CH707447B1 - Device for DC voltage conversion for high transmission ratios. - Google Patents

Abstract

The inventive DC-DC converter is used for conversion between a first voltage and at least a second, smaller voltage. In this case, the DC-DC converter at least one series circuit of capacitors (3.1, 3.2, ...), which form a capacitive voltage divider for a first voltage of the DC-DC converter. The DC-DC converter has a balancing circuit with a plurality of balancing modules (6.1, 6.2,...), The balancing modules (6.1, 6.2,...) Each being used to transfer energy between at least two of the capacitors (3.1, 3.2,. are arranged.

Description

Description: [0001] The invention relates to a DC-DC converter according to the preamble of patent claim 1.

In electrical systems, in addition to the power components are other components, e.g. needed for the regulation or control of the system. These components often have operating voltages that are below the rated voltage of the power circuit. The generation of these auxiliary voltages should be done from the nominal voltage of the system with high efficiency.

For the generation of auxiliary voltages for low power from high input voltages usually several stages of switching converters are connected in series, gradually step down to a smaller voltage to avoid small duty cycles of the individual stages. Due to the high voltage load of the first stage (s) IGBTs must be used as switching elements for this purpose, which are limited to low switching frequencies. For applications in limited space systems, it would be advantageous to use MOSFETs with high switching frequencies to reduce the size of the passive components. In addition, it would be advantageous to carry out the generation of the auxiliary voltage directly in one step.

One such device is an Input Series Output Parallel (ISOP) converter, such as known from: Giri, R .; Choudhary, V; Ayyanar, R .; Mohan, N .; "DC-DC converters with active input voltage and load-current sharing", IEEE Transactions on Industry Applications, vol.42, no.4, pp. 1101-1111, July-Aug. Of 2006.

In this case, the voltage load on the input side is distributed to a plurality of series-connected stages, on each of which a DC-DC converter operates with potential-separated output, wherein the outputs of the DC-DC converter are connected in parallel (Fig. 1). The DC-DC converters must be galvanically isolated for this purpose. Control methods are known which ensure a uniform power consumption of each input stage and thus ensure an equal distribution of the input voltage to the series-connected converter inputs. A disadvantage of this topology is that the transformers required for the galvanic isolation must have a dielectric strength up to a maximum of the full input voltage.

Another possible device is the multilevel converter concept, e.g. out

Torresan, H.D .; Holmes, D.G .; Shraga, I .; «Auxiliary power supplies for high voltage converter systems», 35th Annual IEEE Power Electronics Specialists Conference (PESC), vol. 1, pp. 645-651, June 20-25, 2004, and

Peng, F.Z .; Wei, Q .; Cao, D .; "Recent advances in multilevel converter / inverter topologies and applications", International Power Electronics Conference (IPEC), vol., Pp. 492-501, June 21-24, 2010. The advantage of these converters is that they have no magnetic components in their basic structure and are distinguished by a high power density. Possible embodiments include the diode-clamped topology (FIG. 2) and the flying capacitor topology (FIG. 3). The disadvantage of the diode-clamped design lies in the different high voltage load of the diodes. Therefore, depending on the voltage load, several diodes must be connected in series, which, however, leads to higher conduction losses. In addition, it is only possible with the Flying Capacitor design to ensure uniform loading of the individual capacitors by means of redundant switching states. In the diode-clamped version, an additional balancing circuit is required, which ensures a uniform distribution of the input voltage to the in-line capacitors. In the flying capacitor design, n (n + 1) * n / 2 capacitors are required for n stages, so that each capacitor is subjected to the same voltage stress. However, the capacitance of the individual capacitors must be increased by the factor of the number of capacitors connected in series. Instead of the switch 99, diodes can also be used.

Another possibility for implementing a DC-DC DC-DC converter with a large transmission ratio is the use of so-called switched-capacitor circuits (FIGS. 4 and 5) as known from:

Seeman, M.D .; Sanders, S.R .; "Analysis and Optimization of Switched-Capacitor DC-DC-Converters", IEEE Transactions on Power Electronics, vol.23, no.2, pp. 841-851, March 2008.

