The Idea Factory: Bell Labs and the Great Age of American Innovation
Page 3
AT&T’s ENGINEERS HAD BEEN VEXED by distance from the very beginning. The telephone essentially converted the human voice into an electrical signal; in turn-of-the-century phones this was done by allowing sound waves produced by a voice to vibrate a taut diaphragm—usually a disc made of thin aluminum—that was backed by another thin metal disc. A mild electric current ran between the two discs, which were separated by a chamber filled with the tiny carbon granules Edison had invented. As sound waves from a voice vibrated the top diaphragm, waves of varying pressure moved through the granules below it. The varying pressure would in turn vary the resistance to the electric current running between the metal discs. In the process sound waves would be converted to electric waves. On a simple journey, the electrified voice signal would then travel through a wire, to a switchboard, to another cable, to another switchboard, and finally to a receiver and a distant eardrum. But a telephone voice signal was weak—much weaker and more delicate than a telegraph’s simple dot-dash signal. Even worse, the delicate signal would grow thinner—or “attenuate,” to use the phone company’s preferred term—after even a few miles.
In the telephone’s first few decades, AT&T’s engineers had found that different methods could move a phone call farther and farther. Copper wire worked better than iron wire, and stiff, “hard-drawn” copper wire seemed to work even better. Best of all was extremely thick-gauge hard-drawn copper wire. The engineers likewise discovered that an invention known as “loading coils” inserted on the wires could extend the signal tremendously. Finally, there were “repeaters.” These were mechanical amplifiers that took the sound of a weakening voice and made it louder so the call could travel many miles farther. But you could only install a few repeaters on a line before the advantages of boosting a call’s volume were undone by distortion and the attenuation of the signals. And that left the engineers at a final disconnect. The tricks of their trade might allow them to conquer a distance of about 1,700 miles, roughly from New York to Denver. A great impasse lay beyond.
In 1909, Frank Jewett, now one of the phone company’s senior managers, traveled to San Francisco with his boss, John J. Carty, AT&T’s chief engineer. They found parts of the city still in ruins. As Jewett recalled, “The wreckage of the [1906] earthquake and fire was still only partially cleared away and but the beginnings made on the vast rebuilding operations.”17 The men were there to determine how to repair the local phone system, but they also began discussing the possibility of providing transcontinental phone service—New York to San Francisco—in time for the Panama-Pacific International Exposition of 1914. Theodore Vail, who met Jewett and Carty there, was in favor of making a commitment, since it represented a clear step toward universal service. Carty and Jewett were more circumspect. Together they spent long days and nights debating the problem, usually continuing their discussions far past midnight. The men could see there were enormous, but surmountable, engineering challenges; they would, for instance, need a cable that could be effectively strung across the mountains and desert and survive the weather and stress. But there were also profound challenges of science. “The crux of the problem,” Jewett wrote in describing his conversations with Carty, “was a satisfactory telephone repeater or amplifier”:
Did we know how to develop such a repeater? No. Why not? Science hadn’t yet shown us the way. Did we have any reason to think that she would? Yes. In time? Possibly. What must we do to make “possibly” into “probably” in two years?
And so on night after night without end almost.
Carty and Jewett eventually told Vail they would do it—and the task soon came to be Jewett’s personal responsibility. That was risky on a number of counts. Jewett’s talents were as a manager and social sophisticate; he was quick to apprehend technical problems but not necessarily equipped to solve them. On the other hand, he knew someone who could help.
