Microwave Pulse Modulation
April 1946 Radio News Article

April 1946 Radio News

[Table of Contents] Wax nostalgic about and learn from the history of early electronics. See articles from Radio & Television News, published 1919-1959. All copyrights hereby acknowledged.

Try Googling 'cyclodos' and 'cyclophone' and see what you come up with. I found that Cyclodos is now a German company which makes apparel from recycled inner tubes and tents (among other things), and cyclophone is a weird bicycle-mounted contraption for blasting sound while peddling down the street. In 1946, the terms cyclodos and cyclophone referred to modulator and demodulator tubes, respectively, for pulse-time modulation applications. Fortunately, the science of pulse modulation quickly evolved past such devices. This article goes into quite a lot of detail on the beginnings of pulse modulation techniques developed for radar systems during World War II. It is very informative without going into the gory details of equations that govern the theory.

Microwave Pulse Modulation for Ham Communications

By Robert Endall

A simple analysis of a wartime development which will have widespread ham application in the u.h.f. and s.h.f. bands.

In the past, radio amateurs have always been in the forefront of technical progress in experimental communication on the higher frequencies, and many of the advances in high-frequency communication are traceable directly to amateur activity. Before the war many hams were spending much of their time on the 56- and 112-mc. bands and experimenting with higher frequencies. Wartime progress has so greatly extended the radio frequency spectrum that communication frequencies are now regularly assigned and may be expected to carry useful traffic up to 30,000 mc. Many radio amateurs, especially those who have for the past five years been active in wartime u.h.f. and microwave research, will now want to go on the air at carrier frequencies well up in the microwave region.

The cost of ultra-high frequency equipment, the expense of tubes for generating microwave power, and the difficulty of modulating at u.h.f., have up to now been a handicap to widespread amateur experimentation on these frequencies. The intensive military program of development and production of high frequency equipment has done much to eliminate these difficulties. The application of mass production techniques has greatly decreased the expense of equipment.

The development of disc-seal tubes, magnetrons and klystrons, has greatly simplified the problem of generating microwaves. Now the new technique of pulse-time modulation, which has recently been removed from the secret list by military authorities, introduces a considerable simplification in the problem of modulating microwave transmissions.

A pulse-time-modulated carrier consists of signals of the simplest type - short pulses of r.f. having constant shape and amplitude with variable timing. Because of bandwidth considerations, this type of modulation has its maximum usefulness at microwave frequencies. Although developed primarily for commercial multiplex telephone transmission, pulse-time modulation has distinct advantages which suit it particularly well to amateur communications:

1. Since the carrier is modulated only by being switched on or off to create the pulses, the circuit can be of an extreme simplicity of design not possible with other systems.

2. By making full use of the much wider bands per channel which are not only available, but definitely preferable, at the higher frequencies, this system makes it possible to reduce considerably the influence of parasites of artificial origin and to increase considerably signal-to-noise ratio.

3. Higher peak power and much higher frequencies are attainable than would otherwise be possible because of limitations due to heating of the transmitter tubes.

Thus, at the same time that it effects economy in transmission and simplifies and improves the efficiency of reception, pulse-time modulation provides also a greater degree of static reduction in communication.

In addition, the multiplex properties of pulse-time modulation open up still another application for amateur communication in helping to overcome line-of-sight difficulties to some extent The transmitted pulses are of very short duration and separated by comparatively large spaces, therefore more than one telephone channel may be transmitted on the same carrier provided that the pulses are suitably displaced with respect to one another. It is quite conceivable that by making use of this factor, amateurs may cooperate in setting up long nets and relay systems by means of which each operator will be conducting his own conversation and acting as a relay point for as many as eight or ten other conversations taking place over distances far beyond the limitations of ordinary line-of-sight communication.

Theory of Pulse-time Modulation

The basic theory of pulse-time modulation is extremely simple. It consists of transmitting intelligence by pulses of r.f. having constant amplitude and duration, the instantaneous amplitude of the voice being translated into variation of time intervals between successive pulses. The rate of this variation corresponds to the instantaneous frequency of the signal. The pulses themselves are of very short duration, separated by comparatively large spaces.

The spacing between the successive pulses can be varied in a number of different ways. Two of these possible methods of position-variation which have been used in practice are of interest to amateurs, and will be described.

The unmodulated carrier, consisting of a series of pulses having a constant repetition rate, is shown in (A) of Fig. 3. The individual pulses may be designated by numbers 1, 2, 3, 4, 5, etc., and the pulse repetition rate R per second. Then the time interval between successive pulses is T = 1/R seconds. When modulation is applied, the timing between the pulses is varied by an amount proportional to the modulating voltage, and at a rate corresponding to the frequency of the modulating signal.

