Multi-phase DC-DC converter and method for controlling a multi-phase DC-DC converter

Abstract

In a control method for controlling a multi-phase DC-DC converter (10) having two or more phase circuits (20), each phase circuit (20) has a switch (24) arranged to control an inductor current I L through an inductor (21), the phase circuits (20) being arranged to generate phase currents contributing to a total current to be delivered to an output side of the multi-phase DC-DC converter (10). The method comprises: • switching the two or more phase circuits (20) in a boundary conduction mode (BCM) to generate interleaved phase current pulses having a period length T, a nominal on-time period t on for the switches (24); • repeatedly adjusting the on-time period t on in at least one of the switched two or more phase circuits (20), and for successive phases, to control a pulse length of the inductor current I L to minimise a difference from the period length T.

Description

Technical Field

[0001] This invention relates to the field of power converters, and more particularly to a multiphase DC-DC converter and a method for controlling the multiphase DC-DC converter. Background Technology

[0002] Figure 1 It shows a method for using an input voltage U E Input capacitor C E and input current I E The input side has an output voltage U A Output capacitor C A and output current I A In existing boost converters with unidirectional power transfer on the output side, the inductance value of the inductor or choke is L. S Carrying inductor current I L A diode has a diode current I D There is a switch between the bridging point and the common terminal, and the voltage between the bridging point and the common terminal is U. S This switch is a semiconductor switch, such as a MOSFET switch. The input capacitor filters the inductor current, making the input current the average value of the inductor current. The output capacitor filters the diode current, making the output current the average value of the diode current.

[0003] Figure 1 The relevant voltage and current over time are also shown, illustrating the well-known continuous conduction mode (CCM), where the inductor current remains above zero. In the discontinuous conduction mode (DCM) (not shown), during the time the switch is off (U... S (t) is essentially equal to U A The inductor current drops to zero and remains zero for a period of time until the switch closes again. Closing the switch clamps the voltage at the bridge point to zero, causing the inductor current to rise again. In boundary conduction mode (BCM, also known as transient mode), the goal is to close the switch when the inductor current drops to zero, thus eliminating the period during which it remains zero. To control the circuit in boundary mode, it is necessary to measure the inductor current.

[0004] These converters are known to be combined in parallel to create a multiphase DC-DC converter. Individual phases must be controlled to generate and add interleaved current pulses on the output side to reduce ripple in the resulting output current. In the CCM, the current in the different inductors must also be controlled to ensure uniform current distribution across multiple phases and to prevent current from rising to arbitrarily large values. Depending on the requirements for current dynamics, the corresponding current sensor in each phase must have a correspondingly high bandwidth. Switching losses occur when the switches are turned on and off.

[0005] US 7'884'588 B2 discloses a DC-DC converter with two or more phases and addresses the problem of phases having different electrical parameters: in this case, if each phase operates in BCM mode, they will operate at different switching frequencies, and the ripple current resulting from the sum of all phase currents will vary considerably. The proposed solution is to identify the phase with the lowest switching frequency when operating in BCM mode and operate the remaining phases at the same frequency in DCM mode. Alternatively, to reduce capacitor losses, US 7'884'588 B2 suggests delaying the power switch turn-on time for each phase. This comes at the cost of increased input current ripple, as the phase shift between the individual input currents is no longer optimal.

[0006] Multiphase DC-DC converters can be like Figure 2 The diagram illustrates a construction using, for example, two or more phase circuits that generate interleaved current pulses. Summary of the Invention

[0007] Therefore, one object of the present invention is to create a multiphase DC-DC converter and a method for controlling the multiphase DC-DC converter, which overcomes one or more of the above-mentioned disadvantages.

[0008] In a method for controlling a multiphase DC-DC converter, the multiphase DC-DC converter is arranged to exchange power between the input and output sides, and the multiphase DC-DC converter includes two or more phase circuits.

[0009] Each phase circuit includes a switch and an inductor, the switch being arranged to control the inductor current I through the inductor. L ,

[0010] The phase circuit is arranged to generate a phase current that contributes to the total current to be supplied to the output side of the multiphase DC-DC converter. The method includes:

[0011] • In boundary conduction mode (BCM), two or more phase circuits are switched to generate interleaved phase current pulses with a period length of T, and the nominal on-time period of the switch is t. on ;

[0012] • In at least one of the two or more phase circuits being switched, and for consecutive phases, the on-time period t is repeatedly adjusted. on To control the inductor current I L The pulse length is adjusted to minimize the difference from the period length T.

[0013] Therefore, ripple in the output current can be reduced by interleaving the phase current pulses generated by the phase circuits, while operating the required number of phase circuits in boundary conduction mode (BCM). In contrast to DCM operation, BCM operation reduces switching losses and EMC interference. Furthermore, unlike CCM operation, it eliminates the problem of current rising to arbitrarily large values.

[0014] The period length T and the start-up time period t of each phase circuit can be adjusted. on This is to ensure that the sum of the interleaved phase currents matches the required total current. The period length T of all phase circuits should be the same because their current pulses are to be interleaved. If the phase circuits have the same electrical parameters, the on-time period t... on They will be the same. In fact, changes in these parameters, especially the inductance, will cause the phase currents in the phase circuits to have different rates of change. If their turn-on periods t... on Operating with the same value will again prevent them from operating under BCM. Individually control the on-time period t of each phase. on Allow BCM operations to be performed, even during the initial or nominal start time period t. on It will not cause the inductor current I L It returns precisely to zero after a period length T - because it should theoretically be.

[0015] In subsequent switching cycles, the switching time period t is adjusted based on the deviation between the actual zero-crossing time and the expected zero-crossing time (determined by the cycle length T). on This synchronizes the phase circuits.

[0016] In one embodiment, the control method includes determining a period length T by operating one of the phase circuits referred to as the master phase, which has an on-time period t determined based on the average current to be delivered by the phase circuit. on And operate one or more of the remaining active phase circuits to adapt their timing and period lengths to the timing and period lengths of the main device phase. That is, the remaining active phase circuits adjust their on-time period t. on This is to ensure that they reach the cycle length T determined by the main unit phase. Alternatively, the controller determines the prescribed cycle length T and the on-time period t based on the average current to be delivered by each activated phase circuit. on The initial value is determined, and all active phase circuits are operated to adjust their turn-on time period t. on To achieve the specified period length T.

