Zero-voltage-switching power supply

[0002] This invention relates to electric power supplies, and particularly to a switching power supply capable of a.c. to d.c. voltage conversion, featuring provisions for attainment of closer approximation of the input current waveform to a sinusoidal wave, and a higher power factor, than by the comparable prior art.

[0003] A conversion from an alternating to a direct current is possible by a rectifying and smoothing circuit comprising a rectifying circuit having a diode connected to an a.c. power supply, and a smoothing capacitor connected to the rectifying circuit. This type of rectifying and smoothing circuit possesses the disadvantage, however, of a somewhat poor power factor as a result of the fact that the smoothing capacitor is charged only at or adjacent the peaks of the a.c. voltage of sinusoidal waveform. Another drawback is that it is incapable of adjustably varying the d.c. output voltage.

[0004] Japanese Unexamined Patent Publication No. 8-154379 represents an improvement of the rectifying and smoothing circuit above. It teaches a switching power supply comprising a rectifying circuit, a smoothing capacitor, a d.c.-to-d.c. converter circuit, and an inductive reactor for a higher power factor. The reactor is electrically connected between the pair of output terminals of the rectifying circuit upon closure of a switch included in the d.c.-to-d.c. converter circuit. The desired improvements in power factor and in input current waveform are thus attained, as the current flowing through the reactor varies in amplitude in step with the a.c. input voltage.

[0005] This prior art switching power supply has proved to be not wholly satisfactory, however, particularly in terms of power loss. The advent of a more efficient power supply has been awaited.

[0006] The present invention seeks to reduce the power loss, and hence improve the efficiency, of the switching power supply of the kind discussed above, without impairment of its inherent advantages.

[0007] Briefly, the invention may be summarized as a switching power supply capable of translating a.c. voltage into d.c. voltage. Included are a rectifier circuit connected to a pair of a.c. input terminals for rectifying a.c. input voltage, a transformer having a primary winding, and a rectifying and smoothing circuit connected between the transformer and a pair of d.c. output terminals for providing d.c. output voltage. The rectifier circuit has at least two output conductors. A smoothing capacitor is connected between one of these output conductors of the rectifier circuit and a first extremity of the primary winding of the transformer. A primary switch is connected between a second extremity of the primary winding and said one output conductor of the rectifier circuit. Also included are a first ancillary winding (e.g. tertiary winding of the transformer) electromagnetically coupled to the primary winding and having a first extremity connected to the first extremity of the primary winding and the smoothing capacitor, a second ancillary winding (e.g. quaternary winding) electromagnetically coupled to the primary winding and to the first ancillary winding and having a first extremity connected to a second extremity of the first ancillary winding, a first inductance coil connected between the other output conductor of the rectifier circuit and a second extremity of the first ancillary winding, and a second inductance coil connected between said other output conductor of the rectifier circuit and the smoothing capacitor and electromagnetically coupled to the first inductance coil. Additionally, and perhaps most characteristically of the invention, there is provided an ancillary switch which is connected in parallel with a serial connection of the transformer primary and the primary switch via the first and the second ancillary winding. A switch control circuit is connected to the primary switch and the ancillary switch for on-off control thereof at a repetition frequency higher than the frequency of the a.c. input voltage. The switch control circuit is adapted to turn the ancillary switch on earlier than the primary switch is turned on, and to turn the ancillary switch off not later than the primary switch is turned off.

[0008] Current flows through the first and the second inductance coils as the primary switch is turned on and off at a prescribed switching frequency. The current varies in amplitude in proportion with that of the a.c. input voltage, realizing a higher power factor and a better waveform as in the noted prior art.

[0009] Each time the ancillary switch is turned on earlier than is the primary switch, current flows through the first and the second ancillary winding connected in series with the ancillary switch. Since these ancillary windings are in fact the tertiary and quaternary of the transformer, the current flow through these windings results in a discharge of soft-switching capacitance means conventionally connected in parallel with the primary switch, and hence in a drop in the voltage across the primary switch. Thus the primary switch is to be turned on at zero voltage; that is, the so-called “soft switching” is accomplished. The advantages accruing from such soft switching include less switching loss and less noise production.

[0010] The primary switch is utilized both for improvements in power factor and waveform and for d.c.-to-d.c. conversion. All these objectives are therefore realized in a device of simple, inexpensive and compact design.

[0011] It will also be appreciated that the two ancillary windings for the soft switching of the primary switch can be incorporated with the transformer. The ancillary windings are therefore not to add substantially to the size of the power supply.

[0012] According to a further feature of the invention, the ancillary switch has an ancillary diode connected in series therewith for blocking reverse current flow. The diodes constituting the rectifier circuit are thus blocked from high frequency current due to the switching of the primary switch and so prevented from noise production.

[0013] A still further feature of the invention resides in a bypass capacitor which is less in capacitance than the smoothing capacitor and which is connected between the two outputs of the rectifier circuit. Such a bypass capacitor will expedite the release of the energy that has been stored on the second ancillary winding, or transformer quaternary.

[0014] The two inductance coils are connected in parallel with each other in some preferred embodiments of the invention, and in series with each other in others. Experiment has proved that the power factor is satisfactory either way.

[0015] The above and other objects, features and advantages of this invention will become more apparent, and the invention itself will best be understood, from a study of the following description and appended claims, with reference had to the attached drawings showing the preferred embodiments of the invention.

[0016] FIG. 1 is a schematic electrical diagram of a first preferred form of switching power supply according to the invention;

[0017] FIG. 2 is a schematic electrical diagram, partly in block form, showing in more detail the switch control circuit included in the FIG. 1 switching power supply;

[0018] FIG. 3, consisting of (A) through (H), is a series of diagrams showing the waveforms appearing at various parts of the FIG. 2 switch control circuit in order to explain how the switch control signals for the primary and the ancillary switch of the FIG. 1 device are created according to the invention;

[0019] FIG. 4, consisting of (A) through (M), is a series of diagrams showing the voltage and current waveforms appearing at various parts of the FIG. 1 device in order to explain how the primary switch is soft-switched according to the invention;

[0020] FIG. 5, consisting of (A) through (D), is a series of diagrams showing the voltage and current waveforms appearing at various parts of the FIG. 1 device in order to explain how improvements in power factor and waveform are attained;

[0021] FIG. 6 is a view similar to FIG. 1 but showing a second preferred form of switching power supply according to the invention;

[0022] FIG. 7 is also a view similar to FIG. 1 but showing a third preferred form of switching power supply according to the invention;

[0023] FIG. 8, consisting of (A) through (M), is a series of diagrams showing the voltage and current waveforms appearing at various parts of the FIG. 7 device in order to explain how the primary switch is soft-switched according to the invention; and

[0024] FIG. 9 is a view similar to FIG. 1 but showing a fourth preferred form of switching power supply according to the invention.

[0025] The switching power supply shown in FIG. 1 by way of a preferable embodiment of the invention has a pair of input terminals 1 and 2 which are to be connected to a source, not shown, of commercial alternating current with a frequency of 50 Hz or any other local value. Connected directly to this pair of input terminals 1 and 2 is a noise filter 3 of conventional make comprising inductors and capacitors for removal of the high-frequency noise component from the incoming fixed-frequency alternating current.

