When talk fall on the Loftin-White circuit most thinks of a special DC connection of AF amplifiers. But among radio technicans the Loftin-White circuit is quite another thing. It is a circuit ensuring constant energy transfer between RF amplifier stages over the whole tuning range. It was first described December 24th, 1925. In 1927 you can find a description of the circuit in 'Radio News' published by Hugo Gernsback.

From Radio News March 1927.

The New Loftin-White Constant-Coupling Circuit
By ROBERT H. MARRIOTT

WHAT bids fair to become the outstanding circuit of 1927 is the Loftin-White Circuit, the theory of which is described in this article. The new circuit is remarkable in that it gives constant amplification over the entire broadcast range, without the usual artifices and without the usual "crutches" which have been employed for several years in trying to accomplish this end.

 -EDITOR-

During the past year I became very much interested in a new circuit, because it seemed to me to be the first real improvement in radio-frequency-amplifier circuits since broadcasting began to be the principal use for radio. The theory of the circuit interested me at first; since then my interest has been increased by the practical results that have been obtained from the circuit in actual use. Broadcasting began its very rapid development in 1921. However, the better radio-frequency amplifier circuits which we have used since 1921, date back before that time. For example, in several law-suits, Hartley, Rice and Hazeltine have all carried their dates back before 1919. Hartley, for example, goes back to 1915, and Hazeltine to 1918 with Rice between those two years. Radio-frequency amplifiers were in demand before the radio broadcasting era. They were fairly practical at long wavelengths, but for comparatively shorter waves, such as we use now in broadcasting, their performance was handicapped by an excessive feed-back, which took place through the inherent capacity between the grid end and the plate end, respectively, of the grid and plate circuits. This feed-back produced undesired oscillation in the grid circuit, and subsequent signal distortion. Hartley, then Rice and then Hazeltine (apparently not knowing of Rice), set out to kill the effect of the feed-back by introducing into the grid circuit an opposing feed-back from the plate circuit. In a sense, they did what you may have done if you have had to fight a prairie or forest fire; that is, they produced a back fire. They used varied connection arrangements to produce the back fire; but all three employed the one principle of having two opposing feed-backs from the plate circuit, The use of this principle makes an improvement; but the improvement is not uniformly effective for all broadcast wavelengths. Also, this likeness in principle has produced another controversial patent situation, involving Hartley, Rice, Hazeltine and others.

A DIFFERENT SOLUTION

This new circuit does not operate on the principle of having two opposing feed-back currents from the plate circuit. It operates on the principle of making the feed-back harmless before it starts back to the grid. In other words, the new circuit takes the teeth out of the feed-back. Also, it is uniformly efficient for all wavelengths. And because its principle is different, it is not involved in any controversial patent situation. The new circuit is not the product of the wealth of a large corporation, rich in money, laboratory facilities and picked men from our higher educational institutions. On the contrary, it is the product of two independent inventors. The circuit was produced partly by Edward H. Loftin, who has been in close contact with the practical, theoretical and patent aspects of such circuits for a number of years; first as officer in charge of the radio patent and research section of the United States Navy, and later as a consulting engineer, in private practice. His co-worker, Mr. S. Young White, has been interested in the practical application of radio theories for about fifteen years and, during recent years, in the tedious work of producing practical improvements in broadcast receivers. The two-fold functioning of this circuit involves two theories, one being the explanation of why the feed-back does not produce regeneration, and the other the explanation of why the coupling scheme utilized transfers all broadcast frequencies onward equally well; the

 second is responsible for the description "constant-coupled, non-reactive plate circuit" given the system by its inventors. Those explanations can perhaps be made plainer by referring to circuit diagrams and graphs. Fig. 1 shows one stage of a radio-frequency-amplifier system, the first tube being an amplifier and the second tube a detector. Fig. 2 shows the corresponding Loftin-White circuit, applied between the amplifier tube and the detector tube. Fig. 3 contains graphs of the grid voltage of the first tube and the feed-back voltages in circuits like Figs. 1 and 2. A glance tells us that the two circuits are quite different, Fig. 1 has one plate circuit while Fig. 2 has two. Fig. 1 has one coupling, while Fig. 2 has two couplings between the plate circuit and the grid circuit of the detector tube. Fig. 1 has no condensers in the plate circuit, while Fig. 2 has two condensers in one of its plate circuits. Also, Fig. 3 indicates that the feed-back voltages and resulting voltages in Figs. 1 and 2 are quite different.

TWO PLATE CIRCUITS.

