Università degli Studi di Padova Padua Research Archive - Institutional Repository Negative Feedback, Amplifiers, Governors, and More Original Citation: Availability: This version is available at: 11577/3257394 since: 2018-02-15T15:55:12Z Publisher: Institute of Electrical and Electronics Engineers Inc. Published version: DOI: 10.1109/MIE.2017.2726244 Terms of use: Open Access This article is made available under terms and conditions applicable to Open Access Guidelines, as described at http://www.unipd.it/download/file/fid/55401 (Italian only) (Article begins on next page) Historical by Massimo Guarnieri Negative Feedback, Amplifiers, Governors, and More Massimo Guarnieri he invention of the negative feed- Henry (1797–1878) and Samuel Morse (1873–1961), the holder of a similar back amplifier by Harold S. Black (1791–1872) and was very successful patent of 1916. In the final courtroom T (1898–1983) in 1928 is consid- against the attenuation of telegraph digi- battle in 1934, the Supreme Court ruled ered one of the great achievements in tal signals. in favor of De Forest. Meanwhile, in electronics. In fact, it is listed among Telephone lines, which started to 1922, Armstrong introduced the su- the IEEE Milestones, where it is cred- be laid in the 1880s, were also prone perregenerative receiver, which used ited to Bell Labs. Black was hired by to attenuation. However, their signals a larger part of the signal to obtain Western Electric in 1921 and as- were analog, so regeneration based on an even higher amplification (gain signed to work on the Type C system, a just an electrochemical battery and around 1 million). All these forms of newly introduced three-channel tele- an electromagnetic relay could not be positive feedback could provide a phone network whose push-pull, vacu- used. Early devices trying to regener- high gain, high selectivity, and high um-tube, repeater amplifiers produced ate telephone analog signals consisted quality factor. too much harmonic distortion when of pre-electronics crude amplifiers Electronic amplifiers were soon used connected in tandem [1]. At that time, [2], but the real step forward occurred also as repeaters for regenerating long- telephone networks were spreading, after the invention of the vacuum tri- distance cable telephone signals. With- and Bell Labs emerged quickly as the ode, namely, Lee De Forrest’s Audion, in such a program, Harold D. Arnold major research company in the tele- in 1906 [3]. (1883–1933) of Western phone industry. The extension of lines During the first de- Electric developed an im - over long distances required coun- cade of the 20th century, proved high-vacuum triode EARLY DEVICES teracting signal attenuation, which oc- a number of prolific in- in 1913 that was used in the curred, though at a reduced level, even ventors formulated its TRYING TO 2,900-km transcontinental in lines provided with Pupin’s loading use as a signal detector, a REGENERATE line connecting New York coils, which discretely increased the rectifier, an amplifier, and TELEPHONE and San Francisco, laid line longitudinal inductance L so as to a high-frequency oscil- ANALOG SIGNALS down by AT&T in 1914 and match the Heaviside condition L/R=C/G lator. However, vacuum CONSISTED OF put into service the follow- for distortionless transmission (R be- tubes were costly, also ing year [5]. However, in this PRE-ELECTRONICS ing the longitudinal resistance and C in terms of supplying bat- case, narrow bandwidth in- and G the transverse capacitance and teries, due to high power CRUDE AMPLIFIERS. volved poor linearity. conductance, respectively). demand. It was desirable The Western Electric The idea of the repeater amplifier that a high gain be provided by a single Type C system of 1921 was an advanced had developed from the well-established device. A solution was found in 1912 by solution for telephone networking that telegraph technology in which line atten- Edwin Armstrong (1890–1954), then required several repeater amplifiers uation was counteracted by the place- an undergraduate student at Columbia in tandem. Unfortunately, this ar- ment of repeating stations, consisting University. He conceived the regenera- rangement resulted in an unaccept- of batteries and relays operated by the tive receiver of wireless signals—an able harmonic distortion, and Black weakened arriving signals, at regular amplifier provided with partially pos- was appointed to search for a solution. distances. These regenerating stations itive feedback in which a part of the When Bell Labs was consolidated in allowed telegraph signals to travel over- output signal was fed back. In this way, 1925, by merging Western Electric Re- land as far as needed. This technology the signal was amplified hundreds of search Labs and part of the research was developed in the 1830s by Joseph times by a single device [4]. department of AT&T, Black was en- The system was patented in 1914 gaged in finding measures