Conventional aerospace wisdom holds that stealth and high performance--like matter and antimatter--cannot be friends. Gains in one area usually lead to sacrifices in the other.
Take, for example, the F-117 stealth fighter of Gulf War fame. Its sharp edges, abrupt angles, and flat surfaces deflect radar waves to deliver impressive stealth--while simultaneously flouting the laws of aerodynamics. Though amazing for its time, the F-117's subsonic, non-aerobatic capabilities emphasize invisibility over agility. By contrast, there's the F-16, a sexy-sleek, supersonic air-superiority fighter possessed of outstanding maneuverability--and a quite noticeable radar signature.
In April, engineers from NASA and McDonnell Douglas unveiled an aircraft designed to prove the conventional wisdom wrong: the X-36, a subscale, remotely piloted demonstrator of an extremely agile, very stealthy fighter. Developed in 28 months at a fire-sale price of $17-million, its job is to prove several fundamental but high-risk technologies--all under development during the past five years by engineers at both NASA and industry. The technologies include:
- A low-profile, tailless planform with unique split-aileron control surfaces
- A classified thrust-vectoring system
- High-rate, digital, fly-by-wire flight controls.
"These are all brand new technologies," says Larry Birckelbaw, X-36 program manager at NASA, "and we recognized that any potential customers would be hesitant to incorporate them into an advanced aircraft design without them being sufficiently proven. That's the purpose of the X-36."
The X-36, a 28%-scale model of a hypothetical 40,000-lb class aircraft, actually employs no radar-absorbing materials in its structure. However, the basic configuration would be a stealthy one were the vehicle full size. In addition to being stealthy, the design explores reductions in weight and drag, and improvements in range, maneuverability, and survivability for possible application to future designs.
Aside from its duties in the air, the X-36 also provided McDonnell Douglas with a test bed for demonstrating several innovative engineering and precision manufacturing techniques (see the "Phantom Works" sidebar), such as:
- Low-temperature cured composites
- Low-cost tooling molds
- Unitized, high-speed machined assemblies
- Advanced CAD/CAM, straight from "screen to machine."
Cat-agile, but no tail. The design's most obvious departure from current fighter design is its lack of an empennage. "There are some obvious benefits to removing the tail," says Dave Manley, program manager at McDonnell Douglas. "One is the smaller radar signature; the others are reduced drag and weight." Lopping off the tail saves an immediate 10% in mass and produces a "fairly significant" drop in RCS (radar cross-section). But most importantly, eliminating the tail cuts drag. The X-36 experiences 20 to 30% less aerodynamic drag than an equivalent-size F-15.
On a conventional aircraft, of course, the tail supplies important stability and control forces. These must be generated in another way on the X-36. That other way requires a variety of aerodynamic surfaces. These include two large forward canards for pitch control, and on the outboard trailing edge of each side of the main wing, two pairs of split ailerons. The ailerons' upper and lower surfaces can move separately or together to act as drag rudders (for yaw control) and as ailerons at the same time. Inboard of the ailerons are solid (non-split) elevons, and at the leading edge lie full-span flaps.
"The breakthrough really is the way the canards and split ailerons work together aerodynamically to create the required yaw, roll, and pitch rates," says Birckelbaw. "A lot of the new technology is in the control-law development." These control laws are the heart of a single-channel, fully digital, fly-by-wire flight-control system. It re-sides in a special McDonnell Douglas-developed computer driven by a trio of Texas Instruments (Austin, TX) digital signal processors, said to produce 117 MIPS. Such substantial processing is needed onboard to keep the vehicle's inherent instability in check and to account for scaling laws. An aircraft's response varies with the inverse-square of the scale factor--meaning the X-36 responds 1.89 times quicker than if it were full-size.
"We believe this is the fastest that any control system has had to operate," says Birckelbaw. "If it weren't for recent advances in the hardware for the onboard computer systems and digital flight-control laws, you wouldn't be able to fly a vehicle like this."
Initially, the flight control software will be based on values calculated from wind tunnel and CFD analysis. But engineers intend to measure the vehicle's dynamics during test flights, and then update the control laws with actual values between flights.
