Radio Broadcast (May 1927-Apr 1928)

How Reliable Are Short Waves? Recent experiments of the General Electric Company, of engineers from other companies, and of those interested in private research, have resulted in an explanation of many of the mysteries that have surrounded the short waves. Everyone marvels at the ease with which amateurs communicate with fellow enthusiasts over great distances with small input powers. We have done it ourselves — communicated and marveled both — and great is the "kick" thereof. It is undeniably thrilling to take from one's lamp socket 6o-cycle power of less than one fifth that required to heat the average electric iron, and to feed it into a comparatively simple system of apparatus from which it emerges as radio-frequency energy with which we actually ask a man in South Africa how the weather is there, all the time sitting quietly in our den surrounded by unimposing gear. W hen it is winter in New York it is summer in South Africa, when day here, it is night there, and so on. It is one of the marvels of our time that two people in the security of their homes but separated by 7000 miles can transfer their thoughts instantaneously and economically. Just how good are these short waves? How reliable is communication? How many hours of the day, how many days of the year, can we send messages via short waves from New York to South Africa? The results of several investigations point to the following facts which seem fairly well established. Ten meters (30,000 kc.) is probably the shortest useful wavelength. Below about 20 meters (15,000 kc.) the waves prefer to travel in the daylight, and above that wavelength, night time is the best. Below about 45 meters (6660 kc.) curious "skip distances" occur, resulting in regions beyond which signals are heard but within which they are inaudible. For example, on 15 meters (20,000 kc.) during daylight, transmission is not practicable within a distance of 900 miles, which increases to about 1000 miles at night, although it is possible to transmit signals for reception at points more distant than these figures indicate. At 27 meters (11,000 kc.) the daytime skip distance has been reported as 1000 miles and 450 miles at night, these distances being about the same at 33 meters (9086 kc). The General Electric experiments show that the 32.79-meter wave is no good at all for short distances. A power output of 500 watts on 65.16 meters (4500 kc.) will, however, transmit commercial daytime signals up to 100 miles. Short waves seem necessary for extremely longdistance communication. During daytime, waves of the order of 20 meters should be used; waves below about 45 meters are not much good for short-distance work. Night after night we have heard naa on 37.5 meters pound away at a terrific rate, we have listened to the Marconi beam stations on 26 and substituting this into the equation above for Es: a Strays from the Laboratory /La <ii iaiiiw 6" meters, aga at Xauen, Germany, fw in Paris and others, all bent on getting somewhere in a hurry, and we wonder how soon it will be before the band between 25 and 45 meters is as busy as the long-wave channels. Amateurs in this country who have a channel 1000 kilocycles wide around 40 meters have been blessed with an excellent assignment which at the time it was doled out was thought to be more or less worthless. More and more amateurs are going to 20 meters and with low powers are accomplishing unheard of records in broad daylight. They have not as yet the feeling for this band and for conditions existing there that they had for the 40-meter band, but it will come when they gain the wealth of experience they have amassed on the longer wavelength band. ^Mathematics of the Audio Transformer In most radio work the mathematics is fairly simple; the difficulty comes when it is necessary to put the mathematical theory into practice. For example, the following mathematics is that underlying the theory of the input transformer for audiofrequency amplifiers. The circuit about which this theory is built up is given in Fig. 1, and its equivalent is shown in Fig. 2. In this mathematics, N is the turn ratio of the transformer. Es = NEp = Ig Rg Ep = 'J.Eg — Ip Rp Es = N (jxEg — Nig Rp) IgRg = N ftlEg — NIgRpj Whence Es Rg + N=Rp Rg Ny-Eg Rg + VPp N>Eg I + K = Rp Rg FIG. I Differentiating this equation with respect to N and solving for a maximum, it is found that: N = Rg/Rp 228 Es = •JEg N All of this assumes that the transformer is perfect, i.e., no d.c. resistance, no magnetic leakage, infinite primary and secondary reactance. It shows that under these conditions the voltage delivered to the input of the tube is one half that delivered to the previous tube multiplied by the turns ratio of the transformer and by the amplification factor of the previous tube. If the input impedance is one megohm, 1,000,000 ohms, and the plate impedance of the previous tube is 12,000 ohms, the turn ratio will be equal to the square root of the ratio between these two quantities, viz: N = = 3 approximated J 12,000 In order that a large percentage of the a.c. voltage developed in the plate circuit of the previous tube be available across the primary of the input transformer, the impedance of this transformer must be high. If we want to amplify well at 100 cycles, the input impedance should be not less than 30 000 ohms which means that the primary shouid have no less than 50 henries inductance— which in turn explains why transformers, good ones, cost money, and why a skinny little affair with a few sheets of iron in the core and a little wire on the primary makes radio music sound "something fierce." Furthermore, if the turn ratio is three, and the inductance of a winding varies as the square of the turn ratio, the secondary inductance must be about 450 henries — and when anyone states that the secondary of an audio transformer makes a good output choke he neglects the fact that one cannot wind up an inductance of 450 henries without adding enough d.c. resistance to prevent the last tube from getting any plate voltage at all. MATHEMATICS OF THE OUTPUT TRANSFORMER THE mathematics of an output transformer design is no more difficult than that of the input transformer — and the answer is the same, as the following rigamarole proves. The symbols used in this discussion are the same as for the input transformer. N is the turn ratio of the transformer, and instead of Rg we use Rs. _ Es NEp ls~ r7~ Rs Ep = fiEg — IpRp fJ-Eg Rp4-_Rs N2 Whence I fJ-Eg N;j.Eg" N2Rp +Rs y. Eg Rp V N'2 Rp + Rs NjJ-Eg y'2 (RpNH-Rs) ) FIG. 2