Elevator, aileron and rudder trim in Aerofly...

  • While in some gliders I've flown the trim actually mechanically acts on the stick, through a spring, in most aircraft the trim controls, electrical or mechanical, act on trim tabs on the control surfaces.

    In a Baron 58, a C90, etc... on the ground, with no significant airflow the controls will not move ( the yoke and the rudder will stay put... )

    With dynamic pressure build up, the airflow over the control surfaces and their trim tabs will then force them up / down or left right depending on the control surface and trim input by the pilot, and then we can see the controls moving accordingly in the aircraft.

    Presently when we make aileron, elevator or rudder trim inputs in Aerofly FS 2 that have this trim options, we can see the controls moving even on ground, with the aircraft stopped.

    Limited by Main Thread...

  • And how would you like to model that?

    If you hold the yoke or stick still in flight and change the trim you don't pitch up or down in the real aircraft. You only create more or less force that you need to hold but the airplane does not change its attitude (if you ignore the tiny force that the trim tap actually produces)

    Code
    elevator_angle = c * input_angle

    Full backwards/forward deflection will always give the same response no matter what the trim is. So if I am on the ground an pull full elevator the deflection is the same as it is in flight. If I neutralize elevator the angle is identical to the one in the flight. And most airplanes are constructed to have a linear relationship between stick angle or yoke deflection and the elevator and therfor change in angle of attack and lift.

    So on the one hand you want the full deflection to stay the same, on the other hand you want the center to stay the same, to maintain linearity. This can only be achieved if you actually move the centering position of the joystick/yoke spring and get rid of the simulator trim all together. You would need to trim your actual hardware in front of you. Or you just remove the spring (this is what I do).

    The current trim works by offsetting your input by a little amount so that the center position represents a slight up or down input. Then either have to large control deflection to one side or you clamp it at the maximum mechanically angle, this is what Aerofly does.

    Code
    elevator_angle = clamp( c * ( input_angle + trim ), min_elevator, max_elevator)

    Another option would be to ignore the linearity and change the response to elevator based on the trim setting, which wouldn't feel any good, except at full deflection and neutral.

    Code
    elevator_angle = -c * input_angle^2 + c * input_angle + trim (or something like that)

    Or you don't represent the elevator deflection at all and just model the control column forces. But then it gets really ugly because in some aircraft on the ground the elevator just falls down on its own weight - but this would be the only way to get rid of the limitations of the users hardware...

    Code
    d(elevator_angle) = a * aerodynamic_force( trim ) + b * inertia_forces... + c * input_force

    Long story short: I see no solution that respects all the demands with the current available hardware. You can either do a "force feedback" immitation where the elevator deflection is not directly controlled by the stick, making the elevator itself unrealistic, or you have a 1:1 releation ship between deflection and input and it only gets messy when you add trim.

    What we need is a joystick that can have its center of force at a different location than the center which can dynamically be moved by the computed control forces of the simulator.

    Cheers,
    Jan

    P.S. Just making the elevator look correct wouldn't be a real issue. Just map the unmodified input directly to the graphics and ignore the fact that the physical deflection is no longer identical to the visual deflection. This could go horribly wrong if you just add more and more trim and the simulated physical deflection can't be checked anymore.

    Regards,

    Jan

  • Hi Jan, and thanks for your reply...

    I was more referring to the GA ( Baron 58, even the C90 ) than to the airliners where in case of the Airbus there isn't actually any feedback in terms of control deflection in the real aircraft.

    But on a Baron 58, if you apply aileron trim on ground you will not see the yoke deflect, or if you apply rudder trim you will not see the rudders move, but inflight, with your hands off or in stick free "mode", aileron trimming will actually result in your yoke deflecting CW or CCW, depending on which way you trimmed, because the trim tab will create, over the corresponding control surface, the required force...

    It's a detail, I know, and I also understand the implications in terms of total travel... - you certainly have a good point in your justification. It's actually the same why it also isn't modelled in either MSFS or X-plane, just to give two widely known examples of similar simulation platforms - and the main limitation lays on the hardware we use for our controllers...

    So, best leave it as it is now - probably the most satisfying solution after all Jan.

    Limited by Main Thread...