Swept Wing: Why the Stall Starts at the Tip and Why the Aircraft Pitches Up

Extrados d'une aile en flèche vu d'un hublot, générateurs de tourbillons visibles

In a straight-wing trainer, release the yoke and the nose drops: the penalty for the stall contains its own correction. On a swept wing, the same excess angle of attack produces the opposite: the attitude keeps rising even as the wing loses lift, and nothing brings the nose back down. This is not a design flaw but a geometric consequence: separation starts at the wingtip, well aft of the center of gravity, and the lift deficit it creates acts as a nose-up moment. From washout to the stick pusher, everything a modern wing carries corrects or replaces what sweep took away.

Spanwise flow on a swept wingPlan view of a swept half-wing. On the upper surface, the flow does not follow the chord: it drifts outward, increasingly so with distance from the wing root. The boundary layer therefore thickens as it accumulates toward the tip, a region already the most heavily loaded. Separation appears there first, while the root is still lifting.fuselagechorddrifts toward the tipseparationthe root is still lifting
Sweep moves the problem outward. It is not that the tip is poorly designed: it is the flow that delivers a boundary layer already thickened by everything upstream. The stall therefore starts right where the aileron works.

What sweep does to the flow

Project the freestream velocity onto the leading edge: it splits into a component normal to that edge and a component parallel to it. Only the first one sees the airfoil and generates lift; the second slides along the span without producing anything. At a given true airspeed, the useful dynamic pressure seen by the section is thus reduced roughly by the square of the cosine of the sweep angle: the wing needs more angle of attack for the same load factor, the maximum achievable lift drops, and the stall speed rises. That is the price paid for pushing back the shock wave.

The parallel component is not inert: it carries. The isobars of a swept wing are oblique, and there is a pressure gradient directed from the root toward the tip. The boundary layer, the layer of slowed air stuck to the skin, obeys it all the more readily the slower it is: it drifts outward, and what leaves the root accumulates further out. Toward the tip it is therefore thicker, less energetic, less able to climb an adverse gradient. Sweep thus places the most fragile region exactly where local demand is highest.

On a swept wing, the air flows outward along the wing, making the tips more likely to stall first. This is because the tips are already working harder and the air there is slower and less stable. When the tips stall, the wing loses lift in a way that makes the nose of the plane rise, which can be dangerous.

On an Airbus A320 at 15,000 feet, a stall at 130 knots (IAS) will cause the outer wing panels to stall first due to sweep. The nose will pitch up, and the aircraft will require immediate action to recover, such as pushing the control column forward to reduce angle of attack.

Why the tip arrives first

Three contributions add up, and the manufacturer does not treat them with the same tools. First, boundary-layer drift: it depletes the wall flow of the outer sections before they reach their own critical angle of attack. Next, the lift distribution: the longitudinal offset of the sections modifies the induced field, increasing the effective angle of attack outboard and reducing it at the root, so the local lift coefficient peaks between mid-span and the tip. Finally, taper: a short chord at the tip requires, for a given local load, a higher coefficient — a multiplier of the problem rather than an independent factor.

Design answers with washout. Geometric washout — the root set at a higher incidence than the tip — pushes back the onset; aerodynamic washout does the same by varying thickness ratio, camber, and leading-edge radius along the span. The goal is never to avoid the stall but to choose where it starts: as close as possible to the root, so the wake buffets the tail and gives an identifiable shudder, where the loss of lift costs the least in pitch. In cruise, the outer sections operate at a higher local Mach number and separation can be triggered there by a shock: the trigger changes, not the consequence.

Retreat then advance of the center of pressure during tip stallPlan view. In normal flight, the lift resultant sits aft of the center of gravity and the aircraft is trimmed nose-down. When the wingtips stall, lift disappears from a region located aft, because of the sweep: the resultant moves forward. The moment arm about the center of gravity changes sign and the aircraft pitches up on its own, without any pilot input.center of gravitylift, healthy wingafter tip stallnose-up moment
The pitch-up is not a pilot reaction, it is a geometric consequence. On a straight wing, losing the tips barely shifts the resultant longitudinally. On a swept wing, the lost region is also the furthest aft: the aircraft digs itself deeper into the stall on its own.