[0026] The noise filter 3 is connected on the other hand to a rectifier circuit 4 by way of a pair of output conductors 41 and 42. The rectifier circuit 4 has six diodes D₁, D₂, D₃, D₄, D₅and D₆in combination with three output conductors 43, 44 and 45. The first diode D₁has its anode connected to the first a.c. input conductor 41, and its cathode to the first rectifier output conductor 43, the second diode D₂has its cathode connected to the first a.c. input conductor 41, and its anode to the third rectifier output conductor 45. The third diode D₃has its anode connected to the second a.c. input conductor 42, and its cathode to the first rectifier output conductor 43. The fourth diode D₄has its cathode connected to the second a.c. input conductor 42, and its anode to the third rectifier output conductor 45. The fifth diode D₅has its anode connected to the first a.c. input conductor 41, and its cathode to the second rectifier output conductor 44. The sixth diode D₆has its anode connected to the second a.c. input conductor 42, and its cathode to the second rectifier output conductor 44.

[0027] It will be observed that the third output conductor 45 of the rectifier 4 is grounded. This rectifier output conductor 45 may therefore be sometimes referred to simply as the grounded conductor hereinafter in this specification.

[0028] At 5 in FIG. 1 is shown a transformer having a primary winding N₁, a secondary winding N₂, a tertiary winding N₃and a quaternary winding N₄, all wound around a magnetic core 9 and electromagnetically coupled together. The transformer primary N₁and secondary N₂are opposite, but the transformer primary N₁and tertiary N₃and quaternary N₄are the same, in polarity, as indicated by the dots in FIG. 1. The transformer primary N₁and tertiary N₃have their extremities of the same polarity connected to the first rectifier output conductor 43 via an inductance coil L₂. The junction 10 between the transformer tertiary N₃and quaternary N₄is connected to the second rectifier output conductor 44 via another inductance coil L₁. It is also understood that the transformer tertiary N₃is less in turns than the transformer primary N₁, and that the transformer quaternary N₄is not more in turns than the transformer tertiary N₃.

[0029] A smoothing capacitor C₁, preferably an electrolytic capacitor, has one of its opposite polarity terminals connected to the first rectifier output conductor 43 via the inductance coil L₂. The other terminal of the smoothing capacitor C₁is connected to the grounded third rectifier output conductor 45. It will also be seen that the smoothing capacitor C₁is parallel to the serial connection of the transformer primary N₁and a primary switch Q₁.

[0030] Shown as an insulated-gate field-effect transistor, the primary switch Q₁is connected between the second 44 and third 45 rectifier output conductors via the transformer primary N₁, transformer tertiary N₃and first inductance coil L₁, besides being connected in parallel with the smoothing capacitor C₁via the transformer primary N₁. A diode D_q1is connected reversely in parallel with the primary switch Q₁for its protection. This switch protection diode D_q1, however, could be replaced by the so-called body diode of the primary switch Q₁, or a diode built into it.

[0031] A capacitor C_q1for the soft switching of the primary switch Q₁is connected in parallel therewith. This soft-switching capacitor C_q1should be less in capacitance than the smoothing capacitor C₁. The parasitic capacitance between the two electrodes, drain and source in this case, of the primary switch Q₁could serve the purposes of the soft-switching capacitor C_q1, though.

[0032] The transformer secondary N₂has its pair of opposite extremities connected respectively to the pair of output terminals 11 and 12 of this switching power supply via a rectifying and smoothing circuit 6. The rectifying and smoothing circuit 6 comprises a rectifying diode D₀and a smoothing capacitor C₀. Connected between one extremity of the transformer secondary N₂and the output terminal 11, the rectifying diode D₀is so oriented as to be conductive when the primary switch Q₁is off, and nonconductive when the primary switch Q₁is on. The smoothing capacitor C₀is connected in parallel with the transformer secondary N₂via the rectifying diode D₀. A unidirectional output voltage is thus obtained between the pair of output terminals 11 and 12 for feeding a load 18 connected thereto.

[0033] The reference numeral 7 in FIG. 1 generally denotes an ancillary soft-switching circuit or an ancillary resonance circuit comprising an ancillary switch Q₂shown as another FET, two ancillary diodes D_aand D_b, and a second switch protection diode D_q2, in addition to the aforesaid transformer tertiary N₃and quaternary N₄, which are inductive. Shown connected in parallel with the ancillary switch Q₂for its protection, the diode D_q2is not an essential component of the ancillary soft-switching circuit 7, as the so-called body diode of the FET switch Q₂could serve its purposes.

[0034] The transformer tertiary N₃and quaternary N₄have their polarities so determined as to enable the transformer primary N₁to develop a voltage to cause the soft-switching capacitor C_q1to discharge, in response to the current supplied to the transformer tertiary N₃and quaternary N₄from the smoothing capacitor C₁. There may be cases in which these inductive transformer windings N₃and N₄fail to provide all the inductances needed; in such cases, an additional inductor may be connected in series with the quaternary N₄.

[0035] It will be seen that the ancillary soft-switching circuit 7 includes a serial connection of the transformer tertiary N₃, quaternary N₄, first ancillary diode D_a, and ancillary switch Q₂. This serial connection is in parallel with the serial connection of the transformer primary N₁and primary switch Q₁, and with the smoothing capacitor C₁as well. More specifically, the transformer tertiary N₃has one of its extremities connected to one extremity of the transformer primary N₁, and the other extremity connected to the source of the primary switch Q₁via the transformer quaternary N₄, first ancillary diode D_aand ancillary switch Q₂. It is understood that the ancillary switch Q₂has parasitic capacitance between its drain and source. The first ancillary diode D_ais so oriented as to be forward biased by the voltage across the smoothing capacitor C₁. The second ancillary diode D_bis connected in parallel with the serial connection of the transformer quaternary N₄, first ancillary diode D_aand ancillary switch Q₂. The second ancillary diode D_band second switch protection diode D_q2are both so oriented as to be reverse biased by the voltage across the smoothing capacitor C₁.

[0036] The two inductance coils or inductors L₁and L₂are both intended for a higher power factor and waveform improvement. Wound upon a common magnetic core 17, the coils L₁and L₂are electromagnetically coupled together and polarized in the same direction. The first coil L₁has one extremely connected to the second rectifier output conductor 44, and the other extremity connected to the junction 10 between the transformer tertiary N₃and quaternary N₄. The second coil L₂has one extremely connected to the first rectifier output conductor 43, and the other extremity connected to the smoothing capacitor C₁.

[0037] As shown also in FIG. 1, a switch control circuit 8 has inputs connected to the pair of power supply output terminals 11 and 12 by way of conductors 13 and 14, respectively, an output connected to the control terminal of the primary switch Q₁by way of a conductor 15, and another output connected to the control terminal of the ancillary switch Q₂by way of a conductor 16. The indicia V_g1and V_g2on the output conductors 15 and 16 stand for the switch control signals applied to the switches Q₁and Q₂, respectively, from the switch control circuit 8. Actually, the switch control circuit 8 is electrically connected to the sources of the FET switches Q₁and Q₂, but such connections are not shown as they are considered apparent to the specialists.

[0038] FIG. 2 is a more detailed illustration of the switch control circuit 8. Included is a voltage detector circuit 21 which is connected to the pair of power supply output terminals 11 and 12, FIG. 1, via the input conductors 13 and 14 for providing an output voltage proportional to the power supply output voltage V_o. The output of the voltage detector circuit 21 is connected to one input of a differential amplifier 22, the other input of which is connected to a reference voltage source 23. Comparing the voltage detector output voltage and the reference voltage, the differential amplifier 22 delivers the difference V_r1therebetween both to one input of a first comparator 25 and, via a level setting circuit 26, to one input of a second comparator 27. A sawtooth generator circuit 18 is connected to the other inputs of both comparators 25 and 27. Notwithstanding the showing of FIG. 2, however, the differential amplifier 22 could be photoelectrically coupled to both comparator 25 and level setting circuit 26, and so could the voltage detector circuit 21 to the differential amplifier 22.

