Radio broadcast .. (1922-30)

Report on Experiments With Units of Various Designs CHARACTERISTICS OF A.F. TRANSFORMERS miimiiiii iimimiiiiiiiiimmimimiiiiimi MIIIIIIMHIIIIIHIIIII i: iiiiuiiitiiiMiiiim imimiiiiiitui MiiHimiMiimiiiiiiiimiiiiiiiiiiiiiiiiii itiiuti imniH By PROF. H. M. TURNER Sheffield Scientific School, Yale University S A RESULT of the keen 'competi- tion among manufacturers, stimu- _ lated by the demand of the radio public for better and better quality, there has been marked improvement in the per- formance of audio-frequency transformers during the past few years. The subject has been extensively investigated by com- petent research engineers and designers but the results of their studies come to us in the form of the finished, product with little information'as to the exact changes that have brought about improved perform- ance. There is in no sense a criti- cism of the manufacturer, for such information is his stock in trade. However, one of an inquiring mind desires to know in what way and to what extent the characteris- tics are modified by changes in design or operating condition. What constitutes a good transformer? This, of course, depends upon the use that is made of it. For broadcast reception of music, the primary consideration is quality or faithful reproduction which requires, assuming an ideal loud speaker, that the amplification be essentially independent of frequency over some predetermined band, say from 30 to 8000 cycles. However, with corrijercial loud speakers, considera- ble departure from a flat characteristic is permissible, and in some cases even desira- ble, in order to compensate deficiencies in the reproducing unit. For telegraph pur- poses the primary consideration is intens- ity or large amplification. The band may be much narrower, say from 800 to 1200 cycles, for a peaked characteristic is quite desirable in that it may be made to give a louder response to the desired signal and at the same time greatly reduce interfer- ence from neighboring channels. Unfortunately it is not always possible to obtain excellent quality and high ampli- fication from the same transformer. The response characteristic of transformers, however, may be modified materially by design and operating conditions as follows: the amount of iron in the core; the ratio of turns; the actual number of turns in the windings and their position with respect to each other and to the core, that is, whether the windings are interwoven or the primary or secondary is placed next to the core; the total resistance of the plate circuit; the plate voltage and grid bias of the amplifying tube associated with the primary of the transformer; and, to some extent, the capacity in parallel with the primary and secondary. Object of Investigation The object of this paper is to report on some experiments on audio-frequency transformers of different design and oper- ated under widely different conditions with the hope of at least partially answer- ing some of the questions that occur to one but which in some cases remain un- I I I I I I I I I I I I I 1334567 I- ELECTRON TUBE. VOLTMETER Fig. 1 120 110 100 50 4 6 8 10 EFFECTIVE VOLTAGE Fig. 2 answered due to lack of time or adequate facilities for conducting the necessary experiments. No attempt is made to define or to determine an ideal transformer but rather to show the effect produced upon the amplification characteristic by certain changes many of which are far removed from those encountered in practice. 2.75' FREQUENCY IN KILOCYCLE5 Fig. 4 CORE LAMINATION Fig. 3 Faithful reproduction is dependent not only on the transformer but also on the associated amplifier, loud speaker, and the response characteristics of the receiving ear. But, since the imperfection of the latter are present whether one hears the original program or a loud speaker repro- duction, it may be left out of consideration here. Although the ear must be the final judge of both quality and intensity it is not a satisfactory instrument for labora- tory determinations. In the first place the ear, although sensitive to changes in pitch, is quite tolerant of changes in in- tensity; in other words, variations of 10 to 20 per cent, or more may take place with- out being perceived. This is fortunate for, if the response were directly proportional to the stimulus, an organ of hearing suffi- ciently sensitive to catch the slightest whis- per would be destroyed by sound waves of great intensity, or at least they would be a source of great pain, 'In the second place there would be disagreement among ob- servers. Differences that would be apparent to a trained ear would • be entirely overlooked by one less discriminating. Even the results obtained by a given observer vary from day to day. Thus, for measurement purposes the ear cannot be considered an instrument of precision and for this reason an electron-tube voltmeter was used to measure the amplification as determined by the output voltage of the transformer and its associated amplifying tube. The Electric Circuit The circuit arrangement that was used is shown in Fig. 1. A variable frequency is obtained from an SA oscillator whose out- put transformer at the left supplies a volt- age which is maintained at a constant value, usually 0.25 volt to the grid of a 201A amplifying tube in the plate circuit of which is connected the primary of the transformer. The secondary or output voltage being measured by an electron- tube voltmeter of the type developed by E. T. Dickey of the Radio Corporation. The condenser across the secondary of the transformer, shown dotted, was intro- duced in some of the experiments to be referred to later. The Magnetic Circuit Where the transformers were assembled in the laboratory, a core of 15 mil hipernik was used, the other dimensions being given in Fig. 3. In all cases, unless otherwise indicated, the core consisted of 50 lamina- tions, the exception being Figs. 4 and 7 where the amount of iron was varied. Series I: This transformer consisted of a primary of 7000 turns of No. 40 enameled wire having a resistance of the order of 2400 ohms and a secondary of fREQUENCY-VOLTAGE AMPLIFICATION PRIMARY NEXT TO CORE 3 4 5 FREQUENCY IN KILOCYLES Fig. 5 NOVEMBER 1929 • 43