GB2090078A

GB2090078A - D.c.-to-d.c. Switching Converter

Info

Publication number: GB2090078A
Application number: GB8104494A
Authority: GB
Original assignee: California Institute of Technology CalTech
Current assignee: California Institute of Technology CalTech
Priority date: 1977-09-28
Filing date: 1978-09-28
Publication date: 1982-06-30
Also published as: GB2090078B

Abstract

A d.c.-to-d.c. switching converter for connection between a source 10 and a load 20 comprises an input inductance 34 and an output inductance 36. The inductances 34 and 36 are coupled to form a transformer whereby the ripple in the input current i1 and/or the output current i2 may be reduced. When the matching conditions of the transformer are n=1/k where n is equal to positive <IMAGE> and L11 and L22 are the self inductances of the first and second inductances respectively and k is the coupling coefficient the input current ripple may be reduced to zero and when n=k the output current ripple may be reduced to zero. The convertor may include a capacitor 14 arranged to be alternately switched to transfer energy from the input inductance to the output inductance; it may include an isolation transformer; and the input and/or output inductances may be tapped to extend the step-up or step- down range of the converter. <IMAGE>

Description

SPECIFICATION D.c.-To-D.c. Switching Converter This invention relates to dc-to-dc converters and more particularly to a dc-to-dc switching converter.

Several of the prior-art dc-to-dc converters are described here and some of their deficiencies are discussed. Later, after the present invention disclosed herein has been described, comparison of the invention with the prior-art converters will reveal the superior characteristics of the invention.

All of the three configurations discussed here (conventional buck-boost, cascade connected boost and buck, and double-ended, push-pull converters) have the general DC conversion property, that is increase or decrease of both DC voltage and current. The same property is present in this invention as well, but the input or output current ripple may be reduced to zero, and if desired isolation between the input and output can be obtained.

According to the invention there is provided a dc-to-dc switching converter including a transformer having a first winding and a second winding, said windings being excited in use by separate and proportional pulsating voltage waveforms, the matching conditions of said transformer being substantially n=1/k where n is equal to positive

and L11 and L22 are the self-inductances of the first and second windings respectively, and k is the coefficient of coupling between said first winding and said second winding, whereby the current ripple in said first winding is reduced to substantially zero, or n=k whereby the current ripple in said second winding is reduced to substantially zero.

The invention also provides a dc-to-dc switching converter comprising an input inductance, a storage capacitance, an output inductance and switching means arranged to couple a dc source, the input inductance and the storage capacitance in series whereby to charge the storage capacitance from the source via the input inductance, and to couple the storage capacitance, the output inductance and a load in series to transfer energy from the storage capacitance to the load via the output inductance, the input and output inductances being magnetically coupled whereby to substantially reduce the ripple in the input current from the source or the output current to the load.

In some applications, both the load and the source may be referenced to the same ground. In those applications where the load and the source may not be referenced to a common ground, isolation between the source and the load is provided by dividing the storage capacitance between two capacitors with transformer coupling between the capacitors.

Either of the two inductances may be a tapped inductor to form an auto-transformer. In the input, the autotransformer provides an extended range of step-up of the input DC voltage, and in the output it provides an extended range of stepdown of the input DC voltage.

Certain embodiments of the invention will now be described by way of example and with reference to the accompanying drawings, in which: Fig. 1 is a schematic diagram of a conventional buck-boost converter.

Fig. 2 is a schematic diagram of a conventional transformer-coupled buck-boost converter.

Fig. 3 is a schematic diagram of a conventional cascaded boost converter followed by a buck converter.

Fig. 4 is a schematic diagram of conventional push-pull switching dc-to-dc converter.

Fig. 5 is a schematic diagram of a converter to which the present invention may be applied illustrating and idealizing its switching means, Figs. 5(1) and 5(2) illustrate configurations of the circuit of Fig. 5 with its switch to the left and to the right, respectively, Fig. 5a shows the circuit of Fig. 5 with a bipolar transistor and diode used to implement the ideal switch, and Figs. 5b and Sc show variants of the circuit of Fig. 5.

Fig. 6 is a waveform diagram illustrative of waveforms occurring in the circuit of Fig. 5.

Fig. 7 is a waveform diagram illustrating voltage waveforms occurring on input and output inductors.

Fig. 8 is a circuit diagram illustrative of an embodiment of the invention, and Figs. 8a and 8b show variants of the circuit in Fig. 8. Figs. 8(1), 8(2) and 8(3) are waveform diagrams illustrating the effects of inductive coupling design between input and output inductances in the present invention, and Fig. 8(4) is a generalized circuit diagram of this inductive coupling.

