The trim bug in MSFlight Simulator

[Brett_Henderson 5]

Replying to Brett_Henderson :No, the visual centre as well as the reference have nothing to do with the view axis of the camera.I maintain then that in all situations MSFS hooks the model from an invariable point: its origin (that is the visual centre).When MSFS loads an aircraft model, it loads as well a reference point (the visual centre) that links the model and makes it coherent with the aircraft.cfg file (this is why the ref.datum is set with respect to the visual centre of model). Those who sustain that the model rotates around the varying cg point (like in real world) have not yet given any convincing proof.

Spend countless hours building, and testing a MSFS model ... or just try the tests I've suggested in this thread ... and replay from a tower view (which minimizes the illusion caused because the, "camera" follows the model origin)..Just like a contact point can determine an axis for pitch.. so do the CoG/wing_axpex ...

[b_kimoun 0]

Note that the following reply goes also to mgh :Replying now to n4gix :Sure that at rest , on the ground, the centre of force is at cg and, and forces are in balance( Aircraft weight + ground reaction =0).Now how about the case where the aircaft is in level flight? 2 forces have to balance the aircraft weight : the wing lift and the tail lift. But moreover, the moment of these 3 forces around the rotating point( cg in real world) must also balance and the sum be equal to zero. MSFS uses TBL 404, flaps lift, H tail lift (tail incidence not zero), elevator lift, spoilers lift(negative) to balance the aircraft weight. As for the moment balance issue, MSFS uses Cm0+TBL 473, the flaps pitching moment, the h tail pitching moment (tail incidence not zero), the elevator pitching moment, the spoilers pitching moment, the gear pitching moment and the thrust pitching moment. So, as you can deduce, the location of centre of lift does not rise because it is a concern that has been taken into account in the airfile computation. Same thing for the wing apex. It serves in calculating the airfile, but has no effect if changed after the airfile is ready. The only change I could see is the cg gauge reading (% MAC,since moving the wing apex changes the location of MAC with respect to cg location).As for the centre of lift location, I agree there are 2: The wing lift centre (at MAC/4) and the wing-body-tail lift centre(called also Neutral Point ). The airfile tables 404 (for lift) and 473 (for pitching moment) are implemented using these 2 centres. So far , I don't have a precise idea how.We have to agree finally, that the only force that changes(in location) in an aircraft in flight is the aircraft weight (cg). The aerodynamic forces are supposed to act at a fix centre of lift. These forces vary only in magnitude.Note that the following reply goes also to mgh :Replying now to n4gix :Sure that at rest , on the ground, the centre of force is at cg and, and forces are in balance( Aircraft weight + ground reaction =0).Now how about the case where the aircaft is in level flight? 2 forces have to balance the aircraft weight : the wing lift and the tail lift. But moreover, the moment of these 3 forces around the rotating point( cg in real world) must also balance and the sum be equal to zero. MSFS uses TBL 404, flaps lift, H tail lift (tail incidence not zero), elevator lift, spoilers lift(negative) to balance the aircraft weight. As for the moment balance issue, MSFS uses Cm0+TBL 473, the flaps pitching moment, the h tail pitching moment (tail incidence not zero), the elevator pitching moment, the spoilers pitching moment, the gear pitching moment and the thrust pitching moment. So, as you can deduce, the location of centre of lift does not rise because it is a concern that has been taken into account in the airfile computation. Same thing for the wing apex. It serves in calculating the airfile, but has no effect if changed after the airfile is ready. The only change I could see is the cg gauge reading (% MAC,since moving the wing apex changes the location of MAC with respect to cg location).As for the centre of lift location, I agree there are 2: The wing lift centre (at MAC/4) and the wing-body-tail lift centre(called also Neutral Point ). The airfile tables 404 (for lift) and 473 (for pitching moment) are implemented using these 2 centres. So far , I don't have a precise idea how.We have to agree finally, that the only force that changes(in location) in an aircraft in flight is the aircraft weight (cg). The aerodynamic forces are supposed to act at a fix centre of lift. These forces vary only in magnitude.

