Edwin Howard Armstrong

b 1890 New York City, d 1954 New York City

Armstrong is the most important engineer on this list, and one of greats of the 20th century. Most of the greats are known for one or two key innovations: Widlar and the op amp, Cray and vector supercomputers, Sutherland for both Sketchpad and flight simulation. Armstrong had three: regeneration, superheterodyning, and frequency modulation. He was the last of the line of heroic individual American inventors, and he came to a particularly American end - death by lawyers. His last struggles also bear on questions that are puzzling to this day, such as why AM radio hasn't been swept away by FM.

portrait of Armstrong

Here he is in the late 1940s, looking like a particularly unsympathetic banker, but showing the fierce and uncompromising attitude that ultimately did him in.

The reference here is "Edwin Howard Armstrong - Man of High Fidelity" by Lawrence Lessing, 1956. Lessing was a personal friend, and saw in Armstrong's suicide the triumph of blind corporate power over individual achievment. It's a heated book, but an engrossing read.

Armstrong was born into a large Presbyterian family in Manhattan (something that doesn't happen today) in 1890. His people were energetic and affable, but of no visible creativity. The only thing that distinguished his childhood from that of a million other middle-class kids was an attack of St. Vitus Dance (rheumatic fever) when he was nine. It kept him out of school for two years, and left him with a tic in his shoulder and jaw for the rest of his life. He soon caught up in school but was always a serious kid. He was fascinated by Marconi's adventures and soon learned all there was to know about wireless.

He enrolled in electrical engineering at Columbia, and in 1913, while still an undergrad, made his first great discovery, regeneration. At that time Lee De Forest's audion tube, the first triode vacuum tube, had been around for several years. No one had done much with it, though, because De Forest himself didn't understand how it worked and because it was a really poor amplifier. Armstrong discovered that the gain of a triode amplifier could be enormously increased by feeding some of the amplifier output back into the input, i.e. by using positive feedback. Given enough feedback, the amplifier became a stable and powerful oscillator, perfect for driving radio transmitters. Given a little less feedback, the amplifier became a more sensitive radio receiver than anything else at the time. Other people had come upon feedback at about the same time, but Armstrong characterized it and understood it in a way that made it practical for real use.

De Forest heard of Armstrong's work, and immediately directed his own research into regenerative techniques. He quickly filed patents on variants of the technique, and then in 1920, started attacking Armstrong's patents on it. He was infuriated that such a young man had used his own tubes better than he had. De Forest merits a note on this list of his own, since he founded and drove into bankruptcy 25 different companies. He was also married and divorced three times. His only success was the audion tube, so he was determined to gain control of its use. He was backed up by AT&T, who stood to gain enormously if they could control a fundamental circuit of radio.

The patent fight lasted fourteen years, cost over a million dollars, and went in front of the Supreme Court several times. De Forest's entire evidence of priority was a 1912 note in a lab book that a particular circuit emitted a howl when tuned a certain way. The howl was because feedback caused the circuit to oscillate, but he didn't understand that and didn't pursue it. AT&T's lawyers, however, managed to get a sympathetic ruling from a technically ignorant judge in Philadelphia. Given one such ruling, they could beat any appeals by saying that judgement had already been rendered. They beat the Supreme Court appearances by so obscuring the issues that the Court refused to hear the case.

So they broke Armstrong's patent. AT&T's publicists also went into high gear, and to this day you'll see De Forest credited with regeneration.

Fortunately for Armstrong, he had plenty more ideas where that came from. While working for the Army during World War I, he came up with the superheterodyne, a subtle and elegant technique for improving reception and tuning at the same time. It's quite difficult to build an amplifier that will work at high frequencies, such as the ones radio uses, and also difficult to build a tuning filter than can select a narrow band of frequencies and yet be adjusted across a range of frequencies. The filter must tune in one station and reject all others, but then change to tune in other stations. It's easy, however, to build a tunable oscillator. If the oscillator signal is added to an incoming radio signal, a beat signal will result that has a frequency of the difference between the two. A fixed filter can be built to narrowly select this beat frequency, and pass it on to a low-frequency amplifier. As the oscillator frequency is varied, different radio frequencies will be moved down to the beat frequency and so selected. In other words, a variable oscillator and a fixed, narrow filter can do the work of a variable narrow filter. This technique is used in practically every radio to this day.