[0039] As indicated at (A) in FIG. 3, the sawtooth generator 24 generates a sawtooth voltage V_tat a frequency higher than that of the incoming a.c. voltage V_acbetween the a.c. input terminals 1 and 2, FIG. 1. Comparing this sawtooth voltage V_tand the output voltage V_r1, shown at (A) in FIG. 3, of the differential amplifier 22, the first comparator 25 puts out a rectangular pulse train V₂₅as at (B) in FIG. 3.

[0040] The level setting circuit 26 has two voltage-dividing resistors R₁and R₂for putting out a voltage V_r2which is less than the output voltage V_r1of the differential amplifier 22, as indicated also at (A) in FIG. 3. Comparing the sawtooth voltage V_tand the level setting circuit output voltage V_r2, the second comparator 27 produces an output V₂₇, (C) in FIG. 3, in the form of a train of pulses of greater duration than that of the output pulses V₂₅of the first comparator 25.

[0041] The FIG. 2 switch control circuit 8 further comprises a pulse leading edge detector circuit 28 and a pulse trailing edge detector circuit 29, which are both connected to the output of the first comparator 25, and another pulse leading edge detector circuit 30 connected to the output of the second comparator 27. The pulse leading edge detector circuit 28 puts out a short duration pulse P₁, as at (D) in FIG. 3, in response to the leading edge of each output pulse V₂₅, at (B) in FIG. 3, of the first comparator 25. The pulse trailing edge detector circuit 29 puts out a short duration pulse P₂, as at (E) in FIG. 3, in response to the trailing edge of each output pulse V₂₅of the first comparator 25. The other pulse leading edge detector circuit 30 puts out a similarly short duration pulse P₃, as at (F) in FIG. 3, in response to the leading edge of each output pulse V₂₇, at (C) in FIG. 3, of the second comparator 27. It will be noted from FIG. 5 that the first pulse leading edge detector circuit 28 produces a pulse P₁as at t₂whereas the second pulse leading edge detector circuit 30 produces a pulse P₃as at t₁, a moment shortly before t₂.

[0042] The first pulse leading edge detector circuit 28 has its output connected to the set input S of a first RS flip flop 31. The pulse trailing edge detector circuit 29 has its output connected both to the reset input R of the first flip flop 31 and to the reset input R of a second RS flip flop 32. The second pulse leading edge detector circuit 30 has its output connected to the set input S of the first flip flop 31. The first flip flop 31 has its noninverting output Q connected to the primary switch Q₁, FIG. 1, by way of the switch control circuit output conductor 15. The second flip flop 32 has its noninverting output Q connected to the ancillary switch Q₂by way of the other switch control circuit output conductor 16.

[0043] Thus the first flip flop 31 is to be set when triggered by each output pulse P₁of the first pulse leading edge detector circuit 28, and reset when triggered by each output pulse P₂of the pulse trailing edge detector circuit 29. At (G) in FIG. 3 are shown the resulting primary switch control pulses V_g1produced by the first flip flop 31, for application to the gate of the primary switch Q₁by the first flip flop 31. Each such pulse V_g1lasts as from t₂to t₅. The second flip flop 32 is to be set when triggered by each output pulse P₃of the second pulse leading edge detector circuit 30, and reset when triggered by each output pulse P₂of the pulse training edge detector circuit 29. FIG. 3 indicates at (H) the resulting ancillary switch control pulses V_g2delivered from the second flip flop 32 to the gate of the ancillary switch Q₂. Each ancillary switch control pulse V_g2lasts as from t₁to t₅.

[0044] A consideration of FIG. 4 will make it easier to understand hoist the two switches Q₁and Q₂are driven by the respective switch control signals V_g1and V_g2. One primary switch control pulse V_g1is shown at (A), and one ancillary switch control pulse V_g2at (B), in FIG. 4. The primary switch Q₁is off at t₁when the ancillary switch control pulse V_g2goes high to turn the ancillary switch Q₂on. The primary switch control pulse V_g1goes high at t₂, shortly after t₁, causing conduction through the primary switch Q₁. The time spacing between t₁and t₂must be well calculated to reduce switching loss to a minimum when the primary switch Q₁is turned on.

[0045] Both primary and ancillary switches Q₁and Q₂are shown to go off at the same moment t₅in FIGS. 3 and 4 in this particular embodiment of the invention. In practice, however, the ancillary switch Q₂may be turned off at any moment from t₄in FIG. 4, when the current I_q2through the ancillary switch Q₂becomes zero as at (F) in FIG. 4, and t₅when the primary switch Q₁goes off. The ancillary switch Q₂should therefore be turned off at t₅at the latest. In order to turn off the ancillary switch Q₂at t₄, for instance, as indicated by the broken lines at (B) in FIG. 4, a monostable multivibrator may be employed in lieu of the second flip flop 32, FIG. 2, of the switch control circuit 8, as indicated by the broken lines in the same figure and designated 33. The MMV 33 may be made to produce pulses each lasting as from t₁to t₄in FIG. 4, in response to the output pulses P₃, at (F) in FIG. 3, of the second pulse leading edge detector circuit 30. Such MMV output pulses may be impressed directly to the ancillary switch Q₂over the conductor 16.

[0046] In use of the FIG. 1 power supply the pair of a.c. input terminals 1 and 2 are to be connected to an unshown source of a.c. power, and the pair of d.c. output terminals 11 and 12 to the load 18. The smoothing capacitor C₁will be charged to the desired d.c. voltage V_c1as the primary switch Q₁is turned on and off by the switch control circuit 8. The resulting steady-state operation of this representative switching power supply will be discussed hereinbelow with reference to FIGS. 4 and 5 which show the voltage and current waveforms appearing in various parts of the FIGS. 1 and 2 circuitry.

[0047] FIG. 5 in particular is explanatory of how the instant invention attains improvements in power factor and input waveform. At (A) in this figure is shown the series of primary switch control pulses V_g1applied from the switch control circuit 8 to the primary switch Q₁for its on-off control. The primary switch Q₁is turned on and off during each cycle T of the switch control signal, the cycle T consisting of one pulse, as from t₂to t₃, and one space between such pulses, as from t₃to t₅. The repetition rate of these primary switch control pulses V_g1is now assumed to be 20 kHz. It is also understood that the 50 Hz a.c. voltage V_acis now applied between the pair of a.c. input terminals 1 and 2 as at (D) in FIG. 5.

[0048] As the primary switch Q₁is repeatedly turned on and off as above, and the ancillary switch Q₂concurrently turned on and off as at (B) in FIG. 4, the input current I_inof the rectifier circuit 4 will change in amplitude as at (B) in FIG. 5 in step with the amplitude of the a.c. input voltage V_acat (D) in FIG. 5. Thus the a.c. input current I_acshown at (C) in FIG. 5 will closely approximate a sinusoidal wave, thereby achieving improvements in input waveform and in power factor. Incidentally, as a result of the connection of the first inductance coil L₁to the junction 10 between the transformer tertiary N₃and quaternary N₄, the current I_L1does not flow through the first inductance coil L₁unless the potential at the second rectifier output conductor 44 grows higher than that at the junction 10 due to the voltage V_C1across the smoothing capacitor C₁. FIG. 5 shows the rectifier input current I_inand a.c. input current I_acas flowing from t₁to t₆and from t₈to t₉.