This topology is based on a capacitive voltage divider, in which the output voltage is tapped by one or more capacitors. In order nevertheless to ensure a uniform distribution of the total voltage to the n series capacitors, (n-1) clamp capacitors are alternately switched back and forth between two adjacent capacitors. This type of wiring is known, for example, from battery systems as a cell charge state equalization circuit (balancing circuit) in which the voltage of series-connected battery cells is leveled. The switched-capacitor circuit can also be made resonant by connecting an inductor in series with each clamp capacitor and equalizing the switching frequency to the resonant frequency. The disadvantage of the switched-capacitor balancing circuit lies in the mandatory synchronous control of the circuit breaker.

The object of the invention is to provide an improved DC-DC converter.

This object is achieved by the DC-DC converter according to claim 1.

The DC-DC converter thus serves to convert between a first voltage and at least a second voltage, which is smaller than the first voltage. In this case, the DC-DC converter has at least one series circuit of capacitors, which form a capacitive voltage divider for the first voltage of the DC-DC converter. Furthermore, the DC-DC converter has a balancing circuit with a plurality of balancing modules, wherein the balancing modules are each arranged for the transfer of energy between at least two of the capacitors. Typically, at least one of the capacitors is connected to a consumer, which draws energy from the capacitor, or an energy source, which feeds energy.

Energy taken or fed to the capacitor results in an imbalance in the voltage distribution across the capacitors. With the balancing modules it is possible to achieve an energy balance and thus a voltage compensation between the capacitors. Local regulation and local balancing of the capacitor voltages, as viewed throughout the DC-DC converter, results in a transfer of energy from one capacitor to another, and overall shifting of the energy from the first voltage side, which may be a high voltage side. to the load on the side of the second voltage, respectively, from a power source on the side of the second voltage toward the side of the first voltage.

In one embodiment, a Balanciermodul each compensating elements, and each of the compensation elements is in each case parallel to selectively one of at least two of the capacitors switchable, in particular by means of power electronic switches. In this case, according to one embodiment, the compensation elements each have an energy storage, in particular inductive energy storage, wherein capacitive and / or battery-based energy storage can be used. The compensating elements thus serve for the intermediate storage of energy which is to be reloaded from one capacitor to another.

In one embodiment, a compensating element (e.g., by means of the mentioned switches) may each be selectively connectable to one of two capacitors of a pair of capacitors, the two capacitors of a pair of capacitors being one after another in the series connection of capacitors. Thus, the stress of the compensation elements and thus the requirement for insulation in the compensation elements can be kept low.

In one embodiment, in each case at a common connection point at which an upper terminal of a lower capacitor and a lower terminal of an upper capacitor of the series circuit are connected to each other, a first terminal of a compensating element is connected, and a second terminal of the compensating element is optional an upper terminal of the upper capacitor and a lower terminal of the lower

Capacitor switchable.

This is a Balancierschaltung in the sense of a switched-capacitor converter (English, "switched capacitor") realized. However, other types of modular active balancers may also be used as DC-DC converters:

In one embodiment, a balancing module has a converter as an inverse converter (ie, without galvanic isolation, buck-boost converter) or as a high-buck converter (ie with galvanic isolation, flyback converter, engl., Flyback converter), which between two the capacitors is connected. This converter thus effects the energy transfer between these two capacitors.

An advantage with such a converter-based topology over a switched-capacitor variant is the simpler control, since no synchronicity of the switching of the transistors of the modules is required.

A drive supply of the power transistors of each Balanciermoduls can be generated locally via appropriate control of the input current of the module. This type of supply can also be applied to other circuits as soon as ripple-prone (ripple-prone) currents are present.

