# Why are airliners not inherently “speed stable”?

In a light aircraft, once the pilot has trimmed for a speed, the power controls the pitch (climb/descent rate).

I don't know the technical name for this stability, so I chose "speed stable" for now.

The fly-by-wire (FBW) of the 787 (for example) mimics this behavior:

In flight, the pitch trim switches do not position the stabilizer directly, but make inputs to the [Primary Flight Computers] to change the trim reference speed. The trim reference speed is the speed at which the airplane would eventually stabilize if there were no control column inputs. Once the control column forces are trimmed to zero, the airplane maintains a constant speed with no column inputs. Thrust changes result in a relatively constant indicated airspeed climb or descent, with no trim inputs needed unless airspeed changes.—787 FCOM

I.e., once a speed is trimmed, the top sentence becomes true like in light planes.

So the related questions are:

1. What's the name of this stability?
2. Is it limited only to light aircraft?
3. Is it related to having a trim tab instead of a stabilizer trim? (Just a speculation.)
4. How can an airliner without FBW be designed to be "speed stable"?

An answer doesn't need to address the exact questions in order, I'm just communicating what I'd like to understand.

2019: The impact of the third point still confuses me:

Once the speed is trimmed (and left alone), and power is increased (for example), wouldn't having a trim tab (like on a typical light plane) have more impact on pitch response (speed stability) compared to the now-fixed stabilizer (jet-liner)? In other words, a light plane would pitch up quicker (elevator responding to the increase in airspeed), while a non-FBW jet-liner would pitch up much slower (stabilizer unaffected by airspeed) and gain more speed compared to the earlier trimmed state?

• "Speed stability" is the exact terminology. Some matter from Boeing. Related: Phugoid. – mins Sep 20 '16 at 2:11
• Btw most light aircraft do not retain the EXACT same speed when you change power. Even disregarding "phugoid" oscillations, it's not at all uncommon at least in a high-wing plane for the plane to stabilize at a lower airspeed after you've added more power to climb. If the engine is mounted above the wing as in some pusher-engined light-sport planes or ultralights, the opposite is often true. – quiet flyer Jun 23 at 14:39
• No, I don't think it's related to having a trim tab versus stabilizer trim. – quiet flyer Jun 23 at 14:42
• Both trim tab and stab trim change the free-float angle of the elevator. The effect is the same, so the dynamic response to a thrust change should be identical for both. – Peter Kämpf Jun 23 at 14:46
• @PeterKämpf: So on a jet-liner with a THS, the elevator free-floats? If so, wouldn't an increase in airspeed (from high AoA to a lower one) make the elevator more flush, i.e., less downforce? – ymb1 Jun 23 at 14:52

Light aircraft and transport aircraft behave exactly the same regarding trimming for speed. Trim is for angle of attack, but power affects that trim and phugoid oscillation occurs in speed and pitch, because the feedback is second order.

The Boeing FBW has no trouble mimicking anything. It simply translates the control forces and position exactly as a hydro-mechanical link would, just in electric signals. The aircraft is still stable.

Airbus aircraft are also all stable, but their FBW relaxes the stability by auto-trimming for current speed. The result is that side-stick controls flight path angle and power controls speed.

The only exception to the stability are fighter jets (starting with F-16). Those are aerodynamically unstable to get faster control response and their FBW makes them neutral (that is, the yoke/stick again controls flight path angle).

The reason that large aircraft are often not exactly speed stable is that the line of thrust does not go through the centre of gravity of the aircraft. A change in thrust setting will therefore cause a change of pitch moment.

In small propeller aircraft the thrust is usually acting through or near the centre of gravity and hence the aircraft are speed stable.

