October 1930 Radio-Craft
Wax nostalgic about and learn from the history of early electronics.
See articles from Radio-Craft,
published 1929 - 1953. All copyrights are hereby acknowledged.
Frequency crowding has evidently
been an issue since the early days of radio according to this 1930 article in
Radio−Craft magazine. The situation was really bad in the earliest times
when unfiltered spark type transmitters were the norm. Those pioneers could be credited,
I suppose, with being the first users of wideband communications, but it was not
because they chose to do so. Here author Clyde Fitch discusses the debate over whether
there really were such things as sidebands from modulation (yes, it was questionable
in the day) and makes an argument for their existence based on analysis of various
types of modulation. In particular, he predicts the coming popularity of single
sideband receivers with crystal-filtered channels (also a relatively new concept),
and the need for matching SSB transmitters with... wait for it... carrier and sideband
suppression! He also reviews "The Heterodyne Theory."
The "Stenode Radiostat" System
Fig. 1 - The superheterodyne action above creates a beat-frequency
c, as a transmitter creates a wave d, which is similarly modulated in amplitude.
The "Stenode," on the other hand, uses a frequency-modulated wave e.
The invention which promises to revolutionize radio practice is explained here
By Clyde J. Fitch
Since radio broadcasting struck its stride, the ether has seemed too narrow for
all who wish to put it to use. The system described here promises to make it possible
to put a hundred stations on the air, where there was one before. Yet its adoption
requires not only special transmitters, but costly crystal-controlled receivers;
and will therefore be a matter of time.
Still, it must come; and the alert radio man will therefore desire to familiarize
himself with it at the earliest moment.
The recent announcement that Dr. James Robinson, an English inventor, has perfected
a receiver which gives high-quality reproduction of radio telephony on what is practically
a carrier wave alone (the frequency "channel" required is stated to be but from
20 to 50 cycles wide, instead of the 10,000-cycle channel which is standard in the
United States, or the 9,000-cycle channel used in Europe) has revived throughout
radio circles the discussion of the much-mooted question of "sidebands."
Do sidebands in radio waves really exist? And, if so, what is their nature?
The answer is of the highest practical importance. Throughout the world, nations
and individuals are endeavoring to increase the number of radio stations to the
maximum which is practicable without causing undue interference and, in the case
of broadcast stations, impairment of the tone quality of the reproduced voice and
To explain the points in controversy among radio engineers and physicists, as
well as to describe the principle of the new invention, is the purpose of this article.
What are Sidebands - If Any?
It was once customary, in the early days of radio, to think of a wave as having
only a single frequency, with current and voltage variations proportioned in amount
or amplitude, to the audio-frequency impulses with which it was modulated. But,
after hundreds of stations came on the air, it was obvious that a station of one
frequency interferes with a station on another frequency and that, when a receiver
is tuned so sharply that the interfering station is cut out, something is lost also
from the higher notes of the station to which the set was exactly tuned. We express
this fact by saying that the loss of tone quality is due to "cutting sidebands."
The simple explanation, which satisfied most radio fans who could not go into
the mathematics of the thing, is that: to the carrier-wave frequency there is added
a whole band of frequencies on either side, which are equal numerically to twice
the highest audio frequency used in modulation. That is to say, if a 1,000-kilocycle
carrier wave is modulated by audio frequencies up to 5,000 cycles, there results
a compound wave varying in frequency from 995 to 1,005 kilocycles and, consequently,
10 kilocycles wide. The five-kilocycle bands of frequencies, above and below the
carrier frequency, were known as "upper and lower sidebands."
However, many radio engineers have never been satisfied with such an explanation;
and others, admitting the existence of sidebands, saw no reason why they should
be necessary in broadcasting. Sir Ambrose Fleming, inventor of the two-element vacuum
tube, and a noted mathematical authority, doubts as to whether the sideband is anything
but a convenient mathematical assumption, or has any real existence. Sir Oliver
Lodge, the dean of radio research, says that a deep philosophical question seems
to be involved.
