Ripple cancellation circuit for ultra-low-noise power supplies

This invention relates to direct-voltage power supplies, and more particularly to low-noise or low-ripple power supplies.

Much of the advance in standard of living over the past twenty or so years results from the use of advanced communications, data processing, and environmental sensing techniques. The devices used in such communications, processing, and sensing generally become more useful as their sizes are decreased, such that more of them can be used. For example, computers and cellular phones require ever-smaller elements, and become more capable as the number of devices which can be accommodated increases. Similarly, lightweight and reliable sensors can be used in large numbers in vehicles to aid in control and, in the case of spacecraft and military vehicles, to aid in carrying out their missions.

Most modern semiconductor devices, and other devices important for the above purposes, are generally energized or biased by direct voltages. As devices have become smaller, their powering requirements also advantageously decrease. Unfortunately, a concomitant of low power requirements is often sensitivity to unintended noise or fluctuations in the applied power. It is easy to understand that extremely small transistors, which ordinarily operate at two or three volts, could be destroyed by application of tens of volts. It is less apparent but true that small-percentage variations or noise on the applied powering voltage may result in degradation of the operating characteristics of semiconductor and other devices and the circuits in which they operate, which may adversely affect the performance. It is a commonplace that conventional radio and television receivers will respond to noise on or sudden changes in their supply voltages with aural or visual distortions, or both.

In general, electronic equipments require direct voltages for their power sources. There are two general sources of electrical energy which can be used to provide the power, and these two sources are batteries, which provide direct voltage, and power mains of an alternating voltage. When power mains are the source of electrical energy, it is common to rectify the alternating voltage to achieve a direct voltage. The power mains are used to drive machine motors in addition to electronic equipment, so the mains voltages tend to be higher than the voltages required for electronic equipment, and rectified voltages also tend to be higher than desired or usable. In the past, transformers have been used to convert the mains power to voltages more compatible with electronic equipment. However, transformers operating at 60 Hz tend to be much larger than is desirable in modern miniaturized equipment. It might be thought that there are no problems with the powering of electronic equipment from batteries, which directly provide direct voltage. However, batteries have the same general problem as that of mains powering, namely that available direct voltage does not necessarily correspond with the desired operating voltage. One modern technique for producing voltages for powering electronic equipment is that of use of a switching power supply or switching converter, which changes a direct source voltage to a different direct voltage.

A switching power converter can operate from a direct voltage derived from the power mains or from a battery, and can either increase or decrease the output voltage relative to the input voltage. These switching power converters take many different forms, some examples of which include those described in U.S. Pat. Nos. 4,163,926 issued Aug. 7, 1979 in the name of Willis; U.S. Pat. No. 4,190,791, issued Feb. 26, 1980 in the name of Hicks; U.S. Pat. No. 4,298,892 issued Nov. 3, 1981 in the name of Scott; U.S. Pat. No. 4,761,722 issued Aug. 2, 1988 in the name of Pruitt; and U.S. Pat. No. 5,602,464 issued Feb. 11, 1997 in the name of Linkowski et al.

A power supply according to an aspect of the invention powers a load. A storage capacitor is coupled across the load. A first inductance arrangement is coupled to the storage capacitor, which is coupled across the load, to thereby form a combined circuit. A source of voltage produces a direct voltage component and a time-varying voltage component. The source of voltage is coupled to the combined circuit for producing a flow of current therethrough, which flow of current results in division of the direct voltage component and the time-varying voltage component between at least the first inductance arrangement and the storage capacitor coupled across the load, whereby that portion of the time-varying voltage component appearing across the first inductance arrangement tends to cause a time-varying current flow through the first inductance arrangement. A magnetically coupled inductive arrangement is responsive to the time-varying voltage component appearing across the inductance arrangement, for generating a second time-varying current component in response to the time-varying voltage. The second time-varying current component is similar to the time-varying current flow through the first inductance arrangement. A combining arrangement is coupled to the combined circuit and to the magnetically coupled inductive arrangement, for combining the second time-varying current component with at least the time-varying current flow in such a manner as to tend to oppose the time-varying current flow.