In one embodiment, the compensation elements each have a resonant capacitance and in series to an inductance. In this case, the compensation elements can be inductively coupled to each other, whereby a power transmission between compensating elements can be realized.

By means of the coupled inductors so two Balanciermodule are coupled by means of a transformer and is achieved by resonant operation to minimize the switching losses. In comparison to the described buck-boost topology, fewer, in the best case half as many, inductive components are needed. Furthermore, a higher efficiency results.

In one embodiment, the DC-DC converter has at least one supply unit for supplying power to at least one of the balancing modules, wherein the supply unit is connected in series with the series connection of the capacitors. One or more such supply units can also be implemented without balancing or completely independent of a balancing circuit.

The supply unit may have a series-connected to the capacitors switch and a switch connected in parallel with the supply switch memory circuit having an energy storage and supply diode, for example, a capacitor for smoothing a supply voltage generated in this way.

In one embodiment, the DC-DC converter has a controller which is adapted to vary time portions during which the equalizing elements are connected in parallel to the different capacitors, and thereby to control an energy transfer between the capacitors and thereby in turn the voltages of Regulate capacitors.

In the following, the subject invention based on preferred embodiments, which are illustrated in the accompanying drawings, explained in more detail. Each show schematically:

Fig. 1 shows an embodiment of a known type of DC-DC converter for high input voltage and low output voltage (Input-Series Output-Parallel [ISOP] converter concept).

Fig. 2 shows an embodiment of a known type of DC-DC converter for high input voltage and low output voltage in the form of a multi-level converter in "diode-clamped" design.

Fig. 3 shows an embodiment of a known type of DC-DC converter for high input voltage and low-level output voltage in the form of a multilevel converter in "Capacitor-clamped" version.

Fig. 4 shows an embodiment of a known type of DC-DC converter for high input voltage and low output voltage in the form of a switched-capacitor circuit.

Fig. 5 shows an embodiment of a known type of DC-DC converter for high input voltage and low output voltage in the form of a resonant-switched-capacitor circuit.

Fig. 6 shows a DC-DC converter consisting of balance-boosting elements (buck-boost modules) which each voltage-balance two adjacent capacitors, i. hold at the same voltage.

Fig. 7 Characteristic waveforms of a balance element.

Fig. 8 A device for generating a supply voltage from a ripple-prone current.

Fig. 8a, b embodiments of the DC-DC converter with three Balancierungsgrundelementen with novel devices for generating the supply voltages.

FIG. 9 shows characteristic signal curves of the device shown in FIG. 8.

Fig. 10a, b DC-DC converter with a reduced number of magnetic components.

Fig. 11 Further embodiment of the DC-DC converter with a reduced number of magnetic components.

Fig. 12a, b DC-DC converter with a reduced number of magnetic components and potential separation.

The reference numerals used in the drawings and their meaning are given in the list of reference numerals. Basically, the same parts are provided with the same reference numerals in the figures.

One aspect of the embodiment of the invention is to use a balancing circuit as known from battery management for a high-ratio DC-DC converter. (A balancing circuit does not provide an "output" for the connection of a consumer, but compensates only due to different capacities of the battery cells occurring voltage imbalances). The DC-DC converter is designed so that a first voltage is capacitively divided or multiplied and is tapped by one or more capacitors of the capacitive voltage divider power for the supply of consumers. The use of a balancing circuit on the voltage divider capacitors can then ensure uniform voltage sharing, regardless of the local power tap. a same voltage across all capacitors, to be ensured. The inventive device can also be operated in the reverse power flow direction, in which case a large output voltage is generated as the first voltage from one or more small input voltages as second voltages.