[0017] Given the average current to be delivered, the cycle length T and the turn-on time period t in each phase circuit can be determined by calculation. on Use this period length T and the start time period t.on The control phase circuit will ideally generate the average current to be delivered, eliminating the need for current measurements in each phase. This eliminates the need for current sensors that return a quantitative current measurement, rather than sensors that only detect whether a threshold has been exceeded.

[0018] In reality, the actual average current may not have the exact value specified for the average current to be delivered. However, this will cause a control deviation in the signal affected by the average current, and the controller of that signal can adjust the average current to be delivered accordingly. Typically, this can be accomplished with a controller that has an integral part.

[0019] The supervisory control loop can be input with the total current to be delivered, for example from a total current setpoint, corresponding to the required total current to be delivered by the multiphase DC-DC converter. If the total current setpoint is not reached, the supervisory control loop can adjust the total current to be delivered. This can compensate for deviations between the actual parameters of the phase circuit and their nominal values.

[0020] When operating the phase circuit under BCM, the corresponding switch is turned on, thereby increasing the inductor current I. L Immediately after returning to zero, a current pulse is initiated at the zero-crossing time. There is no inductor current I. L A substantial time period of zero, as in the case of discontinuous conduction mode (DCM).

[0021] The number of phase circuits to be activated can be selected based on the total current required to be delivered to the output side and the maximum and minimum currents that each phase circuit can deliver. The minimum current depends on the voltage ratio, the maximum permissible frequency, and the possible shortest turn-on time period t. on The maximum current depends on the inductor, the upper branch switching unit, especially the diodes in the branch switching unit, and the maximum current the switch can carry, with a safety margin, for example, for controllability. Within these ranges, the number of active phase circuits can be selected. Furthermore, the current in each phase circuit can be selected to be within the range optimal for circuit efficiency or other criteria.

[0022] The exact method for determining the number of phase circuits to be activated is not within the scope of this invention. This invention solves problems such as operating a given number of activated phase circuits and changing the number of activated phase circuits.

[0023] In an embodiment, a multiphase DC-DC converter is arranged to exchange power between an input side including a first input terminal and a second input terminal and an output side including a first output terminal and a second output terminal.

[0024] A multiphase DC-DC converter includes two or more phase circuits.

[0025] Each phase circuit includes an inductor connected between the first input terminal and the bridge point, an upper branch switch unit connected between the bridge point and the first output terminal, and a switch with a parallel freewheeling diode connected between the bridge point and the second input terminal and the second output terminal.

[0026] The control method includes the following steps:

[0027] • Determine the number N of phase circuits to be activated;

[0028] • Determine the start time period t on The initial value;

[0029] • Determine the period length T;

[0030] For at least one of the phase circuits to be activated;

[0031] • Turn on the switch of the phase circuit to make the inductor current I L The number of switches increases over time, and the target on time when the switch should be turned on next is determined.

[0032] • During the activation time period t on Then, turn off the switch, causing the inductor current I to... L The flow passes through the upper branch switch unit and decreases over time;

[0033] • After the current through the inductor returns to zero, at the zero-crossing time, turn the switch on again and repeat the above steps;

[0034] • When repeating the above steps, if the zero-crossing time is before the target start time, then increase the start time period t. on To control the start-up time period t on If the zero-crossing time is after the target activation time, then the activation time period t is reduced. on To control the start-up time period t on .

[0035] In the embodiment, the start-up time period t is determined. on The initialization steps include calculating the on-time period t. on This makes the average current I through the inductor with inductance L... mean It equals a given value. This value can be determined by a supervisory control loop, for example, from the total current setpoint, corresponding to the required total current delivered by the multiphase DC-DC converter. In subsequent operation of the converter, the turn-on time period t is adjusted. on The initial value can be used for feedforward control. This can improve dynamic performance when the required current changes.

[0036] In the embodiment, the turn-on time period t of the phase circuiton The initial value is calculated as

[0037]

[0038] Among them, U IN The input voltage is L, the inductance of the inductor is I, and the input voltage is L. mean It is the average current that is delivered by the phase circuit.

[0039] In this way, during converter operation, the average current to be delivered can be approximated by quantitative measurement without actually measuring the inductor current or other currents. If the actual inductance (unknown) matches the inductance value L of the inductor (a nominal value stored in the controller), then the actual average current will substantially match the average current to be delivered. If not, the discrepancy can be eliminated by specifying the average current to be delivered in the controller.

[0040] In the embodiment, in the activated phase circuit, the length T of each cycle is calculated as follows:

[0041]

[0042] U OUT It is the voltage on the output side.

[0043] Given that the target is to operate within a BCM, the resulting cycle length T is the measurable actual input and output value, as well as the conduction time period t. on The function, and the conduction time period t on It is also a function of the average current to be transmitted.

[0044] In an embodiment, the step of determining the target turn-on time includes calculating the target turn-on time as an offset relative to a reference time, which is the period length T divided by the number of activated phase circuits N.

[0045] The target start time for the next cycle can be calculated before or after the switch is turned on, depending on the environment. The reference time can be determined by the main unit or by a controller that determines the specified cycle length T, as described above.

[0046] In this embodiment, the step of turning the switch back on includes one of the following:

[0047] • Monitor the voltage across the switch and open the switch when the voltage across the switch is zero;

[0048] • Monitor the voltage across the switch and turn on the switch after a predetermined delay when the voltage across the switch falls below a predetermined threshold;

[0049] • Monitor inductor current I L And in the inductor current IL The switch is opened when the current becomes negative due to the reverse current through the upper branch switch unit and then becomes zero.

[0050] • Monitor the current through the upper branch switch unit, especially the diode current, and turn on the switch when the current becomes zero after it becomes negative due to the reverse current through the upper branch switch unit.

[0051] This allows the use of qualitative signals to determine at least the time of near-zero voltage switching, i.e., to determine when the value exceeds a threshold.

[0052] In this embodiment, the inductor current I is measured by measuring the magnetic field of the inductor. L For example, this can be accomplished using a Hall element.

