Radio broadcast .. (1922-30)

Knockabout Wavemeters THE EXPERIMENTER'S ARMCHAIR iiuimiiiimiiiiimi BEFORE GETTING down to the main discussion I should like to make a suggestion as to the usefulness of dividing one's wavemeters into two classes. A good example is found in the practice of the Radio Frequency Laboratories, Inc. The precision meters of this establishment are reserved for precision work; during the preliminary adjustments simple and con- venient knockabout wavemeters, such as shown in the picture (Fig. 1), are used. If one of these is damaged replace- ment and recalibration are easy and prompt. To save guesses, I will explain that the two large coils are wound on General Radio type 277u forms, the variable con- densers are of National Company make, and the two smaller coils are of 3-16th-inch copper tubing. The long baseboard not only carries a calibration chart but also provides room for the long insulating shaft and the National type A vernier dial which materially facilitate readings. The wavelength ranges in meters in this case happen to be 1.85—6.1, 5.5—27.2, 25.0—122.5, and 118.0—565.0. Lamps are placed in the tuned circuit of the two larger wavemeters but are omitted from the smaller ones for the obvious reason that their in- ductance would be a considerable proportion of the total and would be changed on replacement of the lamps. By ROBERT S. KRUSE a complete change-over should be quite painless. What are the opinions of the congregation? Thirty-Megacycle Reception 1 us get down to the main topic. Our material this month comes from C. A. HART, whose location has been in rapid succession at San Diego, New York, San Pedro, and wherever else the U. S. S. California has gone. The corre- WAVELENGTH VS. FREQUENCY The wavemeter charts are in dial setting against wavelength—not frequency. This recalls a discussion held last month. A group of six experimenters agreed thoroughly that frequency was logi- cal as a basis for calculation but was not equally convenient for measurement, and amounted to a positive difficulty during rough preliminary work. The point is, of course, that wavelength is related to the size of things while frequency bears an inverse ratio. During the preliminaries much time is saved if one may estimate the needs without the mental contortion of taking reciprocals of everything. If one insists on talking frequency the only way to avoid such a thing at every turn is to think in reactances instead of inductance and capacities. This in itself involves some detours. At short waves it is furthermore an infernal nuisance to say "Sixty-thousand kilocycles" when "5 meters" will do just as well. I was quite tickled to find that from the Bell Telephone Laboratories there emerged the same opinion and this was printed in the I. R. E. proceedings. I am not quite sure whom the joke is on because one of our I. R. E. Committees has proscribed the meter—and I belong to the Committee! Incidentally, if frequency is preferred lubberly numbers may be avoided by speaking in megacycles whenever dealing with the territory below (I have fallen back on wavelength again) the ordinary broadcast band. I believe Dr. Pickard is the sponsor of the megacycle terminology. It is in sufficiently general use now so that Fig. 1 — Knockabout wavemeters are used at the Radio Frequency Laboratories for preliminary adjustments. This picture shows four such instruments of different wavelength ranges. spondence has accordingly taken some time to develop what follows. Let us first look at Mr. Hart's experiments. A receiver was built up with the conven- tional oscillating detector and a.f. ampli- fier for the purpose of going somewhat into the necessities in the region of thirty megacycles. Because such receivers are notoriously weak as to holding calibration, the next step may seems surprising, but this will be explained presently. The curve of Fig. 2 was made by heterodyning the receiver with a crystal-controlled driver. It will be seen that the range of the receiver was from 27.62 megacycles to 45.16 megacycles (10.85 meters to 6.64 meters). For this range the grid coil con- sisted of one turn and the tickler of 2| turns of No. 18 enameled wire wound on a standard ux tube base and tuned by a " 5-plate" variable condenser. The lengths of connecting leads are shown in the insert diagram of Fig. 2. The ability to take and hold calibration requires the removal of tuning effects from the operator's hands, the antenna, and the regeneration control. If really good permanence of calibration is desired one must also be careful to avoid changes due to shifts of filament and plate voltage, the commonly neglected changes in ordinary r.f. chokes, and, of course, one must keep (and not mistreat) the same tube. Hand capacity was removed by the use of extension controls rather than shielding. The effect of antenna variations or of changing to another antenna is very large when coupling in the customary amateur manner through a condenser connecting the lower end of the antenna to the upper, or grid, end of the tuned circuit. Therefore, magnetic coupling was used with a primary coil of three times the diameter of the secondary. With a 30-foot antenna ade- quate coupling is obtained when this coil is placed about 1J" from the secondary, whereby the capacity effect is made sufficiently small so that the differences between various thirty-foot antenna's are of no consequence. One ac- cordingly needs to carry only a 30- foot piece of wire and put it up in any convenient manner which will keep it reasonably clear of things. The teeth of the regeneration con- trol were drawn by the simple ex- pedient of cutting and trying until a combination was obtained which permitted leaving this control en- tirely alone while at the same time giving continued smooth oscilla- tion when tuning across the whole scale. This is in line with the sound amateur tradition of throwing out those things which make trouble. The necessity of doing this while not forcing the tube accounts for the abnormal tickler. The regener- ation condenser size is not stated but a small photograph suggests that its capacity runs to several hundred micromicrofarads. To secure uniform operation across the tuning range the single- layer chokes common to transmis- sion were avoided in favor of a type likely to give a broader response though possibly not of as high an impedance. Three pairs of 1" bakelite squares were strung on a quarter inch bakelite rod. These were located so as to form three slots \" wide and J" apart. In the narrow slots were placed scramble- wound, windings consisting of 100 turns each of No. 28 s.c.c. wire, the three wind- ings being connected in series. Smaller wire in the same form proved inferior. With this combination smoother oscil- lation was produced than with any com- bination using higher filament voltages and smaller regeneration capacities in the usual manner. The detector plate voltage had to be kept up to 90 and a 5-megohm grid leak was used. The a.f. stages were run at lower plate voltage than the de- tector. Concerning this point Mr. Hart says "This may be explained by quoting from Van der Bijl's Thermionic Vacuum Tubes, page 214. 'The higher plate po- tential, of course, gives a higher amplifica- tion because the plate resistance of the tube is lower. It is seen also that the amplification at 1000 meters is about 3 times as large as the amplification at 100 meters.' Since without amplification there can be no oscillation, it follows that for 10 meters or less the plate potential must be high." With this receiver changes of filament voltage such as ordinarily made had very slight effect on the beat frequency, which is unusual for such receivers. The noise level was lower than the other receivers used in the 15-40 meter region, probably because of the loose magnetic primary JULY • 1929 • 165