[0049] The operation of the FIG. 1 power supply exclusive of the ancillary soft-switching circuit 7 will be explained in some more detail. A current will flow through the path comprising the second rectifier output conductor 44, first inductance coil L₁, transformer tertiary N₃, transformer primary N₁, and primary switch Q₁during the conducting periods T_onof the primary switch Q₁, as from t₂to t₃in FIG. 5. A current will flow at the same time through the path comprising the smoothing capacitor C₁, transformer primary N₁, and primary switch Q₁. The voltage developing across the transformer secondary N₂during these primary switch conducting periods T_onis oriented to reverse bias the diode D₀, holding the same nonconductive. Energy is therefore stored on the transformer 5 during the primary switch conducting periods T_on, as well as on the first inductance coil L₁.

[0050] On the other hand, during the nonconducting periods T_offof the primary switch Q₁, as from t₃to t₅in FIG. 5, both first inductance coil L₁and transformer 5 will release the energy they have stored during each primary switch conducting period T_on, causing current flow through the path comprising the second rectifier output conductor 44, first inductance coil L₁, transformer tertiary N₃, and smoothing capacitor C₁. This capacitor C₁will therefore be charged. A charging current for the smoothing capacitor C₁will also flow through the path comprising the first rectifier output conductor 43, second inductance coil L₂, and smoothing capacitor C₁.

[0051] The current charging the smoothing capacitor C₁during each primary switch nonconducting period T_offwill diminish with time. Unlike the showing of FIG. 3 the charging current of the smoothing capacitor C₁need not flow throughout each primary switch nonconducting period T_off. The inductance values of the coils L₁and L₂may be so determined that the current I_inmay drop to zero as at t₄in FIG. 5 during each primary switch nonconducting period T_off.

[0052] During these nonconducting periods T_off, due to energy release from the transformer 5, there will be induced across the transformer secondary N₂a voltage oriented to cause conduction-through the diode D₀of the rectifying and smoothing circuit 6. The capacitor C₀and load 18 will both be powered through the diode D₀.

[0053] One cycle of switching operation, lasting as from t₂to t₅in FIG. 5, has now come to an end. The same cycle will be repeated after t₅when the primary switch Q₁is closed again. The power supply output voltage V₀may exceed a predefined limit in the course of such repetition of switching cycles. Thereupon the switch control circuit 8 will respond by shortening the primary switch conducting periods T_onto an extent necessary to return the power supply output voltage V₀to normal. The switch control circuit 8 will also respond to an excessive drop in the power supply output voltage V₀, by making the primary switch conducting periods T_onlonger until the output voltage returns to normal.

[0054] Reference is now invited to FIG. 4 again for a study of how the ancillary soft switching circuit 7 functions for the soft switching of the primary switch Q₁. FIG. 4 may be considered an enlargement of a part of the t₁-t₆period of FIG. 5. As will be noted from (A) and (B) in FIG. 4, which show the primary switch control pulses V_g1and ancillary switch control pulses V_g2, both primary switch Q₁and ancillary switch Q₂are off. before t₁in FIG. 4. Therefore, as in the t₃-t₅period in FIG. 5, the diode current I_dowill flow as at (M) in FIG. 4, and the voltage V_q1across the primary switch Q₁and the voltage V_q2across the ancillary switch Q₂will be both held high before t₁as at (C) and (E) in FIG. 4.

[0055] The ancillary switch Q₂is shown turned on at t₁by the ancillary switch control signal V_q2, at (B) in. FIG. 4. The current I_cq1discharged by the soft switching capacitor C_q1will then flow as at (G) in FIG. 4. There are two different paths for the flow of the discharge current, one comprising the soft switching capacitor C_q1, transformer primary N₁, transformer tertiary N₃, transformer quaternary N₄, ancillary diode D_a, and ancillary switch Q₂, and the other comprising the soft switching capacitor C_q1, transformer primary N₁, and smoothing capacitor C₁. The voltage V_q1across the primary switch Q₁will drop as at (C) in FIG. 4 with the flow of the discharge current, dropping substantially to zero at t₂. The transformer tertiary N₃and quaternary N₄, both connected in series with the ancillary switch Q₂, are inductive, so that the discharge current I_cq1of the soft switching capacitor C_q1flows due to the resonance of the capacitance of the capacitor C_q1and the inductances of the transformer tertiary N₃and quaternary N₄. The current I_cq1will gradually increase in absolute value from t₁.

[0056] As indicated at (F) in FIG. 4, the current I_q2through the ancillary switch Q₂will rise in magnitude with the current I_q1. It is thus seen that the zero current switching of the ancillary switch Q₂is accomplished at t₁, with little or no power loss when this switch is turned on. A voltage will develop across the transformer secondary N₂in a direction to reverse bias the output rectifying diode D₀during the ensuing t₁-t₂period when the current I_q2flows through the transformer tertiary N₃and quaternary N₄as at (F) in FIG. 4. The diode D₀will then become nonconductive, with no flow of current I_dotherethrough as at (M) in FIG. 4.

[0057] During this same t₁-t₂period there is some flow of current I_L1, as at (H) in FIG. 4, along the path comprising the second rectifier output conductor 44, first inductance coil L₁, transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, and third rectifier output conductor 45. There is also some flow of current I_L2, as at (I) in FIG. 4, along the path comprising the first rectifier output conductor 43, second inductance coil L₂, smoothing capacitor C₁and third rectifier output conductor 45.

[0058] Upon completion of discharge by the soft switching capacitor C_q1at t₂, the current I_q1shown at (D) in FIG. 4 will start flowing along the path comprising the transformer tertiary N₃, transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, primary switch protection diode D_q1, and transformer primary N₁, due to the liberation of the energy that has been stored on the transformer primary N₁, tertiary N₃and quaternary N₄. The current I_q1is shown at (D) in FIG. 4 as the sum of the current through the primary switch Q₁and the current through its protection diode D_q1. However, this current I_q1will be referred to as the primary switch current for simplicity.

[0059] The current I_db, at (L) in FIG. 4, will also flow during the t₂-t₃period along the path comprising the transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, and second ancillary diode D_b, due to the liberation of the energy from the transformer quaternary N₄. The current I_q2through the ancillary switch Q₂will start diminishing at t₂, as at (F) in FIG. 4, so that the primary switch current I_q1flowing through the primary switch protection diode D_q1will drop to zero at t₃.

[0060] Since the primary switch protection diode D_q1is conductive from t₂to t₃, the voltage V_q1across the primary switch Q₁is approximately zero during this period. The zero voltage switching of the primary switch Q₁is therefore accomplished if it is turned on during the t₂-t₃period. The primary switch control signal V_g1is shown at (A) in FIG. 4 as going high at t₂for ease of explanation. However, in consideration of possible fluctuations in the moment the primary switch Q₁is turned on in practice, it is desirable and practical that the primary switch be turned on midway between t₂and t₃.

[0061] It must nevertheless be pointed out that the primary switch Q₁may be turned on before t₂when the voltage V_q1across the same becomes zero, and not earlier than t₁when it starts dwindling. Switching loss will then diminish to an extent to which the voltage V_q1has dropped at the moment the primary switch is turned on.

[0062] Some reduction of switching loss is also possible if the primary switch Q₁is turned on shortly after t₃. With the primary switch Q₁held open at t₃, the resonance capacitor C_q1will start to be charged at that moment. But if the primary switch Q₁is turned on while the voltage across this capacitor C_q1is still less than the voltage V_q1across the primary switch during its nonconducting periods, then a corresponding reduction of the switching loss will be realized. Broadly speaking, therefore, the primary switch Q₁may be turned on at any moment after t₁when the ancillary switch Q₂conducts, provided that the voltage V_q1across the primary switch is less than that during the nonconducting period before t₁.