The thrust vectoring advantage. Though the vehicle represents a hypothetical full-size supersonic fighter, the X-36 itself isn't supersonic. Engineers optimized neither the engine inlets nor wing-area ratios for high-speed flight. This is far from a handicap, since the most challenging portion of the flight envelope occurs at the slow end. "For a vehicle with no tail, the supersonic control characteristics are not as difficult as the low-speed, high angle of attack handling qualities," says Birckelbaw. Under these conditions, flow over the wings detriorates, reducing control-surface effectiveness. Maintaining flight control is one job for the still-classified thrust-vectoring system.
Yet surprisingly, the X-36 relies only moderately on vectored thrust for most of its handling needs. Thrust vectoring isn't essential for flight--an intentional safety feature. And should it fail, even in a hard-over position, the aircraft possesses sufficient authority from the control surfaces alone to overpower the skewed thrust and make it home. "With the nozzle, the aircraft's agility is equal to or better than that of current fighters," says Manley. "Without it, it's maybe 85-90% of current fighters."
Seems like quite a technological investment for a 10% boost in agility. That's because the thrust vector system's major benefit occurs, quite unexpectedly, during cruise. By allowing the X-36 to maneuver at speed without deflecting any drag-inducing aerodynamic surfaces, the thrust vectoring greatly improves range. Maintaining a clean, buttoned-up configuration during cruise doesn't hurt the aircraft's stealthiness either.
Pint-size performer. Much of the X-36 story revolves around the vehicle's size. A beneficiary of NASA's "better, cheaper, faster" philosophy, the program hopes to show that it's possible to get useful results for less money. "The official word is that we did it in one-third the time at one-fifth the cost," says Birckelbaw, "but I think it was even cheaper than that."
It might come as no surprise that a Lilliputian aircraft commanded a Lilliputian price. But the "one-fifth the cost" figure compares to other subscale programs of the past, such as Himat, a 1970s era remotely controlled test aircraft. In today's dollars, Manley estimates that Himat would cost $75 million. "What we're really after is the aerodynamics and controls integration," he says. "If you put a pilot in it and did it full scale, you're talking hundreds of millions of dollars to do this."
What is surprising is that a government and industry team could work together so well in such a short time, with so little in the way of funds and personnel. The consensus is that the program was a beneficiary of such frugality, not a victim. At its peak, the staff numbered about 40 people, with much sharing between the organizations. "If you walked in during this program at most any time, you couldn't tell who was the contractor and who was the civil servant," says Birckelbaw.
McDonnell Douglas' Phantom Works, where the vehicle was designed and built, applied many low-cost manufacturing methods--driven by necessity as much as the desire to test the new techniques. The low budget also resulted in less government concern about failure and thus less unwanted oversight. "Even if we lose both of the X-36 demonstrators, we would still learn a lot from this program," Birckelbaw explains. "For $17-million, you could even lose lots more and still be way ahead."
Testing begins this fall with a series of 25 flights from Edwards Air Force Base. A ground-based pilot will control the aircraft from a cockpit station in a mobile data-collection van. The pilot gets a view out the front of the aircraft from a camera embedded into what would normally be the aircraft's canopy.
Landing gear allows the X-36 to take off and land on its own--a feature absent from the Himat drone, which had to be dropped from a B-52. Goodyear (Akron, OH) designed and produced sets of special aircraft tires to withstand the high tire rotational speeds X-36 will produce. In an emergency, a parachute situated behind the mock cockpit can be deployed to float the aircraft to earth at the landing gear's maximum vertical-rate limit of 14 ft/second.
Will future programs take the low-budget road being paved by projects such as the X-36, mitigating risk and perhaps enabling a greater number of designs to be studied for less money? Only time will tell. For Birckelbaw, though, the proof of such an approach is already in: "It's the most cost effective way to demonstrate new technologies like these."
