Radio Broadcast (May 1929-Apr 1930)

RADIO BROADCAST begins to fall off badly. The next obvious weak spot is the "scramble" winding of the usual choke, since this permits chokecoil turns to slip down inside the winding and come near other turns that should be removed from them. A systematic winding would seem to be of advantage but trial shows that plain layer winding produces so much inter-layer capacity as to ruin the choke rather completely. This leaves the expedient of a winding with narrow layers of the "universal" or "honeycomb" type, or else a winding with layers but one wire wide! This last sounds foolish but is actually done in the Samson choke which consists, in effect, of spirals (one wire wide) laid side to side. Lest this seem to run to hairsplitting let us show a few concrete figures. To secure these a tuned 222 stage was fed from a 1-meter antenna, the driver being a sinusoidally modulated 50-watt transmitter some distance away. The r.f. output of the 222 stage was fed to a detector and twostage audio amplifier, the a.c. output being measured with an a.c. meter suited to the modulation frequency of 120 cycles. Various chokes were used to shunt-feed the 222 tube whose plate load was the detector-input tuned circuit. Choke Meter Readings (Scale 0-200) used 15 meters 80 meters 360 meters A 1 8 80 B 0 10 50 Samson 5 25 100 C 3 18 90 D 4 6 35 E 6 28 110 The chokes used were as follows: A Commercial tube using a wooden form with three unequal grooves wound full of No. 38 enamel. B Same as A except as to proportions. Wire size about 34. C Three small "Universal" coils 2J" diameter by If" wide, wound with No. 38 single silk. D One Universal coil 16iV' wide and with outside diameter of 2". E Feed through tuned circuit Fortunately, also the lumped types of chokes have so many resonance points that they overlap into a sort of mountain range, giving a fairly uniform high impedance over a wide range and then sinking imperceptibly into the normal Xc and Xl curves. One accordingly has a wide useful range. Though the peaks are not as high as in a simple solenoid, which is very much better over a small range, the voltages involved are low and the circulating cur It will be seen that even the best of the chokes drops off in impedance with wavelength. This is not altogether the fault of the choke, since the tube falls off also. For comparison there is provided the set of figures obtained with the same tube fed through the tuned circuit, as in Fig. 2. It is only fair to say that the chokes used were considerably better than the average. Some of those sold gave no reading at all at 80 meters and values as low as 10 at 360 meters. The circuit used here would work well as long as the choke reactance was high regardless of i!s sign: that is, whether it was capacitive, inductive, or resonant. Dr. E. F. W. Alexanderson (right), consulting engineer of the General Electric Company, examining the memory meter which he developed for depth sounding from airplanes. The memory meter intercepts the radio echo and records the altitude for the observation of the pilot. With Dr. Alexanderson in the picture is his assistant, S. P. Nixdorff. rents due to resonance are accordingly low also. The case for the transmitter is different. Here the voltages are materially higher and serious currents will flow through the choke unless its impedence is very high indeed. If we attempt to run the impedence up to the requisite value by using a "tuned trap" (coil shunted by a variable condenser) we will have large circulating currents and large losses in this trap. Fortunately, the transmitter, unlike the receiver, is not required to work at all wavelengths and we are enabled to do away with tunable circuits, using instead a coil hav ing small C, large L and ample length so as to cut down the field intensity. In general, this becomes a slim solenoid or else a series of small honeycomb coils strung on a rod with their fields additive. In the latter case the sections may be equal or unequal without marked change in performance. The wire size must be as small as possible — frequently a small wire will heat less than a large one. However, the smaller wire must retain the same thickness of covering to secure this result. The effect is even more noticeable if the spacing in turns per inch is kept constant as the wire size is reduced. This is often difficult without the introduction of grooved forms, whose value is dubious indeed in this work. It has frequently been denied that working a choke at resonance is a practical operating condition but the writer can find no case where resonant operation is not satisfactory, given a suitable choke design. This statement has not been checked above 3000 volts and the experimenter may get into difficulties higher up on the scale. A distinct exception must be made in the case of regenerative operation — which is to say either in transmission or in receiving with a regenerative stage or one imperfectly neutralized. Here a resonance peak on a choke is a nuisance since a change of tuning may throw the choke reactance positive or negative with an unpleasantly strong effect on the degree of regeneration. In the case of the receiver this may be dodged by using the types of chokes which have multiple resonances in sufficient number to cause the peaks to tie together throughout working range. Abrupt changes of voltage distribution in the choke may be found but these are harmless. At ultra-high frequencies, corresponding to wavelengths below 20 meters, it is best to arrange circuits so as to minimize the need of r.f. chokes, since it is almost impossible to combine satisfactory choke performance and good mechanical arrangement. Where chokes must be used a special effort must be made to decrease the inter-turn and the end-to-end capacity. Very small "basket weave" or Lorenz coils are fairly satisfactory if the wire is not over number 34 and the leads are continued for several inches with the same wire before reaching larger wire or metal parts. A diameter of -|" works out well and may be made on a form consisting of five slim wire nails driven into a board and beheaded. WHAT MANUFACTURERS MAKE AND BUY have rearranged these data in Table IV. In the preceding paragraphs we have pointed out some of the salient facts obtained from the survey. It is hoped the executives of the various set manufacturing companies will be able to obtain some useful information from the various tables. It should be helpful to compare the practices of the various companies who cooperated in this survey with the practices in one's own plant. Whether or not one makes a particular part will depend on the manufacturing facilities available, keeping in mind always that outside manufacturers who make a specialty of a particular item might, in many cases, be able to build it more economically than could an individual set manufacturer. Parti 238 • (Continued from page 205) cularly in the case of the small set manufacturer, the more parts he purchases on the outside, the more space and time can he devote to the making of complete sets. However, among large set companies with adequate manufacturing facilities available, the tendency, as indicated by our survey, is build as many parts as possible in the plant. The companies from whom data were obtained for the figures in this article are: All-American Mohawk Corporation American Bosch Magneto Corporation Alwater-Kent Mfg. Company Crosley Mfg. Company Day-Fan Electric Company Federal Radio Corporation Chas. H. Freshman Company Grigsby-Grunow Company • AUGUST 1929 • Kellogg Switchboard & Supply Company National Carbon Company Philadelphia Storage Battery Company Sterling Mfg. Company Stromberg-Carlson Telephone Mfg. Company Temple Corporation MAJESTIC SERVICE FOR DEALERS Grigsby-Grunow is now publishing, for dealer distribution, an illustrated rotogravure paper called "The Voice of the Air." It is supplied fortnightly. In addition to many news pictures of general interest, the new publication includes a complete timetable of network programs, pictures of radio stars, and a radio log of all leading stations, divided into four geographical areas. The first issue, it was announced, had an edition of 1,170,000 copies.