What is it with MSFS and taildraggers?

[Tom Clayton]

Gotcha. I'd have to reset the 180 bit and re-test. Unfortunately, I'm dead tired, and my oven's telling my my pizza is ready! I'll make a note and test as soon as I can.

[bazzar]

If you have our Spitfire or our DC-3 in P3D, test the castor wheel there. It works in P3D. If you have our Cessna in P3D you will see the difference between the two quite easily. Have a great Pizza!

[K5083]

It's not torque. The uncommanded leftward excursion is not only independent of crosswind, it's independent of power. You can use a low power setting and a long groundroll, and you'll still get it at 40 knots. As YoYo notes above, it's the result of the rudder suddenly going from zero to 100 percent effectiveness at that speed. In taildraggers it's compounded by the absence of a castoring tailwheel and by persistent ground friction issues - which aren't as bad as they were during the tech alpha, but they're still there.

In other words, it's all about the bugs and has nothing to do with flying technique.

Very true. I thought the sudden swing might be an attempt to model torque, but you are quite right that power setting doesn't affect it.

I question whether it is about rudder effectiveness though, because the swing happens even if your rudder is at neutral.

August

[K5083]

This is correct and can be adjusted in the flight_model.cfg [CONTACT_POINTS], for example here's the Savage Cub:

max_speed_full_steering = 10 ; Defines the speed under which the full angle of steering is available (in feet/second).
max_speed_decreasing_steering = 20 ; Defines the speed above which the angle of steering stops decreasing (in feet/second).

Put a 1 in front of each those numbers and note the difference on take-off:

max_speed_full_steering = 110 ; Defines the speed under which the full angle of steering is available (in feet/second).
max_speed_decreasing_steering = 120 ; Defines the speed above which the angle of steering stops decreasing (in feet/second).

This is very interesting, but it seems unrelated to the wild swing bug, which is related to rudder effectiveness, not tailwheel steering effectiveness.

August

[bazzar]

If the majority of stock aircraft exhibit the same behaviour, no amount of fiddling with the steer angles, elasticity tables or any other file entry is going to help. The problem lies in the base flight model of the sim and there it will remain until somebody fixes it.

[keithb77]

Give this a try... not half bad... https://discord.com/channels/7933762...99794853232641

Can you post a summary?
Unless we're already on that server, just get the "you're in a strange place..." message

[K5083]

If the majority of stock aircraft exhibit the same behaviour, no amount of fiddling with the steer angles, elasticity tables or any other file entry is going to help. The problem lies in the base flight model of the sim and there it will remain until somebody fixes it.

We know. :/ I think we're just trying to understand the bug better.

Meanwhile, with many aircraft, I don't worry about sloppy takeoffs caused by this; just get the plane in the air somehow and enjoy the flight. And on landings. I may rate my approach and touchdown, but I don't beat myself up if the sim shoves me off the runway on rollout. If I want a plausible start-up to shut-down flying experience, I can still play P3D or FSX. Or FS9, for that matter.

August

[jmig]

We know. :/ I think we're just trying to understand the bug better.

Meanwhile, with many aircraft, I don't worry about sloppy takeoffs caused by this; just get the plane in the air somehow and enjoy the flight. And on landings. I may rate my approach and touchdown, but I don't beat myself up if the sim shoves me off the runway on rollout. If I want a plausible start-up to shut-down flying experience, I can still play P3D or FSX. Or FS9, for that matter.

August

To me, that is like saying, "While I love my wife, if I want sex I go to..." No self-respecting wife will think well of that activity and no self-respecting flight simulator/game should require someone to think of and worst follow your statement. This NOT a reflection on you. It is a SUPER BAD reflection on MSFS that anything from FS9 can be seen as better. Asobo has their X-Box. As the song from Evita said, "The Money with keep rolling in." Time for them to fix this giant zit on the end of their nose. What happened to physics and 1000 points of vector contact?

[tiger1962]

This is very interesting, but it seems unrelated to the wild swing bug, which is related to rudder effectiveness, not tailwheel steering effectiveness.

