MULTI-PHASE CONVERTER

Abstract

Disclosed is a multi-phase converter comprising a plurality of electric phases, each of which can be triggered by a switching means. At least one coupling means (100 to 106) is provided for coupling a phase to another phase, each phase having two windings.

Description

BACKGROUND OF THE INVENTION

The invention relates to a multiphase converter.

A power converter, which can also be referred to as a DC-to-DC converter, is designed to convert a DC voltage at the input into a DC voltage with a different voltage level. Such a power converter can also be in the form of a multiphase converter with coupled inductances. Coupled multiphase converters comprise a plurality of phases, wherein each phase is routed through a conductor through which current flows, which conductors are coupled to one another by magnetic coupling means. Current ripple generated by each phase as a result of the magnetic coupling can, however, impair the operation of the multiphase converter, for which reason sections of the individual conductors need to be arranged in a suitable manner with respect to one another in order to avoid, for example, disadvantages associated with electromagnetic compatibility.

Document WO 2012/028558 A1 describes a multiphase converter in which, owing to the coupling of one phase of at least six phases to at least three other phases in magnetic opposition, disruptive mutual influencing of the phases is minimized and a large proportion of the magnetic flux is compensated for. The phases to be coupled are in this case selected such that optimum compensation can be achieved. This takes place by an opposing current profile of the phases. The aim here is for the phases to be magnetically coupled in such a way that the resultant magnetic field owing to the coupled phases is minimized. In this case, a ferrite core is used for coupling the magnetic fluxes in order to profit from the high permeability of the material. Given the coupling proposed here, the phases can be driven in order and independently of one another.

Document DE 10 2010 040 202 A1 discloses a multiphase converter comprising a plurality of phases, which are each drivable by a switching means. In this case, the coupling means is intended to magnetically couple at least one of the phases to at least three other phases.

Document DE 10 2010 040 222 A1 describes a multiphase converter for a plurality of electrical phases, which are each drivable by switching means. In this case, coupling means are provided which magnetically couple at least a first phase to at least a second phase, wherein at least two phases run spatially in one plane.

A similar multiphase converter is described in document DE 10 2010 040 205 A1. In said document, at least two coupling means are provided, wherein at least one of the two coupling means has a lower inductance than the other coupling means.

Various concepts with coupled inductances are therefore known, but at least some of these concepts have different disadvantages, of which some will be mentioned below by way of example. For example, excessive coupling results in EMC problems. An excessively complex design results in high manufacturing costs. If the phase start and the phase end are spatially separate, this results in disadvantages in respect of EMC and efficiency as a result of conductor loops. Furthermore, high core losses during no-load operation cause poor efficiency on a reduced load. A design which is excessively flat requires an excessively large amount of installation space. In addition, a high volume requirement for soft-magnetic material results in high costs.

Care should be taken to ensure that, in the power electronics, there are DC-to-DC converter topologies in which a leakage flux is required explicitly for operation. These include, inter alia, converters with coupled inductances. In special leakage transformers, such as, for example, doorbell transformers, welding transformers, series transformers etc., the leakage flux is used for short-circuit withstand capability and therefore for current regulation. This additional leakage flux can to a small extent be influenced by the type and/or embodiment of the copper winding in the ferrite core.

In the cases where this low, additionally achievable leakage inductance is insufficient, additional inductances, for example with component parts, need to be built into the current path. In addition to additional installation space, this also results in additional costs for component parts and power losses.

A further multiphase converter is known from WO 2009/114873 A1, for example. The DC-to-DC converter described therein comprises a coil comprising a nonlinear inductive resistor, a switching system and an output filter. In this case, adjacent phases are coupled to one another.

EP 1 145 416 B1 has already disclosed a converter for converting electrical energy. It is thus proposed here that the inductor size can be reduced by using coupled inductances. In this case, the coupled inductors are intended to be dimensioned such that the load currents in the subbranches compensate for one another and do not result in any magnetic loading of the inductors. Only the differential current between the individual subbranches then results in a magnetic field.

US 2009/0179723 A1 already discloses a DC-to-DC converter, in which the leakage inductance can be set by means of a distance between two phases to be magnetically coupled to one another which is selected in a targeted manner.

SUMMARY OF THE INVENTION

Against this background, a multiphase converter according to the invention and a coupling concept are proposed.

A coupling concept which is optimized in terms of installation space, costs and efficiency in respect of the requirements for automotive high-power converters is therefore proposed.

It is important that phases are magnetically coupled to coupling means, wherein each coupling means couples in each case at least two phases, wherein each phase comprises in each case at least two turns.

The multiphase converter can be configured in such a way that, by introducing a magnetic leakage flux in a targeted manner, the coupling of the phases is slightly reduced, as a result of which a reduction in the power loss and therefore an increase in the efficiency can be achieved. Thus, interference could also be further reduced. This is achieved by a means for influencing a magnetic leakage flux, which means is arranged between at least two of the phases to be coupled. This solution manages without any additional installation space and is therefore space-saving.

