US3247449A - Transformer apparatus - Google Patents

Description

- April 19, 1966 L. A. MEDLAR 3,247,449

TRANSFORMER APPARATUS Original Filed March l5, 1957 9 Sheets-Sheet l lllll lllllllll April 19, 1966 L.. A *Ml-:DLAR 3,247,449

TRANSFORMER APPARATUS riginal Filed March 15, 1957 9 Sheets-Sheet 2 CIVWL MMF April 19, 1966 A. MEDLAR 3,247,449

TRANSFORMER APPARATS Driginal Filed March 15, 1957 9 Sheets-Sheet 3 *Saai INVENTOR 5w/.s 0. Afm-ahw?) ATTORNEY5 April 19, 1966 l.. A. MEDLAR 3,247,449

TRANSFORMER APPARATUS Griginal Filed March 15. 1957 9 Sheets-Sheet 4 mlm-Ms( ATTORNEY 5 April 19, 1966 4 l.. A. MEDLAR 3,247,449

TRANSFORMER APPARATUS Original Filed March 15, 1957 9 Sheets-Sheet 5 O O /A/W?" M4000 BY 2 E: Qlf

ATTORNEYJ April 19, 1966 9 Sheets-Sheet 6 c sa l N VENTOR fn/As ,aw/fom?,

ATTORNEYS April 19, 1966 L. A. MEDLAR TRANSFORMERVAPPARATUS Original Filed March 15, `195'? 9 Sheets-Sheet 7 oar/7l cappe/vr- I0 Z' CURRENT F 777 777 0 OUTPUT L CONTROL P PAW/971949) Magu ATTORNEYS April 19, 1966 L.. A. MEDLAR 3,247,449

TRANSFORMER APPARATUS Original Filed March 15, 1957 9 Sheets-Sheet 8 ma BYQ, Z( Qd lm ATTORXYS April 19, 1966 L. A. MEDLAR 3,247,449

TRANSFORMER APPARATUS riginal Filed March l5, 1957 S3v Sheets-Sheet 9 6 /7 TUIPBLE TPH/VSFOPMEP SOURCE OF l/P/BLE D. C. VOLT/QSE INV ENTOR EW/s ,rz Mam/JP,

ATTORNEY United States Patent O 3,247,449 TRANSFORMER APPARATUS Lewis A. Mcdlar, Lansdale, Pa., assignor to Fox Products Company, Philadelphia, Pa., a corperation of Pennsylvania Original application Mar. 15, 1957, Ser. No. 646,429, now Patent No. 2,999,973, dated Sept. 12, 1961. Divided and this application Aug. 21, 1961, Ser. No. 132,844

13 Claims. (Cl. 323-60) This invention relates to a transformer system for supplying power from an A.C. input to a load, and, more particularly, to a transformer system for supplying a substantially constant current to a load subject to variations of impedance. This application is a division of my copending application Serial Number 646,429, l'iled March 15, 1957, and now U.S. Patent 2,999,973.

As will be understood after the following explanation, the transformer system of this invention is capable of supplying a large number of diiferent characteristic output. Among such outputs are a characteristic rise of current output with increasing load, a characteristic drop of current output with increasing load, and a substantially constant current output with increasing load. The invention will be more fully described in conjunction with the last-mentioned characteristic, which is preferred, but it will be understood that the invention is not limited to this characteristic.

In the past, several different means of obtaining a substantially constant current output with changing load impedance have been evolved. Perhaps the earliest of these various systems is that using a movable coil, the coil being counter-balanced and automatically adjusted in accordance with changing load to vary the distance between the input and output sides of the transformer and thereby to vary the coupling. This system is still in use today for constant current lighting systems, but it is subject to several disadvantages, among which are poor speed of response, the substantial expense of the system, the bulkiness of the unit, and the use of moving parts subject to wear and eventual shutdown of the system.

Another past-suggested constant current apparatus uses a saturable core reactor. The reactor may be a relatively simple one or may be a complex system utilizing shunts, partial air gaps, etc. This type of system, however, is relatively expensive, as well as being complex. Moreover, it has a lagging power factor.

A third general type of known constant current system is one using a resonance phenomenon. This type of system has taken many forms, but the simplest form is the series resonant circuit, including an inductive reactance and a capacitive reactance connected in series, the two reactances being adjusted to be substantially equal, and the load being connected across one of the reactances. While this type of system provides quite good constancy of output current, it has been found necessary to provide auxiliary means to limit the voltages across the components on open circuit. Moreover, this 4type of system is not readily adjustable to change the level of output current which is to be maintained constant. In order to provide for such change, there must be a change in ethciency, with no moving parts, with no need for auxiliary units to limit component voltages on open circuit,

with relatively easy adjustment of the parts to change the level of output current to be maintained constant, and for a substantially lower cost than many of the previous systems.

The transformer system of this invention is in the nature of a resonant phenomenon, but, as will be obvious from the theoretical analysis to follow, it is far from a simple resonant phenomenon. None of the various embodiments of the invention to be described uses a simple series resonant cir-cuit, and all of the embodiments actually automatically adjust themselves to maintain a substantially resonant condition even with change of only one of the components, such as change in capacity. The system is not dependent 'for its operation upon leakage reactance or on saturation of the core or any of its cornponents, and the system actually is limited and adversely affected in its operation by these ever-present effects.

One important advantage of the transformer system of this invention is the fact that it draws a leading power factor current from the input, rather than the lagging power factor usually obtained with a transformer system.

The apparatus of the present invention is capable of use wherever it is desired to draw a substantially constant current through a variable load, when a constant voltage input is provided. One instance of such use is in lighting systems.

The apparatus of the present invention, generally described, includes an input, an output, and a control electrical circuit, each of these circuits including at least one -coil which may be wound on a transformer core. The coils of the input and output electrical circuits are so inductively related, wound and connected with respect to one another that two magnetic circuits, an input and an output, are formed. The magnetic circuits have vtwo common portions, in one of which the magnetomotive forces generated by the input and output currents aid, and in the other of which the magnetomotive forces generated by these currents oppose. As a result, there is no direct coupling of power between the input circuit and the output circuit. The control electrical circuit includes va capacitor and is inductively coupled to one of the coils of the input and output circuits through one of the common portions of the magnetic circuits. Current through the control circuit generates a magnetomotive force which is opposite in phase and of greater magnitude than the` output magnetomotive force in the common portion through which the coupling takes place. The control magnetomotive force in effect reverses the output magnetomotive force in said common portion of the magnetic circuit and hence couples power from the input to the output through the control electrical circuit.

The invention will now be `described in conjunction with the accompanying drawings, showing preferred ernbodiments thereof.

In the drawings:

FIGS, 1-16 are directed to one general class of the various embodimentsl of the present invention, this class including at least three coils, one of the three coils being wound on each of the three legs of the core. In these figures, A

FIG. l is a schematic diagram of one connection of FIG. 8 is a further vector diagram showing the actual operation of the transformer for changing load;

FIG. 9 is a graph of the tiux density versus magnetic field intensity characteristic curve of the type of ferromagnetic material preferably though not necessarily used with the transformer, with a curve of static permeability of that material;

FIG. l() is a schematic showing ofthe input and output magnetic circuits of the transformer system of this class;

FIG. ll is a schematic showing of the various embodiments to this invention as servo systems;

FIG. 12 is a schematic diagram similar to FIG. 1, but showing the input and output connections reversed, so that the load is connected yto what was formerly the input circuit, and the input is connected to what was formerly the load circuit;

FIG. 13 is a schematic diagram showing a control coil and listing the various possible connections of the control circuit;

FIG. 14 is a schematic showing of one method for adjusting the effective capacity of the control circuit;

FIG. 15 is a schematic diagram showing transformer coupling of the capacitor into the control system;

FIG. 16 is a schematic diagram of an apparatus for adjusting the effective capacity of the control circuit through use of a saturable transformer; and

FIGS. 17-27 are employed `to show various embodiments of a second class of the invention, together with the theory of operation of such embodiments. This class of the various embodiments of the present invention includes at least four coils wound on the transformer core, two coils being wound on each of two of the legs of the core, and with one of the coils from each of the two legs being connected in series with each other, and the two series c-ombinations of the coils being connected into an input and an output circuit, respectively. With this connection, one leg of the transformer core is free of coils.

Of these figures, p

FIG. 17 is a schematic diagram of the basic connection of this class of the embodiments of the invention;

FIG. 18 is an equivalent electrical circuit of the appa ratus of FIG. 17;

FIG. l9 is a schematic'showing of an embodiment similar to that of FIG. 17, but including a separate control coil;

FIG. 20 is a vector diagram showing the actual operation of the transformer of FIG. 19 for increasing loads;

FIG. 21 is a graph of a series of characteristic output vcurrent versus output voltage curves of the apparatus of FIG. 19 for changing values of the capacitor;

FIG. 22 is a schematic showing of the magnetic input and output circuits of this class of embodiments of the invention;

FIG. 23 is a schematic showing of a modification similar to that of FIG. 17, but with the input connected across what were previously the load terminals, and the load connected across what'were previously the input terminals;

FIG. 24 is a schematic showing of the use of the control coil and listing its various possible connections;

FIG. 25 is a schematic showing of a system providing for variation of effective capacity in the control circuit;

FIG. 26 is a schematic showing transformer coupling of the capacitor into-the control circuit;

FIG. 27 is .a schematic showing of the usev of a saturable transformer in the control circuit to vary the effective capacity of the capaci-tor in the control circuit.

Referring first to FIGS. 1-16, it will be evident that all of the various embodiments shown in these figures have their magnetic paths linking the coils defined by a threelegged core. The core has one coil wound on each of the three legs. In the system of FIG. l, representing the basic embodiment of this class of the invention, a core 1 of. ferromagnetic material is preferably employed.

L I. Standard transformer iron embodied in stamped lamina tions may be used, but different core materials can be employed, depending upon the output characteristic and the general level of output current that is desired. Thel core 1 includes three legs, the two outer legs being labelled 2 and 3, and the center leg 4. Reference letters have also been used to further identify the configuration of the various legs. The left outer leg 2 is a channel-shaped element including the arms e-f, f-a, and a-b. The center leg is an I-shaped element, labelled b-e, and the right outer leg 3 is a complementary channel-shaped element having arms b-c, c-d and d-a The portions a-f and c-d of legs 2 and 3 arev parallel to one another and to portion b-e ofvleg 4. These configurations have been used for simplicity in identifying core legs, and the core need not necessarily be so made up.

The structure described provides closed inductive or magnetic paths between portion a-f of leg 2 and portion c-d of leg 3, as well as between portions a-f and b-e and c-a' of legs 2, tand 3, respectively If a ferromagnetic core is employed, as is preferred, the permeability of such pa-ths will be high in comparison with air.

Though the several legs of the core have been shown as co-planar and as having parallel portions a-f, b-e -and c-d, neither of these conditions is necessary to operation of the system. The legs of the core may be aligned so as to define dierent planes which may be parallel or intersecting, as desired, and the legs themselves need not be parallel, as long as there are three legs all coupledtogether by magnetic paths. Moreover, though no air gaps have been shown in the core of FIG. l, it is evident that a core of sheet stampings, assembled in conventional manner, to form a core of the configuration of FIG. 1, would have air gaps therein. Moreover, it is possible that it might be desirable to construct the core with an air or other. non-magnetic gap in one or more of its legs for some speci-al purpose, and such construction is within the scope of this invention. Further, it mightbe desirable to employ a non-ferromagneticY material for the core if the high reluctance of such a core is not important or is otherwise compensated for. Therefore, the invention is not to be considered limited to ferromagnetic cores.

The transformer system of FIG. l includes three electrical circuits, an input, :an output and a control circuit.y

The input circuit includes ya coilLl wound on leg 4 and the conductors 5 and 6 connected to the terminals of coil L1 and connecting the coil across an input source of A.C. voltage 7. In order to obtain constant current action, the input 7 must be of substantially constant voltage.

The output electrical circuit includes coil L2 wound on parallel portion a-f of leg 2, yand coil L3 wound on parallel portion c-d of leg Coils L2 and L3 are connected in series by conductor 8, and their distal ends are connected by conductors 9 and 10, respectively, across -a load l1. Load Il may be of any type, but most satisfactory constant current operation has been found to -Gccur when the load is primarily resistive, rather than reactive. When a substantially unity power factor load is to absorb power from the input, the best constant current action is obtained.

Coils L2 and L3 are so wound on the transformer corev with respect to the direction of primary fiuX and so connected in series `by conductor 8 'that their voltages induced lby the primary flux oppose one another in the ofutput circuit, with the result that the voltage across the load terminals is substantially zero `at no load.

The control electrical circuit 0f FIG. 1 includes a capacitor K connected directly across the terminals of coil L3. In this embodiment of the first class of the invention, one of the output coils also acts as the control coil, the voltage induced across coil L3 driving the control current through capacitor K.

Referring now to FIG. 2, a methematical analysis of the operation of the apparatus of FIG. 1 will be per-V,

formed, so that the operation of that apparatus can be better understood.

FIG. 2 is a schematic diagram of the equivalent electrical circuit of the apparatus of FIG. 1, with the input voltage EP applied across coil L1, L2 and L3 being shown c-Onnected in series but wound in opposite directions, and currents Ip, Io and IL owing in the input, output and control electrical circuits. 'Phe mutual inductances or inductive couplings between the various coils are shown as M2 between coils L1 and L2, M1 between coils L2 and L3, and M3 between coils L1 and L3. The capacity of the capacitor in the control circuit is identied as K, and the load impedance is Z.

Using p-iw, setting up mesh equations for the input, output and control circuits, assuming one:one turns ratios, neglecting ohmic resistance of coils, leakage react 'ance and core loss, and solving for the load current, we obtain the following equation:

the output current m.m.f. Fo tends to cause ilux to circulate through the two outer legs and Fo therefore opposes the .primary magnetomotive force in one leg and aids it in the other leg. There actually are two magnetic circuits, the rst magnetic circuit being traced by the letters a-b-c-de-f-a, this being the output magnetic circuit for the systems in which the load is connected across the series combination of the two outer coils, and the second magnetic circuit being g-b-c-aLe-g and g-b-a-f-e-g, this circuit being the path of ux driven by the primary magnetomotive force in the system in which the input is supplied to the coil wound on the center leg.

When the connections of the input and the load are reversed, as will be described, the'magnetic circuits reverse, lbut the function of thesystem is the same. In the leg on which the control coil is wound, the right leg in FIG. 10, the ouput magnetomotive force and the primary or input magnetomotive force are shown as aid- L1 2 k: 2 l pI +L1L3 M3, o

This obviously is the condition for pure constancy of the load, or output current. Through simple algebraic operadition but it also is certainly not a simple resonance tion: l

L1L3M32 l (3) "ii: L. :IMT

This last equation is obviously one for a resonant condition but it also is certainly not a simple resonance equation.`

We will now proceed to show how the general class of embodiments of this invention represented by the basic circuit of FIG. 1 achieves the contants of current which required that the last two equations be substantially true. Before doing this, however, we will discuss the modication of FIG. 3, which, while operating very similarly to FIG. l, includes a separate control coil L4 wound on leg 3 with coil L3 of the output circuit. In FIG. 3, the series combination of control coil L4 and the capacitor, now called C, is connected directly across output coil L3 on the same leg. It has been found that the only diiference in operation of the system of FIG. 3 over that of FIG. l is caused by the higher Voltage available to drive the current through the capacitor. This higher voltage makes the control current larger, thus allowing the output current, as will be shown hereinafter, to achieve a higher level. We use the identification C, rather than K, in FIG.` 3 to indicate that an actual capacitorI of value C is connected in this circuit. The effect, of using a separate control coil, such as with the apparatus of FIG. 3, may be included in the mathematical analysis for FIG. 2 by substituting for the term K in the above equations ZC, where =l+a, and a=the turns ratio between the control lcoil and coil L3. It will be evidentl from these statements that the only effect of the control coil is to add the turns thereof to the control circuit, and thereby to increase the voltage across the capacitor.

The systems of 'FIGS. 1 through 16 produce magnetic circuits as shown in'FIG. 10. In that figure, the primary or input current produces a primary magnetomotive force FP which tends to cause flux to flow through the center leg and split to flow through the two outer legs. The two output coils being opposed in the output circuit,

ing. This being so, there is no direct coupling of the load into the primary, so that the system would not transform power but for the control circuit. However, the control coil magnetomotive force, FL, is opposite to the output current magnetomotive force and substantially equals twice the output magnetomotive force in the right leg, so as to form a resultant of these two magnetomotive force reversed from the ouput m.m.f. and to reflect the load back into the primary circuit through the center leg. This results in coupling of power between the input and the load through control coil magnetomotive force produced by current in the control coil.

To analyze the physical operation of the apparatus of FIGS. 1 and 3, we now turn to FIG. 4, showing a vector diagram for an ideal transformer of the type shown in the foregoing iigures, with the output shorted. For the purpose of this analysis, an ideal transformer is taken as one which requires negligible magnetomotive force to drive the ux necessary to induce the voltages (that is, has substantially zero reluctance). The ideal transformer also has no saturation effects, and there is no leakage reactance between the windings.

With such a transformer, and the connection of FIGS. l or 3, first considering the control circuit open, the voltages across coils L2 and L3 in the output circuit appear as equal and opposite voltages, labelled in FIG. 4 as Basic E2 and Basic E3. These voltages cancel to produce zero ouput voltage. This is so because in the usual arrangement in the apparatus of FIGS. l and 3, the number of turns of the coil L2 will equal the number of turns of the other coil L3 of the output circuit. However, it

is not necessary that the number of turns of the two coils be equal, because there is an equalizing action between the two coils, so that the two voltages will cancel at short circuit, even though the number of turns of the two coils are unequal. Next consider the control circuit closed, so that the control circuit voltage drives current through the capacitor. The vector diagrams of this and the succeeding figures are shown for the ouput magnetic circuit. In that circuit, the current through the capacitor leads the voltage which drives that current by Since that voltage is substantially in phase (or out of phase) with Basic E3, with shorted output, the control current, and hence the control coil magnetomotive force (m.m.f.) is 90 ahead f Basic E3.

The control current m.m.f. produces an initial circulating ux which links coils L2 and L3 to produce in lphase induced voltages therein which lag by A90" the m.m.f. producing the flux. Consequently, the voltages in L2 and L3 no longerv cancel, but they combine to produce a net E0. This net E3 drives an output current in phase With it, andthe output current in coils L2 and L3 produces an output current m.m.f. in phase with it. The control n andy output current m.m.f.s then combine vectorially to produce a net or unbalance m.m.f. rotated clockwise from the control current m.m.f. This net m.ni.f. generates a circulating flux which links -coils L2 and L3 to produce new induced voltages in the coils which are rotated clockwise from the original voltages therein. This procedure continues until the steady state condition of FIG. 4 is reached. In that figure, it will be noted that the output and control current m.rr1.f.s are shown as substantially less than 180 from each other, for clarity, though they are actually substantially 180 apart. This relationship obtains because negligible net m.m.f. is required to produce the uxes necessary to induce the component voltages. The net m.m.f. is also exaggerated in the figure for ease of examination. Also, the output voltage is substantially zero (though it is shown in the figure), since Z equals zero, so that net E is substantially 90 from Basic E3, and net E3 and the net rr1.m.f. are susbtantially in phase with Basic E3. The output current is of substantial magnitude and in phase with E0.

To show how the system which produces the vector relationships of FIG. 4 operates to prevent changes in the output current, consider what happens if the output current should decrease. Decrease in the output current vector Io of FIG. 4 would immediately unbalance the net m.m.f., causing it to go toward the direction of the control coil m.m.f. The net unbalanced m.m.f. would then generate voltages in coils L2 and L3 rotated counterclockwise from their positions as shown in FIG. 4. The voltages E0 in L2 and E0 in L3 would then combine to yield a net E0 also rotated counterclockwise from the position shown in FIG. 4. This net E0 would drive an output -current Io in phase with it to produce an output current m.m.f. also in phase with E0. The new output current m.m.f. would be more nearly in phase with the control m.m.f., thus increasing the net m.m.f. and therefore the output voltages. With the increase in output voltage, the output current would return to its original value to cause the system to return to the steady state constant current condition of FIG. 4. Note in this figure that the net m.m.f. is susbtantially in phase with Basic E3 and the control and output currents are substantially out of phase.

The vector diagram of FIG. 4 is shown for the ideal transformer at the impedance Z-0. FIG. 5 shows the same transformer, but with the output impedance Z not equal to 0. When resistance is introduced into the load circuit, the output current first decreased, causing the output current m.rn.f. to decrease and the net m.m.f. to swing counterclockwise toward the control current m.m.f. The same action as described immediately above returns the system to the steady state condition of FIG. 4. In this case, however, since an appreciable output voltage is required to drive the output current through the appreciable impedance in the load circuit, the output components of the coil voltages must be larger than before. Nets E3 and E2 therefore rotate clockwise and counterclockwise, respectively, from their positions in FIG. 4, and, since the control m.m.f. must be 90 from net E3, the control m.m.f. rotates clockwise to the position of FIG. 5. In this ideal case, the net rn.m.f. necessary to produce the increased output voltage is still negligible, so that the output current m.m.f. is substantially 180 from the control current rn.m.f., and the net m.m.f. is therefore substantially in phase with net E3, as shown in FIG. 5. However, the output and control m.m.f.s are shown as sub- `stantially less than 180 apart, as if appreciable net m.m.f.

is required, for clarity.

The vectors are also shown in FIG. 5 to indicate the action of the circuit of FIGS. 1 and 3 to maintain a constant current output. Referring again to FIG. 5, if the load impedance decreases, the output current m.m.f. will tend to increase, as shown by the longer vector NEW OUTPUT CURRENT mmf. This action will unbalance the m.m.f.s to rotate the net m.m.f. toward the output current m.m.f., yielding the NEW NET m.m.f.

of FIG. 5. This new net m.m.f. will generate output voltages in'coils L2 and L3 rotated clockwise from the original output voltages in those coils, as shown in FIG. 5. These output voltages combine to produce a new net E0 also rotated clockwise, and, assuming the impedance is resistive, an output current and output current m.m.f. which are likewise rotated clockwise from the original. This output current rn.m.f., labelled correcting in FIG. 5, being more out of phase with the control m.m.f., will reduce the net m.m.f., reducing the output coil voltages and therefore the output current back toward its original magnitude. The system then returns to the original condition, shown in full lines in FIG. 5.

The vector diagram of FIG. 6 is designed to give a complete picture of the action of the ideal transformer, showing all important component voltages and currents thereof, for increasing impedance. It will be noted that as the impedance increases continuously, the vector net E3 will rotate clockwise, decreasing continuously. Since the control current IL is driven by net E3 (and the voltage across the control coil which is in phase with net E3), the control coil current must rotate and decrease continuously with net E3. Likewise, the control m.m.f. will decrease continuously with net E3 and, E0, being substantially 90 behind net E3, will rotate clockwise. The output current is directly dependent upon the control current, so that, as the control current decreases, the output current likewise decreases continuously, describing a semi-circle of diameter equal to the original output current at Zero impedance.

It will also be noted that the primary current Ip is substantially in phase with the control coil current IL, and so leads the primary' voltage Ep. This results in a leading power factor, ordinarily an advantageous condition.

It will be observed from the vector diagram of FIG. 6 that the ideal transformer tends toward a continuously decreasing output current with increasing load impedance, rather than a constant output current. However, an ideal transformer has been stipulated, in which negligible m.m.f. is required to drive the fluxes necessary to produce the voltages in the system, that is, the core material has zero reluctance. The inductances of the coils would be infinite but the mutual inductance between the coils, due to their physical separation, would be finite. Therefore the second term of Equation 2 would be zero so that it could never balance the rst term. Hence the impedance would play a very large part in the operation of the systemh and the output current would vary inversely therewit In actual practice, however, the inductances of the coils are never infinite, and it has been found that the coupling, as represented by M in Equation 2, automatically adjusts itself in correspondence with the value of capacity to make the equation substantially true.

To bring out this compensating feature of practical transformers in which the coupling is not perfect and in which the core has appreciable reluctance, consider next a different ideal'core material having appreciably less than perfect coupling and constant reluctance, but not subject to saturation effects. With such a core material, the net m.m.f. must be'appreciable to drive the circulating flux whichilinks coils L2 and L3. 'There must be relative rotation between the control and output m.m.f.s toward Basic E3 to produce this appreciable net 1n.m.f. from the condition of FIG. 4. In other words, the control and output m.'m.f.s are no longer substantially at 180 to each other, and the net m.m.f. must therefore rotate back toward Basic E3 from the phase relationship of FIG. 4, vwhere is was substantially in phase with net E3.

For this new ideal core material, the relative rotation between the control and output rn.m.f.s required to pro- I duce the linearly increasing net m.m.f. with increasing output yvoltage will tend to maintain the output vcurrent constant. If the reluctance is of proper value, the net m.m.f. will stay midway between Basic E3 and net E3 despite increasing output voltage andclockwise rotation of net E3. The output current then would be perfectly constant over a range of output voltage limited only by leakage between the windings. If the reluctance were too high, of course', the required net m.m.f. would be so high that it would be closer to Basic E3 than to net E3 and the output current would rise with increasing voltage. If the reluctance were too low, the system would approach the first ideal core having negligible reluctance, and the current would drop off. However, it is the apparent reluctance, as represented by the coupling, rather than the physical reluctance, that is important. The value of this apparent reluctance relative to the other parameters of the system, then, determines the characteristic of output current versus output voltage.

In contrast to the labove ideal core materials is a practical ferromagnetic material having changing reluctance and permeability and saturation characteristics. The performance of the system with such a core material will now be examined, referring first to FIG. 7, showing curves obtained With an actual core material having the windings and circuitry of FIG. 3.

FIG. 7 shows four different curves of the output current versus the output voltage obtained for four different input voltages. It will be noted that the output current is relatively constant over a very substantial range extending upwardly from zero output voltage in all cases, but that Variations in the input voltage cause a substantial change in the output current. It will further be noted that the output current decreases from its value at zero output voltage when the impedance increases, reaches a minimum, and then increases again up to a peak. From this peak, the output current decreases continuously and the constant current action is no longer had. It has been found that this continuous decrease in output current occurs because of saturation of the core material. It will be appreciated, then, rather than being dependent upon saturation of core material, the various embodiments of this invention are limited, as far as their constancy of current is concerned, by saturation of the core material. is further evident that the apparent reluctance of the core material is too low at first, resulting in the current decrease, but it then increases to cause the current to rise back to its original value.

In order to explain the .action of the output current in rst decreasing, then increasing to a peak, then decreasing continuously as saturation enters into the picture, we refer next to FIG. 8 showing a vector diagram for an actual transformer core material, with the circuits of FIG. 3 in use.

It will be noted in FIG. 8 that the vector En in L3 rotates clockwise from its original position perpendicular to Basic E3 at zero impedance and that net E3 decreases continuously at first. The control m.m.f., being dependent on net E3, follows this action, likewise decreasing and rotating clockwise. The output current Io, being dependent upon the control m.m.f., likewise rotates clockwise at first and decreases.l At low levels of net m.m.f., the control and output currents are substantially 180 out of phase with each other and the net m.m.f. is in phase with E3, so that the system behaves like the ideal core discussed in conjunction with FIG. 6. However, at higher levels of net m.ni.f., required to produce the higher output voltages for increasing E (in contrast to the ideal core system), In and IL must move toward each other to produce the larger net m.m.f. Io moves counterclockwise to carry E0, and therefore E3, with it. Since IL must al. ways be perpendicular to net E3, being produced by it, the net m.m.f. must move back toward Basic E3 as the output voltage increases and moves counterclockwise with I0. This ymovement causes net E3 to move back toward Basic E3 and increase, so that the control m.m.f. increases,

and the net m.m.f. must increase. The output current,

It) be noted, however, that net E2 increases continuously in the first portion of its locus, though net E3 does not. As lthe output voltage increases continuously, the flux density in the leg carrying L2 rises, being proportional to net E2, so that this leg lbecomes substantially saturated. As the needed net m.m.f. further increases, E0 in L2 and E0 in L3 reverse their clockwise rotation to cause the output current to swing more and more toward the voltage Basic E3, thus increasing net E. When net E3 reaches a value `such that the leg carrying -coil L3 becomes saturated, the output current decreases continuously. Hence, saturation is first caused in the core leg carrying coil L2, and this effect causes unbalance to set up saturation in the leg carrying coil L3, the -control leg.

FIG. 9 is provided to show the action of a transformer in--approaching saturation. The figure is a graph of the intrinsic flux density versus the magnetic field intensity for a typical ferromagnetic transformer core material, t0- gether with a graph of the static permeability for the same material; It will be noted that the permeability first increases rapidly, reflecting the very fast'increase in flux density with increasing field intensity, then the permeability drops with the levelling off of the flux density curve. In the area following the beginning of the decrease in permeability, there is a gradual reduction in the static permeability of the core material. It has been found that the area between the point marked A on the graph and the point marked B on the graph of flux density, for the basic flux density of the core (that is, the density of flux caused by primary current), furnishes the best action for constancy of current, since the permeability of the core material changes the least over this area.

The action of the system of FIGS. l and 3 in producing constancy of current may be better understood by discussing it as a magnetic servo system, which it has been found to be. Referring to FIG. 11, the input to the conventional servo system is a reference or standard. The control m.m.f. of the system of this invention fills the function of the reference or standard. In the servo system, the reference or standard is referred to a differential in which it is compared with the output of a feedback loop, reflecting the action that the output -of the system has taken. The differenial of this invention is the phase relationship between the output and control m.m.f.s. The net or error -between the reference or standard and the feedback loop information, represented by the net m.m.f. in the system of this invention, is supplied to an amplifier. The Ifunction of the amplifier in the system of this invention is performed by the permeability ofthe core material. The output of the amplifier, the amplified error of the servo system, is directed to a power device for influencing the output of the system. The amplified error is equivalent to the circulating fiux of the system of this invention, and the power device consists of the output coils which provide output E0. The regulated output of the power device is the output current I0 in the system of this invention. The output Io is converted by the feedback amplitier (also the output coils) from the input to the amplifier, I0, into the output m.m.f. of the system of this invention. This system results in the control m.m.f. and the output m.m.f. being continuously compared to produce a net This net rn.m.f., through the permeability of the core, is amplified and produces a circulating flux which induces output voltages in the two coils. These output voltages add together algebraically to produce an output current. The output current, then, is carefully regulated by the control ni.m.f. to be maintained in substantial agreement therewith. If the control m.m.f. is changed, the output current Will be correspondingly changed.

It will be appreciated thatthe apparatus of this invention is actually a magnetic servo system. The various embodiments to be explained later herein also behave in the same manner, so that the servo explanation applies to them, also.