Fig. 9 is a circuit diagram of a variant of the invention using VMOS switching, and Figs. 9a and 9b are variants of Fig. 9.

Fig. 10 is a circuit diagram of a variant of the circuit in Fig. 5a which provides for DC isolation and to which the present invention may also be applied.

Fig. 11 is a circuit diagram of a variant of the circuit of Fig. 10.

Fig. 12 is a circuit diagram of a converter having multiple outputs.

Fig. 13 is a circuit diagram of an embodiment of the invention having multiple outputs with isolation and coupled inductors.

Fig. 1 4 illustrates a variant of the circuit of Fig.

10 where the isolation transformer is replaced by an autotransformer with multiple output taps.

Fig. 15 illustrates a variant of the circuit of Fig.

13.

A conventional buck-boost converter is shown in Fig. 1 together with the corresponding pulsating input and output current waveforms, iin and OUtS As seen in Fig. 1, the current drawn by this DC converter is not continuous and "clean" DC; rather it is pulsating, where the pulse amplitude in a higher current (and power) application may be very high. This invariably requires that an input filter F (usually a single section low-pass filter comprised of an inductance L1 and capacitance C1) be put in front to smooth out the substantial current pulse component at the switching frequency drawn from the line supply.In that way, electromagnetic interference (EMI) problems generated by the abrupt variation in energy flow (pulsating current) are reduced, and contamination of the environment by the undesired electromagnetic disturbances is alleviated. In operation, a transistor Q1 is alternately tumed on and off. When it is conducting, an inductance L2 stores energy, and while it is off, the energy stored flows through a diode D1 to a load R. A capacitance C2 across the load smooths the output current to the load.

While the input filter F reduces EMI problems at the input, the output current ioutt as seen in Fig.

1, is still pulsating, thus requiring a large size of output capacitance C2 to smooth out that current ripple. In addition, relatively complicated "floating" switch drive circuitry is required for the transistor in this converter since its emitter is not referenced to ground.

In summary, an important deficiency of this prior art is the fact that both input and output currents are pulsating, with corresponding EMI problems.

The converter of Fig. 1 can be easily updated to obtain a converter with a so-called isolation property that allows the unregulated source voltage ground to be isolated from the load ground. Such a modified transformer-coupled buck-boost converter is shown in Fig. 2, wherein corresponding elements are identified by like reference characters as in Fig. 1.

Note that the ideal DC voltage gain (V/Vg) of the converter in Fig. 1 is given byt

where Drduty ratio (ratio of the one-time to the switching period Ts), or

and D' is the complementary duty ratio,

The corresponding gain of the converter in Fig. 2 is:-

Hence, an additional control over the step-up or step-down of the voltage can be achieved by the transformer turns ratio n.

Note also the reversal of the output voltage polarity in the converter of Fig. 1 since it is inherently inverting, while that of Fig. 2 can have either polarity depending on the direction of coupling of the transformer (dot sign), and corresponding direction of the diode D1.

The same general DC conversion function of Equation (1) can be achieved by simply cascading the two basic converters, namely the boost

and the buck

resulting in the same overall DC gain as given by Equation (1). This suggested converter is shown in Fig. 3 wherein corresponding elements are identified by like reference numbers as in Figs. 1 and 2, but with subscripts 1 and 2 for the two basic converters cascaded.

While this converter has some good properties (both input and output currents continuous, that is, non-pulsating) it has some additional deficiencies. It needs an additional transistor Q2 and diode D2 which cause added DC and switching losses and hence significantly degrade the efficiency of the converter, besides its increase of complexity and number of components. Also at least one of the switching transistors requires floating drive circuitry, hence need for two isolated drive circuits, which further complicates its drive. Also there is no possibility to introduce the isolation property into this converter by simple means, as was the case in the conventional buckboost converter of Fig. 2.

Another commonly used variation of the conventional buck-boost converter which uses a push-pull arrangement of the two converters in Fig. 2 is shown in Fig. 4, where again corresponding elements are identified by like reference characters as in Fig. 2 but with subscripts 1 and 2 for the two basic converters working in parallel and a single output filter, L and C.

This typical push-pull type converter circuit generally requires many circuit components in addition to those shown to compensate for the inherently disadvantageous characteristics of the converter circuit. For example, the input impedance of the conventional push-pull converter tends to be capacitive causing large inrush currents when the circuit is initially energized. Also in this converter, peak transient voltages and currents occur owing to the conduction discontinuity when both switching transistors are biased nonconducting. Large capacity switching transistors and converter power transformer are required to deal with these peak transient currents and voltages. In addition, symmetry correction is frequently necessary to prevent the transformer from drifting into saturation owing to unequal parameter characteristics of the two switching transistors.

This converter normally requires an input filter which still further increases its already high complexity.

