Why is the 707’s yaw-damper-out lateral controllability marginal to nonexistent with flaps and spoilers extended?

Question

According to the NTSB report on the 1973 crash of Pan Am Flight 160, the 707-300C¹ has lateral controllability characteristics ranging from “minimal” to “none” if the flaps and spoilers are extended with the yaw damper inoperative:

Performance data for the Boeing 707-321C^[2] show that lateral control capability may be extremely limited, if not impossible, with an inoperative yaw damper, extended spoilers, and lowered flaps. [NTSB-AAR-74-16, p. 31 (marked as 29).]

On most aircraft, the yaw damper, although useful, can still be rendered inoperative in all flight regimes without causing the aircraft to become uncontrollable. Which makes sense - of the two main uses for the yaw damper (minimising Dutch roll and helping to counter adverse yaw), the first is of little consequence at the low altitudes where the flaps and spoilers are most likely to be deployed simultaneously (as Dutch roll is strongly damped out by the thick, dense low-altitude air), and is not generally a threat to controllability even when weakly damped at high altitude³ (although it excels at making the aircraft’s occupants airsick), while adverse yaw can be countered just as well by manual rudder inputs (something every jetliner pilot will have long since learned how to do, manual rudder usage being necessary for making coordinated turns in the small general-aviation planes pilots first learn to fly in) as by the yaw damper. Granted, the very high induced drag generated by flap and spoiler deployment (mostly the latter) would considerably aggravate the aircraft’s adverse-yaw tendency,⁴ but, assuming that the 707’s rudder was adequately sized to begin with, the aircraft should still remain completely controllable with the proper manual rudder inputs.

Indeed, if the 707-300C (and possibly the other variants as well) is truly uncontrollable in the lateral axis, or almost so, with flaps and spoilers deployed and the yaw damper inoperative, it is somewhat difficult to see how the aircraft could ever be safely flown without a functional yaw damper; takeoff and cruise wouldn’t pose too much of a problem, but landing, as a rule, involves a considerable amount of flap due to the resultant decrease in landing speeds, with spoiler extension being used to slow the aircraft down during the descent, and, if the yaw damper is indeed necessary for the 707(-300C) to be controllable with flaps and spoilers extended, one would expect a loss of control at some point during the descent/approach/landing sequence if it were flown without an operative yaw damper.

Why is the 707 so difficult to control with the yaw damper inoperative and flaps and spoilers extended?

---

¹: The 707-300C was the most-produced civilian 707 variant; the other variants (-100, -100B, -200, -300, -300B, and -400), which differed mainly in fuselage length and engine selection, presumably behave similarly to the 707-300C for the purposes of this question.

²: A Boeing 707-300C built for Pan Am (Boeing customer code *21).

³: I am aware of only one case where Dutch roll actually caused an accident (and the aircraft still would not have crashed were it not for a generous heaping of pilot error). This was a 707-200 training flight in 1959, conducted with the yaw damper off, where the aircraft’s motions were greatly amplified by improper and excessive flight-control inputs by the pilots, causing a loss of control; the aircraft was eventually recovered from an inverted dive, but three of the four engines were torn off at some point in the process, and the aircraft crashed short of an intended emergency-landing site when flight-control damage from the resulting fire forced the remaining engine to be throttled down in order to maintain control.

⁴: Flap deployment increases the induced drag of the wings directly; however, this should not, on its own, cause a large increase in adverse-yaw tendency, as the flaps are located on the inboard portions of the wings (there being more room for them and their associated paraphernalia in the wide, thick wing roots than nearer the thin, narrow tips, and there being less additional wing-bending moment if the increased lift from the flaps is added as close to the fuselage as possible), while the ailerons (the increased induced drag from the downward deflection of which is the cause of adverse yaw) are located on the outer, unflapped portions of the wings, as far from the aircraft’s centerline as possible, in order to achieve the maximum possible roll control authority.⁵ Spoiler deployment increases the wings’ induced drag indirectly, by forcing the aircraft as a whole - including the ailerons - to fly at a higher angle of attack in order to maintain flight, causing an increase in adverse yaw.

