3-point landing is not possible

[dany93]

The 3-point landing is not possible with the current FDM. It is important for short landing (e.g. in the bush). And slats will be useless otherwise. This point is one of the most interesting features of this aircraft.
In order to be able to do a decent 3-point touchdown, we must have the possibility of holding the aircraft during a (not too) short time at 11 to 13 deg pitch up angle at grazing the ground with a weak descent rate. Reminder: it means 13 to 15 deg wing AoA. And this pitch angle is a minimum.

JSBSim Piper J-3 Cub forum.

This issue has been observed from the beginning of this aircraft IIRC and it seems quite difficult to improve, worse to solve. All the more with the numerous aircraft options.

Hopefully for an improvement and after the discussions and suggestions in the forum, I tried and propose these changes:

• Change the CoG location (a bit aft).
  The empty CoG can be between 23 and 28% MAC. Current 13.8" is at 22%. Change for 15.75" is at 25%.
  https://scindeks-clanci.ceon.rs/data/pdf/1820-0206/2015/1820-02061501031Q.pdf
• Change the current wing lift curve (close to Xfoil prediction, 2D) for a measured one from NACA Technical Reports. Slightly extended and smoothed close to the stall angle to take into account for the wing twist.
• Add lift due to horizontal stab incidence
• Add lift contribution from the fuselage (Ysop suggestion).

Unfortunately, it does not completely solve the issue for every engine and wheel options. However, I think it is a noticeable improvement.

• With the J3Cub 65 hp at default load (at full load too), I succeeded several times decent 3-point landings.
  Tricky, but possible. One need to hold the aircraft grazing the ground and progressively pull the yoke full back to reach and hold the pitch angle just in time. Too high above the ground, you lose airspeed and the stall makes the aircraft pitch down with no control. At best, you touch down with the main gear. Too low to the ground, you also touch it with the main gear, with too much velocity, and a bouncing risk.
• With the PA18, or/and with the 35" wheels, 1 or 2 degrees are lacking to do a nice 3 point touch down.

These 3-point touchdown tests should be carried out with the elevator trim at 0 or slightly nose-up. Under FG, the pitch down elevator trim prevents the elevator from reaching 100% travel when the stick is pulled.
Ideally, this should be done with the throttle at 0.

Also, these changes noticeably improve the "nose down" issue for the PA18 at cruise speed (tested close to 88 kts). Pitch angle about -2 deg instead of -5 or -6 deg.

I've made a branch to enable people who wish testing.
I've made a several commits in order to help understanding and make reverting easier. Also, each intermediate commit can be tested.
I've put the replaced code lines under comment tags (not deleted) for the same objective. Even without git, anyone can see and change.

if you think it's an overall improvement, I will make a merge request.

Branch Issue-8

[wlbragg]

@dany93
I'm still testing this. It's very hard to get a definitive feel for this change VS without it. With the right speed I can get a 3 point landing either way. But is it easier one way over the other? I'm still testing under various conditions.

[wlbragg]

Second round of testing, it does seem a little easier and quicker to get the front end up with the changes.I think so much of this depends on a decent head wind and that is the conditions needed to really compare.

[dany93]

It's better to do these tests with no wind. Adding external parameters complicates everything. And wind does not change the touchdown pitch angle. It only changes the ground velocity.

With the right speed I can get a 3 point landing either way

Did you observe the aircraft in external view?(easier in replay mode).
3 point landing means 3 points simultaneously. Even better if you can touch the tail wheel gently first.
See this video https://www.youtube.com/watch?v=hPakbghLe38.
I never did as in the video (which is extreme) with our aircraft, but I can have a 3 point landing with the version I propose (which I hope might still be improved).
With the current master version, I've (almost?) never succeeded at doing this by a soft and well controlled manner. I will check this again.

Viewed from the pilot, observe the dashboard height from the ground (left and right horizon) when at stop. The aircraft has a pitch up angle of 11 to 12 deg with the default gear. You must being able to hold this just before and at touching for a 3 point touchdown.

[dany93]

I tried master again.

You are partly right, it is not impossible to do a 3 point landing with the current master version (tested with the J3Cub, 65 hp, default load). It is only very difficult. You cannot hold the pitch up angle during enough time (2 to 4 sec) to touch down with the 3-point pitch angle. The aircraft looses airspeed and quickly pitches down (that is, tail high). And you touch with the main gear.

The FDM compromise is very difficult.

Which is annoying is that, up to now, none proposes a solution. Is anyone interested in this behavior?

[wlbragg]

Which is annoying is that, up to now, none proposes a solution. Is anyone interested in this behavior?

Yeah, me! :)

It's the main reason I added all the gear change outs and ground cover logic. I really want the STOL behavior these aircraft are known for. Unfortunately, I'm just not versed in the FDM aspect.

[wlbragg]

Why is the ground spin so strong and hard to control on these simulated taildraggers? Two things of note for me.
The ground spin makes it really hard to test. Without a slight head wind it is really hard to test 3 point. With both, it's really hard to test. I have pedals but need to hook them back up. That should help make this easier to test.

[dany93]

I had this observation, the J3Cub in FG is particularly hard to control on the ground. Maybe too much, I don't know why. But it is a well known behavior of taildraggers in reality.
You can improve your control by pulling the stick full backwards (it improves the tail wheel friction).

You can try by changing the tail wheel static and/or dynamic friction (J3Cub.xml line 512) but I don't think it changes much. Unless it would be overwritten elsewhere in Systems?

        <contact type="BOGEY" name="TAIL">
            (...)
            <static_friction>  0.80 </static_friction> <!-- initial 0.60 -->
            <dynamic_friction> 0.60 </dynamic_friction> <!-- initial 0.50 -->

Rudder pedals help a lot, they are almost indispensable for helicopters and taildraggers.

