Dell J,a6ora/ories Record Volume Three OCTOBER, 1926 Number Two The Evolution of the Input Transformer By F. E. FIELD INPUT transformers started and the input side: i.e., the receiving their history but not their pres- element of the repeater. Its function ent characteristic name with the was to connect line and repeater with- commercial installation in 1904 of out excessive transmission loss of telephone repeaters of the electro- voice currents due to unmatched im- magnetic type. Transformers of pedances, and to prevent transmis- various kinds and for different pur- sion of low frequencies incident to poses had been in use in the telephone the use of the repeater on composited system for some years. In general lines carrying telegraph signals. their function was to insulate one This repeating coil had a low ratio portion of a telephone circuit from and a low impedance. another. Thus, a toll line would be When a vacuum tube, suitable as an terminated by one winding of a trans- amplifier of voice frequencies, had former, the other winding of which been developed a need arose for a re- could be connected to the local cir- peating coil with ratios and impe- cuit. The local and long distance dances previously not required in the portions were thus insulated, but the telephone plant. And with this need speech currents in either portion there was a new emphasis upon the were, by induction, repeated in the word "input" and the introduction of other, and hence, the name of "re- the term "input transformer" to de- peating coil" for such a transformer. scribe the coil used to connect the line The ratio of the turns in the two sets and the receiving side of the repeater of windings was, of course, chosen element. with reference to the impedances of The input impedance of a vacuum the lines between which the coil was tube -that is, the impedance between to repeat. the grid and the filament -is very When the mechanical repeater was high, approaching infinity under cer- inserted in a stretch of long distance tain conditions; and a transformer circuit, a transformer called a repeat- was required to transform-i.e., "step ing coil was used between the line up" -the low impedance of the line 433} www.americanradiohistory.com to match the high impedance of the different type, therefore, was de- tube. Development work by E. O. signed by R. L. Jones in the early Scriven was started in February, part of 1914. That transformer, 1913, toward the production of one shown in Figure 3, was of the shell with as high a step -up ratio as could type. It had less distributed capacity be obtained with satisfactory trans- than the straight -core type and there mission of voice frequencies. The was some improvement in the range design then developed had a straight of transmitted frequencies, as is shown in Figure 2 by the characteris- tic dated 1914. This type of trans- former was used in the first demon- stration of the Transcontinental Tel- ephone Line. A considerable number of trans- formers of this type were built for use in the early vacuum -tube repeat- CT] ers. Difficulty was experienced, how- ever, in their manufacture, Fig. 1- Design of 1913; size 9 by 3 inches particu- larly in the winding. With the in- silicon -steel core and was wound in a creasing demand for input trans- number of sections to minimize the formers it was soon desirable to re- distributed capacity, as shown in Fig- design for commercial production. ure i. Before assembly in the re- The type of structure shown in Fig- peater equipment it was mounted in a ure 4 was, therefore, evolved by W. wooden box filled with moisture - L. Casper, and the first model of a proofing compound. Its voltage -am- transformer of this design was built plification - frequency characteristic in March, 1915. It was later known is shown in Figure 2, and indicates a as the 203 -type input transformer. range of transmitted frequencies Its construction is well suited to com- wide enough for intelligible speech. mercial production; the spools and This type of transformer had a windings are readily made and the very large external magnetic field assembly of the core with the spool which proved troublesome in the op- involves very much less labor than eration of the repeater. An entirely the assembly of the earlier design. 