[0063] Since FIG. 4 is explanatory of what is taking place from t₁to t₆in FIG. 5, the primary switch protection diode D_q1is understood to conduct at t₂in FIG. 4. With a drop in the potential at the junction 10 between the transformer tertiary N₃and quaternary N₄, the current I_L1through the first inductance coil L₁will start increasing as at (H) in FIG. 4. This current I_L1will flow along the path comprising the second rectifier output conductor 44, first inductance coil L₁, transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, and third rectifier output conductor 45. As depicted at (I) in FIG. 4, the current I_L2through the second inductance coil L₂will continue flowing during the t₂-t₃period, as during the t₁-t₂period, but dwindle to zero at t₃.

[0064] At t₃, when the primary switch protection diode D_q1becomes incapable of being held conductive, the current I_q1through the primary switch Q₁will become zero and thereafter start flowing positive as at (D) in FIG. 4. Thus, during the ensuing t₃-t₄period, the primary switch current I_q1will flow along both the path comprising the first a.c. input terminal 1, filter 3, fifth diode D₅, first inductance coil L₁, transformer tertiary N₃, transformer primary N₁, primary switch Q₁, fourth diode D₄, filter 3, and second a.c. input terminal 2, and the path comprising the smoothing capacitor C₁, transformer primary N₁, and primary switch Q₁.

[0065] It is understood that energy release from the transformer quaternary N₄comes to an end at t₄, rather than at t₃, with the result that the current I_q2flows during the t₃-t₄period, as at (F) in FIG. 4, along the path comprising the transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, smoothing capacitor C₁, and transformer tertiary N₃. The rectifying diode D₀will be reverse biased by the voltage building up across the transformer secondary N₂when the primary switch current I_q1is going positive as from t₃to t₄, and from t₄to t₅, in FIG. 4. The current I_dothrough this diode D₀will therefore remain zero as at (M) in FIG. 4.

[0066] The current I_q2through the ancillary switch Q₂will be zero as at (F) in FIG. 4, whereas the current I_q1through the primary switch Q₁will flow as at (D) in FIG. 4, from t₄to t₅. During this period, as from t₃to t₄, there will be current flow both along the first path comprising the first a.c. input terminal 1, filter 3, fifth diode D₅, first inductance coil L₁, transformer tertiary N₃, transformer primary N₁, primary switch Q₁, fourth diode D₄, filter 3, and second a.c. input terminal 2, and along the path comprising the smoothing capacitor C₁, transformer primary N₁, and primary switch Q₁. The current on the first path is equivalent to the first inductance coil current I_L1shown at (H) in FIG. 4. The current on the second path is equivalent to the smoothing capacitor current I_C1during the t₄-t₅period at (K) in FIG. 4. The current I_q1through the primary switch Q₁is the sum of the currents on the first and the second path. The first and third diodes D₁and D₃of the rectifier circuit 4 are both reverse biased during this t₄-t₅period by the voltage developing across the second inductance coil L₂, which is electromagnetically coupled to the first inductance coil L₁, so that there is no current flow through the second inductance coil L₂.

[0067] The voltage V_q2across the ancillary switch Q₂is zero from t₄to t₅, as at (E) in FIG. 4. The current I_q2through the ancillary switch Q₂is zero from t₄to t₅, as at (F) in FIG. 4. The ancillary switch Q₂may therefore be turned off at any moment during this t₄-t₅period for both zero-voltage and zero-current switching. The conduction control of the ancillary switch Q₂is shown to end at t₅when that of the primary switch Q₁also ends, in this particular embodiment of the invention. The zero-voltage and zero-current switching of the ancillary switch Q₂is thus accomplished for reduction of switching loss when the ancillary switch Q₂is turned off. In practice, of course, the ancillary switch Q₂may be turned off at any other moment from t₄to t₅.

[0068] As plotted at (D) in FIG. 4, the current I_q1through the primary switch Q₁will drop to zero at t₅when this switch is turned off. The current I_cq1will flow instead into the soft switching capacitor C_q1, as at (G) in FIG. 4, thereby charging the same. The voltage V_q1across the primary switch Q₁will rise with a gradient as at (C) in FIG. 4. The zero-voltage turnoff of the primary switch Q₁has thus been achieved. The current charging the soft switching capacitor C_q1as above will flow both along the path comprising the rectifier circuit 4, first inductance coil L₁, transformer tertiary N₃, transformer primary N₁, and soft switching capacitor C_q1, and along the path comprising the smoothing capacitor C₁, transformer primary N₁, and soft switching capacitor C_q1.

[0069] The voltage across the soft switching capacitor C_q1will gradually build up after t₅when both primary switch Q₁and ancillary switch Q₂turn off. Thereupon the voltage impressed to the transformer primary N₁from the smoothing capacitor C₁will diminish until it becomes practically zero at t₆.

[0070] The transformer 5 will commence liberation of the stored energy at t₆when the primary switch Q₁completely goes off. The diode D₀of the rectifying and smoothing circuit 6 will then be forward biased by the voltage across the transformer secondary N₂, with the result that the diode current I_dowill start flowing as at (M) in FIG. 4. A voltage will develop across the transformer tertiary N₃, which is electromagnetically coupled to the transformer secondary N₂, with the indicated polarity when the primary switch Q₁is off. The potential at the junction 10 between the transformer tertiary N₃and quaternary N₄will therefore be higher than the voltage across the smoothing capacitor C₁by the voltage across the transformer tertiary N₃. The current I_L1through the first inductance coil L₁will start diminishing at t₆, as at (H) in FIG. 4.

[0071] On the other hand, being electromagnetically coupled to the first inductance coil L₁, the second inductance coil L₂will give rise to a voltage forwardly biasing the first and third diodes D₁and D₃of the rectifier circuit 4 with a rise in the potential at the junction 10 between the transformer tertiary N₃and quaternary N₄when the primary switch Q₁is off. The current I_L2through the second inductance coil L₂will start flowing at t₆, as at (I) in FIG. 4.

[0072] Flowing along the path comprising the rectifier circuit 4, first inductance coil L₁, transformer tertiary N₃, and smoothing capacitor C₁, the current I_L1shown at (H) in FIG. 4 will diminish in magnitude with energy release from the first inductance coil L₁until it becomes zero at t₇. As indicated at (I) in FIG. 4, on the other hand, the current I_L2on the second inductance coil L₂, which flows along the path comprising the rectifier circuit 4, second inductance coil L₂, and smoothing capacitor C₁, will gradually rise in magnitude until t₇and dwindle thereafter. These inductance coil currents I_L1and I_L2combine into the smoothing capacitor current I_C1.

[0073] The t₇-t₈period is a repetition of the pre-t₁period in FIG. 4. Due to energy release from the second inductance coil L₂, the current I_L2will flow along the path comprising the rectifier circuit 4, second inductance coil L₂, and smoothing capacitor C₁.

[0074] One cycle of operation comes to an end at t8. Another similar cycle will restart at t₈when the ancillary switch Q₂is turned on again.

[0075] The advantages gained by this particular embodiment of the invention may be recapitulated as follows:

[0076] 1. The primary switch Q₁is turned both on and off at zero voltage, assuring less switching loss, higher efficiency, and less noise production.

[0077] 2. The ancillary switch Q₂is turned on at zero current and off at zero voltage and zero current, resulting in less switching loss and noise production.

[0078] 3. The currents I_L1and I_L2flowing through the inductance coils L₁and L₂change in peak values with the amplitude of the a.c. input voltage V_ac, with consequent improvements in a.c. input power factor and waveform. Such improvements in power factor and waveform are accomplished with the aid of the primary switch Q₁in the d.c.-to-d.c. converter circuit comprising the smoothing capacitor C₁, transformer 5, primary switch Q₁, and rectifying and smoothing circuit 6. The objectives of improved power factor, improved waveform, and output voltage control are realized with the simple circuitry.

