10 hours ago, jcomm said:

Apparently it'll remais as a toroid in XP12... at least as far as I interpreted from one of Austin's videos

Then I will dream on! I only ask for user-assignable side CL/Cd values. Vertical components would be nice too, similar to what Murmur is proposing, but even more rudimentary in that I'd bet rectangles alone would be a big step up. Ironic that we're literally begging for at least a table look up model for the fuselage...

I'm also sure Austin isn't willing to part with his center of pressure experiments done in the Tesla a few years ago. janov made some statements on Org a bit ago about changes coming to side loads on fuselage. It was also a side slip conversation. Of course he couldn't elaborate. The forces are there, they're just smaller than they should be.

As for the rudder locking in side slip, that sounds like something that would show up in a force feedback to the pedals. Just takes more muscle to move it the same deflection angle in the other direction. Am I understanding correctly? That would definitely be an adhoc mod to a flight model if no feedback hardware was being used.

4 hours ago, Murmur said:

I think the fuselage is the most problematic part to accurately simulate in a predictive-type flight model.

One method would be to treat it as two perpendicular flat plates (top view projection + side view projection, a bit like those small indoor RC aircraft)

XPs fuselage is modelled as a varying diameter 3D cylinder that then has wings, gear and "miscellaneous objects" (nacelles, weapons, drop tanks, etc) attached to it. All of which effect airflow (not sure how well lifting bodies are done - F35 models seem to manage OK), drag, turning moment etc.

6 hours ago, Murmur said:

I think the fuselage is the most problematic part to accurately simulate in a predictive-type flight model.

One method would be to treat it as two perpendicular flat plates (top view projection + side view projection, a bit like those small indoor RC aircraft) and use empirical formulas to calculate the aerodynamic centroid. I think Roskam uses a method like that in his texts. I don't remember very well, but maybe also Aerofly FS uses that method?

Anyway, even simplified CFD software like AVL (which does more complex calculations than our flight sims) uses simple body-of-revolution shapes (or similar  simplified shapes) for the fuselage.

Yes Murmur, and a "cross in cross" solution was somewthing we used in Flight Unlimited 2 and 3 to model the aerodynamics of a fuselage.

I used it when fine tuning a Pitts Special for Flight Unlimited which, just like the aerobatic aircraft aircraft in Aerofly FS, did remarkably in knife-edge flight !

 

2 hours ago, mSparks said:

XPs fuselage is modelled as a varying diameter 3D cylinder that then has wings, gear and "miscellaneous objects" (nacelles, weapons, drop tanks, etc) attached to it. All of which effect airflow (not sure how well lifting bodies are done - F35 models seem to manage OK), drag, turning moment etc.

 

But that's just for the 3d fuselage graphics if you only use Plane Maker to model it... For aerodynamics sake, as Austin made clear, only a toroid is considered, just like in MFS new "Modern Flight Model".

Watch the first minutes of the bellow linked video where Austin explains how the fuselage is interpreted in X-Plane's FDM...

Plane-Maker 8: Fuselage! - YouTube

Aerofly FS 2, and most probably 4 too, uses up to 16 sections each with it's own Cl / Cd coefficients. Flight Unlimited allows one to use airfoils to model a fuselage in two orthogonal planes.

Edited July 28, 20223 yr by jcomm
various typos... as usual...

1 hour ago, jcomm said:

But that's just for the 3d fuselage graphics

That's the physics. The graphics is (generally) done by connecting 3D .obj files together - but you can also make the physics objects "solid" - and even attach textures to them (not recommended - that functionality dates back to fairly early XP versions)

This is the 744 physics model in planemaker (show parts but not objects setting in expert - invisible parts), the fuselage part made by those cross sections I just posted.

The "graphics model" wireframe:

making both visible in XP

 

Edited July 28, 20223 yr by mSparks

17 minutes ago, mSparks said:

That's the physics. The graphics is (generally) done by connecting 3D .obj files together - but you can also make the physics objects "solid" - and even attach textures to them (not recommended - that functionality dates back to fairly early XP versions)

This is the 744 physics model in planemaker (show parts but not objects setting in expert - invisible parts), the fuselage part made by those cross sections I just posted.

The "graphics model" wireframe:

making both visible in XP

Watch the linked vídeo, by Austin himself explaining that no matter what shape is used in plane maker the fuselage is always taken as a toroid for the flight dynamics.

47 minutes ago, jcomm said:

Watch the linked vídeo, by Austin himself explaining that no matter what shape is used in plane maker the fuselage is always taken as a toroid for the flight dynamics.

the video shows the same thing, I just had the "Top/Bottom" tab selected which lets you drag the nodes around in 3D space, he is on the cross sections editor which lets you define the diameters by the numbers.

He shows the tab i posted about 12 minutes in.

all the force vectors from something like

are tied onto the physics parts - the physics wireframe I posted which gets made automatically from all the parameters you enter.

