Radio Broadcast (May 1929-Apr 1930)

•Z RADIO BROADCAST Gang Cond. The complete circuit diagram of the Crosley 804 receiver (The Jewelbox). Note the taps on the antenna coil and the arrangement of the detector which is designed to overload before the power tubes so that the distorting blasts from the loud speaker which indicate overloaded amplifiers are eliminated. Fig. 4 shows that distortion from the Jewelbox actually decreases as the input is increased while the conventional grid leak and condenser detector overloads sooner and the distortion increases at a rapid rate. i — 0^ 0.00025 Cables, Plug corresponding to about 1-watt output (171's in push-pull with 4000-ohm load), the power distortion is only 1 per cent. ! As the radio input is further increased, the output voltage increased to about 82 volts (output 1.7 watts) with a distortion of 9 per cent. If the radio input voltage be further increased, the audio output voltage will decrease as shown by the reversing curve. The percentage of harmonics increases slightly and then decreases as the radio input is continuously increased. This is a new idea in receiver design and one which deserves further comment. The Output Circuit The output system of this set consists of two 17lA-type tubesin push pull. Singly, these tubes are capable of delivering about 0.7 of a watt. In push pull they will deliver slightly more than twice this value, about 1.5 watts. If the input voltage to such a push pull stage is increased to give a greater power output, the distortion will be enormously increased. In fact, it will sound terrible! But if some means were provided to prevent a great overload of these out 100 80 tu 40 £ 4 V-I v \ .3 \ — V \ \ \2 s — 500 600 700 800 900 1000 1100 1200 1300 1400 1500 FREQUENCY IN KILOCYCLES Fig. 5 — Sensitivity curves of the Crosley receiver. put tubes, it is apparent that even in the hands of an inexperienced operator, it would be impossible to overload the output system. The receiver can never be ma e to deliver a signal of poor quality ! This novel effect is obtained by proper adjustment of the detector voltage characteristics, so that as the detector overloads it not only fails to contribute greatly to the distortion, but actually serves to reduce the distortion if the input is sufficiently great. 120 FREQUENCY IN CYCLES PER SECOND Fig. 6 — Overall fidelity curve ivith the r.f. stages tuned to 1000 kc. Some idea as to the magnitude of this distortion should be given. A dashed line at 10 per cent, distortion is shown as the "Minimum Perceptible Distortion." This is perhaps a pessimistic viewpoint as many individuals in test have permitted a power distortion as high as 25 or 30 per cent, before noticing it. To an untrained ear, this is perhaps excusable. At low percentages, this distortion is apparent only as a change of timbre, or tone quality, and has no characteristics of discord or poor quality. As indicated, this "Minimum Perceptible Distortion" means that point at which a change in timbre becomes apparent. If further comment were needed it would only be necessary to show the great distortion in some receivers whose owners claim them to have the finest quality on earth! With this idea of what distortion means, the curve b of Fig. 4 may not be judged too harshly. Certainly at low (and probably the normal room volume) there is no choice on a basis of distortion. An output of at least one watt is possible without perceptible distortion. The limit beyond this point is largely a matter of customer discrimination. It can, of course, be pushed until the distortion is apparent to every one, including the neighbors. An automatic overload control such as is shown in curve a is a real contribution toward the improvement of modern radio reception. Sensitivity of Receiver The curves shown in Fig. 5 are the usual sensitivity curves plotted in microvolts per meter. The sensitivity on tap (3) was sacrificed to some extent to provide for an extreme selectivity. As shown previously, in the tuned antenna stage discussion, the combination of the characteristics of the three antenna taps makes the receiver adaptable to almost any receiving condition. A comparison of these curves with those shown in a previous article in the January, 1929, Radio Broadcast will show what an enormous improvement has been made in this receiver. In Fig. 6 is shown a fidelity curve taken at 1000 kc. Throughout a range of approximately five octaves there will be no apparent change in sound intensity. This is unusual, and is accounted for only by those factors of design previously discussed. A comparison with those fidelity curves shown in the above mentioned article is invited. The overall selectivity is such that, except in unusual circumstances, stations on adjacent channels may be received without interference. (And providing thay are on their adjacent channels). Close to strong local stations, satisfactory reception may be had three channels adjacent. In considering the development of this receiver and the results obtained, the Crosley engineers feel that they have not merely brought forth another receiver, but one deserving in every particular of the sobriquet of "new." • JULY • 1929 • • 157