DISTORTION IN TRANSFORMERS

Radio Record, Volume I, Issue 5, 19 August 1927, Page 13

DISTORTION IN TRANSFORMERS

"SOME DIFFICULTIES

(4 Talt from BLO, Melbotrne by

H.

K.

Love

The incessant demand for faithful reproduction of sound waves in all their forms, that ig music, voice, etc., has led investigators to make a deep study of the requirements for audio frequency transformers, ‘The musical scale which the average ‘transformer is called upon to handle runs in a straight line of frequencies of from approximately 16 vibrations or cycles per second up to 8000, The human voice covers the range of, roughly, from 125 cycles to, say, 280 cycles, so that it will be seen that a transformer which may give excellent .reproduction of the voice is only being called upon to amplify a ; very small portion of the frequencies ' that have to be provided for. _ The lowest note of a pipe organ is ‘about 16 cycles per second; this note is so low in frequency .or vibrates so slowly that it is more frequently ‘easier to feel it than to hear it. A transformer is rarely expected to handle ' freauencies of such a low order as 16 ‘per second. The pianoforte range starts from a frequency of about 275 and runs up to: a shade over a 1000-cycles per second... The top notes of the violin go even higher. In addition to the proyision which has to be made in an ‘amplifying transformer for the natural ‘musical frequencies a big margin must .be allowed for what is termed the harsmonic range. This is explained in the | following way:-If the tep ote of Piano is strnck it emits a sound wave of, say, about 1100 cycles, but it also emits harmonic sound waves which are, to put it simply, multiples of the original note both above and below the troe frequency. It is on this account that a big overlap of amplifying ability mist be allowed for in any transformer which may be used to step up or step down audio frequency currents corresponding to anything within the range of frequencies previously mentioned. An Ideal Transformer. It would be ideal if a transformer could be designed which would give a straight line amplification curve over the whole frequency baud, but I am afraid that so far this has not been possible. Most transformers are designed to amplify correctly from 50 cycles to about 6000. Even within this range of frequencies there is a decided drop of as much as 10 degrees between the frequencies of 300 and 50 cycles per second, and on the upper range a drop of 3 or 4 degrees from 4000 to 8000 cycles per second, ‘The best transformer of present-day design should at least have a straight line curve between the frequencies 300 and 4000. This covers the, principal portion of a pianoforte range, most instruments, and a reasonable inargin for harmonic range. ‘The construction of a good transformer, which will be expected to give reasonable performance should be somewhat along the following lines: . CORE. A. Ample cere section with practically no air gap should be provided. INDUCTANCE. B. The inductance of the primary should be high at the average voltages used in wireless sets, say, about 80 henries. RATIO. C. The ratio should not be too high, but as high as possible consistent with low losses at high frequencies, WINDINGS. D. The windiugs should be sectioned or subdivisioned with a yiew to the reduction of self-capacity current and eddy current losses. The points imentioned in A, B, and C are made to ensure faithful amplification at low frequencies, and that mentioned in D ensures good amplification at high frequencies,