Stealth turns 40: Looking back at the first flight of Have Blue

The pole-off

After breaking in as a self-funded effort, Lockheed was invited in the summer of 1975 (along with Northrop and McDonnell Douglas) to the next step of Project Harvey: the "Experimental Survivability Testbed" (XST). DARPA wanted the companies to build full-scale models of their proposed aircraft for wind tunnel and radar return testing, and the agency asked for an uncrewed flyable aircraft based on the design to boot. The winner of this design bake-off would win the contract to build a full prototype for further testing.

McDonnell Douglas' team took a look at the specifications for the XST's radar return threshold and balked. That left Lockheed and Northrop, who were both given the modest sum of $1.5 million to build their mock-ups. The models would be evaluated at the Air Force's Radar Target Scatter testing range at the White Sands proving grounds in New Mexico. To test the model before delivery to DARPA, Lockheed had to turn to former competitor McDonnell Douglas to make use of their radar test range in the Mojave Desert—a move Rich compared to "Buick borrowing Ford's test track." According to Rich's recounting of that test, the radar operator told Rich to check if his model had fallen off the pole... until a crow landed atop it and was detected by the radar.

The Skunk Works team delivered their models to White Sands in March of 1976. "They put our model up on a pole at the radar return range," Burnett recounts. "People were saying, we see your model. But it turns out they did not see the model—they were seeing a lot of other things, not the actual vehicle model. That original shape was not something that could have been made into an airplane, but it definitely showed the technologies were feasible."

In fact, what the Air Force test team picked up was the pole that they provided to mount the models on, which had bigger radar return than both the models. Lockheed and Northrop each spent hundreds of thousands of dollars from their budgets to build new poles with smaller radar returns—poles with 10 decibels less return than the models themselves.

The "Hopeless Diamond" had, as Rich recounted in his memoir, a radar cross section that was a thousandth of that of the D-21 reconnaissance drone—a small uncrewed supersonic drone that had been designed to fly deep into Chinese airspace to monitor China's nuclear tests. Rich said that Overholser told him the design, implemented as a full fighter-sized aircraft, would have the apparent radar cross section of "an eagle's eyeball." During testing at White Sands, the radar return increased by about 1.5 decibels from the droppings left on it by roosting birds.

Northrop's design was similar in concept to the Skunk Works model, having been designed in collaboration with Hughes Radar Systems Group (a partnership that allowed the company to tap into research on radar behavior). But the Northrop XST didn't have the benefit of the computer modeling done with Echo 1. And as a result, Northrop's model had a radar cross section from the side that was 10 times larger than the "Hopeless Diamond" model's.

In the spring of 1976, the next phase of the contract was awarded solely to Lockheed. However, DARPA wanted to keep Northrop's team moving forward on stealth, and the feds urged the company's executives to keep those researchers together. That team went on to design an aircraft called Tacit Blue for another DARPA program, and Tacit Blue became the basis for another stealth aircraft: the B-2 Spirit bomber.

The sum of its parts

Winning the second phase of the XST contract, designated by DARPA with the codename Have Blue, didn't exactly drop a big government-funded development budget in Rich's lap for the next phase—the construction of two piloted prototypes. The project had been passed over to the Air Force at this point for funding, and there was only $20 million available from the discretionary funds for classified programs. To finance the rest of the construction of the two aircrafts, Rich had to convince Lockheed to pitch in another $10 million of the company's own money.

The first Have Blue aircraft wouldn't be a full-blown demonstrator. "There were only about 25 requirements for that first vehicle," Burnett explained. "It was a very simple proof of concept prototype, simplified to the point where it was only what really matters—getting down to what we consider at Skunk Works to be the one miracle that the program needed to go solve, and that was the stealth technology." The first aircraft also carried a rather unstealthy addition: an instrument boom that extended from the aircraft to take measurements during flight.

Since the shape and other details of the aircraft itself were the crux of the problem, Lockheed pulled parts and tooling from other existing aircraft programs to turn the design into an actual flyable aircraft. Most of them were from aircraft built by Lockheed's competitors already in the Air Force's logistics system.

"We were able to take a lot of off the shelf components from other vehicles," said Burnett. "The flight control system was lifted out of (General Dynamics') F-16 design. The engines were out of the (Northrop) T-38, and the nose gear also came off a T-38. And the main gear were from the F-104, if I remember correctly. The ejection seat was from the F-16." To put it all together, tools were pulled from Lockheed's C-5 cargo plane line.

The F-16's flight control system was an important piece of the puzzle, because it was a "fly-by-wire" control system. Since the strange nature of Have Blue design would make the aircraft inherently unstable in flight—more than any human pilot could cope with—it needed a "fly-by-wire" control system to make it work. "Because we knew we had a very unstable vehicle, the F-16 flight controllers made a lot of sense," Burnett said.

The rest of the cockpit layout was decidedly low tech. Aside from the F-16's "side stick" arrangement, the Have Blue aircraft used "a lot of the old-fashioned steam gauges for instrumentation," Burnett explained.

Constant calculus

Fly-by wire systems were relatively new territory in 1976. Taking inputs from sensors, these systems based on analog computers "made up of resistors, capacitors, and inductors to create the second order equations" that calculate corrections, Burnett explained. "They take accelerometer and air data coming in, and are continuously calculating what we want the plane to do in real time," sending corrections to the aircraft's control surfaces to enhance the stability of the aircraft without the pilot having to act. "The trick is to have the feedback system stabilize the airplane so that the pilot has a nice flying aircraft as it appears to him, but let the control system do that stabilizing," said Burnett.

There had been numerous attempts to create "adaptive" controls in the early 1970s. NASA's X-15 rocket plane program experimented with controls to assist the pilot and ease the handling of the aircraft, and they found mixed results. An Air Force Research Lab project called the B-52 Control Configured Vehicle (CCV) in 1972 "really started to look at some of these advanced fly-by wire systems," Burnett said. "That tech led into the YF-16 and the development of the F-16." The YF-16 was the first aircraft designed from scratch with a fly-by-wire system, and its first (unintentional) flight was in January of 1974.

The F-16 and Have Blue obviously had much different shapes, so they required completely different "control laws" to govern how they worked, according to Burnett. Keeping Have Blue flying by just pilot adjustments would be practically impossible because of the reaction times required. The F-16's stability problems were much simpler than those of Have Blue—while the F-16 had some longitudinal instability in subsonic flight, Have Blue was "unstable on multiple axes," explained Burnett. "That was kind of the big (engineering) push (in its development), but the infrastructure of the F-16 control system—the quad-redundant system—was easily tailor able by the manufacturer. So we could go in and change some of the feedback paths and the gains to better reflect the characteristics of our shape relative to an F-16."

Burnett said that the stability problem itself was not extremely complex. "It just takes a good sharp pencil and a good engineer to sit down and figure out where the break points are and how the gains need to change as a function of the characteristics of the phase of flight," he noted. "We fortunately had some phenomenally good engineers to do that."