In Fig. 1 both the radio circuit and the direct-current circuit of the "B" battery are through F to P, L1, the "B" battery and back to F; while in Fig. 2, the "B" battery circuit is from F to P, through the choke coil, Ch, through the "B" battery and back to F, because C3 and C1 will not pass the direct current from the "B" battery. The radio-frequency circuit in Fig. 2 is through F to P, C3, L1, C1, and back to F, because the choke coil, Ch, will not pass much radio frequency current. The plate circuit in Fig. 1 is coupled to the grid circuit of the detector tube only by the inductive relation of the primary, L1, to the secondary, L2; or, we can say, by the mutual inductance of L1 and L2. The plate circuit in Fig. 2 is coupled to the grid circuit of the detector tube, not only by the mutual inductance of Li and L2, but in a second and additional way by the mutual capacity, C1. Mutual inductance is less effective for transferring energy at the higher wavelengths than at the shorter ones. Mutual capacity behaves in an opposite manner; that is, it is more effective for the long-wave broadcasts than it is for the shorter. Also, as the tuning condenser, C2, is increased in capacity for tuning to the long waves, it automatically changes the relative coupling value of C1 making the latter, still more effective for the transfer of long-wave signals. In Fig. 2, the coupling abilities of the mutual inductance and the mutual capacity are adjusted so that they combine to produce the same signal transfer for all wavelengths. This is the needed improvement over Fig. 1; because the latter depends on inductive coupling only, and, therefore, does not transfer all frequencies equally well. PHASE-SHIFTING CONDENSER In Fig. 2, the condenser, C3, is provided to shift the phase of the radio-frequency alternating current in the plate circuit; so that any feed-back that may occur from the plate, P, to the grid circuit of that tube, will be out of phase with the same frequency in that grid circuit. When sufficiently out of phase, it will not add itself to the grid frequency to build up an amplitude out of proportion to the other broadcast frequencies in the grid circuit. Such adding-up does occur in receivers employing circuits like Fig. 1, producing serious signal distortion. Too much feed-back causes the grid circuit components to oscillate of their own accord; and oscillation in a tuned R.F. receiver is ruinous as it produces

distortion and uncontrollable squealing when stations are tuned in. Condenser C3 prevents both the distortion and squealing. In Fig. 3 the graph, A, is intended to represent the alternating voltage in the grid circuit at the frequency with which the grid circuit of the first tube in Fig. 1 is in tune. Now the conclusion to which Loftin and White came, in their study of regenerative and oscillating tube circuits, was that, in tuning the grid circuit of the second tube, the normal tuning adjustment over-shoots the mark; to the extent of increasing the inductive reactance of the first tube's plate circuit to such a point that the voltage feedback from the plate circuit is shifted forward to the position of the solid graph, B. That solid graph, B, is in phase with the graph A and adds to A to make the solid graph C. For example, +5 adds to +10 and -5 adds to -10, in combining these graphs. This adding up of the energy represented by A and the solid graph B produces distortion. That is, A has been made to grow from its normal size to the size represented by the solid graph, C ; while the other different frequencies which go to make up broadcast signals are not built up in proportion (if we follow out the Loftin-White theory) because they do not bear the same relation to the frequencies to which the grid circuits of the first and second tubes are tuned. If the feed-back represented by the solid graph B is large enough, then the solid graph C will become larger, and excessive feed-back or oscillation, will result. Loftin and White, working on this theory that oscillations are produced by the inductive reactance of the plate circuit in shifting the voltage forward in phase, placed capacity reactance in the plate circuit to shift the voltage backwards. The dotted graph, B, represents the feedback shifted backwards in phase, so that when the dotted graph B is added to graph A, the resulting dotted graph C is no larger in amplitude than the original graph A. For example, the dotted graph B is at about 0 when A is at 10, therefore, nothing is added to A. Were the relations, as shown in Fig. 3, maintained between the dotted graph B and the graph, A, the result must be a shifting process rather than a building-up process. That is an interesting theory for the experimenter to speculate on, and may be a field for other inventions. If we had an oscillograph that operated as plainly on a million cycles as the ordinary

 oscillographs operate on sixty cycles, we might be able to see what goes on under those circumstances. Perhaps a tube with an added capacity between the grid and plate, and other elements suitably chosen, could be used at ordinary oscillograph frequencies to indicate what goes on at broadcast frequencies. This, of course, is of more interest to mathematically-inclined experimenters and those who have access to oscillographs and other laboratory equipment.

Index

Next page