Also, braking by pedals is hard but it's a common difficulty.

Could it help if I tried to explain why taildraggers are hard to control on the ground? Or is it something that you already know?

[wlbragg]

I can always use and explanation!

[dany93]

It's a question of stable / unstable equilibrium (balance).

Imagine holding a rod vertical, by hand or by a rope.

• If you hold it by its upper end, any disturbance is corrected and it returns to its initial position. It's stable equilibrium. Its CoG is below (lower than) its holding point.
• You can also have it remain vertical by putting it on its lower end, if this end it flat and perpendicular. But a small perturbation will result in falling down. It's unstable equilibrium. As long as its CoG is exactly at the vertical of the (small) contact end, it holds steady. But a small movement will break this condition.
  It's unstable equilibrium. Its CoG was above (higher than) its holding point.

For an aircraft:

• The inertia force (weight) is applied at the aircraft CoG.
• The main-gear resultant force (rolling resistance, side-slipping resistance) is applied between the two wheels.
  When rolling, a change in direction will give a lateral force (inertia, side slipping from the aircraft axis). The aircraft axis can be across the trajectory.
• The main gear tries to keep the trajectory by applying a lateral force, opposed to the trajectory slipping direction.
• The inertia gives a force applied at the CoG, in the trajectory slipping direction. More accurately, in the centrifugal acceleration direction.
  In a turn or if the aircraft is side slipping, these two forces will not be aligned: they will give a moment.

A tricycle aircraft has its CoG before the main gear (between the main gear and the nose wheel).
As the CoG is in front of the main gear center of forces, these two forces (inertia and main gear) give a moment which tends to bring the aircraft back in line. The inertia makes the weight "pulling" the aircraft. Like for the rod held by its upper end.
A tricycle is in stable equilibrium in case of side slipping on the ground.

A taildragger has its CoG behind the main gear (between the main gear and the tail wheel).
The application point of the inertial force (the CoG) is behind the main gear center of forces.
In case of a trajectory disturbance or side slipping, the moment from the inertial force (centrifugal) and the main gear force (centripetal) result in amplifying the spin movement. The inertia makes the weight "pushing" the aircraft. Like for the rod put on its lower end.

https://www.youtube.com/watch?v=q5trygRQaV0

Uneasy to explain well, sorry for my not very elegant English. Please ask again if not clear.

(EDIT)
See the video from 1:30.
You will understand what is said much better than me.
https://www.flight-training-made-simple.com/post/how-to-groundloop-your-taildragger

[dany93]

How to land a Piper J-3 Cub on a grass runway
See 0:29 ("a small amount of power..."), 0:38 ("The tail wheel should touch down at the same time or just before the mains....")

[wlbragg]

After further testing, I can tell the difference with the latest change. I compared the YASim version to the default version, then tested this version.
Almost impossible with the YASim version to do a three point landing.
The default version, has a very small window on take off that you can drag the rear wheel.
Issue-8 version has a larger window that you can keep the front gear off the ground with the back gear dragging.

That was the most accurate way for me to see the differences. Profiling the aircraft on descent for a three point landing is much harder to distinguish the differences.

There was a short discussion on another post about the sensitivity of the throttle slider. I think this is an area we can improve. The discussion was a twist ring around the base of the throttle that tightened up the force requires to change throttle settings and maybe the speed or sensitivity of travel to change the throttle setting.
In the Cub we have a different throttle control but I can see where a less sensitive throttle (more travel to achieve the desired setting) might help in this instance. You need an exacting speed/power setting to perform this maneuver. We may need to extent the travel of the throttle. More steps, smaller increases over distance moved. I think I need to work on that and see what I can achieve.
I have a friend who has a Cub with the wings off sitting in his garage, I may see if he can give me some details on the throttle mechanism. If there are any adjustments on it.

[dany93]

I agree, the three point landing is particularly difficult with this aircraft in FG.

To do these tests, I have three ways.
1- >Viewed from the pilot, observe the dashboard height from the ground (left and right horizon) when at stop. The aircraft has a pitch up angle of 11 to 12 deg with the default gear. You must being able to hold this just before and at touching for a 3 point touchdown.

This gives me a visual reference for the pitch angle seen from the cockpit for the three point landing.

2- In replay mode, I watch the aircraft in external view.

3- When almost ready for the landing, grazing low and well aligned, I press "V" to watch in external view and do the latest part (tricky too for the controls).

[dany93]

I have still slightly improved the three point landing possibility. I will send it in this branch.

[dany93]

Pushed. "Slightly tweak lift and drag beyond stall AoA". 37a7c6f
https://github.com/wlbragg/J3Cub/tree/Issue-8

Still slightly improved version.
I'm not completely satisfied with the FDM because (among other points) I had to guess a bit CLift and CDrag beyond the stall AoA.
The FDM is not a formal reference because of this, but the resulting 3 point touchdown possibility, stall, beyond stall behavior are closer to that I would expect for this aircraft than the initial FDM. It can be useful at least for you @wlbragg to try and get an idea of what I mean.

Alant seems actively working at calculating coefficients. Hopefully, we can expect that he will find a set which will give a correct behavior for our main issues.

[wlbragg]

Big difference. That is actually making it easy to do a 3 point. I need to test some more, but that's looking and feeling really good.

[wlbragg]

Here are a couple videos showing how easy this is now.

2023-10-09.19-22-18.mp4

Even though this second one was rather a stall to a 3 point, it is what you see in some of these competitions.

2023-10-09.19-21-37.mp4