Armstrong and his wife as newlyweds Here are Armstrong and his wife Marion on their honeymoon in 1923. She was the secretary of the president of RCA, David Sarnoff. He brought one of the first superheterodyne sets with them - the very first beach boom box.

Westinghouse paid hundreds of thousands for the rights to the superheterodyne. Armstrong also made a deal with RCA for a similarly innovative circuit, the super-regenerative receiver. This would drive an amplifier right to the edge of oscillation and then kill the gain to avoid a squeal. If done at more than 20 kHz, the ear couldn't pick up the variation in gain, and a lot more gain could be gotten out of one tube. This wasn't as profound as his other innovations, but it netted him a big block of RCA stock and made vastly more money than all of his other ideas.

In his whole career Armstrong never became an employee. Instead he was on the faculty of Columbia for a salary of one dollar a year, since his patents paid him much more than the university could. He never taught classes, but they were honored to have him. It gave him access to a broad intellectual circle and research resources. He rented lab space in their basement. He never incorporated, and did all his work with only a couple of assistants, some of whom went on to significant work of their own.

The patent fight with De Forest was exhausting and humiliating. When the courts turned against him and the publicists referred to him as "that discredited inventor", he resolved to show them all. This was in the early thirties, when Armstrong himself was in his early forties. Although this is past the creative prime for most engineers, Armstrong came up with his greatest invention of all, frequency modulation (FM).

In conventional, amplitude modulation (AM), the strength of a radio signal is proportional to the strength of the audio signal that is being transmitted. Unfortunately, the natural world is full of similarly modulated signals, which are then heard as static. In FM, the frequency of the main signal is varied instead of its amplitude. Since few natural sources vary in the same way as an FM signal, interference is much less.

FM had been tried in the 20s and rejected. It was thought to be a way to pack more signals into a given frequency band. Detailed mathematical analysis showed that a narrow-band FM signal would always sound worse than an AM signal of the same power.

Armstrong's insight was that an FM signal didn't have to have a narrow range of frequencies. It could vary over a wide range, say five times as wide as an AM signal, and have a far better signal-to-noise ratio. The mathematics for FM had been correct, but confining. By relying on experiment and physical reasoning, Armstrong got beyond the equations. This in turn laid the foundations for information theory, which quantifies how bandwdith can be exchanged for noise immunity.

He first got FM working in 1933. He showed it to RCA, who had most of his licenses, but they were oddly unimpressed. RCA had at that point made vast investments in AM. All of their transmitters and all of the millions of radios that they had sold used it, and the investment was not depreciated. Radios were now commodities, and sold entirely based on price, not quality. RCA told itself that consumers didn't care what the music sounded like from their radios, they just wanted to get it as cheaply as possible. They spent a couple of years evaluating the technique and then declined to license it.

Fine, said Armstrong. He went ahead with licenses to smaller companies. He designed the complete system of transmitter, antennas, and receivers, and set up pilot broadcasting services in New York and New England in 1939. People were entranced. This was the highest quality medium for music of the day, much better than the phonographs, and before the Germans perfected tapes.

RCA immediately struck back by petitioning the FCC to give FM's frequency assignments (around 50 MHz) to television, which was just starting up at the time. Their attack was so obvious, however, that the FCC chairman, a New Dealer named Fly, instead gave the whole band from 44 to 50 MHz to FM. This would have been TV Channel 1, and it's still missing from the TV dial. He also required TV sound to be carried as FM.

Since RCA couldn't lick 'em, they decided to join 'em. They offered Armstrong a million dollars for his patents, but no subsequent royalties. Armstrong refused. Every other licensee paid royalties, and he felt that giving such a deal to RCA would be unfair to the companies that had actually worked with him instead of against him. Before the battle could really get heated up, though, a somewhat larger struggle, World War II, began.