In one embodiment, the source of voltage includes a switch which recurrently applies a raw direct voltage to the combined circuit, and applies a reference potential across the combined circuit during those intervals in which the raw direct voltage is not applied, whereby the time-varying component is a rectangular wave.

In another embodiment, of the power supply, the source of voltage comprises a phase-shifted full-wave switched bridge circuit including first and second tap points across which an alternating voltage is generated, and a transformer including a primary winding connected to the first and second tap points. The transformer also includes a secondary winding across which a varying voltage is generated in response to the alternating voltage. The source of voltage also includes a rectifying arrangement coupled to the secondary winding for converting the varying voltage into a varying or pulsating direct voltage.

In one version of a power supply according to an aspect of the invention, the magnetically coupled inductive arrangement comprises an inductive winding magnetically coupled to the first inductive arrangement, whereby the second time-varying current component is directly generated. In another version of a power supply according to this aspect of the invention, the magnetically coupled inductive arrangement comprises a transformer including a primary winding coupled across the first inductance arrangement, and also including a secondary winding across which a secondary voltage is generated in response to the time-varying voltage component appearing across the first inductance arrangement. An inductor or other inductance means is coupled in series with the secondary winding of the transformer, for producing the second time-varying current component in response to the secondary voltage.

A power supply according to an aspect of the invention, in which the first inductance means and the magnetically coupled inductive means responsive to the time-varying voltage component appearing across the inductance means, for generating a second time-varying current component in response thereto, comprises a unitary arrangement, and the unitary arrangement comprises a magnetic core with first and second spaced-apart magnetic paths through which magnetic flux flows. The first inductance means includes a conductor winding about the first magnetic path, and the magnetically coupled inductive means comprising a conductor winding about the second magnetic path. In a first variant of this arrangement, the magnetic core is in the form of two half-cores, each having a cross-sectional shape in the general form of the letter “U,” spaced apart by a pair of gaps located at the distal ends of the legs, and the first magnetic path comprises one leg of each of the halves together with one of the gaps, and the second magnetic path comprises another leg of each of the halves together with another of the gaps. In a second variant of this arrangement, the magnetic core is in the form of one of an E or pot core in two halves having legs, where each half has a cross-section in the general shape of the letter “E,” which halves fit together with a gap between the center legs of the halves. In this second variant, the first magnetic path includes the center leg of one of the halves of the core, and the second magnetic path includes the center leg of the other one of the halves of the core. In a third variant, the magnetic core is in the form of an E core in two halves, each of which halves has a cross-section defining three legs and a base in the general shape of the letter “E,” which halves fit together with a first gap between the center legs of the halves and a second gap between one pair of outer legs. In this third variant, the first magnetic path includes the one pair of outer legs of the halves of the core and the second gap, and the second magnetic path includes the other of the outer legs of the halves of the core and no gap.

In yet another hypostasis of the invention, the combining arrangement comprises a direct-voltage blocking capacitor. This blocking capacitor may be placed in series with the inductive winding of the one embodiment or in series with the secondary winding and inductor of the other embodiment.

FIG. 1 is a simplified schematic diagram of a switching buck voltage regulator with current ripple cancellation according to an aspect of the invention;

FIGS. 2a , 2 b, and 2 c are amplitude-time plots of voltages and currents which occur in the regulator of FIG. 1 during operation;

FIG. 3 is a simplified schematic diagram of an alternate embodiment of a regulator according to an aspect of the invention;

FIG. 4 is a semipictorial representation of the arrangement of transformer T 1 and inductor L 2 used in the arrangement of FIG. 1 ;

FIG. 5 illustrates one possible arrangement of loosely coupled inductors of FIG. 3 ;

FIG. 6 is a semipictorial representation of an E core or pot core arranged to produce an inductive arrangement for use in FIG. 3 ;

FIG. 7 is an arrangement similar to that of FIG. 5 , except in that an additional flux path with an air gap is provided through the center of the core; and

FIG. 8 is a simplified schematic diagram illustrating another aspect of the invention.