August

I've tested this adjustment in 90 degree 20 knot crosswinds with gusts up to 25 knots in all the editable taildraggers and the tailwheel steering no longer overpowers the rudder before the tail lifts off the ground.
I'm able to make a smooth takeoff on the runway heading after "a little shimmy" (hem-hem!) as the tail lifts. It may not be the perfect solution but it works for me!

Edit: Just tested them again at ROKR Kerama with Live weather: Thunderstorm and 90-degree 27 knots crosswind. Full throttle used straight away in all the planes.
The Savage Cub takes off in no time with no problems. The X Cub needs more runway to build up speed, but is still controllable and takes off heading the right way.
The Pitts, Extra 330 and Cap10 all need plenty of elevator down trim or they try to lift off before even reaching stall speed. There's a bit of weaving about to say the least (I need a LOT more practice at this!) until they reach take off speed but again, they all take off heading the right way.
At no time was there any loss of rudder control, in fact rudder over-control was entirely my own problem due to lack of practice!

[K5083]

OK, I may not understand what you are doing and why you are doing it.

What adjustment did you make? Changing those values from 10 and 20 to 110 and 120? Wouldn't that make the tailwheel steering even stronger, and make the tailwheel steering overpower the rudder even more?

Not that that is a problem. Tailwheel steering should overpower the rudder when the tailwheel is planted, and the main way that I keep my taildraggers straight is to keep that tailwheel planted as long as possible. The rudder, in the 2-point attitude, is just not strong enough to counter the bizarre kick to the side - and again, it's unrelated to crosswinds - that the sim gives you at 40 knots or so. (At least in some planes - I agree the Cubs don't really have big problems - but add-on taildraggers, especially warbirds, are a different story.)

I might try your trick to get more tailwheel steering authority. I'll be especially interested in using it with my port-over warbirds, which have some of the worst swing problems. Those steering commands did not exist in FSX/P3D, so port-overs do not have them in their aircraft.cfgs; I wonder what default values they then take. If I add those to my port-overs, maybe I can fight off the swing a little better while the tail is down.

August

[tiger1962]

OK, I may not understand what you are doing and why you are doing it.

What adjustment did you make? Changing those values from 10 and 20 to 110 and 120? Wouldn't that make the tailwheel steering even stronger, and make the tailwheel steering overpower the rudder even more?

Not quite - the adjustment is to the speed (in feet per minute) at which the tailwheel reaches maximum control effectiveness. The values for the taildraggers are identical to the values for tricycle (nosewheel) planes, which is clearly incorrect.
What I've done is to raise the tailwheel maximum effectiveness speed to a level which the aircraft will never reach before it takes off - the tailwheel will never have it's full effectiveness on the ground so the rudder will always be more effective. I hope.

[K5083]

Huh, that's almost the opposite of the way I interpret those variables. The way I read it, nose/tailwheel steering should be at maximum effectiveness between zero and max_speed_full_steering, then it should decrease linearly to min_avail_steering_pct as you accelerate from max_speed_full_steering to max_speed_decreasing_steering. Then it should stay at min_available_steering_pct at any higher speed. So your adjustment would give you max steering effectiveness at any speed where the tailwheel is likely to be on the ground.

On a slightly different topic, but related, we talked about crosswinds in this thread. A guy just posted an interesting video on the FS forum where he watched the Developer Mode crosswind component readout while taking off. This may be old news to many of you, but I didn't even know about this readout. It shows that there is never any crosswind when you are stationary, and the sim builds it as you accelerate. The crosswind component that this user programmed (20 kts) didn't fully show up until the aircraft was doing 76 kts.

https://youtu.be/Vsmr339kx8c

This is, of course, completely unrealistic and is part, but not all, of our side swing problem. The crosswind builds up as we accelerate instead of being constant as IRL, meanwhile our nose/tailwheel steering effectiveness is going down, and the rudder is no use until it pops in suddenly around 40 kts. There's a span where we have no yaw authority of any kind, depending how the values above are set, and are just at the mercy of any crosswind. And then when the rudder effectiveness kicks in, if we aren't already in the weeds, we have the pedal hard over, and violently overcorrect. What a mess.

