Skunk Works: A Personal Memoir of My Years at Lockheed

9 FASTER THAN A SPEEDING BULLET

The Blackbird, which dominated our work in the sixties, was the greatest high-performance airplane of the twentieth century. Everything about this airplane’s creation was gigantic: the technical problems that had to be overcome, the political complexities surrounding its funding, even the ability of the Air Force’s most skilled pilots to master this incredible wild horse of the stratosphere. Kelly Johnson rightly regarded the Blackbird as the crowning triumph of his years at the Skunk Works’ helm. All of us who shared in its creation wear a badge of special pride. Nothing designed and built by any other aerospace operation in the world, before or since the Blackbird, can begin to rival its speed, height, effectiveness, and impact. Had we built Blackbird in the year 2010, the world would still have been awed by such an achievement. But the first model, designed and built for the CIA as the successor to the U-2, was being test-flown as early as 1962. Even today, that feat seems nothing less than miraculous.

Kelly Johnson’s disappointment over our failure to produce a workable hydrogen-powered airplane had him pouting for a day or two, but he quickly recovered and began lobbying the CIA for a new spy plane to fly over Russia that would be a quantum leap over the U-2 in every way. He assembled the small group of us who had worked on the hydrogen plane and had us brainstorming ideas for new designs and approaches for an airplane that used conventional engines and fuel but still could outrace any Russian missile. “It makes no sense,” he said, “to just take this one or two steps ahead, because we’d be buying only a couple of years before the Russians would be able to nail us again. No, I want us to come up with an airplane that can rule the skies for a decade or more.”

At that point, in April 1958, the U-2 overflights of Russia were in their second year and going well. In fact, it would be two more years before Francis Gary Powers was shot down, and the high priority at the Skunk Works was Operation Rainbow, our attempt to lower the U-2’s radar cross section. But Kelly declared the U-2 doomed. The Russians had made it a matter of national honor to find a way to stop U-2 overflights and were investing billions of rubles in rushing to develop a missile system to do so. Dick Bissell shared Kelly’s glum outlook for the U-2’s future and encouraged him to begin sketching out a successor spy plane. “We’ll fly at ninety thousand feet, and jack up the speed to Mach 3. It will have a range of four thousand miles,” Kelly told a group of us. “The higher and faster we fly the harder it will be to spot us, much less stop us.”

I didn’t know about Richard Bissell of the CIA, but Ben Rich of the Skunk Works reacted to Kelly’s idea with jaw-dropping disbelief. He was proposing to build an airplane that would fly not only four times faster than the U-2 but five miles higher — and the U-2 was then the current high-altitude champion of the skies. A Mach 3 airplane was 60 percent faster than the maximum dash capability of our top-performance jet fighter. Experimental rocket airplanes had flown at blinding Mach 3 speed using powerful thrusters for two or three minutes at a time until fuel ran out. But Kelly was proposing an airplane to cruise at more than three times the speed of sound, that could fly coast to coast in less than an hour on one tank of gas.

Kelly’s audacious idea would probably not have been taken seriously by the CIA coming from anyone other than the boss of the Skunk Works. After all, in 1954 we had built the F-104 Starfighter, the world’s first Mach 2 fighter. So a Mach 3 airplane seemed a logical extension of our skills. However, there was a Grand Canyon — size gulf between designing an airplane like the F-104 that could kick in its afterburners on takeoffs and in dash modes lasting a minute or two, and designing an airplane whose “normal” cruising speed was nearly twice as fast as the fastest fighter’s dash speed. On afterburners, a fighter was burning fuel at a rate four times faster than at cruise speed, so afterburners were saved for combat threat situations — escaping flak after a bombing run or outflying missiles or dogfighting MiGs. We were proposing to fly whole missions on afterburners. The technology confronting us was so far beyond anything on the drawing boards of any other aerospace company in the world that we might as well have been proposing commuter rocket service between the moon and the outer ring of planets.

For openers, to be able to fly sustained at such heights and speeds would require radical departures in how we designed and built propulsion systems.

“Rich,” Kelly said, turning my way, “I’m making you program manager for the propulsion system.” He ignored my stunned expression. “How hot do you suppose the airplane will get at Mach 3 in sustained flight?” he asked me. “Somewhere between a blowtorch and a soldering iron, I guess,” I replied when my voice returned. He nodded. “Probably around nine hundred degrees at the nose,” he said. “Just imagine that kind of thrust! You’re the lucky one. You’ll at least have known laws of physics to guide you. The rest of us are going to have to do some fancy stretching to find out what can work. We start from scratch as if we are building the first airplane, just like the Wright brothers.”

If I had been older and smarter, I would’ve run for the nearest exit. I had to produce a propulsion system more efficient than any other ever designed. I was then only a thirty-two-year-old fledgling, still on probation to prove my worth as a propulsion and thermodynamics engineer among many of my senior colleagues. But I was cocky enough to shrug off Kelly’s challenge and think, A Mach 3 airplane! Why in hell not? Kelly surrounded himself only with the kind of can-do guys that made American aerospace technology preeminent. To him, the word “impossible” was a gross insult.