In FIG. 1 , an unregulated or “raw” direct voltage Vin is applied from a source (not illustrated) to regulator or power supply 10 input terminals 12 1 , and 12 2 . A controllable switch illustrated as a field-effect transistor (FET) Q 1 is controlled, by means which are not illustrated but which are well known in the art, to switch in a recurrent manner. The switching may be periodic or aperiodic, but the effect is to recurrently apply the Vin voltage “across” terminals 14 1 , and 14 2 , as illustrated by plot v 1 (t) of FIG. 2a in the intervals t 0 to t 1 , t 0 ′ to t 1 ′, and t 0 ″ to t 1 ″. Those skilled in the art will understand that the words “across” and “between” as used in electrical contexts have no particular physical meaning as might be ascribed in a mechanical or common context.

As illustrated in FIG. 1 , power supply 10 includes an inductor or inductive arrangement 16 connected in “series” with an output filter capacitor Cout, and the resulting series combination or combined circuit is connected across terminals 14 1 , and 14 2 for receiving the varying or pulsatory voltage v 1 (t). Under the impetus of each voltage pulse in the intervals t 0 to t 1 , t 0 ′ to t 1 ′, and t 0 ″ to t 1 ″ of FIG. 2a , electrical current through inductor L 1 increases, as illustrated in the relevant intervals by plot (I L1 +I N1 ) in FIG. 2 b. In this context, I L1 , represents the magnetizing or inductive current component flowing in inductor L 1 . The increasing current flow through the inductor L 1 in the intervals t 0 to t 1 , t 0 ′ to t 1 ′, and t 0 ″ to t 1 ″ of FIG. 2a flows as current I 0 through output filter capacitor Cout. Since output capacitor Cout is relatively large, its ac voltage is small and most time varying currents flow therethrough. As known to those skilled in the art, the flow of increasing current results, in general, in an increasing output voltage Vout across output filter capacitor Cout, although the current drawn by the load, represented by resistor R L in FIG. 1 , may under some conditions exceed the inductor current, thereby resulting in a net reduction of Vout. The voltage across output filter capacitor Cout is the voltage available to supply the load represented by resistor R L .

There are many ways to view the effects of the pulsating or varying supply voltage v 1 (t) applied across the series combination of inductor L 1 and output filter capacitor Cout. The applied voltage v 1 (t) may be viewed as consisting of a direct voltage component with a pulsatory voltage component superposed thereon. The inductor and capacitor may be viewed as a voltage divider, in which case the direct voltage component of v 1 (t) may be viewed as being developed solely across the output filter capacitor, as in steady-state operation the inductor L 1 cannot develop or withstand a direct voltage. In this voltage divider view, the alternating component of the applied voltage v 1 (t) may be viewed as appearing across the inductance of inductor L 1 , assuming that output filter capacitor Cout has zero impedance. However, filter capacitors do not have zero impedance, so some portion of the applied pulsatory or varying component of the applied voltage v 1 (t) will appear across output filter capacitor Cout. This portion of the pulsatory voltage is then an undesired ripple which is manifest across the load R L . In an alternative view, that portion of the pulsatory or varying applied voltage v 1 (t) which is applied to or across inductor L 1 results in a varying current flow in the inductor, which current also flows mostly through the internal impedance of output filter capacitor Cout, and thereby generates an undesired ripple voltage which appears across the load R L .