[0053] In this embodiment, the step of turning the switch back on includes:

[0054] • When the inductor current I L When the current returns to zero, its direction is reversed, and it continues to flow through the inductor and the upper branch switch unit until the upper branch switch unit is closed, and the inductor current I... L Commutation to freewheeling diode;

[0055] • Turn on the switch.

[0056] This allows for zero-voltage switching of the switch.

[0057] In one embodiment, the upper branch switch unit is composed of, or includes, a diode and is turned off by a reverse charge in the diode, wherein the reverse charge of the diode is selected such that the reverse current through the diode is sufficient to discharge the capacitor between the bridge point and the second input terminal.

[0058] In one embodiment, the upper branch switching unit is composed of, or includes, a diode. The diode acts as a passive switch. It turns on and off based on the current flowing through it. In such an embodiment, the upper branch switching unit can be turned off by a reverse charge established in the diode. The reverse charge of the diode can be selected such that the reverse current through the diode is sufficient to discharge the capacitor between the bridge point and the second input terminal.

[0059] In this embodiment, the upper branch switch unit is composed of, or includes, an active switch. The active switch is turned on and off according to the state of the control signal. In such an embodiment, the upper branch switch unit can be actively turned off when the capacitor between the bridging point and the second input terminal has been discharged.

[0060] In the embodiment, when the inductor current I LAfter the current flows forward through the freewheeling diode, it reverses its direction, is driven by the input voltage, and returns to zero, while the voltage across the switch is at least approximately zero. In addition to zero-voltage switching, this also allows for zero-current switching.

[0061] In an embodiment, to increase the number N of activated phase circuits to N+1 given a total current setpoint, the method includes:

[0062] • During the transition period of length Tpi, the switches of N phase circuits are switched to be turned on at turn-on times 0, dTpi, 2·dTpi, 3·dTpi … (N-1)·dTpi relative to this period, where dTpi = Tpi / N;

[0063] • For a transition period of length Tsi, the target turn-on time of N+1 phase circuits relative to this period is calculated as 0, dTsi, 2·dTsi, 3·dTsi … N·dTsi, where dTsi = Tsi / (N+1);

[0064] • During the transition period, the switches of N phase circuits are switched to turn on at the same time as in the previous transition period relative to this period.

[0065] • During the transition period, for each of the N phase circuits, set the turn-on time period t. on This ensures that the current returns to zero at the corresponding target on-time during the transition period;

[0066] • During the transition period, after the start time of (N-1)•dTpi, the (N+1)th new operation is started with an start time period t. on The switching minimizes (at least approximately) the deviation between the total current of all phase circuits and the total current setpoint during the transition and post-transition periods.

[0067] This allows for the inclusion of an additional phase circuit in the operation of a multiphase DC-DC converter with minimal impact on the quality of the total current. This, in turn, can be part of adapting to load variations: if a higher total current is required, the additional phase circuit can be activated in this way, which initially reduces the average current delivered by each phase circuit. Subsequently, the average current in each phase circuit can be increased, thereby increasing the total current.

[0068] In an embodiment, in order to reduce the number N of activated phase circuits to N-1 given a total current setpoint, the method includes:

[0069] • During the transition period of length Tpd, the switches of N phase circuits are switched to be turned on relative to this period at turn-on times 0, dTpd, 2·dTpd, 3·dTpd … (N-1)·dTpd, where dTpd = Tpd / N;

[0070] • For a transition period of length Tsd, the target turn-on time of N-1 phase circuits relative to this period is calculated as 0, dTsd, 2·dTsd, 3·dTsd … (N-2)·dTsd, where dTsd=Tsd / (N-1).

[0071] • During the transition period, the switches of N phase circuits are switched to turn on at the same time as in the previous transition period relative to this period.

[0072] • During the transition period, for each of the N phase circuits, except for the phase circuit with a target turn-on time of dTpd relative to that period, the turn-on time period t is set. on This ensures that the current returns to zero at the corresponding target on-time during the transition period;

[0073] • During the transition period, for a phase circuit with a target on-time of dTpd relative to that period, the on-time period t of the last pulse of that phase circuit is set. on This minimizes (at least approximately) the deviation between the total current of all phase circuits and the total current setpoint during the transition period and the post-transition period.

[0074] This allows for the removal of phase circuits during the operation of a multiphase DC-DC converter with minimal impact on the quality of the total current. Similarly, this can be part of adapting to load variations to provide a lower total current: first, the average current in each active phase circuit decreases, then one of the phase circuits is deactivated, which increases the average current in each of the remaining phase circuits.

[0075] In an embodiment, for the transition of a multiphase DC-DC converter between discontinuous conduction mode and boundary conduction mode operation, for one or more pairs of phase circuits, the method includes:

[0076] • Operate the corresponding two phase circuits of the pair to generate alternating current pulses of the same shape, each of the two phase circuits generating a zero-current pulse and a period of equal length, the period length T.

[0077] • Turn off one of the two corresponding phase circuits, and generate a current pulse with a period length of T by the other of the two phase circuits in boundary conduction mode.

[0078] This allows for a switch from DCM to BCM, thereby reducing the number of active phase circuits by half. If the number of phase circuits is large enough, the same principle can be applied to three times or more.

[0079] In an embodiment, the method includes, for the transition of a multiphase DC-DC converter between discontinuous conduction mode and boundary conduction mode operation, for each of one or more sets of np-phase circuits, where np is two or more, for each of these sets

[0080] • In discontinuous conduction mode, each np phase circuit of the group is operated to generate a sequence of np pairs of adjacent current pulses with a period length of T, and each phase circuit contributes to one of the current pulses in the sequence.

[0081] • Instead of operating the individual np phase circuits of the group in discontinuous conduction mode, operate one phase circuit in boundary conduction mode to continue the sequence of current pulses with a period length of T.

[0082] The phase circuit that ultimately operates in the boundary conduction mode can be one of the phase circuits that initially operated in the discontinuous conduction mode, or the other.

[0083] Several such sequences can be interleaved, each sequence being generated by a set of such np-phase circuits.