It should be noted that the proposed multiphase converter can be designed for a number of phases. For example, two, three, four, five, six or more phases can be provided. At least two turns are in this case associated with each phase.

Particularly expediently, the means for influencing a magnetic leakage flux is connected to the coupling means in such a way as to conduct magnetic flux. Thus, this functionality could be integrated even during manufacture of the coupling means, for example by pressing the ferrite cores as coupling means directly in the production process.

An expedient development provides that the means for influencing a magnetic leakage flux is arranged centrally between the two phases to be coupled. Particularly expediently, the means for influencing the leakage flux is rectangular or dome-shaped. An expedient development provides that the means for influencing the leakage flux is connected to the coupling means on only one side. This variant can be manufactured in a particularly simple manner since the means for influencing the leakage flux is part of the coupling means and consists of the same material.

An expedient development provides that a gap, preferably an air gap, preferably of the order of magnitude of 1 mm, particularly preferably between 0.3 and 0.5 mm, is provided between the means for influencing the leakage flux and the coupling means. This size avoids any undesired saturation effects beyond a certain tolerance band.

An expedient development provides that the at least one phase is embodied as a round wire. This increases the leakage fluxes and it is thus possible for the interference to be further reduced.

An expedient development provides that a first phase has substantially a planar, U-shaped profile, while a second phase has a substantially rectangular, planar profile. These phases with such a design can be surrounded by coupling means, preferably conventional ferrite cores. As a result, the desired coupling of the phases is achieved in a very simple manner by using a matrix-shaped design.

A number of phases, for example two to five, can be provided. In this case, the coupling means couple in each case one phase to precisely one other phase, to two phases, to three or more phases. It is expedient if the individual phases can still be controlled independently of one another. In this case, at least two turns are associated with each phase.

A particularly expedient development provides that precisely six phases are provided, wherein the coupling means magnetically couple each of the six phases to three further phases of the six phases. This type of coupling ensures firstly that the individual phases can still be controlled independently of one another. In addition, the failsafety of the multiphase converter can be increased owing to the increased level of interconnection of the phases.

An expedient development provides that the coupling means comprises at least two parts, wherein one of the parts has a U-shaped, O-shaped, I-shaped or E-shaped cross section. By virtue of this design, the phases to be coupled can be surrounded by the coupling means in a particularly simple manner. An expedient development provides that a gap, preferably an air gap, is provided between two parts. In this way, the inductance can be influenced particularly easily. An expedient development provides that a plurality of coupling means comprising at least two parts have at least one common part, preferably a metal plate. Thus, fitting could be facilitated since all of the coupling means could be closed by positioning of the plate in only one step.

An expedient development provides that at least two, in particular three, coupling means are provided in order to magnetically couple one of the phases to two further phases, wherein at least one of the two coupling means has a lower inductance than the other coupling means. By suitable selection of the inductance of the coupling means, various aspects can be influenced and optimized. Firstly, the inductance influences the power loss and therefore also the development of heat in the coupling means. A reduction in the inductance also reduces the power losses. In addition, a lower inductance can act as protection against saturation. As a result, coupling means with a lower inductance only enter saturation later at higher currents, with the result that the multiphase converter can be operated for longer in a stable operating state in the event of a fault. Secondly, a high inductance reduces the current ripple. Thus, by selecting the suitable inductance, the distribution of losses, saturation response and current ripple can be optimized.

An expedient development provides that the coupling means which couples one phase to a phase which is driven with a phase shift substantially through approximately 180° has a lower inductance than at least one of the other coupling means. As a result, said coupling means which are generally subjected to a greater load can be reduced in terms of losses such that a lower development of heat is also achieved.

BRIEF DESCRIPTION OF THE DRAWINGS

Further advantages and configurations of the invention result from the description of the attached drawings.

It goes without saying that the features mentioned above and yet to be explained below can be used not only in the respectively specified combination, but also in other combinations or on their own, without departing from the scope of the present invention.

FIG. 1 shows coupling elements which couple phases to one another.

FIG. 2 shows a coupling element in a side view.

FIG. 3 shows the coupling concept.

FIG. 4 shows further coupling elements.

FIG. 5 shows a circuit arrangement.

FIG. 6 shows a schematic illustration of the respective coupling of the phases.

FIG. 7 shows driving and current characteristics in the case of the exemplary embodiment shown in FIG. 5.

FIG. 8 shows a section through a coupling means comprising two coupled phases.

FIG. 9 shows a section along the line A indicated in FIG. 4 in the case of a matrix-shaped design comprising nine coupling means in a plan view.

FIG. 10 shows a section through a coupling means comprising means for influencing the leakage flux comprising two coupled phases in an alternative exemplary embodiment.