Kelly promised to deliver the world’s first Mach 3 airplane to the CIA only twenty months after we signed a contract. That also seemed to me, in my pathetic innocence, a reasonable deadline. After all, it had taken us only eight quick months to deliver the first U-2. Had I really thought about it, in complexity the U-2 was to the Blackbird as a covered wagon was to an Indy 500 race car.

To action-oriented guys like Bissell and Kelly, President Eisenhower often moved too cautiously. In pique, they referred to him as “Speedy Gonzales,” while being forced to cool their heels for weeks or months awaiting Oval Office decisions, whether for approving a particular U-2 mission over Russia or signing off on a new spy plane project. Kelly’s airplane was bound to cost millions, and would be a tough sell. The president was already spending a billion dollars in covert funds on the Agena rocket that would boost our first spy satellite into orbit. Bissell was in charge of that program, too, and the first twelve test firings had all been failures. Lockheed’s Missiles and Space Company in Sunnyvale, California, had that contract, and Bissell asked Kelly to evaluate and reorganize their operation. Kelly set up a mini Skunk Works and, coincidentally or not, the thirteenth test shot was a success. But spy satellites had distinct limitations: their pictures in those days were not very sharp, and their orbits were fixed, so the Russians would learn to hide secrets before each scheduled overflight. By contrast, a spy plane operated on no fixed schedules, could loiter in areas of interest, and could overfly tension spots within hours. Our photography was vastly superior to a satellite’s.

Ike tremendously valued the U-2 photo takes but continually worried about the consequences of a shoot-down. He was attracted to the satellite alternative because he felt it was a less aggressive and threatening way to obtain overhead intelligence. Nations would learn to live in the age of satellites, but a spy plane flight would always be regarded as a provocative and aggressive violation of a country’s territory. So Bissell wisely decided to seek the backing of Ike’s two most influential technology advisers — Dr. James Killian, the president of the Massachusetts Institute of Technology, and Dr. Edwin Land of Polaroid, who chaired the presidential advisory panel on aerial espionage and was the godfather of the U-2 program. In May 1958, Kelly flew to Cambridge to meet with Dr. Land and his associates. At that first meeting he was amazed to learn that the Navy had its own Rube Goldberg blueprint for a high-flying spy plane. Theirs would be a ramjet, lifted high into the stratosphere by a balloon. At 100,000 feet, the pilot would release the balloon, light booster rockets to get his ramjet started, then roar up to 155,000 feet. A Navy commander presented this unique idea to the panel while Kelly sat scribbling figures on a pad. “By my calculations,” Kelly told the group, “in order to lift that ramjet, the Navy’s balloon would have to be over one mile in diameter. Gentlemen, that’s one hell of a lot of hot air.”

A more serious proposal came from Convair, which had also been solicited by Bissell for ideas on a high-flying, high-speed spy plane. They had built the B-58 Hustler bomber, a highly regarded Mach 2 tactical strike airplane, which they presented as a “mothership” that would launch a piloted rocket plane that supposedly could reach 125,000 feet at Mach 4. The piggyback launch concept interested Land, but as months passed and the idea was further refined and tested, it became increasingly obvious that the B-58 could not go supersonic while carrying a smaller bird under its belly. Kelly was also skeptical about whether Convair’s plan could produce a reliable photo platform, and he wasn’t shy in passing along his doubts to his CIA friends.

Over the next year Kelly shuttled back and forth to Washington, meeting with Bissell, Land, and other panel members, offering them our latest designs and radar test data, often returning dejected by rumors that Convair’s proposals promised better performance and radar-cross-section data than ours, even though our first preliminary design drawing looked terrific. It was designated A-1, and showed a striking single-seat, two-engine airplane — a long, sleek, bullet-shaped fuselage with rounded inlets on big engines mounted on the tip of small delta wings that were two-thirds of the way back on the fuselage. One look and even a schoolboy would realize that this bird was designed for blazing speed. But the president was less interested in performance and more intent on pushing for the lowest radar cross section possible. It wasn’t that he just didn’t want to get us shot down — he didn’t want the Russians to know we were even up there.

Kelly argued with Washington that our tremendous height and speed advantages were the most potent factors in making us difficult to detect, but the White House and the CIA were not mollified. So we decided to apply radar-absorbing ferrites and plastics to all the airplane’s leading edges — a first in military aviation. We kept the twin tails as small as possible and decided to try to construct them entirely with radar-absorbing composites — a significant technological breakthrough if we could actually do it. But “hiding” this airplane seemed impossible. The tremendous heat generated in supersonic flight made infrared detection inevitable. How do you hide a meteor? Our Mach 3 airplane would streak across the sky like a flaming arrow.

About six months into the design phase I could see discouragement clouding Kelly’s big round face. Our design was now numbered A-10 and we still were not achieving lower radar-cross-section results than Convair, according to Dick Bissell. So, in late March 1959, we began a series of almost around-the-clock brainstorming sessions to review all our previous work and to somehow find a design that would elude Soviet radar. But it seemed fruitless, and Kelly invited Bissell and a couple of agency radar experts to Burbank for what was to be a showdown “where we stand” meeting. He asked me and two others from the design team to sit in and lend him moral support.

