BIDIRECTIONAL SIGNAL CONVERSION

Abstract

An embodiment of a controller for a multidirectional signal converter is operable to cause the converter to regulate a first signal at a first converter node, and to have a switch timing that is independent of a direction of power transfer between the first converter node and a second converter node. For example, in an embodiment, such a controller may be part of a bidirectional voltage converter that handles power transfer between two loads. Such a voltage converter may have improved conversion efficiency and a smaller size and lower component count as compared to a conventional multidirectional voltage converter. Furthermore, such a voltage converter may be operable with a common switching scheme regardless of the direction of power transfer, and without the need for an indicator of the instantaneous direction of power flow.

Description

SUMMARY

This Summary is provided to introduce, in a simplified form, a selection of concepts that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.

An embodiment of a controller for a bidirectional signal converter is operable to cause the converter to regulate a first signal at a first converter node, and to have a switch timing that is independent of a direction of power transfer between the first converter node and a second converter node.

For example, in an embodiment, such a controller may be part of a bidirectional voltage converter that handles power transfer between two loads. Such a voltage converter may have improved conversion efficiency, and a smaller size and lower component count, as compared to a conventional bidirectional voltage converter. Furthermore, such a voltage converter may be operable with a common switching scheme regardless of the direction of power transfer, and without the need for an indicator of the instantaneous direction of power flow.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of an embodiment of a bidirectional voltage converter and the sources/loads between which the converter is operable to transfer power.

FIG. 2 is a more detailed schematic diagram of an embodiment of the converter stages and transformer of the bidirectional converter of FIG. 1.

FIG. 3 is a timing diagram the switching signals for an embodiment of the converter stages of FIG. 2 operating at a duty cycle of greater than 50%.

FIG. 4 is a plot of the voltage across the first-stage filter capacitor of FIG. 2 versus the current through the first transformer winding of FIG. 2 while an embodiment of the first converter stage of FIG. 2 is operating in a boost mode.

FIG. 5 is a plot of the voltage across the first-stage filter capacitor of FIG. 2 versus the current through the first transformer winding of FIG. 2 while an embodiment of the first converter stage of FIG. 2 is operating in a buck mode.

FIG. 6 is a combination of the plots of FIGS. 4 and 5, and shows a transition of an embodiment of the converter stages of FIG. 2 from the buck mode to the boost mode and vice-versa in response to a change in the direction of power transfer.

FIG. 7 is a timing diagram of the switching signals for an embodiment of the converter stages of FIG. 2 operating at a duty cycle of less than 50%.

FIG. 8A is a schematic diagram of the converter stages and transformer of FIG. 2, and an embodiment of a current sensor coupled to the converter stages for sensing the total first-stage transformer current.

FIG. 8B is a schematic diagram of an embodiment of the controller of FIG. 1 for controlling the converter stages of FIGS. 2 and 8A.

FIG. 9 is a schematic diagram of an embodiment of the converter stages and transformer of FIG. 2, where the second converter stage includes a signal multiplier.

FIG. 10 is a schematic diagram of an embodiment of a bidirectional voltage converter having more than two phases.

DETAILED DESCRIPTION

Bidirectional signal converters, such as bidirectional voltage converters, may be used in applications where power is transferred back and forth between multiple loads. For example, an automotive system such as a gas-electric hybrid vehicle may have a higher-voltage battery for powering the electric drive motors (e.g., one motor per wheel), a lower-voltage battery for powering every other electrically powered component (e.g., lights, radio) of the automobile, and a bidirectional DC-DC voltage converter coupled between these two batteries. During a period of vehicle acceleration, the bidirectional converter may provide power from the lower-voltage battery to maintain a charge on the higher-voltage battery; conversely, during a period of regenerative braking, the power flow may reverse such that the bidirectional converter may provide power from the higher-voltage battery (which is being recharged by the electric drive motors operating as generators) to recharge the lower-voltage battery.

Unfortunately, such bidirectional converters may have problems including poor conversion efficiency, large size and high component count, the need for an indicator of the instantaneous direction of power flow, and a respective switching scheme for each direction of power transfer.

FIG. 1 is a schematic diagram of an embodiment of a portion of a system 10 that includes sources/loads 12 and 14, at least one motor/generator 16 that selectively receives power from and provides power to at least one of the sources/loads, and a bidirectional DC-DC voltage converter 18 that transfers power between the two sources/loads. For example, the system 10 may be an automotive system such as a gas-electric hybrid vehicle. As discussed below, an embodiment of the bidirectional converter 18 may have improved conversion efficiency, a smaller size, and a lower component count as compared to a conventional bidirectional voltage converter. Furthermore, an embodiment of the converter 18 may be operable with a switching scheme that is at least approximately independent of the direction of power transfer, and without the need for an indicator of the instantaneous direction of power flow.

In an embodiment, the sources/loads 12 and 14 are respective first and second batteries, each of which acts as a power source while providing a current to, e.g., charge the other battery, and which acts as a load while receiving a current, e.g., a charging current from the other battery. The first and second batteries 12 and 14 generate respective first and second voltages V₁ and V₂, which may be equal or unequal. For example, if the system 10 is an automotive system such as a gas-electric hybrid vehicle, then the first battery 12 may be a lead-acid battery that generates a lower voltage in the range of approximately 7 Volts (V)-16 V to power, e.g., the vehicle's lights and radio, and the second battery 14 may be a lithium-ion or nickel-metal-hydride (NiMH) battery that generates a higher voltage in the range of approximately 100 V-500 V to power the at least one motor/generator 16 while it is operating as a motor, e.g., to rotate at least one wheel of the vehicle.

The motor/generator 16 is operable as a motor while it is receiving power from at least one of the sources/loads 12 and 14, and operates as a generator while it is providing power to at least one of the sources/loads. For example, of the system 10 is hybrid vehicle and the sources/loads 12 and 14 are batteries, then during vehicle acceleration the motor/generator 16 may act as a motor by receiving power from at least one of the batteries to rotate one or more of the vehicle wheels, and during vehicle braking the motor/generator may act as a generator to recharge at least one of the batteries (sometime called “regenerative braking”).

The bidirectional voltage converter 18 includes first and second bidirectional-converter stages 20 and 22, a transformer 24, first and second current sensors 26 and 28, a controller 30, and first, second, third, and fourth converter nodes 32, 34, 36, and 38 respectively coupled to the sources/loads 12 and 14.

The first and second stages 20 and 22 each include at least one phase 40₁-40_n, respectively, and operate to bidirectionally transfer power between the source/loads 12 and 14 in response to the controller 30; and as discussed below, the converter stages may also operate to step up, step down, or regulate at least one of the voltages V₁ and V₂ at the converter nodes 34 and 36 in response to the controller. For example, assume that the controller 30 causes the converter stages 20 and 22 to regulate voltage V₂ to a level that is higher than the voltage V₁. While power is flowing from the source/load 14 (acting as a source) to the source/load 12 (acting as a load) during a first mode of operation, the first converter stage 20 may effectively step down the voltage V₂ to the voltage V₁ (the transformer 24 may assist in this stepping down as discussed below), and the first and second converter stages may cooperate to regulate the flow of current into the converter node 36 so as to regulate the voltage V₂. And while power is flowing from the source/load 12 (acting as a source) to the source/load 14 (acting as a load) during a second mode of operation, the first stage 20 may effectively step up, or boost, the voltage V₁ to the voltage V₂ (the transformer 24 may assist in this stepping up as discussed below), and the first and second converter stages may cooperate to regulate the flow of current out from the converter node 36 (i.e., from the second stage 22 toward the sources/loads 14) so as to regulate the voltage V₂.

The transformer 24 provides galvanic isolation between the sources/loads 12 and 14, and may also assist the first and second converter stages 20 and 22 with stepping up/down V₁ and V₂. The transformer 24 includes at least one first-stage winding 44₁-44_w, and at least one second-stage winding 46₁-46_w. As discussed below in conjunction with FIG. 2, in an embodiment, the transformer 24 includes one respective first-stage winding 44 and second-stage winding 46 for each pair of converter phases 40. The turns ratio between the windings 44 and 46 determines the level to which the transformer 24 steps up/down V₁ and V₂. For example, a turns ratio of 2:1 would cause the transformer 24 to generate across a second-stage winding 46 a voltage that is twice the voltage that is across a corresponding first-stage winding 44 while power is flowing from the source/load 12 to the source/load 14; likewise, the same turns ratio of 2:1 would cause the transformer to generate across a first-stage winding 44 a voltage that is ½ the voltage across a corresponding second-stage winding 46 while power is flowing from the source/load 14 to the source/load 12.

But because the efficiency (i.e., the ratio of power out to power in) of a transformer may decrease as the turns ratio increases, as discussed below in conjunction with FIG. 2, the first and second converter stages 20 and 22 may be designed to allow the transformer 24 to have a turns ratio as low as approximately 1:1 for improved efficiency of the bidirectional converter 18.

Still referring to FIG. 1, the first and second current sensors 26 and 28 allow the controller 30 to monitor the currents to the sources/loads 12 and 14. For example, where the sources/loads 12 and 14 are batteries, the first and second current sensors 26 and 26 may allow the controller 30 to control at least one charging parameter (e.g., current) of the batteries, and to prevent overcharging of the batteries.

