Vintage Radio: The Life, Decline and Possible Rebirth of AM

Mar 16, 2016

Last month we looked at contributions to the art made by amateur operators, in particular advancements in Amplitude Modulation, or AM, and how it came to give radio its voice. This month, we will look a little deeper into AM, its history, how it works, the corporate politics at its heyday and where it is going.

In the beginning, there was Morse code. After the discovery of radio waves and how to create them, came the question of how to use them for communication. If a transmitter could be switched on and off with a telegraph key, then it could transmit long and short bursts of radio waves corresponding to the long and short dashes and dots of Morse code.  

The receiver would detect these bursts and convert them to audible dashes and dots that the operator could hear. Given how robust a Morse signal is, the large commercial traffic companies like Marconi’s were quite satisfied. However, amateur operators using Amplitude Modulation (AM) to inject audio into a radio signal, not to mention the general public listening to them, were more interested in audio that they could understand, be it voice or music. Thus began the early roots of broadcasting.

But why AM?

Above left: A 630 kHz radio wave as depicted on an oscilloscope. Above middle: An unmodulated radio signal. Above right: The same radio signal as in the middle but modulated by the audio signal on the lower trace.
Credit P. Litwinovich

Pictured above left is a 630 kHz radio wave as it would display on an oscilloscope. If you slow down the trace speed of the oscilloscope into the audio range, so many cycles are displayed that they look like a solid band. This is what an unmodulated radio wave signal, referred to as a “carrier,” would look like. The lower trace has been added to display an audio signal which when present, will be added to the carrier. The image on the right shows what happens to the 630 kHz carrier when a 1000 hz audio tone is used to modulate the radio signal. The audio causes the strength of the carrier, or its amplitude, to increase or decrease corresponding to the applied audio.

AM is natural. When you raise or lower your voice, you are changing the amplitude of your own audio, so to speak.

AM occurs elsewhere in nature. A lightning strike or manmade electrical discharge will produce a burst of electrical noise that varies in amplitude. Since AM radios are designed to detect variations in amplitude, this is why they are prone to interference from such things. AM held sway as the primary method of modulating a radio wave up to WWII, not only for broadcasting, but for all types of radio communications. 

Pre-WWII RCA Console Radio
Credit P. Litwinovich

Every vintage consumer radio, be it standard broadcast or shortwave, up to WWII, received amplitude modulated signals. Nowadays, AM broadcast stations are associated with lower quality audio, but such was not always the case. Receiver design really came of age in the 1930s with the superheterodyne circuit and advancements in loudspeaker design. The grand floor consoles of the late 1930s leading up to WWII were capable of producing audio that was very good, even by today’s standards, the only exception being that they were monaural, as stereo technology was still a ways off.

The politics of AM vs FM 

The first mention of Frequency Modulation (FM) occurs in a paper published by Edwin Armstrong in 1936 (Ref 1). Armstrong is credited with inventing Frequency Modulation, and received patents to that effect.

David Sarnoff, CEO of RCA, had other ideas. RCA was heavily invested in AM radio, which was big business at the time. Not only did RCA produce radios, but received substantial royalties for licensing various AM-related patents. They also had a stake in networks of AM stations. Sarnoff is credited with saying, "I thought Armstrong would invent some kind of a filter to remove static from our AM radio. I didn't think he'd start a revolution -- start up a whole damn new industry to compete with RCA" (Ref 2). 

Sarnoff and RCA did everything possible to interfere with Armstrong’s efforts. Sarnoff jumped on a paper published by John Renshaw Carson, a researcher at Bell Laboratories, who “proved” mathematically that FM would produce an infinite number of sidebands (frequency products) and therefore would not work. Armstrong did not buy the argument and went on to prove that FM does work.

After FM investors had built a number of stations in the 40-50 mHz band, Sarnoff tried to kill FM a second time by convincing the FCC that the 88-108 mHz band would be much better. Although his intent was to waste the money that competitors had spent on the original effort and transmission gear, the plan backfired when it was quickly discovered that propagation characteristics of the 88-108 mHz band were indeed much better suited for FM broadcast. The new band also had room for many more stations.

The actual space used by a modern AM radio station
Credit P. Litwinovich

The decline in AM audio was due more to regulation than to method of modulation. One aspect of radio not understood by most listeners is the concept of occupied bandwidth, or the amount of spectrum that a station uses to transmit its signal. The picture above is the display of an instrument known as a spectrum analyzer. It displays signals within a specified range of frequencies. Here we are looking at a small portion of the AM band, from 580-680 kHz, with each vertical line spaced 10 kHz apart. 

Imagine it as the dial of your radio, with each vertical line representing the possible location of an AM station. In the left picture, our 630 kHz station has no audio and takes up just a sliver of room.  When you mix two signals together, such as audio and a radio frequency signal, you generate additional frequencies that are the sum and difference of both. These products appear on each side of the main frequency and are referred to as sidebands.

