The Idea Factory: Bell Labs and the Great Age of American Innovation

by Jon Gertner

  The meeting didn’t diminish Pierce’s desire to write, however. In the 1930s and 1940s, Karl Darrow, a gifted writer and explainer employed by Bell Labs, had translated in various journals the new ideas of quantum mechanics for a younger generation of physicists. In many respects, Pierce picked up in the 1950s where Darrow left off, but instead of technical journals his venue was books and magazines for a general audience. Often Pierce’s books on technology weren’t nearly as accessible as he imagined; frequently they were disorganized and dense with mathematics. Nevertheless, in addition to his book on traveling wave tubes, he wrote general-interest books, some of them fairly successful, about cathode-ray tubes, electromagnetic waves, acoustics, communications infrastructure, information theory, and transistors. He would say that instead of naming the transistor, he wished he had actually invented it. He also remarked that rather than writing about information theory, he wished that he, rather than Shannon, had thought of it.

  He thought of something else, though. In 1952, Pierce wrote a nonfiction essay for Astounding Science Fiction entitled “Don’t Write: Telegraph.” The piece explored the idea of sending messages to and from the moon. Pierce later noted, “What struck me then was how much easier it would be to communicate between the moon and the earth than across the United States—if only one could put the apparatus in place.” Microwaves travel vast distances in straight lines, which is ideal in space. Here on earth, however, they can’t follow the curvature of the planet. To send microwave signals across the United States, therefore, as the phone company was now doing with its nationwide system of relay towers, required stations every thirty miles or so. “The truth,” Pierce wrote, “is that you could order equipment for an Earth-Moon link from any of several manufacturers.”

  Slowly, the idea evolved in his mind. In October 1954, he was invited to give a talk about space in Princeton at a convention of the Institute of Radio Engineers. Pierce decided he would discuss an idea he had for communications satellites—that is, orbiting unmanned spaceships that could relay communications (radio, telephone, television, or the like) from one great distance to another. A terrestrial signal could be directed toward the orbiting satellite in space; the satellite, much like a mirror, could in turn direct the signal to another part of the globe. Pierce didn’t consider himself the inventor of this idea; it was, he would later say, “in the air.” In fact, unbeknownst to Pierce, Arthur Clarke had written an obscure paper about ten years before suggesting that a small number of satellites, orbiting the earth at a height of about 22,300 miles, could connect the continents. Clarke never developed the idea any further and quickly lost interest in it. “There seemed nothing more that could be said until technical developments had validated (or invalidated) the basic concept,” he later wrote. In Pierce’s talk, however, he made some detailed calculations about satellites. He concluded that orbiting relays might not be financially viable over land; in the United States, the Bell System already had an intricate system of coaxial cables and microwave links. The oceans were a different story. The new cable that Bell Labs was planning for the Atlantic crossing in 1954 would carry only thirty-six telephone channels at tremendous expense and tremendous risk of mechanical failure. A satellite could satisfy the need for more connections without laying more cable.

  One academic in the audience that day in Princeton suggested to Pierce that he publish his talk, which he soon did in the journal Jet Propulsion. “But what could be done about satellite communications in a practical way?” Pierce wondered. “At the time, nothing.” He questioned whether he had fallen into a trap of speculation, something a self-styled pragmatist like Pierce despised. There were no satellites yet of any kind, and there were apparently no rockets capable of launching such devices. It was doubtful, moreover, whether the proper technology even existed yet to operate a useful communications satellite. As Pierce often observed ruefully, “We do what we can, not what we think we should or what we want to do.”

  Thirteen

  ON CRAWFORD HILL

  Ideas may come to us out of order in point of time,” the first director of the Rockefeller Institute for Medical Research, Simon Flexner, once remarked. “We may discover a detail of the façade before we know too much about the foundation. But in the end all knowledge has its place.”1 Flexner was speaking of the biological sciences; he had seen how the individual fruits of research might have little use—and little clarity—but could accrue over time to create a grand idea. The same might be said about any branch of the sciences, or about many of the large projects in the planning stages at Bell Labs. The transatlantic cable, for instance, which had been on the drawing boards for several decades until a variety of developments made it technologically feasible as well as cost-effective, was a good example. For communications satellites, the main idea had come too early. But was Pierce decades early, or just a few years?

  Already, by the time he had given his talk in Princeton, some of the elements needed to create a satellite were available. The transistor—rugged and miserly in its power requirements—was in wide production. It could undoubtedly be useful in an orbiting satellite, Pierce realized. The traveling wave tube, now moving out of its development stage, could likewise be valuable, since it could amplify a multitude of telephone or television channels simultaneously. Much of the challenge in creating satellite communications lay not in the satellites themselves but in building an adequate system for transmitting and receiving signals from the ground, as well as a system for tracking the satellite as it moved across the sky. In this regard, a third existing technology appeared vital. It was the horn antenna, which had been designed by Bell Labs’ Harald Friis at the rural Holmdel lab in southern New Jersey. Horn antennas were already a crucial component in microwave towers across the country: They allowed for the reception of signals in a focused manner that greatly reduced surrounding noise and interference. There was no reason to think they couldn’t be adapted for satellite communications.

