Runway sideslip with FAR

[JANXOL]

 

I seem to be having a little problem with my planes on takeoff. That's three separate designs now, so I have no idea what's wrong, and in fact I can't see much in common between them. They all have the exact same problem though. On the runway, during takeoff roll, a particular speed comes when the aircraft starts leering of the center of the runway. It should be noted that I am using FAR and I'm not really an expert with plane design, even though i enjoy it.

 

Right coming back to the case, I present to you most recent of these troublesome designs. It is in no way a spaceplane, as one can see by its engine. Hence it was designed with low-speed, low-altitude performance in mind. However, the runawy sideslip is fatal on takeoff, the plane ususally rips off a wingtip, rotates around it and comes to stop, sliding backwards.

https://imgur.com/a/zPpRW

 

 

 

 

 

 

I have checked all landing gears, they are straight. I will upload images of previous design with same flow (which was on old career save, earlier KSP version), once I find them.

 

 

EDIT: I have found some pictures of the other mentioned aircraft. Unfortunately I no longer have that save, so can't make any that would show CoM and CoL

Spoiler

I have also found a description of it that I've written. Both for RP purposes and as notes for myself in case i come back to it after a longer break:

Spoiler

Yak-3 is a modification of Yak-2. The fuselage remains the same, while the wings have been considerably enlarged. It is also powered by a different engine, trading gimball for more power. Additional air intakes were added as well. Designed for high altitude flight (contracts). It seems to have all merits and flaws of its predecessor.

Tested service ceiling: 23000m

Tested top speed: mach 2.8 at 12000m

+Very stable in flight, minor pitch oscillations at around 1.5 mach.

+Mediocre turnrate, good acceleration.

-No flaps

-Violent sideslip on takeoff and landing

Takeoff notes: Nose position: elevated at 1.6 degrees. No flaps. Apply full power. At around 60m/s plane starts swerving, correct immediately. Wings have enough ground clearance not to hit the ground if oscilations are under control. After takeoff, situaton immediately stabilizes.

 

Landing notes: Final approach speed at 115m/s. Flare upon passing runway threshold, touchdown main gear first. Touchdown with airbrakes deployed, gear brakes engaged. Focus on rudder control, plane will swerve violently when slowing below 70m/s.

 

Edited December 31, 2017 by JANXOL

[Rocket In My Pocket]

3 hours ago, JANXOL said:

It should be noted that I am using FAR and I'm not really an expert with plane design, even though i enjoy it.

I've never touched FAR, but let me give you my two cents on it anyways.

You should really completely master how planes work in the stock game before moving on to FAR, as anyone trying to help you has no idea if you even understand the basics, let alone all the things FAR changes.

Ultimately, you might get more help in the actual FAR release thread; that would be the place to find someone who actually uses FAR as well and knows if this problem is specific to FAR, or is just a plain ol' stock design issue.

Edited December 29, 2017 by Rocket In My Pocket

[JANXOL]

58 minutes ago, Rocket In My Pocket said:

I've never touched FAR, but let me give you my two cents on it anyways.

You should really completely master how planes work in the stock game before moving on to FAR, as anyone trying to help you has no idea if you even understand the basics, let alone all the things FAR changes.

Ultimately, you might get more help in the actual FAR release thread; that would be the place to find someone who actually uses FAR as well and knows if this problem is specific to FAR, or is just a plain ol' stock design issue.

I have been using FAR since KSP 0.21, so I'd say I at least understand basics. I'm not exactly sure what all of FAR specific data mean, but I have flown planes with and without FAR, though with FAR i haven't built any overly ambitious constructions.

The plane which I am searching for pictures of (earlier KSP version design which also had same problem) did fly. And quite well and stable, minus some pitch oscillations around mach 2. It also had the same issue as this plane (runway leer) but it could be corrected both on takeoff and landing.

 

It is safe to say I know basic principles of plane design. I am also somewhat familiar with real-world aircraft.

 

Edited December 29, 2017 by JANXOL

[OHara]

Make sure you deflect the control surfaces to pitch up for takeoff, to get the weight off the main landing gear.  Your elevators are small, so you might need to trim up significantly. 

If the landing gear is just barely forward of the center of mass, you can let the tail-wheel lift up a bit, but then the rolling resistance of the wheels acts in front of the COM so tends to make any swerves worse.  You depend entirely on the vertical stabilizer to keep you straight.  You need to keep the tail-wheel dragging for stability until you have enough aerodynamic forces  to keep you straight.

With FAR, the takeoff speed is higher so you need to overcome the destabilizing resistance of the main wheels for a longer time.  Make sure you have the main gear positioned so the plane rolling on three wheels is pitched up a little more than you need for takeoff.

The inboard ailerons, probably intended for flaps, are so close to the COM that they won't help much as control surfaces, so you should probably right-click to de-select pitch/yaw/roll control there.  If KSP tries to use them to pitch up, it would mostly push down harder on the main gear.