[Brett_Henderson 5]

Let's examine level flight, albeit simplifying it all, and ignoring that the most significant impact of a force from the elevator is NOT the force itself (relative to pitch-rotation).. but that it alters the main-wing's AoA.Can we agree that if the center-of-lift and CoG are not on the exact, longitudinal point, that the net of just those two vectors will be a pitch where the axis of rotation is somewhere between them ? (FIGURE A)If so, whe have to agree that the longitudinal axis of rotation is a variable, determined by these (and other) forces.Now, since realistically, we know that lift and gravity can never be at the same, logitudinal point, but for the instance when burnt fuel, or moving passengers cause them to coincide.. we'll need another force..FIGURE B shows a CoG in front of the center-of-lift, and a downward force from the tail, in order to maintain a level-flight equilibrium. In this situation, lift has to be the largest of the three vectors, as the other two counter it. Now, as we change the the tail force, pitch will ensue, and I'd argue that the point of pitch-rotation would be nearsest the largest vector (lift).As CoG moves behind the center-of-lift, we have to alter the tail vector to an upward force, and now (maintaining level equilibrium), the CoG, becomes the largest force, as the other two, lesser forces must equal it. And again, as we change the tail force to induce pitch, the center of pitch-rotation is nearer the largest vector..which is now the CoG.My argument is, that the pitch point of rotation is not the CoG, nor the center-of-lift.. but is a variable point.

[n4gix 4,955]

I maintain then that in all situations MSFS hooks the model from an invariable point: its origin (that is the visual centre).When MSFS loads an aircraft model, it loads as well a reference point (the visual centre) that links the model and makes it coherent with the aircraft.cfg file (this is why the ref.datum is set with respect to the visual centre of model).<snipped for brevity>We have to agree finally, that the only force that changes(in location) in an aircraft in flight is the aircraft weight (cg). The aerodynamic forces are supposed to act at a fix centre of lift. These forces vary only in magnitude.

Of course, that's precisely what my screenshot using Model Visualizer* demonstrates. The model's origin (0,0,0) is read from the compiled .mdl file, and becomes the default camera's point of focus. Of course in the case of FSX we can quite easily create a "custom camera," but even in such a case all offsets are relative to that one, fixed position.One of these days -if I can ever make the time for it!- I'll record a short video clip that will clearly show the "forces vector" changing its position during different phases of flight, which should put the matter to rest.* Model Visualizer is a SimConnect utility that will display selected static and dynamic (calculated) values from the running model as reported by the sim in real time. It was developed by Lefteris Kalamaris during the FSX Beta testing, but never released to the general public for some reason. That's a real pity because it is a tremendously useful tool for any developer to have in his/her toolkit! :(

[dmaher 0]

I’ve always assumed the flight model was slightly more abstract.This may be all wrong :( :( but it's how I've come to understand it.All forces are seperated into a vector and a moment acting at one point - the CG. *In the model the various ‘lifts’ don’t generate moments - directly; as you’d expect logically.Their various moments are isolated, controlled, and computed separately.In the model CL's never cause moments.The mathematical model doesn’t ‘look’ like an airplane, but it behaves like one.It’s a single-body model...*The visual model rotates about its origin not the CG - it's a compromise.(The roll behavior makes this very obvious)As the thread suggests - for best results place the origin in the CG range.All surfaces that have a Cm component contribute directly to the body’s total moment about CG.All surfaces that have a CL component contribute directly to the body’s total lift vector @ CG.That's my story anyway... :(

[Brett_Henderson 5]

The visual model rotates about its origin not the CG - it's a compromise.(The roll behavior makes this very obvious)As the thread suggests - for best results place the origin in the CG range.

The visual model rotates (pitch) around an ever-changing point.. as illustrate in my last post, and confirmed by n4gix's (soon to de demonstrated) statement

One of these days -if I can ever make the time for it!- I'll record a short video clip that will clearly show the "forces vector" changing its position during different phases of flight, which should put the matter to rest.

.. the gear-contact point example, and can be easily tested by anyone (per my tower-view test, post #76)... Normal, spot-view references are useless, as the "camera" moves with the origin.