[0084] In an embodiment, for the transition of a multiphase DC-DC converter between boundary conduction mode and discontinuous conduction mode operation, for one or more pairs of phase circuits, the method includes:

[0085] • Do not operate one of the two corresponding phase circuits, and generate a current pulse with a period length of T by the other of the two phase circuits in boundary conduction mode;

[0086] • At the point when the current is zero, the discontinuous conduction mode is switched by operating the corresponding two phase circuits of the pair to generate alternating current pulses of the same shape. Each of the two phase circuits generates a zero-current pulse and a period of equal length, which is the period length T.

[0087] This allows a switch from BCM to DCM, thereby doubling the number of active phase circuits. If the number of phase circuits is large enough, the same principle can be applied to triple or more.

[0088] In an embodiment, the method includes, for the transition of operation of a multiphase DC-DC converter from a boundary conduction mode to a discontinuous conduction mode, for each of one or more sets of np-phase circuits, where np is two or more, for each of these sets

[0089] • Operate a phase circuit in boundary conduction mode to generate a sequence of current pulses with a period length of T in boundary conduction mode;

[0090] • Instead of operating one phase circuit in boundary conduction mode, the individual np phase circuits in the group are operated in discontinuous conduction mode to generate a sequence of np pairs of adjacent current pulses of period length T, thereby continuing the sequence of current pulses of period length T, with each phase circuit contributing to one of the current pulses in the sequence.

[0091] The phase circuit that first operates in boundary conduction mode can be one of the phase circuits that subsequently operate in discontinuous conduction mode, or the other.

[0092] The multiphase DC-DC converter includes a controller that includes voltage sensors arranged to determine the voltage U on the input side of each phase circuit. IN Output voltage U OUT and the voltage U across the inductor L The controller is configured to perform the methods described herein.

[0093] In this embodiment, the multiphase DC-DC converter and controller do not require measuring the current through the inductors and / or switches of the respective phase circuits.

[0094] In an embodiment, in a multiphase DC-DC converter, in at least one phase circuit, the upper branch switching unit includes or is composed of diodes, the diodes having a sufficiently large current to withstand the inductor current I. L After returning to zero, the inductor current I is reversed. L The reverse recovery time is such that the reverse current discharges the capacitance of the switch, freewheeling diode, and parallel capacitor (if present) before the switch is turned on.

[0095] Even considering the inherent capacitance in the switch and / or freewheeling diode, and more generally the capacitance that can be connected in parallel to the switch, this allows for zero-voltage switching of the switch, for example, to reduce switching losses.

[0096] In an embodiment, the reverse recovery time is large enough that the reverse current also discharges the capacitor arranged in parallel with the switch before the switch is turned on.

[0097] The characteristics of a method can be combined with the characteristics of a device, and vice versa. Attached Figure Description

[0098] The subject matter of the invention will be explained in more detail below with reference to exemplary embodiments shown in the accompanying drawings, which schematically illustrate:

[0099] Figure 1This illustrates typical values ​​in existing DC-DC boost converters and CCM operation;

[0100] Figure 2 This illustrates a multiphase DC-DC converter;

[0101] Figure 3 This shows the interlacing of current pulses;

[0102] Figure 4 Showing the inductor current I L Different trajectories;

[0103] Figure 5 This shows the relevant time points when controlling a multiphase DC-DC converter;

[0104] Figure 6 The voltage and current traces of the switches in the phase circuit are shown;

[0105] Figure 7 This illustrates activating the additional phase circuit and inserting its current pulse into an interleaved pulse sequence;

[0106] Figure 8 This demonstrates disabling the phase circuit and removing its current pulses from the interleaved pulse sequence;

[0107] Figure 9 This illustrates the switching between DCM and BCM, and vice versa; and

[0108] Figure 10 A method for controlling a converter in a DCM is shown.

[0109] In principle, the same parts have the same reference numerals in the drawings. Detailed Implementation

[0110] Figure 2 A multiphase DC-DC converter 10 is shown. On the input side, it includes a first input terminal 11 and a second input terminal 12, with an input capacitor 13 disposed between the two input terminals. On the output side, it includes a first output terminal 14 and a second output terminal 15, with an output capacitor 16 disposed between the two output terminals. Two or more phase circuits 20 are arranged to connect the input and output terminals. In this embodiment, the phase circuits 20 are boost converters, each having an inductor 21 disposed between the first input terminal 11 and a bridge point 22, an upper branch switching unit 23 consisting of diodes disposed between the bridge point 22 and the first output terminal 14, and a switch 24 with a freewheeling diode 25 disposed between the bridge point 22 and the second input terminal 12 and the second output terminal 15. The switch 24 is typically a semiconductor switch, such as a MOSFET, SiC, IGBT, GaN, or other known switch types.

[0111] The figure shows an upper branch switching unit 23 composed of diodes; in other words, the upper branch switching unit 23 is essentially a diode. In other embodiments, the upper branch switching unit 23 includes variations, such as an active switch connected in parallel with a diode (example shown in...). Figure 2 The rightmost phase circuit 20 is shown. In fact, all phase circuits 20 typically have the same type of upper branch switch unit 23, or a separate active switch (not shown).

[0112] The phrase "arranged between" means that the corresponding components connect two points in the circuit and can carry the current between the two points according to the state of the components.

[0113] Controller 40 is arranged to control the switching of switch 24 and uses sensors (not shown) to measure, for example, voltage and current in the multiphase DC-DC converter 10. Controller 40 can be configured to control the current supplied to the output side to follow a total current setpoint. Depending on the operating environment of the multiphase DC-DC converter 10, such a setpoint can be determined by a supervisory control loop.

[0114] The total current supplied to the output side is the sum of the phase currents supplied by the phase circuits 20. The phase circuits 20 are operated to generate interleaved current pulses to minimize ripple in the total current. Based on the required total current and the total current setpoint, the controller 40 can determine the optimal number of phase circuits 20 to be activated, thereby supplying the total current when each phase circuit 20 operates under optimal or near-optimal conditions (e.g., with respect to switching losses).

