I have been thinking about a new design for a plane, but have run into a problem where running control linkages to the tail will not be easily done.

I was wondering what would happen if I made the plane with elevon control only, but will still have a more or less conventional tail on it. I'm thinking it should still work okay, but the controls might not be as sensitive, and I may want to make the elevons a little larger.

From : Don Stackhouse

We've done some experiments and analysis of our own on controls of this type. This can be made to work, but it's tricky.

Aileron effects are pretty much unchanged for the bulk of the flight envelope (but be careful at low speeds and/or high angles of attack). If the model had decent roll response with just ailerons (with no help from the rudder or from aileron differential to counteract any adverse yaw), it will probably still have decent roll response. The tricky part begins when you try to get decent pitch control.

To understand the problem, first consider that two things happen to a airfoil when you deflect a conventional control surface that's a part of it:

1. The amount of camber and the shape of the camber line (the curved line running through the exact middle of the airfoil from the leading edge to the trailing edge) changes.

2. The location and angle of its chord line changes relative to the rest of the airframe.

In the case of an all-flying control surface such as a stabilator, or the rudder on a Fokker DR1 Triplane, the camber stays the same, but the chord line relative to the airframe changes.

Next, consider the effects of #1.

An increase in camber MAY (but might not always, especially at our Reynolds numbers) increase the max lift coefficient, and tends to lower the zero-lift angle. Thus, if you kept the chord line constant somehow while you increased camber, and assuming that the camber is still low enough that bad things due to excessive amounts of camber haven't yet begun to set in, the increase in camber will tend to increase the lift coefficient.

The other thing that essentially always happens with an increase in camber is an increase in the aerodynamic pitching moment. Airfoils with positive non-zero camber and no reflex have a nose-down pitching moment. How strong this pitching moment is depends mainly on how much camber there is and how that camber is distributed. If there's a lot of camber near rhe rear of the airfoil, the pitching moment is likely to be higher than if that camber is mostly concentrated further forward. Take special note of that little bit of trivia, it will become crucial in just a moment.

Reflex, such as what we use on "plank" style flying wings, works by adding a dose of negative camber near the trailing edge. Camber changes near the trailing edge tend to have a proportionately stronger influence on pitching moment they do on lift coefficient. If we bend the trailing edge of a cambered airfoil up (i.e.: add some negative camber to that area of the airfoil ONLY), we can reduce the negative (i.e.: nose-down) aerodynamic pitching moment, even to the point of making it become a positive pitching moment, while still keeping the overall camber of the airfoil positive. Some flying wings with only moderate sweep cut the wing with a conventional positive-cambered airfoil and little or no washout, then rig the elevons with just enough "up" to create the necessary amount of reflex for stability.

Now consider item #2 above. The camber line is a straight line from the furthest forward point on the leading edge to the furthest aft point on the trailing edge. If we bend to trailing edge up or down, the aft end of the chord line moves up or down with it. Since the incidence of the airfoil (the angle it is mounted at, relative to the rest of the airframe) is typically measured from the chord line, then if we change the chord line by deflecting a control surface, we also change its incidence.

So what happens when we combine all of this? Lets consider three cases:

A. An airplane with a conventional tail, but no control surfaces on the tail, and narrow-chord elevons on the wing.

B. Same as case "A", but with extremely wide-chord elevons, perhaps 50% to 70% of the wing chord.

C. a "pitcheron" model, where the wings have no control surfaces on them, but both wing panels can pivot about the joiner rod as commanded by their servos. You can think of these as elevons whose area is 100% of the wing area, an "all-flying wing panel".

Case A: When we try to give it a command to pitch up, we drop the elevons both downwards. The incidence angle increases a small amount because of the change in the chord line. This increases the decalage (the difference between the incidence of the wing and the incidence of the tail), creating a nose-up effect. The greater lift from the wing causes a small increase in the downwash from the wing, which tends to raise the nose by pushing the tail down. The increase in camber of the wing causes a LARGE increase in the nose-down pitching moment. Add all this together and these factors tend to cancel, resulting in a very weak net effect on the model's pitch attitude. The effect may be a little bit upwards, but more likely it will be either zero or weakly downwards.

One other sub-case should be looked at here. Consider what would happen if we kept the elevons the same size, but gradually reduced both the tail area and the tail moment arm, until both were equal to zero. The incidence (decalage) effects and downwash effects are both dependent on the size and moment arm of the tail. As the tail gets smaller, they get smaller. When the tail is zero, those effects are also zero.

At that point the only effect left is the pitching moment effect, so if we reverse the elevon travels so they both go up for an "up" pitch command, we should get pretty decent pitch response. The Pibros, Zagi, Pico Jet, etc. all use this approach. Pitch response is excellent. Note, in this particular situation we want to keep the elevons as NARROW in chord as possible if we want to maximize our pitch response, as long as they are still wide enough that they don't become aerodynamically insignificant. The only drawback is that as we move the elevons up to get an "up" pitch command, we are also reducing the overall camber of the airfoil. At the point in the aircraft's operating envelope where we need lift-making capability the most, in low-speed flight, we have just REDUCED the airfoil's lift-making ability. Engineering is a study in tradeoffs.

Case B: Same effects as above, except that with the hinge line much further forward, the effect on pitching moment is small but the effect on incidence is comparatively large. In this case the incidence and downwash effects are more likely to dominate, probably resulting in a weak to moderate nose-up tendedcy when the elevons are drooped.

Case C: There is no change in camber, only incidence. Since for a given camber, the pitching moment of most airfoils vs. angle of attack is relatively constant except near positive and negative stall, there is no significant change in pitching moment. The incidence and downwash changes result in a fairly decent response to the pitch command.

Two other things I should point out about case "C" before everyone runs off to build pitcheron models. First of all, this results in a wing with fixed camber, so you don't get any of the max-lift improving effects of camber changing. For our slow-flier models we're typically operating within a fairly narrow range of airspeeds, so being able to widen the speed range through the use of flaps normally isn't an issue. In addition, at our typical Reynolds numbers, adding camber through flaps in an effort to increase lift often doesn't work anyway. A lot of the old axioms from full-scale and larger models, like "more thickness and/or more camber means more max lift" are no longer true for our models.

The other problem with pitcheron models is the load on the servos. If you try to maximize performance in upright flight by using a cambered airfoil, that airfoil will generate a pitching moment. For a given airspeed it's a constant moment, so it doesn't have any significant effect on pitch control response, but it does put a load on the servo. As speed increases, this moment increases with the square of the airspeed, and at some airspeed it may get large enough to overpower the servos. This is not a big concern for most of our slowfliers, but for larger and faster models, and especially for full-scale aircraft, it can be fatal if the servo sizing doesn't properly account for it. Use of a symmetrical airfoil (camber = zero, so pitching moment = zero) and a proper hinge location (typically at the aerodynamic center of the panel, approx. 25% back from the leading edge at the mean aerodynamic chord) will eliminate this problem in subsonic flight.

If you want to keep the tail fixed and still use flaperons on the wing for both pitch control and lift increase at low speeds, your best bet is probably to use a fairly large tail on a short moment arm, with extremely wide-chord flaperons, or all-flying pitcherons. Even so, it's tricky to get everything to work in combination just right. This treatise merely scratches the surface of the problem, there are a huge number of other incidental effects that influence the outcome as well. Be prepared to do a LOT of experimenting!

Don Stackhouse
DJ Aerotech