When designing an aeroplane, it has to work at the extremes of its performance. A plane is most heavily-laden at the moment of takeoff, when it is flying almost at its slowest. In order to rotate the nose up, the elevator is moved to make the tail push down. The wings have to lift against this downforce as well as the weight of the plane, so they have to be made oversize. But this has nothing to do with stability. To make a plane stable in such pitching motion, all that is needed is for the tail to lift less hard than the wing, to have a downwards component relative to the main lift. For maximum cruise efficiency the plane is balanced so that the tail gives zero lift. Another stable arrangement is to put the tail in front of the wing (the canard configuration) and make it lift even harder.
You might think that when a stable aeroplane stalls, the front surface is lifting harder and so it stalls first. If this is a canard, then the nose will simply drop while the main wing is still flying normally. Thus, a canard type cannot be stalled and is inherently safe. Well, not usually. Firstly, such a design is often subject to phugoid (up-and-down pitching) oscillation which, if allowed to diverge, will eventually cause the main wing to stall if it doesn't break up first. Unhelpfully, it also makes for an inefficient main wing operating below its maximum loading. Tinkering with the aerofoils and stuff to avoid these problems usually means that the wing will sometimes stall first, causing the tail to drop and the canard to stall too. Whatever causes the main wing to stall, it is disastrous because the control surface needed to pull out of the stall, the canard, is also stalled.
Actually, a safe compromise can be reached. Simple workarounds or cheats include; a downward-pointing rocket in the tail to push it back up in an emergency (don't laugh, this is sometimes done, and not just on canards) or a digital flight system which will stop a deep stall happening in any circumstance which doesn't break up the airframe anyway. A more satisfactory solution is a pure control canard with enough movement never to stall: if the wing does, then the canard can still be used to force the nose down hard. Additional measures must be taken to arrange for stability as, unlike the lifting canard, a control canard makes no such contribution. Alternatively, with a lifting canard design it is possible to ensure that phugoid oscillations do not develop and that no significant part of the wing can stall before the foreplane does. With care, a safe design can be produced.
Newtonian dynamics treats the air as a stream of "billiard-ball" molecules striking the underside of the plate and being deflected downwards to push the wing up. But even taking thermal motion into account, so as to allow an equal reduction in pressure above the wing, this model accounts for only about a quarter of the lift actually created. Half the flow deflection takes place behind the wing and Newton's laws cannot explain the pressure differences which cause this. Only the more complex theory used for proper aerofoils can fully explain the lift generated by a flat plate.
Bernoulli's principle is no better than Newton's laws in explaining why the pressures above and below an aerofoil change as much as they do. The slight restriction of space introduced by the aerofoil camber is nowhere near enough to explain the strong acceleration of the air above. Once the local airspeed above and below is measured, Bernoulli's principle does give the pressure difference and hence the lift, but Bernoulli is unable to account for the speed difference.
The air above does travel faster, in fact it accelerates backwards while the air below slows down. So the air above gets to the back a lot earlier than the air at the front. The exact relative time depends on the aerofoil form, its speed and its angle of attack.
When a wing is taxiing on the ground, a stagnation point develops a short way forward of the trailing edge and on the upper surface. Air curls upward round both the leading and trailing edges. No lift is generated yet. As the wing accelerates forward the stagnation point slides back towards the trailing edge until it can attach itself there – the Kutta condition. The upward flow there suddenly becomes a downward flow. The change creates a "starting vortex" which gets left behind above the runway and, more importantly, a counter-rotating circulation is left enveloping the wing. This circulation rises up at the front, accelerates over the top, curls down behind and returns forward underneath. It is superimposed on the main airflow, accelerating the flow above and slowing it below. Bernoulli's principle now tells us that the fast-moving air above the aerofoil must lower the pressure, creating lift. At the same time it draws down more air than is needed for the circulatory flow, and this creates a downward-moving wake. With the net air passing over the wing now being accelerated downwards, by Newton's laws this must generate an upwards, lifting reaction on the wing. Thus, the circulation theory of lift is needed to explain how all this works, even Bernoulli and Newton together are not enough on their own.