• Thrust acting through the centre of gravity is not relevant. It depends on how far it is from centre of drag, and I don't think typical light aircraft is much better in that regard than typical jet transport. – Jan Hudec Sep 21 '16 at 21:45
• @JanHudec All motion in flight is relative to the CoG. DelaLima”s answer is correct. – Koyovis Jun 23 at 19:05
• All acceleration (forces applied) are relative to CG, but in an aerogravimetric environment, these forces come into equilibrium with drag and gravity. Once the object is moving, particularly at higher speeds, aerodynamic forces will predominantly determine the NET center of rotation. Jan Hudec is correct regarding a rapidly moving aircraft. But just as it is in pitch, CG cannot be ignored, as it's effects are larger in slow flight. – Robert DiGiovanni Jun 23 at 23:50
• The flight is trimmed: thrust & drag & all moments about the CoG are in balance. Now power increases => thrust increases below the CoG => total moment about the CoG changes => aircraft needs to be re-trimmed. Not the case when thrust lines up with CoG. – Koyovis Jun 24 at 9:49
• This is perfect for wind tunnel test (xylophone is optional). Take a rectangular plank and offset the balance with a lead weight. Hold the board sideways in airstream and see where you must support it to keep it straight up and down. In a constant velocity stream, it should be right in the middle. But if you increase the velocity of the stream, initially it will torque around the CG, but then re-equilibrate at the center of drag. – Robert DiGiovanni Jun 24 at 11:34

I'll try to approach it from a slightly different angle.

Fundamentally, there is no difference between light aircraft and airliners in terms of stability. 'Normally', that is, on the 'front' of the power curve, if the aircraft is statically stable in pitch (or more precisely, in angle of attack), it will be stable in speed.

If you have a classical trim tab and a free-floating elevator (reversible control), it still behaves largely like an irreversible ('fixed') stabiliser/elevator: as you accelerate, the aerodynamic forces on the elevator and on the trim tab increase proportionally, and the hinge balance remains. (Of course, assuming that the changes are not too great to alter the flow dramatically). There will be a difference in pitch stability with relaxed control, because the elevator will 'flop', but in the end the steady state will be the same.

However, there are substantial 'practical' differences between the light and heavy aircraft which make them behave differently in this regard.

• First, as already mentioned, the thrust line on most modern airliners passes quite far from the centre of drag (and CG). This does matter for the speed changes caused by thrust - which are, unlike pitch changes, are the most practically relevant changes. (It is rare to encounter a wind shear of such duration that speed would re-stabilise). Having thrust line low enough can completely destabilise trim (with respect to thrust changes). In fact, the 737 MAX ordeal is a testament of how important such considerations are.
• Second, and this is more interesting, the so called long period motion (in particular, speed) on heavy aircraft is more decoupled from the short period motion (e.g. pitch) than on light aircraft. In other words, the speed and altitude dynamics is slower with respect to the pitch/roll/yaw dynamics.

This latter deserves some discussion. For the pilot, this makes control more difficult in some ways - but in others, easier. This all depends on the task at hand.

In particular, if we just change trim, a light GA airplane will very quickly find a new equilibrium with speed and climb angle. But a heavy will (naturally) quickly reach the new AoA, and then enter a long sequence of very long period phugoid oscillations. Here engineers can help pilots to make control convenient as they we want.

Enter human factors. In general, research tells us, we achieve the best results if we control the first derivative of the target parameter. Say, if we want to aim a gun, we want a joystick which controls its pitch and azimuth (or yaw if you like) rates proportionally to the applied force.

So, what control is the 'best' for an airplane? Of course, this depends on the aim. Many modern control systems reconfigure themselves for different tasks. But for 'normal' flying, we mostly want to control the flight path angle. This is one integral away from the load factor, and this is why (to a large extent) Airbus stick controls the load factor.

Great? Almost. Not quite. There is one problem with humans. We naturally predict things. When we interact with the world - we walk, we see - we constantly and subconsciously forecast the world the way it's going to be in a few moments. As a result, when the motion is reasonably slow and does not exceed our abilities, we prefer a seemingly more 'difficult' way of control - directly, without an integral between the input and output. Or a blend of them. And we actually get measurably better results. For our little task of setting speed, setting trim (or elevator control in general) for speed control is more 'direct' and faster than thrust control, even though the underlying dynamics may be more complicated. This may be counter-intuitive for most people, but experienced pilots may prefer that, and this is why (my guess) Boeing made it that way.