Phantom in the Works
Atop plywood supports, deep in the heart of McDonnell Douglas' Advanced Systems and Technology Phantom Works division, lies an open-faced fiberglass mold upon which workers cut and lay sheets of composite to air-dry and form the fuselage for America's next-generation fighter-demonstrator, the X-36. Not exactly the stereotyped vision of cost-is-no-object manufacturing everyone imagines. Yet this approach is representative of the lean engineering and production approaches that made possible the manufacture of this sophisticated, inexpensive vehicle.
Formed in May, 1991, Phantom Works is a new high-profile front end for the corporation. Its role: to shorten the transition of technologies and processes from R&D to production. How? Prove the technologies on programs such as the X-36 to mitigate risk before taking them to the main shop floor.
Comparisons to Lockheed-Martin's legendary Skunk Works are inevitable. The difference, says Gerry Ennis, vice president of the Phantom Works Prototype Center, is that, "we not only develop new technologies and processes, we demonstrate them and then move them to production." Harry Stonecipher, president and CEO of McDonnell Douglas, charged Ennis with the difficult task of forming an organization that didn't lurk behind a cloak of secrecy, but rather incubated engineering ideas to be applied corporate-wide. People move in and out of the division regularly, following their programs from design to production. "That way," says Ennis, "their attitudes and ideas can infect the entire organization."
What's it take to be a Phantom Worker? Ennis looks for generalists, engineers with experience in design, analysis, and manufacturing--and a passion for airplanes. "People who understand the big picture produce better designs," he says.
The division always keeps its eye on cost. By focusing on key processes, Ennis feels that both superior quality and lower cost are attainable. "It's like Rolls Royce and Lexus. By focusing on the process and taking out the hand labor, Lexus puts out a car of equal or better quality at a percentage of the cost," he explains.
A Phantom Works-inspired example is the Manufacturing 2005 initiative on the C-17. Engineers completely re-engineered the tail, reducing the number of parts by 85% and fasteners by 82%. High-speed machining--unavailable just a few years ago--combines many separate components into a single machined structure. Cost plunged 50% and the tail's weight dropped by one-fifth.
Other technologies and processes being studied in the Phantom Works include:
- Room-temperature cured composites that eliminate the costly tooling and time needed to bake parts in an autoclave
- Pre-coated fasteners that simplify installation, provide corrosion protection, and save $2-million in labor on each C-17.
- Automatic fiber placement that allows large composite structures to be produced more quickly
- Advanced solid-model CAD (Unigraphics II from EDS) that allows engineers to use a single model for every aspect of the design and production.
Control in tight places
One of the more difficult challenges on X-36 involved the design of the control-surface actuators, supplied by Moog Inc. (East Aurora, NY). To maintain the aircraft's stealthy profile, the vehicle's engineers insisted that no bumps blemish its exterior. Yet the wings were too thin to accommodate conventional actuators.
"The only way we could get everything to fit was to separate the control modules or valve logic from the pistons, and plumb it with hard lines," says John Kopp, design engineer at Moog. Normally, the manifolds and actuators would lie just inches apart. In the X-36, that distance stretched to about 10 ft, producing a load resonance problem because of the excess oil volume. Engineers addressed this by tuning the control electronics and adding special damping circuits.
In all, Kopp relocated the control-module functions for 17 channels into the aircraft's belly--including those for all the wing actuators, canard actuators, thrust vectoring, and landing-gear doors. These are fed by three manifolds, two with six channels and one with five. Each manifold squeezes into a roughly 2- by 8- by 11-inch space.
The pistons for the six-inch-long actuators were too small to accept LDVT position sensors on their centerline--the usual location--so Kopp moved them to separate pistons alongside the main one. Moog farmed out the actuators' production to Emco Fluid Systems (Valencia, CA). Each produces about 1,200-lbf at 3,000 psi and are made entirely of aluminum, except for the pistons, to keep down cost; the entire system, actuators and all, weighs 52 lbs.
By using Unigraphics II, Kopp could easily alter the internal boring to accommodate different strokes in a single design. To date, the company has received half a dozen calls from other customers who want the tiny actuators. And Kopp is busy turning the design into a parametric model to produce quick, custom designs on the fly.