Radio Broadcast (May 1928-Apr 1929)

260 RADIO BROADCAST SEPTEMBER, 1928 FIG. 3 ment and plate power from the house lighting lines. On a load of about 135 watts, the device did not regulate the voltage — but of course it was severely overloaded under these conditions, making the test unfair. The figures above are conclusive evidence that under the conditions for which the device was designed, almost perfect regulation is obtained by its use. We hope the Acme Company gets at least half the million dollar business we predicted for a device of this kind! It is the first such device tested in the Laboratory. THE coils used in present Some Coil day receivers control to a ^Measurements large extent both the selec tivity and sensitivity of those receivers. All receivers with which we are familiar use the well known parallel tuned circuit — the "anti-resonant" circuit of the telephone engineers— shown in Fig. 3. The input to this circuit is usually a small coil acting as the primary. P, of a transformer whose secondary is the inductance, L, which is tuned by the variable condenser, C. The input of the following tube is placed across this shunt circuit. The impedance of such a circuit, at resonance, is expressed as Z where L is the inductance of the coil in henries w is equal to 6.28 times the frequency in cycles r is the series resistance of the coil and condenser in ohms. The impedance at any other frequency than the one to which the coil and condenser are tuned is lower than the above formula gives. At higher frequencies the impedance is essentially capacitive and low; at lower frequencies the impedance is inductive and low. Currents of other than the resonant frequency, then, do not build up such high voltages which may be applied to the following tube. Therein lies the selectivity due to such a tuned circuit. It will be increased as we decrease the series resistance. This is the reason for the "low loss" craze. A good coilcondenser combination is one which has little resistance in it and which has, therefore, a sharp resonance curve. Since a resistance in shunt to this circuit has the same effect as a smaller series resistance, the tube resistance, Rp, of the previous tube tends to decrease the selectivity of the circuit, especially if it is connected across the entire coil. This is one reason why a step-up transformer is used. It may be looked at as a kind of selectivity transformer. Instead of shunting the entire tuned circuit with Rp, which may be 10,000 ohms, we step-up this resistance to about 200,000 ohms by connecting it across only part of the coil or through a small primary coil. Multiplying Rp by the square of the effective turns ratio, N, which usually amounts to about 20, adds a much lower effective series resistance in the circuit than if the 10,000 ohms were across the whole coil. The greater the mutual inductance, the poorer the selectivity; the fewer the number of primary turns, or the looser the coupling to the secondary, the greater the selectivity. It is also true that the greatest voltage amplification will take place in such a system when the effective impedance in the plate circuit of the amplifier is equal to the plate resistance of the tube. That is, for maximum amplification N2 LJ to2 L Cr where N is the turns ratio of the transformer R,, is the plate resistance of the tube C is the capacity of the circuit When these conditions are fulfilled, by means of the turns ratio of the transformer, the maximum voltage gain, K, is ^max VT Under these conditions the series resistance of the tuned circuit is twice as great as it would be if the coil-condenser combination were unattached to the previous tube. The expression K above may be divided into TABLE 1 Coil F k.c. R ohms L D L/D Wire No. B & S Turns inch N Ins. Ind. 3" slight spacing 550 1000 1500 4. 1 6.5 9.5 1.09 3 .36 24 41 45 D.S.C. 170 3" Hammarlund Roberts 550 1000 1500 3.1 4.4 7. 1 1.69 3 .56 22 29 50 3" Special 550 1000 1500 2.6 3.7 6.0 2.25 3 . 75 22 23 52 Coil No. 1 550 1000 1500 7.2 11.1 14. 1 1.44 2 .72 26 56 80 Enam. 240 Coil No. 2 550 1000 1500 3.9 6.0 9.0 2.63 2 1.31 22 34.5 76 D.C.C. 168 Coil No. 3 550 1000 1500 3.7 5.7 8.9 2.1 2 1.07 24 35 74 s.s.c. 170 Note. F = frequency in kilocycles R = high-frequency resistance in ohms L = length of winding in inches D = diameter of coil in inches N = total number of turns Ins. = insulation Ind. = Inductance in microhenries FIG. 5 two parts, one giving the amplification due to the tube jt== and the other giving the amplification contributed by the coil — The factor due the tube is not the same as the mutual conductance, which is iVRp, nor is the coil gain factor equal to the "Q" of the coil, which is Lw/r. The selectivity of the circuit, however, is more nearly proportional to "Q" and so this figure of merit of a coil is useful. In Table 1, the first three coils listed are Hammarlund coils; the remainder are inductances of other manufacturers. This information is sufficient to enable anyone to calculate the "Q" of the coils, to estimate their selectivity factors, and to calculate the maximum voltage amplification that can be secured from a tube and coilcondenser combination of this nature. The effect of coil dimensions, of the L/D ratio, of the spacing of the wires, and of the size of wire will appear in these tables. It has been known for a long time that the most efficient coil is one whose length of winding is equal to the diameter, and it is interesting to note how nearly this rule holds for the coils listed in the tables. As has been pointed out in Professor Hazeltine's patent No. 1 ,648,808, reducing the primary turns to half the number required for maximum voltage amplification at resonance, reduces the resonant values only twenty per cent, compared to 50 per cent, at other frequencies. This means, of course, a material increase in selectivity. And in these days what everyone needs is selectivity! THE diagram in Fig. 4 Another is taken from Patent No. Patent Muddle 1,658,805, issued to Lester L. Jones, on a capacitive-coupling control system, dated Feb. 14, 1928 (original filed Dec. 15, 1922) and in Fig 5 is a diagram taken from Patent No. 1,648,808, issued to L. A. Hazeltine on a wave signalling system, dated Nov. 8, 1927, filed Feb. 27, 1925. We should like some technical minded reader to tell us how thest differ. It will be a good test not only of one's knowledge of radio circuits, but of how patent lawyer's minds work. On the basis of Mr. Jones patent, suit was filed against the Freed-Eisemann Company and an authorized Stromberg-Carlson dealer some time in May, 1928. Keith Henney