You need to have your thinking cap on for this article entitled, "The dispersion transmitter," which appeared in a 1944 issue of Radio News magazine. It might be more aptly entitled, "The Dispersion Antenna," or "How to Use a Phase-Driven Horizontal and Vertical Dipole Antenna Array to Effect a Circularly Polarized Transmission Antenna Which Continuously Scans All Combinations of Azimuth and Elevation Thereby Effectively Covering Every Point on the Earth's Surface Reachable by Ionospheric Reflection." That last title describes the ingenious system devised by author H.W. Kline. A description of tests run to verify the operation is included. He mentions that the circularly polarized signals emitted by the transmitter are usually received in a horizontally polarized orientation after being reflected in the ionosphere. Using the antenna setup for receiving signals does not benefit from the phasing system since, by the author's argument, any long distance received signal will have been horizontally polarized during atmospheric reflection. There would also need to be some means of actively sweeping the array during reception, which there is not.
The Dispersion Transmitter
Fig. 1 - This system is designed so that the radiated field load is swung to all possible vertical angles, as shown, so that skip distances can be avoided.
By H. W. Kline
When planning that postwar rig, amateurs will do well to employ this antenna system to eliminate skip distance effects caused by ionosphere variations. Radio transmissions will be greatly improved.
The following describes a radio-transmitting system developed to distribute nonfading signals to all radial points distant from the transmitter, without skips and regardless of shifts or vagaries of the ionosphere. The system is designed to swing the radiated field lobe to all possible vertical angles between the extreme incident angles of a composite antenna in the manner indicated in Fig. 1.
By continuously changing the resultant vertical incident angle of the radiated field lobe, dispersion of the reflected components from the ionosphere is accomplished and these components are caused to reach all regions on the earth's surface that otherwise would be skipped when the incident lobe is fixed, as in the case of the more common transmitting antenna in use today.
Transmissions of this characteristic also enable getting the signal into the greatest number of receiving antennas, all having different optimum angles of receptivity.
To enable this scanning of the ionosphere, an antenna having the capability of vertically swinging the incident field lobe at a supersonic or at a high-frequency rate, must be used. A simple form of antenna for doing this is described, wherein the scanning rate is synchronized with the carrier frequency. It was particularly designed to enable simple modification of existing transmitters. Its employment has resulted in reports from remote points of unusually strong, nonfading signals, while with no indications of skip transmission from nearby points.
Fig. 2 - 45° phase displaced waves producing unbalance of heating currents in the thermo ammeters. A1 will read high while A2 will read low.
Fig. 3 - Curves used to determine the permissible loading of the power amplifier by two simultaneous loads displaced by 90 degrees.
Fig. 4 - Showing the rising of the horizontal-antenna currents as lagging the rise of the vertical-antenna currents by 90 degrees.
Table 1 - Summarizing the amplifier and antenna-system operating conditions during a period in which initial long-distance tests were made.
Although the system was initially designed to operate on a carrier frequency of 7.15 megacycles, there is the possibility that it opens a path for long distance transmission of ultra-high-frequency carriers where emitted lobes are relatively sharp and depend on exact incident angles, during scanning sweeps, to produce signals at distant points, particularly if there is but one critical angle that must be found.
Although there are many other possible forms, the antenna system described actually consists of a combination of two antennas, a vertical radiator and a horizontal radiator in composite form. The incident lobe radiated from the vertical antenna, alone, is at lowest angle in respect to the earth's plane or about 10 degrees rise. The incident lobe radiated from the horizontal antenna, alone, is at highest angle in respect to the earth's plane or about 90 degrees.
The incident angle of the resultant, or combined field lobes of these two antennas, for all angles lying between the extreme angles, is regulated by varying the excitation between the two antennas. In order to do this at the carrier frequency, it was found necessary to obtain a phase displacement of 90 degrees between the energies fed to the antennas. How this was done is explained in complete detail further on.
The transmitting antenna consists of a half-wave, horizontal dipole, open-center fed with a tuned feeder line approximately one-quarter wavelength long. The dipole was erected one-quarter wavelength above the earth so that the incident angle from it, alone, would be at a very high angle or about 90 degrees in respect to the earth's plane. This same dipole, together with its feed line, was employed as a top-loaded, vertical antenna, with an excitation phase shift 90 degrees from that of the horizontal section. As a vertical radiator, alone, the current maximum occurs at the top end of the feed line or at the junction of the feed line with the horizontal top section. This makes its operation analogous to that of a half-wave, vertical radiator with current maximum at one-quarter wavelength above the earth. The angle for a radiator of this type is very low or about 10 degrees rise above the earth's plane.