[dmaher 0]

The visual model rotates (pitch) around an ever-changing point.. as illustrate in my last post, and confirmed by n4gix's (soon to de demonstrated) statement

Sorry I should have been more specific...it's roll that always acts on the visual model's origin. From what I'm seeing.As for pitch I can't really tell - it may be affected by CG.But I think this would be very difficult to prove one way or other because a good reference point is much-much harder to find.An aircraft with a CG 50' forward of the spinner and a smoke trail at the model’s origin still ‘appears’ to pitch about the model’s origin ( in fly-by view).If the aircraft was pivoting on some interpolated point (in front) the smoke trail should oscillate.But it's makes a smooth path as the model gyrates.I'm less confident about it, but I still think the visual model pitches about it's origin too.Of course my test flies pretty bad ;)

[mgh 89]

As for the centre of lift location, I agree there are 2: The wing lift centre (at MAC/4) and the wing-body-tail lift centre(called also Neutral Point ).

The wing lift centre is NOT generally at MAC/4. As I explained in an earlier post, simple 2-dimensional wing theory puts the centre of lift at the 1/4 chord point only for an uncambered (symmetric) aerofoil. The formula is:C_m = - pi *c - C_l/4 where Cm is the pitching moment about the leading edge and c is the mean camber. The distance of the centre of lift aft of the leading edge is given by:s = C_m / C_l = pi * c / Cl - 1/4 If c = 0 then the wing centre of lift is theoretically at the 1/4 chord point of of the leading edge but otherwise it isn't.This theory is supported by experimental evidence. NACA (NASA's predecessor) measured the NACAxxxx series of wing sections in a wind-tunnel. (NACA Report 460 - The characteristics of 78 related airfoil sections from tests in the variable density tunnel Available here http://naca.central....-report-460.pdf) I've extracted two sets of results from that report. The left-hand one is for the NACA 0012 section which is an uncambered section and 12% thick. The position of the Centre of Lift (cp) agrees with theory and is at 25% of the chord. The right-hand figure is for the NACA 2212 section which is a cambered section with a maximum camber of 2% of the chord located 20% aft of the leading edge and also is 12% thick. The Centre of Lift (cp) also agrees with theory and has moved aft to 80% of the chord as C_l approaches zero, heading for an infinite distance when C_l = 0.The neutral point is NOT wing-body-tail lift centre, The neutral point is the point on an aircraft at which the additional total lift acts following a change in angle of attack. In other words it's the point where there is no change in pitching moment with a change in angle of attack. Its position determines the staic longitudinal stability. If the cg is forward of the neutral point the aircraft is stable, if cg is aft of the neutral point the aircraft is unstable. If the cg's at the neutral point the aircraft is neutral (neither stable or unstable) which is why it's called the neutral point. In the case of a trimmed aircraft, I think we are agreed that the total lift force equals the weight and acts through the cg because and the total pitching moment about the cg is zero. If a statically stable, trimmed aircraft is disturbed by a small upwards angle of attack then this must result in a nose down pitching moment which will tend to restore the angle of attack to its original value. This requires that the additional lift force must act aft of the cg even though the trimmed lift acts through the cg.A real aircraft always rotates about its centre of gravity (unless its physically constrained by, say, being in contact with the ground.) This is an inevitable consequence of the principle that a rigid body is dynamically equiva/ent to a body with its mass concentrated in a point at its cg.

[Brett_Henderson 5]

See FIGURE B in post #109 ..If the CoG is forward of the center-of-lift, it will take a downward force at the elevator to maintain zero pitch, correct?If this zero pitch is during level flight, the combined, downward forces must equal the lifting force, correct ?Are you asserting that a change in the force at the elevator will result in a pitch-change that happens at the lesser CoG force, instead somewhwere nearer the greater, center-of-lift force ?I don't mind being proved wrong, as it means learning something.. but I'll have to disagree. I assert that the pitch-change is neither the CoG, nor the center-of-lift, but somewhere in between (closer to the stronger "fulcrum")..

[dmaher 0]

This requires that the additional lift force must act aft of the cg even though the trimmed lift acts through the cg.