Jewett returned to the University of Chicago in the fall of 1910 to visit his old friend Millikan, and he started the conversation without small talk. Jewett began, “Mr. John J. Carty, my chief, and the other higher-ups in the Bell System, have decided that by 1914, when the San Francisco Fair is to be held, we must be in position, if possible, to telephone from New York to San Francisco.” To get through to San Francisco by the present methods was out of the question, he explained, but he wondered if perhaps Millikan’s work—he pointed to some complex research on electrons—suggested that a different method might be possible. Then Jewett asked his friend for help. “Let us have one or two, or even three, of the best of the young men who are taking their doctorates with you and are intimately familiar with your field. Let us take them into our laboratory in New York and assign to them the sole task of developing a telephone repeater.”18
Here was a new approach to solving an industrial problem, an approach that looked not to engineers but to scientists. The first person offered this opportunity was Millikan’s lab assistant from the oil-drop experiment, Harvey Fletcher, who declined. Fletcher wanted to return home to Salt Lake City to teach at Brigham Young University. The next person was Harold Arnold, a savvy experimentalist who said yes, and who quickly joined the New York engineering group under Jewett.
Within two years Arnold came up with several possible solutions to the repeater problem, but he mainly went to work on improving an amplifier known as the audion that had been brought to AT&T in 1912 by an independent, Yale-trained inventor named Lee De Forest. The early audion was vaguely magical. It resembled a small incandescent light bulb, yet instead of a hot wire filament strung between two supporting wires it had three elements—a metal filament that would get hot and emit electrons (called a cathode); a metal plate that would stay cool and attract electrons (called an anode); and between them a wire mesh, or “grid.” A small electrical current, or signal, that was applied to the audion’s grid could be greatly amplified by another electrical current that was traveling from the hot cathode to the cool anode. Arnold found, through trial and error, the best materials, as well as a superior way to evacuate the air inside the audion tube. (He suspected correctly that a high vacuum would greatly improve the audion’s efficiency.) Once Arnold had refined the audion, he, Jewett, and Millikan convened in Philadelphia to test it against other potential repeater ideas. The men listened in on phone conversations that were passed through the various repeaters, and they found the audion clearly superior. Soon to be known as the vacuum tube, it and its descendants would revolutionize twentieth-century communications.
The transcontinental line, complete with several new vacuum tube repeaters placed strategically in stations along the route, was finished in time for the Pacific exposition, which had been pushed back to 1915. Harold Arnold had improved the design so that the repeaters looked like spherical bulbs, with the three crucial elements inside, sitting upon a base from which three wires emerged. The continental link itself consisted of four copper wires (two for directing calls in each direction) that were strung coast-to-coast by AT&T linemen over 130,000 wooden poles. As a public relations stunt, Alexander Graham Bell, the inventor who had long since stopped having any day-to-day responsibilities at the company he founded, was stationed in New York to speak with his old assistant, Thomas Watson, in San Francisco.
“Mr. Watson, come here, I want you,” the old man quipped, paraphrasing what he had said to Watson on the day the two discovered the working telephone in Boston nearly forty years before.
“It would take me a week to get there now,” Watson replied.
It was a wry bit of stagecraft. For AT&T, it was also an encouraging sign that Vail’s notion of universal service might indeed be possible—at least for customers who could afford to pay about $21 (about $440 in today’s dollars) for a three-minute call to California.19
For Frank Jewett, meanwhile, the cross-country link proved that his cadre of young scientists could be trusted to achieve things that might at first seem technologically impossible. That led him to redouble his efforts to hire more men like Harold Arnold. Jewett kept
writing to Harvey Fletcher, Millikan’s former graduate student who was now in Salt Lake City, sending him every spring for five consecutive years a polite and persuasive invitation to join AT&T. In 1916, Fletcher finally agreed to leave Brigham Young and come work for Jewett. Millikan, meanwhile, didn’t stop serving as the link between his Chicago graduates and his old friend. In late 1917, responding to an offer from Jewett for $2,100 a year, Mervin Kelly, now done counting oil drops, decided that he would come to New York City, too.