Many of the disadvantages and the deficiencies outlined in these converters are overcome by the present invention and in several of its extensions as verified both theoretically and experimentally in the later comparison.

Fig. 5 is a schematic diagram of a new converter topology with which the present invention may advantageously be used as will be described with reference to Figure 8. A DC input voltage (Vg) source 10 has its positive terminal connected to an input inductance 12. The input inductance is connected to a first capacitance 14.

An idealized switch 1 6 is provided for alternately connecting the junction between the first inductance 1 2 and storage capacitance 14, and connecting the junction between the latter and second inductance 18, to the common point between the source and the load. The second inductance 1 8 is connected in series with the load 20. Filter capacitance 22 is connected across the load.

Fig. 5(1) illustrates the circuit configuration with the switch to the left and Fig. 5(2) illustrates the circuit configuration with the switch to the right. Assume that the switch is initially to the right as in Fig. 5(2). Then, capacitance 14 charges up from the source 10 through inductance 12.

When the switch 1 6 is thrown to the left, it connects to ground the positive side of capacitance 14 as shown in Fig. 5(1). As a result, capacitance 14 discharges through inductance 18 into the load 20 and charges capacitance 22.

When the switch is then thrown back to the right, capacitance 14 again commences to charge up while inductance 18 supplies the load current.

Thus it is evident that as the switch 16 is operated alternately to the left and to the right, the capacitance 14 and inductance 18 will supply current to the load. The capacitance 22 aids the inductance, thereby filtering the output current ripple.

Fig. 6 shows in waveform A an on-off switching signal from the switching source 26.

Waveform B represents the input current ii from the source, and waveform C represents the output inductor current i2. Both the input and output currents are seen to have a ripple, but neither is pulsating (periodically falling to zero), as in the prior art. (See Fig. 1.) Fig. 5a shows the circuit of Fig. 5 with the idealized switch 1 6 implemented by a transistor 23 and a diode 24. In operation, assume that the transistor is initially in its nonconductive state.

Then capacitance 14 charges from the source 10 through the inductance since the diode 24 is forward-biased. At this time the collector of the transistor is positive with respect to its emitter so that when a switching pulse arrives to turn on the transistor, it becomes conductive. This has the effect of grounding the side of the capacitance that is connected to the transistor. The diode is thus back-biased, and the capacitance discharges through the inductance 18 into the load. When the switching pulse drops so that the transistor is cut off, the storage capacitance 14 again charges while the output inductance 18 supplies current to the load. Thus the transistor is driven by a train of pulses from a source 26 while the diode 24 responds as a complementary switch to the transistor being switched off and on.

From the foregoing it is evident that the capacitance 14 plays the role of an energytransferring device. Presently used converters store energy in inductance in the form of a magnetic field during one interval and release the energy to the load during a subsequent interval. In the present converter, the energy is accumulated in capacitance in the form of an electric field and then transferred to a load. Owing to significantly higher energy density per unit volume (or weight) of capacitor storage compared to inductive storage, significant size reductions are possible.

The DC voltage gain of the invention can be determined from the following formula:- V/Vg=D/Dt (3) where: Vg is the DC input voltage D is the fractional ON-time of the transistor switch; and D' is the fractional OFF-time of the transistor switch The DC current gain may be determined from the formula:- l2/11=D'/D (4) where: i1 is the DC input current 12 is the DC output current Thus, as seen from the polarity of output voltage in Fig. 5, the present converter is also an inherently inverting converter as is the conventional converter of Fig. 1.

By tapping the input inductor in the new converter of Fig. 5 (as shown in Fig. 5b), an additional step-up of the input DC voltage dependent on the tap ratio nt may be achieved.

Likewise, by tapping the output inductor in the converter of Fig. 5, as shown in Fig. 5c, an additional step down of the input DC voltage is provided through this autotransformer action.

The overall ideal DC gains of the two converters in Figs. 5b and Sc are then

(tapped input inductor) (5)

(tapped output inductor) (6) where the tap ratio nt is as defined in Fig. 5b or Fig. 5c. These two relatively straightforward modifications are useful when a large step-up, or step-down, ratio of the input voltage is required.

The tap ratio nt can be chosen to be less than unity if desired.

Besides these relatively simple modifications, some rather crucial extensions can be made according to the invention, as shown next.

Consider, for example, the voltage waveforms VL1 and VL2 on the two inductances 12 and 18 as defined in Fig. 5. They can easily be found by the fact that the voltage on capacitance 14 is essentially a DC voltage Vc1=VD'. With the transistor turned on:

(7) With the transistor turned off:-

The voltage waveforms VL1 and VL2 are then as shown in Fig. 7 which is a diagram illustrating these waveforms.

