Historical Engineers: Harry Nyquist, a Trail Blazer in Digital Communications
Nyquist’s contributions were the intellectual building blocks essential to the birth of modern information theory
Harry Nyquist's contributions were the intellectual building blocks essential to the birth of modern information theory. His work pushed engineering into new realms of digital communications, revolutionizing the field of telecommunications among many others.
In this article, we'll discuss milestones in Nyquist's life and how his theories were cutting-edge for his time.
Harry Nyquist was born in Nilsby, Sweden on February 7, 1889, according to an encyclopedia entry on his life. Nyquist was an able student, but his family could not afford the cost of education beyond the basics. So, at fourteen, he set to work in construction, saved his money, and a few years later, moved to the United States.
Sketch of Harry Nyquist. Image used courtesy of Chester F. Carlson Center for Imaging Science, RIT
He arrived in 1907, and entered the University of North Dakota in 1912, earning a BSEE in only two years. He earned his master's degree a year later. The next stop was Yale, where he completed a Ph.D. in physics in 1917.
Straight from Yale, Nyquist began his career at AT&T, where he stayed until he retired from the company in 1954 at the age of 65. A few years later, AT&T teamed up with Bell Laboratories to advance the theory and practice of communication. It was a perfect fit for Nyquist.
His early work centered on telegraphy. Nyquist observed that the speed of transmission over a line was proportional to the width of the frequencies. In 1924, he published a paper entitled, "Certain Factors Affecting Telegraph Speed" discussing his observations.
Telegraph Transmission Theory
Then, in 1928, Nyquist published his immortal work, "Certain Topics in Telegraph Transmission Theory." It concerned a now-familiar topic: converting analog signals, such as the human voice, to digital ones and zeros.
Nyquist determined that if one sampled an analog signal at a rate at least twice as fast as its highest frequency component, the analog signal would be perfectly reproduced.
Here’s how it works:
A sine wave to be sampled. Image (modified) used courtesy of Dave Marshall
Now, let's measure it (or, “sample” it) one time each cycle, as below.
A sine wave sampled at a rate of once each cycle. Image (modified) used courtesy of Dave Marshall
Sampling the wave of a frequency equal to the wave would make it seem as though we don’t have a wave, but rather, a constant.
Next, let’s sample it at a rate of 1.5 times per cycle.
A sine wave sampled at a rate of 1½ times each cycle. Image (modified) used courtesy of Dave Marshall
Note that the picture that emerges is a sine wave but at a lower frequency. This is called “aliasing.”
Now, let’s sample our sine wave at twice it’s frequency.
A sine wave sampled at a rate of two times each cycle. Image (modified) used courtesy of Dave Marshall
Note that here, we sampled the sine wave twice in each of its cycles. It’s only when sampling at twice the frequency, the Nyquist Rate, that we get a true picture of the wave we’re trying to measure.
A curious fact is that Nyquist’s original purpose in this research was to find out how much information could be sent down a signal line and then be successfully gleaned at the far end.
Monumental in its own right, Nyquist's work inspired Claude Shannon, whose seminal work, "A Mathematical Theory of Communication" can be said to have ushered in the birth of information theory.
To learn more about Claude Shannon's instrumental place in digital communications, check out the historical engineers we previously focused on him.
The Nyquist Plot
Harry Nyquist is also noted for designing stable control systems. Nyquist plots are used to determine the stability of single loop feedback systems.
Single loop feedback system. Image used courtesy of Erik Cheever, Swarthmore College
Robert Keim's article on how to use a Nyquist plot in AC analysis is the first in a series explaining this concept, to which the entire field of control systems is indebted.
Nyquist is also noted for his fundamental contributions in our modern understanding of thermal noise in communications systems, according to Dennis V. Perepelitsa from MIT.
In his later years, he also came up with a crude but effective embryonic version of the fax machine. It was only in these later years with the invention of the transistor that Nyquist's findings in "Certain Topics in Telegraph Transmission Theory" could be put into practice. His principles now appear in the form of the T1 circuit, still in use today.
Many bright engineering minds like Nyquist's never retire. After his formal retirement from Bell Labs, he remained active professionally until his death in 1976.
It is usually at this point that we ask a question regarding how this historical engineer's work has affected your work as an engineer. But the reality is, without Nyquist's contributions, the world of digital communications would not exist, and we’d all still be mired in analog.
Explore Other Historical Engineers
Historical Engineers: Alexander Graham Bell and the First Telephone
Historical Engineers: George Boole, Pioneer of Algebraic Logic
Historical Engineers: James Clerk Maxwell, Founder of the Electromagnetic Theory
Historical Engineers: Michael Faraday, a Founding Father of Electrical Science
Historical Engineers: Harvey C. Nathanson Leaves Legacy as the Father of MEMs
Information about Nyquist's life and work was pulled from Encyclopedia, the New York Times, Science Direct, Swathmore, and Dave Marshall
Featured image (modified) used courtesy of Chester F. Carlson Center for Imaging Science, RIT and Dave Marshall
Which historical engineer would you like to see covered next? And why? Drop your suggestions in the comments below.
Nice article, but unfortunately the encyclopedia article it draws from is far from accurate. Nyquist didn’t preset “digital ones and zeros”, and didn’t generalize the sampling of signals (much less “huam voice”), in that 1928 paper. Nyquist was concerned with the transmission of pulses over a given bandwidth. His work was generalized many years later with the help of Shannon (though it was discovered by others independently in the mean time).