February / March1932 Short Wave Craft
Wax nostalgic about and learn from the history of early electronics. See articles
from Short Wave Craft,
published 1930 - 1936. All copyrights hereby acknowledged.
These days we would
hardly think of electromagnetic radiation in the 5 cm wavelength realm as being
"quasi-optical" as far as circuit-based manipulation is concerned. Optical wavelengths
begin at around 6,300 Å for red light, which is 6.3x10-5 cm,
or 630 nm. The 5 cm wavelength used an example in a 1932 article in
Short Wave Craft magazine is equivalent to 6 GHz. 6 GHz was an
extraordinarily high frequency to be using for communications back then, and the
author did not intend to liken it to anywhere near visible light. Instead, his terming
its properties as "quasi-optical" referred to how the waves interacted with physical
objects; e.g., reflection, refraction, absorption, and scattering. Barkhausen
oscillations were a popular subject of the era, as I pointed out recently in the
article "The Spook - Another Weird Effect to Haunt TV."
Quasi-Optical Short Waves - Electron Oscillations
Fig. 1 shows path of an electron between filament and
negatively charged plate in an electron oscillator.
Fig. 2 - Hook-up of simple "electron oscillator."
Fig. 3 - How to construct ultra-short wavemeter.
Fig. 4 - Schematic diagram of an electron oscillator.
By C. H. W. Nason
What is the wavelength of a '99 tube? What is the basic action occurring in an
electron oscillator? What conditions or factors govern the frequency of such an
oscillator? How are the signals from such an oscillator received?
The waves from 10 meters down to 5 centimeters are often referred to as the "Quasi-Optical
range" (the inference may be gleaned from the dictionary they are "near-light" waves).
These waves may be treated in many cases as light waves - particularly in the shorter
ranges, where they may be reflected from metallic bodies having either plane surfaces
or specific curvatures for directive transmission or reception. The waves have been
acted upon by lenses made from dielectric materials (Bakelite for example) in such
a manner as to demonstrate the fact that they closely resemble in behavior light
waves, more usually thought of under the classification of "Optical."
By means of special circuits, devised for the utmost of simplicity and efficiency,
the normal vacuum-tube oscillator may be employed with good effect at wavelengths
as low as 1 meter! Below this the stray tube and circuit capacitances are much in
evidence, and the classical circuits are rather hopeless. In the lower range - speaking
now in terms of wavelengths rather than frequencies, to simplify matters - the "electron
oscillations" of Barkhausen and Kurz, and of Gill and Morrell, are most effective.
Although these oscillations are known to the physicist, very little may be found
regarding them in standard works on radio. It is necessary, therefore, that we first
consider the mechanics of these oscillations, before entering into a more detailed
The Mechanism of the Electron Oscillation
Before going further, it is well to state that the most satisfactory tubes for
use in these oscillators are those having concentric elements - cylindrical plates,
etc. These are the '99, the '27, the '52 (more ambitious, of course) and certain
of the tubes provided the government by various organizations during the war. Karplus
in the General Radio Experimenter for May, 1931, indicates success with the G.E.
"CG-1162" which is available from many radio salvage organizations. Electron oscillations
may be obtained with other tubes specially developed for the service but, quite
naturally, our interest rests with those tubes available for experiment.
With the electron oscillators, the determination of frequency no longer rests
upon the inductance and capacitance values of a tuned circuit, but rather on the
electrostatic forces acting upon the individual electrons given off by the filament,
and the resultant time element. Fig. 1 indicates the inter-electrode spacing
of a triode (3-element tube) of the usual character. Note that the grid is positive
and the plate negative, with respect to the filament. An electron, leaving the filament
or cathode, is accelerated toward the grid by virtue of the grid's positive potential.
The majority of these accelerated electrons will pass through the grid's mesh and,
by virtue of their momentum, will travel onward toward the plate until they reach
a point where the negative charge on the plate is sufficiently effective to halt
their flight: they will then assume a backward path, toward the grid. They rejoin
then the other electrons passing toward the grid, following a path somewhat as indicated
in the figure. The length of the path taken and the initial acceleration are the
criteria for determining the frequency of the cycle. The actual A.C. voltages, making
up the oscillatory energy-cycle, are induced by changes in the grid and plate charges
created by the moving electrons.
The original formula of Barkhausen covering the wavelength of the oscillations
- barring factors too complex for inclusion in our discussion - is as follows:
where "d" is the distance between electrodes and "E" the voltage. (The equation
is for the original two-element tube of Barkhausen and Kurz, and not for a triode.)
The Barkhausen oscillations are independent of the circuit constants, to all
intents and purposes.
Circuits for Producing Electron Oscillations
Fig. 5 - Shows antenna placed in focus of a parabolic reflector.
Fig. 6 - Arrangement of directive aerials with reflectors.
Fig. 7 - Receivers for laboratory work may be quite simple,
as shown in diagram above.
Fig. 8-A - How apparatus is connected to make a super-regenerator
"receiver" for electron oscillations.
Fig. 8-B - Receiving circuit for electron oscillator signals
(wave lengths such as 15 inches).
Fig. 2 shows the circuit arrangement of an oscillator for producing the
Barkhausen effect. The Gill-Morrell oscillations are true electron oscillations,
but their frequency is determined by the distance "d" between the elements and the
short-circuiting condenser "C." The change between the two types of oscillations
may be effected at will by altering the circuit conditions. In the Gill-Morrell
effect, the oscillation is due to the timing of the electron's orbits by the oscillating
circuit formed by the Lecher wires. The Gill-Morrell oscillations are much stronger
and are preferable to the simple Barkhausen type. The transition may readily be
obtained by setting the distance equal to one-half the desired wavelength, and adjusting
the voltages for the maximum oscillation. The Lecher wires should be calibrated
directly in centimeters to check the wavelength - remembering of course that the
accuracy is not great. Fig. 3 shows a "trombone" wavemeter for rough measurement
of the emitted wave; this is useful in determining the transition point between
the two effects. By replacing the milliammeter with a crystal detector and phones,
the device may be used to monitor modulated signals.