[0115] Figure 3 The diagram illustrates the interleaving of current pulses, with two phase circuits 20 activated. The figure above shows the inductor current I of the two phase circuits 20 operating under BCM. L In each consecutive current pulse, the inductor current I... L It rises to the positive peak value +I when the corresponding switch 24 is closed. peak The current I is driven by the input voltage across inductor 21 and flows through switch 24. When switch 24 is open, the inductor current I... L The current decreases again, driven by the difference between the output voltage and the input voltage across inductor 21. When switch 24 is open, the inductor current I... L The current flows through the upper branch switching unit 23 and constitutes the phase current supplied to the output side. The reduced inductor current I is allowed before switch 24 closes again. L It drops below zero, reaching a negative peak -I peak This allows for zero-voltage and / or zero-current switching of switch 24, as explained further below.

[0116] The figure below shows the output current I generated by the sum of the two phase currents. AThe output current I A It is filtered by output capacitor 16. Its average current I Amean It can be controlled based on the total current setpoint. It also displays the average value I. Emean Input current I E Obviously, for the switching time period T corresponding to each phase circuit 20... SW The switching frequency, the output current has T SW The period is 2 / 2 and the input current has a variation with respect to its corresponding average value. SW The period of / 2 corresponds to its corresponding ripple frequency. The ripple frequency increases according to the number of activated phase circuits 20.

[0117] Figure 4 This shows the inductor current I for phase circuit 20. L The trajectory changes over time, and the corresponding state SW of switch 24 is observed. When SW is high, switch 24 is closed, i.e., in the conducting state. When SW is low, switch 24 is open, i.e., in the non-conducting state. Given the period length T (corresponding to the switching frequency) between the opening time 31 and the next opening time 31', the opening time period t of switch 24 is selected for closing. on This causes phase circuit 20 to operate under BCM. This is illustrated by BCM mode trajectory 38, where it returns to zero at the next on-time 31'.

[0118] Calculate the start time period t on The period length T is determined such that the average current to be delivered is generated, and the current returns to zero at the end of the period length T. This only requires knowing the inductance value of inductor 21, and further determined by the input voltage and the average current delivered by phase circuit 20.

[0119] More specifically, the turn-on time period t of phase circuit 20 on It can be calculated as

[0120]

[0121] in

[0122] • U IN The voltage on the input side can be measured;

[0123] • L is the inductance value of inductor 21, which can be determined through the design of inductor 21 or by measurement; and

[0124] • I mean The average current supplied by phase circuit 20 can be provided, for example, by controller 40.

[0125] The period length T can be calculated as

[0126]

[0127] Among them U OUT It is the voltage on the output side.

[0128] If two or more phase circuits 20 are to operate synchronously, and their respective phase currents undergo phase shifts depending on the number of activated phase circuits 20, then their period lengths T should be the same. The period length T for all phase circuits 20 can be determined by different methods:

[0129] In one embodiment, one phase circuit 20 is designated as the master device, and the others as slave devices. The master device is operated in a "self-synchronizing" mode. That is, the ideal cycle length T calculated as shown above is not precisely reached, but is determined by the inductor current I. L The actual time to reach zero is determined.

[0130] The period length T, determined in this way by the master device, is then used for the slave device phase circuit 20.

[0131] In another embodiment, the controller 40 together determines the cycle length T of all activated phase circuits 20. This means that an average output current is specified for each phase circuit 20 based on the total current setpoint, and the cycle length T and on-time period t of the phase circuit 20 are adjusted accordingly. on .

[0132] In each of the different methods, for any phase circuit 20 that is not operating in "self-synchronization" mode, there is a problem of operating it in the BCM with a specified period length T:

[0133] Using the start-up time period t calculated as shown above on Ideally, this should produce a BCM mode trajectory 38. In reality, the parameters of phase circuit 20, especially the inductance value, will not be exactly correct or will drift over time.

[0134] If the actual inductance value is lower than expected, or due to other deviations, CCM mode trajectory 37 will be achieved: when switch 24 is closed, the inductor current I... L It will rise more than expected. At the end of the cycle length T, at the next turn-on time 31', switch 24 will operate at the inductor current I. L Turn it on again before it returns to zero. Over several cycles, the inductor current I... L It will continue to rise. This is unacceptable.

[0135] If the actual inductance value is higher than expected, or due to other deviations, DCM mode trajectory 39 will be achieved: when switch 24 is closed, the inductor current I... LIt will rise less than expected. At the end of the cycle length T, at the next turn-on time 31', switch 24 will operate at the inductor current I. L After returning to zero, it will be turned on again. Depending on the delay, electromagnetic interference and switching losses will occur.

[0136] To prevent phase circuit 20 from operating in CCM or DCM and to keep it in BCM, the control period t on Instead of switching on at the next scheduled turn-on time 31', the switch 24 is switched on as in self-synchronizing operation, i.e., when the inductor current I... L After returning to zero, switch 24 is turned on. The time it was turned on is compared with the predetermined next on time 31', and the on time period t is adjusted accordingly. on Controllers such as PID controllers can be used, and more generally, controllers that make the steady-state error zero can be used. Therefore, over a series of periods, the turn-on time period t... on The inductor current I is adjusted so that L Return to zero at the expected next start time 31' corresponding to the expected or predetermined cycle length T.

[0137] The cycle length T can be adjusted or changed by another external control loop so that the average current of the phase circuit 20 and the total current of the multiphase DC-DC converter 10 follow the changes of their respective set points.

[0138] If the total current actually delivered by the multiphase DC-DC converter 10 is not in the ideal state according to the above formula, the external control loop can adjust the total current, thereby adjusting the average current delivered by each phase. This will in turn affect the turn-on time period t. on The period length T is adjusted.

[0139] Figure 5 The relevant timing points for controlling a multiphase DC-DC converter are shown: inductor current pulses from one phase circuit 20 (which may be the master unit) are represented by solid lines, and portions of other pulses (which may be slave units) are represented by dashed lines. The start of one pulse in the master unit is used as a reference time 35 (T) for the other pulses. ref For the nth slave device, the target activation time is 34 (T). target (n) is determined to be T target (n) = (T • n) / N. In the operation of each phase circuit 20, the corresponding actual turn-on time 36 can deviate from the target turn-on time 34, and the controller, as described above, modifies the turn-on time period t for the next cycle. on .