Fig. 6 shows the structure of a DC-DC converter, consisting of a source 1, a capacitive voltage divider 3 with Balancierschaltung and a load 2. The capacitive voltage divider with Balancierschaltung has series-connected capacitors 3.1 to 3.n, between each one inductive element 4.1 to 4.k is connected, which is then connected by means of two switches 5.1a to 5.ka and 5.1b to 5.kb to the ends of two adjacent capacitors. The Balancierschaltung has one or more Balanciergrundelemente or Ba lanciermodule 6.1 to 6.k on. A balancing module 6.1 having in each case three components (an inductive element 4.1 and two switches 5.1a and 5.1b) is repeated at least k = (n-1) times in the case of a capacitive voltage divider with n capacitances. A DC-DC converter with n capacitances thus has k inductive elements and k * 2 switches.

The connection of a load 2 can take place on one or more capacitors 3.1 to 3.n of this voltage divider 3. The control signal of any Balanciermoduls (eg 6.1) for the switches 5.1a and 5.1b is advantageously a symmetrical square wave signal, the two switches 5.1a and 5.1b offset by half a clock period to each other, each with a dead time between the switching off a transistor and the switching on of the next transistor is followed in order to avoid simultaneous conduction of both transistors, which would lead to a short circuit of two voltage divider capacitors safely avoided. The control signals of the various balancing modules 6.1 to 6.k can advantageously be generated independently of each other, the switching frequency either being kept constant or varying depending on the operating point depending on the current flowing to the load 2 and the voltage across the voltage divider capacitors 3.1 to 3.n. can be; the switching frequency of the balancing modules can basically be chosen freely, i. It is not important to pay attention to the same switching frequency or synchronous clocking of the individual balancing modules. An operating mode of the balancing modules with variable switching frequency enables voltage-free switching of the switches and thus loss-minimal operation over the entire range of the first voltage or voltage range of the voltage divider capacitors 3.1 to 3.n and the entire range of load current through load 2. The de-energized switching can then take place when the current direction in the inductive elements reverses in each clock period. It is then a sign-correct current for the resonant transhipment of the parasitic capacitances of the switching elements, which are to be considered as lying parallel to the switching elements 5.1a to 5.ka and 5.1b to 5.kb, available; the period of Umschwingvorgangs occurring during the transhipment is defined by the inductance and the sum of the capacities of the two switches of a Balanciermoduls.

7 shows the characteristic current profile of one of the balancing modules 6.1 to 6.k using the example of the balancing module 6.1 by way of example. It is shown - over a first time axis 7, the current waveform 10 in the inductive element 4.1 and the average current value 13th - over a second time axis 8, the drive signal 11 for switch 5.1a - on a third time axis 9, the drive signal 12 for switch 5.1b. During the dead time, neither of the two switches 5.1a and 5.1b is activated and the parasitic capacitances of the switching elements are transposed. Once the DC-DC converter is loaded by a load 2, the current waveform 10 shows an average value 13 with respect to the clock period.

The power supply required for the control of a Balanciermoduls electronics can either be done galvanically isolated from an external source, or can be done directly in a Balanciergrundelement by an additional circuit. The latter variant is advantageous for a modular design of the overall system, since no external sources are needed. In addition, it is also possible for the control signals of the individual switching elements to generate them centrally in an electronic unit and to control the switching elements synchronously by means of galvanic isolation.

Fig. 8 shows an internal supply unit 19, as used in the DC-DC converter of Fig. 6, but can also be used in other structures. In Fig. 8 for ease of understanding, a DC-DC converter of the type of Fig. 6 with only two voltage dividing capacitors or capacitors, ie an upper capacitor 3.1 and a lower condenser 3.2, and a Balanciermodul with a first switching element 5.1a and a second switching element 5.1b and an inductive element 4.1 shown. The operating principle is based on that a capacitor, for example, the lower capacitor 3.2, which is arranged parallel to the load 2 with a constant load current, the ripple portion of a ripple-like current, as is typically generated by switching power supplies with inductive components absorbs. The supply unit 19 has a switch or supply switch 17, which is arranged in series with the voltage divider capacitors 3.1 and 3.2. In addition, a feed diode 15 and a capacitor 16 are connected in series as an energy store and connected in parallel to the feed switch 17, in such a way that the feed diode 15 is poled in the reverse direction when the supply switch 17 is turned on. The voltage applied to the energy store 16 can be used as supply voltage, e.g. a control electronics, are used. Between the source 1 and the upper capacitor 3.1, a source diode 18 is required because when switching off the supply switch 17, the voltage of the energy storage 16 is added to the voltage of the two capacitors 3.1, 3.2 and then without source diode 18, a return current would flow into the source 1 ,