Referring to Fig. 6, the operation of the composite antenna as a vertical radiator will be considered. In the sketch the horizontal antenna coupling circuit L1C1 is shown decoupled from the amplifier tank for the purpose of more clearly considering operation of the antenna as a vertical radiator. The direction of the charging currents are as indicated by the arrows adjacent to the circuit elements.
A leading antenna current is obtained by utilizing the natural, equivalent antenna capacitance to earth represented by the capacitors dotted and marked as C. Both of the vertical wires radiate as a single, vertical antenna conductor. At any point along the parallel vertical wires, the instantaneous current or voltage is of the same polarity and amplitude in either wire. The top-loading arms are each one-quarter wavelength long and since the currents flow in opposite directions in these arms, the magnetic field from the top section during the vertical mode of operation is substantially cancelled.
Radiation from the parallel wire vertical is approximately ninety percent as efficient as that from an antenna having a physical height of one-half wavelength. Current maximums occur at the junction of the vertical wires with the top-loading arms and at the ground ammeter A3. The antenna has an effective resistance of about 62 ohms and the current due to vertical excitation at A3 is of the order of two amperes.
It should be noted from the foregoing that as far as operation is concerned, the feeders or vertical wires could be shorted at any point or along the entire height without in any way affecting the operation as an efficient, vertical radiator. The coupling circuit L1-C1, used for coupling in energy for horizontal antenna excitation, does not interfere with operation of the vertical radiator for reasons of phasing which are explained later, and for the following reasons:
Since the vertical antenna currents flow in opposite directions in this coupling coil, the effective inductance is close to zero. This being the case, the reactance is negligible and only an insignificant component of effective resistance remains in series with the vertical members. In view of this, the horizontal antenna coupling circuit L1-C1 offers little if any impedance in the path of the vertical antenna currents.
The point on the amplifier tank circuit to which the vertical antenna clip is fixed, depends on the antenna circuit impedance at this point and ground and the 71 percent load limit on the amplifier. The amplifier is loaded by this antenna until 71 percent of the maximum d.c. plate volt-amperes are drawn. The d.c. plate supply voltage and plate current are used as an index of this degree of loading.
From the foregoing, it can be seen that it is definitely not intended for the vertical feed line to be used as a transmission line during those instantaneous periods of vertical antenna excitation. It is intended to operate as a twin-conductor vertical radiator. There are many types of vertical radiators that operate with more than one conductor in the vertical member. There are the cage, fan, pyramid, umbrella, and others. Operation is to be considered in this case as similar to these types except that there is a standing wave on the system one-half wavelength long as shown to the right in Fig. 6. In this instance, the currents flowing in the horizontal arms due to vertical excitation have been neglected because these currents are opposite and the radiated magnetic fields substantially cancel. The design is such that the current maximums are at the junctions of the vertical wires with the horizontal arms and at the ammeter A3.
Since the top-loading section is one-quarter wavelength above ground, the current supplied to the antenna at the ammeters A1-A2, is a small power charging current due to the reason that the antenna at this point is at highest impedance level and therefore must be voltage fed.
It will be attempted now to show how the composite antenna operates as two independent antennas, with noninterfering excitations as both a horizontal and a vertical radiator in which the excitation rises in one while it falls in the other and why this condition is obtained with a phase displacement of 90 degrees between the two exciting energies.
To begin with, it must be realized that the current in the vertical antenna is unidirectional while that in the feeders due to horizontal excitation is bidirectional. The ammeters A1-A2, average the effective values of all the increments of instantaneous current summations resulting in their respective legs. The characteristics of the instantaneous values of current involved are most important for analyzing results in this application.
Fig. 5 - Wiring arrangement showing method of energizing the antenna. The half-wave dipole antenna is erected one quarter of a wavelength above ground.
Fig. 6 - Illustrating the operation of the composite antenna as a vertical radiator.
Let us first consider a condition where the bidirectional, horizontal-antenna currents through A1 and A2 lag in rising by 45 degrees behind the unidirectional, vertical-antenna current. This condition is pictured in graph form in Fig. 2. The instantaneous value of current through either ammeter, at a time interval equivalent to the phase displacement along the X scale, is the sum of the current due to horizontal excitation and that due to vertical excitation. The vertical-antenna current splits in half, regardless of polarity, and flows in the same direction through each instrument. At instants when the horizontal-antenna current bucks the vertical-antenna current through either instrument, that instrument will, if the effective average is maintained, read low. On the other hand, the instrument operating in the adjacent leg will be forced to read high. A study of the graph, Fig. 2, will show that with a phase displacement of 45 degrees, the difference in current readings between A1 and A2 will be of the ratio of 1.27 to 0.566 based on r.m.s. values of the current waves. This was checked during actual phasing of the equipment. The phase shift is accomplished by the tuning of C1, Fig. 6.
