Tubes, randomness, and Brownian motions: or, how engineers learned to start worrying about electronic noise

Abstract

In this paper, we examine the pioneering research on electronic noise—the current fluctuations in electronic circuit devices due to their intrinsic physical characteristics rather than their defects—in Germany and the U.S. during the 1910s–1920s. Such research was not just another demonstration of the general randomness of the physical world Einstein’s work on Brownian motion had revealed. In contrast, we stress the importance of a particular engineering context to electronic noise studies: the motivation to design and improve high-gain thermionic-tube amplifiers for radio and wired communications. Engineering scientists’ endeavors to understand electronic noise started in 1918, when Walter Schottky at Siemens formulated a theory of “shot noise,” current fluctuations owing to the random emissions of discrete electrons in a tube. Schottky’s theory was revised and experimentally tested at Siemens, General Electric, and AT&T during the 1920s, leading to the discoveries of several other types of noise and an increasing interest in the thermal fluctuations in electronic circuits. In 1925–1928, J.B. Johnson and Harry Nyquist at Bell Labs developed a theory of thermal noise for any electrical resistor at a non-zero temperature. Although these studies were initiated to chart the fundamental performance limit of electronic technology, they ended up assisting the empirical determination of individual electronic components’ characteristics.