It is true that an elliptical lift distribution is the most efficient for a given wing span, but in practice the lift distribution differs from the shape of the wing. In particular, tip losses mean that the lift fades away early towards the wing tips, while techniques such as "washing-out" the tips by reducing their angle of incidence further reduces the lift they generate. An example is provided by the famous Spitfire. Conversely, it is possible to tailor almost any blunt-winged planform for an elliptical distribution, although the flight characteristics will not always be pleasant.
Then again, the most efficient wing for a given lifting load does not even have an elliptical distribution. A bell-shaped distribution, with reduced or even negative lift at the tips, reduces induced drag due to wingtip losses and is more efficient overall. But again, the lift distribution and the wing shape are not the same thing. A simple constant-chord, untapered wing is easy enough to tailor by careful profiling and is also easy to manufacture. Don't ever think that a fancy wing necessarily means fancy efficiency.
It is true that if you treat a bumblebee's wing naively as a flapping aerofoil then the numbers don't add up, the wing is just too small. But nowadays we know better. Firstly, at the beginning of the wing stroke the air above and between the wing expands rapidly, lowering the pressure above the wing. As air rushes into the space it sets up a spinning vortex above each wing. These vortices substantially increase the local air speed and keep the pressure low, which greatly enhances lift. Concorde used to raise itself at a sharp angle and create rather similar vortices which allowed it to take off and land at relatively low speeds for such a narrow wing. But you won't see a bumblebee with a "droop snoot", just a proboscis for drinking nectar!
The flying saucer may look neat, but this is actually a problem. It has no far extremities to offer a good moment arm for control thrust, making it an awkward shape for vertical-lift flight. The Harrier or AV-8 "jump jet" was much more effective.
The flying saucer is also an unstable shape for flying forwards; attempts to stabilise it by blowing the jet thrust backwards over the rear part might perhaps work at subsonic speeds, but become wholly impractical at supersonic speeds. American researchers quickly abandoned it in favour of a tail surface. Although you could make an unstable saucer fly with a modern digital control system, the F-35 JSF has better controllability and lower drag, making it an inherently far better way to take VTOL supersonic.
For practicable hypersonic flight the broad saucer shape is, quite simply, a loser as it would eat enormous power and rapidly overheat; something more streamlined like a blunt-ended paper dart is essential.
This myth is as much about structural weight as aerodynamics but it is such a common howler that it bears including here. The truth is that a vertically launched rocket is almost impossible to beat. Its straight, cylindrical body need only be stiffened lengthwise to withstand the forces of vertical acceleration. Its circular section is already the perfect shape to absorb any secondary sideways forces which may occur and the internal pressure of the fuel vapour provides all the stiffening that is needed to resist such sideways forces. But lay it empty on its side and it collapses in the middle. If you don't believe me, first stand an empty toilet roll on end and balance a moderate-sized book on it. Fine. Now lay the roll on its side and watch the book squash it flat. That's the problem with the SSO. Because of course, on its side is exactly how an SSO vehicle comes in to land. So the SSO has to have extra stiffening hoops added all the way along its fuselage and internal fuel tanks. The whole tube must also resist the bending forces of gravity, which means further strengthening the lengthways stiffeners too. It all adds up to a lot more structural weight.
Worse, unlike a step-rocket which periodically drops off its dead weight, the SSO vehicle has to carry it all throughout the long accelerating haul to orbital speed. And when you do the sums, that is far and away the longest and hardest part of the ride.
All that adds up to a huge inefficiency, far more so than most designers realise. Advanced materials might just make it possible one day, but such materials will bring equal efficiency savings to the multi-stage vehicles of today, making them an even harder act to beat. There is no way that an SSO can be anywhere near as structurally efficient and light-weight as a multi-stage orbital vehicle of equivalent technology, and that makes it next to impossible for it to be operationally as efficient either.
Updated 1 Feb 2018