Radio Broadcast (May 1927-Apr 1928)

ASIHk RkOADC.ASThk ShFTTT BY CAKI TECHNICAL problems presented in this section in the past have consisted of numerical problems requiring more or less lengthy solutions, so that only one question could be allowed to an issue. We shall vary these occasionally by a series of questions requiring only brief answers, like those below. The subjects, while not strictly confined to the design and operation of broadcast transmitters and associated apparatus, will have some connection with those considerations of quality in reproduction which are of most interest to the professional broadcast technician. Question i. Why does sharp tuning tend to drop out the high audio frequencies associated with a carrier? Answer. The audio energy of speech or music exists in the side bands accompanying the carrier in question. If, for example, the carrier is of the order of 600 kilocycles, a frequency within the broadcast band, and the audio note being transmitted is 5 kilocycles, the side bands will have a frequency of 595 and 605 kilocycles. A sharply tuned circuit resonant to 600 kilocycles will include perhaps one kilocycle on either side, and will discriminate to a greater and greater degree against the higher audio frequencies, which lie further out from the central or carrier frequency. Hence the 595 and 605-kilocycle currents, in the present example, may be lost altogether, and with them the 5-kiIocycle audio note which they would yield on demodulation. By this mechanism, called "side band cutting," the higher musical frequencies are likely to be lost. Question 2. Why does slight detuning of a sharply resonant circuit, with reference to the carrier frequency, tend to reinforce the higher notes contained in the modulation? Answer. See Fig. 4, which has reference to the same example used to illustrate the answer to Question 1. Now, however, the circuit has been tuned so that the peak of the resonance curve occurs at 602.5 kilocycles instead of at the carrier frequency of 600.0 kilocycles. As a result, the peak of receptivity, after demodulation, has been shifted from currents of frequencies in the neighborhood of zero cycles to currents in the neighborhood of 2500 cycles per second. If, owing to the steepness of the resonance curve, the cut-off characteristic of the circuit becomes serious one kilocycle to either side of the peak, as assumed in the answer to Question 1 above, with the peak set at 602.5 kilocycles the audio frequencies up to 3.5 kilocycles are nevertheless included in the received band. It will be seen that through this mechanism a sharply tuned radio frequency circuit may be used to some degree as an equalizing or frequency correcting device in reception. This observation may be checked in practice. A similar phenomenon is found in some super-heterodyne receivers. FIG. I Question 3. What change in loudness is normally noticeable by the human ear? Answer. A change of 3 TU is usually observed in speech or music even by listeners who are not expecting it. This corresponds to a change in energy of 2:1. A practiced listener looking for a change may detect differences of the order of 1 TU in a sustained note, without much difficulty. This is equivalent to an energy change of about 25 per cent. Question 4. What size steps, in TU, would you allow in a smooth gain control? Answer. Since, according to the answer to Question 3 above, a change of 3 TU is noticeable, the maximum allowable step in a smooth gain control is 2 TU. Question 5. What error do non-technical observers invariably make in estimating relative loudness of sounds? ' Answer. They underestimate radically. As is well known, the human ear follows a logarithmic response characteristic, which is as much as to say that a large increase or decrease in the stimulating energy results in a slight change in the loudness subjectively perceived. Or, more definitely, multiplying the energy by a fixed ratio results only in adding to the loudness an increment proportional to the logarithm of the energy ratio. This is expressed mathematically in the formula for the telephonic transmission unit, which is a measure of subjective loudness: TU = 10 log P1/P2 where Pi and P2 are the powers corresponding to the two telephone currents under comparison. Non-technical listeners, being unaware that hearing is a logarithmic process, usually apply to sounds the standards of measurement and' estimating to which they are accustomed in dealing with distances, for example. A broadcast listener, comparing reception from two stations, one of which is a stage of audio amplification above the other, will say that the first station is "fifty per cent, louder," or "one hundred per cent, louder." Actually one a. f. stage may correspond to 20 TU, or a sound energy ratio of 100. In other words, a non-technical listener is apt to speak of an energy ratio of the order of 2:1 where the actual figure is of the order of 100:1. Commercial Technical Publications AMONG catalogs and technical publicaZ_A tions received we must mention the I V September, 1927 issue of the General Radio Experimenter , a four-page paper sent out monthly to concerns and individuals on the mailing list of the General Radio Company of Cambridge, Massachusetts. [This publication may be secured by our readers by filling out the "Manufacturers' Booklet" coupon appearing in the advertising section in this and other issues. Booklet number 74. — Editor.] Thi firm, as is well known, specializes in communication laboratory and measuring apparatus, such as standards of inductance, resistance, and capacitance, oscillators, oscillographs, audibility meters, bridges, artificial telephone lines, etc. Its engineers have played no small part in reducing radio designing to a respectably exact science. The September, 1927 issue of the Experimenter is of special interest to broadcasters in that it contains an article on "Design and Use of Attenuation Networks," by Horatio W. Lamson. The subject was discussed, it may be remembered, in this department for September, 1927 (Pages 293-294). Mr. Lamson's paper covers the same ground, in part, with the addition of a number of formulae and a detailed description of the General Radio Company's variable attenuation networks. Given, as in Fig. 1, a source of impedance Z and an absorbing circuit or "sink" of the same impedance, joined by an H-network as shown, with currents 10 and I leaving the source and entering the sink, respectively, for a definite number of transmission units of attenuation N we may calculate the arms X and Y from Z k — 1 X = ( ) 2 k + i 2l (• k= — 1) where k = (■) (2) (3) or, in TU k = 10 n/2° = Antilog (4) If, as shown in Fig. 2, we make all five branches of the H-network adjustable by steps, moving five switch arms in unison to the proper switch points through a single control, the characteristic impedance Z of the network may be maintained constant to match the impedances to either side, while the attenuation is varied in steps. This is the principle of the Type 239 Variable Attenuation Network supplied by the General Radio Company. The box is equipped with two multiple switches. In the size which affords a total attenuation of 55 TU, one decade is calibrated in steps of 5 TU, and the other in steps of 0.5 TU. Another size goes up to a total of 22 TU, and in this case the steps are 2TU and 0.2 TU. Characteristic impedances of 600 and 6000 ohms are carried in stock or built to order. Some types are provided with a center tap for the Y-branch, where a ground may be applied. A somewhat older type of variable pad supplied by the same manufacturer is adjusted by the addition or subtraction of fixed sections, which are cut in or out by means of four-pole double-throw switches, as shown in Fig. 3. This is known as the Type 249. The manipulation of this form is necessarily somewhat less convenient, but in many cases it serves the same purpose and the cost is about half of the rotary switch form. There are eight-section boxes affording a total attenuation of 1 10 TU in steps of 1 TU, and six-section boxes with a total of 63 A \ Output Input 5* 4| 3* 2* li oi FIG. 2 235