“There is nothing new under the sun” is a phrase proclaimed in Ecclesiastes. We are at the brink of converting radio and television broadcasting over to digital yet the theories behind these systems have been around for decades.
On a greater scale of scientific study, mathematicians say Isaac Newton knew how to land humans on the moon. He just didn’t have the means to get them there.
At the terrestrial level, the same is true with telephone and wireless communications. Dr. Harry Nyquist developed the mathematical models to digitize audio in the 1920s. In the 1940s a digital telephone system was placed in operation between Washington and London permitting President Franklin Delano Roosevelt and Prime Minister Winston Churchill to converse on a secure line.
There was one drawback to this method of transmission: It took a room full of vacuum tube electronics and about a dozen technicians in both cities to make it work.
Fifty years later the microprocessor and RAM made digital encoding, transmission and reception the preferred method of sending pictures, video and audio.
The same is true for radio broadcasting. Engineers discovered early in the game that amplitude modulation had several limiting factors inherently built into it.
Fig. 1: An extremely linear sawtooth wave was created from the crystal oscillator and applied to the control grid of a triode vacuum tube. With no audio signal applied, the tube is biased to cut off half-way up the slope of the waveform, rapidly dropping the plate voltage and creating a ‘spike’ through a short RC time constant. Carrier change
AM reached its peak when transmitters capable of generating 100 percent modulation came along, creating the greatest audio fidelity and signal-to-noise separation. When Edwin Howard Armstrong brought FM forward, most engineers knew it was a superior system, but it was before its time.
FM had several problems. There was the political battle, with RCA doing all it could to hamper its growth, but there were also technical issues.
AM was and is stable, operating off of a crystal-controlled oscillator; AM is easy to receive. FM requires the carrier frequency to be altered, changed.
FM does not function as often illustrated, like a Slinky expanding or contracting. In reality, these expansions and contractions are changes in the angular velocity of the carrier, creating pairs of sidebands with varying amplitude levels.
The number of pairs of sidebands and their rate of change carry information. Amplitude levels are discarded.
Using the Bessel Function you find that the sum of the sidebands add up to the total power of the transmitter. At times there is little or no power at the carrier frequency.
In the early years of FM, the age of no phase-locked loops and microprocessors, it was difficult keeping these transmitters on the assigned frequency. Many had AFC circuits driving servo motors that mechanically re-tuned the transmitter throughout the broadcast day.
Throughout the 1950s there was really only one reason to keep FM on the air: “storecasting,” using the subcarriers.
In the mid ’50s, most of the FM stations either were owned by universities or were commercial stations simulcasting the AM programming plus a little money-maker called Subsidiary Communication Authorization.
SCA provided background music to businesses and stores. Stations often contracted with companies like Muzak or they created their own service.
Transmitter stability was improved by keeping the main channel modulation low, always keeping some energy at the carrier frequency. Loudness didn’t matter; there were few listeners out there anyway.
A better transmission system was needed. FM needed stability.
Fig. 2: The spikes occur at the same frequency as the crystal oscillator. When audio is applied to the sawtooth wave, the location where the spike occurs is varied depending on the location where the tube is cut off, creating a modulated signal from a crystal-controlled oscillator. Sawtooth approach
Serrasoid modulation was developed in the late 1940s by James R. Day working for Radio Engineering Labs, a company that worked closely with Major Armstrong in the development of FM transmission and receiving equipment. Armstrong endorsed serrasoid FM as a preferred system.
Serrasoid modulation was a development that came from the electronic scanning system developed by National Television System Committee in the 1930s to generate electronic picture scanning. Serrasoid FM created linear sawtooth waves (thus the name) and applied audio to them.
Serrasoid FM actually is a forerunner to digital sampling. Resistive-capacitive timing circuits are used to create the desired outputs.
REL and Gates built these systems through the 1950s and ’60s, reducing their size to a unit about 19 inches square. The Gates unit used nine-pin miniature tubes producing a 10 watt output — often used as a transmitter for Class D educational stations.
An extremely linear sawtooth wave was created and applied to the control grid of a triode vacuum tube. With no audio signal applied, the tube is biased to cut off half-way up the slope of the waveform, rapidly dropping the plate voltage and creating a “spike” through a short RC time constant (Fig. 1).
The spikes occur at the same frequency as the crystal oscillator. When audio is applied to the sawtooth wave, the location where the spike occurs is varied depending on the location where the tube is cut off (Fig. 2), creating a modulated signal from a crystal-controlled oscillator.
When the chain of spikes is fed to a tuned circuit, sine waves are produced. This signal actually is phase modulated and needed some minor adjustments to be converted to FM.
As with most early transmitters, the modulation was done at a low frequency and was narrow-band in nature. The modulated output passed through a series of frequency multipliers to create a wide-band FM broadcast signal. Adding additional subcarriers to this system presented no real problems.
What I find unique about this is how close serrasoid FM is to digital encoding. The crystal oscillator typically operated between 100 kHz and 125 kHz, frequencies well suited for sampling audio.
If the negative spike had been changed to the creation of pulses, digital encoding could take place.
A drawback of serrasoid exciters was poor low-frequency performance. Frequency stabilization was an issue with this system. This along with the introduction of stable PLL circuitry is why serrasoid exciters fell out of favor.
Serrasoid FM was a harbinger of things to come decades later; its exciters were so effective, many broadcasters retrofitted their old transmitters, affording them extra life. This modulation method contributed to the success of FM today.
The author is laboratory director for video technology and communications at the Thomas Jefferson High School for Science and Technology in Fairfax, Va.