Everybody used FM during the war. Armstrong allowed the military to use his patents royalty-free for the duration, a gesture that no company could make and that even he, with his lab expenses, could barely afford. Mobile FM communications were of tremendous value in the thrusts and parries across Europe and the Pacific. He himself worked on continuous-wave radar, but didn't get anything deployed before the war's end.

RCA had not rested in the meantime. In 1945 it and a bloc of other radio companies convinced the FCC to move the FM band from 44-50 MHz to 88-108 MHz, where it is today. The nominal reason was to prevent "ionospheric disturbance", but no one was quite clear on what this was. The TV channel right above 50 MHz with its FM sound channel didn't seem to suffer from it. What it did do, though, was to immediately obsolete all the transmitters and receivers that had been built. To add insult to injury, the FCC also voted to severely limit FM's broadcasting power, and disallow radio relays from central stations to mountaintop antennas. Instead, the FM broadcasters had to send their material over AT&T's coax cables at exorbitant rates.

The overall, and intended, result was to cripple FM broadcasting. These machinations allowed AM to survive even down to the present day. No programming of any audio quality is offered on AM any more. It was obsoleted more than 60 years ago, but is kept alive by inertia and regulation.

After this defeat, Armstrong grimly set out to redesign all his systems. He got them working again at the higher frequencies by 1948, but that was his last engineering achievment. RCA had been building FM receivers using his patents for the previous eight years without paying him a dime. His patents only had two years left to run. It was time for them to pay.

He brought a patent infringement suit against them in 1949. Armstrong himself was called on to be the first witness. RCA's lawyers kept him on the witness stand for an entire year with niggling and irrelevant questions. Another two years elapsed when RCA was called upon to reveal the mountain of research it had done on FM in the 30s. The capstone of this was when David Sarnoff himself claimed that RCA had invented FM all by itself without any help from Armstrong. After a claim like that, Armstrong would hear no talk of settlement.

By 1953, Armstrong's licenses and patents had all expired. His crushing legal bills and research expenses brought him to near bankruptcy. A bitter argument with his wife on Thanksgiving day caused her to leave him. She went to live with her sister in Conneticut.

On January 31, 1954, he wrote a two-page letter to her and left it on his apartment desk. He dressed neatly in an overcoat, hat, scarf, and gloves, and walked out a 13th story window. He hit a third story overhang, and so his body wasn't discovered until the next day.

His wife Marion continued his suits. Unlike Armstrong she was willing to compromise. She settled with RCA for over a million dollars, then went after other companies like Sylvania and CBS that had also infringed his patents. She conclusively won all her cases and collected millions, although the last holdout, Motorola, didn't give up until the Supreme Court ruled against them in 1967, thirteen years after Armstrong's death.

According to Armstrong's biographer, Lessing, true innovation is always the work of individuals or small groups. Radio is built around seven inventions made by only nine people. Plastics are based on the work of five independent scientists, four German and one American, and only the last, Wallace Carrothers, did significant work within a company, Du Pont. (Oddly enough, he was also a suicide.) Atomic energy is the work of a small group of refugee Europeans. Innovation is lightning in the brain, and that doesn't happen in committee meetings.

According to ads that Microsoft ran recently in the Wall Street Journal, a country's true innovators are its companies. It's an extraordinary claim, especially from a company not known for innovation. The corporatist attitude seems to be as alive today in Microsoft as it was in RCA's and AT&T's attempts to discredit Armstrong. RCA was long ago swallowed up by General Electric, and even AT&T is fading these days, but the publicity machine lives on. Only the organization produces; the individual is just a cell in its body.

Most engineers go along with this, for it does take a lot of people to get things done. Lots of people contribute to any project's success - should they all be slighted for the sake of one person's ego? No, but in emphasizing the team over the individual, we risk emphasizing the organization over the engineer. The company becomes paramount and its people become interchangeable resources.

Armstrong was not a company man, yet he gave more to radio than RCA ever did. He gave to his country as well in both world wars. His example shows how Microsoft has it wrong. The true source of a country's innovations are not its companies, but its citizens.

(c) John Redford, Feb-96

Back to "Doomed Engineers"