However, let us consider the matter in the simplest possible manner:
The Heterodyne Theory
Suppose we forget our old ideas of "modulated radio waves" and look at the situation
from a different viewpoint. Take the superheterodyne; we know that the superheterodyne
works, and we claim to understand its theory. Well then, apply the same theory to
the radio broadcast station and our problem is partly solved. This idea was described
by the writer, at some length, in the November, 1929, issue of Radio-Craft.
In accordance with the heterodyne theory, which has been verified by the action
of the many superheterodyne receivers now in use, when two frequencies a and b are
combined, two others or "beat" frequencies are produced, namely, a+b, and a-b. Harmonics
may also appear; but they are of no importance in this discussion.
It will be observed, when analyzing the superheterodyne circuit, that the beat
frequencies a+ band a-b appear only after the frequencies have been passed through
a detector - the first detector. This is shown in Fig. 1. Here the frequencies
a and b are combined or mixed to produce the composite effect shown at c. After
passing through the first detector, the beat or difference-frequency a-b appears.
For simplicity, the sum-frequency a+b is omitted; as, usually, only one beat note
of the superheterodyne is employed.
We can apply this theory to the broadcast transmitter, and thereby explain its
action clearly. In the transmitter, we have the carrier frequency, a' combined with
the audio frequency b', producing the composite effect shown at c' (the frequencies
are added algebraically on a graph paper) and this, in turn is converted into that
shown at d by means of the limiting action of the transmitter's oscillating tubes,
In other words, these tubes also perform a function similar to that of the first
detector in a superheterodyne. The wave that leaves the transmitting aerial is therefore
of the shape indicated at d. This is in accordance with observed facts.
Now if you will compare them, you will note that the wave shape at d is the same
as that at c. Since c is composed of two radio-frequency waves, so also is d. Therefore
d must have, in addition to the carrier frequency, another radio-frequency or "sideband"
This reasoning, at least, ought to prove that sidebands exist. The next question
is, are sidebands necessary?
Of the four frequencies generated at the transmitter (namely the audio frequency,
the carrier frequency, and the two sideband frequencies) all except the audio frequency
are radiated into space. This is because audio frequencies do not radiate very far
and the antenna system is not tuned to the audio frequency. Therefore, three frequencies
reach the receiver in a form something like the composite group indicated at c.
lf we should reverse the process, and subtract the various sideband waves from
the carrier wave (d or c) the result would be a single "sine wave," representing
the unmodulated carrier wave of the station.
Therefore, if we took the sidebands away from the carrier by means of suitable
filter circuits, the result would be an unmodulated carrier wave. In other words,
removing the sidebands would demodulate the wave. Partial removal (as in a highly-selective
receiver) would partially demodulate the wave, and result in poor tone quality.
This is actually the case; there is no question about it.
It is safe to say that a pure sine wave has no sidebands, or harmonics. Altering
the shape of the wave, no matter how, is the same as adding sidebands, or harmonics.
Removing these sidebands or harmonics by means of filter circuits, brings the wave
back to its original pure sine-wave shape.
Carrying our heterodyne example farther, at the receiving set both the upper
and lower sidebands heterodyne with the carrier frequency, and produce beat-notes,
which are intercepted by the detector and represent the original audio frequencies.
Special Transmission Systems
We can even suppress one of the sidebands at the transmitter; then let the other
heterodyne with the carrier at the receiver, and produce the audio frequencies without
impairing the tone quality. A single-sideband system of transmission was described
in the June, 1925, issue of the Proceedings of the Institute of Radio Engineers.
We can even suppress one sideband and the carrier frequency also, at the transmitter,
and radiate the other sideband only. In this case a new "carrier" frequency must
be generated at the receiver, by means of a local oscillator, in order to produce
the heterodyne effect with the received sideband. This oscillator-frequency must
be exact; if it is too near the nearest sideband frequency, the audio-frequency
beat notes will be too low. They will sound something like running a phonograph
record slowly; except that the length of time required to complete the message is
not reduced. If the oscillator frequency is too far from the nearest sideband frequency,
the audio-frequency beat notes will be too high and a "tinny" sound will be the
effect. If the frequency is on the wrong side of the sideband, the high audio notes
will be low, and the low audio notes will be high. This is what is called "inverted
speech"; the system has been demonstrated and described by Bell telephone engineers,
and has been used in transatlantic telephony to obtain secrecy.