The controller 30 may regulate at least one of the voltages and V₂, and, where the sources/loads 12 and 14 are batteries, may control the charging of these batteries, by controlling the operation of the first and second converter stages 20 and 22. For example, the controller 30 may control the switching duty cycle of at least one of the converter stages 20 and 22 as discussed below in conjunction with FIG. 2. Furthermore, the controller 30 may control the converter stages 20 and 22 without “knowing” the direction of power flow. That is, an embodiment of the controller 30 need not receive a signal that indicates the instantaneous direction of the power flow.

Still referring to FIG. 1, the operation of an embodiment of the system 10 is described, where, for example purposes, the system is an automotive system such as a hybrid vehicle, the sources/loads 12 and 14 are batteries (e.g., lead-acid and lithium-ion batteries, respectively), the voltage V₂ is regulated, and the voltage V₁ is unregulated (although the controller 30 may prevent overcharging of the battery 12). Furthermore, the periods of charging and discharging described below are assumed to be short enough such that the charges on the batteries 12 and 14 remain sufficiently high so that another generator (not shown, but typically run by a gasoline engine in the vehicle) need not be activated to recharge them.

During an accelerating mode of operation where the motor/generator 16 acts as a motor to rotate at least one of the wheels of the automotive system 10, the battery 14 provides a load current that drives the motor/generator.

After a period of time that depends on the level of charge on the battery 14, the voltage V₂ begins to decrease below its regulated value.

In response to the voltage V₂ decreasing below its regulated value, the controller 30 adjusts the duty cycle of the first and second converter stages 20 and 22 such that these stages transfer power from the battery 12 to the battery 14 so as to maintain V₂ at approximately its regulated value. Specifically, the controller 30 causes the first and second converter stages 20 and 22 to sink a discharge current from the battery 12 into the first converter node 34, to convert this discharge current into a charging current, and to source this charging current from the second converter node 36 so as to maintain the voltage V₂ at its regulated value by replenishing the second battery 14 with an amount of charge that is approximately equal to the charge that the battery 14 is providing to drive the motor/generation 16.

The charging of the second battery 14 by the first battery 12 may continue as long as the motor/generator 16 requires current to drive the at least one wheel of the vehicle 10.

Next, the driver (not shown in FIG. 1) of the vehicle 10 applies the brakes such that the vehicle enters into what is often called a regenerative-braking mode.

This causes the current being drawn by the motor/generator 16 from the battery 14 to decrease toward zero rather rapidly.

As the current drawn by the motor/generator 16 decreases, the controller 30 maintains V₂ at its regulated level by adjusting the duty cycle of the first and second converter stages 20 and 22 such that the current flowing into the converter node 34 from the battery 12, and the current flowing out from the converter node 36, decrease to compensate for the decrease in the current being drawn by the motor/generator 16.

If the driver (not shown in FIG. 1) of the vehicle 10 continues to apply the brakes, then, at some point, the motor/generator 16 begins to source a current into the battery 14. Therefore, this current from the motor/generator 16 recharges the battery 14.

In response to the current generated by the motor/generator 16, the controller 30 continues to maintain V₂ at its regulated level by adjusting the duty cycle of the first and second converter stages 20 and 22 such that the current flowing into the converter node 34 from the battery 12, and the current flowing out from the converter node 36 to the battery 14, further decrease to compensate for the current being generated by the motor/generator 16.

If the driver (not shown in FIG. 1) of the vehicle 10 still continues to apply the brakes, then, at some point, the current needed to recharge the battery 14 becomes less than the current being generated by the motor/generator 16. Therefore, this “excess current” from the motor/generator 14 causes the voltage V₂ to increase above its desired level unless this excess current is compensated for.

To maintain the voltage V₂ at its regulated level in response to the excess current being generated by the motor/generator 16, the controller 30 adjusts the duty cycle of the first and second converter stages 20 and 22 such that the first and second converter stages convert this excess current into a current for charging the battery 12. That is, the excess current from the motor/generator 16 flows into the converter node 36, and the controller 30 causes the first and second converter stages 20 and 22 to convert this excess current into a charging current that flows out from the converter node 34 and into the battery 12.

Therefore, the bidirectional converter 18 allows the motor/generator 16 to recharge not only the battery 14, but the battery 12 as well.

If the driver (not shown in FIG. 1) of the vehicle 10 still continues to apply the brakes, then, at some point, the voltage V₁ across the recharging battery 12 may equal or exceed a first charging-threshold voltage, thus indicating that the charging current into the battery 12 is to be reduced to a “trickle” so as to apply a “trickle charge” to the battery—trickle charging a battery may prevent damage to the battery caused by, e.g., overcharging.

Therefore, the controller 30 may generate a trickle current to continue the recharging of the battery 12 in a number of ways.

For example, the controller 30 may adjust the duty cycle of the first and second converter stages 20 and 22 so that the charging current flowing out from the node 34 and being monitored by the current sensor 26 does not exceed a specified trickle-value. Or, the controller 30, in addition to regulating the voltage V₂, may also regulate the voltage V₁ to a specified level such that the battery 12 is recharged via an approximately constant voltage applied across the battery.

But limiting the current flowing out from the convert node 34 or regulating the voltage V₁ may allow the excess current from the motor/generator 16 to increase V₂ above its regulated level, because now the converter 18 does “absorb” all of this excess current.

Therefore, the controller 30 may deactivate the motor/controller 16 from generating a current, may control an optional circuit (not shown in FIG. 1) between the motor/generator and the battery 14 to limit or block the current from the motor/generator, or may control another optional circuit (not shown in FIG. 1) between the converter 18 and the battery 12 to generate the trickle current and to divert any additional current flowing out from the converter node 34 to a dissipative load such as a resistor.

If the driver (not shown in FIG. 1) of the vehicle 10 still continues to apply the brakes, then, at some point, the voltage V₁ on the recharging battery 12 may equal or exceed a fully-charged threshold voltage, thus indicating that the charging current into the battery 12 is to be reduced to zero, i.e., terminated.

Therefore, the controller 30 may terminate the current flowing into the battery 12 in a number of ways.

For example, the controller 30 may adjust the duty cycle of the first and second converter stages 20 and 22 so that zero current flows out from the converter node 34.

But this may allow the excess current from the motor/generator 16 to increase the voltage V₂ above its regulated level, because now the converter 18 does not absorb all of this excess current.

Therefore, the controller 30 may deactivate the motor/controller 16 from generating a current, may control an optional circuit (not shown in FIG. 1) between the motor/generator and the battery 14 to limit or block the current from the motor/generator, or may control another optional circuit (not shown in FIG. 1) between the converter 18 and the battery 12 to block current from entering the battery 12 and to divert any current flowing out from the converter node 34 to a dissipative load such as a resistor.

Still referring to FIG. 1 and to the above-described embodiment of the system 10 and to the above-described example of operation of the system, at no time does the above-described embodiment of the controller 30 require a signal from, for example, a microprocessor, to notify the controller of the direction of the converter-node 36 current. By regulating the voltage V₂ regardless of the power-transfer direction, the controller 30 allows a smooth transition of the converter-node-34 and converter-node 36 currents from one direction to the other.

Furthermore, the above-described embodiment of the controller 30 need not change the switching scheme (e.g., switching timing, duty cycle) of the first and second converter stages 20 and 22 in dependence on the power-transfer direction. Instead, the controller 30 may adjust the duty cycle of the converter stages 20 and 22 as needed to regulate the voltage V₂ to a desired level.

Still referring to FIG. 1, alternate embodiments of the system 10 are contemplated. For example, instead of a single controller 30, the bidirectional converter 18 may include multiple controllers to perform the above-described actions. Furthermore, although described as being positive, at least one of the voltages V₁ and V₂ may be negative. Moreover, the system 10 may be other than an automotive system. In addition, at least one of the sources/loads 12 and 14 may be other than a battery, for example a bank of super capacitors. Furthermore, the controller 30 may control the charging of the battery 14 in a manner similar to that in which the controller controls the charging of the battery 12.

FIG. 2 is a schematic diagram of the first and second converter stages 20 and 22 and of the transformer 24 of a two-phase embodiment of the bidirectional converter 18 of FIG. 1. As discussed below, an embodiment of the converter 18 may provide one or more advantages, including:

a) allowing the transformer 24 to have a relatively low turns ratio (e.g., 1:1) for improved transformer efficiency;b) eliminating the need for a pre-regulator circuit on either side of the transformer 24 to reduce the component count and size of the converter 18;c) allowing the transistors to switch under zero-voltage-switching (ZVS) or zero-current-switching (ZCS) conditions in most circumstances for improved efficiency of the converter 18, and to reduce the size of one or more components of the converter in high-frequency applications;d) allowing the first converter stage 20 to operate as a current multiplier (e.g., a current doubler) while the converter 18 is providing a current (e.g., a charging current) to the source/load 12 (FIG. 1) so as to reduce the sizes of at least some of the components of the converter 18;e) allowing the first converter stage 20 to operate as a multiphase boost circuit while the converter 18 is providing a current (e.g., a charging current) to the source/load 14 (FIG. 1) so as to allow elimination of at least one pre-regulator circuit from the converter 18 and to allow a relatively low turns ratio for the transformer 24;f) allowing the controller 30 (FIG. 1) to be constructed from a commercially available power-supply controller, with perhaps minor modifications;g) reducing the ripple-voltage components of the voltages V₁ and V₂ due to the multiphase structure of the converter 18; andh) modulizing the converter 18 to allow phase dropping for improving the light-load efficiency of the converter.

The first converter stage 20 of the bidirectional converter 18 includes phase inductors 50 and 52 having inductances L₁ and L₂, low-side switching transistors 54 and 56, which receive switching signals S₁ and S₂ from the controller 30 (FIG. 1), high-side switching transistors 58 and 60, which receive switching signals S₃ and S₄ from the controller, and a filter capacitor 62 having a capacitance C₁. The inductor 50 and transistors 54 and 58 form a first phase of the converter 18, and the inductor 52 and transistors 56 and 60 form a second phase of the converter. The number of phases (two phases in this embodiment) in the first converter stage 20 may be considered the number of phases in the bidirectional converter 18. For example, one may refer to the converter 18 as a two phase converter of the first converter stage 20 has two phases. As discussed below, the first converter stage 20 operates as a boost converter while power is flowing from the converter node 34 to the converter node 36, and operates as a buck converter while power is flowing from the converter node 36 to the converter node 34.

The second converter stage 22 of the bidirectional converter 18 includes high-side switching transistors 64 and 66, which receive switching signals P₁ and P₂ from the controller 30 (FIG. 1), low-side switching transistors 68 and 70, which receive switching signals P₃ and P₄ from the controller, and a filter capacitor 72 having a capacitance C₂. The transistors 64 and 70 form a first half-bridge of the second stage 22, and the transistors 66 and 68 form a second half-bridge of the converter. As discussed below, the second stage 22 operates as a synchronous full-wave rectifier while power is flowing from the converter node 34 to the converter node 36, and operates as a DC-AC converter (a DC-to-square-wave converter in an embodiment) while power is flowing from the converter node 36 to the converter node 34.

The transformer 24 includes a first-stage winding 44 that may be modelled as having a leakage inductance L_k1, a second-stage winding 46 that may be modelled as having a leakage inductance L_k2, and the transformer itself may be modelled as having a magnetizing (sometimes called a coupling) inductance L_m.

FIG. 3 is a timing diagram of the signals S₁-S₄ and P₁-P₄ of FIG. 2 while an embodiment of the converter 18 of FIG. 2 is operating with a duty cycle greater than 50% to transfer power in other direction. Although in this embodiment the “duty cycle” of the stage 20 and 22, and thus of the converter 18, is defined as the ratio of the logic-high portion of the S₁ switching period to the total S₁ switching period, other definitions of the “duty cycle” are contemplated.

Referring to FIGS. 2 and 3, first is described an operational mode of an embodiment of the converter 18 where the converter has duty cycle of greater than 50% and is transferring power from the converter node 34 to the converter node 36 (i.e., from the first converter stage 20 to the second converter stage 22). In this mode of operation, the first converter stage 20 operates as a boost converter (a two-phase boost converter in the described embodiment), and the second converter stage 22 operates as a synchronous full-wave rectifier. Furthermore, the delay periods dd_x are fixed durations that are independent of the duty cycle, and may be generated by the controller 30 to allow at least some of the transistors to achieve at least approximately ZVS or ZCS as described below. In contrast, the periods D_x depend on the duty cycle.

At a time t₁, the signal S₁ has an inactive-logic-low level, the signal S₂ has an active-logic-high level, the signal S₃ is transitioning from an active logic low level to an active logic high level, and the signal S₄ has an inactive-logic-low level; therefore, the transistor 54, operating as a switch, is off, the transistor 56 is on, the transistor 58 is transitioning from off to on, and the transistor 60 is off. Furthermore, the signals P₂ and P₃ are transitioning from active logic-low to active logic-high levels, and the signals P₁ and P₄ have inactive logic-low levels; therefore, the transistors 66 and 68 are transitioning from off to on, and the transistors 64 and 70 are off.

Because the transistor 54 has been off for at least a delay period dd₁ before the transistor 58 turns on, at least a portion of the boost current flowing out from the inductor 50 is flowing through the body diode of the transistor 58 (the other portion of the inductor 50 boost current, I_firstwinding, is flowing through the first-stage winding 44) to the capacitor 62, and is thus charging the capacitor.

Therefore, while the transistor 58 is turning on, it does so with approximately zero volts (e.g., a diode drop of approximately 0.6 V-0.7 V) across it; in this way, the controller 30 (FIG. 1) causes the transistor 58 to achieve, at least approximately, ZVS, thus rendering the power consumed by this transistor during its switching period relatively low. Therefore, the ZVS of the transistor 58 may improve the efficiency of the bidirectional converter 18 as compared to a conventional bidirectional diode converters.

Also, because the transistor 54 has been off for at least a delay time dd₁ before the transistors 66 and 68 turn on, one of the following two scenarios is possible: 1) the current I_firstwinding flowing through the first-stage winding 44 induces in the second-stage winding 46 a current I_{secondwinding} that is high enough to forward bias the body diodes of the transistors 66 and 68, and to thus flow through this body diode, through the capacitor 72 (thus charging this capacitor), and through the body diode of the transistor 68 back to the winding 46, or 2) the current I_{secondwinding} induced in the winding 46 is not high enough to forward bias the body diodes of the transistors 66 and 68.

Therefore, in the first scenario, while the transistors 66 and 68 are turning on, they do so with approximately zero Volts (e.g., a diode drop of approximately 0.6V-0.7 V) across them; in this way, the controller 30 (FIG. 1) allows these transistors to achieve, at least approximately, ZVS, thus rendering the power consumed by the transistors 66 and 68 during their switching relatively low. Alternatively, in the second scenario, while the transistors 66 and 68 are turning on, they do so with approximately zero current through them; in this way, the controller 30 (FIG. 1) allows the transistors 66 and 68 to achieve, at least approximately, ZCS, thus also rendering the power consumed by the transistors 66 and 68 during their switching relatively low. Therefore, in either scenario, the respective ZVS or ZCS of the transistors 66 and 68 may further improve the efficiency of the converter 18 as compared to a conventional bidirectional converters. Furthermore, because the second scenario (ZCS) may hold even if the transistors 66 and 68 turn on at approximately the same time as the transistor 54, the controller 30 may transition the signals P₂ and P₃ to active high levels at approximately the same time that it transitions the signal S₁ to an inactive low level.

Next, during a period D₁, the signal S₁ is inactive low, the signal S₂ is active high, the signal S₃ is active high, and the signal S₄ is inactive low; therefore, the transistor 54 is off, the transistors 56 and 58 are on, and the transistor 60 is off. Furthermore, the signals P₂ and P₃ are active high, and the signals P₁ and P₄ are inactive low: therefore, the transistors 66 and 68 are on, and the transistors 64 and 70 are off.

Therefore, the boost current from the inductor 50 flows through the on transistor 58, and, therefore, this current, which was previously flowing through the body diode of the transistor 58, continues to charge the capacitor 62, and the voltage V_C1 across the capacitor (the on transistors 56 and 58 couple the capacitor C₁, and thus the voltage V_C1, across the winding 44) causes the current I_firstwinding to flow through the first-stage winding 44; therefore, the magnetically induced current I_{secondwinding} flows through the second-stage winding 46 and the transistors 66 and 68 to charge the capacitor C₂.

Furthermore, an inductor-charging current flows out from through the inductor 52 and into the transistor 56.

Moreover, because the first-stage winding 44 is connected across the capacitor 62 by the on transistors 56 and 58, the voltage across the winding 44 is effectively clamped to the voltage V_C1 across the capacitor 62. This also clamps the voltage across the second-stage winding 46 to V_C1×turns ratio of the transformer 24 (where the turns ratio is 1:1, then the voltage across the second winding 46 is also clamped to V_C1). Therefore, this limits the voltage across, the thus the voltage stress applied to, the transistors 66 and 68. Consequently, this may allow the bidirectional converter 18 to include smaller transistors 66 and 68 as compared to a conventional bidirectional converter.

Still during the period D₁, the boost current from the inductor 50 may remain relatively constant, but the current I_firstwinding through the first-stage winding 44 is increasing due to the voltage V_C1 from the capacitor 62 being applied across the first-stage winding.

Therefore, when the current I_firstwinding through the first-stage winding 44 exceeds the boost current from the inductor 50, a current flows from the capacitor 62, through the transistor 58, and through the first-stage winding to make up the difference between the first-stage winding current I_firstwinding and the boost current. That is the current from the capacitor 62 equals the difference between the boost current from the inductor 50 and the current I_firstwinding. As time passes during the period D₁, the current sourced by the capacitor 62 to the first-stage winding 44 increases, and the boost current from the inductor 50 may stay substantially constant or decrease, although such a decrease, if it occurs, may be negligible.

At a time t₂, the controller 30 transitions the signal S₃ from an active logic-high level to an inactive logic-low level, thus turning off the transistor 58. Furthermore, the controller 30 transitions the signals P₂ and P₃ to an inactive logic-low level, thus turning off the transistors 66 and 68.

During a delay period dd₂, because the current I_firstwinding through the first-stage winding 44 does not change instantaneously, the portion of I_firstwinding supplied by the capacitor 62 before the transistor 58 was turned off (at time t₂) is now supplied through the body diode of the transistor 54. The duration of the period dd₂ may be at least long enough to allow the body diode of the transistor 54 to begin to conduct. Furthermore, the induced current I_{secondwinding} through the second-stage winding 46 flows through the body diodes of the transistors 66 and 68.

Also during the delay period dd₂, an inductor-charging current continues to flow from the inductor 52 through the transistor 56 to ground.

At a time t₃, the controller 30 transitions the switching signal S₁ to an active logic-high level, thus turning on the transistor 54. But because the body diode of the transistor 54 is already conducting per above, this transistor achieves at least approximately ZVS, which may improve the efficiency of the converter 18. Furthermore, instead of transitioning the signals P₂ and P₃ to inactive logic-low levels at time t₂, the controller 30 may so transition P₂ and P₃ at time t₃ to reduce the time that the second-stage-winding current I_{secondwinding} flows through the body diodes of the transistors 66 and 68, and to thus improve the efficiency of the bidirectional converter 18.

Next, during a period D₂, both the transistors 54 and 56 are on, thus effectively connecting together both end nodes of the first-stage winding 44. If the period D₂ is long enough, then the current I_firstwinding through the first winding 44 caused by the discharging of the leakage inductance L_k1 will decay to zero, and thus the current I_{secondwinding} through the second-stage winding 46 will also decay to zero. As discussed below, this may allow the transistors 64 and 70 to achieve at least approximately ZCS.

Then, at a time t₄, the controller 30 transitions the signal S₂ to an inactive logic-low level, and thus turns off the transistor 56. Furthermore, the controller 30 may transition the signals P₁ and P₄ to active logic-high levels to turn on the transistors 64 and 70; if, per above, the current through the second-stage winding 46 has decayed to zero, then the transistors 64 and 70 achieve at least approximately ZCS.

During a delay period dd₃, the boost current from the inductor 52 that was flowing through the transistor 56 before it was turned off now flows toward the first winding 44.

But because the current I_firstwinding through the first-stage winding 44 cannot change instantaneously (from, e.g., zero as discussed above), the voltage at the node between the inductor 52 and the first-stage winding increases until the body diode of the transistor 60 begins to conduct this boost current—the delay period dd₃ may be at least long enough to allow the body diode of the transistor 60 to begin to conduct. This current through the body diode of the transistor 60 charges the capacitor 62.

At a time t₅, the controller 30 (FIG. 1) transitions the switching signal S₄ from an inactive logic-low level to an active logic-high level, thus turning on the transistor 60.

But because the body diode of the transistor 60 is conducting at least a portion of the boost current from the inductor 52 at the time t₅, this transistor achieves at least approximately ZVS, which may thus improve the efficiency of the bidirectional converter 18 as compared to a conventional bidirectional converter.

Also at time t₅, the controller 30 (FIG. 1) may transition the signals P₁ and P₄ to active logic-high levels to turn the transistors 64 and 70 on at the time t₅ instead of at the time t₄.

But even when turned on at the time t₅, the transistors 64 and 70 achieve at least approximately ZVS or ZCS, thus potentially improving the efficiency of the converter 18. If the current −I_{secondwinding} through the second-stage winding 46, which current is induced by the current −I_firstwinding through the first-stage winding 44, is not high enough at the time t₅ to turn on the body diodes of the transistors 64 and 70, then at least this current is low enough to allow the transistors 64 and 70 to achieve at least approximately ZCS. But if the current −I_{secondwinding} through the winding 46 is high enough to turn on the body diodes of the transistors 64 and 70, then the transistors 64 and 70 achieve at least approximately ZVS. Note that −I_firstwinding flows through the first-stage winding 44 in a direction opposite to the direction indicated by the respective arrow in FIG. 2; likewise, −I_{secondwinding} flows through the second-stage winding 46 in a direction opposite to the direction indicated by the respective arrow in FIG. 2.

During a period D₃, the signal S₂ is inactive logic low, the signal S_i is active logic high, the signal S₄ is active logic high, and the signal S₃ is inactive logic low; therefore, the transistor 54 is on, the transistors 56 and 58 are off, and the transistor 60 is on. Furthermore, the signals P₂ and P₃ are inactive logic low, and the signals P₁ and P₄ are active logic high: therefore, the transistors 66 and 68 are off, and the transistors 64 and 70 are on.

Therefore, the boost current from the inductor 52 flows through the on transistor 60, and, therefore, this current, which was previously flowing through the body diode of the transistor 60, continues to charge the capacitor 62, and the voltage −V_C1 causes the current −I_firstwinding to flow through the first-stage winding 44; therefore, an induced current −I_{secondwinding} flows through the winding 46 and the transistors 64 and 70 to maintain the voltage V₂ across the capacitor C₂ at a desired level.

Furthermore, during the period D₃, an inductor-charging current flows through the inductor 50 and the transistor 54 to ground.

Moreover, because the first-stage winding 44 is connected across the capacitor 62 by the on transistors 54 and 60, the voltage across the first-stage winding is effectively clamped to the voltage −V_C1 across the capacitor—the “−” sign indicates that the polarity of V_C1 relative to the first-stage winding 44 causes a current −I_firstwinding to flow through the first-stage winding. This also clamps the voltage across the second-stage winding 46 to −V_C1×the turns ratio of the transformer 24 (where the turns ratio is 1:1, then the voltage across the second-stage winding is also clamped to −V_C1). Therefore, this limits the voltage across, the thus the voltage stress applied to, the transistors 64 and 70. Consequently, this may allow the bidirectional converter 18 to include smaller transistors 64 and 70 as compared to a conventional bidirectional converter.

Still during the period D₃, the boost current from the inductor 52 may remain relatively constant, but the current −I_firstcurrent through the first-stage winding 44 is increasing due to the voltage −V_C1 from the capacitor 62 being applied across the first-stage winding.

Therefore, when the current −I_firstwinding through the first winding 44 exceeds the boost current from the inductor 52, a current sourced by the capacitor 62 flows through the transistor 60 and into the first-stage winding to make up the difference between the current −I_firstwinding and the boost current. As time passes during the period D₃, the portion of the current −I_firstwinding sourced by the capacitor 62 increases, and the boost current from the inductor 52 may stay substantially constant or decrease, although such a decrease, if it occurs, may be negligible.

At a time t₆, the controller 30 (FIG. 1) transitions the signal S₄ from an active logic-high level to an inactive logic-low level, thus turning off the transistor 60.

Also at the time t₆, the controller 30 transitions the signals P₁ and P₄ from an active logic-high level to an inactive logic-low level, thus turning off the transistors 64 and 70.

During a delay period dd₄, because the current −I_firstwinding through the first-stage winding 44 does not change instantaneously, the portion of this current supplied by the capacitor 62 before the transistor 60 was turned off at the time t₆ is now supplied through the body diode of the transistor 56. The duration of the period dd₄ may be at least long enough to allow the body diode of the transistor 56 to begin to conduct.

Also during the delay period dd₄, an inductor-charging current continues to flow from the inductor 50 and through the transistor 54 to ground.

Furthermore during the delay period dd₄, the current −I_{secondwinding} still flowing through the second-stage winding 46 continues to flow through the body diodes of the transistors 64 and 70.

At a time t₇, the controller 30 transitions the switching signal S₂ to an active logic-high level, thus turning on the transistor 56. But because the body diode of the transistor 56 is already conducting per above, the transistor 56 achieves at least approximately ZVS, which may improve the efficiency of the converter 18. Furthermore, instead of transitioning the signals P₁ and P₄ to inactive logic-low levels at the time t₆, the controller 30 may so transition P₁ and P₄ at the time t₇ to reduce the time that the second-stage winding current −I_{secondwinding} flows through the body diodes of the transistors 64 and 70, and to thus improve the efficiency of the bidirectional converter 18.

Next, during a period D₄, both the transistors 54 and 56 are on, thus effectively connecting together the end nodes of the first-stage winding 44. If the period D₄ is long enough, then the current −I_firstwinding through the first-stage winding 44 caused by the leakage inductance L_k1 will decay to zero, and thus the current −I_{secondwinding} through the second winding 46 will also decay to zero. This allows at least approximately ZCS of the transistors 66 and 68 (e.g., at the time t₈ or t₉ per below) in a manner similar to that discussed above for the transistors 64 and 70.

Then, at a time t₈, the controller 30 transitions the signal S₁ to an inactive logic-low level, and thus turns off the transistor 54.

Next, during a delay period dd₅ the boost current from the inductor 50 that was flowing through the transistor 54 causes the body diode of the transistor 58 to conduct.

Then, the above-described cycle repeats.

Still referring to FIGS. 2 and 3, alternate embodiments of the above-described boost operation are contemplated. For example, at least one of the delay periods dd₁-dd₅ may be omitted, although this may reduce the efficiency of the bidirectional converter 18.

FIG. 4 is a plot of the voltage V_C1 across the capacitor 62 of FIG. 2 versus the current I_firstwinding through the first-stage transformer winding 44 of FIG. 2 while an embodiment of the bidirectional converter 18 is operating in a boost mode. The voltage V_C1 is plotted along the x-axis, and the current I_firstwinding, scaled by a value Z₀=1/C₁, is plotted along the y-axis.

Referring to FIGS. 2-4, the operation of an embodiment of the bidirectional converter 18 of FIGS. 1-2 in boost mode with a duty cycle of greater than 50% is again discussed, but this time in terms of the voltage V_C1 across the capacitor 62 and the current I_firstwinding through the first-stage transformer winding 44. As discussed below in conjunction with FIG. 6, analyzing the operation in view of V_C1 and the current I_firstwinding illustrates how an embodiment of the bidirectional converter 18 may smoothly transition from transferring power in one direction to transferring power in the other direction.