In the first type of time modulation, one set of pulses is kept fixed in its time position and serves as a reference set for the time modulation of the other pulses. For instance, in Fig. 3A and 3B it can be seen that the odd-numbered pulses 1, 3, 5, etc., always remain in the same position under one another, while the position of the even-numbered pulses 2, 4, 6, etc., varies. The fixed pulses are the reference (or marker) pulses, and the others are modulated in position with reference to them. Fig. 3B illustrates the pulse position relationships for a single-channel system having a peak modulation of T/5. As shown in the upper part of the diagram, the distance between the modulated pulse and the marker is increased for positive modulation; the lower figure in diagram B shows the corresponding decrease in distance between modulated and marker pulse for peak negative modulation.

The process of modulation may be better understood by reference to Fig. 3E, which shows the pulse-position relationships when the single-channel system is modulated by one complete cycle of a 1000-cycle sine wave. During the positive half-cycle the pulse spacing is increased, as shown, by amounts varying from zero to T/5 according to the modulating voltage; while during the negative half-cycle of modulating voltage the spacing is decreased by amounts varying from zero to T/5. A comparison between pulse positions for the modulated and unmodulated condition can be seen in the middle diagram of Fig.3E, which shows the pulse sequence through a complete position-modulated cycle. The position variations of the modulated pulse for the 1000-cycle modulation are shown in summarized form in the lower figure, which shows the various positions of the pulse at different times in the modulation cycle. A system of the type shown, having a frame period of 125 microseconds (i.e., repetition frequency of 8000 cycles), is capable of transmitting a telephone channel of 3000 cycles.

(This system may be used for multiplex transmission in the manner shown in Fig. 3C. In this case the same marker pulses and the same frame period are used as in the single-channel system just described, but for each marker pulse there are now three modulated pulses each capable of carrying its own independent modulation. Thus, when the B pulses are modulated, the distance between the B pulses and the marker is varied within its limits regardless of the instantaneous position of the A and C pulses. Likewise the positions of the A and of the C pulses are varied regardless of the other channels. In the system illustrated, the A pulses constitute channel I, the B pulses channel II, and the C pulses channel III. When multiplex transmission is used, the marker pulse is made wider than the others, as illustrated, in order to distinguish it from the channel pulses.)

The second type of pulse-time modulation does not make use of a fixed series of pulses. Instead, the pulses are divided into pairs, and the timing between the two pulses of each set varied in accordance with the modulating voltage. Thus, pulses 1 and 2 would be one pair, 3 and 4 another, 5 and 6, etc. For the condition of peak positive modulation as shown in the upper part of Fig. 3D, these pairs are moved closer together while the pairs of pulses 2 and 3, 4 and 5, 6 and 7, etc., are moved farther apart by an equal amount. For the condition of peak negative modulation as shown in the lower part of the diagram, the situation is exactly reversed. Pulses 1 and 2, 3 and 4, etc., are moved farther apart while 2 and 3, 4 and 5, etc., are brought closer together. There is no change in the average pulse rate.

In pulse-time modulation systems the exact shape, duration or amplitude of the pulse has no fundamental importance, although it does affect the signal-to-noise ratio of the system. As in the case of a frequency-modulated system, the amplitudes are made uniform in the receiver by an amplitude-limiting circuit, thus making possible a considerable reduction in noise on the sole condition that the maximum potential due to noise is lower by a certain amount than the maximum amplitude of the received pulses. (In fact, even if the noise amplitude is greater than the signal pulse amplitude, there is still the possibility of time modulating during part of the pulse interval only, and of eliminating the majority of the interference by blocking the receiver except during the extremely short interval when the pulses are actually transmitted.)

Any noise small enough in amplitude to be eliminated by the limiters will nevertheless generate an audible noise at the output of the demodulator, because it also has the effect of advancing or retarding the time position of the leading edge of the desired pulse. This effect is decreased as the steepness of the wave front of the transmitted pulse is increased.

Since both the required bandwidth and the signal-to-noise ratio are determined by the steepness of rise of the transmitted pulse, it is thus possible to strike the best compromise between bandwidth, noise, and receiver-oscillator stability to make most effective use of the high-frequency bands.

Pulse-Time-Modulation Circuits

A number of the circuits which have been developed for time-modulated pulse transmission and reception are of interest to radio amateurs from the viewpoint of experimental communication on the microwave frequencies. Some of these circuits can be used with very little modification, while others can easily be adapted to make them reasonably inexpensive and suitable for ham use. A brief consideration of the existing circuits will serve to illustrate how amateurs can use them for experimental microwave communication, and how they may go about designing their own circuits for high-frequency pulse communication.