Here's it showing element forces

shows the drag being created by the wings, engines gear and fuselage etc, and I'm pretty sure "lifting body" lift created by the hump shown between the wings.

Edited July 28, 20223 yr by mSparks

On 7/25/2022 at 10:53 PM, mjrhealth said:

Is that xplanes fault, or developers simply not going the whole hog. Just look at Carenado. Pretty but they cant be bothered.

That is why I like A2A. Their flight modeling is good for the MS line and the mechanical actualizations are good also. Word has it their Commanche 250 may be coming out in XP after the 2020 model is released.

In a May developement posting I read that the control response adjustment will not be in XP12 because the flight modeling of each plane will be accurate enough to dispense with it. Only control sensitivity will be adjustable for people with unsteady hands. I have them all the way left. It actually "feels" on my CH Yoke that Im fighting forces. I know it's mental but it does feel more like a real plane that way to me.

 

Edited July 28, 20223 yr by Gary1124

7 hours ago, jcomm said:

only a toroid is considered, just like in MFS new "Modern Flight Model"

I'm not convinced it's even a cylinder there. They use the word "box" and even draw it as such, but the element count is described in 2D: 

"40 x 3 for the fuselage, assimilated to an elongated rectangular box"

They've really confused up the nomenclature. Maybe it's a triangular prism? My bet is a 2D sheet oriented normal to the lateral forces. This would obviously ignore lift and drag which doesn't entirely make sense either, given what the structure is used for.

6 hours ago, mSparks said:

drag the nodes around in 3D space, he is on the cross sections editor which lets you define the diameters by the numbers.

There's no doubt that it uses this 3D info. What's revealing in this case is at the 0:55 mark of Austin's video:

"xplane has to use rules of thumb from missile codes..." 

I think if there was variability in the forces along the fuselage centerline, it would show up in the force vectors in the graphical display. What we see right now is a single set of forces near the CG. That alone is a hint to me that all the blood-sweat-and-tears from the designer is being averaged up into a big cylinder. And to handle the huge drag associated with the flat front face of the resulting cylinder, planemaker lets you set an axial Cd value. 

And, to be clear, these missile codes are using experimental data. It's not just some adhoc guess.

Edited July 28, 20223 yr by blingthinger

Well I had a look at the "Cycle Dump.txt" file and apparently the modeling of the fuselage is more detailed than we assumed.

XP uses frontal, side and top areas of the fuselage as the reference areas to calculate fuselage drag, sideforce and lift.

Also, it calculates the centroids of side and top views of the fuselage to determine the point of application of fuselage lift and sideforce.

The point of application is then interpolated between 0 and 90 degress (AoA or Beta) according to the "missile codes" mentioned above (I think Austin might be referring to DATCOM).

 

2 hours ago, blingthinger said:

lets you set an axial Cd value. 

And, to be clear, these missile codes are using experimental data. It's not just some adhoc guess

There's a few "fine tuning" parameters that facilitate the last mile of getting the flight model realistic, Cd's are one of them (there are seperate ones for example for the wheel behaviour on the ground), where for the real aircraft it doesnt just depend on shape, but also things like paint/surface and other items which can have a significant effect on the aircraft behaviour but dont have a place to be defined.

This is also a primary reason XP11 is going to be around for a while yet, because how these behave changes during the XP run, which will lose developers loads of time if they devote significant effort to perfecting them only to have them change in an XP update (this erked lots of developers in XP11.30)

XP11 is stable now.

But fundamentally the aircraft is defined by its real world parameters, then any off spec behavior should be fixed via a bug report. There are lots of experienced developers involved in that (250 quoted in the recent livestream iirc)

Edited July 28, 20223 yr by mSparks

1 hour ago, Murmur said:

The point of application is then interpolated between 0 and 90 degress (AoA or Beta) according to the "missile codes"

Regarding this interpolation: I guess it's kind of a missile.

https://austinmeyer.com/the-rolling-wind-tunnel/

The missile data is used for CL/Cd/Cm, no doubt. Applied as a single point at CG.

 

21 minutes ago, mSparks said:

There's a few "fine tuning" parameters

Which of these parameters are available to the user?

Edited July 28, 20223 yr by blingthinger

1 hour ago, mSparks said:

There's a few "fine tuning" parameters

I guess I'm specifically wondering about datarefs. I'm certain Austin is fine tuning the fuselage to some degree. I doubt it's just a cylinder outright. If there are additional parameters you wizard devs use, I'm all ears!

One of the aspects that I believe I saw mentioned somewhere is surface hidding.

I expect that it get's more detail in the FDM for XP12, because under some circumstances it can play an important role.

Sideslips and relative wind and / or propwash being obstructed by the fuselage, engine nacelles, tail surfaces, are just a few examples.