The meeting was tense and somber; Kelly was typically candid. “We’ve put in six months of intensive design and study, and by God, we know what we’re doing, but we will never get to the point where the president will be happy with the results. I’m convinced that current improvements in Russian radar will allow them to detect any airplane built in the next three to five years. Radar technology is far ahead of antiradar technology, and we’re just going to have to live with that fact. We’ll never achieve the zero degree of visibility the president seems so stuck on. That technology is way beyond what we know how to do at this point. Maybe Convair can deliver it for you. But we can’t.”

Not much more was said. And when the CIA officials left, Kelly said to us, “Well, boys, I think we’re out. Ike wants an airplane from Mandrake the Magician.”

But later he took me aside. “Keep after this, Ben. Maybe Land or someone else will get Ike to see the light.”

We kept working mostly because it was an unusually slack period at the Skunk Works, without too many other competing distractions. By design A-11, in May 1959, we felt we had scored a breakthrough in dramatically lowering the radar cross section of the aircraft. One of the structural designers presented the idea of modifying the bullet-shaped fuselage by adding a chine, a lateral downward sloped surface that gave the fuselage an almost cobralike appearance. Now the underbelly of the airplane was flat, and the radar cross section had magically decreased by an incredible 90 percent.

By July, we decided to lay out a final revised drawing of the entire airplane making full use of the new chine configuration. In those days I shared an office with four others working on the new airplane — aerodynamicist Dick Fuller, two others who did performance and stability control, and my own sidekick in propulsion, a brilliant twenty-four-year-old Caltech grad named David Campbell, an aerothermodynamicist. (Dave was destined for true greatness, but only two years later, during his daily two-mile jog, he dropped dead from a massive coronary; he was only twenty-six years old.)

I was separated by a connecting doorway from the office of four structures guys, who configured the strength, loads, and weight of the airplane from preliminary design sketches. They put skin and muscle onto the original design concept.

After lunch one blazing summer afternoon, the aerodynamics group in my office began talking through the open door to the structures bunch about calculations on the center of pressures on the fuselage, when suddenly I got the idea of unhinging the door between us, laying the door between a couple of desks, tacking onto it a long sheet of paper, and having all of us join in designing the optimum final design to make full use of the chines. My object was simple. I said, “We’re never going to get this design a hundred percent right. We could play around forever. But I think we now know enough to nail it down at eighty percent. And that’s plenty good enough.”

One of the participants later wisecracked that it was like the Russian and American soldiers joining up on the banks of the Elbe River during the last days of World War II. It took us a day and a half; Ed Baldwin did the basic design and Ed Martin the systems. Henry Combs and Ray McHenry did the structures. Merv Heal figured the weights and Lorne Cass the loads. Dan Zuck designed the cockpit, and Dave Robertson handled the fuel system requirements. Dave Campbell and I weighed in with propulsion and Dick Fuller and Dick Cantrell with the aerodynamics. Everyone chipped in with changes and modifications from previous designs. The airplane weighed 96,000 pounds without fuel — keeping it light to maximize fuel consumption and minimize cost — and was 108 feet long with an extremely thin double delta wing attached at mid-fuselage. The wing edge was designed so razor-thin that it could actually cut a mechanic’s hand. We took the long sheet of paper to Kelly Johnson and unrolled it on his desk. We told him, “Kelly, everything is now exactly where it should be — the engines, the inlets, the twin tails. This is probably as close to the best we can come up with.”

It was our twelfth design, number A-12, which would later become its official CIA project designation. Kelly took the design and ran with it to Washington. Throughout midsummer 1959, he shuttled back and forth to CIA headquarters in Langley, Virginia, nearly a dozen times, meeting with Bissell, Land, Allen Dulles, and others, noting at one point in his private journal, “There is a good deal of concern that Speedy Gonzales [Ike] will cancel. Too expensive.”

But the deal was finally nailed on August 28: “Saw Mr. Bissell alone. He told me that we had the project and that Convair was out of the picture. The agency accepts our conditions that our method of doing business will be identical to that of the U-2. Mr. Bissell agreed very firmly to this latter condition and said that unless it was done this way he wanted nothing to do with the project either. He and Allen Dulles stated following conditions: (1) We must exercise the greatest possible ingenuity, an honest effort in the field of radar. (2) The degree of security on this project is, if possible, even tighter than on the U-2, and (3) We should make no large commitments, large meaning in terms of millions of dollars.”

We were being funded to build five A-12 spy planes over the next two years at a quoted price of $96.6 million.

God help us, we were in business.

The CIA code-named the project Oxcart, an oxymoron to end all: at Mach 3, our spy plane would zip across the skies faster than a high-velocity rifle bullet.

