Introduction
Today, Silicon Valley is known as perhaps the most important region in the world for the creation and use of digital technologies. But much of the groundwork for this famous technological hub was established in a remarkable fifteen-year period following the end of the Second World War. During that time, regional firms pursuing innovations in electronics grew rapidly. Engineers created electronic digital computers and made them a new business. Researchers invented the transistor, which quickly changed the entire electronics industry. The Cold War dominated geopolitics, and it was often waged through competition in electronics technologies. Silicon electronics came to the region that, by 1970, would first be called “Silicon Valley.” This story offers a tour of this pivotal era in the creation of Silicon Valley through a selection of special artifacts from CHM’s extensive collection.
Manufacturing the Electronic World
Manufacturing the Electronic World
The creation of vacuum tubes in the early 20th century that could control and amplify electrical currents sparked the new world of electronics. Scientists, engineers, and entrepreneurs in the San Francisco Bay Area explored the new possibilities enabled by tube-based electronics. In the late 1940s, researchers at the Bell Telephone Laboratories in New Jersey introduced a new device, the transistor, which rivaled the vacuum tube, and would reshape the electronic world.
Audion
In 1906, inventor Lee De Forest created the “Audion” tube, which allowed the amplification of electrical signals. In the 1910s, researchers modified De Forest’s designs to work with high vacuums inside the tubes, expanding their performance and applications. For several years, De Forest worked on vacuum tube electronics research at the Federal Telegraph Company in Palo Alto, California. With the technology, the Bay Area became a critical hub for radio broadcasting. This is an early example of a De Forest Audion.
Eitel-McCormack
In the mid-1930s, two amateur radio hobbyists, Bill Eitel and Jack McCullough, started a new and ambitious vacuum tube manufacturing company on the San Francisco Peninsula. Selling under the brand “EIMAC,” Eitel-McCullough rapidly became a major national supplier of vacuum tubes to electronics professionals, radio hobbyists, and the US military. The skilled workforce, manufacturing expertise, supply chains, and infrastructure of vacuum tube makers like Eitel-McCullough would prove vital for the further development of electronic and computing industries in the region. This is an EIMAC 4-65A tetrode vacuum tube, an amplifier designed for use in radio transmission.
Hewlett Packard
In 1939, Bill Hewlett and David Packard founded their eponymous startup in a one-car garage in Palo Alto, near Stanford University, where the pair had studied electrical engineering, particularly vacuum tube electronics. The firm’s initial product line was precision audio oscillators, which used vacuum tubes to produce exact tones. This is a model 200C oscillator, introduced in 1940. Hewlett Packard grew to become a major producer of electronic test equipment, and after 1960 of instrumentation, components, and computing. With its distinctive employee-centered culture, HP became an important training ground for generations of engineers, technical staff, and managers in the region.
Varian
In the late 1930s, the brothers Russell and Sigurd Varian were working in Stanford University’s Physics Department with Professor W. W. Hansen. Drawing from Hansen’s work and the prevailing technology of vacuum tube electronics, the brothers created the “klystron,” a specialized tube capable of producing powerful microwaves. Klystrons proved essential for radar developments during the Second World War, and afterwards for radar of all kinds, long distance communications, satellite links, and physics research. In 1948, the Varian brothers, Hansen, and several others launched a startup called Varian Associates in Palo Alto near Stanford to manufacture klystrons. Soon, the region became a center for microwave electronics research and manufacture. After 1960, Varian became a major player in instrumentation, ion implantation, computers, and medical equipment. This is a Varian model V-82 klystron, likely from the mid 1950s, for producing X-band radar waves.
Whirlwind I
The earliest electronic digital computers emerged in the 1940s, with researchers and engineers using vacuum tubes as the critical component. Tubes were used for a variety of purposes, including the “on/off” switches at the heart of digital operation. In the early 1950s, a variety of vacuum tube based commercial and experimental computers were created. One of these experimental computers, the Whirlwind I, was built at the Massachusetts Institute of Technology using military funding. Whirlwind was hugely influential for a generation of computer researchers and programmers. This is a section of Whirlwind’s processing racks from 1951.
