Please help me better understand what causes a jet airliner (such as the Boeing 737-500) to roll toward the "weaker" engine when power is reduced on one side.

Such as in the situations featured in 6:36 through 7:54 and 21:50 through 23:20 in this "Mentour Pilot" YouTube video about the January 9, 2021 crash of Sriwijaya flight 182. (Links to set to start at segments: 1, 2.)

My specific questions center around how the yaw damper affects the situation, compared to what we'd see in a twin-engine airplane with no yaw damper.

Is the situation simply that the asymmetrical thrust creates a yaw torque that directly leads to the aircraft flying in a slipping condition, as would be measured by a yaw string at the nose? Which then creates a roll torque via aerodynamic coupling between slip and roll, due to sweep and dihedral? Is this the dominant source of the roll torque toward the weaker engine?

How is the yaw damper responding to the asymmetrical thrust condition? Is it attempting to maintain a zero-sideslip condition, as would be measured by a yaw string? Or as would be measured by a slip-skid ball (inclinometer)? But is unable to do so? Why is it unable to do so?

Or is it likely that the sideforce from the deflected rudder, acting high above the CG of the aircraft, is in fact the dominant source of the roll torque toward the weaker engine? (It seems that this is what we might expect if the yaw damper was in fact able to maintain a zero-sideslip condition, as measured either by a yaw string, or by the slip-skid ball.)

Or alternatively, could the effect of the jetwash on the airflow near the wing be the dominant source of the roll torque toward the weaker engine?

  • $\begingroup$ Related: aviation.stackexchange.com/questions/6391/… -- note that the answers do not seem to agree about exactly what the inputs to the yaw damper actually are. $\endgroup$ Feb 17 at 14:42
  • $\begingroup$ What we're really talking about here is skid, not slip. But from an engineering perspective , as opposed to a pilot's perspective, any condition where the nose is pointing differently than the airplane is actually moving, yaw-wise, is often referred to as slip or sideslip, without intending to specify whether nose is yawed toward "inside" of actual direction of curving flight path (skid), or toward outside (slip-- also applies to linear flight path.) (There's even an ambiguous area where flight path is curving toward high wingtip-- do you call that a slip, or a skid?!?) $\endgroup$ Feb 20 at 14:46

3 Answers 3


The reason is that yaw dampers normally have only limited authority, maybe 1/4 to 1/3 the available rudder travel, since they are only intended to counteract aileron adverse yaw and dampen any dutch roll tendency.

It's just sensing lateral accelerations and working the rudder to remove them; it doesn't know why. All that happens is the yaw power of asymmetric thrust, beyond a point, will exceed the YD's range and ability to correct.

If you ease back the thrust on one engine while in cruise, the YD will apply rudder to keep the "brick" (skid ball) centered, but it will soon hit its authority limit at some level of thrust differential, and unless you start to apply more rudder with your feet, the airplane will start skidding toward the low thrust engine. With swept wings, any amount of yaw, especially at high speed, you start to get a roll into the yaw immediately.

  • $\begingroup$ Got it. Last two paragraphs of question then become superfluous as those effects are presumably only minor relative to the rest. $\endgroup$ Feb 17 at 15:08
  • 2
    $\begingroup$ Yes. It's like you hired a troll who sits in a closet with a skid ball and rudder pedals, and you tell him just keep the ball centered at all times. On FBW airplanes the YD function is integrated into the multiple FBW computer channels, and I don't recall if its function is still limited the way a mechanical one is. You also often have rudder travel limiter system to limit pilot input at high speed for structural protection. Or, as in the CRJ200, the 3 rudder hydraulic PCUs simply can't push hard enough to get full travel above a speed, and a mechanical travel limiter wasn't necessary. $\endgroup$
    – John K
    Feb 17 at 16:18
  • $\begingroup$ This is interesting though, including link in first comment below that answer -- aviation.stackexchange.com/a/6393/34686 -- (doesn't pertain directly to main thrust of question, which you've answered, but it seems that sometimes yaw rate is an input as well? ) $\endgroup$ Feb 17 at 16:24
  • $\begingroup$ Also of interest -- aviation.stackexchange.com/questions/90943/… -- statement "Because it would oppose any rudder applied to counteract the swing from a failed engine, it was only engaged..." -- why? I'm figuring what's really being counteracted is the pilots attempt to quickly "swing" the nose back to a condition closer to zero slip, after it's already "swung" into a slipping condition? Again because yaw rate is an input? Not sure, or maybe there is just something wrong with statement-- $\endgroup$ Feb 17 at 16:26
  • $\begingroup$ But that's another question--may have to ask it-- $\endgroup$ Feb 17 at 16:30

Step back a little. It really is much easier then you are making it out to be. You know that a rudder induces yaw by introducing asymmetrical drag around the vertical axis, then to damp unintended yaw just means using some counter-rudder. That is all a Yaw damper does. Its sensors detect unintended yaw and operate the rudder (or part of it) to counter the effect. A rudder is not the only way of inducing yaw, anything that adds an asymmetrical force around the vertical axis will induce yaw, which is why without a yaw damper, you have to 'step on the peddle' when rolling a plane. If you have a yaw damper, it will step on the peddle for you. However, there are limits to how much yaw a damper can overcome. Push one engine to max or pull the other to idle and you have a lot of asymmetrical force to counter. As the aircraft yaws in response, the thrust from the dominant engine continues the accelerate the rate of rotation and if it is not addressed early enough, it may induce yaw beyond the ability of the yaw damper or even the rudder to address.