Kelly had sold the idea brilliantly, but now it was up to us peons to deliver the goods. One of the great strengths of the place was the combined experience of Kelly’s most senior and trusted engineers and designers, who, among other attributes, were walking parts catalogs. But suddenly we were all operating in the dark, struggling by trial and error, like Cro-Magnons trying to look beyond the cooking fire to the first steam engine.

All the fundamentals of building a conventional airplane were suddenly obsolete. Even the standard aluminum airframe was now useless. Aluminum lost its strength at 300 degrees F, which for our Mach 3 airplane was barely breaking a sweat. At the nose the heat would be 800 degrees — hotter than a soldering iron—1,200 degrees on the engine cowlings, and 620 degrees on the cockpit windshield, which was hot enough to melt lead. About the only material capable of sustaining that kind of ferocious heat was stainless steel.

For security and convenience, Kelly kept those of us working on his airplane jammed together in one corner of our old Building 82, a remnant of the bomber factory from World War II, in which we had built the U-2 and the F-104 Starfighter before it. From the original four he had approached on this project, we had now grown to a modest fifty or so, seated at back-to-back desks, where, like the early U-2 days, privacy surrendered to incessant kibitzing, teasing, brainstorming, and harassment. Some wag hung the sign PRIVACY SUCKS. My three-man thermodynamics and propulsion group now shared space with the performance and stability control people. Through a connecting door was the eight-man structures group, who designed the strength and load characteristics of the airplane. Their “dean” was that irascible genius Henry Combs, who had been with Kelly since World War II and helped him build the classic two-engine P-38 Lightning interceptor. Henry and I could have reached through the doorway and shaken hands. And, of course, he relished offering his unsolicited advice and opinions to the young whippersnapper running thermodynamics and propulsion.

“Ben Rich,” he teased, “how in hell do you propose getting our stainless-steel monstrosity up to speed at Mach 3? You’ll need inlets the size of the Holland Tunnel.” Then he chuckled sardonically.

Of course, Henry was no happier with the prospect of building a steel airplane than I was. More weight meant more internal support structure, more fuel, less range, and less altitude. Back in 1951 he had recommended to Kelly that we use a rare alloy called titanium for the white-hot exhaust nozzles on the afterburner of the supersonic F-104 Starfighter. So Henry Combs now was mulling the pluses and minuses of building the world’s first titanium airplane. It would be a huge risk. On the positive side, titanium was as strong as stainless steel but only half its weight and could withstand blast-furnace heat and tremendous pressures. Titanium’s tensile strength would allow us to make our wings and fuselage paper-thin. But to build a high-performance aircraft out of such an unproven exotic material was inviting potential disaster. “Unpredictability is a guarantee that we’ll be in the soup on this one from start to finish,” Henry predicted dourly. I knew he was right. Meanwhile, Kelly was already pondering the titanium idea himself. “Any material that can cut our gross weight nearly in half is damned tempting,” Kelly told Combs, “even if it will drive us nuts in the bargain.”

Only one small U.S. company milled titanium, but sold it in sheets of wildly varying quality. We had no idea how to extrude it, push it through into various shapes, or weld or rivet or drill it. Drilling bits used for aluminum simply broke into pieces trying to pierce titanium’s unyielding hide. This exotic alloy would undoubtedly break our tools as well as our spirits. At one of our daily seven a.m. planning sessions in Kelly’s office, I volunteered some unsolicited advice about how we could use a softer titanium that began to lose its strength at 550 degrees. My idea was to paint the airplane black. From my college days I remembered that a good heat absorber was also a good heat emitter and would actually radiate away more heat than it would absorb through friction. I calculated that black paint would lower the wing temperatures 35 degrees by radiation. But Kelly snorted impatiently and shook his head. “Goddam it, Rich, you’re asking me to add weight — at least a hundred pounds of black paint — when I’m desperately struggling to lose even an extra ounce. The weight of your black paint will cost me about eighty pounds of fuel.” I said, “But, Kelly, think of how much easier it will be to build the airplane using a softer titanium, which we can do if we lower the heat friction temperatures on the surface. Adding a hundred pounds is nothing compared to that.”

“Well, I’m not betting this airplane on any damned textbook theories you’ve dredged up. Unless I got bad wax buildup, I’m only hearing you suggest a way to add weight.”

Overnight, however, he apparently had second thoughts, or did some textbook reading on his own, and at the next meeting he turned to me as the first order of business. “On the black paint,” he said, “you were right about the advantages and I was wrong.” He handed me a quarter. It was a rare win. So Kelly approved my idea of painting the airplane black, and by the time our first prototype rolled out the airplane became known as the Blackbird.

Our supplier, Titanium Metals Corporation, had only limited reserves of the precious alloy, so the CIA conducted a worldwide search and, using third parties and dummy companies, managed to unobtrusively purchase the base metal from one of the world’s leading exporters — the Soviet Union. The Russians never had an inkling of how they were actually contributing to the creation of the airplane being rushed into construction to spy on their homeland.