The principles of pulse-modulation transmitters and receivers can best be understood by considering them from the viewpoint of the basic principles of radio transmitter and receiver design. Fig.  2 shows block diagrams giving the essential details of the most general types of transmitter and receiver. These block diagrams apply to all three types of modulation now in general use - amplitude modulation, frequency modulation, and pulse-time modulation.

For any type of modulation, the transmitter consists essentially of the following sections: (a) an audio amplifier, (b) the modulator, (c) a high-frequency oscillator for generating the carrier, (d) a radio-frequency power amplifier. The receiver can be divided into the following sections: (a) a radio-frequency amplifier, (b) a high-frequency local oscillator and converter, (c) an intermediate-frequency amplifier and a series of limiters, (d) a converter or demodulator to restore the audio characteristics of the original signal, (e) audio filters and an audio amplifier to bring the signal to the desired level. In u.h.f. receivers, the r.f. amplifier (a) is generally omitted and the signal from the antenna fed directly to the high-frequency mixer.

The circuits for the three types of modulation differ from one another primarily in the modulating and demodulating circuits, and in the manner in which the modulation is applied to the output stage of the transmitter. In an AM transmitter, the modulation is generally applied to the power amplifier or to an intermediate power amplifier. In an FM transmitter, the modulation is applied before the power amplifier to the r.f. oscillator, whose frequency is caused to vary by an amount proportional to the instantaneous audio amplitude. A pulse-time modulated transmitter uses no power amplifier, the oscillator supplying the carrier power directly to the antenna. The modulator in this system serves the function of converting the audio amplitude into time-modulated pulses, which are applied to either the grid or plate circuit of the power oscillator so that the carrier is generated in short position-modulated pulses.

The essential difference between AM receivers and receivers for FM and pulse-time modulation is in the use of limiters. The AM receiver is a linear system as regards both the audio signal and the noise input, whereas the use of limiters in FM and pulse-time modulation receivers makes possible a considerable reduction in the noise output. Pulse-time modulation possesses the further advantage that oscillator tuning and stability are much less critical than in FM reception.

Modulation and demodulation, which is where pulse-time communication differs basically from AM and FM, may be accomplished by many different methods:

(a) One pulse-time system of the first type - ie., using a fixed series of marker pulses - makes use of two tubes, known as the "cyclodos" and the "cyclophone," developed specifically for this purpose. The essential features of these tubes are shown schematically in Fig.  10. Both tubes make use of an electron beam which, by means of an ordinary circular sweep circuit, is made to strike the aperture plate in a circular path. (For ordinary telephone communication, 8000 revolutions per second is a suitable frequency of rotation for the electron beam.) The aperture plate contains radial slits, so that during the time the beam is passing each slit the electrons go through and strike the collector segments, while at all other times they are intercepted by the aperture plate. Thus, once in each complete rotation of the electron beam a short pulse of current is passed to each collector segment. The tubes shown in the diagram, having aperture plates with four slits, would be suitable for a multiplex system having three channels, one slit serving for the marker pulse and the others for the three time-modulated signals.

In the modulating tube (the cyclodos) the slits are placed at an angle to the radius of the circular plate, as shown in Fig. 10A. It can be seen from the diagram that because the slits are tilted at an angle, the time when the beam crosses the open slit changes either forward or backward as the radius of rotation of the beam changes. Thus, audio-frequency amplitude variations may be converted into pulse-time variations by varying the instantaneous voltage on the deflecting plates and thereby changing the radius of rotation of the electron beam. The dotted circle in Fig. 10A represents the circular pattern obtained when all three channels are unmodulated. The solid path represents modulation in all three channels - instantaneous peak positive modulation in channels 1 and 3, negative modulation in channel 2. (The marker pulse, of course, is always unmodulated.)

In the demodulator tube (the cyclophone) shown in Fig. 10B, the slits are placed along the radius of the circular aperture plate instead of being tilted. To convert the time-modulated pulses into amplitude variations, the electron beam is made to rotate in a circular path, but the control grid is kept sufficiently, negative so that the beam is normally cut off except when one of the time-modulated pulses arrives. The pulses are applied to the grid and, depending upon where the electron beam is directed with respect to the slit at the time when the unblocking pulse arrives, there will be a variation in the amount of beam current passing through the slit and reaching the collector segment.

At the present time the cyclodos and cyclophone tubes are not commercially available. When they do become available the cost of the tubes, although expected to be moderate, will be a factor in determining their suitability for amateur use. For the time being, a reasonable substitute can possibly be improvised for amateur service by making use of an ordinary cathode-ray tube and phototube arrangement. This can be accomplished by cutting slits in a piece of black cardboard which fits over the face of the cathode-ray tube. Light shields and phototubes are placed in front of the slits so that each phototube records the light coming through one slit. In this manner, the slotted piece of black cardboard is made to act as an aperture plate and the phototubes take the place of the electron-collecting segments.