FIG. 11 shows an alternative exemplary embodiment of a coupling means comprising an air gap and means for influencing the leakage flux.

FIG. 12 shows a section through a coupling means comprising dome-shaped means for influencing the leakage flux comprising two coupled phases in a further alternative exemplary embodiment.

FIG. 13 shows a section through a coupling means comprising rectangular means for influencing the leakage flux comprising two coupled phases in a further alternative exemplary embodiment.

FIG. 14 shows a section through a coupling means comprising rectangular means for influencing the leakage flux comprising two coupled phases in a further alternative exemplary embodiment, in which first and second phase are always arranged alternately.

FIG. 15 shows a section through a coupling means comprising means for influencing the leakage flux comprising two coupled phases in the form of round wires in a further alternative exemplary embodiment.

DETAILED DESCRIPTION

The invention is illustrated schematically in the drawings on the basis of embodiments and will be described in detail below with reference to the drawings.

The figures show the compact and flat design of the four-phase system of coupled inductances with adjustable leakage flux.

In this system, in each case two phases are coupled to one another per coupling element or core. In the four cores, each phase is therefore magnetically coupled to its predecessor and successor (in time). By virtue of the embodiment with in each case two turns per phase, the magnetic saturation of the core cross section is markedly reduced. This in turn results in markedly lower core losses and therefore in less complexity in terms of cooling.

Owing to the fact that no air gap is required in the useful flux path in coupled systems, high inductance values can also be realized with the two turns. Owing to the increased inductance, any core can be provided with an air gap as protection against saturation.

In comparison with embodiments with only one turn, the current which can be tolerated as imbalance/load splitting between the phases is doubled. The additional segments are used to introduce the desired leakage flux. This can take place during manufacture by grinding the air gap. The design is simple, and therefore stamped or bent parts can be inserted into the lower core half and can be closed from above with an identical second core half. Additional winding techniques are dispensed with.

If the installation space available does not permit a flat extent, the 2×2 matrix can also be folded to form a narrow and more compact design.

This results in a plurality of advantages for low-voltage automotive converters having high powers and efficiencies:

1) high efficiency owing to low copper resistance,
2) low quiescent losses and thus high efficiency on reduced load,
3) high dynamics owing to coupling,
4) easily adaptable leakage flux and thus optimizable EMC response to the load demands,
5) small conductor loop since each phase can be passed directly out of the core where it is merely guided. As a result, it can be terminated directly at the switching cell with a capacitor, which is advantageous for EMC.

FIG. 1 shows coupling means 100, 102, 104 and 106, which each couple two phases to one another, wherein each phase comprises two turns.

FIG. 2 shows the coupling means 102 comprising a first turn 110, a second turn 112, a third turn 114 and a fourth turn 116. The first turn 110 and the second turn 112 form a first phase, and the third turn 114 and the fourth turn 116 form a second phase. These two phases are magnetically coupled to one another in the coupling means 104.

This coupling is illustrated once again in FIG. 3. It is important that the turns of the phases can be arranged next to one another and one above the other. Furthermore shown are turns 120 and 122, which form a phase, and turns 130 and 132, which form a phase.

FIG. 4 shows a further illustration of coupling means 150, 152, 154 and 156, which each couple two phases, which each comprise two turns.

The design of a multiphase converter 10 is illustrated in terms of circuitry in FIG. 5. The multiphase converter 10 described by way of example here comprises six phases 11 to 16. Each of the phases 11 to 16 can be driven individually via corresponding switching means 21 to 26, in each case comprising a high-side switch and a low-side switch. Each current of the phases 11 to 16 flows, owing to magnetic coupling to three further phases, through three inductances Lxx, which magnetic coupling is effected by the corresponding coupling means 31 to 39. A first coupling means 31 couples the first phase 11 to the second phase 12 magnetically, with the result that an inductance L12 results for the first phase 11, and an inductance L21 results for the second phase 12. A sixth coupling means 36 couples the first phase 11 to the sixth phase 16 magnetically, with the result that an inductance L16 results for the first phase 11, and an inductance L61 results for the sixth phase 16. A seventh coupling means 37 couples the first phase 11 to the fourth phase 14 magnetically, with the result that an inductance L14 results for the first phase 11, and an inductance L41 results for the sixth phase 16. A second coupling means 32 couples the second phase 12 to the third phase 13 magnetically, with the result that an inductance L23 results for the second phase 12, and an inductance L32 results for the third phase 13. A ninth coupling means 39 couples the second phase 12 to the fifth phase 15 magnetically, with the result that an inductance L25 results for the second phase 12, and an inductance L52 results for the fifth phase 15. A third coupling means 33 couples the third phase 13 to the fourth phase 14 magnetically, with the result that an inductance L34 results for the third phase 13, and an inductance L43 results for the fourth phase 14. An eighth coupling means 38 couples the third phase 13 to the sixth phase 16 magnetically, with the result that an inductance L36 results for the third phase 13, and an inductance L63 results for the sixth phase 16. A fourth coupling means 34 couples the fourth phase 14 to the fifth phase 15 magnetically, with the result that an inductance L45 results for the fourth phase 14, and an inductance L54 results for the fifth phase 15. A fifth coupling means 35 couples the fifth phase 15 to the sixth phase 16 magnetically, with the result that an inductance L56 results for the fifth phase 15, and an inductance L65 results for the sixth phase 16.

