Radio Broadcast (May 1929-Apr 1930)

How and Why Good Receivers Are Made Better ENGINEERING BEHIND A CROSLEY SET By KENNETH W. JARVIS Kenneth W. Jarvis Every season has brought forth upon the market new radio receivers; new in appearance, new in operations, and new in claims as to their great superiority over all previous models. Just what provokes this customary offering on the altar of public service? Technically, it is a result of new inventions or from more intensive engineering development. From the merchandise angle, the public demand for something new and different is an impelling motive. It is the purpose of this article to show how such a recent set (the Crosley Jewelbox) was developed and how its claims to superiority are justified. The "average" radio receiver of last season comprised an untuned radiofrequency amplifier and two or more tuned stages of radio-frequency amplification. This was followed by a grid-leak-condenser type detector using a heater tube, such as the 227. The first audio stage, as did the radio stages, used a low-voltage filament tube such as the 226. The output system employed one or two low-impedance tubes such as the 171a. The set was transformer coupled throughout, good design insuring a satisfactory fidelity. The volume control was a variable input system to the untuned r.f. stage. The First Improvement With this as a starting point, changes can be suggested and tried out. The first proposal is to change the input system. The advantages of the untuned input system are very real or so many manufacturers would not have used it. Its greatest advantage is in producing a unicontrol set, as the antenna capacity cannot affect any of the tuning units. It is economical and conservative of space and material. A properly designed choke coil can be used to provide a greater amplification on the lower radio frequencies and thereby flatten out the usually sloping sensitivity curve. It provides a convenient place to operate a volume control. However, in comparison with a tuned input system, the untuned amplifier shows several gross faults. It is relatively insensitive. If the average amplification over the entire broadcast band is greater than one, the designer may consider himself fortunate. The untuned system does not contribute in the least to the selectivity of the receiver, a factor which is requiring major consideration in these days of congested broadcast traffic. Due to the fact that no vacuum-tube amplifier is absolutely linear in characteristics, this untuned stage may produce peculiar effects. The second harmonic of strong local stations may be generated in this tube and 18 o U 5 16 < 14 UJ SS 12 10 then the station may be received at two places on the dial. Obviously, this can occur only with stations operating between 550 and 750 kilocycles. (The second harmonic of 750 kc. is 1500 kc, the limit of the tuning range of the receiver). Due to this same non-linear characteristic, and the lack of selectivity, a strong local station may modulate a weak distant station, and therefore be heard whenever these weak stations are tuned-in. There is certainly room for improvement here. The tuned stage has major advan The respective merits of untuned and tuned antenna systems, "power" detection and conventional detection with the two-stage audio amplifiers, as well as the engineering data behind the Crosley Jewelbox receiver will be found in this article by Mr. Jarvis, formerly of the Crosley Radio Corporation and now Chief Engineer of the Sterling Mfg. Co., of Cleveland, Ohio. The Editor. tages in the increased amplification and selectivity. However, the cost is greater and it adds an extra control to the receiver. Measurement and experiment have shown that the proper utilization of these advantages far outweighs the disadvantages. A comparative idea may be obtained from a study of Fig. 1. The dashed line shown is an amplification curve of a good chokecoupled untuned radio-frequency amplifier. In this, and the other curves in this figure, the amplification is measured from the input voltage in the dummy antenna system to the grid of the first tube. Antenna Input Design IN the design finally adopted for the antenna system, a coil with three taps at 3, 15, and 36 turns was used. These are numbered (3), (2), and (1), respectively, in Fig. 1 > ** y _ <•"*" Uin *a otage r~T 1. Tap No. 3 is used with extremely large antennas, or when exceptional selectivity is required. However, even with this small coupling, the average amplification is much better than the untuned stage. Tap No. 2 is used for the average antenna, and in the hands of the average owner is connected once and is never changed. Tap No. 1 is for an extremely small or indoor antenna. On a normal antenna, this tap will give a greater sensitivity to low radio frequencies than Tap No. 2. The relative efficiency of these curves compared with the untuned stage needs no comment. In addition to the great increase in sensitivity, a proportional gain in selectivity is made. After determining that the tuned input was desirable, the mechanical problem had to be solved. Due to the antenna capacity, this stage could not be made to rack exactly with the other stages. In the hand of an unskilled operator, it is extremely desirable that the input tuned stage always be tuned approximately, as the great increase in selectivity in this stage results in greatly decreased sensitivity unless the tuning condenser is maintained approximately in resonance. This makes some form of a four-gang condenser (there are four tuned circuits in this receiver) necessary, with the required adjustment on the first tuned circuit to compensate antenna variations. Various combinations were tried to accomplish this compensation, and the simple means of rocking the "stator" was finally adopted as being the best. This rocking movement is obtained by a worm wheel drive on a molded gear segment mounted on the stator of the condenser. The great reduction in gear ratio makes the tuning of this selective circuit very easy. Due to the way the condenser tuned circuits follow each other, only a slight adjustment of this control is necessary to bring the receiver to the peak of its sensitivity and selectivity. Another change which may be noted here is in the use of heater-type tubes throughout. The a.c. hum voltage is lowered greatly when 227-type tubes are substituted for the 226 type. This is due to decreased modulation of the radio-frequency amplifiers by the filament power-supply voltage, and to the lowered hum voltage in the first a.f. stage. The volume control is arranged by varying the grid bias on the radiofrequency amplifiers This cannot be done on the 226-type tubes due to the resulting hum modulation. In general, a volume control operates more satisfactorily when the sensitivity of the radiofrequency amplifiers is reduced, due to the reduction of the hiss produced in the tubes themselves. Thus, the use of the heater tubes allows this better type of volume control to be used. 600 700 i 900 1000 1100 1200 FREQUENCY IN KILOCYCLES 1300 1400 Fig. 1 — An idea of the increased amplification and selectivity obtained with a tuned input stage may be gained from a study of this graph. New Detector Circuit BY far the biggest change in the new receiver lies in the use of the bias-type detector. As very little data has been pub • JULY 1 929 • • 155