  By late 1956, about two years after Pierce’s talk at Princeton, some other necessary elements for his idea had sprung into existence. Pierce realized that the batteries in an orbiting satellite would wear out unless the satellite could tap into a dependable and inexhaustible power source. Bell Labs’ silicon solar cell had been celebrated until the silicon strips had proven too expensive and too inefficient for use in rural phone installations. But satellites seemed to be the perfect problem for this solution. Since there was no other way to recharge a satellite’s batteries in space, the expense or inefficiency of solar cells was no hindrance.

  Just as important was another development, which originated mainly from the work of a former colleague of Pierce’s. Charles Townes had come to Bell Labs from Caltech in 1939, just after Pierce; Townes was the curious recruit who had used the Labs’ travel advance that year as a means to fund an adventurous bus tour for himself, taking him south from California through Mexico, and then up again through Texas, North Carolina, and ultimately the Bell Labs offices on West Street. Townes was an experimental physicist. Regarded as brilliant by his colleagues at Bell Labs, he had been reluctant to spend his career in an industrial setting, and had left Bell for a position at Columbia University a few years after World War II ended. In the late 1940s, Townes’s work was concentrated in the field of microwave spectroscopy—analyzing the “spectra” (or energy emissions) of excited molecules as a means of studying their makeup. By the early 1950s, he had moved on to something known as “stimulated emission.” Townes believed that under the right circumstances, and inside a complex apparatus he’d spent years building in his Columbia laboratory, he could (1) take a gas such as ammonia and direct its molecules through a slit to form a narrow stream; (2) move the narrow stream through an electric field; and then (3) have the stream finally enter something called a “resonant chamber.” In the chamber, these ammonia molecules, now in a highly excited state, would become increasingly excited—to the point where they could generate a beam of highly focused energy.2

&nbs
p; Townes faced skepticism for his project, even from his boss at Columbia, the Nobel laureate I. I. Rabi. But then his machine actually started working. When it came time to name the device, he and his colleagues dubbed it the maser, which stood for “microwave amplification by stimulated emission of radiation.” Within a few years, physicists (including Townes himself) would try building similar devices that emitted energy in the form of light rather than microwaves. These devices were often called optical masers, until that name fell out of fashion. Afterward, they were simply known as lasers.

  Townes’s maser became operational in 1954. By 1957, a team at Bell Labs had built a more sophisticated maser that appeared to have some use in communications.3 Its functional part looked like a huge metal test tube, about eight inches in diameter and a couple of feet long; a crystal about the size of a fingernail (rather than ammonia gas) was placed in a cavity at the bottom of the tube that was cooled to about minus 456 degrees Fahrenheit by a bath of liquid helium. The crystal was then bombarded with electrical energy from a power source called a “pump.”4 Pierce hadn’t recognized that the maser might be useful for his satellite project, but one day his old friend Rudi Kompfner, home sick in bed, realized that the maser could be ideal in amplifying the faint signals coming from a satellite in orbit. Its sensitivity and fidelity would surpass any other known device. “As soon as I could get out of bed,” Kompfner said, “I rushed over to tell John, and he agreed right away that the maser would make all the difference.”5

  So now there were transistors, the horn antenna, the traveling wave tube, solar cells, and the maser. Even with the right electronic components, though, communications satellites weren’t going anywhere yet. There was still no proof that aeronautical engineers had developed rockets that could propel the idea into space. Proof arrived dramatically in October 1957 when the Soviet Union launched its Sputnik satellite.6 A year later the United States, fearful of losing ground to the opposing superpower, launched its first satellite, Explorer I. The event was followed soon after by the National Aeronautics and Space Act, which created a new national space agency called NASA. With the technology as well as the money now available, the ideas Pierce had been writing about for the past few years were now plausibly real. When someone asked him for his reaction to the Sputnik launch, Pierce said, “It’s like a writer of detective stories going home and finding a body in his living room.”7

  AS THEY WAITED for the gun to fire, the communications engineers poised at the starting line of the satellite era had two choices before them. Pierce had proposed that they could first explore the frontiers of space by launching either a “passive” satellite or an “active” one. A passive satellite would merely be a reflector of sorts that circled the earth in a low orbit, perhaps a thousand miles up, that engineers could bounce signals to and from. A transmission from California, for instance, could be directed at an angle toward the traveling satellite, which in turn could passively reflect the signal at a different angle back down to earth, perhaps to a receiving station in New Jersey. An active satellite, on the other hand, contained batteries, transistors, antennas, and vacuum tubes, with which it could amplify a signal received from earth before sending it back down. Essentially, it followed the same principles as the microwave relay towers already stationed around the country—except this relay was orbiting in space.