However the mechanism which generates the ripple across the output filter capacitor is viewed, the ripple is undesirable. According to an aspect of the invention, an additional current is generated, which ideally is equal in magnitude and opposite in phase to the alternating component of the current through the inductor L 1 , and this additional current is supplied to output filter capacitor Cout together with the inductor L 1 current, in a phase or polarity which cancels or offsets the alternating component of current. In effect, the output filter capacitor “sees” only a direct current flow because the time-varying currents in inductor L 1 , winding Ni and auxiliary inductor L 2 add to zero. Since no alternating current component flows through the internal impedance of output filter capacitor Cout, no ripple voltage can be generated across the capacitor. Of course, nothing is perfect, so there will necessarily always be some difference between the compensating ripple current and the ripple current actually flowing in the inductor L 1 and output filter capacitor Cout which will prevent total cancellation, but significant ripple current reduction should result.

In FIG. 1 , a diode D 1 has its cathode connected to terminal 14 1 , and its anode connected to terminal 14 2 . Those skilled in the art recognize this as a “freewheeling” diode, which is maintained in a nonconductive condition during those intervals in which the raw supply voltage is coupled through switching transistor Q 1 , corresponding to intervals t 0 to t 1 , t 0 ′ to t 1 ′, and t 0 ″ to t 1 ″ of FIG. 2 a. During those intervals when switching transistor Q 1 is nonconductive, the energy stored in inductor L 1 tends to cause current to continue to flow in the path including Cout and D 1 , with the result that D 1 becomes forward-biased and allows the inductive current to continue flowing in the intervals t 1 to t 0 ′, t 1 ′ to t 0 ″, and after t 1 ″. When diode D 1 is conductive, its voltage drop is small, and may be viewed as being zero for purposes of this analysis. Since the energy stored in inductor L 1 is the motive force for the current IL 1 , the current during intervals t 1 to t 0 ′, t 1 ′ to t 0 ″, and after t 1 ″, the magnitude of the current decreases, as illustrated in FIG. 2 b. Thus, the current flow through inductor L 1 includes a varying component which increases during those intervals in which voltage is applied by v 1 being positive, and which decreases during those intervals in which diode D 1 conducts and a voltage of opposite polarity is applied to inductor L 1 by output capacitor Cout.

In FIG. 1 , a transformer T 1 includes a primary winding designated N 1 and a secondary winding designated N 2 , poled as indicated by the standard dot notation. The primary winding N 1 is connected across inductor L 1 , so that transformer T 1 is energized by that varying component of the applied voltage appearing across inductor L 1 , which in most cases will be the principal portion of the varying component of the applied voltage. The varying component of voltage applied to primary winding N 1 of transformer T 1 transforms to the secondary N 2 side of the transformer. The voltage applied to primary winding N 1 of Transformer T 1 may be viewed as being similar to the pulsatory or varying component of the voltage applied to terminals 14 1 , and 14 2 , so the voltage across secondary winding N 2 may be viewed as a surrogate for the varying component of the applied voltage v 1 , except for that minor portion appearing across output filter capacitor Cout. The dotted end of secondary winding N 2 is connected to terminal 14 2 . The voltage appearing across the secondary winding N 2 , which is a surrogate for the applied varying voltage component, is applied to a second inductor for inductance arrangement L 2 , which generates a current which is a surrogate for the varying component of current through inductor L 1 . Those skilled in the art will know how to select the parameters of transformer T 1 and inductor L 2 so as to cause the surrogate varying current to substantially equal the varying current component in inductor L 2 plus the current in the primary of transformer T 1 .

A solution for selecting L 2 when N 2 and N 1 are given is

L 2 = L 1 ⁡ ( N ⁢   ⁢ 2 N ⁢   ⁢ 1 ) ⁢ ( 1 - N ⁢   ⁢ 2 N ⁢   ⁢ 1 ) 1

where L 1 , L 2 , N 1 , and N 2 all have real, positive values.

The three currents are combined by coupling the “output” ends of inductors L 1 and L 2 together with transformer primary winding N 1 at a junction point 18 corresponding to the juncture of “serially” connected inductor L 1 and output filter capacitor Cout. In order to avoid the application of direct voltage from junction point 18 to the serial combination of inductor L 2 and secondary winding N 2 , which might result in the flow of excess current to ground, a direct voltage blocking capacitor Cb is placed in the serial connection. As illustrated, blocking capacitor Cb is placed between inductor L 2 and tap point 18 , but Cb could also be placed between N 2 and L 2 , or alternatively between N 2 and ground or connection 14 2 .