for improv- Digital Object Identifier 10.1109/MIE.2017.2726244 and was later the object of a long and ing the linearity and stability of such Date of publication: 21 September 2017 bitter litigation with Lee De Forest repeater amplifiers. The requirements 50 IEEE INDUSTRIAL ELECTRONICS MAGAZINE ■ SEPTEMBER 2017 were, by far, well beyond the limits of early operational amplifiers. It has been the technology of the time. used in biomechanics, bioengineer- After six years of persistent search, ing, cybernetics, computers, artificial during which he benefited from advice limbs, and many other devices. The from coworker Ralph Hartley (1888– transition to solid-state electronics 1970) and attended an inspiring lec- even increased its importance. Both ture by Charles Steinmetz (1865–1923), electrical and mechanical controls have Black conceived the groundbreaking largely grown on the concept of negative idea of the negative feedback ampli- feedback. It is exploited in psychology, fier one morning in August 1929, while physiology, sociology, and ethology, he was riding the Hudson River Ferry namely in studies on the behavior of the to work. (Figure 1) [6]. By December, brain, living beings, animal flocks, and he had fully demonstrated the sound- human societies [11]. ness of the concept by achieving, in a However, Black was not the first to single amplifier, a distortion reduction exploit the concept of negative feed- of 100,000 to 1 (50 dB) over a band- back. In 1788, at the dawn of the in- width of 4–45 kHz. Tandem tests were dustrial age, James Watt (1736–1819) performed in early 1930 on a system invented the centrifugal governor for reg- simulating a 12,000-km connection. FIGURE 1 – The page from the New York ulating in a negative-feedback scheme The test showed an excellent speech Times on 2 August 1927 in which Black the steam input and, thus, the engine sketched his early ideas on the negative quality, with an attenuation of about feedback amplifier. (Photo courtesy of Guill- shaft speed, preventing run away 12,000 dB, balanced by repeaters. ermo Carpinter.) (Figure 2) [12]. He was inspired by a A patent was required in 1928 and re- similar device already used in wind- leased in 1937 [6]. Black’s negative of 1948, in which he specified that mills to regulate the distance between feedback was seminal to subsequent negative feedback occurs when “the millstones, although not on the feed- work done at AT&T by Harold Nyquist information fed back to the control back basis. (1889–1976), who developed the sta- center tends to oppose the departure In 1840, Astronomer Royal George bility criterion for feedback systems in of the controlled from the control- Airy (1812–1878) developed a feedback- 1932 (Nyquist’s stability theorem) [7], ling quantity” [10]. As a matter of fact, controlled motor for a Greenwich tele- and at Bell Labs by Hendrik Wade Bode the negative feedback amplifier played scope to compensate for the rotational (1905–1982), who conceived the epony- a key role in the success of the long- motion of the earth. The design was mous diagram in 1938 that allowed distance telephone and television based on differential equation analy- taking full advantage of Black’s inven- network. And the application of the sis. An early general study of feedback tion [8]. These studies constituted the negative feedback has not been lim- control was proposed in the 1868 paper foundation of information theory, de- ited to telecommunications. Industrial “On Governors” by James Clerk Maxwell veloped in the following years mostly at and military applications as well as ser- (1831–1879), which discussed Watt’s Bell Labs and AT&T by Nyquist, Bode, vomechanism and drives use it widely. centrifugal governor and other devices Hartley, and others, up to pivotal con- Negative feedback formed the basis of by engineers and technicians of the day tributions by Claude Shannon [13]. The theory was further (1916–2001) in the years 1938 developed by other control and 1948. Frederick Terman pioneers, notably English- (1900–1982) of Stanford Uni- Canadian Edward John Routh versity was among the first to (1831–1907), German Adolf Hur- use the negative feedback prin- witz (1859–1919), Russian Alek- ciple in electronics since 1939. sandr Lyapunov (1857–1918), William R. Hewlett (1913–2001) Slovakian Aurel Stodola (1859– and David Packard (1912– 1942), and the great French 1996), two of Terman’s most mathematician Henri Poincaré brilliant students, designed (1854–1912). It seems that if one their audio frequency oscillator wanted to become a control 200A based on negative feed- pioneer, it was convenient to be back, starting the rise of the born in the 1850s! Hewlett-Packard Company [9]. Despite extensive exploi- A general definition of nega- tation during the Industrial tive feedback was given by Revolution, negative feedback FIGURE 2 – The centrifugal governor was invented by James Watt Norbert Wiener (1894–1964) in in 1788.