From tip stall to nose-up moment

On a swept wing, the outer sections are not just laterally distant: they are markedly aft. When the outer part of the half-span stops lifting, the point of application of lift shifts toward the root, hence forward. A significant force whose point of application moves forward relative to the center of gravity produces a nose-up moment. The mechanism is geometric and has no equivalent on a straight wing: there, the same loss shifts the center of pressure laterally and almost not longitudinally — it produces roll, not pitch-up. The tailplane adds to this: the downwash reaching it changes, most often in the direction that reinforces the tendency.

A silhouetted aircraft with swept wings in flight against a sunset sky.
Swept wings stall at the tips, causing pitch-up tendencies.

The decisive point is the sign of the loop. A pitch-up increases the angle of attack, a higher angle of attack extends the separation inward, a more extensive separation moves the point of application further forward. This is not an offset to be recovered with trim, it is a divergence: in this regime, the slope of the pitching-moment curve with respect to angle of attack changes sign, and the aircraft is no longer statically stable in pitch. The nose will not come down by itself because nothing in the airframe asks it to.

There remains asymmetry: a surface condition, a slight sideslip is enough to make one wing depart before the other. The roll occurs precisely where the ailerons are, in the already separated zone. Lateral authority is thus degraded exactly when it would be needed, and a control surface deflected down on the low wing increases the local angle of attack there.

Stall: low-taper straight wing versus tapered swept wing
Criterion Low-taper straight wing Tapered swept wing
First separation Root or mid-span Outer half-span
Longitudinal effect Nose-down moment Nose-up moment, self-sustaining
Lateral effect Moderate, ailerons in attached flow Roll departure, ailerons in separated zone
Natural warning Clear and early buffet Late buffet, mistaken for turbulence
Tailplane at high angle of attack Authority generally retained Possible blanking depending on tailplane height
What must be added to demonstrate the required behavior Often nothing, or a stall strip at the root Shape fixes, warning system, often a pusher or angle-of-attack limiting
High-mounted tailplane in the wake of the stalled wingSide view at high angle of attack. The thick, turbulent wake of the stalled wing rises behind the aircraft. A tailplane mounted high on the fin sits right in it: it receives only a disorganized, low-energy flow. The elevator then loses most of its authority, at the exact moment when releasing the controls to reduce angle of attack would be needed.relative windseparated waketailplane in the wakethe elevator commands air with no energy left
The second stage gives no warning of its own. As long as the tailplane stays in clean flow, the yoke keeps its authority and recovery is routine. Once it enters the wake, the same input produces nothing more — and it is the aircraft’s geometry, not the maneuver, that decides the threshold.

The second stage: when the tailplane drops out

As angle of attack increases, the wake of the wing and fuselage thickens and rises. What follows depends on the position of the horizontal tail. A low-mounted tailplane stays out of the wake longer and retains usable dynamic pressure; a tailplane mounted high on the fin finds itself, at high angle of attack, exactly where the wake is heading. Once it enters it, local dynamic pressure collapses, elevator effectiveness with it, and the nose-down moment is no longer available. The aircraft settles into a stable state rather than a transient: very high angle of attack, high rate of descent, elevator with no effect. This is the deep stall, with no aerodynamic recovery.

Thrust enters the equation. On most installations, the thrust line passes below the center of gravity: adding power produces its own nose-up couple. Depending on the propulsion architecture, an intake positioned at high angle of attack can operate in badly disturbed flow, and the thrust actually delivered is then not necessarily what the throttle indicates. Adding power before reducing angle of attack is therefore not a neutral action, which is why recovery procedures give priority to reducing angle of attack over adding power.

The trap. On a swept wing, neutral stick is not a neutral state: in the pitch-up regime, angle of attack keeps growing on its own, and hesitation costs authority — all the faster the higher the horizontal tail is mounted. Two reflexes learned elsewhere become harmful: countering roll with ailerons, while that control surface works in a separated zone and further loads the low wing; and adding power before having unloaded the wing. A trim trap compounds this: a long autopilot flight near the limit can leave the aircraft trimmed nose-up, and the elevator must then overcome the trim before it can be useful.

