Radio Broadcast (May 1929-Apr 1930)

Interesting Data on Their Design and Use CHARACTERISTICS OF R. F. CHOKE COILS By ROBERT S. KRUSE Consulting Engineer R. S. Kruse In a recent article in Radio Broadcast, the author made the general statement that it is sometimes quite troublesome to have an r.f. choke present a capacity reactance instead of the inductive reactance that one usually attributes to it. A request has been received for a more specific statement on this point. Unfortunately it is difficult to make any statement that is both specific and general. More or less qualification is necessary and the reader is asked to keep in mind the necessity of some analysis of his particular circuit before deciding if the remarks apply there. Rather obviously the choke will act as an inductance at low frequencies, since its distributed capacity is of no importance then. As the frequency is raised one eventually arrives at the point where the choke is in resonance ; that is, the inductance and the self-capacity of the choke tune it to the frequency being fed to it. The result is that the choke presents a high impedance to the r.f. (which is desirable), but at the same time tends to develop a large "circulating current" inside the choke itself. This circulating current can be thought of as going in one direction through the inductance of the choke and returning through its self-capacity. Since these constants are distributed it follows that the circuit is not a simple one but instead is capable of showing multiple resonances. On one side of each of these resonances the choke presents (to the external circuit) an inductive reactance; on the other side it presents a capacitive reactance — both being high in normal designs of chokes. Further general statements become impossible; we must consider the various types. An Incorrect Assumption From what has been said one may assume that it is necessary always to operate in the inductive-reactance regions, avoiding the resonant points and those of capacitive reactance. This assumption is not correct. In many cases it is of little consequence whether the reactance is positive or negative, as long as it is high 0.01 iw=.t„. To Amplifier Fig. 2 — In this circuit the r.f. tube was fed through the tuned circuit of the following tube. Somewhat higher output was obtained because of the relatively higher load impedance. enough — and for some applications it does not even need to be high ! Furthermore, the resonance points need not be avoided in the more common cases for reasons that are given in the following paragraphs. Let us first take the case of the simple cylindrical choke, having one layer of wire, the length of the coil being perhaps 4 times the diameter, and the wire size small. Measured alone, this coil will show prominent resonance at a frequency such that a half wave is standing on it (voltage at both ends and none at the center), likewise at 3 times and 5 times this frequency, above which the resonances become less distinct. It will be noted also that the alleged third harmonic is not exactly at 3 times the lowest resonance frequency, likewise the fifth is not at exactly 5 times the fundamental. This is the usual action of a resonant circuit with distributed constants; UX-222 Detector Tuned Circuit of Transmitter Fig. 1 — The effectiveness of the choke between the detector and amplifier was noted by the signal obtained from the two-stage amplifier following the detector. antennas act similarly. The reason is that at each system of resonance we have a different current distribution which is equivalent to rearranging the current carried by each of the various small portions of the coil's self-capacity. The effect of these self-capacities accordingly varies from that at the fundamental. So far we have dealt with laboratory conditions— a coil hung out in space and measured by a weak field cautiously coupled into it. This is not the way the equipment will be used. Instead we will connect apparatus to the two ends of the coil and feed r.f. to the coil through this apparatus. This is quite different from the process of magnetically manufacturing the voltage in the coil itself, and again we find the elusive self-capacities shifting about so as to alter the seeming resonance points. It follows that the goodness or badness of a choke depends not only on the choke but also on the suitability of the job for the candidate. This leaves us badly at sea. We have not determined anything and have discovered several new variables. Until one has learned something by trial and error it is usually impossible to generate useful theory. Let us see, therefore, what practice teaches us, then perhaps we can apply the theoretical speculation to step on toward a better type — to be given practical test. Let us leave the single-layer coil for a moment. We find that where the r.f. voltage across a coil is high, or where a very high impedence is necessary, it is essential that the coil have physical length. If we do not have length — distance between the two terminals — we obviously will have fairly strong electric fields directly between the ends of the coil without much regard to the manner in which the wire wanders about. If the frequency is high there will be entirely too much bypassing due to this end-to-end capacity of the coil, entirely without any reference to the turn-to-turn capacity or the layer-to-layer capacity. Very well — when is this serious and what's to do about it? The reply to this is that it is serious only for receiving equipment working with shunt-fed screen-grid tubes, short-wave sets working below 50 meters, and for transmitting sets using shunt feed of plate supply or grid bias. The cure depends on which of the cases we consider. Screen-Grid Circuits Shunt feeding screen-grid tubes is fairly common practice, frequently with small regard for the effect on the gain of the stage. Unless one knows what the load of the tube looks like it is hard to say if the feed-choke sjhould have a capacity or inductive reactance. One knows with certainty, however, that it should have a very high reactance. Even at broadcast frequencies this means that the inductance must run to many millihenries and the distributed capacity to a very few "micromikes." The average receiving r.f. choke entirely fails to supply such a combination, since it consists of a mere "scramble winding" divided into sections. This sort of choke works well enough when it is required only to keep r.f. out of an audiotransformer primary or to perform some other chore in connection with a 201a or a 227 tube. Inspection of the tube constants suggests that equally good work at the same wavelengths with a 222 or 224 tube will require ^ the self-capacity and 10 times the inductance — or something of that general sort. Such an improvement is hardly to be hoped for, hence we must either put up with the nuisance of feeding through a tuned (tunable) circuit or else do the best we can. The obvious way to improve the choke seems to be that of lengthening it and making it slimmer so as to reduce self-capacity. Unfortunately, one is limited as to space and this process does not help much before the inductance Clip . Fig. 3 — A common type of multi-section choke. The slots were wound full of No. 38 wire, and when placed near a heterodyne wave meter, resonances were found at or near the following frequencies: 982, 1022, 1620, 1930 kc. Other small resonance peaks were also found. • AUGUST 1929 • • 237