In addition, a controller for controlling the voltage to energy storage 16 is required, which is advantageously designed as a hysteresis and controls the supply voltage via appropriate control of the supply switch 17. FIG. 9 shows characteristic signal profiles of a supply unit and the mode of operation of the voltage regulation of the energy store 16, for example over horizontal time axes, i. - Over a first time axis 20, the current waveform 24 of the inductive element 4.1 and the constant load current 25 through the load 2; over a second time axis 21, the voltage actual value 27 of the energy store 16 and the upper 26a and lower 26b hysteresis limit; - Over a third time axis 22, the current waveform 28 through the dining diode 15 and the energy storage 16; - Via a fourth time axis 23, the drive signal 29 for the supply switch 17th

For ease of understanding, the current waveform 24 is shown without resonant transients, i. E. no dead time is used between the drive signals for switches 5.1a and 5.1b. In addition, the load current 25 through

Load 2 is shown as being ideally constant and it is assumed that the energy store 16 is e.g. is loaded by a control electronics (this load is not shown).

The controller operates in such a way that the supply switch 17 is turned off when the supply voltage 27 at the energy storage 16 falls below a lower hysteresis limit 26b and is turned on when the upper hysteresis limit 26a is exceeded. The voltage 27 at the energy store 16 increases as soon as the feed switch 17 is switched off and the first switching element 5.1b is switched on at the same time and the current 24 through the inductive element 4.1 is greater than the load current 25 and therefore current flows into the energy store 16. It can be generated for each Balanciermodul easily and independently of other Balanciermodulen a supply voltage of a local control electronics and the drive circuits of all switches of Balanciermoduls.

For an expanded understanding, Fig. 8a shows an embodiment of the DC-DC converter of Fig. 6 with, by way of example, three balancing modules, each having the internal supply unit shown in Fig. 8, i. The system shown has several internal supply units 19.1 to 19.3. Fig. 8b shows an alternative embodiment of the circuit of Fig. 8a. Advantageously, the source diode 18 can be omitted here. However, additional start-up units 31.1 to 31.3 are needed for the start-up of the circuit to ensure that the energy stores 16.1 to 16.3 are charged to a minimum voltage, i. each with a start-up unit 31.i each one of the energy storage 16.i is loaded. For further operation, the start-up units 31.1 to 31.1 are no longer needed and are advantageously used again only when an energy store 16.i falls below its minimum voltage. An internal structure of a start-up unit 31.i can be designed in many ways, according to its function to maintain a minimum voltage in the associated energy storage.

An alternative to the new type of DC-DC converter shown in Fig. 6 is shown in Fig. 10a. Advantageously, this requires a lower number of inductive components and switching elements. The structure is again based on a capacitive voltage divider with voltage divider capacitors 3.1 to 3.m, wherein in the simplest case there are four series-connected voltage divider capacitors 3.1, 3.2, 3.3, 3.4; The following description is limited to such a simplest arrangement. Between the capacitors 3.1 and 3.2, a resonant capacitor 30.1a branches off, the second end of which is connected to an inductance 60.1a, which is magnetically coupled to an inductance 60.1b, as represented by the coupling 65.1. The second end of the inductance 60.1a is connected to the series connection of two switches 50.1a and 50.1b, the switch 50.1a branching off from the upper terminal of the first voltage divider capacitor 3.1 and switch 50.1b to the connection point of the second voltage divider capacitor 3.2 to the third voltage divider capacitor 3.3 is connected.