In Fig. 7, the waveform 30 is the waveform on inductance 12 and the waveform 32 is the waveform on inductance 18 when the converter shown in Fig. 5 is operating. From the waveforms shown in Fig. 7, it becomes clear that the two inductances can be coupled according to the invention as in a transformer, without affecting the basic DC conversion property provided the transformer has a one-to-one turns ratio. This is easily achieved by choosing a transformer having the primary and secondary winding with the same number of turns and with the direction of coupling shown by dots in Fig. 8, which illustrates a converter according to the invention wherein the inductances 14 and 18 are respectively the primary winding 34 and the secondary winding 36 of a one-to-one transformer.Other components of the circuit which serve the same function as the components on Fig. 1 are identified by the same reference numerals. A single-pole, double-throw switch 40 represents the functions of the diode and transistor in the circuit of Fig. S. The switching converter shown in Fig. 8 is the only switching converter (besides the straightforward cascade-connected boost and buck converters shown in Fig. 3) in which transformer coupling of the input and output inductance is possible. In other two-inductance converters (buck or buck-boost with input filter), this cannot be achieved.

While the basic conversion property is not affected by this transformer coupling, and all of the advantages of the optimum topology switching converter of Fig. 5 are retained, several additional benefits are achieved in the circuit of Fig. 8. One benefit is that either the input or the output current ripple can be reduced by an order of magnitude or more in comparison with their non-coupled counter-part. In fact, either input or output current ripple can be made zero. This has been verified both theoretically and experimentally. This allows further significant reduction in size and weight for the same ripple performance. A second benefit is that instead of two cores for two non-coupled inductors, a single core for the transformer may be used, thus further reducing the size, weight, and component count 3f the converter.Hence, the switching converter shown in Fig. 8 has the simplest possible structure (single 1:1 transformer, commutation capacitance 14, and switch 40), and yet it achieves the maximum performance (both input and output current are non-pulsating) in a topology which offers the smallest possible size and weight and highest efficiency.

Detailed analysis of the effect of this inductor coupling upon the current ripples shows that the ratio of the transformer primary and secondary leakage inductance plays the major role in the degree of reduction of the input and output current ripples. At one extreme, in the limiting case when primary leakage inductance becomes zero (for tightly coupled transformer) the output secondary current may become essentially DC with a zero switching ripple. At another extreme, when the two leakage inductances approach equality (usually for loosely coupled transformer), the reduction in the switching current ripple is equally divided between transformer primary and secondary, cutting approximately in half the current ripples in the original uncoupled converter.This balanced reduction represents, in a way, the "worst-case improvement" over the uncoupled configuration, since the unbalanced case (obtained by closely matching effective transformer turns ratio and coefficient of coupling) leads to an order of magnitude reduction in output current ripple, for example.

Both of these effects (loose and tight coupling) have been experimentally verified and shown to agree with theoretical predictions. With proper design of the transformer, the current ripple can be reduced to zero (DC current) at either the input or the output. For the latter, n=k and for the former n=1/k, where

and L11 and L22 are the respective primary and secondary self-inductances, and k is the coupling coefficient.

Figs. 8(1), 8(2), and 8(3) illustrate in waveform diagrams the advantage of coupling the input and output inductances, input and output currents, ii and i2, are shown in each Figure for three different conditions of inductive coupling. In Fig. 8(1) the condition of no inductive coupling (k=O) is first shown in dotted lines to illustrate the ripple in typical input and output currents. The solid lines in Fig. 8(1) then show the balanced reduction of ripple, the "worst-case improvement" referred to above, where n=1. By design on the inductive coupling for the condition n+1, either the input or the output ripple may be reduced to zero.If n is set equal to k (the value of the coefficient k is always less than 1), the input and output current waveforms of Fig. 8(2) can be achieved with zero ripple in the output current for a particular value of k. Similarly, by design of the inductive coupling for the condition n=1/kthe input and output current waveforms of Fig. 8(3) can be achieved with zero ripple in the input current for a particular value of k. The balanced condition (n=1) of Fig.

8(1) is thus clearly a compromise in reduction of ripple. Half of the reduction is achieved in the input current, and half in the output current. By unbalancing that condition in favor of the condition n=1/k, all of the reduction of ripple is achieved in the input current, ii as shown in Fig.

8(3). If n is increased further from that condition of zero input ripple or if the coefficient of coupling is further increased such that n > 1/k is satisfied, the input current will again have ripple, but with 1800 phase shift from the ripple that would otherwise be present for the balanced condition.

