304
length used be somewhat longer' than the fundamental of the aerial, which is the usual condition of practical wireless telegraphy.
In all the discussions up to this point the use of sustained or undamped radio frequency current has been assumed. The generators indicated by the symbol E in the diagrams have been supposed to be radio frequency alternators of the Fessenden type, which produce continu- ous alternating current of a definite ra- dio frequency depending only upon the speed of the machine. Such an alternator forces any attached circuit to oscillate at the ma- chine's generating frequency, but the amount of the current set up in the circuit depends strictly upon the dynamo's voltage and the circuit's im- pedance to that frequency. Transmitters of this general type are coming into wider use day by day, as is seen from the work of the Gold- schmidt, Fessenden and Tele- funken companies. The cir- cuit effects described are substantially identical with alternating current circuits at commercial
��Popular Science Monthly
���Fig. 5
��those in operating power-distribution fre- quencies of 25 or 60 per second ; in the radio work, however, resonant or zero- reactance effects are made useful, and condensers are used directly in the cir- cuits. In low-frequency practice, reso- nance is usually carefully avoided and series condensers are almost never used. By far the- greatest number of radio telegraph transmitters in use today are of the spark condenser-discharge type. The circuit behavior in these senders is somewhat different from that in the sus- tained wave alternator transmitters, but most of the basic principles already ex- plained hold true. The main difference arises from the fact that with the alter- nator the frequency of the oscillations developed depends entirely upon the speed of dynamo and is independent of the circuits connected to it, while in the spark transmitter the frequency depends mainly upon the capacity and inductance of the discharging circuit.
Consider for a moment the arrange-
��ment of Fig. 4. Here an antenna A, which possesses inductance, capacity and resistance, has connected between it and the earth E a spark gap 5". Across the spark gap, by means of terminals TT, a high voltage transformer, induct- ance coil or other charging source is connected. If the potential of this charging source gradually increases, a current flows into the antenna and, be- cause of its electrostatic capacity, this aerial system takes a charge. If the voltage continues to rise until the elec- trical pressure is so great that the air between the spark gap terminals at 6^ breaks down, a spark will pass and the elec- tric charge previously im- pressed upon the aerial will rush to earth. In an ordinary antenna this discharge to earth will be such that the electrical inertia of the system will cause the charge to "over- shoot," in a sense, and the an- tenna will take on a polarity opposite to that which it had originally but somewhat weak- er. The insulating properties of the air gap 5" are not regained in the brief time of the charge's passage, and so the current rushes up to the antenna once more; at each swing or partial elec- trical oscillation the electromagnetic in- ertia due to inductance causes the effect of "overshooting, and the oscillations continue until the energy of the original charge is used up. The electrical phe- nomenon is in many ways similar to the mechanical effects which may be ob- served when a weight at the top of a springy rod (which has its lower end clamped in a vise) is swung back and forth.
Consider such a mechanical system, as shown in Fig. 5. If the weighted end A is pulled to the right by drawing on the light thread B, the spring C will be more and more strained until at last a point is reached at which the thread snaps. This is a fairly close analogy to the straining of the air in the spark gap S, Fig. 4, as the charging voltage grad- ually increases to the breaking point. Referring again to Fig. 5, as soon as the "charge" of mechanical energy placed in
�� �
