Spaceplanes randomly veering off of runway.

[Darkeldar]

It veers either right or left, with or without SAS. Sometimes the spaceplane survives coming off a left side veer and I've been able to take off. And trying to correct the veer usually results in a worse crash as the correction ends up as a sharper turn than expected, rolling the plane. I was trying to see if the Lazor mod could manage my cruising flight, but I haven't been able to take off and get to altitude to attempt to use these features.

Edited November 23, 2013 by Darkeldar

[Smidge204]

Are any of your landing gear not perpendicular to the runway? That is, did you use symmetry to place two wheels on a fuel tank? Wheels at angles can cause problems.

=Smidge=

[vetrox]

The reasson i have found is that on bigger craft. If stuff wobbles (wings, fuselage etc) then it qill make you veer off everywhere. In the same principle as rocket design, add some struts. Even if its strutting the wings to the fuselage to reduce wobbley wobbley.

[Kielm]

I've had a similar problem.

I'll bet it's a single engine plane with wheels on the wings?

The issue could be that one of wheels (either the left or right) is not always in contact with the ground. Friction caused by having one of them in contact but not the other will veer the plane left or right. Correcting it can add a sudden friction to one side and remove it from the other, causing a severe sideways pull on your plane if done at speed.

Try reducing the number of wheels down to three if it's a small plane. If that doesn't cut it, or you need more wheels, try keeping the wheels on the fuselage only. The distance from the centre of thrust/mass to the drag will be reduced, meaning the resultant moment (force) from any drag left or right should be smaller.

[linkxsc]

Angled wheels, A wheel being slightly off center (turn on angle lock), unstable aircraft (make sure your CoG is in-front of your CoL), Not enough lift (only really shows a problem when oyur plane gets really fast and still can't get off the ground)

[Traches]

It could be a few things. First check that your gear is perfectly straight up and down, and pointing perfectly forwards. Other things that will cause that: Your tail lifting off before your nose, your plane sits nose-down on the runway which creates downforce and puts enough weight on the gear to make it wobble, or your gear/whatever it's mounted to straight up isn't strong enough.

There's an excellent post that addresses those issues as well as others you might have:

http://forum.kerbalspaceprogram.com/threads/52080-Basic-Aircraft-Design-Explained-Simply-With-Pictures

[Darkeldar]

So Moar Struts! Spike4 already has a tricycle landing gear set up. The plane does twist a good bit when it rolls off the side of the runway. so I'll see if that helps.

CoL is located inside the rear of the CoM sphere, but it is pointed slightly forward.

I may go back to the longer gear on the front to raise my nose again, though.

Edited November 22, 2013 by Darkeldar

[Zephram Kerman]

I avoid the situation by putting CoM, CoL, and main wheels very close together. If the plane can rotate at 50 m/s, it doesn't have a chance to wander away.

[Rusty6899]

Do you use angle snap to place your central wheel? That could be it.

[gslarmour]

I've found that added IRW's at the fore and aft of the plane help stabilization. Sure, they add some weight, but they tend to counterbalance each other in my experience.

[thereaverofdarkness]

A solution that works for me is to watch the runway very carefully and catch the craft as it starts to veer, adjusting it very slightly to keep it from doing so. Your craft needs to be fairly stable already for this to work.

[Darkeldar]

Adding moar struts stiffened up the spaceplane enough that it was able to takeoff. I put the taller landing gear back on so I don't need the entire runway.

Now I'm off to the Lazor thread to ask my questions there.

[Murph]

Very often I think this can be caused by a plane lacking sufficient structural rigidity, so flexing causing what should be a good wheel geometry to become bad. A couple of carefully placed struts can often help fix that aspect of it. You can watch for it flexing when it's initially dropped onto the runway, and if you apply pitch/yaw/roll controls while it's stationary. If you see much flex in either case, that could be the root cause of the problem. The stock Aeris 4A, for example, shows notable flex around the RCS tank, docking port, and IAS.

I disagree quite strongly with adding reaction wheels to solve stability on the runway during takeoff runs. To me, that suggests that there's a far more fundamental issue with the structure and/or balance of the plane, if you need extra torque just to hold it steady during the takeoff run. To me, the reaction wheel torque should only really be important once you're above the altitude range where aerodynamic control surfaces can operate effectively, sized appropriate for control from 30km up to orbit, and in orbit. That's assuming we're talking about a modestly sized craft which looks vaguely like a plane, of course. Giant or bizarre craft may well require a more unconventional approach.

Another issue which is difficult to be certain about, but I became highly suspicious of, is the highly erratic prograde vector which can sometimes be seen on the nav ball when the plane rolls slightly after physics starts, but before you have actually applied power, etc. I can't be certain, but I think that if you turn on SAS at the wrong moment, while the vector is jumping around, SAS can lock onto some wildly inappropriate vector, and actually drag the plane off trying to follow it. The solution I use is to hold the brakes on, then turn on SAS, apply power, then release brakes. It's difficult to say with certainty, but I'm pretty sure that holding the brakes on like that has eliminated most random runway excursions on takeoff for me (for craft which are otherwise generally ok for takeoff runs).

Edited November 23, 2013 by Murph

[HeleleMama]

On 2013/11/23 at 2:38 AM, Kielm said:

I've had a similar problem.

I'll bet it's a single engine plane with wheels on the wings?

The issue could be that one of wheels (either the left or right) is not always in contact with the ground. Friction caused by having one of them in contact but not the other will veer the plane left or right. Correcting it can add a sudden friction to one side and remove it from the other, causing a severe sideways pull on your plane if done at speed.

Try reducing the number of wheels down to three if it's a small plane. If that doesn't cut it, or you need more wheels, try keeping the wheels on the fuselage only. The distance from the centre of thrust/mass to the drag will be reduced, meaning the resultant moment (force) from any drag left or right should be smaller.

Confirmed, "Wheels on wings" is the cause of my  veering problem. Thanks..

[Tokoshoran]

On 11/22/2013 at 4:45 PM, Darkeldar said:

Adding moar struts stiffened up the spaceplane enough that it was able to takeoff. I put the taller landing gear back on so I don't need the entire runway.

Now I'm off to the Lazor thread to ask my questions there.

Yeah, struts allow a flat surface to connect to, and if you put enough off to the sides they'll allow you to spread your wheels out, you can even turn while driving without needing to worry about much, in case something forces you to veer anyways, or are driving a "car" around for surface readings.

[Mark Kerbin]

I found the easiest way to deal with this is to keep the weigh (ie fuel tanks) more central. The farther out they are, the more bouncy you start to get on the outer wheels causing  Rapid unplanned disassembly.

[Lego_Prodigy]

All of my own plane designs suffer from this once they get up to about 160 m/s, whether it be flying wing, SSTO, or just a cruising plane.

All of my planes have at least two engines, as thats more reliable as I've found than only one engine due to more even distribution of thrust along the CoM.