• "..we achieve the best results if we control the first derivative of the target parameter." We achieve the best results if we control the actual target parameter. We achieve faster results if the derivative of the parameter is part of the feedback loop. – Koyovis Jul 1 at 23:16
• No, or not quite, if by 'best' we understand the lowest integral tracking error. As I tried to explain, if the signal is fairly slow, then yes, we tend to prefer a more direct control (which inherently requires mental (or augmented, as you pointed out) derivation - that is, prediction - if there is any inertia (dynamics) in the system). But for a faster signal - say, typical frequencies of short period motion of most aircraft - controlling the derivative is the best. Think of aiming a gun at a moving target, for example. – Zeus Jul 2 at 1:59
• Aiming a gun at a moving target: the signal to be controlled is velocity. – Koyovis Jul 2 at 9:39
• Exactly. Angular velocity, of course. – Zeus Jul 3 at 0:54
• ah, you wish to improve on the statement. The point was: We achieve the best results if we control the actual target parameter - in the moving object case, velocity. In a static target case: position. The derivative signal in the feedback loop improves frequency response. The motion sensors in our inner ear detect angular acceleration, Peripheral vision detects angular velocities. just for the purpose of frequency response improvement in control tasks. Problem with regulating the phugoid id that the accelerations are below the inner ear detection threshold. – Koyovis Jul 3 at 2:03

All the forces and moments acting upon an aeroplane, with the following main groups:

• Wing/fuselage/engine pods, indicated with index w. Components are $$N_w$$ and $$T_w$$ of the total aerodynamic force $$R_w$$ acting upon the aerodynamic centre of the wing/fuselage/engine pod group; and the moment $$M_{ac_w}$$
• Horizontal tail. Analogous to the wing group: $$N_h$$ and $$T_h$$ acting upon the aerodynamic centre of the horizontal tail group $$ac_h$$; and the moment $$M_{ac_h}$$
• The propulsion installation. Contributions of propeller or jet engines are thrust $$T_p$$ along the propeller plane or jet exhaust angle; and the perpendicular force $$N_p$$ which occurs when the propeller or jet intake has a local angle other than zero.

With the aeroplane in trim, the moments about the y-axis are:

M = $$+\;M_{ac_w} + N_w(x_{cg} - x_w) - T_w(z_{cg} - z_w) \;+$$

$$\quad\;\;+\; M_{ac_h} + N_h(x_{cg} - x_h) - T_h(z_{cg} - z_h) \;+$$

$$\quad\;\;+\; (N_p \cdot cos\; i_p + T_p \cdot sin\; i_p) \cdot (x_{cg} - x_p) \;+$$

$$\quad\;\;+\; (T_p \cdot cos\; i_p - N_p \cdot sin\; i_p) \cdot (z_{cg} - z_p) = 0$$

Now increase thrust => $$T_p$$ and $$N_p$$ increase. If $$(z_{cg} - z_p)$$ ≠ 0 and/or $$(x_{cg} - x_p)$$ ≠ 0, the change in thrust will create a change in moment, which will need to be counteracted with a change in $$N_h$$.

So I'm concurring with DeltaLima's answer.

The flight is trimmed: thrust & drag & all moments about the CoG are in balance, that is what the trim setting does, whether it is a trim tab or stabiliser or a little umbrella sticking out of the cockpit. Now power increases:

• in case of a jet with underslung engines, thrust increases below the CoG;
• the total moment about the CoG changes: de-stabilises for increased thrust (nose-up moment);
• the aircraft needs to be re-trimmed.

Not the case when thrust lines up with CoG, for instance in little PPL aeroplanes or twin tail engine aircraft like the MD-80 and the Fokker 100.

By the way, the term speed stability is mostly used with reference to the reaction of the aircraft to a sudden gust in the x-direction. In order to be certified for flight, the reaction must alway provide a stabilising moment. Not sure if this sort of thrust-trim feedback has a specific name, trim stability might be appropriate.

