I was on a flight with a regional airline a while ago in a Bombardier airplane (if I recall correctly) where the airplane would slowly and gently rock its wings just a little bit on the whole flight. I wasn't a long flight but wasn't very short either at roughly 2 hours.

It was very comfortable and the rocking made for a very gentle cradle feeling 🙂

My question is, why would that be? Is it economical to do so for any reason? Is it possible that the airplane had no autopilot and the pilot was flying it by hand? What could cause such a behaviour?

  • 4
    $\begingroup$ Sounds like the dutch roll mode was working. Were you high and fast at the time? $\endgroup$
    – MikeY
    Commented Mar 31, 2019 at 23:52
  • 1
    $\begingroup$ Are you talking about this? It is pretty common for airplanes. $\endgroup$
    – havenoidea
    Commented Apr 1, 2019 at 4:29

1 Answer 1


I can give you two theories. One is it's a very very mild case of "overcontrolling" of the autopilot roll servo as it chases a perfectly wings level condition and continuously applies a teeny bit too much left, ooops, then a teeny bit too much right, oops, and back and forth.

The aileron control circuit is cables running to hydraulics at the ailerons, with a spring loaded roller in a curvy V shaped cam device that centers each side of the aileron cable circuit. The breakout force to move the control circuit out of neutral (forcing the roller from the center V of the cam) is fairly high and this can actually be a problem for pilots hand flying at higher speeds where the ailerons are quite snappy - applying enough force to break out the circuit and move the wheel left gives too much, and back the other way... too much... and back and forth you go. It can be a pain.

The autopilot servo is basically a little electric winch connected into the aileron cable circuit, doing the same thing the pilot does. It could be that something in the electronic control loop, or mechanically within the servo, can cause this back and forth oscillation on some airplanes, sometimes. It may be related to wear and backlash within the autopilot servo.

The other theory is it's more or less something similar, but in the yaw mode with the yaw damper not able to fully dampen out a dutch roll mode because the amplitude was within the yaw damper's sensitivity band. Dutch roll is rolling due to yaw, which in swept wing jets is very strong, such that it can set up continuous self energizing oscillation. It could also be related to wear and backlash in the yaw damper actuator, which is a pair of electric linear actuators (little electric screw jacks), in the push rods going to the rudder hydraulics.

If you saw the wing tip just going straight up and down, it's mostly likely the first theory. If you could see the winglet moving in a little circle (indicating rolling and yawing), it's more likely the second theory.

I can really only theorize because the issue was never fully explored on the RJ program as this phenomenon was random, very minor and not dangerous, and was never complained about by operators because only someone sitting near the wings that can look straight out at the winglet ever notices it. It was never bad enough to get anyone sick for example, at least on the RJ fleet.

  • 5
    $\begingroup$ For those who want to know the little tiny left, little tiny right would be a bad hysteresis. Similar to how you ac/heat does not come on right at the correct temp. It waits for it to be out of a range, higher for heat to turn off, lower for ac to turn off. $\endgroup$
    – NDEthos
    Commented Apr 1, 2019 at 6:36
  • 2
    $\begingroup$ @NDEthos: Are they really using simple thermostat-style controllers instead of properly configured PID controllers? $\endgroup$
    – Michael
    Commented Apr 1, 2019 at 7:49
  • 2
    $\begingroup$ I just used a thermostat as an example. Even with a PID I would assume the concept of hysteresis still applies. $\endgroup$
    – NDEthos
    Commented Apr 1, 2019 at 13:21
  • $\begingroup$ But shouldn't an aircraft be roll-stable at cruise naturally? $\endgroup$ Commented Apr 1, 2019 at 14:53
  • $\begingroup$ It's way more complicated with swept wing airplanes because of the strong roll/yaw couple caused by the sweep. The sweep is a bit like having way too much dihedral; any yawing motions set up sympathetic rolling motions (setting up a kind of S shaped flight path - dutch roll) and they may dampen themselves out, but in some configurations the damping is too weak, or is negative (dutch roll), necessitating a yaw damper system that works the rudder to actively stop or prevent yawing movements. I suspect that the condition you were seeing was a feedback or hysterisis loop within the YD system. $\endgroup$
    – John K
    Commented Apr 1, 2019 at 16:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .