Radio Broadcast (May 1928-Apr 1929)

170 RADIO BROADCAST JANUARY, 1929 c 0.8 a 6a+9b+10c (j.h high frequencies, although many experimenters have advocated the use of a condenser because of improved low-frequency response. What is the cause? At high frequencies the transformer can be looked upon as in Fig. 2 in which L, the primary inductance, is shunted by all the distributed and stray capacities, C, of the circuit. Li is the leakage inductance, RP is the tube's plate resistance and R is the C-bias resistor. Now the leakage inductance and the capacity form a series-resonant circuit. The effect of L shunted across C is negligible, since it is a high-impedance shunted across a lower impedance. When the resistor, R, is not bypassed a considerable voltage is developed across it, due to the resonant current tlowing through the resistance in this circuit. This voltage is introduced into the amplifier so that it is out of phase with the voltage from the signal? on the grid of the tube. In other words it detracts from the amplification at this frequency. When the resistor is properly bypassed, the voltage drop here is greatly reduced, and, of course, the out-of-phase voltage introduced into the amplifier is reduced, so normal gain is experienced. Whether or not the resistor should be bypassed depends upon conditions. For example, there is usually a tendency for an amplifier to sing at the point where the capacity resonates with the leakage inductance. The tendency for the amplifier to sing, due to this resonance condition, then, is decreased when an out-of-phase voltage is introduced, due the C-bias resistor. Some amplifiers which do not sing when the bias resistor is not shunted, do sing when it is bypassed. In the case of the amplifier measured in the Laboratory, there was a gain of 6tu at 6000 cycles — which went a long way toward making up the usual loss at this frequency, due to only ordinary side-band cutting in the r.f. amplifier. Cycles 60 100 2000 1000 6000 Without C TU -.5 -1.0 -2.5 -6.0 With C TU +.5 0 0 0 -.7 The task of Editing Radio Copy FEW writers for and readers of a magazine like Radio Broadcast realize the complexity of the tasks of its editorial and technical staff. Let us consider a how-to-make-it article, perhaps on a prominent kit from a well-known manufacturer. The kit comes to the Laboratory, is tested, accepted, or turned down. Then the article is looked over, diagrams are checked against lists of parts, the photographer is called in, and, after the result of his labor comes to the office, an "over-lay" is made, that is, the photograph is overlaid with a thin sheet of paper and the various condensers, resistance, coils, etc., are marked with letters and numbers which correspond with the list of parts and the circuit diagram. Here is where trouble begins. The set, the list of parts, the diagram, and the article which comes from the outside, from the kit manufacturer perhaps, seldom — may we say, never? — check. On the diagram a resistor may be marked as 50,000 ohms, in the list of parts it is 100,000 ohms, and a^N< 9a+10b L= a2N2 ia +11 c FIG. 1 in the article itself it may not be mentioned, or it may have a third value. Which is correct? A recent article came to the office — late as usual — and still later came the receiver. The list of parts did not check either the diagram, or the receiver, although the manufacturer claimed we would have no trouble because the material was "just as he sent to his would-be purchasers." After exchanging several telegrams and longdistance phone calls, a list of parts, a diagram, and a photograph were assembled which checked — but unfortunately this list will not check anything the manufacturer sends out. What is the reader to do? Why cannot the manufacturer check his material before it gets into print? We have a bulletin sent out by a well-known manufacturer, this time describing a B-power system which makes it possible to "get away with" smaller filter condensers. The diagram sent out with the bulletin gives one value of condenser; the circuit diagram gives another. Which is correct? Who knows? We have another yarn from a publicity writer of a nationally known organization — we are going to turn it down — in which the list of parts gives several items which do not appear on the circuit diagram, and the diagram gives two items which do not appear on the list of parts. The Technical Staff feels that its responsibility is to the reader. It will get up a list of parts which will work properly as evidenced by a test in the Laboratory — and if the manufacturer is foolish enough to send out material which not only conflicts with what we print but which contradicts his own printed matter, it is his own fault. Incidentally, every receiver and power supply is tested in the Laboratory before it is described in the magazine — some of them several times, as well as many aggregations of apparatus whose descriptions never see the printed page of this magazine. THE CONFLICT between simplicity and accuracy rages in the soul of every technical man. Is there not some short cut to a mathematical or laboratory investigation that will give sufficient accuracy? What Empirical Rules and Formulas factor can we neglect in order to obtain the result sooner, and still not have the bridge we are designing fall down? It has been said that nearly any result may be obtained from a mathematical analysis of a given problem, providing the proper assumptions are made — and, as every physicist knows, many, many, problemscannot be solved completely at all. There are always some factors which must be neglected in favor of others. In the hunt for simplified methods, certain empirical rules and formulas appear. Several have recently been published which are very interesting. One was in Experimental Wireless (England) September, 1928, and this was reprinted in Lefax, October, 1928. It related to a simple tube tester with which a service man could quickly determine the value of tubes. In the course of many measurements Marcus G. Scroggie, who developed the tube tester, discovered that the following expression would "work" with all modern tubes with a fair degree of accuracy. Ro = 0.6 V where Rp is the plate resistance of the tube IE is the plate current V is the "lumped voltage" on the plate. The lumped voltage is the effective voltage on the plate of the tube; that is, it is the sum of the voltages Ep (the voltage due to the plate battery) and Eg (the voltage due the grid). For example, if a tube has 90 volts on the plate, a C bias of 4.5 volts, and an amplification factor of 8, the lumped or effective voltage on the plate is V = Ep+m-E,, = 90+8(— 4. 5) = 54 so that the effective voltage on the plate is 54 volts. Now, if the tube has a plate current of 2.5 milliamperes under these conditions, the plate resistance, RP, can be obtained from Mr. Scroggie's formula. We have taken this formula and computed the figures below which give the measured and calculated Rp of several tubes and the discrepancy between them. The table follows: Tube Rp (meas) Rp (calc) % 199 15,500 14,500 —6.45 201a 11,000 12,900 +15.4 226 9,400 8,700 —7.45 112a 5,000 5,400 +8.0 171 2,000 1,750 —12.5 210 5,000 4.825 — 3.5 250 2,100 1,430 —32.0 FIG. 2 Another set of empirical formulas appeared in October Proceedings I.R.E. and were developed in the Hazeltine Laboratories by Harold A. Wheeler. They relate to the inductance of three types of coils illustrated in Fig. I. The inductance of the coils may be calculated with good accuracy by using the formulas on the diagrams. For example, the formula for the multi-layer coil is accurate to within 1 per cent, if the three terms in the denominator are about equal, the formula for the solenoid inductance is accurate to 1.0 per cent, if the length of winding is greater than 0.8 times the diameter, and the formula for the single-layer spiral or helical coil is accurate to 1.0 per cent, if the dimension (c) is greater than 0.2 the dimension (a). All of the dimensions must be in inches to be used in these formulas. The inductance of such coils may be computed with one setting of this slide rule, and without consultation of complicated tables or correction factors. — Keith Henney