As discussed above in conjunction with FIGS. 2 and 3, at the time t₁ (FIG. 3), which corresponds to point 80 of FIG. 4, the controller 30 transitions the signal S₃ to an active logic-high level to turn the transistor 58 on, and the current I_firstwinding is, for example purposes, assumed to be zero due to the transistors 54 and 56 connecting together the end nodes of the first-stage winding 44 before t₁. Assuming that I_firstwinding is zero is a valid assumption where the delay time dd₁ between the falling edge of S₁ and the rising edge of S₃ is long enough to allow the body diode of the transistor 58 to conduct so that this transistor achieves at least approximately ZVS.

During the period D₁ of FIG. 3, which corresponds to the curve 82 in FIG. 4, the capacitor 62 begins to charge, and, therefore, V_C1 begins to rise, due to a first portion of the boost current from the inductor 50 flowing through the transistor 58 and into the capacitor. Also, the current I_firstwinding begins to increase, and is equal to a second portion of the boost current from the inductor 50.

Because the on transistors 56 and 58 apply the voltage V_c1 across the first-stage winding 46, the current I_firstwinding continues to increase.

At a point 84 of the curve 82, I_firstwinding begins to exceed the boost current through the inductor 50.

Therefore, this excess portion of I_firstwinding—this excess portion being the difference between I_firstwinding and the boost current through the inductor 50—is sourced by the capacitor 62, thus causing V_C1 to begin to decrease (i.e., the capacitor 62 is discharging).

It is assumed that the delay dd₂ is short enough that it can be ignored for purposes of this analysis, such that at the time t₃ of FIG. 3 and at a point 86 of FIG. 4, the controller 30 (FIG. 1) transitions the signal S₁ to an active logic-high level and the signal S₃ to an inactive logic-low level to turn on the transistor 54 and to turn off the transistor 58.

Therefore, because the capacitor 62 is isolated from the first-stage winding 44, the voltage V_C1 remains at a constant value, and because both transistors 54 and 56 are on to couple together the end nodes of the first-stage winding 44, the current I_firstwinding quickly decays to zero along a line 88 of FIG. 4, such that the state of the bidirectional converter 18 has returned to the point 80 of FIG. 4.

Assuming, for purposes of explanation, that the inductance L₁ of the inductor 50 is significantly (e.g., ten times or more) greater than the leakage inductance L_k1 of the first-stage winding 44, then one may model the inductance 50 as a current source during the period D₁.

Therefore, making this assumption, one may show that the curve 82 and the points 84 and 86 lie on a circle having a center 90 at a point (V₂·TR, Z₀I_inductor50) and a radius R₁ given by the following equation:

R ₁=√{square root over ((V_c01−V₂·TR)²+(Z₀I_inductor50)²)}{square root over ((V_c01−V₂·TR)²+(Z₀I_inductor50)²)}  (1)

where V_C01 is the value of V_C1 at point 91 of FIG. 4 when I_firstwinding=0, and TR is the turns ratio of the transformer 24 (FIG. 2) looking at the first-stage side from the second-stage side.

And an angle θ₁ between the line 88 and a radius R to the point 91 is given by the following equation:

$\begin{array}{cc}{\theta }_1=\arctan \frac{V_{C01}-V_2·\mathrm{TR}}{Z_0I_{\mathrm{inductor}50}}& \left(2\right)\end{array}$

In an embodiment, the point 91 is not coincident with the point 80 because the voltage across the first-stage winding 44 must exceed V₂·TR before a nonzero current I_firstwinding begins to flow. This is because the on transistors 66 and 68 effectively clamp the right side of L_k1 (FIG. 2) to V₂·TR, so in order to cause a current to flow through the first-stage winding 44, the voltage on the left side of L_k1 must be greater than V₂·TR, even if only by a small amount (the amount shown in FIG. 4 may be exaggerated for clarity). Therefore, for example, where the transistor 58 turns on at the beginning of the delay period dd₁ such that it does not achieve ZVS, then the voltage V_C1 across the capacitor may increase to V_C01+V₂·TR before a nonzero current I_firstwinding begins to flow through the first-stage winding 44.

In another embodiment, the point 91 is substantially coincident with the point 80. If the transistor 58 turns on at the time t₁, and thus achieves ZVS, the diode drop across the body diode of the transistor 58 may increase the voltage at the left side of L_k1 enough so that a nonzero current I_firstwinding begins to flow at substantially the same time as the capacitor 62 begins to charge (i.e., at substantially the same time as the voltage V_C1 begins to increase). Therefore, in such an embodiment equations (1) and (2) reduce to the following equations.

$\begin{array}{cc}{R}_1=\sqrt{{\left(V_{\mathrm{co}1}-V_2·\mathrm{TR}=0\right)}^2+{\left(Z_0I_{\mathrm{inductor}50}\right)}^2}=Z_0I_{\mathrm{inductor}50}& \left(3\right)\\ {\theta }_1=\arctan \frac{V_{C01}-V_2·\mathrm{TR}=0}{Z_0I_{\mathrm{inductor}50}}=0& \left(4\right)\end{array}$

In yet another embodiment where a nonzero current I_firstwinding begins to flow before the capacitor 62 begins to charge, the point 91 may be above, if only slightly, the point 80 on the line 88, in which case the angle θ₁ remains equal to zero, and the radius R₁ is reduced from its value in equation (3) by the magnitude of I_firstwinding when the capacitor 62 begins to charge (i.e., when V_C1 begins to increase). This situation may occur if the transistor 58 turns on at the time t₁, and the voltage across the body diode of this transistor causes a nonzero current I_firstwinding to flow before the body diode begins to conduct.

Still referring to FIG. 4, continuing on, during the period D₂, the current I_firstwinding through the first-stage winding 44 remains at zero, and, therefore, the operating condition of the bidirectional converter 18, particularly of the first converter stage 20, remains at the point 80 of FIG. 4.

Next, for purposes of this analysis, it is assumed that the delay dd₃ is short enough to be ignored, such that at the time t₅ the controller 30 (FIG. 1) transitions the signal S₂ to an inactive logic-low level and the signal S₄ to an active logic-high level, thus turning off the transistor 56 and turning on the transistor 60.

During the period D₃, which corresponds to the curve 92 in FIG. 4, the capacitor 62 begins to charge, and, therefore, V_C1 begins to rise from the point 91, due to a first portion of the boost current from the inductor 52 flowing through the transistor 60 and into the capacitor, and the current −I_firstwinding begins to increase, and is equal to a second portion of the boost current from the inductor 52. As explained above in conjunction with FIGS. 2-3, the current −I_firstwinding is negative because it is flowing through the winding 44 in a direction opposite to its direction during the period D₁; this is why the curve 92 is in the lower-right quadrant of the plot of FIG. 4. Furthermore, for purposes of this analysis, it is assumed that the inductance L₂ of the inductor 52 equals the inductance L₁ of the inductor 50.

Because the voltage −V_C1 is across the first-stage winding 44, the magnitude of the current −I_firstwinding continues to increase.

At a point 94 of the curve 92, the magnitude of −I_firstwinding begins to exceed the boost current from the inductor 52.

Therefore, this excess portion of −I_firstwinding, which is equal to the difference between the magnitudes of −I_firstwinding and the boost current through the inductor 52, is sourced by the capacitor 62, thus causing the magnitude of −V_C1 to begin to decrease.

For purposes of this analysis, it is assumed that the delay dd₄ is short enough that it can be ignored, such that at the time t₇ of FIG. 3 and at a point 96 of FIG. 4, the controller 30 (FIG. 1) transitions the signal S₂ to an active logic-high level and the signal S₄ to an inactive logic-low level to turn on the transistor 56 and to turn off the transistor 60.

Therefore, because the capacitor 62 is isolated from the first-stage winding 44, the voltage −V_C1 remains at a constant value, and because both transistors 54 and 56 are on, the current −I_firstwinding quickly decays to zero along a line 98 of FIG. 4, such that the state of the bidirectional converter 18 has returned to the stable point 80 of FIG. 4 where the current −I_firstwinding is zero and the voltage −V_C1 is unchanging.

In an embodiment where the inductance L₂ of the inductor 52 is significantly (e.g., ten times or more) greater than the leakage inductance L_k1 of the first-stage winding 44, then one may model the inductor 52 as a current source during the period D₃.

Therefore, making this assumption, one may show that the curve 92 and the points 94 and 96 lie on a circle having a center 100 at a point (V₂·TR, −Z₀I_inductor52) and a radius R₁ given by the following equation:

R ₁=√{square root over ((V_C01−V₂·TR)²+(Z₀I_inductor52)²)}{square root over ((V_C01−V₂·TR)²+(Z₀I_inductor52)²)}  (5)

where V_C is the value of V_C1 at the point91 of FIG. 4 when −I_firstwinding=0.

And an angle θ₂ between the line 98 and a radius R to the point 91 of FIG. 4 is given by the following equation:

$\begin{array}{cc}{\theta }_2=\arctan \frac{V_{C01}-V_2·\mathrm{TR}}{Z_0I_{\mathrm{inductor}52}}& \left(6\right)\end{array}$

As discussed above, in an embodiment, the point 91 is not coincident with the point 80 because the voltage across the first-stage winding 44 must exceed zero before a nonzero current −I_firstwinding begins to flow. This is because the on transistors 64 and 70 effectively clamp the right side of L_k1 (FIG. 2) to −V₂·TR, so in order to cause a current to flow through the first-stage winding 44, the magnitude of the voltage on the bottom side of the first-stage winding must be greater than V₂·TR, even if only by a small amount (the amount shown in FIG. 4 may be exaggerated for clarity). Therefore, for example, where the transistor 60 turns on at the beginning of the delay period dd₃ such that it does not achieve ZVS, then the voltage V_C1 across the capacitor may increase to V_C0+V₂·TR before a nonzero current −I_firstwinding begins to flow through the first-stage winding 44.