[0079] 4. The ancillary circuit 7 for the soft switching of the primary switch Q₁makes use of the windings N₃and N₄which are incorporated with the transformer 5 as its tertiary and quaternary, thereby avoiding too much increase in the size and cost of the power supply.

[0080] 5. By reason of the connection of the first inductance coil L₁to the junction 10 between the transformer tertiary N₃and quaternary N₄, the current I_L1does not flow through this first inductance coil even if the primary switch is closed, unless the potential on the first rectifier output conductor 44 grows higher than that at the junction 10. The first inductance coil current I_L1does not flow for this reason during the t₀-t₁, t₆-t₈, and t₉-t₁₀periods in FIG. 3. Although this is disadvantageous purely from the standpoints of waveform and power factor improvements, it should also be taken into account that power loss does not occur at the first inductance coil L₁as long as there is no current flow therethrough. Higher efficiency may therefore be attained without sacrifice in waveform and power factor through adjustment of the turns of the transformer primary N₁, tertiary N₃and quaternary N₄and of the periods of current flow through the first inductance coil L₁.

[0081] This alternative form of switching power supply features a rectifier circuit 4_ain substitution for the FIG. 1 rectifier circuit 4, and an addition of two reverse-blocking diodes D₁₁and D₁₂and a bypass capacitor C₂. All the other details of construction are as set forth above with reference to FIGS. 1 and 2.

[0082] The modified rectifier circuit 4_adiffers from its FIG. 1 counterpart 4 in having but four diodes D₁-D₄, instead of six, which are of the same connections as their equivalents in the FIG. 1 circuit 4. As the modified rectifier circuit 4_adoes not have the diodes D₅and D₆of the FIG. 1 circuit 4, the junction 40 between the first and third diodes D₁and D₃is connected to both first and second rectifier output conductors 43 and 44. These rectifier output conductors 43 and 44 are connected to the inductance coils L₂and L₁via reverse-blocking diodes D₁₂and D₁₁, respectively. The reverse-blocking diodes D₁₁and D₁₂could be connected on the output sides of the inductance coils L₁and L₂, as indicated in phantom outline in this figure.

[0083] The bypass capacitor C₂is connected between the third rectifier output conductor 45 and the junction 40 between the rectifier diodes D₁and D₃. The higher frequency component produced by the operation of the primary switch Q₁does not flow through the four rectifier diodes D₁-D₄but to the bypass capacitor C₂. There is therefore less noise production by the rectifier diodes D₁-D₄. Inexpensive low-frequency diodes find use as the four rectifier diodes D₁-D₄purely for rectification of the low-frequency input voltage V_ac, because no high-frequency current is to flow through these diodes D₁-D₄. The reverse-blocking diodes D₁₁and D₁₂should be high-frequency devices capable of response to the switching frequency of the primary switch Q₁.

[0084] Fundamentally, the FIG. 6 power supply operates the same way as does the FIG. 1 device and so gains the same advantages therewith. The advantages peculiar to this embodiment are as set forth previously.

[0085] A comparison of FIGS. 1 and 7 will reveal that this third preferred embodiment differs from that of FIG. 1 in: (a) the construction of another modified rectifier circuit 4_b; (b) the connections of inductance coils L₁and L₂; and (c) the addition of reverse-blocking diodes D₁₁and D₁₂.

[0086] The rectifier circuit 4_bis similar to the FIG. 6 rectifier circuit 4_ain having four diodes D₁-D₄but differs therefrom, and from the FIG. 1 rectifier circuit 4 too, in having but two output conductors 43 and 45. This pair of output conductors 43 and 45 are for production of a full-wave rectified output voltage.

[0087] Intended for improvements in power factor and waveform as in the foregoing embodiments, the two inductance coils L₁and L₂are electromagnetically coupled together, being both wound on a magnetic core 17, and polarized in the same direction. The first inductance coil L₁has one extremity thereof connected to the first rectifier output conductor 43, and the other extremity connected to one extremity of the second inductance coil L₂and, via the first reverse-blocking diode D₁₁to the junction 10 between the transformer tertiary N₃and quaternary N₄. The other extremity of the second inductance coil L₂is connected to the smoothing capacitor C₁via the second reverse-blocking diode D₁₂. The two inductance coils L₁and L₂may be formed by tapping a winding at 40′. The second reverse-blocking diode D₁₂could be connected between the tap or junction 40 and the second inductance coil L₂.

[0088] The FIG. 7 power supply operates essentially the same way as that of FIG. 1 as the switch control pulses V_g1and V_q2, shown at (A) and (B) in FIG. 8, are impressed respectively to the primary switch Q₁and ancillary switch Q₂from the switch control circuit 8 while the commercial sinusoidal a.c. voltage V_ac, (D) in FIG. 5, is being applied to the pair of a.c. input terminals 1 and 2. The input current I_in, (B) in FIG. 5, of the rectifier circuit 4 will change in amplitude with that of the a.c. input voltage V_ac, with the result that the a.c. input current I_ac, (C) in FIG. 5, closely approximates a sinusoidal wave for improvements in power factor and waveform.

[0089] Since the first inductance coil L₁is connected via the reverse-blocking diode D₁to the junction 10 between the transformer tertiary N₃and quaternary N₄, the current I_L1will flow through the first inductance coil L₁when the potential on the first rectifier output conductor 43 is higher than that of the junction 10 which is due to the voltage V_c1across the smoothing capacitor C₁. The rectifier input current I_inand a.c. input current I_acare shown to flow from t₁to t₆, and from t₈to t₉, in FIG. 5.

[0090] The operation of the FIG. 7 power supply exclusive of the ancillary soft switching circuit 7 will be explained in some more detail. Current will flow along the path comprising the first rectifier output conductor 43, first inductance coil L₁, first reverse-blocking diode D₁, transformer tertiary N₃, transformer primary N₁, and primary switch Q₁during each conducting period T_onof this switch Q₁, as from t₂to t₃in FIG. 5. Current will flow at the same time along the path comprising the smoothing capacitor C₁, transformer primary N₁, and primary switch Q₁. The voltage appearing across the transformer secondary N₂during each such period T_onwill be oriented to reverse-bias the diode D₀of the rectifying and smoothing circuit 6, holding this diode nonconductive. Energy will therefore be stored on the transformer 5, as well as on the first inductance coil L₁.

[0091] During each nonconducting period T_offof the primary switch Q₁, on the other hand, as from, t₃to t₅in FIG. 5, both first inductance coil L₁and transformer 5 will release the energy they have stored during the preceding conducting period T_on. Current will then flow along the path comprising the first rectifier output conductor 43, first inductance coil L₁, first reverse-blocking diode D₁₁, transformer tertiary N₃, and smoothing capacitor C₁thereby charging this capacitor. A charging current will also flow along the path comprising the first rectifier output conductor 43, first inductance coil L₁, second inductance coil L₂, second reverse-blocking diode D₁₂, and smoothing capacitor C₁. The current charging the smoothing capacitor C₁during each primary switch nonconducting period T_offwill diminish with time. Although FIG. 5 shows the charging current I_inas flowing throughout each primary switch nonconducting period T_off, this current may be made to cease flowing before the end of the nonconducting period, as at t₄in FIG. 5. The inductance values of the coils L₁and L₂may be readjusted for such termination of the charging current flow earlier than the end of each nonconducting period T_off.

[0092] As the transformer 5 liberates the stored energy during each primary switch nonconducting period T_offthere will be induced on the transformer secondary N₂a voltage oriented to cause conduction through the diode D₀of the rectifying and smoothing circuit 6. The result will be power delivery to both capacitor C₀and load 13. Another cycle of operation will restart at t₅in FIG. 5 when the primary switch Q₁reconducts. The output voltage V_oof the FIG. 7 power supply is held constant by the same method as in that of FIG. 1.