An input current IE is distributed among the six phases 11 to 16. A capacitor as filter means is connected to ground at the input. The outputs of the phases 11 to 16 are combined at a common summation point and connected to ground by means of a capacitor (not designated), as filter means. Then, the output current IA is present at the common output-side summation point. The inductances Lxx respectively coupled to one another are oriented with a different winding sense with respect to one another as indicated by the corresponding points in FIG. 1.

In this case, at least two turns or windings can be provided for each phase.

FIG. 6 illustrates systematically how the six phases 11 to 16 are coupled to one another by corresponding coupling means 31 to 39. As already described in connection with FIG. 1, both adjacent phases are coupled to one another and also, in addition, the phase offset through 180°. An adjacent phase is understood to mean one which is driven directly previously or subsequently in time, i.e. whose switch-on times are directly preceding or succeeding in time. In the exemplary embodiment, the designation of the phases 11 to 16 is selected such that the phases 11 to 16 are driven successively corresponding to the numbering, i.e. in the following order (figures correspond to the reference signs of the phases): 11-12-13-14-15-16-11 etc., in each case phase-shifted through 60 degrees or through T/6 (360 degrees/number of phases), where T is the period of a drive cycle. This order is also shown in FIG. 2 and FIG. 7. That is to say that the start times for the various phases 11 to 16 are phase-shifted through in each case 60 degrees or shifted in time through in each case T/6. Admittedly the respective phase is switched off again after the time duration T/6 in FIG. 7 (PWM ratio 1/6). Depending on the desired voltage ratio, the disconnection could be earlier or later up to permanently on Te, depending on the desired PWM signal (between 0% (permanently off, Te=0) and 100% (permanently on, Te=T), based on a period T).

The graph shown in FIG. 7 shows the time characteristics of the drive signals 52 for the respective switching means 21 to 26 of the corresponding phases 11 to 16 and the current characteristics in phases 11 to 16. The switching means 21 to 26 energize the associated phases 11 to 16 successively for in each case one sixth of a period T, for example by a PWM signal, and are then in the freewheeling mode. The resultant current characteristics of the individual phases 11 to 16 are shown by way of example below this. The period T of the drive signals 52 is of the order of magnitude of 0.01 ms, for example. The start times for the different phases 11 to 16 are phase-shifted through in each case 60 degrees or shifted in time through T/6. The start time of the second phase 12 with the corresponding drive signal 52 of the second switching means 22 is t=0 and is switched off again after 1/6 T (depending on the desired PWM ratio). The start time of the third phase 13 adjacent to the second phase 12 is T/6, the start time of the fourth phase 14 is 2T/6, and so on. Admittedly the respective phase is switched off again after T/6 (PWM ratio 1/6) in FIG. 7. Depending on the desired voltage ratio, however, the disconnection could take place earlier or later, up to permanently on, depending on the desired PWM signal (between 0% (permanently off) and 100% (permanently on)). That is to say that a plurality of phases 11 to 16 could also be energized simultaneously at a specific time if this is required by the desired voltage ratios. The start times are temporally offset, however.

FIG. 8 shows the design of the coupling means 31-39 by way of example with reference to the first coupling means 31, which magnetically couples the first phase 11 and the second phase 12. The first coupling means 13 comprises a substantially E-shaped first part 44 and a plate-shaped second part 43, which form the coil cores. The outer limbs of the first part 44 with an E-shaped cross section are all equal in length, with the result that they can be closed, for example by adhesive bonding, by the plate-shaped (I-shaped cross section) second part 43 without an air gap. In each case one means 81 for influencing the leakage flux is arranged between the central limb and the outer limb of the E-shaped part 44 of the coupling means 31. Said means is part of the first part 44 and is likewise rectangular and oriented in the same way as the outer limbs. However, the means 81 for influencing the leakage flux is slightly shorter than the outer limbs, with the result that an air gap 96 is formed with respect to the second part 43 in the positioned state. The second phase 12 is arranged between the left-hand outer limb of the first part 44 and the means 81 for influencing the leakage flux, and the first phase 11 with a current flow in the opposite direction is arranged between the means 81 and the central limb. The first phase 11 is located between the other side of the central limb and the further means 81 for influencing the leakage flux. The second phase 12 is arranged with a current flow in the opposite direction between the further means 81 and the right-hand outer limb of the first part 44. The outer limbs of the first part 44 have an area A1 in plan view. The central limb of the first part 44 preferably has an area 2*A1 in plan view. The limbs of the means 81 for influencing the leakage flux 81 have an area A2. For reasons of simple manufacture, the coupling means 31 to 39 are each constructed from the two parts 43, 44, as described. The outer limb and the central limb are connected to the second part 43 in such a way that a magnetic circuit is closed. Thus, only small gaps, for example of the order of magnitude of approximately 10 μm, are permitted. In order that the means 81 build up the desired leakage flux, in the exemplary embodiment the air gap 96 between the ends of the means and the second part 43 in the positioned state is selected to be of the order of magnitude of 1 mm, preferably between 0.3 and 0.5 mm. The air gap 96 is also dependent on the geometry of the means 81 for influencing the leakage flux, in particular on the area A2. Given an area A1 of approximately 100 mm2 and an area A2 of likewise 100 mm2, the above-cited region of the air gap 96 has proven successful.