How can an airliner without FBW be designed to be "speed stable"?

If we mean No Deviation From Trim: with a given$$(z_{cg} - z_p)$$ and $$(x_{cg} - x_p)$$, which are parameters of aircraft configuration and loading, the moments about the CoG are a linear function of thrust. An automatic trim system would incorporate altitude feedback, and can be implemented in any flight control system regardless of FBW or mechanical cable input.

All above is about static stability. The dynamic response to a disturbance are usually two-fold:

• A fast response, too rapid for pilot reaction, which must be strongly positively damped. The horizontal tail being perpendicular to local flow provides this strong damping.
• A long period response, the phugoid, which must dampen out to a new equilibrium position if not corrected by the pilot.

The equilibrium position is the key to the OP question, and is a parameter of the static stability consideration. Disturbance in thrust moment must be counteracted by an aerodynamic moment in order to maintain the old equilibrium position.

• Really talking about two ways of testing for the same thing: the effect of thrust line on pitching torques. Your way works just as well. Best test may be to do both, slow down, deploy flaps and slats, then firewall the throttle. – Robert DiGiovanni Jun 24 at 16:12
• The picture (and text) creates an impression that the aerodynamic forces (N, T) actually act at the aerodynamic centre. Of course, they don't, and hence $M_{ac}$, but on the surface, the text may contribute to the common confusion between AC an CP. By the way, you have to have both $z$ in the second bracket of the first line of M. – Zeus Jun 26 at 1:27
• @Zeus It states clearly that the aeroforces per group act upon the aerodynamic centre of each group. – Koyovis Jun 26 at 1:56
• That's not my point. The forces don't really act there. Lift 'acts' at the centre of pressure, not AC. It is convenient to consider them acting at AC (or CG for that matter), but then we need to introduce an additional moment (which is $M_{ac}$). Not explaining the nature of this moment may be confusing. It's a minor complaint though; I'm just tired of explaining the difference to those who ought to know (most pilots, for example), and I'm worried about them reading this and misinterpreting it. – Zeus Jun 26 at 2:18
• @Zeus It is even more complicated than that, bacuse aerodynamic loads are distributed loads that act over the whole wetted surface. Stability & Control models the behaviour of the aircraft using the AC, otherwise there are simply too many variables. From an aerodynamic standpoint your statement that lift can be modelled as acting in the centre of pressure is of course correct, analogous to gravity and the centre of gravity. – Koyovis Jun 26 at 2:58

For simplicity, we shall assume we are dealing with a longitudinally stable aircraft that is a small increase in angle of attack will cause the pitching moment on the aircraft to change so that the angle of attack decreases. Similarly, a small decrease in angle of attack will cause the pitching moment to change so that the angle of attack increases.

If you increase the thrust, and If it is in line with CG, you will increase as a start the speed, thus the lift generated at the wing which will cause a climb, similarly, on civil aircrafts, the increase in speed also increases the down force on the horizontal stabilizer which will cause a pitching moment up and also cause a climb. This is why a pilot needs frequently, to retrim an aircraft after every change in power and attitude.

Is it related to having a trim tab instead of a stabilizer trim? (Just a speculation.)

With respect to Boeing FBW you are perfectly right, it acts as if it was like you are saying, as a proof, in the B777 simulator (probable the same for the B787), if you display the flight control page in flight, and without moving the column, you just act on the thumb switch for a short moment, you notice the elevators moving, as if you were acting on the column; afterwards, and only afterwards, the THS will move and the elevators will return to neutral.

With respect to Airbus philosophy,it is different(there are no thumb switches, you may act directly on the wheel which is very easy to move, but normally you don’t touch it in flight, just on ground for setting the THS for takeoff).

On Airbus the FBW management is different, The Z axis depends on the side stick which gives a load factor order. When the stick is not touched the load factor is equal to 1, that is if stabilized on a flight level, any thrust increase will only increase the speed without any increase of the lift that is the computers will act on the elevators to prevent any pitching effect due to the speed increase. Similarly if the thrust is decreased the AOA will increase to maintain the flight level within the maximum allowable AOA. With respect to the trim which is automatic always in flight, any action on the stick gives a load factor order(above 1, or below 1) which, through the computers, will immediately act on the elevators, later on the stab will automatically take over and the elevators will go to neutral. During climb or descent, if the thrust is modified, the FBW will maintain the path within the protection envelop of the angle of attack and of the speed limits.

