Kerbal Space Program

If you could provide a mission with this craft already on Eve, I'd appreciate it. That way, I could test this in a pure stock+DLC installation.

I did test this craft on Kerbin and never got the lift vector switch. HOWEVER, I have seen such a switch on Kerbin with some of my own designs. The underlying cause is that motorless servos/rotors are no longer frictionless bearings. Thus, the top and bottom halves of a powered rotor sitting on such a "bearing" do not rotate at the same speeds. The blades closest to the "bearing" are always slower than the others. This results in the following behaviors:

• Contra-rotation using 1 powered and 1 unpowered rotor no longer totally cancel torque. Observe that your craft here will roll about its longitudinal axis even in space with the blades flat and no aerodynamic forces being generated at all.
• Because the blades are moving at different speeds relative to the air, they require different amounts of pitch to provide the same thrust. For a certain combination of blade speed, blade pitch, and air density, thrust always reverses for any set of blades.

That's why this happens. Because the blades are moving at different speeds but have the same pitch (the set closest to the "bearing" is always slightly slower), the slower, lower set reaches this critical situation and reverses thrust.

So, this is not a bug with rotors, it's a bug with motorless rotors/servos no longer being frictionless. This is particularly crippling for autogyros.

The problem doesn't manifest on Kerbin. I think it has to do with the thick atmosphere on Eve. I'll attach a mission that starts on Eve.

I'm not sure about your diagnosis however, I think it might be a separate issue. The vectors flip back and forth abruptly like a sign got reversed; there is also the framerate slowdown to single digits.

I can confirm the strange reversal of lift, using the mission file attached earlier.
You can see the lift change sign, on all blades simultaneously, even with collective set very near zero blade-pitch.
The alt-F12 option to show lift in right-click menus shows the same reversal of sign as do the alt-F12 lift vectors.

I cannot see any significant friction (mentioned in the comment above) in the free rotor in the design attached above, nor in my own free-rotating pairs of counter-rotating props. So, I suppose we should put the title back to describe the problem that the attachments demonstrate.