In Fig. 4 there is illustrated a more complex arrangement of the original
figure, showing the oscillator circuit. To this, it will be seen, there has been
added an antenna; this should be positioned exactly one-fourth of a wavelength away
from the bridging condenser. This places the antenna approximately in the center
of the Lecher wire, where the Gill-Morrell oscillations are used, and at an indeterminate
distance, depending upon the wavelength of the oscillations generated. The antenna
is formed by two copper or brass rods clamped to the Lecher wires by a movable slide.
They should be 1/4-wavelength long, and might be "tromboned" for ready variation.
The grid-current meter should be of the order of 0-100 milliamperes while a 0-1-ma.
meter may be used in the plate circuit. The. oscillatory current may be measured
by a 100-ma. thermal milliammeter; the condensers are 0.001-mf. mica units. The
R.F. chokes are simply wound from annunciator wire on a broom handle, and slid off.
They are somewhat like the pretty curlicues that we used to employ to connect up
buzzers and what not, before we had "wireless" to play with. In tuning the transmitter,
the maximum oscillation is indicated by a maximum reading of the plate milliammeter.
The antenna may be backed by a parabolic reflector of the type shown in Fig. 5,
with the antenna situated at the focal point. (See the preceding issue of Short
Wave Craft - page 254, Dec.-Jan., for details, of a parabolic curve.) Other types
of directive antennas may be employed by the "ham" desirous of going deeply into
the operation of the system. A reference to the article by Yagi, in the June, 1928,
I. R. E. Proceedings will yield much data on the use of directive antennas at such
short wavelengths. The directive effect of the reflector may be greatly increased
by employing an arrangement such as that indicated in Fig. 6, at both transmitter
and receiver; the rods used in the director chain are 1/2-wavelength long, and spaced
1/4-wavelength apart. The metallic reflector may be replaced by a system of five
1/2-wave length rods arranged in a parabola, at the focus of which the antenna is
Receivers for Electron Oscillations
Because of the relatively low frequency stability obtained, little success will
be achieved with electron oscillators for communications where straight C.W. is
employed; although with A.C. on the filament, the hum modulation will be so great
as to alter these conditions by creating an interrupted continuous-wave effect.
Receivers for intramural (laboratory reception) work may be quite simple - as shown
in the two circuits shown in Fig. 7. It is also possible to achieve high sensitivity
in the receiver system by means of the "super-regenerator." Such a circuit arrangement
is shown in Fig. 8-a.
The most practical arrangement is that of employing another electron oscillator,
almost identical with that employed as it transmitter. Indeed, a "changeover," of
such type that a single oscillator may be used for both transmission and reception,
may readily be effected.
The most logical arrangement is that shown in Fig. 8-b, where the circuit
constants are clearly indicated. The antenna is approximately 1/4-wavelength in
dimension, and attached directly to the plate of the tube. Here it might be mentioned
that any good tube socket may be used, and that "de-basing" of the tubes (as usual
when extremely short waves are desired from the more usual oscillators) is unnecessary.
The output of the receiver may be taken by means of a pair of phones, or by a transformer
feeding a standard A.F. amplifier in the plate circuit of the receiver, as shown.
The best tube to be used in the receiver is perhaps the '99, because of the extreme
Phone Modulation With the Electron Oscillator
Telephonic modulation of the electron transmitter is achieved in the plate circuit,
by substituting a modulation or microphone transformer for the phones shown in the
receiver schematic. It is not necessary to provide a speech amplifier, although
it is best to do so where long-range operation is desired. It should be remembered
that a variation in the plate voltage of the oscillator effects a frequency change
rather than - or as well as - an amplitude change. The modulation achieved is not
so perfect, therefore, as in the case of the usual oscillators. Telephone communication
has been achieved up to distances of about twenty miles with "electron transmitters,"
and telegraphic communication is possible over much greater distances. When we consider
the fact that all possible forms of modulation involve a frequency shift, it is
surprising that good quality can be obtained. Nevertheless, the quality of speech
is quite good.
Tubes and Voltages to Be Used
The following table gives some idea of the voltages to be applied for various
tubes and the oscillation wavelength to be expected. All tubes of a given class
do not function as electron oscillators and many tubes must be operated with trick
filament potentials. With the '27 the filament voltage should be rather low; whereas,
in the case of the '99, best results seem to be obtained when the filament is completely
deactivated and operated at a high voltage. The grid current is high in all cases,
and the frequency range may be limited by the voltage which can be applied without
melting the grid - or by the effects of grid emission.
45-50 cms. (17.7 to 19.6 inches)
40 to 75 cms. (15.7 to 29.5 inches)
40 to 75 cms. (15.7 to 29.5 inches)
The whole outfit may be laid out on a breadboard, with Lecher wires about one
meter (39.37 inches) long, made from heavy copper rod, mounted on G.R. stand-off
insulators. Copper clamps spaced with Bakelite strips may be used to provide riders
for the short-circuiting meter, or condensers, so that they may be readily slid
along the Lecher wires. The plate and grid voltages should be made continuously
variable by the use of potentiometers, and the necessary meters should be provided
for taking readings.
All work with electron oscillations is of a highly experimental nature, and no
specific data can be provided with a sure-fire operation guaranteed. The experimenter
undertaking this work should have had considerable experience with radio equipment,
if any hope of success is to be held out to him; it is no game for the tyro.
Posted May 21, 2020
(updated from original post on 7/12/2015)