Effectively, but the simulation model uses a strict reference system.This same force can/should be described as a vector and moment at the CG.The forces act through the CG.So this small increase in AOA will...For the wing…CMvsAlpha will shift right increasing Moment@CG.CLvsAlpha will shift right increasing Lift@CG.For the horizontal stabilizer…++AOA will decrease Moment@CG.++AOA will increasing Lift@CG.And so on for all components of the body…The two aspects of the force are separate.I think the structure of the airfile reflects this well too.

[b_kimoun 0]

My reply to Brett:I am really unable to see what you seem able to see. In fact I tried the Tower View but "infortunately", again, the model rotates with respect to visual centre (please look at the pictures enclosed). The pictures are about the experiment I tried yesterday to confirm what I am sustaining (model rotation around the visual centre is really a compromise in MSFS): - The model I used is a transparent model (2k mdl file) - AP engaged. - Altitude 5000 ft. - Wing span =0 ( this condition disturbs the model and makes it rotate randomly around itself). - The XYZ axes are drawn with green lights. - The red light is the cg. - The white light is the visual centre of model.Note: Picture 1 shows the axes at the visual centre, pictures shows the axis at the cg location. Both cases lead to the same conclusion: the model rotates around the visual centre.Please note that on ground, the pitching ( that you gave as example) on contact points is again always the cg rotating with respect to the contact point (main gear for instance) considered.

[b_kimoun 0]

I’ve always assumed the flight model was slightly more abstract.This may be all wrong :( :( but it's how I've come to understand it.All forces are seperated into a vector and a moment acting at one point - the CG. *In the model the various ‘lifts’ don’t generate moments - directly; as you’d expect logically.Their various moments are isolated, controlled, and computed separately.In the model CL's never cause moments.The mathematical model doesn’t ‘look’ like an airplane, but it behaves like one.It’s a single-body model...*The visual model rotates about its origin not the CG - it's a compromise.(The roll behavior makes this very obvious)As the thread suggests - for best results place the origin in the CG range.All surfaces that have a Cm component contribute directly to the body’s total moment about CG.All surfaces that have a CL component contribute directly to the body’s total lift vector @ CG.That's my story anyway... :(

100% correct dmaher.Once the airfile is calculated (using the right locations and dimensions) and is ready for use, for MSFS, the lift components mean Cl's +TBL404, and the pitching moment components mean the related Cm's+TBL 473. We do not have then anymore to care about where the centre of aerodynamic forces are, except of course for the cg point, where aircraft weight applies. The position of the latter concerns the Weight & Balance issue of aircraft. For instance, for the Boeing 737-800, cg safe range is 6%MAC to 36%MAC. The mid range is 21%, choosing then the visual centre of model(pivot) at 25% MAC is a good compromise to have the illusion of the visual model pivoting aroung a cg varying within the range [6% MAC; 36% MAC].

[Brett_Henderson 5]

I gotta do a better job of explaining myself..Even from a tower view, the "camera" focuses on, and moves with, the the model origin.. so any screen-shot based examples are useless. Wouldn't matter if the point of rotation was a mile away, if the camera is tied to a fixed point..My test, is to take all of the cfg longitudinal coordinates, and subract 200 from them.. effectively placing the model origin ~200 in front of the CoG, and center-of-lift.Place this model on a runway and execute a normal takeoff.. then do a 50' AGL fly-by and then enter a steep, abrupt climb.... replay them both from a DISTANT tower view (not zoomed in on the aircraft)I'm going to finally do this experiment, today... and will report back..Mean-time.. If the the takeoff rotation happens at the model origin, it will look quite normal from the tower-view. If it happens at a calculate point accounting for CoG and center-of-lift; the takeoff will look more a sweeping elevation of the visual model. Same for the fly-by.. If the pitch happens at model-origin, it will appear quite normal.. But if the pitch happens at the CoG/center-of-lift variable point, the visual model will sweep dramatically upward, as though the whole model were but a nose-cone on a giant airplane... and all we can see is that nose-cone.