Even before the first titanium shipment arrived, many of us were already worrying that building this particular airplane might just prove too difficult, even for the Skunk Works. Wind tunnel tests of our mock-up amazed us all by indicating that, at Mach 3, intense friction heating on the fuselage would actually stretch the entire airframe a couple of inches! The structures people struggled like medieval alchemists to find rare and exotic metals that could withstand such blowtorch temperatures. They recommended that the hydraulic lines be of stainless steel; for the ejector flaps they found a special alloy called Hastelloy X; and they recommended making our control cables out of Elgiloy, the material used in watch springs. Plumbing lines would be gold-plated since gold retains its conductivity at high temperatures better than silver or copper. Kelly just fumed watching our materials costs rocket into the stratosphere.

There was simply no way to cut any corners. We discovered that there was no off-the-shelf, readily available electronics — none of the standard wires, plugs, and transducers commonly used by the aviation industry could function at our extreme temperatures. There were no hydraulics or pumps, oils or greases that could take our kind of heat. There were no escape parachutes, drag chutes, rocket-eject propellants, or other safety equipment that could withstand our temperature ranges, and no engine fuel available for safe operation at such high temperatures. There was no obvious way to avoid camera lens distortions from fuselage heat flows, and no existing pilot life-support systems that could cope with such a hostile, dangerous environment. We would even be forced to manufacture our own titanium screws and rivets. By the time the project ended, we had manufactured on our own thirteen million separate parts.

Cannibalization had been a house specialty at the Skunk Works on every airplane we had ever built before this one. To save cost and avoid delays, whenever possible we would use engines, avionics, and flight controls from other aircraft and cleverly modify them to fit ours. But now we would even have to reinvent the wheel — literally. Our fear was that the rubber tires and folded landing gears might explode as the heat built in flight. We took our problem to B. F. Goodrich, which developed a special rubber mixed with aluminum particles that gave our wheels a distinctive silver color and provided radiant cooling. The wheels were filled with nitrogen, which was less explosive than air.

The airplane was essentially a flying fuel tank carrying 85,000 pounds of fuel — more than 13,000 gallons — in five noninsulated fuselage and wing tanks that would heat up during supersonic flight to about 350 degrees; we turned to Shell to develop a special, safe, high-flash-point fuel that would not vaporize or blow up under tremendous heat and pressure. A lighted match dropped on a spill would not set it ablaze. The fuel remained stable at enormous temperature ranges: the minus 90 degrees experienced when a KC-135 tanker pumped fuel into the Blackbird at 35,000 feet, and the 350 degrees by the time the fuel fed the engines. As an added safety precaution, nitrogen was added to the fuel tanks to pressurize them and prevent an explosive vapor ignition.

The fuel acted as an internal coolant. All the heat built up inside the aircraft was transferred to the fuel by heat exchangers. We designed a smart valve — a special valve that could sense temperature changes — to supply only the hottest fuel to the engines and keep the cooler fuel to cool the retracted landing gear and the avionics.

One day Kelly Johnson came to me looking as happy as a little kid who had just received a free World Series ticket. “I found a guy in Texas who claims to have developed a special oil product that can withstand nine hundred degrees,” he said. “He’s sending a sample overnight.”

Poor Kelly. A big canvas sack of crystal powder arrived the next day. The powder changed into a lubricant at 900 degrees. Oiling our engines with a blowtorch just wouldn’t make it for us, so we turned to Penn State’s excellent petroleum research department to develop a special oil, which they eventually did, but at a price that made it imperative that not one drop be wasted. A quart of our oil was more expensive than the best scotch malt whiskey. We use 10–40 motor oil in our cars when wide temperature ranges are anticipated; our oil was more like 10–400.

Slowly, but expensively, we began to problem-solve. Kelly offered a hundred-dollar reward for any idea that saved us ten pounds of weight. No one collected. He offered five hundred bucks to anyone who could come up with an effective high-temperature fuel-tank sealant. No one collected that dough either, and our airplane would sit on the tarmac leaking fuel from every pore. But fortunately the tanks sealed themselves in flight from the heat generated by supersonic speeds.

Our crown of thorns was designing and building the powerful engine’s inlets — the key to the engine’s thrust and its ability to reach blistering speeds. This became the single most complex and vexing engineering problem of the entire project. Our engines needed tremendous volumes of air at very high pressures to be efficient, so Dave Campbell and I invented movable cones that controlled the velocity and pressure of the air as it entered the engines. These spike-shaped cones acted as an air throttle and actually produced 70 percent of the airplane’s total thrust. Getting those cones to function properly took about twenty of the best years off my life.

I had a staff of three (by Skunk Works standards that was almost an empire). On the air-conditioning team, I had two engineers to help design the internal cooling system to safeguard the camera bays and the avionics and landing gear systems. The cockpit environment also presented a unique problem: without effective and fail-safe cooling the pilot could bake a cake in his lap. And as head thermodynamicist, that problem fell in my lap.

We designed the cockpit air-conditioning to bleed air off the engine compressor and dump it through a fuel air cooler, then through an expansion turbine, into the cabin at a frigid minus 40 degrees F, which lowered the ovenlike 200-degree cockpit to a balmy Southern California beach day. Developing these systems took us a year of frustrating trial and error.