Furthermore, a resonant capacitor 30.1b branches off between the capacitors 3.3 and 3.4, whose second end is connected to an inductance 60.1b (which, as mentioned above, is magnetically coupled to an inductance 60.1b). The second end of the inductor 60.1b is connected to the series connection of two switches 50.2a and 50.2b switches, wherein the switch 50.2a branches off from the upper terminal of the first voltage divider capacitor 3.3 and switch 50.2b is connected to the lower terminal of the second voltage divider capacitor 3.4.

Through this interconnection coupled coils occurs over each coil or between the coils realizing and located on a magnetic core turns only a certain number of stages of the voltage divider fraction of the total voltage. The isolation voltage can thus advantageously be lower than for systems in which directly starting from the total voltage of the source 1, e.g. by means of a flyback converter (which also has two magnetically coupled coils as an inductive element), the voltage required at the load 2 is generated, can be selected.

According to the invention, the presented circuit can be modified in that any coils are magnetically coupled. For example, a magnetic coupling of four each (instead of, as described above, of only two) coils is possible, or all coils can be magnetically coupled. In both cases, the number of required magnetic cores is advantageously reduced.

For voltage dividers with m capacitors (m even and greater than or equal to 4) are p = (((m-4) / 2) +1) coupled inductors, as many resonant capacitors and m switch needed. The number of resonant capacitors can be halved by setting a half capacitance resonant capacitor only in series with a coupled inductor terminal. In the topology shown in Figure 10a, the load can only be tapped off at the connection points between the voltage divider capacitors where there is no series connection of a resonant capacitor and a coupled inductor. In Fig. 10b, a variant is shown, in which the load tap can also be done at the connection points of the voltage divider capacitors, in which there are coupled inductances. For this purpose, the resonance capacitor on the other side of the coupled inductance must have half the capacitance value.

The control signals of the switches are synchronized in such a way that two adjacent switches each receive a drive signal offset by half a clock period. The drive signal is a symmetrical square wave signal with 50% duty cycle, the derived control of the switch advantageously has the dead time described above.

In a practical implementation of the invention is to be noted that the coupled inductors in practice do not have an ideal coupling. As a result of this non-ideal coupling (coupling factor less than 1), the current flowing through the coupled inductors generates a stray field in addition to the main field. For illustrative purposes, an inductance, hereinafter called stray inductance, may be thought of as being connected in series with the coupled inductor, which is responsible for the generation of the stray field, e.g. between 30.1a and 60.1a. In an inventive operation of the circuit, the stray inductances of two coupled inductors together with the resonant capacitors in series form a resonant circuit with a specific resonant frequency. For an advantageous operation of the circuit according to the invention, the switching frequency of the switches 50.ka or 50.kb is to be chosen such that it corresponds to the resonant frequency. According to the invention, the circuit can also be modified in such a way that, instead of discrete resonance capacitors, the parasitic capacitances of the switching elements which are parallel to them can be used.

The flocculation process of the circuit according to the invention can be carried out so that the voltage of the source 1 is gradually increased to nominal voltage. During this process, the load 2, which may be e.g. a DC-DC converter can remain switched off and the voltage of the source 1 is divided according to a capacitive voltage divider on the capacitors 3.1 to 3.m on. After reaching a minimum voltage, the switches of the balancing modules can be clocked according to the invention and the load can be connected. In order to prevent overvoltages occurring at the capacitors of the voltage divider 3 during the run-up operation, so-called suppressor diodes (TVS) can be connected in parallel with the individual voltage divider capacitors, which serve to limit the voltage.

Alternatively, the new type of DC-DC converters of Fig. 6 and Fig. 10a / b can also be modified so that the source 1 is connected to only a portion of the existing voltage divider capacitors 3.i, as shown in Fig. 11. In addition, the DC / DC converter shown in FIGS. 10a / b and 11 can have one or more potential separations, such that the potential-separated regions of the DC / DC converter are connected to one another only via an inductive or magnetic coupling 65.i, as shown in FIG. 12a / b to see.

