That New Black Magic

In the early years of the cold war, enter Kelly Johnson and an clean sheet of paper—long enough to accommodate an 80-foot wingspan.

Clarence L. “Kelly” Johnson of Lockheed’s secret “Skunk Works” facility, chief designer of the U-2. (USAF)
Air & Space Magazine | Subscribe

On Friday, November 26, 1954, Lockheed Aircraft's chief engineer, Clarence L. "Kelly" Johnson, called Edward Baldwin, Elmer Gath, and three other trusted engineers into his office, which was then at the Burbank Airport, near Los Angeles. Baldwin specialized in structures, Gath in engines. Starting the following Monday morning, Johnson said, they would be working on a project for the Central Intelligence Agency. It was so secret that the next 18 months to two years would have to constitute a blank spot on their resumes. Within eight months, Johnson said, they were going to build the world's first dedicated spyplane.

Even after World War II, the United States had gathered aerial intelligence by stuffing a fighter or bomber full of cameras and assigning it the prefix "R" for "reconnaissance." But in late 1953, the Air Force recognized the need for aircraft designed specifically to perform long-range aerial reconnaissance. The cold war had changed the military situation dramatically. Soviet nuclear weapons, ballistic missiles, and the economic and military structures that supported them were hidden behind a heavily defended curtain surrounding a landmass that spanned 11 time zones. The United States, fearful of a devastating nuclear surprise attack, was frantic to know what was going on. And the key to peering deep into forbidden territory was as obvious in the early 1950s as it had been in the Civil War, when the Army of the Potomac had launched observation balloons: You had to get up high.

In 1953 Bell and Fairchild had been invited to submit designs for an all-new high flier, while Martin was tasked with modifying the B-57, its version of the English Electric Canberra. All three were burdened with armor, systems, and heavy, high-G-tolerant structure, and Johnson knew that even a big wing could lift an airplane to 70,000 feet only if its weight were radically reduced. As an airplane climbs, its true airspeed increases in a linear manner until engine thrust begins to fall off. But indicated airspeed, which measures the force of the relative wind, diminishes constantly as the density of the air decreases in a long climb. At 70,000 feet, an airplane might be scooting through the air at an actual speed of 500 mph, but the pilot's airspeed indicator may register only about 125 mph--closer to the speed at which sailplanes cruise at lower altitudes. To the wing, it feels as if there's less wind, so it produces less lift and supports less weight.

When Johnson learned that the Air Force was soliciting three other companies for a high-altitude reconnaissance aircraft, he submitted his own unsolicited entry, the CL-282. "First, we had to study the problem of what was needed," Johnson would recall when I interviewed him almost three decades later. "How far did it have to fly? How high did it have to go to get away from the fighters? How high did it have to go to get around the missiles? And having come up with our guesses in that category, we made a proposal to the Air Force to make this very lightly loaded, high-aspect-ratio vehicle that would fly over 70,000 feet and 3,500 miles."

The CL-282 was basically an F-104 Starfighter with exceptionally long sailplane-style wings. The ratio of the wings' span to the average width of their airfoil--their "aspect ratio"--would be high, which meant they would create very little drag at the wingtips for the amount of lift they produced, thereby ensuring long range and endurance. To minimize weight, the CL-282 had only one engine and flew without armor, pressurization, an ejection seat, or even landing gear. The airplane was so simple in concept that it suggested sublime Oriental understatement. The CL-282's ability to reach high altitude was also its chief means of protection: The best Soviet fighters could reach perhaps 45,000 feet--nearly five miles lower.

Despite the entreaties of a board of scientists and engineers led by Allen F. Donovan of Cornell, the Air Force flatly opposed the CL-282 and sent Johnson a letter rejecting it. "They proved conclusively" that what would become the most successful and longest serving spyplane in history was "impossible" to build, Johnson later said, with a triumphant smirk. Well, not exactly impossible; difficult, maybe. The Air Force strongly favored the Bell X-16, a twin-engine, armored, fully pressurized design, and the stretched-wing RB-57D. But Lockheed found a customer in the CIA and had the U-2 built and flying in time to cause the cancellation of the X-16. The RB-57D, at best an interim aircraft, flew some operational missions but was doomed by fatigue cracks in its wings.

Johnson's CL-282 had gotten a second chance. Under pressure from President Dwight Eisenhower and despite his own misgivings, CIA director Allen Dulles adopted the idea and assigned the project to Richard M. Bissell Jr. An economist from Yale, Bissell readily admitted that he knew nothing about aeronautics, so he turned the design and manufacturing operation completely over to Lockheed and put together a small, tight-knit operation that got things done quickly.

Having conceptualized the essential design, Johnson assigned about 50 engineers at the Lockheed Advanced Development Company, known by then as the Skunk Works, the task of filling in thousands of crucial details. He told Baldwin that he wanted 600 square feet of wing with an aspect ratio of 10 or 10.5 to 1. It wound up being 10.67 to 1. That meant the wing would be at least 10 times longer than its average width. The wingspan, which would come out to 80 feet, was so long that the wingtip ran off Baldwin's drawing board. "Are you sure you want a wing that looks like this?" Baldwin remembers asking Johnson.

