Quote:

Could you explain this ? I'm trying to figure out what kind of airflow gets around the wings, and I don't understand why aelerons have any influence there.

Quote:

That's also one of the main plots of a good flight sim: Allow you to test what you would never test in real life, and show you the most precise result possible.

I just hope they fix the damn collision detection... not realistic when you're flying a C-5A Galaxy and taking off over a C-130 and when youre about 15 feet over it you crash...Or when you just bump a building. CRASH

When it comes to crash detection, this is what happens.

This is my theory [and a plausible one]:

The sim renders an invisible crash-detection box around the aircraft and the size of the box is determined by the farthest contact point indicated by the "aircraft.cfg" file of the aircraft. I'm not sure it's either a box or a sphere of influence. But if the contact point dictate the size then it's most likely a sphere since the contact point can is being used to indicate radius and radius is a term used to measure the circumfrence of the sphere.

If this is true, then the sphere of influence must also extend upwards and downward in the same distance as the farthest contact point. Or maybe the distance downwards is probably affected by the location of the landing gear when extended.

I don't know, but it seems logical doesn't it? It could also explain an-225's situation. If this is the case, then it's only a matter of moving the contact points a little closer to the aircraft's center.

As for the case of the buildings, that's easily explained.

According to gmax gamepack sdk provided by Microsoft, FS will render invisible crash-detection "BOXES" around the scenery objects exported from Gmax to FS9. The boxes are then "STACKED" like blocks to conform to the "GENERAL" shape of the scenery object. However, the boxes will "NOT" be molded to the exact shape of the scenery object. For example, the boxes will be stacked like a "step" pyramid to confirm to the shape of the house's pointy roof. As a result, the remaining parts of the crash boxes around the scenery object will stick out like sore thumbs. But then again, the crash boxes on a square building will extend outwards anyways. Thus the inevitable "crash" will occur even when your plane's wingtip is only five feet away from the building.

Quote:



You'd be surprised...

Actually, I don't think this bubble thing applies.  When you make an aircraft, I you make no contact points, the aircraft would go through anything.  Once, I flew a 777 through 2 close-by buildings, side-ways, and I made it through.  If the spherical bubble rule applied there, i would turned of the sim cuz of the anoying 'crash' sign.

I think FS only cares about the contact points, cuz you hav point_class_wheel, point_class_float, and point_class_scrape.  I believe that any false collision detection is only caused by errors in the terrain collsion mesh, or badly placed contactpoints.

by da way, i think this reason is highly plausible too.

I actually agree with your invisible box/sphere theory there Katahu. Wouldn't it be nice if the collision detecion actually scanned the aircraft  model for perfect collision? Anyway I hope we get something like Battlefield 2's crash detection.

I tell ya, ya know how MS made FSEdit for FS, well, maybe they should make a short program which would actually load the visual model and put it in 'worst-case' crash scenarios, and automatically generate contact points for the aircraft.

Much simpler, and definatly accurate contact points.

But MS are probably not going to do that.  If any of you MS guys are watching this particular thread...

!LOOK HERE!

In FS9, when the wheels are overstressed, (and quite often should be ripped off) the wheels stay stationary, but when u try to land, they go thru the ground.  Stupid.  Hope they fix that in FSX.

By da way, the URL thing under ur avatar, not working...
Properly...