(The coefficient k is easily increased by decreasing the airgap of a cut core on which the inductors are wound.) By unbalancing the condition of Fig. 8(1) in the opposite direction in favor of the condition n=k, all of the reduction of ripple is achieved in the output current, i2, as shown in Fig. 8(2). If the value of k is further increased from that condition of zero output ripple, the output current will again have ripple, but with 1 800 phase shift from the ripple that would otherwise be present for the balanced condition. This coupling technique may be used to equal advantage in other converters having two inductors, such as the prior art converter of Fig. 3.

Fig. 8(4) illustrates a generalization of this inductive coupling technique. The switching voltage source v5 produces a nonpulsating input current ii across the primary inductance L1 and a switching voltage source av8 produces a nonpulsating output current i2 across the secondary inductance L2. It is possible to reduce the current ripple to zero at either end of a single transformer because both the primary and secondary in the new converter configurations disclosed herein are excited by stiff switching waveforms v5 and av8, where v9 and avS correspond to the respective switching waveform VL1 and VL2 of Fig. 7, and a is a constant.This results from the switching source avS for the secondary having been generated by the alternate energy transfer path through the storage capacitance C1 (Fig. 8), as described with reference to Figs. 5(1) and 5(2). By proper design of the transformer inductive coupling, balanced or unbalanced reduction of ripple currents may be achieved as in Fig. 8(1), Fig. 8(2) or Fig. 8(3).

Figs. 8a and 8b illustrate variants of the circuit of Fig. 8. The idealized switch is implemented by a transistor 23 and diode 24, as just suggested, but more important than that, the primary winding 34 is tapped in the variant of Fig. 8a to extend the step-up range of the input voltage. In Fig. 8b, the secondary winding is tapped to extend the stepdown range of the input voltage. This is analagous to the variants shown in Figs. 5b and Sc to the embodiment shown in Fig. 5a.

Realization of the switch 40 in Fig. 8 with a bipolar transistor and diode is only one practical way of implementation. Recent technological advances in vertical metal oxide semiconductor (VMOS) power transistors now permit a true bidirectional realization of the switch 40, and allow full advantage to be taken of the complete symmetry, as shown in Fig. 9. Two n-channel VMOS transistors 42 and 44 replace the diode 24 and transistor 23, respectively, of Fig. 5a. They are alternately switched on and off by a switching signal source. Presently only n-channel VMOS transistors are commercially available. This then necessitates the replacement of the switching diode by a VMOS power transistor 42 as shown with its drain (D) grounded. As a consequence this requires an isolated voltage drive for that transistor since the potential of its source (S) is floating.However, the two out-of-phase isolated voltage drives for two VMOS transistors 42 and 44 can be easily achieved by use of optical couplers OCA and OCB in a switch drive processing scheme as shown in block 46. Another possibility is to use a signal processing transformer with two secondary windings to obtain isolated voltage drives.

The practical implication of the switch 40 (Fig.

8) realization by VMOS transistors is that now the same physical configuration of Fig. 9 can be used to obtain both negative and positive output voltage with respect to ground, depending upon whether source voltage and load are used as shown in Fig. 9 or in the interchanged position, with corresponding polarity change of source voltage when applied on the right side. Hence the same converter can serve as both a positive and a negative power supply, a feature not present in any other converter.

However, very often a switching converter is used as a part of a closed-loop switching regulator, in which interchange of the source voltage and load position would be inconvenient since the feedback signal is taken from a fixed output position. Nevertheless by adding a single manual switch SM, as shown in Fig. 9a, this difficulty is circumvented. For the position of the switch shown, positive input voltage +Vg is applied and negative output voltage obtained.

However in the opposite position of the switch for cross-connection of the VMOS transistors, negative polarity input voltage Vg results in positive polarity output voltage, while the position of the load does not change. In either case voltage-regulating feedback may be provided.

Hence both positive and negative regulated voltage power supplies are obtained with the single switching, regulator configuration shown in Fig. 9a.

While the isolated voltage drives in Fig. 9 may seem quite complex, the expected availability of p-channel VMOS power transistors in the near future will significantly reduce the complexity of the drive circuitry as suggested in Fig. 9b, wherein both voltage drives are referred to ground and corresponding, out-of-phase (non-overlapping) drives can be easily achieved by a conventional signal processing technique.

Referring back to Fig. 9 again, similar functioning components, as are shown in Fig. 8, receive the same reference numerals. For the purpose of complete symmetry, capacitance 22a is connected across the input voltage source 10.

The input voltage source 10 and the load 20 may be interchanged, with the connections being made at the location shown with the polarity of the voltage source reversed when applied at the right side. In the case where the voltage source and the load are at the locations shown in Fig. 9, energy flow occurs from left to right. When they are interchanged, energy flow occurs from right to left.