"Yeah. That's about right."

"Well, I ran off the end of the board."

"Put a little patch of paper up there to show what it looks like," Johnson answered. "Then you'll have to redraw it because the blueprint machine won't handle anything wider than 42 inches."

Altitude would be the U-2's best defense, but altitude also constituted its single most difficult engineering challenge. Its engine, rated at more than 10,000 pounds of thrust at sea level, would produce only about 700 pounds of thrust at altitude. Hydraulic systems were heavy, so Johnson eliminated the customary hydraulic boost for the controls. To save time and cost, they used the bucket seat and the control yoke from a P-38. Some of the pilots recruited to fly the first U-2s disliked the yoke (which they associated with transport aircraft), but it took the strength of both arms to fly the airplane, and the additional leverage of the yoke was necessary.

Johnson saved even more weight by designing the airplane for load factors of only 2.5 Gs, a fraction of that for normal combat aircraft. Instead of using a wing spar that passed through the fuselage, the wings, which also carried almost the entire fuel supply, were simply bolted on. This would turn out to be an ingenious solution, for the airplane would spend little time in turbulent air. (At a CIA symposium on the U-2 in September, gleeful officials reported that a recent structural evaluation indicated the current airframe is good for over a hundred years' more service.)

And in his pursuit of weight reduction Johnson also eliminated landing gear. He wanted the CL-282 to take off from a wheeled dolly and land on its belly, which would act like a skid. But reality, in the person of flight test engineer Ernest Joiner, intruded. Joiner told his boss that the flight test program would quickly fall apart if the airplane had to have its belly repaired after every landing.

It was therefore decided to install a dual-wheel main landing gear and tailwheel in the U-2's fuselage and use flexible struts, or pogos, as they came to be called, to prop up the wingtips during takeoff, after which they would fall away. But that created another problem. The U-2's all-important payload, a very large, heavy camera, was to be carried in a so-called Q-bay behind the cockpit. The engine would be right behind the bay. The only place to put the forward landing gear was therefore between the engine's air intake ducts, which formed a pair of pants whose legs straddled the Q-bay. This was not the ideal place to put the forward landing gear, Baldwin says. It was too far forward and made landing somewhat tricky. But the engineers were stuck with it.

The engine was a modified version of the reliable Pratt & Whitney J-57, which powered F-100s and B-52s. But here, too, there were birthing problems. The P-31 version of the engine, specially adapted to high-altitude flying, was dedicated to another program for the Air Force. The P-37 version, the first U-2 engine, was not designed for altitudes above 65,000 feet. It therefore tended to flame out repeatedly above 57,000 feet. Pilots dreaded the prospect of having to descend to 35,000 feet, where the MiGs and missiles waited, to restart their windmilling engines.

Then there was the oil problem. Because the atmospheric pressure at altitude was so low, oil leaked through the J-57's seals and got into the U-2's air conditioning and de-fogging systems. Engineers calculated that during an operational mission, 64 quarts of oil--the maximum capacity of the system--might be lost. The de-fogging ducts sprayed the windshield with hot air from the engine's compressor, and during long flights a gradually thickening coat of oil would form on the glass. This was solved by providing pilots with long sticks with diaper cloth attached to the ends so they could wipe the windshields clean. Someone even got the idea of welding a small metal box on the de-fog line and stuffing it with Kotex to absorb the oil. But the hot air was under so much pressure that it bent the box out of shape, says Lockheed's Bob Murphy, who was involved in many aspects of the U-2 and SR-71 programs. The problem ended only when the P-31 engine, which required less oil and was optimized for high altitude, replaced the model 37 in 1956.

Another worry was jet fuel: The standard JP-4 and JP-5 would boil away at high altitude. General James Doolittle, an executive with Shell Oil, had been a technical advisor to the reconnaissance community, and he persuaded Shell to develop a new jet fuel, designated JP-7, that had very low volatility. Production of JP-7 required most of the stocks of the petroleum products the company used to manufacture insect sprays, and although few Americans knew why, there was a nationwide shortage of bug spray in 1955.

In the near vacuum of the upper atmosphere, pilots also required special protection so that their body fluids would not bubble and boil. To this end, the David Clark Company of Worcester, Massachusetts, devised a partial-pressure suit, which was the first of its kind for keeping pilots alive in near-space conditions. This even led to the first specialized food and water provisions. Pilots could push a tube through a little hole in the face mask and suck on sweetened water or cheese- and bacon-flavored food mixtures squeezed from soft containers.

Undoubtedly the most daunting problem faced by U-2 pilots was the infamous "coffin corner," the terribly tiny margin between Mach-shock and stall buffet. At altitudes above 65,000 feet, the first U-2s had an interval of only six knots (7 mph) indicated airspeed between the onset of Mach buffet and a stall. In other words, the difference between the U-2's slowest flying speed and its fastest was only six knots. The margin was so narrow that Lockheed test pilots reported that in a bank, a U-2's inner wing could be stalled while its outer one was buffeting wildly from excess speed.