What the manufacturer sets against it: shape first, artifice second

The first line of defense is geometric and invisible in flight: washout, limited taper, and the spanwise distribution of leading-edge radius decide the local lift-coefficient profile along the span, hence where the first separation occurs. The second acts on the flow through two distinct principles: fluid fencing, a longitudinal vortex that opposes boundary-layer drift — wing fence, sawtooth, vortilon; and reinjection of momentum at the wall, the principle of vortex generators. The third belongs to high-lift devices: the staggering of leading-edge devices, and the fact that an inboard panel loses its flow before the outer ones, are design decisions as much as kinematic ones.

The fourth line is artificial and appears only if the first three are not enough. Angle-of-attack probes feed a threshold calculation depending on configuration, Mach number, and sometimes de-icing status: the first threshold triggers the warning, with a demonstrated margin before the actual stall; the second, where it exists, commands the pusher, which imposes the nose-down action the airframe cannot produce on its own. Certification logic reads in this direction: the text does not mandate a pusher, it requires a behavior — a clear warning with sufficient margin, a controllable aircraft during and at recovery from the stall, no uncontrollable pitch-up. If aerodynamics alone does not deliver it, a device does, and it enters the certification basis: redundancy, source monitoring, and inhibition conditions become requirements in their own right. On fly-by-wire aircraft, angle-of-attack limiting fills the same role.

What each device addresses, and what it does not
Device Mechanism targeted What it does not fix
Washout Local lift-coefficient profile along the span Costs lift; fixed at construction
Fence, sawtooth, vortilon Fluid fencing against boundary-layer drift Does not re-energize a depleted boundary layer
Vortex generators Reinjection of momentum at the wall Little effect if separation is shock-induced
Root stall strip Force the first separation inboard Does not remove the outer separation beyond it
Angle-of-attack warning Make the limit perceptible before it is crossed Has no effect on the trajectory
Pusher or angle-of-attack limiting Supply the missing nose-down moment Depends on its sources; can be inhibited or lost

At high altitude, the same mechanism with less margin

At low altitude, the sequence reads clearly separated: buffet, warning, stall. At high cruise, it compresses. The low-speed buffet boundary and the Mach buffet boundary move closer together as altitude increases, and the load-factor margin available between them shrinks.

Two practical consequences follow. A reading problem: the first cue is a buffet easily mistaken for turbulence. An authority problem: equivalent airspeed is low, so dynamic pressure on the controls is too, damping is reduced, and the aircraft responds less crisply. Recovery does not change in nature: reducing angle of attack costs altitude, and available thrust will not do the job. Ice on the leading edge, finally, shifts where separation starts: the order calibrated for a clean wing is no longer guaranteed, hence the recalibration of thresholds when ice-protection is in use.

The right reflex. Address angle of attack before everything else and hold the input until the symptoms disappear: it is the one parameter that governs the others. Accept the loss of altitude as the price of the maneuver. Keep the ailerons nearly neutral until the wing is unloaded, then level the wings with measured inputs. Do not fight the pusher: it does what the airframe cannot do on its own. Disconnect the autopilot and check the elevator trim, which can be working against you. Power comes after unloading, and the buffet margin is reassessed before resuming the level flown before.

What to check for your own type

Five points vary from one type and one operator to another, and are found in approved documentation rather than in a memory of a course. Which probes feed the warning and the pusher, and what the system does in case of disagreement. The inhibition conditions of the pusher and how to override it: force required, cutout device. The margin separating the warning from the actual stall, and how it shifts with configuration or icing. The exact order the type’s procedure gives to reducing angle of attack, leveling the wings, adding power, and configuring. Finally, what the minimum equipment list allows with a warning system, a pusher, or a probe unavailable.