[0093] What follows is the operational description of the ancillary soft-switching circuit 7 with reference had to the waveform diagram of FIG. 8. This figure is explanatory of what is taking place in the FIG. 7 power supply during part of the t₁-t₆period of FIG. 5. As will be noted from (A) and (B) in FIG. 8, both primary switch Q₁and ancillary switch Q₂have been off before t₁. Operation during this pre-t₁period is similar to that of the t₃-t₅period in FIG. 5. The diode current I_doflows as at (M) in FIG. 8, and the primary switch voltage Q_q1and ancillary switch voltage V_q2are both high as at (C) and (E) in FIG. 8.

[0094] At (G) in FIG. 8 is shown the current I_cq1due to the discharge of the soft-switching capacitor C_q1upon conduction of the ancillary switch Q₂at t₁. The discharge current I_cq1will flow both along a path comprising the soft-switching capacitor C_q1transformer primary N₁, transformer tertiary N₃, transformer quaternary N₄, ancillary diode D_a, and ancillary switch Q₂, and along a path comprising the soft-switching capacitor C_q1, transformer primary N₁, and smoothing capacitor C₁. The voltage V_q1across the primary switch Q₁will diminish until it zeroes at t₂, as at (C) in FIG. 8.

[0095] Since the transformer tertiary N₃and quaternary N₄, both connected in series with the ancillary switch Q₂, are inductive, the soft-switching capacitor discharge current I_cq1flows due to the resonance of the capacitance of the capacitor C_q1and the inductances of the transformer tertiary N₃and quaternary N₄. The absolute value of this discharge current I_cq1will start rising gradually at t₁, as at (G) in FIG. 8. As indicated at (F) in FIG. 8, moreover, the current I_q2through the ancillary switch Q₂will also rise gradually. Thus is accomplished the zero-current turning-on of the ancillary switch Q₂at t₁.

[0096] As the current I_q2starts flowing through the transformer tertiary N₃and quaternary N₄at the beginning of the t₁-t₂period, there will develop across the transformer secondary N₂a voltage oriented to reverse-bias the output rectifier diode D₀. The current I_dothrough the diode D₀will drop to zero, as at (M) in FIG. 8, upon nonconduction of this diode.

[0097] During this same t₁-t₂period there will be some flow of current I_d11, at (H) in FIG. 8, along the path comprising the first rectifier output conductor 43, first inductance coil L₁, first reverse-blocking diode D₁₁, transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, and second rectifier output conductor 45. There will also be some flow of current I_d12, at (I) in FIG. 8, along the path comprising the first rectifier output conductor 43, first inductance coil L₁, second inductance coil L₂, second reverse-blocking diode D₁₂, smoothing capacitor C₁, and second rectifier output conductor 45.

[0098] Upon completion of discharge by the soft switching capacitor C_q1at t₂, the current I_q1will start flowing as at (D) in FIG. 8 due to the release of the energy that has been stored on the inductive transformer primary N₁, tertiary N₃and quaternary N₄. The path of the current I_q1comprises the transformer tertiary N₃, transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, primary switch protection diode D_q1, and transformer primary N₁. The current I_q1is shown at (D) in FIG. 8 as sum of the drain-source current of the primary switch Q₁and the current through the protection diode D_q1. However, the current I_q1will be referred to as the primary switch current for simplicity.

[0099] The current I_dbshown at (L) in FIG. 8 will also flow during this t₂-t₃period due to energy release from the transformer quaternary N₄. The path of the current I_dbcomprises the transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, and second ancillary diode D_b. The current I_q2through the ancillary switch Q₂will start diminishing at t₂, as shown at (F) in FIG. 8, so that the primary switch current I_q1will become zero at t₃. The voltage V_q1across the primary switch Q₁is remains approximately zero from t₂to t₃because its protection diode D_q1is conductive during this period.

[0100] It is thus seen that the primary switch Q₁can be turned on at zero voltage during the t₂-t₃period of FIG. 8. The primary switch control signal V_g1is shown at (A) in FIG. 8 as going high at t₂solely in order to expedite explanation. In practice the primary switch Q₁may be turned on intermediate t₂and t₃in consideration of possible fluctuations in turn-on control.

[0101] It is not a prerequisite, however, that the primary switch Q₁be turned on at or after t₂when the voltage V_q1across the same becomes zero. The primary switch Q₁may be turned earlier than t₂, but later than t₁when the voltage V_q1starts decreasing. Switching loss will be reduced even then to an extent depending upon the drop of the voltage V_q1from its pre-t₁value. Some reduction of switching loss is also possible if the primary switch Q₁is turned on shortly after t₃. The resonance capacitor C_q1will start to be charged at t₃if then the primary switch Q₁remains off. But switching loss will lessen if the primary switch Q₁is turned on when the voltage across the capacitor C_q1is still less than the voltage V_q1during the pre-t₁period when the primary switch is off. Broadly speaking, therefore, the primary switch Q₁may be turned on not earlier than t, when the ancillary switch Q₂is turned on and while the voltage V_q1across the primary switch Q₁is less than that during the pre-t₁period.

[0102] Since FIG. 8 is explanatory of what is going on from t₁to t₆in FIG. 5, the current I_d11through the first reverse-blocking diode D₁₁is shown at (H) in FIG. 8 as start rising at t₂when the primary switch protection diode D_q1conducts with a consequent drop in the potential at the junction 10 between the transformer tertiary N₃and quaternary N₄. This current I_d11will flow along the path comprising the first rectifier output conductor 43, first inductance coil L₁, first reverse-blocking diode D₁₁, transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, and second rectifier output conductor 45. As indicated at (I) in FIG. 8, the current I_d12through the second reverse-blocking diode D₁₂will continue flowing from t₂to t₃, as from t, to t₂, but dwindle to zero at t₃.

[0103] The current I_q1through the primary switch Q₁will become zero at t₃when the primary switch protection diode D_q1goes off, as at (D) in FIG. 8. The primary switch current I_q1will flow positive during the ensuing t₃-t₄period, both along the first path comprising the first rectifier output conductor 43, first inductance coil L₁, first reverse-blocking diode D₁₁, transformer tertiary N₃, transformer primary N₁, primary switch Q₁, and second rectifier output conductor 45, and along the second path comprising the smoothing capacitor C₁, transformer primary N₁, and primary switch Q₁.

[0104] Energy release from the transformer quaternary N₄is understood to terminate at t₄, rather than t₃, in this particular embodiment of the invention. For this reason the ancillary switch current I_q2is shown at (F) in FIG. 8 to flow from t₃to t₄along the path comprising the transformer quaternary N₄, first ancillary diode D_a, ancillary switch Q₂, smoothing capacitor C₁, and transformer tertiary N₃.

[0105] While the primary switch current I_q1is going positive as from t₃to t₄, and from t₄to t₅in FIG. 8, there will develop across the transformer quaternary N₂a voltage that will reverse-bias the diode D₀of the rectifying and smoothing circuit 6. The current I_dothrough this diode D₀will therefore remain zero as at (M) in FIG. 8.