August

[YoYo]

This is, of course, completely unrealistic and is part, but not all, of our side swing problem. The crosswind builds up as we accelerate instead of being constant as IRL, meanwhile our nose/tailwheel steering effectiveness is going down, and the rudder is no use until it pops in suddenly around 40 kts. There's a span where we have no yaw authority of any kind, depending how the values above are set, and are just at the mercy of any crosswind. And then when the rudder effectiveness kicks in, if we aren't already in the weeds, we have the pedal hard over, and violently overcorrect. What a mess.

Nice found. Yes it could be also that the rudder works ok, but wind it appears out of nowhere and of varying intensity instead of being constant. This causes a completely unreal experience that you can not overcome with absolutely any technique.

[K5083]

Yeah, this helps to partly explain it, but I still don't think we're all the way there.

The MSFS crosswind component builds up gradually, according to the video, between 0 and 80 kts. That is NOT what we experience with the veer in the sim. For one thing, the veer occurs whether there is a crosswind or not. For another, it doesn't build up gradually, it kicks in violently at 40 kts or so. If it built up gradually, we probably could learn to handle it better.

I remain agnostic as to whether it is sudden rudder authority or something else that kicks in at 40 kts and induces the veer. To me it is just an "invisible hand." It does not act like it is related to power setting, to rudder deflection, or to wind.

August

[Alan_A]

Very true. I thought the sudden swing might be an attempt to model torque, but you are quite right that power setting doesn't affect it.

I question whether it is about rudder effectiveness though, because the swing happens even if your rudder is at neutral.

August

A good point - it reminded my about this thread from the MSFS forum, one that I've linked to before. Short version - user Alec246 did some extensive tests with the MSFS wind tunnel feature in dev mode, and determined that rudder sensitivity is about 10 times what it ought to be. So that - in combination with the tailwheel issues and the crosswind that increases with aircraft speed - seems to contribute to the mayhem. If your rudder is only slightly deflected, it's going to have an exaggerated effect when it kicks in.

Caveat - this thread is a few months old, and I believe the rudder sensitivity was toned down in one of the previous updates - maybe 3 or 4. So it's not as extreme as described. But it still doesn't seem right.

There's also this, about propwash over the tail. Again, its old - goes back to January - and may not be current. But like the one above, it adds one more factor to consider.

[tiger1962]

I remain agnostic as to whether it is sudden rudder authority or something else that kicks in at 40 kts and induces the veer. To me it is just an "invisible hand." It does not act like it is related to power setting, to rudder deflection, or to wind.

I think the "invisible hand" can only be the tailwheel steering effectiveness. I misunderstood the contact points definitions earlier and your interpretation is correct - the longer the tailwheel is in contact with the ground, the more steering control you should have, but the max_speed_full_steering and max_speed_decreasing_steering figures are incorrect for taildraggers and that's why we get the "invisible hand" causing uncontrollable veering off the runway.

So what I've actually changed in my settings is that the tailwheel stays at maximum effectiveness until the tail lifts and the sim responds fully to rudder input for steering/yaw control.

[K5083]

Alan: Yes, I never saw that thread, but I understand about the rudder sensitivity. I think it might be dialed down to about 5x what it should be rather than 10x since SU4, but it's still too much. You can dial down the sensitivity of your rudder pedals, but since you also use them for tailwheel steering, that kills your steering. My solution, since I got an X56, has been to map a mini-stick to the rudder as well as my pedals. The pedals are dialed way down in sensitivity for controlling the actual rudder, and I use the mini-stick as a tiller to steer the nose/tailwheel when it is on the ground. If you have a twist axis on your joystick that you don't use because you have pedals, that could also be made into a tiller.

