They’re Not Winglets

They're not winglets.

Throughout the history of aeronautics, people have experimented with various ways of ending wings. The most common tip treatment has been some variation on a simple rounded cap -- for a long time Mooneys dispensed even with that, wearing their sawed-off tips with disdainful pride -- but marginal gains in one area or another have been claimed for upturned or downturned tips, tip plates, tip fences, winglets and whatnot.

All of these are weapons in the war against the tip vortex.

The tip vortex cannot be destroyed. Airplanes remain aloft because the pressure on the top of a wing is less than that on the bottom, and air near the wingtip inevitably tries to correct the imbalance by flowing outward along the bottom surface and spilling around the tip. This cannot be prevented by any sort of gadget affixed to a wingtip, any more than the flooding of New Orleans could have been prevented by putting vortex generators on top of the levees.

The drag associated with the production of lift -- the so-called induced drag -- manifests itself as the work done in setting the tip vortex into motion. It has a certain minimum theoretical value for a wing of a given span producing a given amount of lift. What wingtip treatments try to do is minimize the effort wasted in producing the vortex -- that is, to produce as weak a vortex as possible as efficiently as possible.

The simplest, most direct and most inarguably effective tip treatment is to put the tip farther away -- that is, to increase wingspan. Wingspan is limited by a lot of factors, however; wingtip devices try to get the same effect as a span increase without actually adding any span.

Nature, a vast research project that has been going on for billions of years, is a good place to begin one's investigation of almost any topic. Soaring birds, we observe, come in two broad classes. Some -- generally the pelagic ones -- have wings of long span and narrow chord; in other words, a high aspect ratio. This is the expected configuration, generally understood to reduce induced drag. But some -- hawks and vultures -- have broad, short wings with tip feathers that they can open up, like a Venetian blind, for soaring flight. These splayed tip-feathers appear to enhance soaring performance considerably, and efforts have been made to duplicate their effect on airplanes. A Swiss aerodynamicist, Dr. Ueli La Roche, reports very promising results -- such as increases of 50 percent in effective aspect ratio -- in both model and full-scale tests, a claim that would be hard to swallow if we did not have the example of vultures to encourage us in this as well as in many other aspects of our lives.

Another source of guidance is the wind tunnel. A wing section extending all the way from one tunnel wall to the other produces no tip vortex and no induced drag. Except for minor interference effects where the wing meets the walls, it is similar to a wing of infinite span. It occurred to early experimenters that the efficiency of wings might be increased by attaching a wall to the wingtip and thereby impeding spillage. It turned out that a small wall did no good, and a large one, a chord length tall or so, added so much drag that it canceled itself out. Nevertheless, the 1930s' practice of placing vertical surfaces at the tips of the horizontal stabilizer received some encouragement from an expectation of increased longitudinal stability and elevator effectiveness.

It was in the mid-1970s, during the first "fuel crunch," that the famous NASA aerodynamicist Richard Whitcomb came up with the idea of the winglet, which has turned out to be the most conspicuously visible of his many contributions to aerodynamics (we have him to thank for the transonic area rule and the supercritical airfoil as well). Experimentally affixed to the wings of an Air Force KC-135 tanker, Whitcomb winglets yielded a 5 percent improvement in cruising fuel consumption. Affixed to the wingtips of countless airplanes since then, they have had an unmeasured effect on fuel consumption but have produced a marked gain in stylishness.

Beech referred to the winglets on the Starship (which did double duty as vertical stabilizers) as "tipsails." That name, despite a strong odor of marketing-department hokum, actually described very aptly how winglets work. Because of the spiraling path of the air streaming around the wingtip, the flow above the tip is angled inward. The lifting force produced by a properly oriented airfoil in this oblique flow inclines slightly forward, producing a net thrust that offsets some of the drag penalty represented by the tip vortex. The winglet is, in fact, exactly analogous to the sail of a boat tacking upwind. It captures some of the energy in the tip vortex and uses it to pull the airplane forward.

Whitcomb's original winglet design consisted of a small element, like a tooth, that projected downward from the leading edge of the wingtip, and a larger element offset aft by about a third of the wingtip chord. This arrangement, while perhaps optimal, was comparatively difficult to build; it has been largely supplanted in modern applications by a simple upturned tip, aligned with the wing and often so tall as to make me wonder how much rotative energy can really be found that high up and that far forward. We may be drifting into the realm of faith-based aerodynamics here, or of the car-wash effect -- that is, the common observation that the engine seems to run better after the car has been washed.

Not everything that is done at a wingtip is done to reduce induced drag. That's why my wingtip extensions are not winglets. Their purpose is quite different: to increase dihedral effect.

Dihedral is the spanwise angle of wings with respect to the horizontal plane. For wings that consist of a single straight panel, dihedral is easily defined; it gets more complicated when wings are curved (like those of Concorde) or they consist of several panels set at different angles. But the important thing about dihedral is not the arrangement or location of the tilted panels -- which can even slope downward rather than upward, as we see in airplanes with swept high wings -- but the so-called dihedral effect, which is the tendency to lift the upwind wing in a sideslip.

Dihedral effect is important for two reasons. One is that it enables the pilot to steer the plane with the rudder; this is a convenience when fiddling with maps and so on, although between GPS and autopilots the inconvenience of unfolding paper maps while keeping airplanes upright may be becoming a thing of the past. The other is that it provides natural lateral stability, keeping the airplane level in hands-off flight.

When I first flew Melmoth 2 in 2002, I found that my homebuilt, which I had provided with a rather minimal three degrees of dihedral, exhibited weak dihedral effect. You could pick up a dropped wing with rudder, but just barely. This was exactly the possibility I had anticipated when I left the wingtips unfinished. Repairing to my aerodynamics software, I used a model of the wing alone (since I was investigating a change rather than an absolute quantity, it wasn't necessary to include the fuselage and empennage), set in yawed flight, to investigate the amount of rolling force generated by various tip treatments.

I knew that two characteristics, upward tilt and leading-edge sweep, ought in theory to provide the needed boost. After trying half a dozen variations, I settled upon 45 degrees of sweep and 30 degrees of dihedral, and a 30-inch span increase rather than 20. In theory the increased span should have improved the L/D ratio by 5 percent and the rate of climb by 1 percent. Neither change, if indeed they occurred, was discernible. But the improvement in dihedral effect was, and I have spent many happy hours since then, when the wing leveler was inoperative for one reason or another, steering with my toes.

Also read these related stories:

Melmoth Flies… Again!

Melmoth 2: A Personal Airplane

Cleaning Up Melmoth

Five Years With Melmoth 2