⁵: Indeed, flap deployment could (depending on the aircraft and the exact flight conditions) even help in this regard, as the increased lift from the deployed flaps allows the aircraft as a whole to fly at a lower angle of attack, reducing the aircraft’s adverse-yaw tendency; the 737, a later Boeing aircraft, shows an example of this.

Answer 1

Dutch roll is rolling, yawing and pitching together in a sort of Lazy 8-like pattern, and if divergent or really severe, there are very few humans that can make the perfectly timed in-phase inputs required to not make the motions worse, without a lot of advance training to get the pilot inside the response loop, which nobody does, so it's not a case of bad piloting when an airplane crashes due to a dutch roll event.

Since DR is related to the very powerful roll/yaw coupling of swept wing airplanes, the coupling is worse flaps down than flaps up, as the flapped wing presented straighter into the airflow by the yaw gets an even bigger lift boost compared to the trailing wing than with flaps up; it wants to roll even harder at the same yaw angle and speed (you learn this pretty fast the first time you do an engine cut simulation on a full flap go around and are late with the rudder). On top of all that what natural yaw damping there is, is reduced by the lowered airspeed.

I believe that most swept wing airplanes with dutch roll tendencies are worse flaps down than flaps up, although not necessarily bad enough to become uncontrollable. I know that the CRJ Regional Jet line has a stable dutch roll mode flaps up (you can excite a DR and if you do nothing it will settle down by itself... eventually) but is mildly unstable flaps down, although this isn't bad enough to prevent it from being dispatchable for 10 days with a Y/D channel inop with no restrictions (one channel working, and failure of that is not considered an excessive risk over a 10 day exposure period). Other aircraft may have MELs that allow single Y/D channel operation only with a flap extension restriction. It would be interesting to see what the 707's MEL says.

Also, adverse yaw is just one potential excitation mode. Turbulence can also excite a yawing motion that can get a dutch roll going and is probably a bigger one than AY.

In the case of the 707, its dutch roll mode flaps down was clearly divergent and severe enough to be considered uncontrollable by a normal pilot, and I'm pretty sure it's not the only one.

Answer 2

Why is the 707 so difficult to control with the yaw damper inoperative and flaps and spoilers extended?

It is a matter of time constants and stability of responses:

• Stable responses return to equilibrium position after a disturbance. As in a damped $2^{nd}$ order response. The amount of damping determines the number of overshoots before returning to equilibrium.
• A neutral response continues to sway indefinitely after a disturbance, at constant amplitude and time period. There is no damping.
• An unstable response sways with diverging amplitude away from equilibrium. There is negative damping.

An active controller, such as a pilot behind the flying controls, can generate control inputs that return an uncontrolled response to equilibrium more rapidly. How easy it is for them to do that, depends on the amount of damping in the system itself: too much damping, and their inputs must help the system to return faster. Too little damping, and they must counteract what the system does - but the systems oscillates, so their inputs must oscillate as well.

Humans, like all control systems, have a certain time delay between sensing a motion and being able to generate an effective control input. If the sway period is too rapid, this time delay causes the input to come at a time where it does not counteract the motion. To an extent, humans can learn how to time and dose their inputs in order to control an unstable system, if the time period is long enough - a helicopter in the hover is a classic example.

...it is somewhat difficult to see how the aircraft could ever be safely flown without a functional yaw damper;

• Landing without yaw damper is possible if the spoilers are not used.
• Landing without using the spoilers is possible on a stable approach.

It's just that the pilots need to know that the yaw damper is inoperative and that therefore the time constant of their responses must be adapted. Whether they are trained to do so in that particular situation is a different matter...

Answer 3

While each aircraft is different, especially for ones with high amount of sweep, some general trends can still be predicted:

1. The Dutch-roll damping decreases with decreasing speed (Ref. Etkins, Dynamics of Flight, Wikipedia).

2. Increasing flap increases the dihedral effect (Ref. ESDU 80034), which worsens the Dutch-roll damping once again (Ref. Etkins, Dynamics of Flight, Wikipedia).

MIL-STD-1797B specification recommends at least a damping ratio of at least 0.02 for all phases of flight for a Handling Quality Level 2, which is often the minimum threshold for a landing task (Level 3 is the minimum for continued safe flight). If the Dutch-roll damping falls too low, it can certainly be dangerous.