US3612988A

US3612988A - Flux-gated voltage regulator

Info

Publication number: US3612988A
Application number: US857906A
Authority: US
Inventors: Cravens L Wanlass
Original assignee: Individual
Current assignee: Individual
Priority date: 1969-09-15
Filing date: 1969-09-15
Publication date: 1971-10-12
Anticipated expiration: 1988-10-12
Also published as: FR2140283B1; FR2140283A1; DE2126819A1; GB1329018A

Abstract

A voltage regulator in which an inductor having a saturable core is connected in series with a capacitor, the output being taken from across the inductor. After startup, the combined effect of the input voltage and the capacitor cause the core of the inductor to periodically switch from a nonsaturated to a saturated condition and vice versa so that the impedance of the inductance switches correspondingly from a high to a low value. The result is a square wave voltage developed across the inductor that has a constant amplitude regardless of input voltage variations if the input frequency remains constant.

Description

United States Patent [72] Inventor Cravens L. Wanlass 9871 Overhill Drive, Santa Ana, Calif.

92705 [21] Appl. No. 857,906 [22] Filed Sept. 15, 1969 [45] Patented Oct. 12, 1971 x [54] FLUX-GATED VOLTAGE REGULATOR 19 Claims, 8 Drawing Figs. [52] U.S. Cl 323/43.5 R, 323/50, 323/60, 323/76, 323/91 [51] Int. Cl G05f1/l4, G05f 3 04 [50] Field of Search 323/6, 43.8, 47, 48, 50, 51, 52, 60, 76, 91 [5 6] References Cited UNITED STATES PATENTS 2,593,651 4/1952 Bird 323/43.5 X 2,605,457 7/1952 Peterson 323/6 X Primary ExaminerJ. D. Miller Assistant ExaminerA. D. Pellinen AttorneyLyon & Lyon ABSTRACT: A voltage regulator in which an inductor having a saturable core is connected in series with a capacitor, the output being taken from across the inductor. After startup, the combined effect of the input voltage and the capacitor cause the core of the inductor to periodically switch from a nonsaturated to a saturated condition and vice versa so that the impedance of the inductance switches correspondingly from a high to a low value. The result is a square wave voltage developed across the inductor that has a constant amplitude regardless of input voltage variations if the input frequency remains constant.

PATENTEDDBT 12 ml 9 ,6 1 2, 9 88 sum 1 8F 2 CONTEOL FLUX-GATED VOLTAGE REGULATOR BACKGROUND OF THE INVENTION A commonly used type of voltage regulator for power applications is the ferroresonant or Sola transformer. In such a device, a magnetic core is provided with a shunt leg having an airgap. Primary and secondary windings are wound on the core on either side of the shunt leg. The operating point of the device is established such that the core leg on which the secondary is wound'saturates in each half cycle of the input. The remainder of the flux will then pass through the shunt leg. A capacitor is connected in parallel with the secondary winding to form a resonant circuit and the resultant output is a square wave having a relatively well-regulated amplitude over the operating range of the device.

The ferroresonant transformer is satisfactory in operation for many applications and has found wide acceptance. However, it is quite bulky and heavy due to the inherent necessities of the core and windings. The same need for a relatively large volume of core material and windings also makes the device relatively expensive.' Another disadvantage of the Sola type device is that although it will current limit if theinput exceeds the rated value, the amplitude at which thecurrent is limited is several times higher than that of the operating value. This can cause severe damage to certain kinds of loads.

SUMMARY OF THE INVENTION According to the present invention, a voltage regulator is provided which requires less core material than previously known devices and also requires only one winding instead of two, although if desired a second winding may be provided for control purposes. The regulator is more efficient and smaller than such previously known devices as ferroresonant transformers. The regulator of the present invention provides a square wave output so that it is very useful, for example, in DC power supplies, and has automatic overload protection. These characteristics are obtained by utilizing the saturation properties of an inductor connected in series with a capacitor, the inductor' being caused to saturate when the input voltage reaches a predetermined amplitude and unsaturate upon charging of the capacitor. The resulting square wave voltage across the inductor, or a portion of it, provides the regulated output voltage. Since the device is frequency responsive it can be used as a filter which, of course, will still produce a regulated output and which is capable of substantial energy transfer despite high selectivity.

It is therefore an object of the present invention to provide a voltage regulator that is relatively inexpensive, highly efficient and reliable.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a partially schematic and partially perspective view of a first embodiment of the regulator of the present invention;

FIG. 2 illustrates the waveforms present at various times in the operating cycle of the regulator of FIG. 1;

FIG. 3 illustrates the 8-H curve of the core of the inductor used in the present invention;

FIG. 4 is a schematic diagram of a DC voltage regulator employing the regulator of FIG. 1;

FIG. 5 is a schematic diagram of a variable output voltage regulator employing the regulator of FIG. 1;

FIG. 6 is a partially schematic and partially perspective view of a modification of the regulator of FIG. 1;

FIG. 7 is a schematic diagram of a second embodiment of a voltage regulator according to the present invention; and

FIG. 8 is a schematic diagram of a closed loop DC voltage regulation system employing the regulator of FIG. 7.

DESCRIPTION OF THE INVENTION Turning now to FIG. 1, there is illustrated the voltage regulator of the present invention. An AC voltage is applied to input terminals 10 and 11 which are connected to a series circuit comprising an inductor l2 and a capacitor 13. The inductor 12 comprises a winding 14 wound on a magnetic core 15. As illustrated, the core 15 is generally rectangular in configuration and typically would be constructed of a pair of mated C-cores. However, the configuration of the core is not important to the operation of the regulator of the present invention and any other core structure such as standard transformer laminations could be used. A double-pole, single-throw switch 16 has a first contact 17 connected to the upper end of winding 14 .and a second contact 18 connected to a tap 19 on the winding 14 so that movement of the arm 16' of the switch 16 from contact 17 to contact 18 will cause a number of the turns of the winding 14 to be bypassed. Output tenninals 20 and 21 are respectively connected to the upper end of the winding 14 (when switch 16 is in its normal position, i.e., arm 16' is engaging contact 17) and a second tap 22 thereon. The positioning of this tap 22 is determined by the output voltage desired, that is, the winding 14 operates as an autotransformer for purposes of producing an output voltage.

The operation of the voltage regulator shown in FIG. I will now be described, reference being made to FIGS. 2 and 3. Before the arm 16' of switch 16 is moved from contact 17 to contact 18, the inductance 12 presents a very high impedance to the input voltage with the result that very low current would flow and there would be a very low output voltage. This is because the core l5 of the inductor 12 is unsaturated and the inductance of the inductor is correspondingly very high. This condition will be maintained until the switch arm 16' is moved from contact 17 to contact 18. Upon such movement of the switch arm 16, the number of turns of the winding 14 is decreased with the result that the inductance presented to the source is considerably reduced and a relatively high current flows through the remaining turns of the winding 14. This current results in a high fiux density being generated in the core 15, causing it to, saturate and operation of the regulator to commence. If desired, the C-cores making up the core may be ofi'set to provide a very high flux density where they are joined. It should be understood that the circuit may be started in any way that provides sufi'lcient energy to saturate the core. For example, energy could be directly transferred to the capacitor so that it becomes charged. Alternatively, the core could be saturated by use of an additional winding on the core that is momentarily energized to produce sufficient flux density. A few turns of the inductor can also be momentarily shorted to institute operation of the circuit. Since the inductor 12 has been storing energy in accordance with the formula E= b LI, where E represents energy, L inductance, and l current, the sudden change of inductance from a high value to a low value when saturation occurs results in a transfer of energy from the inductor 12 to the capacitor 13 causing the latter to charge. The low impedance of the now saturated inductor 12 also permits current to flow essentially directly from the input to assist in charging the capacitor 13. The switch 16 is now returned to its normal position, i.e., the arm 16' engaging contact 17 and the regulator goes into steady state operation.

The values of the circuit parameters are not critical; it will be apparent, however, that proper operation of the regulator requires that the value of the inductance be large enough to prevent initial saturation thereof by the source alone and that the capacitor be large enough to store sufficient energy to cause saturation of the core of the inductor when added to the input voltage in the manner previously described.

Referring to FIG. 2, the waveforms present in the system of FIG. 1 are illustrated. As can be seen, the capacitor voltage V and the inductor voltage V are quite large and out of phase with one another by approximately Also illustrated are the input voltage V and the input current I It should be understood that the amplitude and phase relationships are illustrative only and presented as an aid in understanding the invention since they vary somewhat in accordance with different output currents, input voltage amplitude, etc.

Assume the circuit is operating at point A at FIG. 2. At this point, the impedance value of the inductor 12 is high since it is not saturated. There is a positive voltage across the capacitor 13 which is held fairly constant since the capacitor cannot discharge rapidly due to the high value of the inductance. The operating point on the B-I-l curve of the inductor 12 is shown in FIG. 3 at approximately point A. The current waveform of FIG. 2 shows the current through the series circuit of the inductor 12 and capacitor 13 due to the voltage across the inductor 12 caused by the charge on the capacitor 13 in series with the applied line voltage.

This operating condition remains approximately the, same until the input voltage begins to change polarity. As this occurs, the voltage across the inductor is increased to a point where the inductor can no longer sustain the voltage across it as a function of time (i.e., volt-seconds) and the core 15 of the inductor l2 saturates. This is shown at point B on FIG. 3. Upon saturation of the inductor 12, its inductance suddenly switches to a low value, allowing the capacitor 13 to discharge with the result that the current through the inductor 12 becomes very large as shown in FIG. 2 at point B. Shortly after this point, the total energy in the system is contained in the inductor in the form of A LI as the voltage across the capacitor 13 has dropped to zero.

As can be seen from FIG. 2, the current has been in the same direction during this switching operation. When the capacitor 13 discharges, the inductor 12 becomes the energy source and begins to recharge the capacitor in the opposite, i.e., negative, polarity. This is due to the fact that the inductor l2 builds up a voltage opposite to that of the applied voltage in attempting to keep the current constant. The capacitor 13 continues to'be charged by the energy stored in the inductor and by the' line voltage.

It appears that it is at this time that the input power source transfers energy to the system to make up for energy lost into the load, etc. This energy transfer probably occurs when the voltages across the inductor and capacitor are both low. As the capacitor l3'charges'still further the current through the inductor and capacitor begins to decrease. This process continues until the current through the inductor 12 and thus the ampere turns decreases to the point where the inductor again goes out of saturation. When this occurs the impedance of the inductor l2 suddenly increases to a large value so that the charging of the capacitor is stopped and the capacitor voltage is held essentially constant. This point of operation is shown in FIG. 2 as point C. The capacitor 13 now becomes the source of energy and drives a small current in the opposite direction through the inductor 12..This biases the inductor 12 at point D on the hysteresis curve of FIG. 3.

The process is now set to repeat its operation when the input voltage again changes its polarity and increases its magnitude due to its AC nature. When this input voltagereaches a sufficient magnitude, it causes the inductor 12 to saturate in the opposite direction as shown at point B on FIG. 3. The process is repeated over and over again, each polarity reversal of the input voltage causing the inductor to be triggered into saturation. The result is a voltage across the inductor 12 that has a constant RMS value. Since the output is taken across a portion of the inductor, this voltage represents a predetermined percentage of the voltage across the inductor and its amplitude is correspondingly maintained constant.

The regulated output voltage occurs from the fact that for a given operating frequency, core material, and winding, the 8-H characteristics of the device are fairly constant and therefore the volt-seconds required to cause the inductor 12 to saturate is also fairly constant. In other words, the volt-second value resulting from addition of the line voltage after its change of polarity to the voltage already established across the inductor exceeds the inductors ability to maintain it with the result that the inductor saturates. Since the slope of the applied AC voltage is greatest when it crosses zero, a reasonable change in the amplitude of that voltage will not significantly alter the volt-second value resulting from this voltage. Consequently, the saturation of the inductor will be triggered at essentially the same amplitude point of the input voltage, re-

gardless of the maximum amplitude this voltage might reach. In other words, if the amplitude of the input should rise, the instant when the inductor switches will change slightly. However, the instant when it switches again (i.e., on the next half cycle) will also be correspondingly changed. The time interval between these two switching points will remain constant and consequently the voltage will be constant since the voltsecond value remains constant. The only variation in output voltage that occurs will take place during the time when the input changes from one value to another but this variation will be insignificant.

' A change in the input voltage will result in a phase shift-t0 the left in FIG. 2 if the voltage increases and to the right if it decreases. However, since this phase shift is constant for any given input voltage, it can be seen that the time interval will remain constant and so will the triggering voltage. Accordingly, the output voltage taken across the inductor is substantially independent of line voltage changes, and, within the limits of the device, load current changes. As will be noted from FIG. 2, the peak voltage value across the inductor is not absolutely constant but rather reflects to a small degree the input voltage waveform. This does not significantly affect the RMS regulation discussed and has only a very limited effect on peak or average regulation.

If desired, electrical isolation can be achieved by trapsformer coupling the output circuit to the inductor 12. Such a circuit is shown in FIG. 4 where similar elements are identified with the same reference numerals as were used in FIG. I. As can be seen, a winding 25 is wound on the core 15 with the winding 14. The output of the winding 25 can be used directly as an AC output or can be suitably rectified and filtered as shown to produce a DC output at

terminals

26 and 27. If desired, a plurality of windings having different numbers of turns can be wound on the core 15 so that difi'erent output voltages can be obtained. The outputs of these windings may be rectified or left as AC.

FIG. 5 shows a modification of the circuit of FIG. 1 which permits the output voltage to be varied. Again, the same reference numerals are used for the same elements. Here, the secondary winding 25 is connected to an autotransformer 28, the tap 29 of which can be moved to produce a variable voltage. Of course, if desired, the output across the terminals 30 and 31 can be rectified if a DC potential is required. As previously pointed out, the circuit of the present invention is particularly useful in providing a regulated DC voltage as the square wave output of the device is easily rectified and filtered.

Turning now to FIG. 6, there is illustrated a modification of the regulator of FIG. 1. As was the case with FIG. I, the regulator basically comprises an inductor 40 coupled in series with a capacitor 41 between a pair of input terminals 42 and 43. A switch 44 is provided having a movable switch arm 44' movable between an upper contact 45 and a lower contact 46 to selectively bypass a number of turns on a winding 47 wound on a ferromagnetic core 48. The operation of the device so far described is identical to that of the device shown in FIG. 1. In addition to the main core 48, the inductor 40 is provided with an auxiliary core 49 on which is wound a control winding 50. As can be seen from FIG. 6, the auxiliary core 49 is generally U- or C-shaped with its ends 51 and 52 engaging the main core 48 in such a way that the auxiliary or bias flux generated in the auxiliary core 49 as a result of a current in the winding 50 will, in completing its magnetic circuit, pass through a portion of the core 48 in a direction substantially orthogonal to the primary flux in the core 48 resulting from a current in the winding 47. The core 49 is, of course, fitted to the core 48 as carefully as possible so as to eliminate to the extent possible any airgaps existing between the core 48 and the ends 51 and 52 of the core 49.

As will be apparent, the auxiliary or bias flux in the core 48 will effectively reduce the cross section of the core 48 available to the primary flux. Since the fluxes are orthogonal and symmetrical on each half cycle, no flux coupling will occur between the two windings 47 and 50.

As can be seen, the magnitude of the current in the winding 50, and consequently the orthogonal bias flux generated in the core 48 between the ends 51 and 52 of the core 49 will affect the flux-carrying capacity of the core 48 as seen by the primar-y'flux therein and consequently will affect the point at which the inductor 40 will trigger, i.e., saturate. This phenomena can be used to construct a closed loop regulator as shown in FIG. 6 by the simple expedient of providing a sensing and control circuit 53 which senses the voltage appearing across

output terminals

54 and 55 and provides a corresponding control current to the winding 50.

Assume, for example, that a change in output load or input line frequency causes the output voltage to start to increase in magnitude. This increase is sensed by the circuit 53 which increases the control current in the winding 50 accordingly. This increase in control current will cause an increase in bias or control flux in the core 48 between the ends 51 and 52 of the core 49 with the result that the cross section of the core 48 is effectively reduced. This reduces the point at which the inductor 40 triggers and this in turn causes the output voltage to decrease. The contrary would be true if the output voltage originally decreased. While an AC regulator is shown, it should be understood that a regulated DC output voltage could easily be obtained by the use of a suitable rectifier circuit. It should be understood that, if desired, the output voltage could be taken from a second winding wound on the core 48 in the manner shown in FIGS. 4 and 5.

FIG. 7 shows an embodiment of the present invention in which a pair of toroidal cores 60 and 61 are used in place of the single core shown in the previous figure. Basically, the regulator of FIG. 7 operates in the same manner as the regulator of FIG. 6. The input voltage is applied through

terminals

62 and 63 to a series circuit comprising a capacitor 64 and an inductor 65 made up of a winding 66 wound on the core 60 and a winding 67 wound on the core 61. The output is taken across a portion of each of the

windings

66 and 67 and appears at

output terminals

68 and 69. As was the case previously, the regulator is started by means of switches 70 and 71 which can be operated to bypass a portion of each of the

windings

66 and 67. If desired, only a single such switch can be used but it is preferable to maintain the device in the balanced condition.

The output of the regulator can be controlled by controlling the volt-second capability of the cores ,60 and 61. This can be accomplished by means of

windings

72 and 73 wound on these cores and connected to a control source 74. In order that no induced voltage be introduced into the control circuit, the

windings

72 and 73 are balanced and are wound in bucking relationship. As will be apparent, the presence of a control current in the

windings

72 and 73 will affect the volt-second capability of the cores 60 and 61 and accordingly will affect the point at which the inductor 65 triggers, therefor correspondingly affecting the output voltage produced across the

terminals

68 and 69. As will be apparent to those skilled in the art, any core in which the windings do not directly couple but in which the current in one winding can be used to control the triggering time of the inductor winding can be used in place of the two cores shown in FIG. 7.

It will be apparent that the regulator shown in FIG. 7 is particularly suitable for use in a closed loop regulating system. Such a system is shown in FIG. 8. In this Figure, an unregulated AC voltage is applied to input

terminals

80 and 81 which are connected in series with a circuit comprising a capacitor 82 and an inductor 83 made up of a first winding 84 and a second winding 85, these windings being wound on cores such as the cores 60 and 61 shown in FIG. 7. Operation of the regulating circuit is initiated by means of

switches

86 and 87 in the manner previously described.

Control windings 88 and 89 are wound on the cores in the same manner as illustrated in FIG. 7 and receive a control current from a sensing and control circuit 90. The output voltage, rather than being taken from across a portion of the inductor 83, is developed across

windings

91 and 92 which are wound on the cores and flux-coupled to

windings

84 and 85 respectively. The

windings

91 and 92 are connected in series to the input of a bridge rectifier 93, the output of which is filtered by a capacitor 94 and appears across the

output terminals

95 and 96. The input of the sensing and control circuit is taken across these output terminals and 96. It is believed that the operation of this circuit is self-evident in view of the descrip--.

" In the foregoing description, it has been pointed out that a.

closed loop circuit can be provided by varying the point at which the inductor triggers, this variation being the result of a' change in the magnetic circuit of the inductor. Since the regulator of the present invention is input frequency sensitive, the triggering of the inductor, and hence the output voltage, can also be varied by varying the input frequency. This can be accomplished, for example, by rectifying the input voltage ancl applying it to a conventional DC to AC converter whose output frequency can be varied in accordance with a control signal. Typically, such converters operate in the kilocycle range and employ silicon-controlled rectifiers or power transistors to produce a square wave output of the desired frequency. The output frequency of such a converter can be controlled by a control signal derived from the output voltage in a conventional manner. In this manner, a closed loop is provided, any change in output voltage being counteracted by a change in input frequency to the device. v

It should be understood that the theories of operation set forth herein are believed by the inventor to best describe the physical phenomena present in the operation of the regulator of the present invention. These theories should, however, be taken only as the best presently available and are not meant in any way to limit the scope of the present invention.

From the foregoirig'description it can be seen that a simple, compact and reliable regulator has been provided. Since the output is a square. wave, the circuit not only regulates RMS voltage but also does a satisfactory job of regulating average and peak voltage. If the circuit is overloaded, it will cease to operate since the inductor 12 will no longer saturate as required because the energy removed from the circuit per cycle is greater than that added per cycle. This results in the inductance of inductor 12 remaining in the high-impedance state with the result that the output voltage and the output current are reduced to very low values. The power handling capacity of the system as a function of size and weight is excellent as is its efficiency.

The invention may be embodied in other specific forms not departing from the spirit or central characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive.

What is claimed is:

1. A voltage regulating circuit comprising an input adapted to be connected to a source of AC voltage; inductor means having saturable core means, capacitor means; means coupling said inductor and said capacitor means in series across said input; an output coupled across at least a portion of said inductor; and means external of said source operable to cause said core means to initially saturate, said inductor means having an inductance sufficiently large to prevent said source alone from initially saturating said core means.

2. The circuit of claim 1 wherein said operable means is operable to reduce the inductance of said inductor.

3. The circuit of claim 2 wherein said operable means comprises switch, means connected across a portion of said inductor.

4. The circuit of claim 1 wherein said capacitor means has a capacitance large enough so that it can store sufficient energy to cause saturation of said core in conjunction with said input.

5. An electrical circuit comprising an input adapted to be connected to a source of AC voltage; inductor means having saturable core means; capacitor means; means coupling said inductor means and said capacitor means in series across said input; an output coupled across at least a portion of said inductor means; said capacitor means being capable of being charged to a voltage sufficient, when added to said AC voltage, to develop a volt-second value across said inductor means greater than the volt-second capacity of said inductor means so that said inductor means will be driven into saturation in opposite directions on each half cycle of said AC voltage.

6. The circuit of claim wherein said core means comprises a single core and wherein a multitum winding is wound on said core to form said inductor means.

7. The circuit of claim 6 wherein said single core has an auxiliary core mounted thereon, said auxiliary core having a control winding wound thereon, application of current to said control winding causing the inductance and volt-second capacity of said inductor means to vary.

8. The circuit of claim 7 wherein said auxiliary core is mounted such that flux generated by said control winding will pass through said single core in a direction transverse to the direction of flux generated therein by said multitum winding.

9. The circuit of claim 8 wherein means are provided for sensing the voltage across said output, said sensing means being coupled to said control winding to supply a control current thereto.

10. The circuit of claim 5 wherein said core means comprises a plurality of cores and wherein a multitum winding has a portion thereof wound on each of said cores to form said inductor means.

11. The circuit of claim 10 wherein control winding means are wound on said cores whereby the inductance and voltsecond capacity of said inductor means may be varied.

12. The circuit of claim 11 wherein means are provided for sensing the voltage across said output, said sensing means being coupled to said control winding means to supply a control current thereto.

13. The circuit of claim 5 wherein said output is directly coupled to said inductor.

14. The circuit of claim 5 wherein said output includes a second winding wound on said core.

15. The circuit of claim 5 wherein rectifying means are coupled to said output.

16. A method for producing an output voltage across a saturable core inductor connected in a series circuit with a capacitor, comprising:

a. charging said capacitor to establish a first voltage across b. applying a voltage to said series circuit having a magnitude, duration and polarity sufficient, when added to said first voltage, to develop a volt-second value across said inductor greater than the volt-second capacity of said inductor to cause said inductor to saturate;

c. discharging said capacitor through said inductor;

d. recharging said capacitor in the opposite polarity until said inductor comes out of saturation;

e. reversing the polarity of said applied voltage to again cause saturation of said inductor; and

f. periodically repeating steps (c) through (e).

17. The method of claim 16 further comprising varying the volt-second capacity of said inductor means to vary said output voltage.

18. The method of claim 16 further comprising varying the volt-second capacity of said inductor means in response to said output voltage to maintain said output voltage at a constant value.

19. The method of claim 16 further comprising the step initially reducing the volt-second capacity of said inductor means to cause said applied voltage to saturate said inductor means and charge said capacitor.

Claims

3. The circuit of claim 2 wherein said operable means comprises switch means connected across a portion of said inductor.

6. The circuit of claim 5 wherein said core means comprises a single core and wherein a multiturn winding is wound on said core to form said inductor means.

8. The circuit of claim 7 wherein said auxiliary core is mounted such that flux generated by said control winding will pass through said single core in a direction transverse to the direction of flux generated therein by said multiturn winding.

10. The circuit of claim 5 wherein said core means comprises a plurality of cores and wherein a multiturn winding has a portion thereof wound on each of said cores to form said inductor means.

11. The circuit of claim 10 wherein control winding means are wound on said cores whereby the inductance and volt-second capacity of said inductor means may be varied.

15. The circuit of claim 5 wherein rectifying means are coupled to said output.

16. A method for producing an output voltage across a saturable core inductor connected in a series circuit with a capacitor, comprising: a. charging said capacitor to establish a first voltage across it; b. applying a voltage to said series circuit having a magnitude, duration and polarity sufficient, when added to said first voltage, to develop a volt-second value across said inductor greater than the volt-second capacity of said inductor to cause said inductor to saturate; c. discharging said capacitor through said inductor; d. recharging said capacitor in the opposite polarity until said inductor comes out of saturation; e. reversing the polarity of said applied voltage to again cause saturation of said inductor; and f. periodically repeating steps (c) through (e).