Tiger, I'm looking forward to trying your steering speed adjustment and seeing if it helps me to keep the aircraft centered until there's enough rudder authority. In an hour or so I "get home from work" (i.e., unplug the zoom mic and cam, and plug in the game controllers) and can try this out. Will report on results, especially with port-overs.

August

[cavaricooper]

Can you post a summary?
Unless we're already on that server, just get the "you're in a strange place..." message

Sorry- https://flightsim.to/file/5061/savag...k-in-the-skies

I'm a trifle biased, but it's not horrible In fact, I'd say it's pretty great for what the Sim currently allows... and will get better as the Sim does.

C

[lagaffe]

Hello,

I used this method to adjust the behavior of the Menestrel HN-433/434 and the result is almost perfect !
As it is, I can get it to take off with a minimum of correction on the rudder (it's only a 32hp VW) and on landing the width of the runway is more than sufficient, I would say even half the width.

I will publish a v1.6.0 with this flight model and some corrections on altimeter function, exterior beacon apparence is also better

[tiger1962]

That's great news! I look forward to flying the Menestrel again in MSFS. I also have fond memories of the Druine Turbulent by frenchvfr...

[K5083]

Here is kind of a stupid and obvious trick, but one which I took a while to catch on to, for controlling these crazy veers more easily.

Go to General Options/Camera and switch "Camera Shake" to OFF in the cockpit camera.

This is the option that bumps your virtual head around in response to aircraft movement. The worst thing about it is that besides shifting the eyepoint, it also rotates your view on the bank axis (like you are cocking your head). The other stuff I can put up with, but in real life, I have never encountered an acceleration that has caused me to tilt my head to the side like that. Or if it does, my brain corrects for it without me ever noticing.

The effect of this is to exaggerate deviations from straight and level, causing you to overcorrect. With respect to the bank angle, in the heat of the moment it is easy not to notice that what is tilting is not the horizon relative to the cockpit, it is the cockpit relative to your monitor, which requires no control correction.

I had this same issue with Accu-Feel in P3D and FSX, too, and turned off this one feature of that add-on, which I otherwise like a lot.

As a TrackIR user, I already get plenty of immersive eyepoint movement from my unconscious head movements, often in instinctive reaction to what is happening in the plane, so I don't need Camera Shake to exaggerate it. I can see that Camera Shake might add realism if you don't have head tracking or VR.

August

[YoYo]

I'm starting to get a bit of the impression that with the SU5, this behavior and the oversensitivity of the rudder has been slightly improved (or maybe I'm lucky that I still encounter a lack or a very weak wind in my Real Weather ).

[Milton Shupe]

All the default tail dragger A/C seem manageable. The two that I have that are difficult to handle are the Flying Iron Spitfire and the Milviz Corsair. But it doesn't seem to me that it's the rudder authority that kicks in at ~40 knots, but the torque or p-factor, or whatever it is that causes the plane to swing left. If you just motor on down the runway with no rudder input, it will go straight until about 40 knots, then you're along for the ride. And it isn't crosswind, as others have pointed out. I've managed to tame the Spitfire during take off by setting 7 degrees right rudder trim before takeoff. During the take off roll, apply back elevator and right aileron. She takes off with no fuss at all. Landings are still "expensive" though... I wish there was a way to observe the aerodynamic data in the sim. FS9/X/P3D have a nice program for that. Forget the name of it. Then we could see what's "kicking in" at 40 knots. There is supposed to be a "virtual wind tunnel" available if you're in dev mode but I haven't been able to get it to work.

AFSD will allow you to monitor the data.

Flight Simulator Aircraft Dynamics and NavData (sors.fr)

[YoYo]

Btw. do you know something more about it? Asobo/Microsoft working on it? I dont know that its reported already on Forum or Zendesk of MSFS.
Maybe we can vote for solving this issue?

[Tom Clayton]

I need more practice with tail draggers myself. I can take off fine, but the last time I flew the Xcub, I ground looped it on landing. I managed to keep it from completely spinning around, but I ended up swinging around and off the runway at a right angle.