In operation of the arrangement of FIG. 1 , the switching of Q 1 produces a pulsatory or varying voltage v 1 (t) as described in conjunction with FIG. 2a , with the result that a total current (I L1 +I N1 ) flows as illustrated in FIG. 2b , with the I L1 component of current flowing through inductor L 1 , and with the IN 1 component flowing through the primary winding N 1 of transformer T 1 . The flow of primary current iN 1 , of FIG. 2c in transformer T 1 results in a flow of varying current i L2 through secondary winding N 2 and through inductor L 2 . Comparing current (I L1 +I N1 ) of FIG. 2b with current i L2 of FIG. 2c shows that they are about equal in magnitude and of opposite phase or polarity, so that the result of their addition at tap point 18 is cancellation of the time-varying component of current. With no varying component of current flowing through output filter capacitor Cout, no ripple voltage is generated thereacross which can appear across the load being energized.

FIG. 3 is a simplified schematic diagram of an alternate embodiment of an aspect of the invention. Elements of FIG. 3 corresponding to those of FIG. 1 are designated by like reference alphanumerics. Generally, the arrangement of FIG. 3 substitutes loosely coupled windings for first inductor L 1 , transformer T 1 , and second inductor L 2 . In the arrangement of L 1 of FIG. 3 , N 1 represents an inductive winding having an inductance equivalent to the inductance of winding L 1 of FIG. 1 . Winding N 2 of FIG. 3 is magnetically coupled to winding N 1 , to thereby produce a resulting voltage in winding N 2 . However, winding N 2 of FIG. 3 is also inductive, at least in part by virtue of its loose coupling to winding N 1 , and therefore also inherently includes the inductive property which is provided in the arrangement of FIG. 1 by separate inductor L 2 . Thus, the arrangement of FIG. 3 operates essentially identically to the arrangement of FIG. 1 .

FIG. 4 is a semipictorial representation of the arrangement of transformer T 1 and inductor L 2 used in the arrangement of FIG. 1 . In FIG. 4 , the core is represented by two C sections or halves 410 a, 410 b defining a gap 412 between legs 410 a 1 and 410 b 1 . Winding N 1 is wound onto one leg of the core, and winding N 2 is wound over winding N 1 , thereby providing substantial magnetic coupling. Inductor L 2 is illustrated as a separate winding on a toroidal magnetic core. Capacitor Cb is also shown. By contrast, FIG. 5 illustrates the arrangement of loosely coupled inductors of FIG. 3 . In FIG. 5 , the core 501 is illustrated as two halves 410 a and 410 b defining a gap 412 1 between legs 410 a 1 and 410 b 1 and a corresponding gap 412 2 between legs 410 a 2 and 410 b 2 . Winding N 1 , corresponding to the main inductor L 1 , is illustrated as being wound on the left leg 410 a 2 , 410 b 2 of the core, and winding N 2 is illustrated as being wound on the right leg 410 a 1 , 410 b 1 of the core. The magnetic coupling between windings N 1 and N 2 is reduced relative to that of the arrangement of FIG. 4 , and the uncoupled inductance of each winding is greater. As illustrated in FIG. 5 , capacitor Cb is connected directly to winding N 2 .

FIG. 6 is a semipictorial representation of the use of an E core or a pot core (seen in cross-section) designated 601 to produce an inductive arrangement for use in the arrangement of FIG. 3 . In FIG. 6 , the coupling between windings N 1 and N 2 is reduced relative to what it might otherwise be by the spatial separation of the windings. The core 601 is in the form of two halves 601 a, 601 b, each of which has the general shape of the letter “E,” with upper half 601 a having outer legs 601 a 1 and 601 a 2 , and a center leg 610 a, and with lower half 601 b having outer legs 601 b 1 and 601 b 2 and a center leg 610 b. The gap 612 between center legs 610 a and 610 b in the central portion of the core is set to give the correct value of inductance L 1 . FIG. 7 is an arrangement 700 generally similar to that of FIG. 5 , except in that an additional flux path 710 a, 710 b with an air gap 712 is provided through the center of the core 701 . Winding N 2 is wound on legs 701 a 2 , 701 b 2 . The additional flux path 710 a, 710 b, 712 can be used to affect or decrease the coupling between windings N 1 and N 2 in a manner controlled by the dimension of the air gap, thus increasing the effective value of the equivalent L 2 . Such a magnetic shunt insures that, for most applications, the correct value of L 1 can be obtained by controlling the air gap 714 on the left leg 701 a 1 , 701 b 1 while the correct value of L 2 can be obtained by shunting coupling flux through the center leg under control of its air gap, while still maintaining the correct turns ratio L 1 /L 2 .

Those skilled in the art will recognize that the arrangements of FIGS. 5 , 6 , and 7 provide for loosely coupled windings which will exhibit more uncoupled inductance than the N 1 /N 2 windings of FIG. 4 . Consequently, the arrangements of FIGS. 5 , 6 , and 7 can provide performance equivalent to that of FIG. 4 .

FIG. 8 is a simplified schematic diagram illustrating another aspect of the invention. In the arrangement of FIG. 8 , the voltage applied to the inductor-capacitor “series” circuit does not come directly from a controllable switch as in FIGS. 1 and 3 , but rather comes by way of a rectifier arrangement. In FIG. 8 , 810 represents a full-wave bridge circuit including plural controllable switches. As known to those skilled in the art, these switches can be operated in a number of modes. For definiteness, the switches of FIG. 8 are operated by a controller (not illustrated) in a phase-shifted mode, in which the switches are rendered conductive in a manner such as to minimize the voltages across the switches during at least one of turn-on and turn-off. The result of these operations is to produce an alternating voltage across a primary winding N 1 of a transformer 812 . The alternating voltage applied to primary winding N 1 of transformer 812 causes an alternating voltage to be generated across the secondary winding, illustrated as separate windings N 2 a and N 2 b , with a tap point 814 therebetween. A pair of diodes or rectifiers R 1 and R 2 are illustrated in FIG. 8 , with their anodes connected to the ends of secondary windings N 2 a and N 2 b , respectively, which are remote from tap 814 . The cathodes of rectifiers R 1 and R 2 are connected together and to an inductive winding L 1 . Inductive winding L 1 is connected in “series” with an output filter capacitor Cout, as in FIG. 3 . An inductive winding L 2 is loosely coupled to winding L 1 as described in conjunction with FIG. 3 , and is connected to reference tap 814 and by way of a blocking capacitor Cb to a junction point 818 . With the described arrangement, a voltage having both direct and varying components appears between reference tap 814 and input terminal 814 1 . The alternating voltage is manifest across the series combination of L 1 and Cout, as described in conjunction with FIGS. 1 and 3 , and the arrangement of winding L 2 coupled to point 818 tends to cancel the alternating or varying current components in inductor L 2 . This, in turn, reduces the magnitude of the alternating current components flowing in capacitor Cout, with consequent reduction in the voltage ripple or noise appearing at the load terminals 20 1 , and 20 2 .

It should be emphasized that the arrangement for cancellation of alternating current components may be used in the case in which an alternating sine wave is rectified to produce “pulsating direct voltage,” corresponding to a sequence of unidirectional half-sine-waves. In general, any alternating voltage waveshape that generates an ac current in inductor L 1 can be cancelled using the invention.

Thus, speaking very generally, a low-ripple power supply includes a storage capacitor coupled across load terminals, and an inductor connected to a source of voltage including a varying or pulsatory component and a direct component, for causing a flow of current to said capacitor through the inductor. The varying component of the inductor current flowing in the capacitor results in ripple across the load. A winding is coupled to the inductor for generating a surrogate of the varying inductor current. The surrogate current is added to the inductor current to cancel or reduce the magnitude of the varying current component. This cancellation effectively reduces the varying current component flowing in the storage capacitor, which in turn reduces the ripple appearing across the load terminals.

More particularly, a power supply ( 10 ) according to an aspect of the invention is capable of powering a load (R L ) coupled to load terminals ( 20 1 , 20 2 ). A storage capacitor (Cout) is coupled across the load (R L ) terminals ( 20 1 , 20 2 ). A first inductance arrangement (L 1 ) is coupled to the storage capacitor (Cout), which is coupled across the load (R L ) terminals ( 20 1 , 20 2 ), to thereby form a combined circuit (L 1 , Cout). A source of voltage (Vin, Q 1 , D 1 ) produces a direct voltage component and a time-varying voltage component. The source of voltage (Vin, Q 1 , D 1 ) is coupled to the combined circuit (L 1 , Cout) for producing a flow of current therethrough, which flow of current results in division of the direct voltage component and the time-varying voltage component between at least the first inductance arrangement (L 1 ) and the storage capacitor (Cout) coupled across the load (R L ) terminals ( 20 1 , 20 2 ), whereby that portion of the time-varying voltage component appearing across the first inductance arrangement (L 1 ) tends to cause a time-varying current (i L1 ) flow through the first inductance arrangement (L 1 ). A magnetically coupled inductive arrangement (T 1 , L 2 ; 310 ) is responsive to the time-varying voltage component appearing across the inductance arrangement (L 1 ), for generating a second time-varying current component (i L2 ) in response to the time-varying voltage. The second time-varying current component (i L2 ) is similar to the time-varying current flow (i L1 ) through the first inductance arrangement (L 1 ). A third time-varying current component (i N1 ) proportional to i L2 flows in the primary of the transformer. A combining arrangement (Cb, 18 ; Cb, 818 ) is coupled to the combined circuit (L 1 , Cout) and to the magnetically coupled inductive arrangement (T 1 , L 2 ; 310 ), for combining the second time-varying current component (i L2 ) with at least the time-varying current flow (i L1 ) in such a manner as to tend to oppose the time-varying current flow. This may be viewed as a combining of the second time-varying current component (i L2 ) and the third time-varying current (i N1 ) with the time-varying current flow (i L1 ) in such a manner as to tend to oppose the time-varying current flow.

In one embodiment, the source of voltage (Vin, Q 1 , D 1 , 810 , 812 , R 1 , R 2 ) includes a switch (Q 1 ; 810 , 812 , R 1 , R 2 ) which recurrently applies a raw direct voltage to the combined circuit (L 1 , Cout), and applies a reference potential (diode drop, for example) across the combined circuit (L 1 , Cout) during those intervals in which the raw direct voltage is not applied, whereby the time-varying component is a rectangular wave.

In another embodiment, of the power supply ( 10 ), the source of voltage (Vin, Q 1 , D 1 , 810 , 812 , R 1 , R 2 ) comprises a phase-shifted full-wave switched bridge circuit ( 810 ) including first ( 811 1 ) and second ( 811 2 ) tap points across which an alternating voltage is generated, and a transformer ( 812 ) including a primary winding (N 1 ) connected to the first ( 811 1 ) and second ( 811 2 ) tap points. The transformer ( 812 ) also includes a secondary winding (N 2 a , N 2 b ) across which a varying voltage is generated in response to the alternating voltage. The source of voltage (Vin, Q 1 , D 1 , 810 , 812 , R 1 , R 2 ) also includes a rectifying arrangement (R 1 , R 2 ) coupled to the secondary winding (N 2 a , N 2 b ) for converting the varying voltage into a varying or pulsating direct voltage.

