Radio Broadcast (May 1928-Apr 1929)

SSZEHE KROAnCASIhK SrhTTT KY (AMI Broadcast Transmitters IN THE May, 1928, issue of the Proceedings of the Institute of Radio Engineers Mr. I. F. Byrnes presents a paper on " Recent Developments in Low Power and Broadcasting Transmitters." Mr. Byrnes is a member of the staff of the General Electric Company, and his paper is a resume of the recent work of the G. E. transmitter experts in this field. The latter portion is of particular interest to broadcasters. The design requirements in modern broadcast transmitter construction are summarized by Byrnes as follows: 1. The carrier wave must be readily maintained within 500 cycles of the assigned frequency, and its frequency must not fluctuate with modulation. 2. The electro-acoustic characteristics of the equipment must be such that the output of the set is as faithful a copy as possible of the input. (The frequency transmission characteristic from microphone to antenna must be sensibly horizontal within the principal portion of the audio band.) 3. The circuits and tubes must be engineered so that the equipment will have ample overload capacity and modulation peaks will not cause distortion through "overshooting" at any point. Illustrations of a 1 k.w. broadcast transmitter (ET-3633) are given in the paper. Normally this transmitter is built for operation on a frequency between 666 and 1200 kilocycles, although it may readily be modified for the other frequencies within the broadcast band. A condenser transmitter is supplied in the studio. The case of the condenser transmitter contains the first audio tube, which is of the ux20 1 a type. There follow in the audio chain one other UX-201A amplifier, a ux-210 amplifier, two uv-211 amplifiers and four uv-851 modulators. In the radio chain the tubes comprise a uv-211 as the master oscillator, two uv-211's as intermediate radio amplifiers, and one uv-851 in the output stage. A uv-211 is also supplied as an oscilloscope rectifier to indicate the percentage of modulation. The ratio of modulators to radio amplifiers (4:1) will not be a surprise to those readers who recollect the article on "Modulation" in this department for July, 1926, and the reference to Kellogg's " Design of Non-Distorting Power Amplifiers" in the May, 1925, issue of the Journal of the A. I. E. E. For any sort of deep modulation it is necessary to supply about as much plate power to the modulators as to the oscillators or power amplifiers in the modulation stage, and as a modulator normally takes about a quarter of the power drawn by the same type of tube delivering radio-frequency energy, the answer is obvious. In the 1 k.w. transmitter the design of which is being outlined the provision of four modulators in parallel results in a very low-impedance bank which, in combination with a good-sized speech reactor, retains the low audio frequencies and keeps the characteristic flat at the lower end. The first three stages of audio amplification in the transmitter proper (following the UX-201A associated directly with the microphone and housed in the "bullet," as the cylindrical case of the earlier designs was called by the poets of the General Electric Company) are resistancecoupled. The condenser tube, however, delivers its output through a step-down transformer, so The power so lost is naturally great in the case of that the lead between the "bullet" and the set, the larger stations. The cooling systems of such being a connection between low impedance cir installations are built to get rid of around 200 cuits, may be made reasonably long (a few hun kw. — more or less of this capacity being used dred feet). The last 211 stage is reactance according to the power at which the transmitter coupled to the modulators. The third and fifth is run. The total dissipation may be calculated in audio stages have volume control potentiometers two ways, the results of which should agree, on the inputs. When the input from a line is One method requires a knowledge of the total used to modulate the set the first audio tube is flow of water and the rise in its temperature ow dropped out by means of a switch. The line, ing to its contact with the anodes. In a specific in other words, is expected to provide as much example, let us assume a water flow of 150 gallons energy as the condenser mike plus two 201 a a minute and a rise in temperature of 5 degrees tubes. Centigrade, corresponding to 9 degrees Fahren The radio-frequency chain of this transmitter heit. A U. S. standard gallon weighs 8.33 pounds starts with a master oscillator. Shielding is em at 72 degrees Fahrenheit. For the accuracy of our ployed to stabilize the frequency, and the coup calculation no correction need be made for an ling of the successive radio-frequency circuits other temperature. The water circulation is reduces harmonic radiation. There is no crystal therefore 1249.5 — 'e* us sav 12 5° pounds per control in this transmitter, however. minute. Now, a B.T.U. (British thermal unit) is Power is supplied to the plates from a motor the heat required to raise the temperature of generator set comprising three units : the motor, one pound of pure water one degree at 62 degrees excitor, and a two-commutator high-voltage gen Fahrenheit. In the case of this water system, erator. A separate two-unit motor generator therefore, we are carrying off, each minute, 1250 supplies d.c. filament energy. The audio filaments times 9, or 1 1,250 B.T.U. But from other relations are heated from a storage battery. These mea we know that 1 kilowatt-hour equals 3415 B.T.U. sures, with the addition of suitable filters, guard One kilowatt-minute, therefore, would corre against noise. spond to about 56.9 B.T.U. Dividing this figure The oscilloscope is similar to the oscillographs into the value of 1 1 ,250 B.T.U. per minute secured found in physical laboratories, except that it above, we get 197.7 kilowatts as the total heat contains only one vibrator and the light source loss in the tubes. The same total should result if is an incandescent lamp instead of an arc. It the individual losses of the various units are es is consequently unsuited for the taking of pho timated electrically and added up. Roughly, the tographs. The operator, looking at the revolving anode efficiency of an oscillator or radio fre mirror, sees a wiggling line the amplitude of quency amplifier in a broadcast transmitter may which shows to what degree the carrier is being be reckoned as 65 per cent. The anode efficiency modulated. The rectified carrier supplies a ref of a modulator may be more like 30 per cent, erence amplitude, while the wiggling line is the Taking, for example, a tube with a nominal audio component, likewise rectified. A loud oscillator rating of 20 kilowatts, we may find it speaker is also connected in this circuit, so that drawing a plate current of 2.0 amperes at 15,000 it provides both visual and audio monitoring volts. Of this 30 kw. input to the plate only about of the radiated energy. Fig. 1 shows the con two thirds will be converted into oscillating nections schematically. energy; about 10 kw. is heat loss. To this should Sets of the ET-3633 type are in use at cyj in be added the I2R of the filament, which in the Mexico City, and at Cornell and St. Lawrence case of such a tube will correspond roughly to Universities. 1 kw. (about 50 amperes at 20 volts). The total 21. water-cooled vacuum tubes {Part 2) volts> the 8rid being biassed negatively about an efficiency of about 20 per cent, in terms of Ieaving 6 „ kw as heat loss from the electrical 1 average carrier power in the antenna over input to the plate> to which must be added the power drawn from the a.c. mains. A transmitter former figure of , kw for the filament> 0r 7.0 kw. rated at 50 kw. m the antenna, for instance, will in al, similar calculations for the various tubes take around 250 kw. from the power company. in the transmitter should result in a total equal Most of this energy has to be carried off in the to the Ioss indicated by the temperature rise of 267