It should be noted that there is much interference between the currents involved and it is doubtful if operation is correct with this phase displacement. This condition also is obtained with any phase displacement other than 90 degrees.
When the effective average of all the instantaneous values of currents flowing through ammeter A1 are equal to the effective average of all the instantaneous values of currents flowing through A2, the horizontal-antenna currents must be rising at a point which lags the rise of the vertical-antenna current by 90 degrees. Addition of the graphs in Fig. 4 will show the r.m.s. values of currents to be equal in both instruments, or approximately 0.997 amperes. The combinations of currents will be different when this phase displacement is other than 90 degrees and the ammeters A1 and A2 will not read alike. Consider a condition where the phase displacement is zero or 180 degrees. The current, on the argument graph, would be zero in one ammeter while it would be twice the crest value in the other. There is no telling what the operation of the system would be under this mode.
Referring to the graph, Fig. 4, showing the rising of the horizontal-antenna currents as lagging the rise of the vertical-antenna current by a time interval corresponding to 90 degrees, it will be noted that at the instantaneous positions marked "Y" on the horizontal-antenna current curve, the vertical-antenna current is zero. This being the case, there is no interference due to the vertical-antenna current and the composite antenna operates as an independent horizontal radiator. At the points of instantaneous vertical-antenna currents marked "Z," the horizontal-antenna currents are zero and, therefore, there is no interference from the horizontal-antenna currents and the composite antenna operates as an independent, vertical antenna.
When this condition is obtained by tuning C1, due to the symmetrical swinging of the current waves either side of the zero axis, ammeters A1 and A2 will read equal values of effective currents. Therefore, it is obvious that the current falls in one antenna as it rises in the other. The radiated fields are lower in one while they rise in the other and vice versa. The combination of the emitted fields is such that a major lobe continuously changes its vector direction in a vertical quadrant allowing scanning of the ionosphere in a continuous manner between the extreme angles of the two antennas incorporated in the composite antennas.
Fig. 5, shows a wiring sketch of the system when coupled to the final power amplifier of the transmitter. It should be noted that the phasing and loading of the amplifier was obtained by two different methods of coupling. For the horizontal antenna, mutual-inductance coupling was employed while the vertical antenna was fed directly from the amplifier tank as an autotransformer. In this way a 90-degree phase shift was obtained and the amount of loading depended on the degree of coupling in each case.
It will be noted-that this phase shifting method is quite similar to that employed in the discriminating circuits of FM receivers.
The available carrier power, at the power amplifier, must be distributed equally into the two members of the composite antenna. The permissible degree of loading was obtained by plotting sine curves of the phase angles for two waves displaced by 90 degrees. By hypothetically extending the sine scale on the Y axis and calling this scale the current scale, the resultant instantaneous current curve was plotted as a function of the displaced curves. Fig. 3 shows how this was arrived at.
The result of this deduction indicated that two systems could be coupled to a common amplifier and, providing a phase displacement of 90 degrees was employed, each could load the amplifier to within 71 percent of full load while the combined loading of the amplifier would be but 100 percent.
At certain intervals during each cycle, the current would rise in one system while it fell in the other and vice versa. This was the operation desired to cause the radiated resultant field lobe to vary through a 90-degree angular sweep, in synchronism with the carrier frequency.
The d.c. plate power to the class "C" amplifier stage of the transmitter was 350 watts at full load. At full load, 250 watts at 7.15 megacycles was obtained in a dummy antenna. The plate efficiency was therefore 71.5 percent. As resonant conditions were maintained in regard to the antennas, the combined or actual loads were analogous to that of the dummy antenna.
From the foregoing it was found that when either antenna system being coupled to the amplifier stage alone caused a dissipation of 71 percent of the d.c. plate power (percentage from curve of Fig. 3), it was considered as having its share of the available output power from the amplifier.
In this manner each antenna, individually, was coupled to absorb 71 percent of the power from the power-amplifier stage. With both antennas coupled, the simultaneous load was found to be actually as anticipated or 100 percent on the amplifier (0.3 amperes at 1165 volts).
Table 2 - A representative report on the operation of the transmission system.
The tuning of this transmitting system was found to be very simple and straightforward. The horizontal LC circuit was first tuned and then decoupled from the amplifier's tank. The vertical antenna clip was then run along the amplifier's tank until 71 percent of full load plate current was obtained. The amplifier tank was kept tuned during this process.
Following the above adjustment, the horizontal-antenna tuned circuit was then coupled to the amplifier tank until full load current was drawn at the amplifier plates. At this coupling point, by disconnecting the vertical antenna clip and after checking tank tuning, it was found that the degree of coupling of the horizontal antenna also placed a 71 percent load on the amplifier. With the two systems coupled, simultaneously, the amplifier reached full load plate current only.