In one version of a power supply ( 10 ) according to an aspect of the invention, the magnetically coupled inductive arrangement (T 1 , L 2 ; 310 ) comprises an inductive winding (L 2 ) magnetically coupled to the first inductive arrangement (L 1 ), whereby the second time-varying current component is directly generated. In another version of a power supply ( 10 ) according to this aspect of the invention, the magnetically coupled inductive arrangement comprises a transformer (T 1 ) including a primary winding (N 1 ) coupled across the first inductance arrangement (L 1 ), and also including a secondary winding (N 2 ) across which a secondary voltage is generated in response to the time-varying voltage component appearing across the first inductance arrangement (L 1 ). An inductor (L 2 ) or other inductance means is coupled in series with the secondary winding (N 2 ) of the transformer (T 1 ), for producing the second time-varying current component in response to the secondary voltage.

A power supply according to an aspect of the invention, in which (a) the first inductance means and (b) the magnetically coupled inductive means responsive to the time-varying voltage component appearing across the inductance means, for generating a second time-varying current component in response thereto, comprises a unitary magnetic arrangement ( 500 , 600 , 700 ). This unitary magnetic arrangement ( 500 , 600 , 700 ) comprises a magnetic core ( 501 , 601 , 701 ) with first and second spaced-apart magnetic paths through which magnetic flux flows. The first inductance means includes a conductor winding about the first magnetic path, and the magnetically coupled inductive means comprising a conductor winding about the second magnetic path. In a first variant of this arrangement, the magnetic core ( 500 ) is in the form of two half-cores ( 410 a, 410 b), each having a cross-sectional shape in the general form of the letter “U,” spaced apart by a pair of gaps ( 412 1 , 412 2 ) located at the distal ends of the legs, and the first magnetic path comprises one leg ( 410 a 2 , 410 b 2 ) of each of the halves ( 410 a, 410 b) together with one of the gaps ( 412 2 ), and the second magnetic path comprises another leg ( 410 a 1 , 410 b 1 ) of each of the halves ( 410 a, 410 b) together with another of the gaps ( 412 1 ). In a second variant of this arrangement, the magnetic core ( 600 ) is in the form of one of an E or pot core in two halves ( 601 a, 601 b) having legs ( 601 a 1 , 601 a 2 , 610 a, 601 b 1 , 601 b 2 , 610 b), where each half ( 601 a, 601 b) has a cross-section in the general shape of the letter “E,” which halves ( 601 a, 601 b) fit together with a gap ( 612 ) between the center legs ( 610 a, 610 b) of the halves ( 601 a, 601 b). In this second variant, the first magnetic path includes the center leg ( 610 a) of one of the halves ( 601 a) of the core ( 601 ), and the second magnetic path includes the center leg ( 610 b) of the other one ( 601 b) of the halves of the core ( 601 ). In a third variant, the magnetic core ( 701 ) is in the form of an E core in two halves ( 701 a, 701 b), each of which halves ( 701 a, 701 b) has a cross-section defining three legs ( 701 a 1 , 701 a 2 , 710 a, 701 b 1 , 701 b 2 , 710 b) and a base ( 701 ab, 701 bb) in the general shape of the letter “E,” which halves ( 701 a, 701 b) fit together with a first gap ( 712 ) between the center legs ( 710 a, 710 b) of the halves ( 701 a, 701 b) and a second gap ( 714 ) between one pair ( 701 a 1 , 701 b 1 ) of outer legs. In this third variant, the first magnetic path includes the one pair of outer legs ( 701 a 1 , 701 b 1 ) of the halves ( 701 a, 701 b) of the core and the second gap ( 714 ), and the second magnetic path includes the other ones ( 701 a 2 , 701 b 2 ) of the outer legs of the halves ( 701 a, 701 b) of the core ( 701 ) and no gap.

In yet another hypostasis of the invention, the combining arrangement comprises a direct-voltage blocking capacitor (Cb). This blocking capacitor (Cb) may be placed in series with the inductive winding (N 2 ) of the one embodiment or in series with the secondary winding (N 2 ) and inductor (L 2 ) of the other embodiment.