[0106] From t₄to t₅the current I_q2through the ancillary switch Q₂will stay zero as at (F) in FIG. 8 whereas the primary switch current I_q1will continue flowing as at (D) in FIG. 8. During this period, as from t₃to t₄, there will be current flow both along the first path comprising the first a.c. input terminal 1, noise filter 3, first diode D₁, first inductance coil L₁, first reverse-blocking diode D₁₁, transformer tertiary N₃, transformer primary N₁, primary switch Q₁, fourth diode D₄, noise filter 3, and second a.c. input terminal 2, and along the second path comprising the smoothing capacitor C₁, transformer primary N₁, and primary switch Q₁. The current over the first path above is equivalent to the current I_d11, shown at (H) in FIG. 8, through the first reverse-blocking diode D₁₁. The current over the second path is equivalent to the smoothing capacitor current I_c1of the t₄-t₅(period at (K) in FIG. 8. The primary switch current I_q1is the sum of the currents flowing along the two paths defined above. A voltage that reverse-biases the second reverse-blocking diode D₁₂will develop across the second inductance coil L₂, which is electromagnetically coupled to the first inductance coil L₁, during this t₄-t₅period. There is therefore no current flow through the second inductance coil L₂.

[0107] The ancillary switch voltage V_q2is zero, as at (E) in FIG. 8, and so is the ancillary switch current I_q2, as at (F) in FIG. 8, from t₄to t₅. Consequently, the ancillary switch Q₂may be turned on at any moment during this period, for both zero-voltage and zero-current switching. Both primary switch Q₁and ancillary switch Q₂are understood to go off at the same moment, as at t₅in FIG. 8, in this particular embodiment, so that the zero-voltage and -current turnoff of the ancillary switch Q₂is accomplished. Of course, the ancillary switch Q₂could be turned off at t₄, as indicated by the dashed line at (B) in FIG. 8, or at any other moment from t₄to t₅.

[0108] The primary switch current I_q1will drop instantly to zero at t₅, as at (D) in FIG. 8, when the primary switch Q₁is turned off. Thereupon the soft-switching capacitor C_q1will be charged, and the primary switch voltage V_q1will rise rather abruptly, as at (C) in FIG. 8. The primary switch Q₁has thus been turned off at zero voltage. The current charging the soft-switching capacitor C_q1will flow both along the path comprising the rectifier circuit 4, first inductance coil L₁, first reverse-blocking diode D₁₁, transformer tertiary N₃, transformer primary N₁, and soft-switching capacitor C_q1, and along the path comprising the smoothing capacitor C₁, transformer primary N₁, and soft-switching capacitor C_q1.

[0109] Both primary switch Q₁and ancillary switch Q₂are shown to go off at t₅. Thereafter, with a gradual rise in the voltage across the soft-switching capacitor C_q1, the voltage impressed to the transformer primary N₁due to the smoothing capacitor C₁will diminish substantially to zero at t₆.

[0110] The transformer 5 will start liberating its stored energy at t₆when the primary switch Q₁goes totally off. With the consequent development, across the transformer secondary N₂, of a voltage forwardly biasing the diode D₀, the current I_d0will flow therethrough as at (M) in FIG. 8. A voltage will also develop with the indicated polarity across the transformer tertiary N₃when the primary switch Q₁is off, as this transformer tertiary is electromagnetically coupled to the secondary N₂. Therefore, when the primary switch Q₁is off, the potential at the junction 10 between the transformer tertiary N₃and quaternary N₄will be higher than the voltage across the smoothing capacitor C₁by an amount equal to the voltage across the transformer tertiary N₃. The current I_d11through the first reverse-blocking diode D₁₁will therefore start diminishing at t₆. A voltage forwardly biasing the second reverse-blocking diode D₁₂will develop across the second inductance coil L₂with a rise in the potential at the junction 10 between the transformer tertiary N₃and quaternary N₄when the primary switch Q₁is off, as the second inductance coil L₂is electromagnetically coupled to the first inductance coil L₁. The current I_d12will thus start flowing through the second reverse-blocking diode D₁₂at t₆, as at (I) in FIG. 8.

[0111] Flowing along the path comprising the rectifier circuit 4, first inductance coil L₁, first reverse-blocking diode D₁₁, transformer tertiary N₃, an smoothing capacitor C₁, the first reverse-blocking diode current I_d11will diminish with energy liberation from the first inductance coil L₁, until it becomes zero at t₇, as at (H) in FIG. 8. On the other hand, flowing along the path comprising the rectifier circuit 4, first inductance coil L₁, second inductance coil L₂, second reverse-blocking diode D₁₂, and smoothing capacitor C₁, the second reverse-blocking diode current I_d12will rise gradually until t₇and dwindle thereafter, as at (I) in FIG. 8. The first and second reverse-blocking diode current I_d11and I_d12constitute in combination the smoothing capacitor current I_c1shown at (K) in FIG. 8.

[0112] The t₇-t₈period is a repetition of the pre-t₁period in FIG. 8. The second reverse-blocking diode current I_d12will flow with energy release from the second inductance coil L₂during this period. The current I_d12will flow along the path comprising the rectifier circuit 4, first inductance coil L₁, second inductance coil L₂, second reverse-blocking diode D₁₂, and smoothing capacitor C₁.

[0113] One cycle of switching operation, lasting from t₁to t₈, has now come to an end. The same cycle will repeat itself after t₈.

[0114] The FIG. 9 switching power supply is akin to that of FIG. 7 except for the addition of a bypass capacitor C₂similar to that used in the FIG. 6 embodiment. The bypass capacitor C₂is shown connected between the rectifier output conductors 44 and 45 in FIG. 6 but between 43 and 45 in this FIG. 9 embodiment.

[0115] As has been set forth in conjunction with the FIG. 6 embodiment, the higher frequency component produced by the operation of the primary switch Q₁will not flow through the four rectifier diodes D₁-D₄but to the bypass capacitor C₂. There will therefore be less noise production by the rectifier diodes D₁-D₄. Inexpensive low-frequency diodes find use as the four rectifier diodes D₁-D₄which need to serve purely for rectification of the low-frequency input voltage V_ac. The reverse-blocking diodes D₁₁and D₁₂should be high-frequency devices capable of sufficient response to the switching frequency of the primary switch Q₁.

[0116] Notwithstanding the foregoing detailed disclosure, it is not desired that the present invention be limited by the exact showing of the drawings or the description thereof. The following, then, is a brief list of possible modifications or alterations of the illustrated embodiments which are all considered to fall within the scope of the invention:

[0117] 1. The bypass capacitor C₂employed in the FIGS. 6 and 9 embodiments could be used in the FIG. 1 embodiment as well. Such a capacitor might be connected between the rectifier output conductors 43 and 45 and/or between the rectifier output conductors 44 and 45. However, in such cases, reverse-blocking high-frequency diodes might be connected in series with the inductance coils L₁and L₂.

[0118] 2. The transformer 5 could be an autotransformer.

[0119] 3. A rectifying and smoothing circuit could be connected to the junction between transformer primary N₁and primary switch Q₁for provision of a step-up power supply.

[0120] 4. The primary switch Q₁could be made bidirectional, and the diode D_q1omitted.

[0121] 5. The second ancillary diode D_bcould be omitted if the current I_q2through the ancillary switch Q₂dropped to zero at t₃in FIGS. 4 and 8.

[0122] 6. A capacitor could be connected between the output of the first inductance coil L₁and the grounded conductor 45 in both FIGS. 6 and 9 for filtration of high frequency noise.

[0123] 7. The first ancillary diode D_awas unnecessary if the ancillary switch Q₂were equipped with the reverse-blocking function.

[0124] 8. An insulated-gate bipolar transistor and other semiconductor switches could be used in place of the FET switch Q₁.

[0125] 9. The invention could be applied to a forward switching power supply as well, as by modifying the rectifying and smoothing circuit 6 so as to cause conduction through the diode D₀during the conducting periods of the primary switch Q₁and by changing the polarity of the transformer secondary N₂.