1926 ----- ......... i.02-3 --' i 9 A ilk V ` 100 1000 10000 FREQUENCY (LOGARITHMIC SCALE) Fig. 2 -How the voltage -amplification characteristics of our transformers have been improved in thirteen years 434} www.americanradiohistory.com From an electrical standpoint the chine. Also, the inertia of the mov- new transformer was practically the ing parts of the winding machines equivalent of the preceding design then in use was so large that wire and was entirely satisfactory for the smaller than No. 31 could not be used transmission of speech. without an excessive number of break- The 203 -type transformer, how- ages. Due to these limitations on the ever, was large, heavy and expensive. To obtain the desired transmission for voice frequencies a very high im- pedance was required and that meant many thousands of turns of fine insu- lated wire. Since the Manufacturing Department had not at that time de- veloped the technique of winding No. 4o black -enamel insulated wire it was necessary to use much larger wire in- sulated with enamel and silk. The amount and kind of wire required, therefore, were directly responsible for the size and weight and much of Fig. 4-Design of 1915; size 6%4 by 454 the cost. inches In 1916 the problem of designing smaller and cheaper repeater equip- windings, an amplification as large ment placed upon the coil develop- as was desired could be obtained only ment group the necessity of develop- at a sacrifice of the low frequencies. ing corresponding transformers. In It was not desirable, however, to im- this development the use of standard pair the quality of the repeater oper- ation in this manner. Up to that time (1916) the use for input transformers had been con- fined practically to telephone repeat- ers. In these, the amplification of frequencies below 135 cycles and above 2300 cycles was not desirable; in fact, it was decidedly objectionable, and means were provided elsewhere in the repeater to reduce the amplifi- The band Fig. 3- Shell -type design of 1914; size 6yß cation of these frequencies. by 4 inches of frequencies which the trans- former was required to transmit was parts and standard methods of con- therefore defined at about 200 cycles struction were adhered to as far as to 2300 cycles and the trend of the possible in order to secure a low cost. development had been along the lines The toroidal construction of Figure of producing smaller and cheaper 5 was selected. The number of turns transformers rather than in attempt- of wire, however, which could be ing to extend the range of trans- wound on the core was limited by the mitted frequencies. size of the shuttle on the winding ma- Furthermore, since it was desirable 13 5} www.americanradiohistory.com Fig. 5- Toroidal transformer, mounted on a block 6 x incites to maintain the input impedance of A number of designs were pro- telephone repeaters approximately duced in the years 1916-1918, but only constant over the transmitted fre- two of them combined advantages of quency- range, each of the transform- cheapness and efficiency which en- ers so far mentioned was designed to abled them to survive. They are be used with a non -inductive resist- the 201 and 213 types shown in Fig- ance shunt across its windings and ure 6. Many of the amplifiers trans- the characteristics shown were taken mitted frequencies below and above under that condition. The resistance the previously adopted range of 200 also served to flatten out the ampli- to 2300 cycles, for considerable prog- fication characteristic, producing a ress was made in extending the fre- more uniform amplification over the quency range. frequency range. The progress in amplifier design Soon after the United States en- made during these years led to the tered the World War the demand decision in 1918 to improve the qual- for amplifiers of various kinds added ity of the repeaters on the transcon- a great stimulus to the development tinental telephone circuit. Since the of input transformers, especially of quality of the transmission of the the smaller and lighter types. Their voice frequencies depended largely on development required not only the so- the input transformer, that part of lution of many theoretical problems the repeater was selected as affording of design but also the development in the greatest possibility of improve- the Manufacturing Department of ment. Naturally the types of trans- the technique of winding small sizes former which were investigated were of insulated wires. This accomplish- those recently developed for small ment at Hawthorne was one of the amplifiers. Of these the 201 type was principal contributions to the realiza- selected, since it was small, inexpensive tion of small, highly- efficient input- and readily adaptable to quantity transformers. production. Furthermore, the Manu- 436k www.americanradiohistory.com facturing Department had had ex- This transformer, as well as those perience in making this type and had mentioned later, is designed to be developed methods of winding. used without any shunt resistance and It was next evident that very little the characteristics shown for it were more could be done in reducing the taken under that condition. size of the input transformer and that Further development in the design any further improvement would have of transformers, with the object of to he in extending the frequency improving the amplification at the range or reducing the cost.