In many applications of dc-to-dc switching converters, it is necessary to incorporate DC isolation between the input and output circuits of the converter. This invention may be very simply modified to achieve this result and at the same time to provide a capability for multiple outputs with different polarities and magnitudes as shown in Fig. 10. There, a circuit diagram is shown of a modification of the circuit shown in Fig. 5a which provides DC isolation between the input and output circuits of the converter. Those elements in Fig. 10 which perform the same function as in Fig.

5a are identified by the same reference numerals.

The modification comprises using two capacitances 50 and 52 in place of the capacitance 14, and the insertion of a one-to-one transformer 54 where the primary 56 connects between the capacitance 50 and the voltage source 10 and the secondary 58 connects between the capacitance 52 and the load 20.

This circuit arrangement operates according to the same principles described in connection with Fig. 5. The sum of the voltage of the two capacitances 50 and 52 equals Vg/D', and the individual capacitance voltages adjust to maintain balanced volt-seconds in the transformer 54. That is, there is automatic volt-second adjustment and there is no problem of operating point "creep" in the transformer.

When the transistor 23 is not conducting, capacitance 50 charges up through primary winding 56, inducing a voltage in secondary winding 58 which charges up capacitance 52.

Diode 24 is conductive at this time. When the transistor 23 becomes conductive, capacitance 50 discharges through primary winding 54, inducing a voltage in secondary winding 58. The voltage across capacitance 52 and the voltage induced in the secondary winding have a polarity to be additive, and they in turn charge up inductance 1 8 and supply current to the load.

When transistor 23 next becomes nonconductive, capacitance 50 charges up from the voltage source 10 causing current flow through primary winding 56 and inducing a voltage across secondary winding 58 which charges capacitance 52 up again. At the same time, inductance 1 8 discharges through the load 20, maintaining a nonpulsating output current. The capacitance 22 connected in parallel with the load filters the ripple in the output current.

While Fig. 10 shows a converter, using a transformer having a one-to-one conversion ratio, there is no reason why the converter may not have a transformer with a 1 :N turns ratio with relative polarity such that the output DC voltage is positive instead of negative, and is scaled by a factor N compared with its value in the original non-isolated converter.

Fig. 11 is a schematic diagram of a converter wherein the isolation transformer 54, which previously had a one-to-one ratio of primary and secondary windings is replaced by a transformer 60, having a primary winding 62 and a secondary winding 64 wherein the ratio of these windings is 1 :N and the relative polarity of these windings is reversed, as contrasted to the relative polarity of the windings of the transformer 54 shown in Fig.

10, as represented by the dots adjacent to these windings. The components of the circuit arrangement shown in Fig. 11, which perform the same function as those shown in Fig. 10, have the same reference numerals applied thereto. Fig. 11 provides an increased voltage output over the arrangement shown in Fig. 10 and a positive input and positive output voltage as well.

Fig. 12 is a schematic diagram of a converter having multiple outputs of various scaling ratios N1, N2, etc. The secondary winding 68 has a tap at a location 70. To obtain the voltage ratio 1 :N1, a capacitance 72 provides a function analogous to that of capacitance 52 in Fig. 10. An inductance 74 produces a function analogous to that of inductance 1 8. Diode 76 is connected between the tap 70 and the junction between capacitance 72 and inductance 74 and performs a function analogous to diode 24. Capacitance 78 is connected across the load 80 and performs the function analogous to the one performed by capacitance 22.

Similarly, to obtain the voltage ratio 1 :N2, connection is made to a tapping point 82. The circuit is the same as described, including the capacitance 84 connected in series with an inductance 86. A diode 88 connects the junction of capacitance 84 and inductance 86 to the tapping point 70. A capacitance 90 is connected across the load 92.

To obtain the voltage ratio 1 :N3, capacitance 94 is connected in series with inductance 96.

Diode 98 is connected between the junction of the capacitance and the inductance to the tap off point 70. Capacitance 100 is connected across the load 102.

Fig. 13 shows an embodiment of the invention wherein isolation between input and output is provided as well as coupling of both output inductances back to the input inductance. The isolation is achieved using transformer 104, wherein the secondary winding 106 has a tap at a point 108, to provide two ratios 1,N1 and 1 :N2.

There is a capacitance 110 connected in series with an inductance 112. A load 114 is connected to receive the output of the entire secondary winding 106 and a capacitance 11 6 is connected across the load. A diode 118 is connected between the junction of capacitance 110 and inductance 112, and the load.

For the second output, capacitance 120 is connected in series with an inductance 122.

Inductance 122 is connected to the other end of the winding 106 through a load 124. Capacitance 126 is connected across the load and diode 128 is connected between junction of capacitance 120 and inductance 122 and the other side of the load. The distinction between the circuit arrangement shown in Fig. 13 from the one shown in Fig. 12 is that in accordance with the invention the input inductance 1 30 is coupled to the respective output inductances 112 and 122.