[0140] If there is no master phase and a cycle length T is specified for all active phase circuits 20, then all active phase circuits 20 operate as slave devices.

[0141] The situation shown in the previous figure, and the methods used to determine the period length T and the start-up time period t on The calculation is an approximation that does not consider the detailed current trajectory exactly before and after the turn-on time 31. This is acceptable because, on the one hand, as... Figure 6 As explained in the context, the charge and current involved in the switching operation are much smaller than the charge and current over the entire cycle. On the other hand, the small error caused by the approximation will affect the control of the on-time period t. on To correct it.

[0142] Figure 6 The inductor current I is shown. L The voltage V across switch 24 of phase circuit 20 S The trajectory. The voltage V across switch 24. S The voltage is the same as that at bridge point 22. During the off-time 32, the voltage rises and the current begins to decrease, flowing through switch 24 and the upper branch switch unit 23.

[0143] According to one embodiment, switch 24 is in the inductor current I L Turning on after the voltage drops to zero: The inductor current I is driven by the difference between the input and output voltages. L The value becomes negative, and the degree to which this negative value is determined by the closing time of the upper branch switch unit 23. When the upper branch switch unit 23 is composed of a diode, this time depends on the reverse recovery charge of the diode. When the upper branch switch unit 23 blocks the inductor current I... L At this time, it switches to the freewheeling diode 25 of switch 24. Freewheeling diode 25 becomes conductive and the voltage V across the switch... S The voltage drops to zero. When the voltage is zero, switch 24 is turned on during conduction time 31. Ideally, the inductor current I at this time... L Returning to zero again. As a result, switch 24 turns on at zero current, reducing EMC interference, and turns on at zero voltage, reducing switching losses. The figure shows that, except after turn-on time 31, the inductor current I... L In addition to rising again, if switch 24 is not turned on, the current and voltage will be measured using traces I' and V'.

[0144] Therefore, the voltage V across switch 24 can be used to determine the voltage. S Threshold detection is used to trigger the connection. When the voltage V S A value of zero can trigger a switch. Alternatively, the voltage V across switch 24 can also trigger a switch. S Below a predetermined threshold V greater than zero S_tresh Then a scheduled time was delayed by Tdel This can trigger a switchover. The predetermined time delay and threshold can be determined based on the parameters of the phase circuit and stored in the controller 40. Based on a threshold V greater than zero... S_tresh The trigger moves the point that has exceeded the threshold to an earlier time point, thus allowing compensation for the processing time required by controller 40. Alternatively, the turn-on can be triggered by threshold detection of the voltage at bridge point 22, which is typically the same as the voltage V across switch 24. S same.

[0145] Alternatively, the current I in the inductor can be used. L Threshold detection is used to trigger the switch-on. For this purpose, the inductor current I can be monitored. L The current itself, or through the upper branch switching unit 23, because it is related to the inductor current I before commutation. L The same applies. The inductor current I can be monitored by monitoring the magnetic field of inductor 21. L .

[0146] To ensure that both current and voltage are simultaneously zero or close to zero, the diodes constituting the upper branch switching unit 23 can be selected to have corresponding reverse recovery times. The reverse recovery time determines the inductor current I. L The time is negative. The diode is selected such that, under normal operating conditions, both current and voltage are zero.

[0147] Figure 7 The diagram illustrates activating an additional phase circuit and inserting its current pulses into an interleaved pulse sequence. Given a total current setpoint, in the pre-transition cycle, the current pulses are higher, with a cycle length Tpi longer than the post-transition cycle, which has a cycle length Tsi. The transition cycle has the same cycle length Tpi as the pre-transition cycle and coincides with the pulse period of the phase designated as the master device. The other activated phases (slave devices) are then operated as follows:

[0148] • During the transition period, the on-time (relative to the cycle) is the same as in the pre-transition period, and

[0149] • Their on-time period t on They are selected so that their next turn-on time (the time when the current returns to zero) is required in the post-transition cycle.

[0150] The newly activated phase circuit 20 is inserted as the last of the transition phases, after the last of the slave phases and before the master phase. Its next turn-on time 31' is required within the transition cycle. Its only free parameter is its turn-on time period t. on : Startup time period t onThe period length T is determined, and T then starts to regress from the next turn-on time 31' to determine the first turn-on time 31 when activating the new phase circuit 20. Turn-on time period t on It is selected to minimize the deviation of the total current of all phase circuits 20 from the total current setpoint.

[0151] Figure 8 This illustrates the disabling of a phase circuit and the removal of its current pulses from an interleaved pulse sequence. Given a total current setpoint, during the pre-transition period, the current pulses are lower, and the period length Tpd is shorter than the period length Tsd of the post-transition period. The transition period has the same period length Tpd as the pre-transition period and coincides with the pulse period of the phase designated as the master. Other phases besides the phase following the master (slave phases) are operated as follows:

[0152] • During the transition period, the activation time (relative to the cycle) is the same as the cycle before the transition, and

[0153] • Their on-time period t on They are selected so that their next turn-on time (the time when the current returns to zero) is required in the post-transition cycle.

[0154] Like other slave devices, the phase circuit 20 following the master device is also turned on at turn-on time 31, just as it is during the pre-transition cycle. Similarly, the only free parameter it uses to shape the last pulse is its turn-on time period t. on : Startup time period t on Determine the period length T and the point where the current returns to zero, ending the last pulse. Start-up time period t on It is selected to minimize the deviation of the total current of all phase circuits 20 from the total current setpoint.

[0155] Figure 9 The diagram illustrates the switching between DCM and BCM, and vice versa: The diagram above shows four phase circuits A, B, C, and D operating in pairs using DCM. In each pair (AC and BD), two phase circuits generate alternating pulses; one phase circuit (A or B) generates a current pulse, while the other phase circuit (C or D) does not provide current, and vice versa. The pulses from the two pairs are interleaved. Switching to BCM can be done individually within each pair by deactivating one of the two phase circuits (C and D) and operating the other (A and B) to generate two pulses in BCM. The overall current pattern remains unchanged.

[0156] Switching from BCM to DCM can be accomplished in a similar manner, i.e., by switching one phase circuit 20 from BCM to DCM and thereby omitting the alternating pulses, and activating another phase circuit 20 to provide the omitted pulses.