Philmore Crystal Radio Detector
Early in the era of vacuum tube electronics and radio, researchers, entrepreneurs, and hobbyists started putting to use an important class of materials: semiconductors. Semiconductors were defined by the way they conducted electricity, that is, so-so. They stood between “conductors” like metals that allowed currents to flow, and “insulators” like glass that largely prevented them. Some chemical elements were semiconductors, and some compounds were as well. At the opening of the twentieth century, experimenters found that placing a metal wire against a piece of semiconductor formed a “junction” that allowed current to flow only one way through it. This effect could be used to make a radio detector, turning the radio waves into a direct current audio signal that could be heard on headphones or a speaker. It was the start of the role of semiconductors in electronics. This is a radio detector made by Philmore in the mid 1920s, using galena (lead sulfide) crystal as the semiconductor.
The Transistor
Semiconductor materials were the basis for a new kind of electronic device, first created at the end of 1947. Since the 1920s, researchers had explored making an equivalent to the vacuum tube, but as a solid device without the need for a high vacuum inside of a glass enclosure. Such a solid device would avoid some of the reliability and power requirements of tubes. None of these experiments proved successful. At the end of 1947, researchers at the Bell Telephone Laboratories in New Jersey, John Bardeen and Walter Brattain, showed that a device made from two closely spaced metal contacts on a piece of the semiconductor element germanium could amplify an electrical current, like a vacuum tube. Quickly named the “point-contact transistor,” this device was much smaller than a vacuum tube, required much less power, could amplify, oscillate, and switch, and was now a rival of vacuum tubes for the future of electronics. The following year another Bell Labs researcher, William Shockley, devised a new form of “junction transistor” made by chemically altering semiconductor materials in precise ways. In the following decades, transistors became the basis for all electronics and the dramatic spread of computing. These photographs showing the first point-contact transistor, Bardeen, Brattain, and Shockley, and an early production version of the point-contact transistor were collected by Steve Allen, who then worked for transistor manufacturer Fairchild Semiconductor.
From Invention to Production
From Invention to Production: Point-Contact and Junction Transistors
Western Electric Pre-Production Germanium Point-Contact Transistors
The Bell Telephone Laboratories was the central hub for research and development for American Telephone and Telegraph (AT&T), a company that, at the time of the invention of the transistor, was a regulated monopoly providing telecommunications for the United States. Western Electric was a subsidiary of AT&T, responsible for manufacturing all the equipment for the system, from home telephones to relays, and from vacuum tubes to switching stations. With the advent of the transistor, Western Electric became one of its earliest and largest manufacturers. An important customer for Western Electric’s point-contact transistors was the US military. By 1958, Western Electric had sold more transistors to the military than they had produced for use in the telephone system. These are rare pre-production samples of the Western Electric Type 1729 germanium point-contact transistors, manufactured between 1951 and 1953.
Germanium Products Corporation Germanium Grown-Junction Transistors
As a regulated monopoly, AT&T was in constant negotiation with the federal government about its pricing and products. One settlement between the two, finalized in 1956 but anticipated much earlier, mandated that AT&T provide royalty-free licenses for all its patents at reasonable rates and that it refrain from any business but telecommunications. This meant that transistor technology was widely available, and Bell Labs actively supported the transfer of its technology to other businesses. One of these, Germanium Products Corporation, began production of the newer junction type transistors, made by “growing” junctions by chemically altering germanium as it was formed into crystals. These are very rare examples of Germanium Products grown-junction germanium transistors made in the early to mid 1950s. This RD2525 model was introduced in 1953 and was an improvement on an earlier model used in the first commercial product to incorporate transistors: the Sonotone 1010 hearing aid.