  At least in theory, active satellites were better than passive ones. The signals directed from earth to a passive satellite are reflected out in all directions, for instance, so that those received at any one point on the ground might be so faint—perhaps a millionth of a billionth of the originally transmitted signal—that the task demanded extraordinarily sensitive equipment (huge horn antennas, expensive masers, and the like) for receiving even a single voice transmission. That wouldn’t be the case with an active satellite, which could broadcast a broad band of strong, accessible signals and likewise receive instructions from ground stations on what signals to carry. (An active satellite could also carry television signals far more easily than a passive one.) On the other hand, Pierce considered himself conservative about any satellite gambit. He wasn’t certain that Bell engineers knew enough yet to build a foolproof and durable active satellite—one that could operate for more than a few weeks or months. He also knew that the small research department at Bell Labs, unlike the huge development department, lacked the manpower and budget necessary for an active project. “There’s a difference, you see, in thinking idly about something, and in setting out to do something,” he explained to an interviewer in the early 1960s. “You begin to see what the problems are when you set out to do things, and that’s why we thought [passive] would be a good idea.” A passive satellite, he added, probably wouldn’t be useful in terms of the business of communications. But it tested the possibility of orbiting relays before they were developed into something more. A passive satellite, in other words, was an experiment. Ten years earlier, Pierce had witnessed how problems with the transistor didn’t show up until the device entered the development and production stage. Here, too, was a relatively low-risk opportunity to confront and solve the practical challenges of this new technology—“to get one’s hands dirty”—before making a slew of big and expensive mistakes.8

  In Pierce’s caution one could discern some of the management wisdom of his mentor, Harald Friis, the pipe-smoking, Danish-born chief of Bell Labs’ small Holmdel lab, where most of the Labs’ microwave research was conducted. Friis often asked his young scientists, “Are you sure this isn’t too big a ball of wax?”9 He had a gentle way of helping them see if they could sharpen their vision and perceive the difference between what was doable and what was not. The passive satellite was certainly a big ball of wax, but the active satellite, in Pierce’s view, was too big a ball of wax. Indeed, Pierce soon found out that the military, through their Advanced Research Projects Administration (ARPA), were thinking about building an extremely expensive active satellite called Advent. “The tendency of ARPA has been to project elaborate and complicated schemes,” Pierce wrote derisively in a memo to Jack Morton after visiting with ARPA’s directors.10 Pierce thought the project was too much, too soon. He believed—correctly, it turned out—that Advent was hurtling toward failure.11

  All through 1958, Pierce’s main advocate for building a passive satellite at Bell Labs was his old friend Rudi Kompfner, the traveling wave tube inventor he had first met in England during the war. At the start, the two men had no precise idea of how they would get such a project to happen; nor did they have any sense yet whether Mervin Kelly, then in the last year of his Labs presidency, would agree to put several million dollars behind the idea. But various aspects of the project fell into place with surprising speed. Early that year, for instance, Pierce and Kompfner found out that a government engineer named William O’Sullivan, stationed at Langley Field in Virginia, was proposing various atmospheric tests by putting a huge aluminum-clad Mylar balloon—100 feet in diameter, but weighing only 136 pounds—into orbit about 800 or 1,000 miles above the earth. It struck Pierce as an ideal design for a passive satellite, and when he and Kompfner actually obtained samples of the balloon, they discovered, much to their satisfaction, that it would reflect 98 percent of the radio waves directed at it.

  O’Sullivan couldn’t get anyone to launch his balloon. But in the summer of 1958 Pierce and Kompfner began discussions with William Pickering, the director of the Jet Propulsion Laboratory in California, who was intrigued by the project. As Pierce recalled, Pickering “agreed to supply a west-coast ground station” that would communicate with another ground station in New Jersey “if we could get the balloon launched.”12 The men still didn’t have a way to fire the balloon into space. But when NASA officially took over the Jet Propulsion Laboratory, just a few months later, NASA’s new chief, Keith Glennan, put his support behind the passive satellite launch, too. That problem now seemed solved.

  Still, there had been the challenge of garnering support at Bell Labs during 1958. Bill Baker, a
chemist who had become the head of research, supported Pierce’s satellite idea. So did Baker’s boss, Jim Fisk, now the executive vice president of the Labs. But during the summer, Mervin Kelly, relying partly on the mathematical analysis of Claude Shannon’s old friend and office mate Brock McMillan, told Pierce one day during a lunch meeting that it was a nonstarter. McMillan’s calculations had been woven into a three-page memo on the project, written by one of Kelly’s deputies, that was hostile to both the economics and technology of the satellite project.13 Whether Kelly’s opinions were the starting point for the memo or were merely reinforced by it wasn’t wholly clear. Pierce later would say that Kelly had a negative view of the satellite project from the very start.

  Kelly may have been swayed by arguments within the Labs that satellite communications were too expensive to justify, or he may have worried that Bell Labs’ entrance into the satellite business might upset the delicate state of relations between AT&T and the federal government. AT&T, he believed, should not be in the space business. But all of these concerns may have been magnified by Kelly’s opposition to the kind of innovation that might later be described as “discontinuous.”14 Bell Labs had just completed the successful transatlantic cable; the future of communications to Europe and beyond appeared to reside in new and better cables. These would be incremental innovations. In such a vision of the future, orbiting satellites weren’t only a risky and unproven technology; they were also—at least to a telephone executive with a well-defined, step-by-step ten-year plan for improving the system—a strange sideways leap.