And as Ben R. Rich, the engineer who succeeded Johnson as head of the Skunk Works, said, "The shuddering felt the same whether it was because of flying too fast or too slow, so a pilot had to keep totally alert while making corrections." Once a U-2 pilot reached 70,000 feet and 400 knots true airspeed, he tried very hard to stay right there.

"The original U-2 was a difficult airplane to fly," says Garfield J. Thomas, vice president of Reconnaissance Systems at what is now Lockheed Martin. "It was a lot of work." Knowing that, the inventors of the revolutionary camera designed it to be automatic. "The pilot really had very little to do with the camera," adds Thomas. "The camera was usually pre-programmed and set up. When he reached a certain area, in the old days, he'd just throw a switch."

The camera itself was the result of a remarkable collaborative effort between Edwin "Din" Land, inventor of the Polaroid Land Camera, and James G. Baker, a Harvard-educated astrophysicist whose interest in optics went back to the 1930s. Land, a longtime member of the reconnaissance community's inner circle, led a group of presidential science advisors known as Project 3. In the U-2, he saw an airplane that could carry a camera good enough to count Soviet bombers and resolve the controversial "bomber gap," and it was Land who introduced the concept to President Dwight Eisenhower. Baker designed the U-2's camera, which carried a mile of specially developed, ultra-thin Eastman Kodak film. The film itself weighed around 300 pounds and had to be spooled on tandem nine-inch-wide rolls that fed in opposite directions on parallel tracks to maintain the airplane's center of gravity (see "Captured on Film," above).

The so-called Type B camera was fabricated by Hycon Corporation in California. Fitted into the Q-bay, which was pressurized to 0.25 atmosphere (one atmosphere is a unit equal to the pressure of the air at sea level: 14.7 pounds per square inch), it was mounted on a hatch that actually formed a section of the skin of the airplane. The U-2 (and its supersonic successors, the A-12 and SR-71) would also come with noses that could be replaced with others carrying different sensors, the way lenses are changed on cameras.

Baker, now a gentleman of considerable years, still spends most nights doing experiments in his basement laboratory in New Hampshire ("Don't call before 10 a.m.," he warns). He credits Walter Pierstorff, the general manager of the Schott Company in Mainz, West Germany, with supplying "excellent optical glass of many types." And, Baker adds, there were "no questions asked." The result was a lens that could pick out a basketball from over 13 miles in the sky. The master optician recalls "jousting" with Kelly Johnson over how much space the camera was going to have and how much it would weigh. Johnson allowed him about 500 pounds.

Baker does not seem as impressed with his creation as the rest of the world is. The B camera, he says, was simply an evolutionary development arising from other work that went back to World War II. He is too modest. The best the older cameras could achieve was 20- or 25-foot resolution at 33,000 feet. At more than twice that altitude, they would be useless, especially for intelligence purposes, in which photo interpretation required 10-foot resolution. Baker's challenge was to design a camera that would be four times better than anything ever built.

Baker also had to meet Kelly Johnson's weight limits, so he replaced a heavy and bulky prism used to scan to the left and right of the airplane's course with a single mirror mounted within a swiveling housing. The assembly followed an automatic sequence to capture overlapping images of a swath of ground that stretched from horizon to horizon.

For the pilots, the airplane was a handful to fly, uncomfortable for many hours, and although initially out of reach of Soviet weapons, that small luxury would not last long. To make it less visible, additional measures were tried. The first, and most obvious, was a flat, midnight-blue paint scheme to match the dark sky. It actually didn't blend all that well, but dark blue was better than polished aluminum, which had led to reports of lights in the sky over the Nevada desert where the U-2 was first tested. ("Pastels are the best stealth colors," Ben Rich once observed, "but real men don't fly pink jets.")

Another technique involved swathing the underside with a metallic grid called a Salisbury Screen, then covering it with black foam rubber to capture and dissipate radar microwaves. The third technique, tested in a program aptly named Dirty Bird, called for adding metal "standoff" posts from tail to nose and connecting them with wires set at precise distance from the skin. MIT's Lincoln Laboratory said the wires would cancel or scatter the radar energy reflected off the skin. Neither scheme worked, but the posts were especially disastrous. Some wires snapped during flight tests and lashed the aircraft like whips. More important, they cut the U-2's performance so badly that they nullified the reason for building it. "They went out and flew it," recalls Garfield Thomas, "and it was so draggy they would never get to altitude and never have any range." Rich said simply, "They made it look like a rake."

Even before the U-2's first missions, in June 1956, the CIA worried that its operational life over the Soviet Union would be no more than two years because of rapid improvements in Soviet air defenses. They were wrong; it would be nearly four years before Francis Gary Powers was downed. The airplane designated with a lowly "U" for "utility" to hide its true purpose would enter Air Force service in 1957, photograph Soviet missiles entering Cuba, and, in the hands of Nationalist Chinese pilots, penetrate the People's Republic of China. Today it flies frequent missions over Iraq in support of the United Nations' surveillance of that nation's efforts to produce weapons of mass destruction. But it was the knowledge that a U-2 would eventually get nailed that drove Kelly Johnson to invent its high-flying supersonic replacement.

Comment on this Story

comments powered by Disqus