January 1939 Proceedings of the I.R.E., pp. 12-15 Observations on Sky-Wave Transmission on Frequencies Above 40 Megacycles* D. R. GODDARD , ASSOCIATE MEMBER, I.R.E. Summary - The results of daily observations at Riverhead, L. I., N. Y., since September, 1937, of European 40-to 45 megacycle transmitters are reported. Measurements of field strength were made on English, French, and German television signals. Multipath propagation of the English video-frequency channel was observed optically and the difference in path length determined. This paper deals briefly with the results ob- tained by systematically observing and meas- uring the field strength of television signals from England, France, and Germany. The English transmitters located at Alexandra Palace, London, operated on 41.5 megacycles for the sound channel and 45 megacycles for the picture channel. The frequency of the French sound trans- mitter at the Eiffel Tower was 42 megacycles and the Berlin sound-transmitter frequency was 42.5 mega- cycles. A rhombic antenna 45 feet above ground and di- rected towards London was used for these measure- ments. Its length was 400 feet per side and it was arranged so that the dimensions of the major and minor diagonals could be readily changed. This was done so as to facilitate matching the antenna to the vertical arrival angle of the signal. The effective height of the antenna system was about 20 meters. Antenna adjustments were made by comparing various settings of the rhombic antenna to a refer- ence dipole 45 feet above ground. As comparisons * Decimal classification: R113.7. Original manuscript re- ceived by the Institute, May 25, 1938. Presented before joint U.R.S.I.-I.R.E. meeting, Washington, D. C., April 29, 1938; presented before New York meeting, November 2, 1938. R.C.A. Communications, Inc., Riverhead, L. I., New York were made on the London 41.5-megacycle signal the results gave the optimum setting for that signal. This setting corresponded to a vertical arrival angle of roughly 7 degrees. Fig. 1 shows the receiving equipment used. In the foreground is a television receiver with a small camera mounted over the oscilloscope. Only the video-frequency amplifier and Kinescope controls were used as it was thought desirable for these ex- periments to have available greater flexibility than the radio circuits of this set provided. Therefore, the receiving standing to the left of the one just described was designed by Bertram Trevor of this department. This set provided automatic or manual volume con- trol, a minimum noise equivalent of about 30 micro- volts, with a band width somewhat less than 5 mega- cycles, and two diode outputs, one giving a "positive" and one a "negative" image. On the bench is the signal generator and receiving equipment used for signal-strength measurements. Most of the observations took place between 9:45 A.M. and 11:30 A.M., E.S.T., as that appeared to cor- respond approximately to the afternoon schedules of all three countries. On several occasions, however, the transmitters continued on into the afternoon, usually on tone modulation. On November 19 and 20 the English audio-frequency transmitter was operated from 5 A.M. to 8 A.M., and 9 A.M. to 11 A.M., E.S.T. On both occasions the signal was first heard at Riverhead a few minutes before seven. On the latter date, however, the signal disappeared at 7 A.M. and did not reappear until 9:20 A.M. October 16 from 4:00 P.M. to 4:30 P.M. was the only occasion on which the 41.5-megacycle English signal was heard on the evening schedule corresponding to 3:45 to 5:00 P.M. E.S.T. Fig. 2 shows the peak signal strength in decibels above or below one microvolt per meter of the English audio- (41.5 megacycles) and video-fre- quency (45 megacycles) signals for every day during the winter of 1937-1938 that either or both were heard. The small crosses indicate holidays on which no observations were made. The uppermost curve is a plot of F2-layer virtual height as broadcast by the National Bureau of Standards. These measurements are taken each Wednesday at noon, E.S.T. Directly below this curve is a plot of the critical frequency of the F2 extraordinary ray. The data for this curve were supplied by John Howard Dellinger of the National Bureau of Standards. Some of these values were not obtained by actual critical frequency meas- urements but are close approximations. Inspection of Figs. 2 and 3 indicate that a strong signal is not necessarily accompanied by a high critical frequency or low layer height. However, the month of November produced the most consistently strong signals and was characterized by a uniformly high critical frequency. One interesting case was that of February 14. On this day, probably due to a magnetic storm, the noontime critical frequency dropped to 10,150 kilocycles and yet the English 45- megacycle channel was heard faintly and the English 41.5-megacycle channel was quite strong. On De- cember 1, however, the critical frequency rose to 14,700 kilocycles and only a weak signal was ob- served on 41.5 megacycles and the 45-megacycle channel went unheard. Of course, it should be pointed out that the criti- cal-frequency and layer-height measurements were made at Washington, D. C., at noon while most of the signal-strength measurements were made from one to two hours earlier in the day. Furthermore, as the signals coming from Europe probably traversed the Atlantic Ocean in two hops, the places in the ionosphere that caused the signals to return to the earth were something like 1800 and 4500 kilometers northeast of Washington. An attempt to correlate the maximum useable fre- quencies taken from the weekly broadcast of iono- sphere data from station WWV appears in Fig. 4. The lower half of this figure represents the maximum useable frequency interpolated for a distance of 2700 kilometers plotted for each Wednesday during the winter of 1937-1938. Twenty-seven hundred kilo- meters was used as it represents half the distance between Riverhead, L. I., N. Y., and London. The upper half of the figure is a plot of maximum signal strength observed on the two English channels for the same days that the ionosphere measurements were made. The