The foregoing explanation has been directed to the l ll formof the invention shown in FIG. 3, but it has been emphasized that it applies also to the embodiment of FIG. 1. Actually, the only difference between FIG. 1 and FIG. 3, besides the increase of driving voltage available to the control coil circuit, and thus the increase in control m.m.f., is the possibility of different connections for the control coil circuit. Various connections possible ttor this circuit will be shown in FIG. 13, but first refer to FIG. 12 in which it is illustrated that the positions of the load and'input in the system may Ibe reversed from those shown in FIG. 1. In FIG. 12, the input is shown as connected across leads 9 and 10, and the load is connected across leads and 6, thus putting the input across the series combination of coils L2 and L3, and the load across coil L1. The action of the system of FIG. 12 is very similar to that of FIG. 1, yielding very good constancy of current.

A Now referring to FIG. 13, control coil L4 is shown connected in series with capacitor C, but the leads of this series circuit are not connected across L3, as inFIG. 3. It is indicated that this control circuit may include any source of voltage, including L4 alone. In other words, the `series combination of L., and C may be connected across the coil L1, coil L3, coil L3, coils L3 and L3, any other source of voltage external to the system, or the two leads may `be shorted together. The important thing is that the control impedance be a capacitive reactance, no matter what its source of voltage. However, it has been Ifound that when the source of voltage for the control irnpedance is so related to the primary voltage that the control current and the primary voltage are substantially 90 out of phase at zero load, the greatest range of constan-cy for varying output voltage is obtained.

With these various connections of the -control coil circuit -shown in FIG. 13, many diiere'nt types of output responso may -be obtained.

In commercial use of` the embodiments of this class of the invention, it is advantageous to provide some means for controlling the level-of output current which is to be maintained constant. Since the output current is directly dependent upon the control coil current, the control coil current may be changed to change the level of output current. This may be done by changing Ithe value of the capacity, but it is more convenient, especially at such high voltages as may obtain with a system of this type, to change the effective value of capacity, rather than the actual value. One way of changing the effective value of the capacity is shown in FIG. 14, in which a variable tnductance 15 is shown shunted directly across capacitor K. Variation of the magnitude of the inductance of 15 will change the effective capacity across coil L3, and thus the control coil current.

It is also possible to couple the control capacity into l the system by use of a transformer, rather than connecting a capacitor directly in series with the control coil. This is exemplified in FIG. 15 by use of a 4transformer T, having capacitor C connected across its secondary, and its primary connected across coil L3. Moreover, the capacitor might be coupled into the control circuit through a variable transformer, or Variac, thus enabling the effective capacitive reactance to be changed by change in the transformer turns ratio.

Another Way of coupling the capacitor into the system without direct connection of the capacitor to the control coil, and with the possibility of variationof the effective capacity in the control coil circuit, is shown in FIG. 16. In that figure, capacitor C is connected across the secondary of a saturable transformer 21. Two coils 22 and 23 wound on the outer legs of the three leg transformer 21'are provided with direct current through a source of variable D.C. voltage 24. The coils are connected in series, and phased in such manner as to cause a D.C. m.m.f. to direct flux through the outer legs of transformer 21 only. An increase in direct current through these coils increases the magnetomotive force and hence the magnetic field intensity in the outer leg portions of the transformer so as to increase the saturation of these portions of the core material. Secondary coil 20 is coupled through the center leg to primary coil 2'5 of the saturable transformer. The primary coil 25 ris connected directly across coil L3 of the constant current transformer system. Since the equivalent circuit of a transformer includes an iron core inductance across an ideal transformer, thus shunting the capacity, an increase in direct current through coils 22 and 23 by increasing the saturation of part of the magnetic path of the transformer reduces this inductance, so as to decrease the effective capacity of the control circuit.

It was stated in the beginning of this specification that the action of the system in maintaining the output current constant is independent of the value of the capacity in the control coil circuit. This is a very surprising thing, for it is evident from `a consideration of equation (2) that the only quantity thatA could possibly vary automatically with changing C to maintain the equality is the mutual inductance, and, more specifically, the coeicient of coupling' forming one factor of the mutual inductance. The coefiicient of coupling of the two output coils has actually been found to vary automatically to compensate for changing C to maintain the system substantially in the resonant condition defined by this equation. Naturally, however, there are theoretical limits to variation in this coefiicient, depending upon the physical spacing of the coils, for the coefiicient can never reach unity, and the higher the percentage of leakage flux the lower the coetiicient. For the connection of the apparatus `in which the single coil on the center leg is the primary of the transformer, the M that is involved is that between the primary coil and one of the secondary coils, whil-e, for the connection of the apparatus in which the two coils on the outer legs constitute the primary, the M that is involved'is that between the two primary windings. Tests have shown that the range of variation of capacity for which current is maintained constant is substantially unlimited for the second connection but limited for the first connection. Apparently the criterion is whether the M involved is between coils between which power is transferred,'e.g., between the primary and control coils. In the first connection, power is transferred between the single primary coil and the control coil, which is also one of the secondary coils, and, since the M involved is that between these two coils, the range of variation is limited by leakage between these coils.

In the second connection, power is not transferred between the two primary coils, which are the coils between which the M involved exists, so that leakage between these two coils does not limit the range of variation.

. Hence, for practical purposes, the range of variation of the second connection (using two primary coils) is unlimited, so that the level of output current can be changed to any value desired by change in the effective capacitive reactance in the control circuit, and the current will be maintained constant at that value for changing load, Without change in any other component. In contrast, for the first connection (using only a single coil as primary), the range of variation is limited by leakage.

It will be further evident from the above that, though the apparatus described functions to maintain constancy of current through some sort of resonance, the resonant circuit is not of any simple type. Further, exact equality of capacitive and inductive terms of Equation 2 is not necessary for independence of load, so that the system need operate only near pure resonance. Moreover, the characteristics of an iron core usually employed for regulating purposes, such as non-linearity and saturability are not necessary to operation of the apparatus of this invention and are actually limiting to its operation. Actually an air core, or some non-magnetic core, could be used, thus avoiding problems with saturation, though the 13 high reluctance of the magnetic paths linking the several coils and the high leakage would be undesirable for many purposes.

Another surprising feature of all the embodiments of this invention is their independence of variation 4in frequency of the input, as far as keeping the circuit in resonance is concerned. Of course, ordinary resonance systems are'extremely dependent on frequency for main- 'tenance of the resonant condition, but the resonant circuits described herein are completely independent 4of frequency, except that the output current level at which constancy is maintained is changed with change in frequency. In other words, as input frequency is changed, the system is maintained in resonance, though the level of output current is changed.

It will be understood that the use of the term magnetic paths in the accompanying claims does not restrict such claims to ferromagnetic cores forming such paths, since air or nonmagnetic material might form the paths. Magnetic paths will be understood to deiine the courses of ilux lines linking the coils referred to.

To explain the second class of embodiments of .the present invention, reference is now made to FIGS. 17 through 28. All of these figures have application to a transformer system including at least four coils, two coils in each of the output and input circuits connected in series, with one of the output and one of the input coils being wound on each of two of the legs of the transformer core. The third leg of the core is unoccupied by a coil. The core ispreferably or advantageously of magnetic material, so as to exhibit relatively low reluctance as compared with air, and stamped laminations of standard transformer iron may make up the core. Actually, the core maybe of non-ferromagnetic material, or air, as long as the high reluctance and leakage of such a core is compensated for by the other parameters of the system or is not too important. v

Such a system as described immediately above is shown in FIG. 17 in which the input source of A C. voltage is connected across series-connected coils L1 and L2. These coils are wound on outer legs 31 and 32 of a shell-type transformer core 33. The left outer leg includes portions b-a-f-'e of the transformer, the center leg portion b-e, and the right outer leg includes portion b-c-d-e of the core. The portions a-f, b-e, c-e of the three legs are substantially parallel to one another.

As indicated in connection with the first class of embodiments, it is not necessary that the legs of the core be co-planar, or parallel, as shown in the drawings, or that the core be free of air gaps. It is only necessary that closed magnetic paths be formed between all coils of the core. v v

Wound on the same leg 31 with coil L1 is a .second coi l L1', and a similar arrangement obtains for leg 32, on which coil L2', is wound. Coils L1' and L2' are connected in series by lead 34, and their distal ends are connected by leads 35 and 36 to a load 37. As with the various embodiments of the iirst class of embodiments of this invention, load 3.7 may be of any type, but if constancy of current is desired, it is preferred that the load be mostly resistive in character, and, for best constancy, the load should be substantially of unity power factor.