• Thank you for taking the time. The Boeing augmentation does not answer why (and if) it is needed compared to a light plane. Judging by that comment, thrust-line is not really the main factor. – ymb1 Jun 23 at 17:53
• When Boeing developed the 777 they were very critical towards Airbus FBW which has been developed from an « « engineering » point of view as a continuity of the Concorde FBW design. Boeing took into accounts many pilots opinion who conservatively insisted on the longitudinal stability and effectively too on the speed stability as you say, not to disturb this stability the thumb switch acts on the elevators as if the pilot is moving the column, of course this action is afterwards neutralized by the THS. – user40476 Jun 23 at 19:36
• It tuned out to behave as a light plane, this is a consequence of pilots desire. « At that time » Boeing had little experience with FBW on civil aircraft and « were » « respectful »of pilots opinion. – user40476 Jun 23 at 19:36

Once the speed is trimmed (and left alone), and power is increased (for example), wouldn't having a trim tab (like on a typical light plane) have more impact on pitch response (speed stability) compared to the now-fixed stabilizer (jet-liner)?

The trim tab is just one piece of the system, with the following all impacting longitudinal stability for just the stick-fixed condition.

• CG position
• flat plate area (drag coefficient)
• wing area
• aspect ratio
• wing $$C_{L \alpha}$$
• Oswald's Coefficient
• tail area
• tail aspect ratio
• tail $$C_{L \alpha}$$
• geometry of wing, tail, and CG (how far apart, any step down or up for tail WRT to the wing)
• wing $$C_{M \alpha}$$
• Atmosphere density
• Aircraft moment of inertia about Y axis
• Aircraft mass
• Prop versus jet
• Constant speed prop versus fixed pitch

Now throw in stick-free, and you get your trim tab, center of pressure versus center of rotation of the tail, and any springs and bobweights in the control system. In short, its complicated.

1. "Speed stable" is static stability under thrust. If aircraft is faster than trim speed it will pitch up aerodynamicly from increased tail downforce, and pitch down from slower speed. This is a relationship between center of gravity and NET center of pitching torques from aerodynamic and thrust forces.

2. No, it works the same for hand held gliders on up to 747s.

3. You can use a trim tab, elevator adjustment, and/or a horizontal stabilizer decalage change (incidence relative to wing) to adjust the amount of negative or positive lift your tail creates. To be "speed stable", you want a slightly forward CG set (from wing CP) and downforce on tail. Thrust force vector must not render aircraft staticly unstable.

4. FBW, hydraulic boost, manual pulleys and cables, R/C servos all do the same thing, that is deflecting control surfaces when commanded by pilot (or computer). The key is properly designing the horizontal stabilizer and elevator to serve its dual function as weather vane to hold wing AOA where you want it, and second wing to change pitch (by altering Net center of lift).

Angling the thrust vector is commonly done in aircraft when center of drag (mainly the wing) is offset from the location of the engine mount. If increased thrust tends to pitch the aircraft up, down angle ahead of center of mass (tractor) or up angle behind center of mass (pusher), with the effect of helping lower the nose, helps counter act the effects of the asymmetry. Nose down thrust angle is commonly seen in the high wing trainer designs.

The Boeing "Some matter" link specifically states "speed stability" is tested starting in level trimmed flight with a constant thrust setting. The "speed stability" is essentially changing the negative lift of the entire tail to produce a higher or lower staticly stable trim speed.

Unfortunately, thrust angle does significantly affect pitch stability and trim settings and there for must be strongly considered in design parameters, particularly when there is a potentially large variation in static stability as seen in cargo/passenger carriers.