For maximum ripple reduction and minimum size and weight, the inductance coupling ratio should be the same as the respective isolation transformer ratios for each output; however, benefits intermediate between the non-coupled inductances and equal ratio-coupled inductances can be obtained by use of unequal ratios.

Note that in the isolated switching converters of Fig. 10 and Fig. 11, the bipolar transistor and diode combination can again be substituted by the combination of two VMOS power transistors as previously shown in Fig. 9. However, to preserve the isolation property of the converter the isolated voltage drives of Fig. 9 should be implemented. Hence, the single configuration can be used (with addition of manual switch SM) to obtain both positive and negative, isolated, regulated output voltages.

Note also that if the multiple output feature is desired without the need for the isolation, a simple modification of the basic isolated version of Fig. 10 can be made. Namely, by use of an autotransformer 60 with several secondaries (tapoff points) instead of the isolation transformer in Fig.10, the converter of Fig. 14 is obtained. The advantage of using the configuration of Fig. 14 is the same as that coming from replacement of an isolation transformer by an autotransformt er, namely, the reduction in the size and weight of the transformer since the same winding is part of both primary and secondary windings. However, the isolation property is lost.

Fig. 1 S illustrates another variation of a dc-todc converter with an isolation transformer as in Fig. 13 and coupling of both output inductances back to the input inductance, as in Fig. 13 with tapping of the input and/or output inductances for extension of step-up and/or step-down voltage ranges as in Figs. 5b and 5c. Here the input inductance winding 130 is tapped for extended step-up range and one output inductor winding is tapped for an extended step-down range. While such an arrangement of both step-up and stepdown range extensions in a single converter of a multi-converter circuit is not likely to be required for any application, it does serve to illustrate tapping the coupled input and/or output inductances in a multi-converter configuration.

It is now appropriate to compare the coupled inductor extension of the new converter (Fig. 8) with a simple transformer-coupled version (Fig. 2) of the conventional buck-boost converter in Fig. 1.

As seen in Figs. 2 and 8, both converters consist of the same components: a single transformer, capacitance C, (capacitance C2 is not essential for the proper operation of the new converter in Fig.

8), and a single switch realized by the bipolar transistor-and-diode combination. However, while both converters have the simplest possible structure, the new converter of Fig. 8 has several important advantages. For example, the "flyback" converter of Fig. 2 has both the output and the input current pulsating, with resulting very poor ripple performance. By contrast, the coupledinductor extension (Fig. 8) has switching current and voltage ripples substantially reduced (by an order of magnitude as shown in Fig. 8(1)) over the already small ripple in the circuit of Fig. 5, and current ripple may be reduced to zero at either the input or output by proper design of the inductive coupling between the inductors, as shown in Figs.

8(3) and 8(2), respectively. The direct consequence of its nonpulsating output current is also the higher voltage gain and efficiency throughout the duty ratio variation. In addition, the transformer in the conventional "flyback" converter acts essentially as an inductor, since the energy is stored in the primary during one part of the switching cycle, and delivered to the output through the secondary during the other part of the cycle. In the transformer of the new converter of Fig. 8, however, the two windings function simply as the two separate original inductors for the DC current and as an AC transformer for ripple and other AC considerations. An obvious saving in both size and weight also results by placing the two inductor windings of the converter of Fig. 5 on the same core.

In still another comparison with the straightforward cascade connection of the conventional boost converter followed by the buck converter (Fig. 3), it may be seen that while the isolation and multiple output feature can be easily incorporated in the new converter, as shown in Fig. 13, by addition of a single transformer, this is not possible in the prior-art converter of Fig. 3.

However, the coupling of the inductors may still be implemented with the resulting ripple reduction benefits.

In summary, the new switching converter achieves some important advantages which are otherwise not achievable in conventional switching converter structures. The additional benefits are: 1. Coupling of the inductances in the new converter substantially reduced both input and output current ripple as well as output voltage switching ripple.

2. Implementation of the ideal switch S in Fig.

9a by two VMOS power transistors allcws the same converter to achieve a dual function, and to serve as both a positive or a negative regulated voltage supply.

3. Insertion of a single transformer in the structure of the new converter results in the highly desirable isolation property, together with multiple inverted or noninverted output capability.

Accordingly, there have been described and shown herein novel and useful circuit arrangements of dc-to-dc converters. Although only a limited number of particular embodiments have been illustrated and described, it is recognized that modifications and variations may readily occur to those skilled in the art.

Consequently, it is intended that the claims be interpreted to cover such modifications and variations.

