Plane roll stability

[aluc24]

Hello,

In KSP 1.04, I'm trying to make SSTO spaceplanes. I have already made 3 or 4 successfully. I always use MechJeb autopilot.

The problem I'm having is that sometimes, my planes do not have enough roll stability. Meaning that even a slight roll movement will make it roll from side to side, and MechJeb autopilot is unable to "calm down" these roll oscillations.

I usually have a big wing aft will ailerons (that control roll and nothing more), smaller elevators at the front of the spaceplane (control only pitch), and and rudder at the tail that controls yaw.

What I've tried already:

1. Making ailerons smaller or larger;

2. Moving them closer or further from the centerline;

3. Making main wings at an angle (slight V shape).

Sometimes the plane has roll stability, sometimes it doesn't. I can't pinpoint the exact cause of it, and I've been guessing for weeks. It's a bit of shooting in the dark now.

Can anybody help?

[pincushionman]

3 (called dihedral) will not work unless you have more than one wing section along the length ("beam") of the wing. This is because in KSP the aerodynamic forces are applied at the wing root, not its center. So the moment arm is very short. With multiple wing sections the outer ones have their forces applied partway down the wing, so it kinda works.

[aluc24]

Okay, I see, so how can I fix roll stability then?

[cantab]

https://en.wikipedia.org/wiki/Dihedral_(aeronautics)

https://en.wikipedia.org/wiki/Aircraft_dynamic_modes

https://en.wikipedia.org/wiki/Dutch_roll

This might help. Some particular things to notice:

Starting with level flight, if a small roll angle tends to steepen in one direction that's an unstable "spiral mode", or what you'd normally think of as instability in roll. This is caused by too little dihedral effect. Now a slightly unstable spiral mode that only slowly steepens the bank isn't a problem.

If on the other hand a small roll angle tends to wobble back and forth, and becomes difficult to handle, that's probably "Dutch roll" and is caused by too much dihedral effect. Again a mild Dutch roll that dampens out isn't really an issue, it's a problem when it's unstable and gets bigger and bigger.

Understanding which problem you have will obviously help.

Besides the things you mentioned, vertical placement of the wings and the wing sweep can have an effect. I had one tailless, canardless delta spaceplane with problematic roll stability that I corrected by increasing the wing sweep a little, meaning the trailing edge was slightly swept back. That was in old FAR though.

[RexKramer]

Does the problem become worse as your speed increases? If so, reducing the size of roll control surfaces may help. In other words, use smaller or less ailerons. If you are already down to the smallest aileron, move them inboard to get closer to the COM if possible.

[Thingymajigy]

Uh, you could try turning off Mechjeb

[Val]

... MechJeb autopilot is unable to "calm down" these roll oscillations. ...

In my experience this is always because of too much aileron control authority.

My suggestion is to make them even smaller and closer to centerline, than you already tried, or to disable/remove them entirely and let the canards be both elevators and ailerons.

[AeroGav]

Also, something which the tutorials fail to mention is that yaw instability causes roll instability. Make sure you have a large enough tail fin, far enough aft of the CG. This is an issue on canard designs, which often have long noses but short tails.

[Hodo]

Are you using FAR or stock?

If you are using FAR I would suggest checking your roll stability.

[DancesWithSquirrels]

I've also run into this, with too much aileron or with angled vertical tails - V-tails in particular can be bad if you've left roll enabled.

Use a smaller aileron out at the ends of the wings, make sure you have enough vertical tail for stability, turn off roll on your vertical control surfaces and yaw on your ailerons (use the tweakables in the SPH) - all of those will help.