 

[bewing]

In general, the usual culprit is landing gear drag with the ground. It's important to understand that not all problems with airplanes are aerodynamic ones! Sometimes, they are "automotive" problems, instead. Sometimes you can overcome automotive problems by adding extra aerodynamic control, but there are other ways. If you were to remove all the aerodynamic surfaces from this vessel and then roll it down the runway, you would most likely see the same problem -- or maybe even worse.

So: ground drag. Basically you want to consider this vessel as a rover, and not as a plane.

Stability comes from having  low drag ahead of the CoM, and high drag behind the CoM. The design you show above has 2 gear ahead, and one behind. The one behind may be off the ground by the time you see your issue.

So in this case, the problem is that the gear that are ahead of the CoM have "too much" drag, and they are so far ahead of the CoM that the torque from that drag is enough to spin your plane. So, as I said, 2 answers: either add enough aerodynamic drag at the back end to overwhelm the issue. Or, go into the PAW for the front landing gear, change their Friction setting to Manual, then grab the slider and reduce their friction setting to .4 or below (which is my preferred solution).

 

[JANXOL]

42 minutes ago, OHara said:

Make sure you deflect the control surfaces to pitch up for takeoff, to get the weight off the main landing gear.  Your elevators are small, so you might need to trim up significantly. 

If the landing gear is just barely forward of the center of mass, you can let the tail-wheel lift up a bit, but then the rolling resistance of the wheels acts in front of the COM so tends to make any swerves worse.  You depend entirely on the vertical stabilizer to keep you straight.  You need to keep the tail-wheel dragging for stability until you have enough aerodynamic forces  to keep you straight.

With FAR, the takeoff speed is higher so you need to overcome the destabilizing resistance of the main wheels for a longer time.  Make sure you have the main gear positioned so the plane rolling on three wheels is pitched up a little more than you need for takeoff.

The inboard ailerons, probably intended for flaps, are so close to the COM that they won't help much as control surfaces, so you should probably right-click to de-select pitch/yaw/roll control there.  If KSP tries to use them to pitch up, it would mostly push down harder on the main gear.

 

Thank you for your input. I have tried larger elevators nad scaled up front gear to lift the nose, undortunately that didn't help. The flaps are below the wing, the rest is ailerons, and you have a point in saying these are too close to CoM.

 

 

34 minutes ago, bewing said:

In general, the usual culprit is landing gear drag with the ground. It's important to understand that not all problems with airplanes are aerodynamic ones! Sometimes, they are "automotive" problems, instead. Sometimes you can overcome automotive problems by adding extra aerodynamic control, but there are other ways. If you were to remove all the aerodynamic surfaces from this vessel and then roll it down the runway, you would most likely see the same problem -- or maybe even worse.

So: ground drag. Basically you want to consider this vessel as a rover, and not as a plane.

Stability comes from having  low drag ahead of the CoM, and high drag behind the CoM. The design you show above has 2 gear ahead, and one behind. The one behind may be off the ground by the time you see your issue.

So in this case, the problem is that the gear that are ahead of the CoM have "too much" drag, and they are so far ahead of the CoM that the torque from that drag is enough to spin your plane. So, as I said, 2 answers: either add enough aerodynamic drag at the back end to overwhelm the issue. Or, go into the PAW for the front landing gear, change their Friction setting to Manual, then grab the slider and reduce their friction setting to .4 or below (which is my preferred solution).

 

Thank you for the suggestions. I have made tests and made sure that the tail wheel is well on the ground when the swerve occurs. Since it's a prop plane, I was even wondering if proppeler inertia is simulated, since the plane always swerves left. However, the other design mentioned was a jet with tricycle gear and had the exact same issue. I will try the friction adjustment.

 

 

So far I have also noted that the swerve always occurs at 39m/s. It feels like its corresponding to the red stability derivative there, but I have no idea why or how to counteract it (if that's even right.)

 

Spoiler

And then all green:

EDIT: Lowering friction on forward gear postponed the swerving, but they still appeared during the roll, this time at about 46 m/s.

 

 

 

Edited December 29, 2017 by JANXOL

[overkill13]

I had two variants of a Mk2 Spaceplane using a tricycle landing gear setup.  One had a pair of Mk2 passenger modules and the other had two small Mk2 cargo bays in their place.  For reasons I have yet to explain, the passenger variant would swerve around 60m/s on the runway while the cargo variant, regardless of how much or how little cargo was loaded, would never swerve.  I suspected excessive nose gear loading and placed two gear side by side under the nose of the passenger variant and it solved the problem.  Not elegant, but I couldn't find any configuration that did not suffer from the drifting and have yet to understand why a heavy cargo version with more mass on the nose gear, avoided the issue.

[foamyesque]

This looks like (though I can't see the gear in your screenshot) a classic wheelbarrow problem, which is almost always caused by an interaction between your aerodynamic forces and your front gear. Specifically, it occurs when your main aerodynamic lift is behind your gears and so as you speed up, the gears take more and more loading. This moves your center of drag further and further forward and rapidly overwhelms vertical stabilizers. Sometimes you can get away with just tweaking the friction of the wheels, but that always feels like a slapdash patch that doesn't really address the root of the issue.