Claims

1. A d.c.-to-d.c. switching converter including a transformer having a first winding and a second winding, said windings being excited in use by separate and proportional pulsating voltage waveforms, the matching conditions of said transformer being substantially n=l/k where n is equal to positive

2. A converter as claimed in claim 1 comprising an input inductance arranged to be connected in series with a source and an output inductance arranged to be connected in series with a load for supplying d.c. thereto, said input and output inductances being coupled to form said transformer.

3. A converter as claimed in claim 2 wherein said input and output inductances are coupled on a single core.

4. A d.c.-to-d.c. switching converter comprising an input inductance, a storage capacitance, an output inductance and switching means arranged to couple a d.c. source, the input inductance and the storage capacitance in series whereby to charge the storage capacitance from the source via the input inductance, and to couple the storage capacitance, the output inductance and a load in series to transfer energy from the storage capacitance to the load via the output inductance, the input and output inductances being magnetically coupled whereby to substantially reduce the ripple in the input current from the source or the output current to the load.

5. A converter as claimed in claim 4 wherein said input and output inductances are coupled to form a transformer, the matching conditions of said transformer being substantially n='/k where n is equal to positive

and L11 and L22 are the self-inductances of the input and output inductances respectively, and k is the coefficient of coupling between said input inductance and said output inductance, whereby the current ripple in said input inductance is reduced to substantially zero, or n=k whereby the current ripple in said output inductance is reduced to substantially zero.

6. A converter as claimed in claim 5 wherein said input and output inductances are coupled on a single core.

7. A converter as claimed in claim 4, 5 or 6 wherein said input inductance is a tapped inductor to form an autotransformer, the switching means being arranged to connect the tap to a source return current path whereby to extend the step-up range of the input voltage.

8. A converter as claimed in any of claims 4 to 7 wherein said output inductance is a tapped inductor to form an autotransformer, the switching means being arranged to connect the tap to a load return current path whereby to extend the step-down range of the input voltage.

9. A converter as claimed in any of claims 4 to 8 wherein said storage capacitance is divided into two capacitances, and including an isolation transformer having a primary winding connected between one capacitance and a return current path for said source and having a secondary winding connected between the other capacitance and a return current path for said load.

10. A converter as claimed in claim 9 wherein one end of said primary winding is connected in series sequence with a first storage capacitance and the input inductance to said source, and the other end of said primary winding is connected to the source return current path to form a primary series circuit, and one end of said secondary winding is connected in series sequence with a second storage capacitance and the output inductance to said load and the other end of said secondary winding is connected to the load return current path to form a secondary series circuit.

11. A method of operating a d.c.-to-d.c.

converter having input and output inductances coupled to form a transformer, comprising producing separate and proportional pulsating voltage waveforms in said input and output inductances, the matching conditions of said transformer being substantially n='/k where n is equal to positive

and11 and L22 are the self-inductances of the input and output inductances respectively, and k is the coefficient of coupling between said input inductance and said output inductance, whereby the current ripple in said input inductance is reduced to substantially zero, or n=k whereby the current ripple in said output inductance is reduced to substantially zero.

Claim 13 deleted Claims 4 to 10 deleted and appendant claims renumbered

11. A method of operating a d.c.-to-d.c.

converter having input and output inductances coupled to form a transformer, in which method separate and proportional pulsating voltage waveforms are produced in said input and output inductances, the matching conditions of said transformer being substantially n=1/k where n is equal to positive

and L,1 and L22 are the self-inductances of the input and output inductances respectively, and k is the coefficient of coupling between said input inductance and said output inductance, whereby the current ripple in said input inductance is reduced to substantially zero, or n=k whereby the current ripple in said output inductance is reduced to substantially zero.

12. A d.c.-to-d.c. converter as claimed in claim 1 and substantially as hereinbefore described with reference to the accompanying drawings.

1 3. In a converter, a transformer of appropriate design having a first winding and second winding, each winding being excited by a separate and proportional pulsating voltage waveform, thereby to reduce ripple in the current into said first winding to zero, or to reduce ripple in the current into said second winding to zero, depending upon the matching conditions in the design of the transformer, said matching conditions being n=ak for zero ripple in said second winding and n=a/k for zero ripple in said first winding, where n is equal to positive

and L11 and L22 are self-inductances of the respective first and second windings, k is the coefficient of coupling between said first winding and said second winding, and a is a positive constant of proportionality greater than, less than or equal to unity.

New Claims or Amendments to Claims filed on 12/10/81 and 12/3/82 New or Amended Claims:

2. A converter as claimed in claim 1 comprising an input inductance arranged to be connected in series with a source and an output inductance arranged to be connected in series with a load for supplying d.c. thereto, said input and output inductances forming said first and second windings and being coupled to form said transformer.