Philco Germanium Surface-Barrier Transistor
One of the key characteristics of a transistor is its speed, or how quickly it can transform from “on” to “off.” The faster the transistor, the quicker a switch made using them, or the better the transistor can act to make or amplify signals, like radio frequencies. In short, the speed of a transistor had a lot to do with how a transistor could be used. Philco, a major electronics firm, developed a very fast junction transistor that involved a complicated process of etching the germanium to very thin dimensions and then alloying materials to it. These surface-barrier transistors were used to make early transistorized computers and in a variety of communications applications by the military and others. These are very rare and early production devices from 1952 or 1953.
Texas Instruments Silicon Grown-Junction Transistor
Up until 1954, transistors had been made of the chemical element and semiconductor germanium. However, the element silicon had long been of interest for electrical and electronic uses. Theoretically, a silicon transistor could have advantages over germanium in terms of reliability and performance. But the high temperatures required to work with silicon presented a real challenge. These challenges were met in 1954. At Bell Labs, Morris Tanenbaum used the grown-junction approach to create a silicon transistor. Very soon thereafter, and independently, Gordon Teal —who had done pathbreaking work on grown junction technology at Bell Labs —created a silicon grown-junction transistor at his new employer, Texas Instruments. Texas Instruments quickly took the market for silicon grown-junction transistors, and by the end of the decade silicon was the material of choice for transistors and electronics. This is a very early production example of Texas Instruments’ first silicon grown-junction transistor.
Sonotone 1010 Hearing Aid
Miniaturization has always been an important goal in the development of hearing aids. For decades, hearing aid makers sought to use fewer, smaller, and less power-hungry vacuum tubes to make the devices more comfortable, portable, and effective for their users. Hearing aid makers were the ultimate early adopters of the first commercial transistors because of their smaller size, lower power requirements, and robustness relative to most vacuum tubes. This Sonotone 1010 hearing aid, introduced in late 1952, and was the first commercial product to use a transistor. It contained two very small vacuum tubes and two Germanium Products Corporation germanium grown-junction transistors for processing and amplifying sound.
Regency Model TR-1 Radio
While military and communications uses of the transistor predominated in the early 1950s, many organizations began to search for new, and potentially much larger, markets for transistors in commercial products. The Sonotone 1010 hearing aid had shown that commercial products were possible. At Texas Instruments, Executive Vice President Pat Haggerty decided on all of the firm’s electronics activities. In 1954, he set the goal of making a miniaturized, portable radio using transistors. With the popularity and ubiquity of radio in this era, Haggerty saw a large potential commercial market that could drive demand for vastly more transistors. By the end of the year, Texas Instruments engineers had developed a new germanium grown-junction transistor that fit the bill and could be manufactured in large volume, . and helped design a radio with the new transistor. The radio was introduced by their manufacturing partner Regency as the Regency TR-1. The radio unlocked a huge demand for transistorized radios and helped to create a large, commercial market for transistors.
Early Transistorized Computers
TX-0 Plug-In “Bottle” Modules
In the first half of the 1950s, MIT’s Lincoln Laboratory was a military-funded organization working to transform the experimental Whirlwind computer (see above) into an IBM-produced mainframe for the military’s air defense system. Researchers at the Lab were also pushing forward ideas for new components, new computers, and new computer applications. In 1955 and 1956, one group decided to explore the potential for making computers out of transistors. They created TX-0, “Transistorized Experimental Computer Zero,” on the pattern of Whirlwind but made with transistors instead of vacuum tubes. They used over three thousand Philco germanium surface-barrier transistors packaged into clear plastic “bottles” containing one or several transistors, sometimes with other components. These are original transistor bottles used in TX-0. The computer’s impact was enormous at MIT and beyond. Within four years, Ken Olsen, a key player in making TX-0, founded the Digital Equipment Company and introduced the hugely influential PDP-1 computer to the market in 1961.