FIG. 9 now shows a schematic illustration of the matrix-shaped physical design of the concept shown in FIG. 6. As already described in conjunction with FIG. 1, each phase 11 to 16 is magnetically coupled to three further phases. For this purpose, by way of example for the first phase 11, three separate coupling means 31, 36, 37 are provided. The first coupling means 31 couples the first phase 11 magnetically to the second phase 12. The sixth coupling means 36 couples the first phase 11 to the sixth phase 16. The seventh coupling means 37 couples the first phase 11 to the fourth phase 14. The preferred coupling principle and coupling means 31 has already been described in conjunction with FIG. 4. The nine separate coupling means 31 to 39 are preferably in the form of planar coil cores, for example ferrite cores, which each have two cavities. In each case two conductors or phase sections, physically separated from means 81 to 89 for influencing the leakage flux, of two phases to be coupled are surrounded in these cavities in the coupling means 31 to 39, which phases have different current directions in these sections. The cavities accommodate the two phases (11, 12) to be coupled. However, they could also be at least partially filled with a non-ferromagnetic material.

The plan view in FIG. 9 shows that the phases 11 to 16 have only two shapes. The first, third and fifth phase 11, 13, 15 is U-shaped. The second, fourth and sixth phase 12, 14, 16 is meandering. In the exemplary embodiment, the phases could be embodied as insulated flexible round conductors, which are arranged within the coupling means 31 to 39 in the same plane.

By way of example, the means 80 to 89 for influencing the leakage flux have different shapes. The embodiment shown on the left in FIG. 10 has two dome-shaped structures, which are arranged between the phases 11, 12. The dome-shaped structures 80 protrude from both sides between the phases 11 and 12, but without touching one another. The ends of the means 80 for influencing the leakage flux are arranged spaced apart from one another, with the result that the magnetic circuit is not closed, but the leakage flux is deflected toward the ends. The dome-shaped structures are part of the second part 44.

The right-hand structure of the means 80 for influencing the leakage flux shown in FIG. 10 has a rectangular cross section, is arranged between the two phases 11, 12 and is part of the second part 44. The end of the means 80 for influencing the leakage flux is oriented in the direction toward the central limb of the second part 44, but without touching it. As a result, no useful magnetic flux is produced. Instead, the leakage flux is guided in a targeted manner by the means 80 and can be influenced in a targeted manner via the distance or air gap with respect to the central limb of the second part 44.

The first phase 11 and the second phase 12 are now magnetically coupled to one another by the first coupling means 31. The antiparallel current conduction indicated results in the resultant magnetic field being kept as low as possible, with the result that the size of the coupling means 31 can be minimized. In addition, insulation 45 is provided in each case between the first phase 11 and the second phase 12 to electrically isolate the two phases 11, 12 from one another and in each case with respect to the coupling means 31.

The exemplary embodiment shown in FIG. 11 differs from that shown in FIG. 10 in that the central limb of the E-shaped first part 44 has an air gap 64 in the direction toward the second part 43. The means 80 for influencing the leakage flux are plate-shaped and are arranged between the phases 11, 12. In the exemplary embodiment, the ends of the means 80 for influencing the leakage flux are in each case oriented in the direction of the central limb of the second part 44, without touching it.

The exemplary embodiments shown in FIGS. 12 to 14 differ from the preceding exemplary embodiments, in particular in terms of the arrangement of the phases 11, 12. The phases 11, 12 to be coupled are in each case arranged next to one another and are strip-shaped. In FIG. 12, the means 80 for influencing the leakage flux are dome-shaped. Furthermore, the two sections of the first phase 11 are isolated by the central limb of the second part 44, i.e. are arranged adjacent to one another.