broken lines represent the 45-mega- cycle signal and the solid lines represent the 41.5- megacycle channel. The small circles indicate no signal heard for that day. From these data it may be seen that at no time was 45 megacycles indicated as useful, the highest values being 43.4 and 43.7 megacycles on December 1 and 22, respectively. On December 1 the voice chan- nel was heard faintly and the video-frequency chan- nel not at all, while on December 22 both frequencies were quite strong. Another interesting case is that of October 27, a day having the relatively low maximum useable frequency of 35 megacycles. On that occasion both the 41.5- and the 45-megacycle signals were fairly strong. On March 23, there was a severe mag- netic storm decreasing the useable frequency to 15.8 megacycles. On November 5 the signal strength of the English voice channel rose to about 56 decibels above one microvolt per meter. Computation indicates that, neglecting the effect of the ground near the receiving and transmitting antennas, the field strength at Riverhead from the 3-kilowatt English transmitter should have been about 40 decibels above one micro- volt per meter. At most, the effect of the ground at both antennas would have increased the field at Riverhead by 12 decibels making the expected field 52 decibels above one microvolt per meter or 4 decibels less than the peak values actually measured. Of course, the field-strength measurements probably include an error of a few decibels, but even so indica- tions were that the attenuation over the path must have been at times nearly zero. Possibly there may have been a concentrating or focusing effect of some nature. The fading on all four signals observed was usually very deep and rather rapid. During days of very strong signals, however, the fading was quite slow, occasionally remaining constant for nearly a minute at a time. Selective fading on the voice channels occurred rarely and was invariably accompanied by a deep dip in signal strength. Two-receiver diversity reception very effectively removed the distortion pro- duced by this selective fading. These European television signals have been re- ported heard on a number of occasions from as far west as Phoenix, Arizona. Clyde Criswell, located near Phoenix, has reported hearing all the afore- mentioned signals during the winters of 1936-1937 and 1937-1938. G. W. Kenrick at San Juan, Puerto Rico, reported hearing the French, German, and English voice channels several times during the past winter. On most of these occasions the signals were also heard at Riverhead. On one occasion the rhombic antenna used for these observations gave a very much weaker signal than a standard short-wave fishbone antenna di- rected towards London. Usually the rhombic antenna gave several times the signal strength observed on these fishbone antennas. This condition lasted from about 10:30 to 11:00 A.M. on February 15. This condition may have been due to the signal arriving over a path other than the great-circle path from London. A deviation of but a few degrees from this path would considerably reduce the voltage picked up by the rhombic antenna while it would have slight effect on the fishbone antennas as they were not designed for these frequencies. Criswell reported ob- bserving variations in horizontal arrival angle with his rotatable Reinartz beam antenna. He also re- ported usually obtaining a stronger signal from a southeasterly direction during times of weak signals from London. This same condition was observed at Riverhead during the winter of 1936-1937. In this connection it might be mentioned that an amateur at Peeksill, N. Y., operating on the 28-megacycle band was observed to have "around the world" echo. This occurred on December 12 at about 10:45 A.M., Eastern Standard Time. A few minutes later an amateur in Holland was heard calling the operator at Peeksill. At about the same time of the morning of February 17 "around the world" echo was heard on the second harmonic (37.8 megacycles) of a Rocky Point, L. I., N. Y., transmitter operating on 18.9 megacycles. Before closing, mention should be made of the re- sults obtained with the Kinescope shown in Fig. 1. On February 18 the English video-frequency channel became strong enough to synchronize the Kinescope sweep circuits and allow glimpses of the picture being transmitted. Usually these pictures consisted of numerous images superimposed on one another in- dicating two or more paths of propagation. The path conditions were continually changing and occasion- ally a single picture would appear quiet plainly and with good detail. Fig 5 shows an attempt to photo- graph this multipath phenomenon. It shows the front view of a man's head and shoulders. As can be seen there are two images and computation shows that the horizontal displacement represents a time delay of about 3.5 microseconds which corresponds to a difference in total length of the two paths from London of something less than 3000 feet. ACKNOWLEDGMENT The helpful suggestions of Mr. Martin Katzin are gratefully acknowledged. Added in Proof: Recent study indicates that the Lorentz(1) polarization term probably should be in- cluded in the computation of maximum useable fre- quencies. Application of the Lorentz term would in this case increase the predicted maximum useable fre- quencies by about 20 per cent. This may be shown graphically either by reploting the maximum usable frequency curve or, as in done in Fig. 4, by drawing the horizontal solid and dotted lines opposite 34.5 and 37.5 megacycles of the ordinate scale. These val- ues represent 20 per cent less than the voice and video frequencies for the London transmitters. Now the correspondence between predicted maxi- mum usable frequencies and observed signal condi- tions is somewhat improved. Using these horizontal lines as references, prediction for the 45-megacycle channel rises from 38 per cent correct to 73 per cent correct, and for the 41.5-megacycle signal increases from 46 per cent correct to 77 per cent correct. (1) H. G. Booker and L. V. Berkner, "Constitution of the iono- sphere and the Lorentz polarization correction, Nature, vol. 141, pp. 562-563; March 26, (1938).