[Brett_Henderson 5]

Quick note.. before anyone is as silly as me, and starts subracting 200 from dozens of coordinates..I'm pretty sure that just moving the reference-datum back 200' accomplishes this :blush:EDIT: .. and since it's that easy..I'm gonna try several differences.. the larger, the more dramitic the effect (if there even is an effect)

[b_kimoun 0]

The wing lift centre is NOT generally at MAC/4. As I explained in an earlier post, simple 2-dimensional wing theory puts the centre of lift at the 1/4 chord point only for an uncambered (symmetric) aerofoil. The formula is:C_m = - pi *c - C_l/4 where Cm is the pitching moment about the leading edge and c is the mean camber. The distance of the centre of lift aft of the leading edge is given by:s = C_m / C_l = pi * c / Cl - 1/4 If c = 0 then the wing centre of lift is theoretically at the 1/4 chord point of of the leading edge but otherwise it isn't.This theory is supported by experimental evidence. NACA (NASA's predecessor) measured the NACAxxxx series of wing sections in a wind-tunnel. (NACA Report 460 - The characteristics of 78 related airfoil sections from tests in the variable density tunnel Available here http://naca.central....-report-460.pdf) I've extracted two sets of results from that report. The left-hand one is for the NACA 0012 section which is an uncambered section and 12% thick. The position of the Centre of Lift (cp) agrees with theory and is at 25% of the chord. The right-hand figure is for the NACA 2212 section which is a cambered section with a maximum camber of 2% of the chord located 20% aft of the leading edge and also is 12% thick. The Centre of Lift (cp) also agrees with theory and has moved aft to 80% of the chord as C_l approaches zero, heading for an infinite distance when C_l = 0.The neutral point is NOT wing-body-tail lift centre, The neutral point is the point on an aircraft at which the additional total lift acts following a change in angle of attack. In other words it's the point where there is no change in pitching moment with a change in angle of attack. Its position determines the staic longitudinal stability. If the cg is forward of the neutral point the aircraft is stable, if cg is aft of the neutral point the aircraft is unstable. If the cg's at the neutral point the aircraft is neutral (neither stable or unstable) which is why it's called the neutral point. In the case of a trimmed aircraft, I think we are agreed that the total lift force equals the weight and acts through the cg because and the total pitching moment about the cg is zero. If a statically stable, trimmed aircraft is disturbed by a small upwards angle of attack then this must result in a nose down pitching moment which will tend to restore the angle of attack to its original value. This requires that the additional lift force must act aft of the cg even though the trimmed lift acts through the cg.A real aircraft always rotates about its centre of gravity (unless its physically constrained by, say, being in contact with the ground.) This is an inevitable consequence of the principle that a rigid body is dynamically equiva/ent to a body with its mass concentrated in a point at its cg.

I think the aerodynamic details you are pointing to are due to some slight differences between the different litteratures on the subject.In fact, wing centre of lift at MAC/4 is considered as a good compromise and is often used. I think that this assumption holds also for MSFS (the Concorde excepted). As for Neutral Point, I often read that this refers to the centre of lift of the whole aircraft( combination of all lifting surfaces, body included). True that changes in overall lift due to change in aircraft alpha acts at this point. And this is why a cg fwd NP makes the aircraft stable (any change in alpha rises a delta lift aft cg and thus a stabilizing nose down pitching moment). Right, at NP aircrafts are neutrally stable and the pitching moment slope is zero. But how do all these work in MSFS?I assume that in MSFS, when a trimmed aircraft is disturbed, TBL 473 enters in action to bring it back to stability, but at the same time Lift changes since there will be also a change in TBL 404 by the same disturbance. At this stage, it seems that there must be some link and coherence between TBL 404 and TBL 473. I frankly do not have knowledge about how these table are built, and even about what lift is referred to by TBL 404. Because, even with incidence=0 and elevator deflection =0, h tail develops a lifting component (of course when alpha is not zero, since it is a symmetrical airfoil).Now as for the pivot problem (cg or visual centre ?), we see that MSFS manages (by extra pitching moment from Cm_dt* trim tab deflection) to get the necessary equilibrium around what I can call the hidden arbitrary cg, that is the model visual centre.