List of Reference Numerals [0045] 1 source 2 load 3.1 voltage divider capacitors 4.1 inductive element 5.ii switching element 6.1 balancing element 7 time axis 8 time axis 9 time axis 10 current curve 11 switching signal 12 switching signal 13 average current value 14 dead time 15 diode, supply diode 16 capacitor, energy storage 17 switching element, supply switch 18 diode, Source diode 19 Supply unit 20 Timeline 21 Timeline

Claims (10)

22 Time axis 23 Timeline 24 Current profile 25 Average current value 26a Upper hysteresis limit 26b Lower hysteresis limit 27 Voltage curve 28 Current profile 29 Switching signal 30. ii Resonance capacitor 31.1 Startup unit 50.N Switching element 60.ii Coupled inductance 65.1 Magnetic coupling Patent claims 1. DC-DC converter for converting between a first voltage and at least a second voltage, which is smaller than the first voltage, wherein the DC-DC converter at least one series circuit of capacitors (3.1,3.2, ...), which comprises a capacitive voltage divider for a first voltage of the DC-DC converter, and characterized in that the DC-DC converter comprises a balancing circuit with a plurality of balancing modules (6.1, 6.2,...) and the balancing modules (6, 6, 6,...) each for transferring energy between at least two of the capacitors ( 3.1,3.2, ...) are arranged. 2. DC-DC converter according to claim 1, wherein as a component of the DC-DC converter to at least one of the capacitors (3.1,3.2, ...) a consumer (2) is connected, which the capacitor (3.1,3.2, ...) draws energy, or an energy source, which feeds energy into the condenser (3.1, 3.2, ...). 3. DC-DC converter according to claim 1 or 2, wherein the plurality Balanciermodule (6.1,6.2, ...) each have compensation elements and each of the compensation elements in each case parallel to optionally one of at least two of the capacitors (3.1, 3.2, ...) is switchable , 4. DC-DC converter according to claim 3, wherein the compensation elements each have an energy storage, in particular inductive energy storage (4.1.4.2, ..., 60.1a, 60.1b, ...). 5. DC-DC converter according to claim 3 or 4, wherein a compensating element is in each case selectively connectable to one of two capacitors of a capacitor pair (3.1, 3.2, 3.2, 3.3, 3.3, 3.4, ...), wherein the two capacitors of a capacitor pair are such which aufeinan follow in the series connection of capacitors. 6. DC-DC converter according to claim 5, wherein in each case at a common connection point at which an upper terminal of a lower capacitor (3.2) and a lower terminal of an upper capacitor (3.1) of the series circuit are connected to each other, a first terminal of a compensating element (4.1) connected a second terminal of the compensating element (4.1) is selectively switchable to an upper terminal of the upper capacitor (3.1) and to a lower terminal of the lower capacitor (3.2), wherein the upper capacitor and the lower capacitor capacitors from the amount of in Claim 1 mentioned capacitors. 7. DC-DC converter according to one of claims 3 to 6, wherein the compensation elements each have a resonant capacitance (30.1a, 30.1b, ...) and in series an inductance (60.1a, 60.1b, ...) have. 8. DC-DC converter according to one of claims 3 to 7, wherein compensating elements (30.1a, 60.1a, 30.1b, 60.1b, ...) are inductively coupled together. 9. DC-DC converter according to claim 1 or 2, wherein the plurality Balanciermodule each having an inverter converter or a buck-boost converter, which is connected between two of the capacitors. 10. DC-DC converter according to one of the preceding claims, comprising at least one supply unit (19) for supplying power to at least one of the balancing modules (6.1, 6.2, ...), wherein the supply unit in series for series connection of the capacitors (3.1, 3.2, ... ), and in particular the supply unit has a supply switch (17) connected in series with the capacitors (3.1, 3.2,...) and a memory circuit (15, 16) connected in parallel with the supply switch (17) with an energy store (16) ,