TRADIC
At the same time that TX-0 was being built in the context of military computing at MIT’s Lincoln Laboratory, Bell Telephone Laboratories announced the creation of its first transistorized computer, called TRADIC, for “Transistorized Airborne Digital Computer,” in 1955. Like TX-0, TRADIC was built with military funding. It used over six hundred germanium point-contact transistors (like the Western Electric transistors above) and over ten thousand point-contact diodes. This photograph from 1955 shows TRADIC installed in a US Air Force airplane where it provided bombing and navigation control.
The University of Manchester Transistor Computer
In the United Kingdom, the University of Manchester was a critical site for the development of early electronic digital computers, starting in the mid 1940s. In 1953, researchers there succeeded in operating one of the first computers to be built using transistors, known as the “Manchester TC,” or “Manchester Transistor Computer.” This first prototype used 92 germanium point-contact transistors. A second prototype, operational in 1955, used over twice as many transistors. In this CHM event from September 2000, Richard Grimsdale, one of the researchers who built the Manchester Transistor Computer, discusses its history.
Superconducting Electronics
Cryotron Assembly #36
In the early 1950s, electronics researchers could scarcely have imagined tthat transistors, and integrated circuits built from them, would underly digital computing and indeed almost all of electronics for the next seven decades. Rather, the early successes of the transistor inspired many of them to search for the next great electronic device that would replace the transistor. At MIT, the young professor Dudley Buck created what he thought was just such a device in 1954: the cryotron. In this fast switch based on the phenomenon of superconductivity, Buck— and soon many other researchers —saw a rival to the transistor in terms of speed, size, and low power even though it required very low temperatures (-452 degrees Fahrenheit). While superconducting electronics did not displace the transistor, it has become the main approach for quantum computing in the 21st century. This is one of the first cryotrons made by Dudley Buck.
Cryotron Circuit
In 1954, soon after Dudley Buck invented the fast superconducting switch he named the cryotron, one of his close associates, Albert Slade, made a complete circuit using the new device. Slade had learned of the cryotron at the National Security Agency (NSA), where Buck had also worked and remained a consultant. Slade’s circuit, a ring oscillator, proved the utility of superconducting electronics. This is Albert Slade’s 1954 cryotron circuit.
Growth in the Valley
Growth in the Valley
Views of Stanford University, c. 1957
Frederick “Fred” Terman was a dynamic, entrepreneurial professor of electrical engineering at Stanford University in Palo Alto, California, from the 1920s into the 1940s. Having earned his doctorate at MIT, and having seen the close connections there between industry and the academy, Terman followed this model at Stanford. With his encouragement, his students founded Varian and Hewlett-Packard. Startingn the late 1940s, Terman was a powerful dean, who did much to develop the University into an engineering research leader, especially in electronics. This included developing a vast industrial park for new companies with ties to the University. The close ties between industry and the academy in the Bay Area continue to this day. The dramatic growth of Stanford is documented in these photographs by Fred English circa 1957, commissioned by the University. They show the newly opened Dinkelspiel Auditorium and an aerial view of the auditorium and its position in the central campus. Another aerial view shows the Stanford Research Institute, home to the university’s contract research for the government, military, and industry. Finally, a pastoral landscape shows one site considered for the development of the Stanford Linear Accelerator Center, a high energy physics laboratory founded in 1962.
Lockheed Missiles and Space Company
By the mid-1950s, the Lockheed Aircraft Corporation had three decades of experience in designing and manufacturing airplanes and had undergone an incredible expansion in Southern California, where it was headquartered, during World War II. Contracts for the US military expanded during the Cold War, including winning contracts to develop the submarine-launched Polaris nuclear missile for the US Navy and the top-secret Corona reconnaissance satellites for the US Air Force and CIA. To contend with these two huge programs, Lockheed established the Lockheed Missiles and Space Company in Sunnyvale, California. By 1956, it was the region’s largest employer, with 28,000 employees. These efforts required substantial computing resources, and the firm also created a computing center in Sunnyvale, boasting IBM’s powerful 7094 mainframes. So great were the computing demands, that the firm established a computing research group in Palo Alto to develop advanced software. These documents describe new programming languages developed there in the mid-1960s.