The output coil L2, has a capacitor K connected across its terminals.

The system of FIG. 17 thus is formed into three electrical circuits, an input, an output and a control circuit. The input circuit includes coils L1 and L2 and is connected across input 30, while the output circuit includes coils L1 and L2' and is connected acrosss load 37, and the control circuit includes capacitor K and coil L2. The input and output coils must be so phased that the ux generated by current through one of the input and output circuits Hows through the center leg, while flux generated by current through the other vcircuit bypasses the 14 center leg. Obviously, the primary flux, then, can bypass the center leg, or flow through it.

An equivalent electrical circuit for the system of FIG. 17 is shown in FIG. 118, in which the input voltage El, is applied across the series combination of the coils L1 and L2, causing primary current Ip to ow in the input circuit. Through the transformer action, control coil current IL flows in the control circuit, and output current Io -flows in the load circuit. Mutual inductances, or inductive couplings, between the various coils are shownV by M1 between coils L1 and L2, M2 between coils Lland L1', M2 between coils L2 and L2', M5 between coils L1' and L1', M6 between coils L1 and L2', and M9 between coils' L1 and L2'. The load impedance is represented by Z, and the capacity by K.

Using p=]`, setting up mesh equationsfor the input, output and control circuits, and solving for .the load current, we obtain the following equation:

2 Li-L2 :I I El@ MLILFM, L1+L2+2M 41J+Zlp K+ LILFM,

This equation represents the use of several simplifications, all of which are perfectly valid. We have assumed a onezone coil ratio, and have neglected ohmic resistance of coils, leakage reactance, and core loss. We have also assumed that L1=L2' and L2=L2', since the coils are wound on the same legs and have the same number of turns. This fact is not necessarily true but the equations are simpler by omission of constants if it is true. We have further assumed M2=L1 and M3=L2, because,'the coils 4being wound -on the same legs, there will be substantially perfect coupling between these coils. We have further assumed that M=M1=M9=M5=M6, because there are only two legs carrying the coils.

It is noted that the impedance Z appears in this equation only in one term of the denominator. For the output current to be independent of the impedance, the multiplier of the impedance must be 0, so that we obtain the following equation:

2 u: i K+ L1L2-M2 This obviously is the condition for pure constancy of the load, or output, current. Through simple anithmetical operations on the last equation, we obtain the following equation:

A [fully-M2] 1 It will be noted that this equation is similar to one for a parallel resonant condition.

From actual operation of the system of FIG. 17, i-t is [known that the `out-put current is substantially constant for av range of variation of load impedance upwards from zero. There-fore, this last equa-tion must be substantially rtrue in the system. However, also for the system of FIG. 17, it has been found that changes in the capacity of the control circuit make no change in the constancy of current obtained with the system, only changing the value of the output current. This is of course in extreme contrast with the operation of the conventional parallel resonant circuit, because if one of the react-ances of such a cincuit is changed, the opposite sign reactance must be correspondingly changed to compensate and maint-ain resonance. Evidently, the system adjusts itself automatically to remain in resonance by `adjusting the sign and magnitude of the mutual coupling between coils on the two outer legs. In other words, over a range of variation of the effective capacity K of lthe system of FIG. 1-7, the mutual inductance M changes between -a range of a negative value through zero to a positive value. How great is the magnitude `of the extremes is not known, but it is put, or load, electrical circuit.

known that M swings through zero between positive and negative.

`It will further be noted that the output current Io is directly dependent upon the capacity of the control circuit capacitor K, so that increase in the effective capacity K causesa direct increase in the output current.

The system of FIG. 1 does not use a separate coil in the control circuit. That system is subject to modification to obtain the transformer system of FIG. 19 inl which -a fifth coil, L3, is wound on one of the outer legs, and the series combination of L3 and capacitor C is connected across output coil L2'. This system operates very simiylarly to the system of FIG. 17, the difference in operation being only that occasioned by the larger voltage available to drive control current IL through the capacitor. In efect, then, this increases or magnifies the effect of the capacitor on the system.' The use of this separate control coil L3 may be interpreted into the equations above by substitution of the actual value of the capacity C in the equation using the equivalency K=2C, where =1i `and, a equals the turns ratio between the control coil L3 and the secondary coil L2'.

The actual operation of the system of FIG. 17 and FIG. 19 is best explained by reference to FIG. 20, showing a vector diagram of the various currents, voltages, and magnetomotive forces, for increasing impedance, using an actual transformer core material. The voltage across Ithe coils L1 and L2 of the input electrical circuit is shown in the vector diagram as Epik, which is actu-ally the vector resultant of two out of phase voltages. This primary voltage generates a magnetomotive force which drives a prim-ary flux, the Iprimary flux generating voltages in the output circuit coils L1 and L2. These two coils are so wound and connected in series with respect to the direction of primary ux that their voltages oppose in the out- The two voltages generated in the output coils by the primary ux are shown as Basic Ef* and Basic Eg* and are in phase opposition. The voltage Basic Ez* combines with a similar voltage across the control coil L3 to provide a driving voltage for `a control coil current IL. This current, because of the capaci-tor in the control circuit, is 90 out of phase With the driving voltage. The control coil current generates a control magnetomotive force FL in phase with it, which in turn causes an initial circulting flux to flow between the two outer legs of the magnetic circuits. As described in conjunction with -the first class of embodiments, this circulating ux generates voltages in output coils L1 and L2 lwhich add together to produce a net output voltage E0.

, This net voltage drives an output current L, here shown in phase with the output voltage, because we are assuming that the loadv is highly resistive. The output current through the Vtwo coils L1 and L2 generates an output magnetomotive force Fo in phase with it, `and the output magnetomotive force and the control magnetomotive torce, substantially 180 out of phase, as shown in the vector diagram, combine to produce a net magnetomotive force.

The net magnetomotive force generates a circulating flux linking the two coils of the output circuit and producing voltages therein shown as E infLl and E0 in L2. These voltages again combine to produce the net outputl voltage E0. From the vector diagram it will be noted that `as the impedance increases, the output voltage En increases continuously, but changes its phase. The output current first decreases slightly, while rotating clockwise, then reverses its rotation and increases to a peak, until it reaches a level at which it begins to decrease continuously with further increase in output voltage. action of the output current with increasing impedance is shown in FIG. 2l, illustratingthe output current versus output voltage for three different values of effective capacity in the control circuit. For the highest curve particularly, it will be noted that the output current first decreases, then increases to a peak, and then drops oir con- This tinuously for increasingV output impedance. The action of the output current first decreasing and then increasing can be explained by reference to the equation for the output current, as well as by the theoretical analysis given for the rst class of the invention. In the equation, it will be noted that one term in the numerator is directly dependent for its value on the difference between the inductances of the two coils on the two outside legs. The inductances of such coils are of course dependent directly upon the voltages across these coils, so thatthey are equal at Eo=0, and, at relatively high levels of magnetic field intensity at which transformer systems of this type are operated, an increase in voltage causes a decrease in inductance of a coil. From FIG. 20, it is obvious that the voltage across coil L2 decreases at first with increasing impedance, while the voltage across L1 increases. Therefore, it is obvious that the inductance L2 increases as the voltage net E2 decreases, While the opposite action takes place for the other coil in the output circuit. From the equation for the load or output current, the output current thus decreases at first. However, net E2', after decreasing for a time, then increases, while net E1 decreases. T his' reverses the direction of change of the inductance, so that the output current increases till it reaches a peak. This peak apparently is the region in which saturation of the magnetic core material takes place. As saturation enters into the picture (occurring first in the leg not carrying the control coil and then `affecting the other outside leg), it requires more and more magnetic eld intensity to obtain even a slight increase in iiux density in the core material. The output current then drops oli continuously with increasing load. This sequence emphasizes the disadvantageous action of saturation on the system. If there were no saturation, the range of constancy would only be limited by leakage. The range within which the transformer operates is shown in FIG. 9 for the embodiments of the first class,

and the same range is used for the second class.