Two
WEST TO EAST
Fletcher and Kelly were joining a company whose size and structure seemed positively bewildering. AT&T was not only a phone company on its own; it contained within it a multitude of other large companies as well. Each region of the country, for instance, had its own local phone company—New England Telephone, for example, or Pacific Bell. These organizations were owned in large part by AT&T and provided service for local phone customers. But these so-called local operating companies didn’t manufacture the equipment to actually make phone service work. For that there was Western Electric, another subsidiary of AT&T. On its own, Western Electric was larger than almost any other American manufacturing corporation. Its factories built the equipment that consumers could see (such as cables and phones), as well as equipment that was largely hidden from sight (such as switchboards). Finally, there was a third branch of AT&T. Neither the local phone companies nor Western Electric maintained the long-distance service that connected different regions and states together. For that, there was AT&T’s Long Lines Department. Long Lines built and provided long-distance service to customers.
Both AT&T and Western Electric had large engineering departments. There was a certain amount of duplication—and sometimes friction—between the two. Generally speaking, the standards and long-term goals of the Bell System were determined by engineers at AT&T. Western Electric’s engineers, in turn, invented, designed, and developed all new equipment and devices.1 In 1916, the year before Fletcher and Kelly arrived, Frank Jewett was appointed the chief of Western Electric’s engineering division, which put him in charge of about a thousand engineers. Western’s main building was located on West Street in New York City, on the western fringe of Greenwich Village, in an immense thirteen-story yellow-brick redoubt that looked out over the tugboat and ferry traffic of New York Harbor. The engineers on the waterfront comprised a twentieth-century insurgency in a receding nineteenth-century world. The fragrance of coffee beans drifted through the large sash windows of the plant from the roasting factories nearby. A rail line, serving the busy harbor docks, stretched north and south along West Street in front of the building. “The trains ran along West Street carrying freight to the boats,” an employee there in the 1920s recalls. And oftentimes, “at dusk, a man with a lantern on horseback led the trains.”2
Under Jewett, Western engineers worked mainly in expansive open rooms floored with maple planks and interrupted every few dozen feet by square stone pillars that supported the building’s massive bulk. The elevators were hand-operated. All told, the rambling West Street plant comprised over 400,000 square feet—a figure that did not include the building’s rooftop, which was also used by chemists for testing how various lacquers and paints and metals withstood the elements. In their first days at the Western Electric shop, Kelly and Fletcher encountered a small city of men, along with a number of female assistants. Vast rooms of the building were dedicated to diagramming new devices for production—men in crisp white shirts, sleeves rolled above their elbows, bent over rows and rows of drafting tables. Before a device was ready for the drafting room, though, it would have to pass through a lengthy and rigorous development process. West Street was a warren of testing labs for phones, cables, switches, cords, coils, and a nearly uncountable assortment of other essential parts. There were chemical laboratories for examining the properties of new materials, such as alloys for wire and sheathing for cables; there were numerous shops, meanwhile, cluttered with wires and dials and batteries, where legions of employees spent their days testing the effects of electrical currents and switching combinations or investigating new patterns of circuitry. Large sections of the labs were also devoted to the perfection of radio transmission, for it was believed (by Jewett’s boss, John J. Carty, especially) that wireless transmission would be a thing of the future, a way “to reach inaccessible places where wires cannot be strung,” or a way to someday create a commercial business linking New York to, say, London.
There was no real distinction at West Street between an engineer and a scientist. If anything, everyone was considered an engineer and was charged with the task of making the thousands of necessary small improvements to augment the phone service that was interconnecting the country. Yet the company now had a small division of men working in the department of research with Harold Arnold. This department was established just after Arnold began his work on a cross-country phone repeater; it had grown slowly and steadily in the time since. Frank Jewett and John J. Carty viewed the research team as an essential part of the phone company’s business strategy.3 These young scientists, many of whom came through Millikan, were encouraged to implement Theodore Vail’s long-term vision for the phone company—to look beyond the day-to-day concerns that shaped the work of their fellow engineers (to think five or ten years ahead was admirable) and focus on how fundamental questions of physics or chemistry might someday affect communications. Scientific research was a leap into the unknown, in other words. “Of its output,” Arnold would later say of his group, “inventions are a valuable part, but invention is not to be scheduled nor coerced.” The point of this kind of experimentation was to provide a free environment for “the operation of genius.” His point was that genius would undoubtedly improve the company’s operations just as ordinary engineering could. But genius was not predictable. You had to give it room to assert itself.