In another embodiment, the point 91 is substantially coincident with the point 80. If the transistor 60 turns on at the time t₅, and thus achieves ZVS, the diode drop across the body diode of the transistor 60 may increase the voltage at the bottom side of the first-stage winding 44 enough so that a nonzero current −I_firstwinding begins to flow at substantially the same time as the capacitor 62 begins to charge (i.e., at substantially the same time as the voltage V_C1 begins to increase). Therefore, in such an embodiment equations (1) and (2) reduce to the following equations.

$\begin{array}{cc}{R}_1=\sqrt{{\left(V_{\mathrm{co}1}-V_2·\mathrm{TR}=0\right)}^2+{\left(Z_0I_{\mathrm{inductor}52}\right)}^2}=Z_0I_{\mathrm{inductor}52}& \left(7\right)\\ {\theta }_1=\arctan \frac{V_{C01}-V_2·\mathrm{TR}=0}{Z_0I_{\mathrm{inductor}52}}=0& \left(8\right)\end{array}$

In yet another embodiment where a nonzero current −I_firstwinding begins to flow before the capacitor 62 begins to charge, the point 91 may be below, if only slightly, the point 80 on the line 98, in which case the angle θ₂ remains equal to zero, the radius R₁ is reduced from its value in equation (7) by the magnitude of −I_firstwinding when the capacitor 62 begins to charge (i.e., when V_C1 begins to increase), and the point 91 for this mode is different than the point 91 for the above described mode (i.e., there are effectively two points 91). This situation may occur if the transistor 62 turns on at the time t₅, and the voltage across the body diode of this transistor causes a nonzero current −I_firstwinding to flow before the body diode begins to conduct.

Still referring to FIG. 4, one may see that where the duty cycle of the bidirectional converter 18 (FIG. 2) in boost mode is greater than 50%, the plot of the current I_firstwinding and the capacitor voltage V_C1 follows a “distorted” figure eight, starting from point 80, to the point 91, along the curve 82 to the point 86, from the point 86 along the line 88 back to the point 80, from the point 80 to the point 91 and along the curve 92 to the point 96, and from the point 96 along the line 98 back to the point 80. A similar smooth transition may be shown for the other scenarios (i.e., where the point 91 coincides with the point 80, or where the point 91 has a nonzero coordinate along the y-axis). Therefore, the current I_firstwinding smoothly transitions from one direction to another through the zero-current point 80; consequently, the current I_{secondwinding} through the second-stage winding 46 likewise smoothly transitions through its zero point. And, as discussed below in conjunction with FIG. 6, such a smooth transition of I_firstwinding and I_{secondwinding} may also occur when the direction of power transfer changes.

Referring again to FIGS. 2 and 3, now is described a buck operational mode of an embodiment of the converter 18, where the converter has a steady-state duty cycle of greater than 50% (i.e., S₁ and S₂ are logic high for more than 50% of the switching period), and is transferring power from the converter node 36 to the converter node 34 (i.e., from the second stage 22 to the first stage 20). In this mode of operation, the first converter stage 20 operates as a buck converter that provides power to (e.g., charges) the source/load 12 (FIG. 1), and the second converter stage 22 operates as a synchronous DC-AC converter. As discussed below, despite the change in the direction of power transfer, the switching timing, for example, the duty cycle, of the converters stages 20 and 22 may be at least approximately the same as described above during power transfer from the node 34 to the node 36 for a duty cycle >50%.

At the time t₁, the signal S₁ is inactive low, the signal S₂ is active high, the signal S₃ is transitioning from active low to active high, and signal S₄ is inactive low; therefore, the transistor 54 is off, the transistor 56 is on, the transistor 58 is transitioning from off to on, and the transistor 60 is off. Furthermore, the signals P₂ and P₃ are transitioning from active low to active high, and the signals P₁ and P₄ are inactive low; therefore, the transistors 66 and 68 are transitioning from off to on, and the transistors 64 and 70 are off.

Because the transistors 54 and 56 were both simultaneously on prior to a time t₀, the current I_firstwinding through the first-stage winding 44 is assumed to be zero, for purposes of this explanation. Likewise, the current I_{secondwinding} through the second-stage winding 46 is also assumed to be zero.

Because the transistors 66 and 68 turn on at time t₁ when the current I_{secondwinding} is zero, these transistors 66 and 68 achieve at least approximately ZCS, which may improve the efficiency of the bidirectional converter 18 as compared to a conventional bidirectional converter.

As the transistors 66 and 68 turn on, the current −I_{secondwinding} begins to flow from the capacitor 72, through the transistor 66, through the second winding 46, and through transistor 68 back to the capacitor 72.

Because the transistor 58 also turns on at the time t₁ when −I_firstwinding is zero, the transistor 58 at least approximately achieves ZCS, which may improve the efficiency of the bidirectional converter 18.

Alternatively, the transistors 66 and 68 may turn on at the time t₀ per the dashed lines of FIG. 3 such that a current −I_firstwinding may begin to flow before the transistor 58 turns on. Assuming that the delay period dd₁ is long enough to allow the body diode of the transistor 58 to begin conducting, then the transistor 58 may instead achieve at least approximately ZVS, which may improve the efficiency of the converter 18.

During the period D₁, the signal S₁ is inactive low, the signal S₂ is active high, the signal S₃ is active high, and the signal S₄ is inactive low; therefore, the transistor 54 is off, the transistors 56 and 58 are on, and the transistor 60 is off. Furthermore, the signals P₂ and P₃ are active high, and the signals P₁ and P₄ are inactive low: therefore, the transistors 66 and 68 are on, and the transistors 64 and 70 are off.

Because the second-stage winding 46 is connected across the capacitor 72 by the on transistors 66 and 68, the voltage across the second-stage winding is effectively clamped to the voltage V₂ across the capacitor. This also clamps the voltage across the first-stage winding 44 to V₂ times the turns ratio TR of the transformer 24 as viewed from the second-stage side (where the turns ratio is 1:1, then the voltage across the first-stage winding 44 is also clamped to V₂). Therefore, this limits the voltage across, the thus the voltage stress applied to, the transistors 56 and 58. Consequently, this may allow the bidirectional converter 18 to include smaller transistors 56 and 58 as compared to a conventional bidirectional converter.

Initially during the period D₁, a buck current that is greater than −I_firstwinding flows through the inductor 50 into the converter mode 34. Therefore, a current from the capacitor 62 flows through the transistor 58 and through the inductor 50 to make up the difference between −I_firstwinding and the buck current, thus discharging the capacitor and causing V_C1 to decrease.

Furthermore, a discharging current flows through the inductor 52 and the transistor 56 into the converter node 34.

Still during the period D₁, the current −I_firstwinding from the first-stage winding 44 is increasing due to the on transistors 66 and 68 applying the voltage V₂ from the capacitor 72 across the second-stage winding 46.

At some point during the period D₁, the current −I_firstwinding exceeds the buck current flowing through the inductor 50. Therefore, a first, excess portion of the current −I_firstwinding from the first stage winding 44 charges the capacitor 62, thus causing V_C1 to increase, and a second portion of the current −I_firstwinding flows through the inductor 50 as the buck current. By controlling the duty cycle of the bidirectional converter 18, the controller 30 (FIG. 1) causes the buck current through the inductor 50 to have a value that regulates the voltage V₂. That is, the controller 30 “bleeds” enough charge off of the capacitor 72 to maintain V₂ at a desired level.

At a time t₂, the signals S₃, P₂, and P₃ transition from active high to inactive low levels, thus turning off the transistors 58, 66, and 68. Alternatively, the controller 30 (FIG. 1) may turn off the transistors 66 and 68 slightly after (e.g., at time t₃) turning off the transistor 58. In this case (and even in the previous case), the portion of −I_firstwinding that was flowing through the on transistor 58 may continue to flow through the body diode of the off transistor 58 during the delay between the turn off of the transistors 66 and 68 and the transistor 58.

During the delay period dd₂, because the buck current through the inductor 50 does not change instantaneously, at least the portion of the buck current previously provided by the capacitor 62 is now supplied through the body diode of the transistor 54. The duration of the period dd₂ may be at least long enough to allow the body diode of the transistor 54 to begin to conduct.

Also during the delay period dd₂, an inductor-discharging current continues to flow from ground, through the transistor 56 and the inductor 52.

At the time t₃, the controller 30 transitions the switching signal S₁ to active high, thus turning on the transistor 54. But because the body diode of the transistor 54 is already conducting per above, the transistor 54 achieves at least approximately ZVS, which may improve the efficiency of the converter 18.

Next, during the period D₂, both the transistors 54 and 56 are on, thus effectively coupling together the end nodes of the first-stage winding 44. If the period D₂ is long enough, then the current −I_firstwinding through the first-stage winding 44 caused by the leakage inductance L_k1 will decay to zero, and thus the current through the second-stage winding 46 will also decay to zero. As discussed below, this allows the transistors 64 and 70 to achieve at least approximately ZCS.

Then, at the time t₄, the controller 30 transitions the signal S₂ to an inactive low level, and thus turns off the transistor 56.