It would be possibleto go through a vector analysis similar to that of FIGS. 4 and 6, for this second class of embodiments, but, since the operation of this class of embodiments of the invention in achieving current constancy is quite similar to that of the first class of embodiments, the vector analysis will not be repeated.

Referring now to FIG. 22, vthe magnetic circuits of the apparatus of FIGS. 17 and 19 include a first and a second magnetic circuit. One circuit includes portions a-b-c-de-f-a of the core, this being the output magnetic circuit for the connections of FIGS. 17 and 19. The other magnetic circuit includes portions g-b-c-d-e-g and g-b-a-fe-g of the core, this -circuit being the input magnetic circuit. for the connections of FIGS. 17 and 19. The input magnetomotive force is shown as Fp in FIG. 22, while the output'rnagnetomotive force is shown as F0. It will be noted that the two magnetic circuits have two common portions, including parallel portion a-f and parallel portion c-d. In parallel portion a-f the two magnetomotive force are in opposition, while in parallel portion c-d the twomagnetomotive forces are aiding. This condition would result in failure to couple power between the input and the load, if there were no control circuit in the system. However, the control circuit, because of the phase change caused bythe use of the capacity in the system, generates a control magnetomotive force FL which is larger than and in phase opposition with the output magnetomotive force. The control magnetomotive force in this leg is thus reversed from Fo, and power is coupled between the input and the load through the control circuit.

In FIG. 1l above, the operation of the embodiments of the first class of this invention has been further described as that of a servo system. The same description applies to the embodiments of this second class of the invention.

It will be obvious thatthe load could be connected across coils L1 `and L2, and the input connected across coils L1 and L2', as shown in FIG. 23. These connections yield the Same results as obtained with the original connections shown in FIG. 17. As noted above, the

input and output coils must be so phased that the primary ilux and the output lux follow two circuits having an uncommon portion-the center leg. However, either ilux may be the one which follows the center leg. Further, the capacity K can lne-connected across L2 of FIG. 17, in place of L2', with the same results.

FIG. 24 shows the use of a separate control coil for the apparatus of the preceding gures. Control coll L3 is wound on the same leg with coils L2 and L2', and the series combination of this coil and capacitor C may be connected` across any appropriate source of voltage, as indicated in the gure. For instance, the series combination may be lconnected across L2', L2, L1, L1', Ll-i-L2, L1l-L2, any other source of voltage, or the leads for this -control circuit may be shorted on each other, so

that the driving voltage for the control current is obtained l only from control coil L3. The constancy of output current can-be maintained with connection of the control circuit across any source of voltage, but substantially similar phase relationships to those shown for FIGS. 17 and 19 are preferable for best range of constancy. In other words, when the source voltage for the control impedance is so related to the primary voltage that the control current and the primary voltage are 90 out of phase at zero load, the range of output voltage during which the current is maintained constant is the greatest.

For commercial use of the circuits of the preceding tigures it will be desirable that the level of output current I 'which is to be maintained constant may be adjusted. AIn

order to provide for such adjustment, the valve of the capacitor may be directly changed, but, particularly at the high voltages at which a system of this type will probably be operated, it is advantageous to adjust the effective value of the capacity, rather than the actual capacitance. In order to permit this, the system of FIG. 25 provides a variable inductance 40 shunted directly across the capacitor K. Adjustment of the inductance of the variable inductor will changethe effective capacity in the control circuit.

It will also be evident that the capacity need not be directly coupled into the control coil circuit, but may be coupled in any appropriate manner. FIG. 26 shows a circuit in which the capacity is coupled by means of a transformer T into a control circuit. The capaci-tor C is shunted acrossthe secondary coil of the transformer, while the primary of the transformer is connected across coil L2'. Moreover, the capacitor might be coupled into the control circuit through a variable transformer, or Variac, thus enabling the effective capacitive reactance to be changed by change in the transformer turns ratio.

Another way of providing for variation or adjustment in the effective capacity of the control circuit, without direct coupling between the capacity and the control circuit, is shown in FIG. Z7, in which capacitor C is connected across the secondary coil 41 of a saturable transf-ormer 4|2. The level of saturation of the transformer is controlled by direct current flowing from a source of variable voltage 43 through series-connected coils 4-4 and 45 wound on opposite legs of the core of the transformer. The prim-ary coil 47 of the transformer couples the capacity into the control circuit by connection of the primary coil directly across coil L2 on the constant current transformer core. Change of the value of the D.C. voltage from source 46 causes a change in current through coils 44 and 45, resulting in a change in the saturation of the outer legs of the core and a corresponding change in the primary exciting current. Thereby, a change in the effective capacity in the control circuit is obtained in the manner described in conjunction with FIG. 16.

As explained above, all of the classes of embodiments of the invention have a substantially unlimited constancy action with variation in capacity of the control circuit capacitor. When leakage reactance is neglected, all units described herein have unlimited constancy. This surprising result is achieved by automatic variation in the mutual inductance of coils on separate legs to compensate for changes in capacity, so that resonance is substantially maintained. As far as is known, the action of such systems in varying mutual inductance automatically, without adjustment of parts, is novel. Also, as explained above, a ferromagnetic core is not essential to the operation of this inventiontand some of its characteristics actually limit the range of operation. An air coreversion of one of the classes of embodiments was tested at high frequency and it was found that the range of constancy was much greater than that indicated by.iron core results, thus proving that saturation limits the range. The range in an air core embodiment appears limited by leakage, rather than saturation.

As will be evident, autotransformer arrangements, rather than those using unconnected primary and secondary coils, may be used. For instance coil L2 of FIG. y17 could be omitted and one end of coil L2 connected to L1 while the other is connected to the side of the load remote from the connection of coil L1 thereto. With coils L1 and L2 phased to oppose in the secondary circuit, the apparatus would work in the same manner as that shown in FIG. 17.

It will be obvious that many minor variations could be made in the elements of the various embodiments of this invention shown and described without departure from -the spirit of the invention. For instance, the positions of the various coils on the respective legs of the transformer core could be changed about and the characteristics fundamental to the system still be maintained. IFrom the description of the various embodiments, it will `be evident that all of these embodiments have the following in common:

The embodiments include two magnetic circuits, which have two common portions, in one of which common portions the input and output magnetomotive forces aid, and in the other of these two common portions the input and output magnetomotive forces oppose, so that there is no direct coupling between the input and the load. The control magnetomotive force in effect reverses the total or resultant magnetomotive force in its common portion from fthe direction of the output m.m.f., so as to couple power from the input to the load through the control electrical circuit. 'From the above, it will be' obvious that this invention is not to be considered limited to the various embodiments shown and described, but rather only by the scope of the appended claims.

1. A transformer system for supplying power to a load from an A.C. input comprising a three-legged core of ferromagnetic material, Ithe legs of said core being of such configuration and so aligned with each other that substantially closed high permeability paths are `formed between all of the legs; at least three coils wound on said core, each of said coils being wholly wound on a different one of the legs and said yiirst and second of said coils being connected in series; an input, an output, and a control electrical circuit, all passing current, the input circuit including one of (a) the third of said coils and (b) the series combination of said `first and second coils and being connected across said input, the output circuit including the other of (a) said third coil and (b) the series combination `of said rst and second coils and being connected across said load; said input andvoutput electrical circuits forming with said core an input and an output magnetic circuit, one including all three of said legs and the other including the first and second of said legs but bypassing at least the major poi-.tion of the third leg, said input and output electrical circuits being so inductively related to said rst and second legs that the input and output magnetomotive forces generated bythe input and output currents aid in one of said rst and second legs and oppose in the other of said rst and second legs; and a capacitive reaotance, said control electrical lcircuit including said capacitive reactance and being in ductively coupled to the leg on which one of said first and second coils is Wound, said one coil being wound on one of said first and second legs; whereby no power is directly coupled between said input and output electrical circuits, but the control current producing a magnetomotive force in said one of said first and second legs of phase and magnitude to produce a resultant of the control and output magnetomotive forces opposite to the output magnetomotive force in said last-mentioned leg and thereby to couple power from the input to the load through said control electrical circuit.