  • $\begingroup$ @quietflyer - that is correct. Balanced flight with two engines, and thrust from each is parallel to direction of motion, so yaw forces are balanced. Add yaw and thrust is no longer parallel or balanced which increases yaw further. $\endgroup$
    – Paul Smith
    Feb 17 at 15:06
  • $\begingroup$ @quietflyer - yaw is rotation around the vertical axis. Yaw rate and sideslip are different things again. $\endgroup$
    – Paul Smith
    Feb 17 at 15:10
  • $\begingroup$ Let us continue this discussion in chat. $\endgroup$ Feb 17 at 16:34
  • $\begingroup$ @quietflyer - Your specific question "How does yawing in and of itself create a thrust asymmetry,..." The application of thrust (or drag) that does not go through the center of lift induces a turning moment. In twin (and quad) engine planes that moment is balanced and parallel to the velocity vector. When the aircraft yaws, the thrust is no longer parallel to the velocity vector so introducing another turning moment. $\endgroup$
    – Paul Smith
    Feb 17 at 17:03
  • $\begingroup$ New content has now been posted to the chat link chat.stackexchange.com/rooms/142952/…, and I'd encourage any ASE users wishing to make "discussion"-type comments to this question to join in there. $\endgroup$ Feb 17 at 17:17

is it likely that the sideforce from the deflected rudder, acting high above the CG, is in fact the dominant source of the roll towards the weak engine?

It was noticed that the aircraft actually had roll input in the opposite direction of the way in actually was rolling due to thrust asymmetry "skidding" the aircraft.

above cross-over speed, roll control is sufficient to counter-act rolling forces created by a rudder "hard-over", or full rudder deflection.

So, it seems the yaw damper, with its limited travel, is less responsible than the wing sweep and dihedral, and any roll resulting from rudder inputs by design can be counter-acted with the yoke.

"raise the dead"

Slipping away from a dead engine helps counter-act yawing forces from the live one.

could the effect of jet wash ... be the dominant source of roll torque?

This is actually more true of wing mounted propellers. The live engine prop wash over the wing will cause it to roll into the engine out side. It would be interesting to compare the rolling effect of a straight wing prop vs a swept wing jet "live" side.

what are the control inputs if one engine completely shuts down?

This would seem to be the path to fully understand what control options competant pilots would have in this situation.

  • $\begingroup$ One little observation-- "It was noticed that the a/c actually had roll input in the opposite direction of the way in actually was rolling due to thrust asymmetry "skidding" the aircraft."-- yes, due to the autopilot. But not full aileron input-- the autopilot was limited to 19 degrees of yoke travel for roll (that is stated at 21:06, just before the start of the second segment mentioned in question.) So I don't think that simply be virtue of being over "crossover speed", we can be certain that the rudder wasn't responsible. However, the accepted answer has made a good case for this. $\endgroup$ Feb 18 at 14:04
  • $\begingroup$ @quietflyer yes, I will edit to "somebody has to fly the plane". Interesting that coordinated rudder and aileron input into the live engine can be helpful. The note is a good point. $\endgroup$ Feb 18 at 14:50
  • $\begingroup$ @quietflyer we need the engine out procedure for the 737-500. Above "cross-over" speed, a fully deflected rudder can be controlled with aileron input. In this case, can an engine out asymmetry be controlled with good rudder and aileron. I would think so. $\endgroup$ Feb 18 at 14:55
  • $\begingroup$ Cross-over speed relates authority of the ailerons and the rudder. VMC relates authority of the rudder to oppose a failed engine. Speed increases control authority, but the speed at which the first two balance isn't necessarily (or typically) the speed at which the latter two balance. In the Classic 737, your clean crossover speed is around 200 to 210 knots; your VMC would be well below that. $\endgroup$
    – Ralph J
    Feb 18 at 17:19
  • $\begingroup$ @RalphJ that is excellent info. The 737 engine out procedure number 1 is "maintain directional control". Perhaps the OP is pointing out that automated systems could have done more to control the engine out. But this leads to a deeper question, is lack of work load as dangerous as too much. Especially in the second video, the pilots did not seem to be focused on flying the plane until it was too late. $\endgroup$ Feb 18 at 19:18

You must log in to answer this question.

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