The final phase adjustment was made with the tuning capacitor across the horizontal antenna coupling coil. By observing the r.f., line currents at A1 and A2, Fig. 5, at a little to one side of peak tuning, the r.f. current in the feeders increased at Al while it decreased at A2. On the opposite side of peak tuning, the current decreased at A1 while it increased at A2. There was a tuning point near peak tuning where the currents in the feed lines were equal. At this point the system was locked, it being assumed that correct phasing was obtained.
At this operating point the horizontal-antenna system was found to operate independent of the vertical-antenna system. There seemed to be a great increase in radiation efficiency as local monitors gave indications of greatly increased field strengths. The heterodyned signal at receivers had a distinct, sharp note.
Together with the vertical scanning of the radiated lobe is a circular polarization shift of 90 degrees caused by combining a vertical with a horizontal antenna in this instance. Emission of a circularly polarized wavefront was not considered harmful as it has been found that all transmissions after reflection from the ionosphere, arrive at remote receiving points substantially horizontally polarized, regardless of polarization at the point of origin. The effect of raising and lowering of the resultant lobe, however, enables scanning the ionosphere so that regardless of the position of a receiver, at a distance from the transmitter it will intercept a pulse of signal per scan at the instant when the angles of incidence and reflection fit the base line between transmitter and receiver.
Summary of Operating Conditions
Table 1 summarizes operating conditions during a period in which initial long-distance tests were made.
Table 2 is representative of operation of this transmission system on the specified wave band. Due to the start of war, further tests could not be carried on.
The following detailed tests made with Tegucigalpa, Honduras are representative of the procedure employed to test the system.
On February 6, 1941, at 9:40 p.m., Eastern Standard Time, HR1AT, Tegucigalpa, Honduras transmitted a prolonged CQ. These signals were intercepted with some difficulty at W2DKE, Schenectady, N. Y., the station owned and operated by the writer .
The experimental Ionoscan antenna was first connected as a center-fed, half-wave dipole, one-quarter wavelength above earth. The frequency of operation was 7115 kilocycles. Extreme care was exercised to make sure that full power was applied to this antenna system. With this connection, the transmitted lobe was determined to be at very high angle or about 90 degrees. HR1AT was called at 9:43 p.m. for 3 minutes continuously. There was no response.
At 9:47 p.m., HR1AT was again heard sending a prolonged CQ which lasted about two and one-half minutes. The Ionoscan antenna was quickly switched over to the top-loaded, vertical radiator connection and tuned for full loading of the final amplifier stage. Under this condition, the emitted lobe was previously determined to be of very low angle or approximately 10 degrees. At 9:50 p.m., HR1AT was again called for three minutes continuously with this radiator. There was no reply.
At 9:55 p.m., HR1AT was again heard transmitting a CQ lasting about one and one-half minutes, The full Ionoscan antenna connections were then made and the loadings quickly checked (all adjustments having been predetermined) and HR1AT was again called for one minute.
HR1AT replied immediately, reporting signals received at RST 589, QSB nil. Decoding the signal report for the benefit of the uninitiated, the report would read as follows: "Perfectly readable, strong signals of purest note with no trace of fading."
Summary of Results
Long-distance communications with the system were limited in that the system was installed only at the Schenectady end. For this reason, communications could be established only with those stations which produced readable signals in Schenectady. The receiving of signals at Schenectady from long-distance stations depended on the skip-distance conditions favoring the received signal, atmospherical limitations, etc. Usually the signals from long-distance stations were characterized by low intensity and both rapid and periodic fading, Schenectady is situated in the bed of what was once a fair-sized glacial lake (Lake Albany). It is only 400 feet above sea level and is completely surrounded by mountainous regions. Signals coming "in" must jump these mountains, while those going out must do likewise. The Ionoscan antenna enables this to a fair degree.
In most every case, distant signals are located by laborious hunting of the weak signals in the band, with the further effort of straining to make sense out of their rapidly fluctuating signals amidst difficult atmospheric conditions. It was quite evident, when contacts were made, that signals from Schenectady were being received with far greater ease. It is of interest to note also that the power radiated by HR1AT, Tegucigalpa, Honduras was 250 watts or substantially equivalent to that used at Schenectady.
In view of the foregoing, it seems reasonable to assume that at certain intervals during the combination of currents in the composite antenna system, the direction of the radiated field was caused to sweep through an angle that coordinated the reflected components from the layer with the optimum sensitivity angle of the receiving antenna of HR1AT in the Honduras. The effect was so pronounced that immediate reply was enabled due to the distinction of the signals transmitted from W2DKE at Schenectady.