced drag. There is no aerodynamic effect seen on the wing coefficients other than lift and induced drag. They do affect the calculation of wing center of lift in the absence of table 1534, but I highly recommend setting this manually.
Wing: chord, area, sweep, dihedral, and apex also affect the calculated CG MAC position. They don't affect CG lift balance, because that is controlled by the Center of Lift to empty CG relationship. What affects the pitching moment of a wing is the Center of Lift distance from CG set in 1534. In a properly designed airfile, the CG distance from the static Center of Lift for a given CG%MAC charted on the real aircraft, will match the MSFS CoL to CG distance at the same FS indicated CG%MAC.
Wing_dihedral affects the MAC height. Dihedral + wing_vertical_position is how the sim finds the MAC height, and if needed the CoL height.
mgh said:
Incidentally, I've done a further experiment when I eliminated everything in the aircraft.cfg file from wing_root_chord to rudder_area inclusive. FSX calculated wing_root_chord as wing-area divided by wing_span, set control surface areas to zero, and set other default values as before. Despite this the C172 still flies and manoeuvres!
You can (and I have) designed an MSFS/P3D aircraft with NO CONTROL MOMENTS, that is controlled by lift only and vice versa---100% moments, 0% lift. This will absolutely affect your test results. You don't even need the surfaces if moments are set.
Remember I said it's all modular? Well, moments affect moments (trim position etc), but not AoA, while lift affects AoA but not moments. Lift and Moments are summed onto the airplane.
mgh said:
Given that FSX creates returns these default values, I shall need hard evidence to convince me that it also uses another set of default values internally.
You are correct, it doesn't use default values. It will either read the legacy tables (they all work for backwards compatibility), or if not present, produce ZEROS. Zeros are a valid parameter in the sim. If you set your CL table to 0 it wont fly. But if you set your control surfaces to zero without first zeroing out your moments, it will still respond to control inputs.
Again,
We must understand the inter-dependencies before assuming a valid test!!!!!!!!
But let's add,
We must understand the proper way to achieve the desired results!!!!!
There's always been a ton of debate about things like wing_sweep having no aerodynamic effect. Why would it be needed? The aero effect of a swept wing is modeled in the lift table, cm versus mach table, cm versus AoA table, and the wave drag versus mach table. The actual wing_sweep variable only affects induced drag. It has no effect on the wings performance.
I would say this point brings me back to the original post here. Do those variables have any effect? Without testing, who knows....but who really cares. Everything you need to custom model winglets is already provided in a modular format.