(b) A method of accomplishing pulse-time transmission and reception without requiring the use of the cyclodos and cyclophone is shown in Figs. 7 and 8. A block diagram explaining the operation of the transmitter is shown in Fig. 7A with the basic details of the modulating circuit - the oscillator clipper, the marker pulse generator, and the pulse position modulator - given in greater detail in Fig. 8A. The recurrence frequency is determined by the 8000-cycle oscillator, which at the same time provides a waveform suitable for starting the marker pulse generator and the pulse position modulator circuits for each channel. The marker pulse is obtained from the 8000-cycle sine wave by means of the oscillator clipper and the marker pulse generator. The pulse position modulator consists of a double triode connected as a biased multivibrator, and a single triode used as a pulse generator. Its operation may be explained briefly as follows: /p>

In the normal state (represented by the upper set of voltage values in the schematic) the second section of the multivibrator conducts current. The application of the marker pulse causes the first section to become conducting and the second section non-conducting, even after the initiating pulse has ceased. This condition persists until current flowing through the 3.3 megohm resistor charges the coupling condenser sufficiently to permit the plate current to flow in this section of the tube, at which point a rapid reversal takes place. Modulation is produced by variation of the potential to which the 3.3 megohm grid resistor is connected, by connecting it to the output of an audio amplifier tube. The mean length of the square wave generated by the multivibrator is adjusted to place the channel pulse in its nor-mal unmodulated position by means of the variable load resistor in the plate circuit, which adjusts the time constant of the multivibrator. The transient in the pulse generator tube when the reversal takes place causes the position-modulated pulse, by pulsing into one-half cycle of oscillation the tuned circuit formed by the inductance L, and the input capacity of the following tube. The remaining features of the transmitting circuit are quite conventional, as can be seen from the block diagram.

The operation of the receiver circuit can be understood from the block diagram in Fig. 7B and the demodulator schematic in Fig.  8B. The pulse position modulation is converted to amplitude variation by first converting it to pulse length modulation, then filtering out the voice frequencies below 3000 cycles from the pulse length-modulated signal. This conversion is accomplished by means of a multivibrator circuit. For each multiplexed channel there is a gate pulse which is generated in a fixed relationship to the marker pulse, as shown. When the voltage of the modulation pulse is superimposed on the gate pulse, the multivibrator is triggered and remains operative until the end of the gate pulse interval. This results in an output square wave whose leading edge is varied at the voice frequency. The remaining features of the receiver circuit are quite conventional, as can be seen from Fig. 7B.

The circuits for pulse-time modulation and demodulation employing conventional tubes are seen to be somewhat more complicated than those making use of the cyclodos and cyclophone tubes, especially for multiplex communication. However, for single-channel operation the circuit is not too complex, and it has the advantage of using standard tubes which are readily available. At the same time, it affords the radio amateur an interesting opportunity to experiment with extremely useful pulse and non-standard waveform techniques.

(c) Circuits for pulse transmission and reception by variation of the timing between pairs of pulses without the use of a reference pulse are shown in Fig. 9. In the transmitting circuit the pulses are generated by connecting the grids of the pulsing tubes in push-pull, and the plates in parallel. The manner in which the pulsing and pulse timing is accomplished can be seen from the wave-shapes shown in the diagram. With no modulation applied, one or the other of the pulsing tubes passes current at all times except when the input from the constant-frequency source is near the zero value. For near zero values, both tubes are simultaneously cut off and a short pulse of relatively high potential is passed into the transmitter. With no modulation these pulses are uniformly spaced. When modulation is applied the bias potential of both grids is varied in opposite directions, thus drawing alternate pulses together. Reversing the polarity of the modula-tion potential causes the pulses to be pushed apart.

The demodulator in the receiver makes use of two pulsing oscillators of the type used to produce sawtooth potentials in television receivers, each adjusted to the same frequency as the oscillator in the transmitter, with a common plate resistor to make them tend to operate 180° out of phase. The received pulses synchronize the operation of the oscillators, alternate pulses controlling each oscillator. Thus, when a received pulse advances the pulse of one oscillator, it retards the pulse of the other. Since the received pulses modulate the phase or timing of the oscillations of each pulse oscillator, but not the frequency, they vary the average current through the tubes. Therefore once the oscillators have dropped into synchronism with the received pulses, the average plate current will be modulated by the pulse modulation. This provides the original audio signal which was transmitted.

From the above discussion of the theory of pulse-time modulation and from the circuits that have been described, the amateur who desires to experiment in microwave communication should be able to design and construct a pulse-time modulation rig to use on these frequencies.

Posted March 16, 2015