Now, "reading between the lines": the speed stability test is performed under thrust (as compared with 0 thrust gliding), so if thrust is kept constant and aircraft is slowed down by pitching up, some very important information can be gleaned from the flight data: how much does thrust line affect static stability! Under what conditions will thrust line render aircraft staticly unstable!

So this test must be pushed to the point of failure to determine proper thrust line, aft CG limit, and proper tail design. It must be done with the most rigorous downwashing, highest AOA, maximum power, lowest speed possible to determine when static stability is lost.

Not just to pass, but to pass with an adequate safety margain.

Future drag saving design considerations may be along the line of the Citation X. There is a possibility of reducing wing root drag by having the engine intake smooth the turbulent airstream in this area. One might even consider completely burying the nacelles in the wing roots.

But "saving drag" by reducing the size of the all important horizontal stabilizer seems silly if other options are available.

• " If increased thrust tends to pitch the aircraft up, down angle ahead of Center of Lift (tractor) or up angle behind center of lift (pusher)" I feel using center of mass would be much clearer here. Using the CoL is needlessly confusing, as the aircraft does not turn around it per se. – AEhere supports Monica Jun 26 at 13:37
• Noted and edited to center of mass. Vickers Valiant type might be worth a look. – Robert DiGiovanni Jun 26 at 14:45

I love the way everybody avoids talking about gliders when speaking of pitch stability, which is essentially the same thing as speed stability. Yes, changing thrust does affect where an aircraft balances itself, but gliders also have stability, with no thrust at all. You can't really understand the physics of flight without understanding gliders.

Any aircraft with a conventional tail has horizontal stabilizers that act as inverted wings, converting forward movement (rearward air flow) into the opposite of lift. They push down on the tail.

Nose down too much, and like a bicycle going down a steep hill, you speed up. Speeding up increases lift on the wings and increases downward pressure on the horizontal stabilizers. The tail goes down. The nose goes up.

Nose up too much and like a bicycle rolling up hill, it slows down. The wings lift less, the horizontal stabilizers push down less and the nose drops. The airplane oscillates between too much nose up or nose down, each time going a little less extreme, until it settles at one pitch and one speed. Weather can disrupt this some, but for the most part, hands off the stick and the plane flys itself at one pitch angle and one speed. This is stability.

Change the weight, and the airframe balances at a different point. Change the Center of Gravity and again, the airframe finds a different balance point.

Canard designs achieve stability by different means. Since the horizontal stabilizer is up front, instead of pushing the tail down, it has to hold the nose up. It achieves stability by having a steeper angle of attack than the main wing. To engineer this feature, you adjust the size of the wing so that it's carrying a heavier load per surface area, requiring the higher angle of attack to hold the nose up.

Since lift is determined by a combination of Angle of Attack and airspeed, the smaller, steeper canard reacts more to changes in airspeed than the main wing, so that if you fly faster, the nose goes up, and if you fly slower, the nose drops down. Do it right, and the canard stalls before the main wing, causing the nose to drop, so that it becomes impossible to stall the main wing.

The canard is more efficient than the conventional tail design because all horizontal surfaces are increasing lift. The conventional tail essentially creates artificial weight at the back of the aircraft, making the main wing work harder to compensate.

But canard designs are so good at flying that they are harder to land. The approach tends to be more shallow, and they require longer landing strips.

• "Speeding up increases lift on the wings and increases downward pressure on the horizontal stabilizers. The tail goes down. The nose goes up." -- the problem with this line of argument is that it tends to suggest that if you add weight to the CG of the glider, so that it tends to fly faster, it will have some tendency to trim to a higher angle-of-attack, which isn't true. The real mechanism behind speed stability is more complicated and includes the fact that if the flight path tends to curve upward for any reason, gravity has now gained a component that acts parallel to the drag vector – quiet flyer Jun 26 at 16:58
• The real mechanism behind speed stability is more complicated and includes the fact that if the flight path tends to curve upward for any reason, gravity has now gained a component that acts parallel to the drag vector-- see aviation.stackexchange.com/questions/29746/… – quiet flyer Jun 26 at 16:59