Our engines were the only items off the shelf, so to speak. Kelly agreed with me that if we started from scratch to invent our engines, we would be hopelessly late in delivering the first Blackbird. We chose two Pratt & Whitney J-58 afterburning bypass turbojets, designed in 1956 for a Navy Mach 2 fighter-interceptor that had been canceled before the start of production. But the engine, which would need major modifications for our purposes, had already undergone about seven hundred hours of testing before the government cut off its funding. Each of these engines was Godzilla, producing the total output of the Queen Mary’s four huge turbines, which churned out 160,000 shaft horsepower. Using afterburners at Mach 3, the exhaust-gas temperatures would reach an incredible 3,400 degrees.

This propulsion system would not only be the most powerful air-breathing engine ever devised but also the first ever to fly continuously on its afterburners, using about eight thousand gallons of fuel an hour. To build this system to our needs and specifications, P & W’s chief designer, Bill Brown, who had worked closely with us on the U-2, agreed to construct a separate plant at their Florida manufacturing complex exclusively for developing this extraordinary engine. The CIA unhappily swallowed the enormous development costs of $600 million. Brown preached teamwork and pledged an unprecedented degree of partnership with the Skunk Works in general, and with me and my team in particular, to design their compressor to match my airflow inlet. This close partnership between the engine builder and the airplane manufacturer was unusual in an industry where the engine people and the airplane manufacturers often used each other as scapegoats if an airplane failed to live up to its potential. Abandoning this kind of adversarial posturing led to achieving the most powerful engine system coupled to the highest-performance inlets at these high Mach numbers that has ever been attained.

Bill Brown also offered us access to one of the largest and costliest computer systems of the day, the IBM 710. The system was state-of-the-art for its time and about as sophisticated as today’s commonly used handheld calculators. But, like us, the Pratt & Whitney team would problem-solve mostly by what Kelly jokingly referred to as “my Michigan computer”—the battered old slide rule he had been using since his university days at Michigan.

Despite the unprecedented power of those two massive engines, they supplied only 25 percent of the Blackbird’s thrust at Mach 3, a fact Bill Brown hated to admit. The inlets produced most of the propulsive thrust by supplying the air required by the engine at the highest pressure recovery and with the lowest drag. At supersonic cruising speed, each of our two inlets swallowed 100,000 cubic feet of air per second — the equivalent of two million people inhaling in unison. Hydrocarbon fuels like kerosene burn at high pressure, but at 80,000 feet, the air density is only one-sixteenth the density at sea level, so we used the inlets to pump up compression, before burning the air-fuel mixture inside the engine and then expanding it through a turbine and finally refiring it with tremendous thrust through the afterburner.

The only way to get energy out of the air is to pump pressure into it or to burn it. Our unique movable inlet cone, shaped like a spike, acted as an air throttle by regulating the airflow into the inlet across the spectrum of speeds from takeoff to climb to maximum cruise speed. Operated by our revolutionary electronic measuring sensors, which recorded speed and angle of attack to position the spikes precisely, the movable spikes were fully extended about eight feet out from the inlets on takeoff and gradually retracted by as much as two feet into the inlet interior as the airplane gained maximum supersonic speed.

At 80,000 feet, the outside air temperature was about minus 65 degrees F. As the inlet sucked in the air at Mach 3 through narrowed openings that compressed it, the air heated to 800 degrees. The bypass turbojet engines took the heated and high-pressure air (40 psi) and squeezed it further in a compressor, heating it to about 1,400 degrees F. At that point fuel was added to heat the air inside the burner to 2,300 degrees F. This supercharged air was then expanded through the turbine, before being fed into the roaring afterburners, superheating the combustible mix of gas and air to 3,400 degrees F, just 200 degrees below the maximum temperature for burning hydrocarbon fuels. The white-hot steel nozzle spit out its fiery plume in the form of diamond-shaped supersonic shock waves. Even in the frigid upper atmosphere, the air boiled at 200 degrees F for a thousand yards behind those booming engines. This unprecedented propulsive power sped the Blackbird at an unbelievable two-thirds of a mile a second.

About six months into our wind tunnel testing, I went to Kelly with joyful results: the inlets produced 64 percent of the airplane’s full-throttled power. The precise shaping of the inlets and our unique movable air throttle cones, or spikes, allowed us to achieve an astounding 84 percent propulsion efficiency at Mach 3, which was 20 percent more than that of any other supersonic propulsion system ever built.

Developing this air-inlet control system was the most exhausting, difficult, and nerve-racking work of my professional life. The design phase took more than a year. I borrowed a few people from the main plant, but my little team and I did most of the work. In fact the entire Skunk Works design group for the Blackbird totaled seventy-five, which was amazing. Nowadays, there would be more than twice that number just pushing papers around on any typical aerospace project.

Having today’s high-speed computers would have accelerated the design process and simplified much of our testing, but perfection was seldom a Skunk Works goal. If we were off in our calculations by a pound or a degree, it didn’t particularly concern us. We aimed to achieve a Chevrolet’s functional reliability rather than a Mercedes’s supposed perfection. Eighty percent efficiency would get the job done, so why strain resources and bust deadlines to achieve that extra 20 percent, which would cost as much as 50 percent more in overtime and delays and have little real impact on the overall performance of the aircraft itself?