I would suggest taking the horizontal stabilizer off and putting a (minor) fixed angle on the wings, five degrees or so, potentially less given your relatively large wing area. Then adjust them along the length of the craft so's to put their CoL just ahead of your CoM, and put your nose gear directly underneath that (as you're building a tail-dragger). Once you've done that, you can then put back the horizontal stabilizer to pull the overall CoL back behind the CoM for standard stable flight.

The result, adjusted properly, should be a tail-dragger that will rotate to level and then take off with minimal or no command input beyond the throttle.

 

 

Edited December 31, 2017 by foamyesque

[JANXOL]

1 hour ago, foamyesque said:

This looks like (though I can't see the gear in your screenshot) a classic wheelbarrow problem, which is almost always caused by an interaction between your aerodynamic forces and your front gear. Specifically, it occurs when your main aerodynamic lift is behind your gears and so as you speed up, the gears take more and more loading. This moves your center of drag further and further forward and rapidly overwhelms vertical stabilizers. Sometimes you can get away with just tweaking the friction of the wheels, but that always feels like a slapdash patch that doesn't really address the root of the issue.

I would suggest taking the horizontal stabilizer off and putting a (minor) fixed angle on the wings, five degrees or so, potentially less given your relatively large wing area. Then adjust them along the length of the craft so's to put their CoL just ahead of your CoM, and put your nose gear directly underneath that (as you're building a tail-dragger). Once you've done that, you can then put back the horizontal stabilizer to pull the overall CoL back behind the CoM for standard stable flight.

The result, adjusted properly, should be a tail-dragger that will rotate to level and then take off with minimal or no command input beyond the throttle.

 

 

I will try that. There are more pictures in the imgur link at the top of the first screenshot, gear should be visible there. I am adjusting things here and there, and while I have gotten to a point where it takes off (barely), it still swerves a lot on the runway. Interestingly, it seems to be stable on landing.

 

 

I have also added pictures of the other aircraft mentioned to the first post. This one is a turbojet tricycle gear, also has the swerving issue. The gear is aligned to the ground, its the adjustable gear mod, hence the suspension is angled, but the wheel is straight.

[foamyesque]

9 hours ago, JANXOL said:

I will try that. There are more pictures in the imgur link at the top of the first screenshot, gear should be visible there. I am adjusting things here and there, and while I have gotten to a point where it takes off (barely), it still swerves a lot on the runway. Interestingly, it seems to be stable on landing.

 

Yeah, looking at the other images, your nose gear seems to be too far beyond your CoM. For a tail-dragger you want 'em just barely in front of your CoM, just enough so's that you don't faceplant on landing.

Edited January 1, 2018 by foamyesque

[Zeiss Ikon]

I can't say for certain that this is applicable to KSP (a lot of real world solutions aren't), but here goes:

For model airplanes, a taildragger (i.e. conventional gear, that is, two wheels slightly ahead of COM and one under the tail, often physically linked to or even mounted on the rudder) has a strong tendency to ground loop -- this is exactly what you're describing, where the aircraft swerves to one side during the ground run, either on takeoff, or after landing.  The solution to this is to build the main gear with a small amount (two or three degrees is enough) of toe-in.

The way this works is that when the aircraft inevitably starts to get a little out of alignment from its velocity, the wheel that's being pushed forward (by the yaw rotation) and downward (by the torque reaction to the side force from the gear) will generate more drag than the one that's having weight taken off.  This is because, with the toe in angle, it's scrubbing on the surface more than the other.  This drag acts as a brake and tends to restore the aircraft into a nose-forward attitude, automatically.

The reason I'm not sure if this will work in KSP is because I don't know how well KSP models the friction between a wheel and a surface.  If the friction isn't related to the "normal force" (as we used to call it in Statics & Dynamics courses), then toeing in the main gear won't do anything to help, but if more weight on a wheel makes it grip better (yet still allows it to slip a bit), this is the easiest solution to keep a "conventional gear" aircraft straight until you can pull the nose up and lift off.  As a bonus, it'll also keep the thing straight after a reasonably non-crash-like landing.

[FrostedShoe]

I don't use FAR, but I have a similar problem, planes never quite staying on centre.

You can adjust friction using advanced tweakables, I use it to lower the friction on leading landing gear.

You could try getting the plane into the sky with a rocket just to see if it can fly in a straight line, it might be torque from slightly uneven aero surfaces or engine balance. Most of my planes have a tendency to drift even though I make them symmetrical - no idea what that is all about.

[Gordon Fecyk]

5 hours ago, FrostedShoe said:

You can adjust friction using advanced tweakables, I use it to lower the friction on leading landing gear.

This helps a lot. My FAR-designed space planes always drift to one side or another unless I adjust the front gear's friction control. It might be 'wheelbarrowing,' or lifting at the rear while still dragging up front, putting undue force on the front gear.

Edited January 2, 2018 by Gordon Fecyk