IT HAPPENED LAST MAY, DURING an air race in South Africa. An airplane was descending toward a turn point in a valley when the pilot of a following airplane saw what appeared to be paper scattering behind it. An instant later, a shattered wing separated from the fuselage, falling to earth a hundred yards from the main wreckage. The airplane was a Czech-built Aveko VL-3 Sprint, an NTCA (non-type-certificated aircraft) that bills itself as the "world's fastest ultralight." The terminology is confusing, because international regulations concerning small sporting aircraft have not been brought into alignment. In the United States, ultralights are very light and slow aircraft, often with sail-like fabric wings; their empty weight is limited to about 250 pounds and their speed to 55 knots; some have conventional aerodynamic controls, while others are controlled by the operator shifting his weight with respect to the wing. Elsewhere in the world, ultralight means what we call "light-sport aircraft," or LSA; these are aircraft that may weigh up to 1,320 pounds gross and cruise at up to 120 knots. This is the category to which the VL-3, a sleek 100 hp machine that is available with retractable gear and looks almost exactly like a Lancair 235, belongs. Although it must be restricted in various respects to qualify as an LSA in the United States (where it is sold as a Gobosh 800XP), the Sprint (vl-3.com) claims a maximum cruising speed of 145 knots and a never-exceed speed of 162 knots.
The accident "has caused much discussion and outright concern," a South African friend wrote to me, "about flying in the yellow arc and at or beyond VNE."
If the accident inspired "outright concern" about flying beyond VNE, then the daring pilot of the accident airplane will at least not have died entirely in vain. VNE, the redline or never-exceed speed, is a subject about which many pilots are understandably confused, but it ought to be pretty clear that the words beyond and never exceed do not belong in the same phrase.
Logically, any airplane ought to have an absolute maximum permissible speed, even if it is only the speed at which the nose begins to melt, because it stands to reason that in any airplane there is bound to be some high speed beyond which it is not safe to fly. But VNE is not that speed; it is a lower speed, selected to provide a margin of protection against the vagaries of air, wear and tear, instrumentation and human error.
The U.S. rules for certifying aircraft are found in several different locations, depending on the type of airplane, but Part 23 — a subpart of Title 14, Aeronautics and Space, of the Code of Federal Regulations — is typical. Descended from the earlier CAR 3, under which some types still in use today were certified, Part 23 came into force in 1965 and governs the certification of airplanes of up to 12,500 pounds gross weight, as well as commuter aircraft of up to 19,000 pounds. (Part 25, which is similar to Part 23 in many respects, applies to transport aircraft.) Light-sport aircraft are approved under a "consensus standard" developed by the American Society for Testing and Materials, a nongovernment body of experts that designs industry standards for just about everything; the ASTM standard, like other aircraft certification standards all over the world, is largely based on Part 23.
Section 335 of Part 23 defines several important speeds. The first is the design cruising speed, which, for airplanes with wing loadings of less than 20 pounds per square foot — that is, most small single-engine airplanes — must be 33 times the square root of the wing loading. For a wing loading of 16 psf, VC would be four times 33, or 132 kias. This speed is a purely formal requirement, used to provide a framework for structural and other decisions; many airplanes can't actually cruise at their "design cruising speed." But the design cruising speed provides a basis for another speed, VD, the design dive speed, which is generally 1.4 times VC; and VD is in turn the basis for VNE, the never-exceed speed, which is nine-tenths of VD. If an airplane can't achieve its design VD in flight, then the dive speed attained in flight test, VDF, replaces VD, and VNE is, again, nine-tenths of it.
These speeds, among others, define various corners of the flight envelope, which in turn determine the required strength of major structural components. But absolute speed affects many aspects of a structure besides obvious things like wing spars and fuselage skins. For example, cowlings must be sufficiently stiff to not bulge or blow apart under the internal pressure of ram air, which, at 175 knots, is about 100 pounds per square foot. Canopy latches and hinges must be strong enough to resist the considerable lift developed by a curved surface at high speed. Higher speeds imply lower angles of attack, and it's even conceivable that a wing with a lot of built-in washout or twist could fail above a certain speed because the angle of attack of the outer panels becomes negative, and they begin to push downward, subjecting the lower spar cap, which is the smaller, to an excessive compression load.
All of these loadings are due to air pressure, which grows in proportion to the square of the speed — double the speed produces four times the force; they are functions of the indicated airspeed, not the true. Now, VD, which is an indicated speed, is by definition a safe speed; the forces at VNE are just 81 percent of those at VD (0.9 x 0.9 = 0.81), and so there is a comfortable margin of safety, so far as structural strength is concerned, at VNE.
But there is a complication that muddies the water considerably. It is flutter. Flutter is a vibration that may be augmented by aerodynamic forces. It is the one challenge to aircraft structures that does not increase gradually with speed. It is possible for a structure to perform normally right up to a certain speed and then, with a gain of two or three more knots, to explode into fragments in a split second. That is what most likely happened to the South African VL-3. The accident has not yet been investigated, but it has the earmarks of wing flutter induced by a vibrating aileron.
Flutter is affected by a number of factors, one of which is the true, not the indicated, airspeed. As you will have immediately perceived, this fact raises a logical difficulty. VNE, the redline on the airspeed indicator, is an indicated airspeed, but the critical flutter speed may be a true airspeed. So the margin separating VNE from the critical flutter speed gets smaller as you gain altitude. Furthermore, if you get really high up, the difference can be larger than the margin that separates VD from VNE, simply because the difference between indicated and true airspeed is greater than 10 percent.
That doesn't mean the airplane will flutter, because VD is not the critical flutter speed. Manufacturers are not required to determine the critical flutter speed for each design, but only to demonstrate that it is free of flutter up to VD and that there is good reason to believe, based on various kinds of ground tests and mathematical analyses, that it will remain so up to 1.2 times VD. It is noteworthy that the section of Part 23 regarding flutter, 23.629, makes no mention of altitude. The cumulative margin between VNE and 1.2 times VD is 33 percent, and this probably provides a good cushion in all normal operations, but if I were to ride a wave to 35,000 feet in a 172, I would not be in a hurry to peg the airspeed at redline on the way back down.
The flutter criteria in 23.629 are quite complex. They attempt to take into account externalities like wear and tear, hidden damage and even in-flight failure of various components. But the kind of testing and validation that the FAR envisions is outside the scope of amateur builders and, I suspect, of many manufacturers of kits, ultralights and LSAs. Flight flutter testing is expensive and dangerous, and ground vibration testing, which goes a long way toward ensuring freedom from flutter, eliminates only the danger, not the expense. I haven't conducted a poll, but I suspect that the flutter testing of most homebuilts and LSAs has been limited to "pulsing" the controls in flight — that is, hitting the stick or kicking the rudder — at gradually increasing speeds while looking for any softness in the damping of the resulting control movements, and while wearing a recently packed parachute. This method is fine as far as it goes, but it's far from perfect.
There are few sharp breaks in aerodynamics. Flying in the yellow ("smooth air") arc, you are obviously safer near the green end than near redline, even though the whole arc is equally yellow. You are not infinitely less safe flying one knot above redline speed than one knot below it. All transitions are gradual, with a single exception: the transition into flutter.
In certified, manufactured airplanes that are not very old and have not been modified or repaired, you are safe at redline speed in smooth air at any altitude below the service ceiling. In uncertified airplanes, more caution is in order; you should shy away from redline speed at high altitude, especially in types of which there have not been many examples flying for many years, and you should not exceed the redline speed at any altitude. That's why it's called the "never-exceed speed." After all, which part of "never" is hard to understand?