The picture on the right shows the same station modulated by music that is limited by a filter to 10 kHz.  The station actually requires a bandwidth that begins at 620 kHz and ends at 640 kHz; in other words it uses a total of 20 kHz of space. Based on this, the next closest station would need to be on 600 kHz or 660 kHz to avoid interfering with each another. Actual FCC rules require four channels of separation, placing the closest stations at 590 kHz or 670 kHz.  

AM is an efficient user of spectrum compared to FM or TV, as the occupied bandwidth is directly proportional to the frequency of the audio imposed on it. The average human can hear frequencies from about 20 Hz to 20 kHz. As we age, the upper limit tends to fall to about 10 kHz. When there were only a few radio stations in any given area, they could afford to modulate with frequencies up 20 kHz, which is better than a modern FM station. As more and more stations went on the air, and had to be fit closer and closer to each other on the band, they began to interfere with one another. Eventually the FCC had to restrict modulation to 10 kHz to fit them all in. This still was not bad though.

Around 1990, in an effort to further reduce interference between and to allow yet closer placement of AM stations, the FCC accepted the recommendation of the National Radio Standards Committee (NRSC) to limit AM audio to 7.5 kHz. This caused significant loss of fidelity in AM station audio quality. While acceptable for news and talk, it made music sound flat and dull.

To make matters worse, some stations broadcasting with the In Band On Channel (IBOC) digital audio system have voluntarily reduced their analog audio to 5 kHz to make room for additional digital bandwidth. You can easily identify the ones doing this; their audio sounds like first generation internet audio, similar to telephone grade audio.

FM uses much more bandwidth. The variation in frequency of an FM station must include allowances for the frequency of the audio and the volume of the audio, plus the multiplex stereo encoding. A single modern analog FM signal, which can reproduce frequencies to 15 kHz, requires about 200 kHz of bandwidth, the same space as 20 AM channels, or with modern spacing, about 5-8 AM radio stations. A single TV channel, analog or digital, requires 6 mHz of bandwidth, enough space for the equivalent of 5 entire AM broadcast bands.

Credit P. Litwinovich

Besides audio quality, there are other major problems plaguing AM stations.


In the early days, interference primarily came from a nearby motor, such as a vacuum cleaner, passing electric train or the occasional thunderstorm. Nowadays, nearly all electronic devices, computers, light dimmers, computer controlled appliances, gas discharge and LED street lights, and many more radiate noise that can interfere with AM reception.  

Cheap low-quality radios, the kind that Atwater Kent was quoted as refusing to manufacture, contribute greatly to the problem. Many manufacturers no longer put any effort into designing radios with quality AM sections in them. The selectivity and sensitivity of many of them is an outright disgrace as far as I am concerned. Modern radios can contain digital audio processing filters that effectively remove static if -- and it is a big if -- the manufacturer wants to add a bit more cost to the radio to add such circuitry.


In an attempt to remain financially viable, many AM stations have resorted to brokered time, infomercials, shock jocks or an endless stream of canned satellite programs.


The current hybrid of analog and digital IBOC signals has worked fairly well on FM, but not so much on AM, where it has resulted in more interference, especially at night.

So is there any hope for the future of AM?

Maybe. The FCC is taking steps in what is known as the AM Revitalization Act. Where available, many AM stations will be allowed to add a small FM signal to cover part or all of their coverage area. Rules regarding distance separation, many of which were written in the 1920s when receivers were not very selective, are being changed. This will allow many stations to increase power, helping overcome some of the noise issues. It will now be easier to move a station to an area where there is less competition or can be better served by the station. In recent years, the FCC has tightened Part 15 rules, which specify limits to how much interfering noise can be generated by a non-radio device such as a computer. But with current budgets, this has gone largely unenforced as have efforts to encourage manufacturers to build radios to higher standards.

Other suggestions have been made. Get rid of the 50 and 100 kW clear channel stations. Originally they were needed to reach deep into rural America where no other service existed. Now many are run like local stations. Someone 200 miles away does not need to hear traffic and weather reports for the city that these stations could serve well with 10 kW. When small stations go dark from too much competition, don’t reissue their licenses; thin the ranks to what the market will bear. Either go all digital, or all analog: the hybrid mix is not living up to expectations. And of course, better programming.

AM broadcasting is 100-year-old technology. Not a bad run for any technology, but considering how much it gave us, I would hate to see it go.

Collector’s tip: An AM radio from the 1930s though the mid-1960s, properly restored, will in general sound better and, with the exception of a few modern high-end receivers, perform better than most modern radios. Its sound quality will only be inhibited by the reduced bandwidth of modern AM stations.


1. Armstrong, E. H. (May 1936). "A Method of Reducing Disturbances in Radio Signaling by a System of Frequency Modulation". Proceedings of the IRE (IRE) 24 (5): 689–740. doi:10.1109/JRPROC.1936.227383.