Silicon Comes to Silicon Valley
Silicon Comes to Silicon Valley: Shockley Semiconductor
Early 1950s Automation and Robotics: Squee
The opening of the 1950s saw a tremendous enthusiasm in business and technical communities for the concept of automation, which promised highly automatic factories built around feedback systems using electronic instruments, computers, and controls. At this same time, a variety of researchers became interested in robots built using these same electronic systems and feedback principles. At the Bell Telephone Laboratories, William Shockley—fresh off his invention of the junction transistor —was also pursuing a wide-ranging vision of the “automatic trainable robot,” incorporating electro-optical sensors, as the means for the widespread automation of US industry. He urged his superiors at Bell Labs to significantly reorient the organization toward this vision. When they refused, Shockley left. In 1951, computing researcher and enthusiast Edmund Berkeley orchestrated the creation of Squee, “an electronic robot squirrel.” Using electro-optical sensors, Squee could locate and collect objects and move to select locations. This is Edmund Berkeley’s Squee.
Early 1950s Automation: Beckman Instruments’ Synchro
Like William Shockley at Bell Labs, the early 1950s saw southern California technologist-entrepreneur Arnold O. Beckman also deeply engaged in the automation movement. His firm, Beckman Instruments, was a major producer of electrical and electronic components, including synchros. Much like an electrical motor, synchros could precisely measure rotation, and, crucially, synchronize rotational positions across multiple connected synchros. They were important components for military targeting systems, feedback control systems, and analog computers. This is a Beckman Instruments synchro used by Dick Norberg for electronic analog computing in the early 1950s.
Early 1950s Automation: Beckman Instruments’ Analog Computer
In addition to synchros, Beckman Instruments was a leading producer of precision potentiometers, the crucial component for building electronic analog computers. As a result, Beckman’s company began to manufacture analog computers as well and was one of the largest producers in the US by the mid-1950s.. Beckman expanded his operations into the Bay Area, with groups manufacturing digital counters as well as advanced biomedical instruments. He was particularly interested in using analog computing systems for industrial automation, like controlling the operations in a chemical plant. This is an electronic analog computer built by Beckman Instruments in the 1950s.
1955: Shockley Approaches Beckman About Automation
In early 1955, William Shockley approached Arnold Beckman about the possibility of using his patent on electro-optical controls—the basis for Shockley’s “automatic trainable robot” vision —for Beckman Instrument’s efforts in automation. While Beckman decided to pass on the offer, the two men continued to talk. Soon, the conversation turned to Shockley’s desire to start a new company to pursue the latest in transistor manufacturing and to develop new innovations in semiconductor electronics. These are Arnold Beckman’s original files about Shockley’s initial offer around the electro-optical controls.
1955: Arnold Beckman and William Shockley Create the Shockley Semiconductor Laboratory
By the close of 1955, Arnold Beckman and William Shockley had signed a business agreement that brought silicon electronics to what would later, become known as “Silicon Valley, USA.” The pair agreed to establish the Shockley Semiconductor Laboratory of Beckman Instruments, Incorporated. Shockley’s new laboratory, which he successfully lobbied to open in Palo Alto, would pursue the large scale, automated manufacture of an advanced transistor —the diffused silicon junction transistor —and also conduct research on new, cutting-edge semiconductor devices. The deal specified that if the new laboratory was successful after two years, it could be spun out of Beckman Instruments as a new independent firm. These are Arnold Beckman’s original files about the establishment of the Shockley Semiconductor Laboratory.
1957: The Shockley Semiconductor Laboratory
Shockley’s new semiconductor organization launched in a rented space in Palo Alto, but it quickly moved to a larger location, more suited to experimentation and pilot production, at 391 San Antonio Road on the border of Palo Alto and Mountain View. This remodeled Quonset hut was the first silicon electronics laboratory in the region, and Shockley began to staff it with young, accomplished PhDs and experienced engineers from the East Coast. These recruits were excited by the company’s plan to manufacture diffused silicon junction transistors, then the cutting-edge, and to work with Shockley to push that edge further. But Shockley’s poor management skills soon put these goals in tension. Although many in the lab wanted to maintain focus on the primary goal of making the diffused silicon junction transistor, Shockley divided effort between the silicon transistors and a new device of his own, the four-layered diode, which he believed had tremendous promise. This is a photograph of the Shockley Semiconductor Laboratory, circa 1957.
