Radio Broadcast (May 1929-Apr 1930)

(A) NO THICKNESS WAVE METERS/ MILL 1 0.0515" 198.0 3.85 2 0.0642" 167.5 2.61 3 0.0410" 159.0 3.88 4 0.0410" 159.0 3.88 5 0.0575 " 155.0 2.66 6 0.0367 " 142.8 3.89 7 0.0535" 139.0 2.60 8 0.0535" 139.0 2.60 9 0.0480" 124.0 2.59 10 0.0460" 120.0 2.61 Fig. 3 fundamental of the crystal, in order to keep it oscillating. The second circuit was first tuned to 79.5 meters — the crystal's second harmonic. The point of maximum voltage occured at only one point on the dial of C2, for any one position of Ci. However, the magnitude of the second-harmonic voltage did vary, as C<i was tuned, in the manner of curve E2 of Fig. 5. As the condenser in the fundamental tank circuit was brought up to the crystal wave, the second harmonic voltage at first increased, reached a maximum, and then dropped off until the oscillations again broke. The interesting point is that maximum harmonic voltage is not obtained at the same setting of the fundamental circuit dial as gives maximum fundamental output. When the trap, L2C2, was tuned to the third harmonic of the crystal, the same type of radio-frequency voltage output curve, E3, resulted. The curves for a perpendicular-cut crystal, No. 7, were in no way dissimilar to those for the parallel-cut specimen. harmonics that are a goodly proportion of the fundamental can be obtained directly from the crystal oscillator, so that it is possible to eliminate one or more frequency multiplying amplifiers in a transmitter layout. It may be noted that a grid leak was used on the oscillator tube. The relative strength of harmonics could probably be changed, and favorably, by using grid bias. A more intensive study on the matter of harmonics should certainly prove fruitful. Radio-frequency voltages across the tuned circuits were measured, instead of reading the tank-circuit currents, because the voltages are more indicative of the true output of the oscillator. Voltage, only, is useful in feeding the next amplifier. The tank-circuit current for any one harmonic could be made to vary widely by merely changing the ratio of inductance to capacity, so that readings of current would mean little. The radio-frequency voltages were measured by impressing them on the grid of a vacuum tube, and then bucking out with direct voltage until the plate current returned to its initial value of 100 microamperes. The direct voltage, minus a small correction for initial bias, was, of course, equal to the peak of the radio frequency. The voltmeter tube was fitted with long leads to the battery and meter assembly, so that it could be brought right to the tuned circuit which was to be measured. By-pass condensers to filament were connected at the tube. The picture on the next page and the diagram of Fig. 2 give further details of the electrical and mechanical designs of the meter. Frequency variation in crystal oscillators 140 120 100 60 40 20 Parallel Vs. Perpendicular Cuts THE maximum voltages obtainable at the optimum tuning of both condensers, for harmonics up to the fifth, were determined. In Fig. 6, the curves , show for four crystals, two parallel-cut and two perpendicular-cut, how the strengths of harmonic and fundamental voltages compare. The fundamental was measured across circuit L1C1, and the harmonic voltages were measured across circuit L2C2 when it was tuned to the given harmonics. Under the test conditions, the parallel-cut crystals gave a greater magnitude of radiofrequency voltage than did the perpendicularcut plates. Also, any given harmonic of the parallel-cut crystal was a greater percentage of the fundamental than was the corresponding harmonic of the other type. In either case, A.C. 40 45 50 55 60 65 DIAL SETTING, FUNDAMENTAL TANKCIRCUIT 70 11 POWER PACK CRYSTAL OSCILLATOR RECEIVER BEAT-FREQ. OSCILLATOR CRYSTAL OSCILLATOR POWER PACK TT AC. Fig. 4 Fig. 5 is a problem whose difficulty of solution varies directly with the frequency. That is, a broadcast station on 500 kc, having to maintain its frequency within 500 cycles of that which was assigned, has to work within limits of 0.1 per cent. A station assigned a frequency of 1500 kc, and required to hold that within 500 cycles, has to maintain an error of less than 0.03 5 per cent. Thus, as the operating wavelength of the station goes down, the percentage accuracy which has to be maintained for a given permissible frequency variation increases. The use of harmonics of a crystal oscillator does not improve matters in this respect, except indirectly. If the frequency variation of a 400-meter crystal is 50 cycles, say for a certain change in temperature, the frequency variation of its fourth harmonic, at 100 meters is 200 cycles. A 100-meter crystal whose temperature was changed the same amount would have exactly the same frequency variation— i.e., 200 cycles. The change of frequency with temperature is due to the expansion of the crystal; and the above conclusions are evident from inspection of the equations for thermal expansion. Indirectly, however, frequency stabilization is aided by using crystal harmonics, because the thicker 140 120 100 < o 80 > ill of 60 40 20 \« \ V N \ \ *, 142 Me ter Crystal Cut \ ♦ x 1 X 139 J. 139 -METER PERPEND1CULAR 159-MET CUT ER PAR CRYST . \LLEL Cl JT CRYi TAL Ei / / ll ✓ r \ I 1 ^Ez — ✓ — ✓ » // 1 E3/ / 1 i \Es \ E, y HARMONIC Fig. 6 crystals are much easier to handle and to mount. Matching Crystals TWO pairs of crystals were cut and ground to approximately zero beat. One pair, crystals No. 3 and No. 4, were cut parallel to a face of the raw quartz. By careful hand grinding with fine carborundum, and continual checking against each other as oscillators, the frequencies of the crystals were brought within one or two hundred cycles. The fundamental wavelength of the two was 159 meters. Two other crystals, No. 7 and No. 8, cut perpendicular, were ground and matched in the same way at a fundamental wavelength of 139 meters. In order to measure frequency variation of the crystal, the layout depicted in Fig. 4 was utilized. The second harmonics of the crystals in the 50-to1 00-meter band were to be used for controlling the transmitter, so the receiver was tuned in this band to the second harmonics of the two crystals. The crystal oscillators were spaced some 15 feet across the room, and were run from entirely separate A and B supplies. The frequency of the audible beat note, present at the receiver output, due to the two harmonics, was measured with a beat-frequency oscillator. For this purpose a 10-mmfd. semi-circular plate condenser was used as a vernier on the beat-frequency oscillator. When calibrated against a Western Electric 8a audio oscillator, a linear frequency variation of slightly more than 300 cycles was had over the dial. Thus, by beating the sound output from the receiver against that from the calibrated audio source, changes in frequency of one cycle in the crystal oscillators could be detected. One of the oscillators was left 75 30 31 32 33 34 TEMPERATE! RE-DEG.C Fig. 7 36 • may, 1929 page 24 •