Now comes the demonstration of Dr. Robinson's "Stenode" (narrow path) system;
and it has been hailed as disproving the existence of sidebands by its operation,
apparently without them. However, the receiver, which is demonstrated together with
a transmitter made by the same inventor, depends upon the principle of "frequency-modulation"
rather than that of "amplitude-modulation," which we have described as producing
The idea of frequency-modulation is an old one - so old that patents issued on
such a system have already expired. But there was then no commercial opportunity
for its use. Several experimenters have worked on the possibility of a crystal-controlled
receiver which would make possible reception of a greater number of stations, each
on a frequency corresponding exactly to one to which the receiver would be held;
this was discussed at the time of the institution of the Federal Radio Commission.
But Dr. Robinson has incorporated the crystal control of the receiver into an intermediate-frequency
amplifier, and thereby is able to receive many frequency-modulated transmissions
with a single crystal.
Fig. 2 - This diagram indicates the essential parts of the
"Stenode" transmitter, the action of which depends on the variation of the capacity
of the crystal Q and modulation thereby of the transmitted frequency. (Adapted from
Fig. 3 - The "Stenode" receiver shown differs from ordinary
superheterodynes in the circled filter, which accepts frequency variations not exceeding
a hundred cycles. By the process indicated in Fig. 4, these are turned into
a full band of audio frequencies.
Referring again to Fig. 1, at e' we have the frequency-modulated wave; compare
it with the amplitude-modulated wave d. In one, the current is constant in frequency
and varies in amplitude or height. In the other, the current remains constant, while
the frequency is varied; as shown by the varying distance between successive peaks.
How this is accomplished is shown in schematic form in Fig. 2, representing
a portion of the Stenode transmitter. The speech currents, impressed through the
microphone input on the amplifier, produce corresponding mechanical vibrations of
a plate, which acts like a loud-speaker diaphragm operated by the electromagnet
The width of the air gap separating the plate from the quartz crystal Q is therefore
varied at audio frequencies; and this affects the period of vibration of the crystal,
which is the master control of the oscillator tube. With this system, a very slight
frequency-modulation is needed - only about 1/300th of one per cent., as will appear.
The variations of the crystal's oscillating frequency are in this order.
At the receiving end, we find the superheterodyne circuit of Fig. 3. As
stated above, the purpose of employing a double-detection circuit is to use a single
quartz crystal to govern reception at all broadcast frequencies. Otherwise, it would
be necessary to have a separate crystal for every station to be received-a physical
and economic absurdity. In addition to this, it is almost impossible to bring two
crystals so exactly into unison as would be required at transmitter and receiver.
The quartz crystal circuit acts as a highly-selective filter or "gate" (hence
the name Stenode ) through which only a very narrow band of frequencies may pass
- less than a hundred cycles wide. If the received wave is of constant frequency,
no sound will be heard from the speaker; but if it is varied in frequency, at intervals
comparable to the periods of audio frequencies, a response will be heard.
Consult Fig. 4. Here we have the selectivity curve of the quartz crystal
of the response is plotted against frequency. At the bottom, T indicates the signal
wave-train, which has been frequency-modulated at an audio frequency indicated by
the curve A. The response of the crystal, to this variation in radio, frequency,
is to pass current in the manner indicated by the curve W. This modified signal
is then fed into the grid circuit of the second detector tube, as indicated at the
The result is that the presence of the crystal has changed the frequency-modulated
waveform of the input into an amplitude-modulated wave W; which is detected in the
usual manner and converted into a suitable audio frequency for reproduction.
Fig. 4 - The frequency-modulated wave T corresponds to audio
modulation A. Passing through a crystal, with response characteristic C, it
is weakened in proportion to its variation from the frequency f3, giving the amplitude
variations shown in W.