The Original Silicon Start-up of Silicon Valley: Fairchild Semiconductor
844 Charleston Road: Fairchild Semiconductor
Dissatisfied with Shockley’s leadership and the technical direction of Shockley Semiconductor Laboratory, eight of the organization’s top scientists and engineers resolved to leave the firm after a failed appeal to Arnold O. Beckman. They believed that, together, they possessed the skills required to manufacture the double-diffused silicon junction transistor that had been their initial aim. After failing to find a company to hire the group en masse, they eventually decided to create their own firm, Fairchild Semiconductor, with backing from technophile-industrialist Sherman Fairchild through one of his companies, Fairchild Camera and Instrument. These candid photographs show four of the eight cofounders of Fairchild Semiconductor at its first headquarters, 844 Charleston Road in Palo Alto, barely two milesfrom Shockley’s lab. Pictured are Julius Blank, Vic Grinich, Gordon Moore, and Robert Noyce.
Venture Capital: Arthur Rock
Through a family connection, the eight dissidents in Shockley’s lab connected with Arthur Rock, an investment banker working on the East Coast. It was Rock who opened the group’s eyes to the possibility of forming a new firm instead of being hired by an established company. Despite a tremendous number of rejections, Rock’s perseverance resulted in the venture investment by Sherman Fairchild. Soon, Rock relocated to San Francisco, where he opened one of the first venture capital partnerships in the region. Across the 1960s and 1970s, Rock raised funds from the cofounders of Fairchild Semiconductor and others and made important investments in new technology startups in what became known as Silicon Valley, providing an important model. This is a video recording and transcript of a conversation between Arthur Rock and journalist John Markoff at the Computer History Museum in 2007 covering some of this history.
How Fairchild Initially Made Silicon Transistors
The cofounders of Fairchild Semiconductor quickly assembled the means to manufacture double-diffused silicon junction transistors, essentially a “chemical printing” technology. Small ingots of pure, single crystal of silicon were created using a machine known as a crystal puller. These were sliced, like a cucumber, into thin “wafers” of silicon. Polished and cleaned, these wafers were coated with a chemical known as a photoresist, which changed its properties on exposure to light. Using photographic techniques, the wafer was exposed to a geometric pattern of light, making the photoresist in the exposed places more durable. Powerful acids were used to etch away photoresist and silicon material, transferring patterns to the wafer. Between these patterning steps, the wafers were placed in high temperature furnaces and carefully exposed to different chemicals, which diffused into the silicon, making different regions with distinct chemical—and thus electrical —properties. These altered regions formed the transistor, with patterned metal layers added to form electrical connections. The chemically printed wafers were then cut apart into individual “die,” each with one transistor structure. The die were then placed and wired into protective packages, allowing them to be soldered into a final circuit. This photograph shows the various stages of the process, and this oral history with Gordon Moore,who was responsible for developing much of the diffusion technology at Fairchild in this period, contains rich detail about these early efforts.
The Work of Crystal Pulling
Initially, the cofounders of Fairchild Semiconductor built their own crystal pullers to supply their manufacturing line with the silicon wafers needed to create transistors. However, as soon as it was possible they switched to buying silicon ingots and wafers from independent suppliers that specialized in their production. At these suppliers, the hourly workers who operated crystal growers were often women with no prior experience in electronics or manufacturing. Patricia Anderson was one of these operators who worked for the supplier Elmat in the late 1960s and early 1970s. This is a portrait of Anderson from this timeframe, along with a set of crystal growing instructions she kept, a crucible that contained the molten silicon from which ingots were pulled, and the transcript of her oral history with the Computer History Museum.