[0157] More generally (not shown), in the same manner, a phase that generates an integer multiple of an adjacent pulse sequence by DCM operation can be replaced by a single phase by BCM operation, and vice versa.

[0158] This switching from BCM to DCM or vice versa can be applied when the load or the required total current necessitates it. For example, when a relatively small current is required, the maximum frequency or minimum on-time period t... on BCM may not be allowed, so DCM must be used.

[0159] Figure 10 A method for controlling a multiphase DC-DC converter 10 in a DCM is illustrated. The purpose is to trigger pulses from one or more phase circuits 20 operating in the DCM, given a total current setpoint to be delivered by the converter. A sequence of phase current pulses of substantially the same shape is generated as follows:

[0160] • Determine the peak current I that defines the pulse shape Peak and period length T;

[0161] • Integrator to correspond to peak current I Peak The value of the product of the period length T and the period length T is initialized;

[0162] • Triggered in one of phase circuits 20 with a peak current I Peak A current pulse with a period length T;

[0163] • The integrator integrates the total current setpoint value 41, and its integrator output 42 tends to zero;

[0164] • When the integrator output 42 reaches zero, the next current pulse in the other phase circuit 20 is triggered by the flip-flop pulse 43, the integrator is re-initialized, and the program continues the previous step.

[0165] Peak current I can be selected Peak The range of the cycle length T is limited by hardware and operational considerations. Relevant parameters for selection may include the maximum switching frequency, minimum pulse length, and switching losses.

[0166] The trigger pulse 43 is multiplexed on the activated phase circuit 20. The sum of the current pulses will correspond to the total current setpoint 41. Due to the simple structure, the total current setpoint 41 can be tracked with low delay.

[0167] Figure 10 The example in the diagram shows the integrator initialized to a negative value and then integrated in the positive direction. It can be understood that the same principle can be implemented with inverted signs and different scaling values ​​of the signals involved.

[0168] In typical applications, the following values ​​can be presented.

[0169] • Period length T: corresponds to frequencies from 10 kHz to 700 kHz, especially from 20 kHz to 400 kHz.

[0170] • Number of phases N: 6 to 12 phases.

[0171] • Peak current per phase: up to 120 A or up to 200 A or higher.

[0172] • Average current per phase: up to 60 A or up to 100 A or higher.

[0173] • Total current: Up to 600 A, up to 800 A or higher.

[0174] • Output voltage: 200 V to 800 V.

[0175] • Inductance value of inductor 21: 5 to 20 microhenries or 8 to 15 microhenries.

[0176] Although the invention has been described in this embodiment, it should be clearly understood that the invention is not limited thereto, but may be embodied and practiced differently in other ways within the scope of the claims.

Claims

1. A control method for controlling a multiphase DC-DC converter (10) arranged to exchange power between an input side and an output side, the multiphase DC-DC converter (10) comprising two or more phase circuits (20). Each phase circuit (20) includes a switch (24) and an inductor (21), the switch (24) being arranged to control the inductor current I through the inductor (21). L , The phase circuit (20) is arranged to generate a phase current that contributes to the total current to be supplied to the output side of the multiphase DC-DC converter (10), the method comprising: • In boundary conduction mode (BCM), two or more of the phase circuits (20) are switched to generate interleaved phase current pulses with a period length of T, and the nominal on-time period of the switch (24) is t. on ; • In at least one of the two or more phase circuits (20) that are switched, and for consecutive phases, the on-time period t is repeatedly adjusted. on To control the inductor current I L The pulse length is adjusted to minimize the difference from the period length T; The method further includes one or two of the following steps in a sequence for the transition of the multiphase DC-DC converter (10) from discontinuous conduction mode to boundary conduction mode and from boundary conduction mode to discontinuous conduction mode, respectively. For the transition of the multiphase DC-DC converter (10) from discontinuous conduction mode to boundary conduction mode, for each of one or more sets of np phase circuits (20), np is two or more, for each of these sets • The individual np phase circuits (20) of the group are operated in discontinuous conduction mode to generate a sequence of np adjacent current pulses with a period length of T, each phase circuit (20) contributing to one of the current pulses in the sequence. • Instead of operating the individual np phase circuits (20) of the group in discontinuous conduction mode, one phase circuit (20) is operated in boundary conduction mode to continue the sequence of current pulses with a period length of T; For the transition of the multiphase DC-DC converter (10) from boundary conduction mode to discontinuous conduction mode, for each of one or more sets of np phase circuits (20), np is two or more, for each of these sets • Operate a phase circuit (20) in boundary conduction mode to generate a sequence of current pulses with a period length of T in boundary conduction mode; • Instead of operating one phase circuit (20) in boundary conduction mode, each np phase circuit (20) of the group is operated by operating the np phase circuit (20) in discontinuous conduction mode to generate a sequence of np adjacent current pulses of period length T to continue the sequence of current pulses of period length T, with each phase circuit (20) contributing to one of the current pulses in the sequence.

2. The control method according to claim 1, comprising adjusting the turn-on time period t based on the deviation between the actual zero-crossing time of the phase current and the desired zero-crossing time determined by the period length T. on This synchronizes the phase circuits.

3. The control method according to claim 1 or 2, The multiphase DC-DC converter (10) is arranged to exchange power between an input side including a first input terminal (11) and a second input terminal (12) and an output side including a first output terminal (14) and a second output terminal (15). The multiphase DC-DC converter (10) includes two or more phase circuits (20). Each phase circuit (20) includes an inductor (21) connected between the first input terminal (11) and the bridge point (22), an upper branch switch unit (23) connected between the bridge point (22) and the first output terminal (14), and a switch (24) with a parallel freewheeling diode (25) connected between the bridge point (22) and the second input terminal (12) and the second output terminal (15). The control method includes the following steps: • Determine the number N of phase circuits (20) to be activated; • Determine the start time period t on The initial value; • Determine the period length T; For at least one of the phase circuits (20) to be activated; • Turn on the switch (24) of the phase circuit (20) to make the inductor current I L The switch (24) is incremented over time, and a target opening time (34) is determined for the switch (24) to be opened next. • During the opening time period t on Then, the switch (24) is turned off, causing the inductor current I to... L The flow passes through the upper branch switch unit (23) and decreases over time; • After the current through the inductor (21) returns to zero, at the zero-crossing time (33), the switch (24) is turned on again and the above steps are repeated; • When repeating the above steps, if the zero-crossing time (33) is before the target opening time (34), then by increasing the opening time period t on To control the opening time period t on If the zero-crossing time (33) is after the target opening time (34), then the opening time period t is reduced. on To control the opening time period t on .