. . .
JOINING WESTERN ELECTRIC, even as a PhD in physics, entailed indoctrination in the phone company’s ways. In Harvey Fletcher’s first year he was taught to climb telephone poles, install telephones, and operate switchboards. Kelly’s experiences must have been similar, but his arrival also coincided with the company’s deepening involvement in building equipment for the military during the final years of World War I. He and his wife, Katherine, lived in a small apartment on Edgecombe Avenue in upper Manhattan, where she would look out the window each day to follow the construction of the Cathedral of St. John the Divine, located on a hill a few dozen blocks south. Kelly, meanwhile, began work in Harold Arnold’s division by sharing a lab office with a physicist named Clinton J. Davisson, whose friends called him Davy. Davisson was an almost spectral presence at the Labs. Taciturn and shy, he was physically slight. “His weight never exceeded 115 pounds,” Kelly recalled, “and for many years it hovered around 100.” Kelly believed Davy was quiet for a reason. He needed to minimize superfluous activity or argument so he could husband his “low level” of energy. Only by doing so, Kelly believed, could Davy direct it, vigorously, toward experimentation.
The two men were a peculiar contrast: the antic and robust Kelly paired with the wraithlike and slow-moving Davisson. Yet it didn’t take long for Kelly to discover he was impressed. Davisson was a midwesterner, too—he was born in Bloomington, Illinois—and like Kelly he owed a debt to Millikan at Chicago, who had championed his career and had helped him win academic appointments at Purdue and Princeton before he came to Western Electric. Also, Davisson was a gifted experimentalist who had an almost unwavering commitment to what Kelly would later define as basic research—that is, research that generally had no immediate application to a product or company effort but (as in Davy’s case) sought fundamental knowledge regarding the deeper nature of things, such as the behavior of electrons.
At Western Electric, Davisson’s passion, not to mention his manner, made him something of an oddity. Industrial labs were less interested in basic research—that was better left to the academics—than in applied research,
which was defined as the kind of investigation done with a specific product or goal in mind. The line wasn’t always distinct (sometimes applied research could yield basic scientific insights, too), but generally speaking it was believed that basic research preceded applied research, and applied research preceded development. In turn, development preceded manufacture.
In Kelly and Davisson’s first years of 1917 and 1918, the military demanded workable technology in Europe—radio sets, cable lines, and phones produced in mass quantities and built to a higher standard than the ones used in the home market so as to withstand the stresses of battle. Kelly and Davisson were assigned to work on resilient vacuum tubes, which were still so new to communications that they hadn’t yet entered mass production. “The relatively few that were required for extending and maintaining [phone] service,” Kelly would remember, “were made in the laboratories of the [Western Electric] Engineering Department.” Thus on West Street the tubes needed to be designed and built, with the help of a team of expert glassblowers, and then tested for defects, one at a time. It was a development shop, in other words, with an eye on rapid deployment for urgent military needs. Until the end of the war, there wouldn’t be time for applied research, let alone basic research.
Kelly and Davisson worked together “in an atmosphere of urgency,” as Kelly recalled.4 “The rapid tempo of the work, with the necessity of accepting partial answers and following one’s nose in an empirical fashion, were foreign to [Davisson’s] way of doing things.” Still, Davisson seemed to accept the cut-and-try approach, along with the switch from research to development, without complaint. In a way, he and Kelly had largely regressed to the old inventive traditions of Edison. But in the process Kelly was learning some things about Davisson. If the Western Electric engineers in the tube shop confronted a baffling question, they would approach Davy, who would give a deep and thoughtful and ultimately convincing response—though it sometimes took him days to do so. Increasingly, Kelly recalled, he and the rest of the staff went to Davy as a matter of last resort. Western’s physicists, like Kelly, could easily understand whether a new tube, or a new tube design, worked or failed, in other words. But they couldn’t always easily understand why.