The exemplary embodiment shown in FIG. 9 differs from that shown in FIG. 8 only in terms of the plate-shaped configuration of the means 80 for influencing the leakage flux.

The exemplary embodiment shown in FIG. 9 differs from that shown in FIG. 8 in that the first and second phases 11, 12 are now always arranged alternately.

The exemplary embodiment shown in FIG. 11 differs from that shown in FIG. 9 in that the phases 11, 12 are in the form of round conductors.

DESCRIPTION OF THE EXEMPLARY EMBODIMENTS

The described exemplary embodiments function in the manner described in more detail below. Multiphase converters 10 or DC-to-DC converters with high powers without any particular requirements in respect of insulation can preferably be realized in polyphase arrangements. As a result, the high input current I_E, for example of the order of 300 A, is distributed among the various six phases 11 to 16 with of the order of 50 A in each case. By subsequently superimposing the individual currents to give an output current IA, a lower AC component can be achieved. Then, the corresponding input or output filters as shown in FIG. 5, illustrated by way of example as capacitors, can be correspondingly small. The driving of the phases 11 to 16 takes place sequentially, i.e. successively, with the result that the switch-on times are phase-shifted in each case through 60 degrees (or temporally through T/6) (in the case of the six-phase system described), as has been shown already in more detail in FIG. 11. Depending on the desired voltage ratios, the respective phases 11 to 16 are energized for different durations. The corresponding high-side switch of the switching means 21 to 26 is closed for this purpose. The phases 11 to 16 are not energized when the corresponding low-side switch of the switching means 21 to 26 is closed. Alternatively, those phases 11 to 16 whose switch-off times are directly preceding or succeeding could also be considered to be adjacent. Then, the corresponding switch-on points would be selected variably depending on the desired PWM signal.

In each case one phase 11 is now magnetically coupled to at least three further phases 12, 14, 16, to be precise in such a way that the DC components of the individual phases are each compensated for as much as possible by other phases. This reduces the resultant magnetic field, with the result that the design of the coupling means 31 to 39 or of the magnetic circuit now only needs to be substantially for the magnetic field generated by the AC component. As a result, the coupling means 31 to 39, such as coil cores, for example, can be dimensioned so as to be correspondingly small, which results in considerable savings in terms of coupling material, mass and costs. In particular the installation space can thus be greatly reduced.

In addition to the two phases which are adjacent in respect of driving (switch-on/switch-off times), the third phase to be coupled is now preferably selected in such a way that disruptive mutual influencing of the phases is minimized. The selection is performed in such a way that optimum compensation of the DC component is achieved. In this case, it has emerged that, in addition to the adjacent phases (+/−60 degrees phase shift of the switch-on times in the case of six phases; the adjacent phases for the first phase 11 would therefore be the second phase 12 and the sixth phase 16), the phase with a phase shift of 180 degrees (for the first phase 11 this would be the fourth phase 14) is also particularly suitable since a very high degree of elimination of the DC component results there. The two currents through the coupled phases 11, 14 flow in opposition in the seventh coupling means 37. The resultant current Ires for the magnetization of the coupling means 37 is in this case only triggered by the difference in the currents Ires. The DC fields cancel one another out to a large extent. The reduced DC component makes itself positively noticeable for the geometry of the coupling means 31 to 39, which can now manage with a lower volume. In the case of six phases 11 to 16, the coupling shown in FIG. 5 has proven to be particularly suitable.

Magnetic Coupling

In principle, two phases can be coupled magnetically by virtue of the two phases being guided with antiparallel current conduction through a rectangular or annular coupling means 31 to 39. It is essential that the coupling means 31 to 39 is capable of forming a magnetic circuit.

This is possible in the case of a substantially closed structure, which can also include an air gap. Furthermore, the coupling means 31 to 39 consists of a material conducting a magnetic field with a suitable permeability.

The coupling concept on which FIG. 6 is based can be explained by way of example with reference to FIG. 9. It is essential that the phases to be coupled (in FIG. 8 these are the first phase 11 and the second phase 12) are driven with opposite current flow. The respectively corresponding magnetic fields cancel one another out substantially in respect of their DC component, with the result that predominantly only the AC component now contributes to the magnetic field generation. As a result, the corresponding coupling means 31 to 39 can be smaller and it is possible to dispense with an air gap.

Coupling Means Design

The coupling means 31 to 39 are means for inductive coupling, such as, for example, an iron or ferrite core of a transformer on which the phases 11 to 16 to be coupled generate a magnetic field. The coupling means 31 to 39 closes the magnetic circuit of the respective two coupled phases 11 to 16.