As it happened, we achieved 70 percent efficiency within the first half year of our work, but to tweak it above that to our target of 80 percent took an additional fourteen months. Of primary concern was where to precisely locate the supersonic shock wave within the inlet walls. That was the key to achieving maximum efficiency, because the shock wave in the wrong place in the inlet would block incoming air, causing energy loss, drag, and in a worst-case scenario — stall.

I logged hundreds of hours testing inlet shapes and cone models at NASA’s Ames Research Center at Moffett Field in Northern California, a giant complex of high-speed wind tunnels. That became my second home, where I spent weeks at a time using their largest, most-powerful supersonic wind tunnel, a twenty-foot-long, ten-foot by ten-foot rectangular chamber powered by a gigantic compressor capable of driving an ocean liner, and a three-story cooling tower holding tens of thousands of gallons of water. Running Mach 3 pressures for several hours at a time drained so much electricity needed by local industry that we were forced to test only late at night, working usually until dawn. Wind tunnel tests cost us $10,000 to $15,000 an hour and we ran up a stupendous bill because our models were tested from every angle and on more than 250,000 separate measuring points, across a broad range of Mach numbers and pressures.

But Kelly preached that a precise model, even one like ours that was one-eighth the size of the real inlet, would provide precise measurements for the full-size model as well. So our wind tunnel testing was critical to the airplane’s success and usually ended at sunrise, when our exhausted little group of analysts finished computing the previous night’s test results. Nowadays, such calculations can be performed in a mini-second by supercomputers.

Kelly was now so desperate to save weight that he upped the ante to one hundred and fifty bucks to anyone who could save him a measly ten pounds. I suggested we inflate the Blackbird’s tires with helium and give each pilot a preflight enema. Kelly tried the helium idea, but helium bled right through the tires. The enema idea he left to me to try to promote among the pilots.

One bet easy to collect was that we would never have this airplane flying on time. By the end of 1960, we were over budget by 30 percent and Kelly was forced to concede we would be at least a year late in getting the Blackbird into the sky.

The biggest delay was in my bailiwick. We had contracted the inlet control mechanism that would move and position the cones to Hamilton Standard, which was also doing the fuel control system for Pratt & Whitney. The trouble was that the pneumatic inlet controls they devised were not responding quickly enough. We had spent $18 million to develop this system, but after more than a year the problem was unsolved. Finally, I took the matter to Kelly. I said, “Kelly, I think we’ve got to cut our losses and find someone else to get the job done.”

He cussed and agreed. We went to a company called Air Research and they developed an electronic control that saved our bacon. And they did it in less than a year. Meanwhile Pratt & Whitney was struggling with a slew of problems that were putting them further and further behind schedule. Dick Bissell and his assistant, John Parangosky, watched in anguish as our delays and costs mounted. In pique, Parangosky had begun referring to the P & W engine as the “Macy’s engine,” and complained to Bill Brown, their program manager, rather unfairly, “If we gave as much money to R. H. Macy’s, they could build that engine in time for Christmas.” By mid-1960, the agency decided to crack a mean whip. Kelly was called on the carpet and told he had to accept a CIA engineer bird-dogging in his shop and looking over all shoulders. Kelly blew up. He knew the agency was only trying to cover their own asses. But Parangosky, whom Kelly liked and trusted, flew in from Washington and warned him that if he refused the request there was a grave likelihood that the CIA would cancel the contract entirely.

Kelly fumed. “No, John! I’m not gonna have one of your spies poking into my business. Bissell promised me you guys would keep hands off and let me do this thing my way just like the U-2.”

“Kelly, be reasonable. We won’t get in your way. We just want someone here you can trust and we can too.”

He suggested a very bright and able engineer on the CIA’s payroll named Norm Nelson, whom Kelly had known from World War II days, and both liked and respected. Kelly sulked but ultimately surrendered. “Well,” he said, “I’ll let in Norm Nelson, but not another goddam person. You got that? Besides, you tell Nelson he can have a desk, a phone, but no chair. I expect him in the shop, not sitting on his fat duff.”

Nelson, who arrived in the spring of 1960, became the first outsider ever allowed a place inside Kelly’s realm. He gave Norm a free hand and actually took suggestions because he respected Norm’s judgment. We all knew that he reported directly to Bissell, and he knew that we kept him out of certain meetings and didn’t let him in on everything. But Norm was sharp. I recall one meeting in 1961, when Kelly told Norm that the agency had given us an additional $20 million to develop wing tanks on the Blackbird to extend its range. Norm did some quick calculations and figured out we would extend the airplane’s range only eighty miles. “Fooling around with wing tanks at this point will be more trouble than it’s worth,” Norm insisted. Kelly said, “You’re probably right.” Kelly Johnson sent back the $20 million that afternoon.