It will also be seen that, if the frequency of the transmission varies too much,
or shifts beyond the points f1 or f3, the station will fade. This
would appear to be the case in actual practice. One of the greatest problems of
the Stenode Radiostat is to maintain a constant average frequency at the transmitter.
Slight shifts in frequency will impair the audio quality. The heterodyne oscillator
at the receiving end must be carefully adjusted and maintained constant, in order
to keep the intermediate frequency at the proper average value - f2 in
Fig. 4. These problems are being, or have been, solved by Dr. Robinson.
Delicacy of Tuning
The extreme selectivity of this set has made it necessary to incorporate unusual
tuning controls. The ordinary type of vernier-dial control is entirely too crude;
the slightest touch would tune the station in and out. Therefore, a single-plate
condenser has been placed in parallel with the main tuning condenser; and this single-plate
condenser is equipped with a slow-motion control, so that one complete turn of the
knob varies the capacity only 1 mmf. Even with this delicacy of tuning, a slight
turn will pass completely by a station, from silence to silence; for the width of
the band passed by the quartz gate is not over 50 cycles - or one two-hundredth
of that used in the ordinary broadcast receiver. This extreme selectivity, of course,
is due entirely to the quartz crystal, and not to the fact that a superheterodyne
circuit is employed. For, if the quartz is removed and a piece of inert material,
such as glass, substituted, energy passes through (by virtue of the electrostatic
capacity of the electrodes when this capacity is not neutralized); and the set functions
as an ordinary superheterodyne, with the degree of selectivity common to superheterodynes.
When the quartz is in place, however, it acts as a gate, allowing only the narrow
20- to 50-cycle band to pass. To prevent capacity coupling through the electrodes,
a balanced-bridge circuit is employed. In this bridge circuit the electrode capacity
is neutralized by the condenser NC2; so that the only energy that can possibly reach
the second detector must pass through the quartz, by piezoelectric action. In Dr.
Robinson's laboratory models, an intermediate frequency of 107 kilocycles is employed.
It is evident that the ordinary, broadly-tuned receiving set will not give reproduction
when tuned in to a frequency-modulated station. But the question is, will the Stenode
receive an amplitude-modulated station? Theoretically, no. But tests prove otherwise,
and we are interested in facts. These show that, in the modern amplitude-modulated
station, a slight frequency-modulation also exists. And this slight frequency-modulation
is sufficient to actuate the Stenode. The test, however, would not prove that sidebands
are not necessary.
One of the main points of this article is to show that the results obtained with
the Stenode Radiostat fail to prove that sidebands do not exist or are unnecessary.
It is true that the Stenode receives ordinary stations with a great degree of success
- and it receives those stations whose frequency is controlled by quartz crystals
much better than the others, which fade, or are hard to tune in. What the tests
do prove is that the ordinary type of station, even though quartz-crystal controlled,
is frequency-modulated as well as amplitude-modulated. The slight frequency-modulation
which exists is due, probably, to some unexplained circuit conditions.
The amplitude-modulated component of the wave could be entirely eliminated; its
existence is detrimental, since it causes the low notes to be disproportionally
amplified; since the quartz circuit is one of low damping and the persistence of
the lower notes is considerable. Hence, a special filter is used in the Stenode's
audio amplifier to compensate for this low-note frequency distortion.
How about sidebands when using a frequency-modulated transmitter? Primarily,
we radiate waves ranging in frequencies from f1 to f3, (Fig. 4)
a difference of 20 to 50 cycles. This band of waves we know to exist. But, since
changing the wave-shape adds sidebands, or harmonics, as stated previously, other
frequencies may be introduced. To what extent, if any, these widen the band is a
problem for the mathematicians. Dr. Robinson's tests indicate that the effect must
be negligible. At least, the improvement over the old method is enough to allow
some hundred times as many stations to transmit at the same time without interference.
How soon the system will be placed in commercial use, is difficult to predict.
It would necessitate revolutionary changes in all receiving sets, as well as transmitters,
and the expense would be enormous. Television, which requires the greatest space
in the ether, may be the first to benefit by this system.
Posted July 31, 2023
(updated from original post