4. The control method according to claim 1 or 2, wherein, The turn-on time period t of the phase circuit (20) on The initial value is calculated as Among them, U IN The input voltage is L, the inductance of the inductor (21), and I is the voltage on the input side. mean The average current to be delivered by the phase circuit (20).

5. The control method according to claim 1 or 2, wherein, In the activated phase circuit (20), the corresponding period length T is calculated as follows: U OUT It is the voltage on the output side.

6. The control method according to claim 3, wherein, The step of determining the target turn-on time (34) includes calculating the target turn-on time (34) as an offset relative to the reference time (35) by dividing the period length T by the number of activated phase circuits N.

7. The control method according to claim 3, wherein, The step of turning the switch (24) back on includes one of the following: • Monitor the voltage across the switch (24) and turn on the switch (24) when the voltage across the switch (24) is zero. • Monitor the voltage across the switch (24) and turn on the switch (24) after a predetermined time delay following a predetermined threshold. • Monitor the inductor current I L And in the inductor current I L The switch (24) is turned on when it becomes zero after becoming negative due to the reverse current through the upper branch switch unit (23). • Monitor the current through the upper branch switch unit (23), and turn on the switch (24) when the current becomes zero after it becomes negative due to the reverse current through the upper branch switch unit (23).

8. The control method according to claim 7, wherein, The current through the upper branch switch unit (23) is the diode current.

9. The control method according to claim 3, wherein, The steps of turning the switch (24) back on include: • When the inductor current I L When the current returns to zero, its direction is reversed and it continues to flow through the inductor (21) and the upper branch switch unit (23) until the upper branch switch unit (23) is closed and the inductor current I... L The current is switched to the freewheeling diode (25); • Turn on the switch (24).

10. The control method according to claim 9, wherein, The upper branch switch unit (23) is composed of or includes a diode (23) and is turned off by a reverse charge in the diode, wherein the reverse charge of the diode is selected such that the reverse current through the diode is sufficient to discharge the capacitor between the bridge point (22) and the second input terminal (15).

11. The control method according to claim 1 or 2, comprising, in order to increase the number N of activated phase circuits (20) to N+1 given a total current setpoint, • During the transition period of length Tpi, the switches (24) of N phase circuits (20) are switched to be turned on relative to the period at the turn-on time (31) 0, dTpi, 2·dTpi, 3·dTpi … (N-1) ·dTpi, where dTpi = Tpi / N; • For a transition period of length Tsi, the N+1 phase circuits (20) are calculated relative to the target turn-on time of the period as 0, dTsi, 2·dTsi, 3·dTsi … N·dTsi, where dTsi=Tsi / (N+1); • During the transition period, the switches (24) of the N phase circuits (20) are switched on at the same on time (31) as in the previous transition period relative to this period; • During the transition period, for each of the N phase circuits (20), the turn-on time period t is set. on So that the current returns to zero at the corresponding target on time (34) in the post-transition period; • During the transition period, after the on-time (31) of (N-1)•dTpi, the (N+1)th new operation is initiated with an on-time period t. on The switch (24) minimizes the deviation of the total current of all phase circuits (20) from the total current setpoint during the transition period and the post-transition period.

12. The control method according to claim 1 or 2, further comprising reducing the number N of activated phase circuits (20) to N-1 given a total current setpoint, • During the transition period of length Tpd, the switches (24) of N phase circuits (20) are switched to be turned on relative to the period at the turn-on time (31) 0, dTpd, 2·dTpd, 3·dTpd … (N-1)·dTpd, where dTpd=Tpd / N; • For a transition period of length Tsd, the N-1 phase circuits (20) are calculated relative to the target turn-on time of the period as 0, dTsd, 2·dTsd, 3·dTsd … (N-2)·dTsd, where dTsd=Tsd / (N-1); • During the transition period, the switches (24) of the N phase circuits (20) are switched on at the same on time (31) as in the previous transition period relative to this period; • During the transition period, for each of the N phase circuits (20), except for the phase circuit (20) with a target turn-on time of dTpd relative to the period, the turn-on time period t is set. on , so that the current returns to zero at the corresponding target on time (34) in the post-transition period; • During the transition period, for the phase circuit (20) with a target turn-on time of dTpd relative to the period, the turn-on time period t of the last pulse of the phase circuit (20) is set. on This minimizes the deviation between the total current of all phase circuits (20) and the total current setpoint during the transition period and the post-transition period.

13. The control method according to claim 1 or 2, comprising determining the period length T by operating one of the phase circuits (20) referred to as the main device phase, having an on-time period t determined according to the average current to be delivered by the phase circuit (20). on And operate one or more of the remaining active phase circuits to adapt their timing and period lengths to the timing and period lengths of the master device phase.

14. A multiphase DC-DC converter (10) including a controller (40) comprising a voltage sensor arranged to determine the voltage U at the input side of each phase circuit (20). IN Output voltage U OUT The voltage U across the inductor (21) L The controller (40) is configured to perform the method according to claim 1.

15. The multiphase DC-DC converter (10) according to claim 14, wherein, In at least one phase circuit (20), the upper branch switching unit (23) includes or is composed of diodes, said diodes having a sufficiently large size to withstand inductor current I. L After returning to zero, the inductor current I is reversed. L The reverse recovery time is such that the reverse current discharges the capacitance of the switch (24), the freewheeling diode (25), and any parallel capacitors present before the switch (24) is turned on.

16. The multiphase DC-DC converter (10) according to claim 15, wherein, The reverse recovery time is large enough that the reverse current also discharges the capacitor arranged in parallel with the switch (24) before the switch (24) is turned on.

Images