But our biggest problem was about as easy to conceal from Norm as a pregnant pachyderm on top of a flagpole. Norm Nelson came aboard just as we were starting to build a mock-up of the fuselage-cockpit section, which would contain more than six thousand parts, for heat testing inside an oven. To our horror, we discovered that the titanium we were trying to use was as brittle as glass. When one of the workers dropped a piece off his bench, it shattered in a dozen pieces. The trouble was diagnosed as poor quality-control procedures in the manufacturer’s heat-treatment process — a problem that caused endless delays, forcing us to reject 95 percent of the titanium delivered and set up a rigorous quality-control procedure.

For an outfit that detested red tape, we now found ourselves wallowing in bureaucratic procedures. We sample-tested for brittleness three out of every ten batches of titanium received and kept detailed records of millions of individual titanium parts. We could trace each part back to the original mill pour, so if a part went bad later on, we could immediately replace other parts from that same batch before trouble developed.

We also learned the hard way that titanium was totally incompatible with many other elements, including chlorine, fluorine, and cadmium. When one of our engineers drew a line on a sheet of titanium with a Pentel pen, he discovered that the chlorine-based ink etched through the titanium just like acid. Our mechanics working on the engine installation used cadmium-plated wrenches to tighten bolt heads. When the bolts became hot, the bolt heads just dropped off! It took intensive detective work to zero in on the cadmium contamination culprit, and we quickly removed all cadmium-plated tools from toolboxes. Even the routine matter of drilling a hole became a nagging frustration. When machining standard aluminum, a hundred holes could be drilled without resharpening the bit. With titanium, we had to resharpen every few minutes and were forced to develop special drills using special cutting angles and special lubricants, until we finally were able to drill more than 120 holes before having to resharpen. But it took us months of painstaking experimentation to get that far.

Miles of extrusions were required to produce an aircraft the size of the Blackbird, so it was necessary to invest time and more than a million dollars in new state-of-the-art precision drills, cutting machinery, powerheads, and lubricants.

For each problem solved, two or three others suddenly cropped up. We were stunned when spot welds on panels began to fail within six or seven weeks. Some intensive sleuthing revealed that the panels had been welded during July and August, when the Burbank water system was heavily chlorinated to prevent algae growth. The panels had to be washed after acid treatment, so we immediately began using only distilled water. During heat tests, the wing panels warped so badly they looked like potato chips. We worked for months to find a solution and finally used corrugated panels that allowed the metal to expand without warping under immense heat friction. At one point an exasperated Kelly Johnson told me: “This goddam titanium is causing premature aging. I’m not talking about on parts. I’m talking about on me.”

We set up training classes for machinists using titanium for the first time and a research operation for developing special tools that would make their jobs easier. Between the new machines and the training, our bean counters figured that ultimately we saved $19 million on the production program.

Still, unforeseen problems kept increasing the costs, and only a few days after the November 1960 elections, which brought John F. Kennedy into office and took the Republicans out, Kelly returned from a week’s vacation and found a wire awaiting him from Dick Bissell, inquiring about what it would cost the government to cancel the Blackbird program. “I am very afraid,” Kelly noted in his private log, “about what will be Kennedy’s attitude toward the program, its overall cost increase, which is very high on all fronts, and the fact that our Russian friends have now come up with a new Tall King radar which appears to be capable of detecting a target about one-third the size that we are able to accomplish with the Blackbird. With all this we have made remarkable strides in reducing the radar cross section, and our experts say we would have about one chance in 100 of being detected, with practically no chance of being tracked.”

Our chief chemist, Mel George, helped us to develop special antiradar coatings loaded with iron ferrites and laced with asbestos (long before it became a dirty word) to be able to withstand the searing heat from the tremendous friction hitting the leading edges of the airplane. These coatings were effective in lowering the radar cross section and comprised about 18 percent of the airplane’s materials. In effect, the Blackbird became the first stealth airplane; its radar cross section was significantly lower than the numbers the B-1B bomber was able to achieve more than twenty-five years later.

To save time and money and maintain high quality standards, we did our own milling and forging and at one time approached the ability of our vendor’s plants to roll parts to precise dimensions. We even developed our own cutting fluid that would not corrode titanium. To prevent oxidation of the titanium — which caused brittleness — we welded in specially constructed chambers with an inert nitrogen gas environment. In all we had about twenty-four hundred trained fabricators, machinists, and mechanics working on the project, all of them specially trained and carefully supervised. And at the height of production, in the mid-1960s, we employed a huge force of nearly eight thousand workers and delivered one Blackbird per month. While we were trying to build that first airplane, the unions were giving Kelly fits because he ignored seniority rules and chose the best workers, so Kelly had the union heads cleared and walked them through the plant and showed them the airplane. He said, “Gents, this airplane is vital for our nation’s security. The president of the United States is counting on it. Please don’t get in my way here.” They backed off.

For security and other reasons, the airplane was assembled in various buildings in the complex. One unique, extremely time-saving technique was to build the fuselage on the half shell. The left half and right half were assembled independently to create easier worker access, then fit together and riveted into place. That was a major first in aircraft manufacturing.