During the delay period dd₃, the buck current that was flowing through the transistor 56 before it was turned off now flows through the body diode of the transistor 56.

At either time t₄ or t₅, the controller 30 transitions signals P₁ and P₄ to turn on the transistors 64 and 70. Because these transistors have no current flowing through them, they achieve at least approximately ZCS, which may improve the efficiency of the converter 18.

At time t₅, the controller 30 transitions S₄ active high, and thus turns on the transistor 60. If the controller 30 transitioned P₁ and P₄ active high at time t₄ then the current I_firstwinding (induced by the current I_{secondwinding} flowing through the second-stage winding 46) begins to flow through the body diode of the transistor 60 by time the t₅ such that this transistor achieves at least approximately ZVS. Alternatively, if the controller 30 transitions P₁ and P₄ active high at the time t₅, then the transistor 60 achieves at least approximately ZCS. In either scenario, the efficiency of the converter 18 may be improved.

During the period D₃, the signal S₂ is inactive low, the signal S₁ is active high, the signal S₄ is active high, and the signal S₃ is inactive low; therefore, the transistor 54 is on, the transistors 56 and 58 are off, and the transistor 60 is on. Furthermore, the signals P₂ and P₃ are inactive low, and the signals P₁ and P₄ are active high: therefore, the transistors 66 and 68 are off, and the transistors 64 and 70 are on.

Therefore, initially during the period D₃, the buck current through the inductor 52 is larger than the current I_firstwinding, and thus the buck current discharges the capacitor 62 via the on transistor 60.

Furthermore during the period D₃, an inductor-discharging current flows from ground, through the transistor 54, and through the inductor 50.

Moreover, because the second-stage winding 46 is connected across the capacitor 72 by the on transistors 64 and 70, the voltage across the second-stage winding is effectively clamped to the voltage V₂ across this capacitor. This also clamps the voltage across the first-stage winding 44 to V₂×TR of the transformer 24 (of the turns ratio is 1:1, then the voltage across the first-stage winding 44 is also clamped to V₂). Therefore, this limits the voltage across, the thus the voltage stress applied to, the transistors 56 and 58. Consequently, this may allow the bidirectional converter 18 to include smaller transistors 56 and 58 as compared to a conventional bidirectional converter.

Still during the period D₃, the current through the first-stage winding 44 is increasing due to the voltage V₂ from the capacitor 72 being applied across the second-stage winding 46.

Therefore, when the current I_firstwinding through the first-stage winding 44 exceeds the buck current through the inductor 52, a current equal to the difference between I_firstwinding and the buck current through the inductor 50 flows through the transistor 60 and into the capacitor 62, thus charging the capacitor and increasing V_C1. As time passes during the period D₃, the portion I_firstwinding to the capacitor 62 increases.

At the time t₆, the controller 30 (FIG. 1) transitions the signal S₄ from active high to inactive low, thus turning off the transistor 60.

Also at time t₆, the controller 30 transitions the signals P₁ and P₄ from active high to inactive low, thus turning off the transistors 64 and 70. Alternatively, the controller 30 may not turn off the transistors 64 and 70 until time t₇.

If the current I_firstwinding continues to flow during the delay period dd₄ after the transistor 60 turns off (e.g. due to the discharge of the leakage inductance L_K1 or the transistors 64 and 70 still being on), then a portion of I_firstwinding that exceeds the buck current through the inductor 52 flows through the body diode of the transistor 60 and into the capacitor 62.

Likewise, of the current I_{secondwinding} continues to flow after the transistors 64 and 70 turn off (e.g., due to the discharge of the leakage inductance L_K2), then I_{secondwinding} flows through the body diodes of the transistors 66 and 68.

Furthermore during the delay period dd₄, because the buck current through the inductor 52 does not change instantaneously, this current begins to flow through the body diode of the transistor 56. The duration of the period dd₄ may be at least long enough to allow the body diode of the transistor 56 to begin to conduct.

Also during the delay period dd₄, an inductor-discharging current continues to flow from ground, through the transistor 54, and through the inductor 50.

At the time t₇, the controller 30 transitions the switching signal S₂ to active high, thus turning on the transistor 56. But because the body diode of the transistor 56 is already conducting per above, the transistor 56 achieves at least approximately ZVS, which may improve the efficiency of the converter 18.

Next, during the period D₄, both the transistors 54 and 56 are on, thus effectively coupling together the end nodes of the first-stage winding 44. If the period D₄ is long enough, then the current through the first-stage winding 44 caused by the leakage inductance L_k1 will decay to zero, and thus the current through the second-stage winding 46 will also decay to zero. This allows the transistors 66 and 68 to achieve at least approximately ZCS (e.g., at time t₈ or t₉), which may improve the efficiency of the converter 18.

Then, at the time t₈, the controller 30 transitions the signal S₁ to an inactive low level, and thus turns off the transistor 54.

At either the time t₈ or t₉, the controller 30 transitions the signal P₂ and P₃ active high to turn on the transistors 66 and 68, which achieve at least approximately ZCS per above.

At the time t₉, the controller 30 transitions the signal S₃ active high to turn on the transistor 58, which achieves at least approximately ZVS (e.g., if the transistors 66 and 68 turn on at the time t₈) or ZCS (e.g., if the transistors 66 and 68 turn on at the time t₉), which may improve the efficiency of the converter 18.

Next, the above-described cycle repeats.

Still referring to FIGS. 2 and 3, alternate embodiments of the above-described buck-operating mode are contemplated. For example, at least one of the delay periods dd₁-dd₅ may be omitted, although this may reduce the efficiency of the converter 18.

FIG. 5 is a plot of the voltage V_C1 across the capacitor 62 of FIG. 2 versus the current I_firstwinding through the first-stage transformer winding 44 of FIG. 2 while an embodiment of the bidirectional converter 18 is operating in a buck mode. The voltage V_C1 is plotted along the x-axis, and the current I_firstwinding, scaled by a value Z₀=1/C₁, is plotted along the y-axis.

Referring to FIGS. 2-3 and 5, the operation of an embodiment of the bidirectional converter 18 of FIGS. 1-2 in buck mode with a duty cycle of greater than 50% is again discussed, but this time in terms of the voltage V_C1 across the capacitor 62 and the current I_firstwinding through the first-stage transformer winding 44. As discussed below in conjunction with FIG. 6, analyzing the operation in view of V_C1 and the current I_firstwinding illustrates how an embodiment of the bidirectional converter 18 may smoothly transition from transferring power in one direction to transferring power in the other direction.

As discussed above in conjunction with FIGS. 2 and 3, at the time t₁ (FIG. 2) or at a time t_o, which is the delay time dd₁ before the time t₁, and which corresponds to point 110 of FIG. 5, the controller 30 transitions the signal S₃ to an active logic-high level to turn the transistor 58 on, and the current I_firstwinding is, for example purposes, assumed to be zero due to the transistors 54 and 56 effectively coupling together the end nodes of the first-stage winding 44 before the time t₀.

During the period D₁ of FIG. 3, which corresponds to the curve 112 in FIG. 5, the capacitor 62 begins to discharge, and, therefore, V_C1 begins to fall, due to a first portion of the buck current to the inductor 50 flowing from the capacitor through the transistor 58.

At a point 111, the current −I_firstwinding begins to increase in magnitude and is equal to a second portion of the buck current to the inductor 50. In an embodiment, the point 111 does not coincide with the point 110 because the voltage V_C1 drops by an amount V₀₀₂ (the magnitude of V₀₀₂ may be exaggerated in FIG. 5 for purposes of illustration) before a non-zero current −I_firstwinding begins to flow. As discussed above in conjunction with FIG. 4, this is because a non-zero voltage drop across the leakage inductance L_k1 may be needed for a non-zero current −I_firstwinding to flow.

At a point 114 of the curve 112, the magnitude of −I_firstwinding begins to exceed the buck current through the inductor 50.

Therefore, the excess portion of −I_firstwinding, this excess portion being the difference between the magnitude of −I_firstwinding and the buck current through the inductor 50, flows through the transistor 58 to the capacitor 62, thus causing V_C1 to begin to increase (i.e., the capacitor 62 is charging).

It is assumed that the delay dd₂ is short enough that it can be ignored for purposes of this analysis, such that at the time t₃ of FIG. 3 and at a point 116 of FIG. 5, the controller 30 (FIG. 1) transitions the signal S₁ to an active logic-high level and at signal S₃ to an inactive logic-low level to turn on the transistor 54 and to turn off the transistor 58.

Therefore, because the capacitor 62 is isolated from the inductor 50, the voltage V_C1 remains at a constant value, and because both transistors 54 and 56 are on to couple together the end nodes of the first-stage winding 44, the current −I_firstwinding quickly decays to zero along a line 118 of FIG. 5, such that the state of the bidirectional converter 18 has returned to the point 110 of FIG. 5.

Assuming, for purposes of explanation, that the inductance L₁ of the inductor 50 is significantly (e.g., ten times or more) greater than the leakage inductance L_k1 of the first-stage winding 44, then one may model the inductance 50 as a current source (sourcing current to the converter node 34) during the period D₁.