The selection of the material for the coupling means 31 to 39 and the permeability does not play such a significant role for the coupling. If no air gap is used, the permeability of the magnetic circuit increases, as a result of which the inductance of the coil becomes greater. As a result, the current increase becomes flatter and the current waveforms come closer to the ideal direct current. The closer the waveforms come to a direct current, the lower the resultant current difference between the two phases which are guided (in opposition) through a core as coupling means 31 to 39. The complexity involved for filters is thus reduced. On the other hand, a system without an air gap has a very sensitive response to different currents between the phases 11 to 16. Although the system is inclined to enter saturation in the case of relatively small current faults, it is still quite stable as a result of the multiple coupling. In principle, air gaps with different dimensions can be selected in order to distribute the losses uniformly among the coupling means 31 to 39. Coupling means 31 to 39 with a lower inductance L also have, in principle, lower power losses.

In order to arrive at a good compromise between high permeability (small air gap->less current ripple) and a high degree of robustness (with air gap->high current ripple), different air gaps can be provided. In this way, the power losses of the coupling means 31 to 39 can also be influenced in such a way that desired criteria, for example uniform distribution of the power losses, are met.

In the exemplary embodiment shown in FIG. 9, the coupling means are provided with an air gap in one of the diagonals (either coupling means 31, 38, 34 or 37, 38, 39). This results in a high level of protection against saturation and, associated therewith, protection against uncontrolled current increase with only three coupling means 31, 38, 34 or 37, 38, 39 with an air gap (which results in a higher current ripple) on all phases 11 to 16. In the case of a large degree of imbalance between the phases 11 to 16 or else in the event of failure of a plurality of phases 11 to 16, only individual coupling means 31 to 42 would enter saturation, but not all coupling means 31 to 39 of one phase at a given current.

A further variant would be to design the coupling means 31 to 39 with different air gaps within the structure. The coupling means (in the exemplary embodiment shown in FIGS. 1, 3 and 5, these are the coupling means with the reference symbols 37, 38, 39) which are subjected to greater, increased magnetization owing to the driving which is phase-shifted through 180 degrees (as arises as a result of coupling of the first phase 11 to the fourth phase 14 by the seventh coupling means 37; coupling of the second phase 12 to the fifth phase 15 by the ninth coupling means 39; coupling of the third phase 13 to the sixth phase 16 by the eighth coupling means 38), could be reduced in terms of their loading, for example, by adaptation or provision of an air gap. This would reduce the total core losses.

In addition, it would be possible in the matrix concept in each row/column to provide a coupling means 31 to 39 with a relatively large air gap or gap. As a result, this coupling means 31 to 39 provided with an air gap would enter saturation first at relatively high currents, with the result that further improved stability in the event of a fault is provided. For reasons of stability, it would be advantageous to guide each phase 11 to 16 through at least one coupling means 31 to 39, which enters saturation later than the other coupling means 31 to 39 in this phase as a result of the provision of a lower inductance L, which could be achieved by the provision of an air gap.

The exemplary embodiment shown in FIG. 11 shows an example of a coupling means 31 provided with an air gap 64. For this, the central limb of the E-shaped first part 44 is designed to be slightly shorter than the outer limbs, with the result that an air gap 64 is produced in the direction toward the second part 43. Alternatively, provision could be made for the limbs of the E-shaped first part 44 to be equal in size, but for an air gap to be provided between the ends of the limbs and the second part 43, for example by means of a nonmagnetic film. Measures which make it possible to achieve the desired inductance L of the respective coupling means 31 to 39, for example by provision of suitable air gap(s) at the suitable points, are customary to a person skilled in the art.

Design of the Phases

The use of only two geometric shapes of the phases 11 to 16 as illustrated in plan view in FIG. 5 is particularly advantageous in terms of manufacturing technology. One basic shape in this case has a U-shaped profile. The second basic shape is substantially rectangular or meandering. The sections shown can be in the form of strip conductors in the form of leadframes, integrated in corresponding conductor tracks in a printed circuit board or embodied as round conductors.

Further magnetic coupling of the individual cores of the coupling means 31 to 39 to form a large total core can result in further savings by virtue of, for example, a single cover plate 43 being provided for all lower parts 44 of the nine coupling means 31 to 39.

Means 80 to 89 for influencing the leakage flux

In FIG. 12, the useful magnetic flux which passes through the ferromagnetic coupling means 43, 44 in each case about two phases 11, 12 to be coupled magnetically is indicated by continuous arrows 92. The magnetic leakage flux is indicated by dashed arrows 94. The magnetic leakage flux passes through non-ferromagnetic material, for example through air or a plastic or insulator surrounding the phases 11, 12. The means 80 for influencing the leakage flux are now designed in such a way that they deflect the leakage flux in a targeted manner between two phases 11, 12 to be coupled. This takes place by projections consisting of ferromagnetic material, which are arranged spatially between the phases 11, 12. These projections are connected to the actual ferromagnetic coupling means 43, 44 in such a manner as to conduct magnetic flux. The projections can be parts of the coupling means 43, 44. However, separate ferromagnetic parts could also be provided.

