Radio Broadcast (May 1928-Apr 1929)

106 RADIO BROADCAST DECEMBER, 1928 O To AF <Z>+A FIG. 9 The compute schematic diagram of the experimental breadboard receiver which is pictured at the head of this article. Note the thorough filtering of all the screen-grid and plate leads, the volume control which varies the screen-grid voltage, and the absence of shielding except the single metal plate. selves, the complete circuit diagram of this setup is given in Fig. 9, as well as a photograph from which the physical considerations can be seen. One coil of each of the pairs was mounted in slots so that the coupling could be varied, thus varying the width of the pass band. The distance between coils used in making the curves was marked so that it could be referred to during operation of the receiver. T his receiver was operated for two evenings in conjunction with a Silver-Marshall type 682-250 pack with remarkable results from a quality standpoint. Several unbiased observers stated that it was the finest quality of reception they had ever heard, noting particularly the excellent timbre of the high notes of the piano and organ. It would be even safe to say that some of the curse is removed from soprano solos when the overtones are freely admitted by a band filter. By placing a milliammeter (0-3) in the plate circuit of the detector, the double hump of the curve could be noted at the lower broadcast frequencies. Selectivity was ample for Chicago conditions, good clear spaces being obtained between the local stations in a location where local field strengths were very great. In these clear spaces on the dial two out-of-town stations could be heard weakly, but the amplification of the system was not sufficient to provide a good signal. The receiver in the form indicated could be called an excellent local receiver of the highest quality. To extend the scope of this model beyond local reception, two possibilities present themselves for the future. The first would be the addition of band-pass filter stages. This does not appear feasible except for those who can bear the expense and the necessary difficulties attendant with the matching of the large number of stages which would be involved. The other possibility is the use of a broad amplifier having no manual adjustments and equally responsive over the whole broadcast band. Many will recall the amplifiers of this type used in the early days of broadcasting when the tuned radio-frequency receiver came to the rescue. Perhaps with more general knowledge of electrical theory better success could be obtained with this type of circuit than in the past. The two tuned circuits above present sufficient selectivity to form very fine receivers with a good "untuned" amplifier of this type. CONCLUSIONS \ A/H1LE this is not intended to be a "final ' * hearing" on the subject of band-pass amplifiers, it would appear that sufficient ma + 45 +135 FIG. 7 The dotted curves in Fig. 6 were taken with this circuit, which is an auto-transformer in which part of the coil is used as primary and all of it as secondary. terial has been presented to arrive at the following conclusions with regard to circuits of the type discussed: (1) That the coils must be carefully matched and of low-loss construction in order to obtain a 1 1 1 1 ft 1 u 1 1 1 / — 1— 1 f I ! WL 600 KC. (A) +10 +20 good band-pass filter effect. Good coils are usually large physically, so good band-pass filters can be expected to have a considerable amount of bulk in their finished form. (2) That with good coils, better selectivity per stage can be attained than is to be had with the same coil operating as a conventional radiofrequency transformer. In general, it can be said that this increase will not be great enough to compensate the greater cost of the band-pass tuner stage, which is at least double that of the single circuit. (3) That the width of the band passed will vary with frequency when attempting to tune the broadcast band by means of fixed coils and variable condensers. This follows from the theory of the device, and was confirmed by the curves which show that with a 1 per cent, coupling the band passed was 10 kc. wide (total) at 1000 kc; 6 kc. wide at 600 kc. This constant percentage relationship prevents the attainment of the correct band width except over a small range of frequencies without changing the coupling between the coils. A similar situation has been long tolerated in radio-frequency transformers, however, so this is not a serious consideration. (4) That it is not possible to cascade sufficient of these stages to obtain the degree of amplification that is usual in sensitive radio receivers without prohibitive cost and constructional difficulties. (5) Far superior tone quality can be obtained by the use of a band-pass device than from the use of simple resonant circuits. It is undoubtedly on this last point that the band-pass filter will find a prominent place in the radio art. LIST OF PARTS THE apparatus used by the writer in making this investigation is given below. There is nothing special in any of the parts, and similar apparatus would perform as well. The coils, as mentioned above, must be of low-loss construction. The list follows: Li — 1 S-M r.f. choke coil, No. 275; L2, L3, L4, L5 — 4 S-M inductances, No. 140 (minus primary); Ci, C2, C3. Q — 1 Four-gang condenser, 0.0003 5-mfd.; C5, Ce, Cn — 3 Parvolt condensers, 2-mfd.; C7, C8 — 2 By-pass condensers, 0.1 mfd.; C9 — 1 Carter condenser, 0.0001 5-mfd.; C10 — 1 Condenser, 0.00 1 -mfd.; Ri — 1 Filament resistor, 10-ohm; R2 — 1 Potentiometer, 3000-ohm; R3 — 1 Filament ballast, 10-ohm; R4 — 2 Filament ballasts, 10-ohm; Rs — 1 Durham grid leak, 1 -megohm; 8 — Fahnestock clips; VTj, VT? — 2 Screen-grid tubes; Det — 1 Detector tube, 312-type. ' * V 1 ■ 1 l\ ' i 1 1 I \ \ \ ** / 1 1 1 1 1 1 1 ( B) 1000 KC. +10 +20 FIG. 6 The result — note that in each case the dotted curve, which represents a conventional transformer with only the secondary tuned, has poorer selectivity than the double-tuned transformer.