2. The apparatus of claim l in which said input electrical circuit includesl said third coil and said output electrical circuit includes the series combination of said first and second coils.

3. The apparatus of claim 1 in which vsaid input electrical circuit includes the series combination of said first and second coils and said output electrical circuit includes said third coil.

4. The apparatus of claim 1 in which said control electrical circuit includes said one of said first and second coils and said capacitive reactance is shunted across said one coil. Y j

5. The apparatus of claim 1'in which said control electrical circuit includes a control coil Wound on said one of said first and second legs.

6. The apparatus of claim-5 in which the series combination of said control coil and said capacitive reactance is connected across a source of voltage.

7. The apparatus of claim 6 in which the series combination of said control coil and said capacitive reactance is connected across at least one of said first, second, and third coils.

8. The apparatusof claim 5 in which the series cornbination of said control coil and said capacitive reactance is connected across the one of said first and second coils' wound on the same leg with it.

9. The apparatus of claim S in Iwhich the capacitive reactance is shunted across the control coil. Y

10. The apparatus of claim 8 in which said control electrical circuit includes a variable inductive reactance shunted across -the capacitive reactance to permit variation of the effective capacity of the capacitive reactance and hence of the output current.

11. The apparatus of claim 8 in which said control electrical circuit includes a saturable core transformer, at leas-t one coil wound on the core of said transformer, a source of D.C. voltage connected to said last-rnen tioned coil and variable to change the degree of saturation of the transformer core, a secondary coil wound on said transformer core and connected across said capacitive reactance, and a primary coil inductively related to said secondary coil.

12. The apparatus of claim 1 in which said control 55 electrical circuit includes a transformer having a primary and a secondary coil, said capacitive reactance being conv 20 nected across said secondary coil, and said primary coil being inductively coupled to said secondary coil.

13. A transformer system for supplying a substantially constant current to a load over a range of variation of said load upward from Zero from a substantially constant voltage A C. input comprising at least three inductively coupled electrical coils, a first and a second of said coils being connected in series; an input, an output, and a control electrical circuit, all passing current, the input electrical circuit including one of (a) a third of said coils and (b) the series combination of saidy first and second coils and being connected across said A C. input, the output electrical circuit including the other of (a) said third coil and (b) the series combination of said first and second coils and being connected across said load; the coils of said input and said output electrical circuits being linked by magnetic paths forming an input and an output magnetic circuit having two common portions and said input and output electrical circuits being so inductively related to said common portions that the input and output magnetomotive forces generated by the input and output currents aid in one of said common portions and oppose in the other; and a capacitive reactance, said control electrical circuit including said capacitive reactance and being inductively coupled to a first onevof:` said common portions; whereby no power is directly coupled between said input and output electrical circuits, but the control current producing a magnetomotive force in said first common portion of the magnetic circuits of phase andmagnitude to produce a resultant of the control and output magnetomotive forces opposite to the output magnetomotive force in said rst common portion and thereby to couple power from the input to the load through said control electrical circuit, said capacitive reactance being of adjustable magnitude and the output current being directly proportional thereto so as to permit adjustment of the level of electrical current to be maintained constant, said transformer system automatically compensating for changes in the magnitude of the capacitive reactance to maintain itself substantially in resonance through compensating changes in the magnitude of the mutual inductance of said coils.

References lCited bythe Examiner UNITED STATES PATENTS 1,599,570I 9/ 1926 Lucas 323-60 2,195,969 4/1940 Minor S23-60 X 2,207,234 7/1940 Bohm 323-60 X V2,212,198 8/1940 Sola 323-60 X 2,305,153 12/1942 Fries 323-60 X 2,403,393 7/ 1946 IPeterson 323-60 2,512,976 6/1950 Smeltzly- 323-6 2,605,457 7/1952 Peterson 323-6 X 2,811,689 10/1957 Balint 323-60 X LLOYD MCCOLLUM, Primary Examiner'.

WARREN E. RAY, Assistant Examiner.

Claims (1)

1. A TRANSFORMER SYSTEM FOR SUPPLYING POWER TO A LOAD FROM AN A.C. INPUT COMPRISING A THREE-LEGGED CORE OF FERROMAGNETIC MATERIAL, THE LEGS OF SAID CORE BEING OF SUCH CONFIGURATION AND SO ALIGNED WITH EACH OTHER THAT SUBSTANTIALLY CLOSED HIGH PERMEABILITY PATHS ARE FORMED BETWEEN ALL OF THE LEGS; AT LEAST THREE COILS WOUND ON SAID CORE, EACH OF SAID COILS BEING WHOLLY WOUND ON A DIFFERENT ONE OF THE LEGS AND SAID FIRST AND SECOND OF SAID COILS BEING CONNECTED IN SERIES; AN INPUT, AN OUTPUT, AND A CONTROL ELECTRICAL CIRCUIT, ALL PASSING CURRENT, THE INPUT CIRCUIT INCLUDING ONE OF (A) THE THIRD OF SAID COILS AND (B) THE SERIES COMBINATION OF SAID FIRST AND SECOND COILS AND BEING CONNECTED ACROSS SAID INPUT, THE OUTPUT CIRCUIT INCLUDING THE OTHER OF (A) SAID THIRD COIL AND (B) THE SERIES COMBINATION OF SAID FIRST AND SECOND COILS AND BEING CONNECTED ACROSS SAID LOAD; SAID INPUT AND OUTPUT ELECTRICAL CIRCUITS FORMING WITH SAID CORE AN INPUT AND AN OUTPUT MAGNETIC CIRCUIT, ONE INCLUDING ALL THREE OF SAID LEGS AND THE OTHER INCLUDING THE FIRST AND SECOND OF SAID LEGS BUT BYPASSING AT LEAST THE MAJOR PORTION OF THE THIRD LEG, SAID INPUT AND OUTPUT ELECTRICAL CIRCUITS BEING SO INDUCTIVELY RELATED TO SAID FIRST AND SECOND LEGS THAT THE INPUT AND OUTPUT MAGNETOMOTIVE FORCES GENERATED BY THE INPUT AND OUTPUT CURRENTS AID IN ONE OF SAID FIRST AND SECOND LEGS AND OPPOSE IN THE OTHER OF SAID FIRST AND SECOND LEGS; AND A CAPACITIVE REACTANCE, SAID CONTROL ELECTRICAL CIRCUIT INCLUDING SAID CAPACITIVE REACTANCE AND BEING INDUCTIVELY COUPLED TO THE LEG ON WHICH ONE OF SAID FIRST AND SECOND COILS IS WOUND, SAID ONE COIL BEING WOUND ON ONE OF SAID FIRST AND SECOND LEGS; WHEREBY NO POWER IS DIRECTLY COUPLED BETWEEN SAID INPUT AND OUTPUT ELECTRICAL CIRCUITS, BUT THE CONTROL CURRENT PRODUCING A MAGNETOMOTIVE FORCE IN SAID ONE OF SAID FIRST AND SECOND LEGS OF PHASE AND MAGNITUDE TO PRODUCE A RESULTANT OF THE CONTROL AND OUTPUT MAGNETOMOTIVE FORCES OPPOSITE TO THE OUTPUT MAGNETOMOTIVE FORCE IN SAID LAST-MENTIONED LEG AND THEREBY TO COUPLE POWER FROM THE INPUT TO THE LOAD THROUGH SAID CONTROL ELECTRICAL CIRCUIT.

Images