The means for influencing the leakage flux could have a rectangular cross section. Owing to the particularly simple geometry, such an arrangement can be produced easily and inexpensively.

Alternatively, the means 80 to 89 for influencing the leakage flux can also have a dome-shaped structure. This is understood to mean, rather than a rectangular structure, a structure which tapers toward the end. There could be a continuous, for example parabolic, rounded or circular, transition to the coupling means 44. The domes can be implemented directly during the production process (pressing) of the ferrite cores.

The means 80 for influencing the leakage flux is arranged between the phases 11, 12 to be coupled. Said means could protrude only from one side of the coupling means 44 to the opposite side of the coupling means 43, as shown in FIGS. 9 to 11. Alternatively, the means 80 could protrude from two sides of the coupling means 43, 44 between the phases 11, 12. Preferably, the ends of said means are opposite one another, as indicated in FIG. 8. The ends of the means for influencing the leakage flux are arranged spaced apart from one another, with the result that the magnetic circuit is not closed, but the leakage flux is deflected toward the ends.

Preferably, the means 80 for influencing the leakage flux are arranged on the axis of symmetry of two conductors to be coupled. Preferably, the cross section of the means 80 for influencing the leakage field is also embodied to be axially symmetrical, in relation to this axis of symmetry.

The magnetic leakage flux passes between the ends of the means 80 (FIG. 12) or the end of the means and the coupling means 43, 44 (FIGS. 13 to 15) in a gap length 86 in a nonferromagnetic medium. The signal characteristics of the multiphase converter can be further optimized by virtue of the geometry of the means 80 and the resultant gap length 86.

As shown in FIG. 15, the phases 11, 12 have now been embodied as round wire laid next to one another, instead of a flat wire (one above the other: FIG. 10). This increases the leakage fluxes and interference can be further reduced.

Coupling means 43, 44 and means 80 for influencing a leakage flux have been made from the material 3C95. Furthermore, a gap 96 of the order of magnitude of 1 mm, for example, has been provided. With this selection, the current ripple/current rate of rise could be further reduced. Saturation effects can be eliminated by virtue of this gap 96.

The described multiphase converter 10 is particularly suitable for use in a motor vehicle electrical distribution system, in which in particular dynamic load requirements are of subordinate importance. In particular for such comparatively sluggish systems, the described design is suitable.

In the core model used at present, the leakage flux is set by an additional leakage limb, which is introduced between the two turns. Owing to the design of the core, the response can be adjusted individually to the application. Main parameters are in this case the two air gaps which can be defined in an application-specific manner.

By virtue of the introduction of more than one turn per coupled phase, the core losses which arise as a result of the remagnetization of the core can be greatly reduced. In this case, turns numbers of 2 or 3, or at least in the low range, are expedient in order to keep the turns losses owing to the winding length low.

In the case of the coupling means shown in the figures, more than two turns per phase can be provided.

Claims

1. A multiphase converter comprising a plurality of electrical phases (11 to 16), which are each drivable by one of a plurality of switches (21 to 26), wherein at least one coupler (31 to 39; 100 to 106; 150 to 156) for coupling a phase (11 to 16) to a further phase (11 to 16) is provided, wherein two turns (110 to 116; 120; 122; 130; 132) are provided per phase (11 to 16).

2. The multiphase converter as claimed in claim 1, wherein the multiphase converter is configured for at least two phases.

3. The multiphase converter as claimed in claim 1, in which at least one of the at least one coupler (31 to 39; 100 to 106; 150 to 156) has at least one means (80 to 89) for influencing a magnetic leakage flux.

4. The multiphase converter as claimed in claim 2, in which the means (80 to 89) for influencing the leakage flux is connected either on only one side or on two sides to the at least one coupler (31 to 39; 100 to 106; 150 to 156).

5. The multiphase converter as claimed in claim 2, characterized in that a gap is provided between the means (80 to 89) for influencing the leakage flux and the at least one coupler (31 to 39; 100 to 106; 150 to 156).

6. The multiphase converter as claimed in claim 2, in which the means (80 to 89) for influencing the leakage flux is rectangular.

7. The multiphase converter as claimed in claim 1, in which at least one phase (11 to 16) is U-shaped and the phase (11 to 16) coupled thereto is meandering.

8. The multiphase converter as claimed in claim 1, in which the plurality of switches (21 to 26) drive the phases (11 to 16) sequentially, and the first phase (11 to 16) is magnetically coupled to at least one further phase (11 to 16), which is driven directly beforehand and afterward.

9. The multiphase converter as claimed in claim 1, in which a phase (11 to 16) is magnetically coupled to at least one further phase (11 to 16), which is driven with a phase shift substantially through approximately 180°.

10. The multiphase converter as claimed in claim 2, in which the means (80 to 89) for influencing the leakage flux is dome-shaped.