The grey area is there. On an aircraft with angle-of-attack protection, the limit normally cannot be exceeded and the crew never sees the natural behavior of its wing; the day the control law degrades, they inherit a swept wing with its pitch-up and its tip stall, without ever having felt that coupling. Training happens in the simulator, whose aerodynamic model beyond the flight-test envelope remains an extrapolation. For those preparing a type rating, what is expected has shifted: it is no longer a list of speeds to demonstrate, but knowing where separation starts on that particular wing, why the nose does not come down, and what remains when a device is missing or the leading edge is no longer clean. This reasoning is developed in training, and type-specific cases are compared on the forum.

The applicable certification specifications, the type’s flight manual, and the operator’s approved documentation prevail at all times over this article, whose purpose is to explain a mechanism, not to serve as a procedure.

Questions fréquentes

Where on the span should the first separation be expected, and what shifts that location?

On an uncorrected tapered swept wing, the first separation appears toward the outer half-span: the boundary layer drifts from the root toward the tip under the oblique pressure gradient, the induced field increases the effective angle of attack there, and the reduced chord requires a higher local lift coefficient for the same load per unit span. The design then deliberately shifts that location toward the root, through washout, limited taper, and sometimes a stall strip. Two factors shift it back in service: at high cruise, a shock on the outer sections can trigger separation there before the critical angle of attack is reached elsewhere; and a contaminated leading edge cancels the order calibrated for a clean wing. That is why warning thresholds are recalibrated when ice protection is in use.

Is a shot of trim enough to contain the pitch-up tendency?

No, because this is not a shift in equilibrium but a divergence: in this regime, the slope of the pitching-moment curve with respect to angle of attack changes sign, so that any increase in angle of attack produces another. Trim, which shifts an equilibrium point, has no useful effect on a moment that is self-sustaining — only reducing angle of attack restores the flow on the outer sections and closes the loop. Trim can even work against the crew: a prolonged autopilot flight near the limit sometimes leaves the aircraft trimmed nose-up, and the elevator must then overcome the trim before it can be useful. Checking its position after disconnecting the autopilot is therefore part of the maneuver.

What exactly does a stick pusher do, and why don't all aircraft have one?

The pusher supplies the nose-down input the airframe cannot produce on its own when the slope of the pitching-moment curve changes sign near the stall. It is not mandated by the regulations: these require a behavior — a clear warning with sufficient margin, a controllable aircraft during and at recovery, no uncontrollable pitch-up or excessive roll. An aircraft whose aerodynamics demonstrate this behavior does not need a pusher; on fly-by-wire aircraft, the same requirement is most often met by an angle-of-attack limitation within the flight-control laws.

What changes when the tailplane is mounted high on the fin in this scenario?

At high angle of attack, the wake of the wing and fuselage thickens and rises, and a horizontal tail mounted high on the fin sits right in its path. Once immersed, it loses dynamic pressure and the elevator its effectiveness: the nose-down moment needed for recovery is no longer available. The aircraft can then settle into a deep stall, a stable state at very high angle of attack with a high rate of descent and no aerodynamic recovery. It is this type of configuration that led to the widespread adoption of artificial angle-of-attack reduction devices, shape alone no longer being sufficient to demonstrate the required behavior.

Should roll be countered with ailerons during a swept-wing stall?

No, not before reducing angle of attack. The roll comes from an asymmetric separation that occurs precisely where the ailerons are; deflecting a control surface down on the low wing increases the local angle of attack there and worsens the separation. The sequence is to unload first, firmly and until the symptoms disappear, then to level the wings with measured inputs once the flow is restored.

Why is the margin smaller at high cruise altitude?

Because the low-speed buffet boundary and the Mach-related boundary move closer together as altitude increases, which reduces the load-factor margin available between them. The first separation can then be triggered by a shock on the outer sections rather than by angle of attack alone. Added to this are a low equivalent airspeed, hence less effective controls and reduced damping, and the fact that reducing angle of attack unavoidably costs altitude, with available thrust not being enough to restore the situation.

Commentaires

Leave a Reply

Découvrez l’écosystème FL330

FL330, c’est toute l’aviation professionnelle — explorez nos autres espaces.