Diffusion Technology
At Fairchild Semiconductor, cofounder Gordon Moore was responsible for building the critical diffusion furnaces used to manufacture silicon transistors. His background as an experimental chemistincluded glassblowing and the careful control of various instruments which proved essential to the task. Further, Moore used the furnaces to carefully determine the optimal diffusion procedures for reliable production. This photograph, likely from Fairchild Semiconductor, shows a diffusion furnace from the mid-1960s in operation, with a “boat” holding multiple silicon wafers placed inside of the furnace tube.
After Chemical Printing: Assembly and Test
After the creation of the transistor “die” through chemical printing, the next phases of transistor production required intensive labor. Through the 1960s and beyond, much of this labor was performed by women collecting hourly wages. These photographs from a Fairchild Semiconductor manufacturing plant in the mid 1960s show women workers engaged in assembling the transistor die into their packages and the subsequent testing of the transistor to make sure they met operating specifications and reliability standards. In this era, far more workers were involved with assembly and test than with the other stages of transistor production.
Fairchild’s “Mesa” Transistor
Fairchild Semiconductor opened for business in October 1957, the same month that Sputnik 1 was launched into orbit by the Soviet Union. The Soviet success underscored that much of the Cold War would be conducted as an ongoing rivalry in military technologies like nuclear weapons, aircraft, missiles, radar, and now satellites, all reliant on advanced electronics. Fairchild Semiconductor’s cofounders developed their diffused silicon transistor precisely for this expanding military aerospace market. In less than a year, they brought their advanced silicon transistor to market as the “2N696,” selling nearly half a million dollars’ worth to eager laboratories and companies serving US military aerospace demands. Informally, the 2N696 and other Fairchild transistors of this early period were also called “mesa” transistors, for their etched profiles evoked the geological mesa formation of the American Southwest. These are photographs of the Fairchild Semiconductor 2N696 in the collection of the Computer History Museum, along with two brochures on the device and the manufacturing technology circa 1958.
Using the Mesa: Computers for the Minuteman
Just before the founding of Fairchild Semiconductor, the United States Air Force began development of a new, long-range intercontinental ballistic missile (ICBM) for delivering nuclear weapons. The new missile—the Minuteman 1—would use solid rather than liquid fuel and would also incorporate an electronic digital computer for guidance functions. A newly formed division of the aircraft manufacturer North American Aviation called Autonetics had won the contract for developing the missile’s computer, and sought robust, fast-switching transistors to build it. Fairchild Semiconductor won a major victory in the electronics industry by securing the contract to supply the silicon transistors to Autonetics. This is a circuit board from the Minuteman 1 onboard computer manufactured by Autonetics and containing Fairchild’s silicon transistors.
World-Changing: The Planar Process
In the spring of 1960, Fairchild introduced a new diffused silicon transistor to the market, called the 2N1613. Behind this new device lay a new revolutionary process for making transistors. Fairchild cofounder Jean Hoerni, a physicist, had first conceived a new process for the chemical printing of technology in late 1957 and wrote a description of it in his patent notebook. Rejecting received wisdom about transistor making the idea was to use the layers of silicon dioxide— quartz—that readily formed on the surface of silicon wafers as a way to both form and then protect the transistor structures made on them. Hoerni returned to his ideas in early 1959 as a possible way to solve reliability problems that had emerged with Fairchild’s mesa transistors. Quick experimentation proved Horni’s intuition correct, and Fairchild soon shifted its emphasis and manufacturing to this new “planar process.” While silicon planar transistors became a huge market success, with most of the wider transistor industry adopting it rapidly, it also became the basis for another chapter in the history of Silicon Valley: the emergence of the planar integrated circuit, the microchips that underly our digital world today. This is Jean Hoerni’s patent notebook, now in the collection of the Computer History Museum, with his ideas for the planar process starting on page 3, along with a close-up photograph of a die for a planar silicon transistor, and a Fairchild Semiconductor publication about the planar process from 1961.