Therefore, making this assumption, one may show that the curve 112 and the points 111, 114, and 116 lie on a circle having a center 120 at (V₂·TR, Z₀I_inductor50) and a radius R₂ given by the following equation:

R ₂=√{square root over ((V_c02−V₂·TR)²+(Z₀I_inductor50)²)}{square root over ((V_c02−V₂·TR)²+(Z₀I_inductor50)²)}  (9)

where V_C02 is the value of V_C1 at the point 111 of FIG. 5 when −I_firstwinding=0 as discussed above.

And an angle θ₃ between the line 118 and a radius R₂ to the point 111 is given by the following equation:

$\begin{array}{cc}{\theta }_3=\arctan \frac{V_2·\mathrm{TR}-V_{C02}}{Z_0I_{\mathrm{inductor}50}}& \left(10\right)\end{array}$

In another embodiment, the point 111 may be substantially coincident with the point 110. For example if the transistor 58 turns on at the time t₁ as described above, and thus achieves ZCS, the diode drop across the body diode of the transistor 54 may increase the voltage drop across L_k1 enough so that a nonzero current −I_firstwinding begins to flow at substantially the same time as the capacitor 62 begins to discharge (i.e., at substantially the same time as the voltage V_C1 begins to decrease). Therefore, in such an embodiment equations (9) and (10) reduce to the following equations.

$\begin{array}{cc}{R}_2=\sqrt{{\left(V_{\mathrm{co}2}-V_2·\mathrm{TR}=0\right)}^2+{\left(Z_0I_{\mathrm{inductor}50}\right)}^2}=Z_0I_{\mathrm{inductor}50}& \left(11\right)\\ {\theta }_3=\arctan \frac{V_{C02}-V_2·\mathrm{TR}=0}{Z_0I_{\mathrm{inductor}50}}=0& \left(12\right)\end{array}$

In yet another embodiment where a nonzero current −I_firstwinding begins to flow before the capacitor 62 begins to discharge, the point 111 may be below, if only slightly, the point 110 on the line 118, in which case the angle θ₃ remains equal to zero and the radius R is reduced from its value in equation (11) by the magnitude of −I_firstwinding when the capacitor 62 begins to discharge (i.e., when V_CS begins to decrease). This situation may also occur if the transistor 58 turns on at the time t₁, and the voltage across the body diode of the transistor 54 causes a nonzero current −I_firstwinding to flow before the transistor 58 begins to conduct.

Still referring to FIGS. 1-2 and 5 and continuing on, during the period D₂, the current −I_firstwinding through the first-stage winding 44 remains at zero, and, therefore, the operating condition of the bidirectional converter 18 remains at the point 110 of FIG. 5.

Next, for purposes of this analysis, it is assumed, that the delay dd₃ is short enough to be ignored, such that at the time t₅ the controller 30 (FIG. 1) transitions the signal S₂ to an inactive logic-low level and the signal S₄ to an active logic-high level, thus turning off the transistor 56 and turning on the transistor 60. Alternatively, the controller 30 may transition the signal S2 to an inactive logic-low level at and the signal S4 to an inactive logic-high level at the time t₄.

During the period D₃, which corresponds to the curve 122 in FIG. 5, the capacitor 62 begins to discharge, and, therefore, V_C1 begins to fall from the point 111, due to a first portion of the buck current to the inductor 52 flowing from the capacitor through the transistor 60, and the current I_firstwinding begins to increase, and is equal to a second portion of the buck current to the inductor 52. Furthermore, for purposes of this analysis, it is assumed that the inductance L₂ of the inductor 52 equals the inductance L₁ of the inductor 50.

Because the voltage V₂ is across the second-stage winding 46, the current I_firstwinding continues to increase.

At a point 124 of the curve 122, I_firstwinding begins to exceed the buck current through the inductor 52.

Therefore, this excess portion of I_firstwinding, which is equal to the difference between I_firstwinding and the buck current through the inductor 52, is supplied to the capacitor 62, thus causing V_C1 to begin to increase.

For purposes of this analysis, it is assumed that the delay dd₄ is short enough that it can be ignored, such that at the time t₇ of FIG. 3 and at a point 126 of FIG. 5, the controller 30 (FIG. 1) transitions the signal S₂ to an active logic-high level and the signal S₄ to an inactive logic-low level to turn on the transistor 56 and to turn off the transistor 60.

Therefore, because the capacitor 62 is isolated from the inductors 50 and 52, the voltage V_C1 remains at a constant value, and because both transistors 54 and 56 are on, the current I_firstwinding quickly decays to zero along a line 128 of FIG. 5, such that the state of the bidirectional converter 18 has returned to the stable point 110 of FIG. 5 where the current I_firstwinding is zero and the voltage V_C1 is unchanging.

In an embodiment where the inductance L₂ of the inductor 52 is significantly (e.g., ten times or more) greater than the leakage inductance L_k1 of the first-stage winding 44, then one may model the inductor 52 as a current source (sourcing current to the node 34) during the period D₃.

Therefore, making this assumption, one may show that the curve 122 and the points 111, 124, and 126 lie on a circle having a center 130 at a point (V₂·TR, Z₀I_inductor52) and a radius R₂ given by the following equation:

R ₂=√{square root over ((V_C02−V₂·TR)²+(Z₀I_inductor52)²)}{square root over ((V_C02−V₂·TR)²+(Z₀I_inductor52)²)}  (13)

where V_C02 is the value of V_C1 at the points 111 of FIG. 5 when I_firstwinding=0 as discussed above.

And an angle θ₄ between the line 126 and a radius R₂ to the point 111 of FIG. 5 is given by the following equation:

$\begin{array}{cc}{\theta }_4=\arctan \frac{V_2·\mathrm{TR}-V_{C02}}{Z_0I_{\mathrm{inductor}52}}& \left(14\right)\end{array}$

The point 111 may be non-coincident with the point 110 where the transistors 56 and 60 turn off and on, respectively, at substantially the same time.

In another embodiment, the point 111 may be substantially coincident with the point 110. For example if the transistor 60 turns on at the time t₅ as described above, and thus the transistor 60 achieves ZCS, the diode drop across the body diode of the transistor 56 may increase the voltage drop across L_k1 enough so that a nonzero current I_firstwinding begins to flow at substantially the same time as the capacitor 62 begins to discharge (i.e., at substantially the same time as the voltage V_C1 begins to decrease). Therefore, in such an embodiment equations (13) and (14) reduce to the following equations.

$\begin{array}{cc}{R}_2=\sqrt{{\left(V_{\mathrm{co}2}-V_2·\mathrm{TR}=0\right)}^2+{\left(Z_0I_{\mathrm{inductor}52}\right)}^2}=Z_0I_{\mathrm{inductor}52}& \left(15\right)\\ {\theta }_3=\arctan \frac{V_{C02}-V_2·\mathrm{TR}=0}{Z_0I_{\mathrm{inductor}52}}=0& \left(16\right)\end{array}$

In yet another embodiment where a nonzero current I_firstwinding begins to flow before the capacitor 62 begins to discharge, the point 111 may be above, if only slightly, the point 110 on the line 128, in which case the angle θ₄ remains equal to zero, the radius R₂ is reduced from its value in equation (15) by the magnitude of I_firstwinding when the capacitor 62 begins to discharge (i.e., when V_C1 begins to decrease), and there are affectively two points 111, one above the point 110 and one below the point 110 as discussed above. This situation may also occur if the transistor 60 turns on at the time t₅, and the voltage across the body diode of the transistor 56 causes a nonzero current I_firstwinding to flow before the transistor 60 begins to conduct.

Referring to FIG. 5, one may see that where the duty cycle of the bidirectional converter 18 (FIG. 2) in buck mode is greater than 50%, the plot of the current I_firstwinding and the capacitor voltage V_C1 follows a “distorted” figure eight, starting from point 110, to the point 111, along the curve 112 to the point 116, from the point 116 along the line 118 back to the point 110, from the point 110 to the point 111, along the curve 122 to the point 126, and from the point 126 along the line 128 back to the point 110. Therefore, the current I_firstwinding smoothly transitions from one direction to another through the zero-current point 110; consequently, the current I_{secondwinding} through the second-stage winding 46 likewise smoothly transitions through this zero point. And, as discussed below in conjunction with FIG. 6, such a smooth transition of I_firstwinding and I_{secondwinding} may also occur when the direction of power transfer changes.

FIG. 6 is a combination of the plots of FIGS. 4 and 5, and shows a transition of an embodiment of the bidirectional converter 18 of FIG. 2 from the buck mode to the boost mode and vice-versa in response to a change in the direction of power flow. For example purposes, it is assumed that the inductances L₁ and L₂ of the inductors 50 and 51 of FIG. 2 are approximately equal.

One may see that any change in the direction of power flow between the converter nodes 34 and 36 of FIG. 2 always causes the state of the bidirectional converter 18 to pass though the zero-current point 80, 110, where the currents I_firstwinding and I_{secondwinding} through the first and second-stage transformer windings 44 and 46 are approximately zero. Therefore, transitions between directions of power flow are smooth in that they do not cause, or attempt to cause, a step change in transformer current or capacitor voltage.

For example, referring to FIGS. 1 and 6, assume that the system 10 is an automotive system such as a gas-electric hybrid vehicle, the source/loads 12 and 14 are batteries, and at an arbitrary time while the vehicle is moving, the motor/generator 16 is drawing a current from the battery 14 to rotate the vehicle wheels, and that the bidirectional converter 18 is operating in the boost mode along the curve 82 to transfer power from the battery 12 to the battery 14 so as to regulate V₂ to a desired level.

Next, assume that a driver of the vehicle 10 applies the brakes such that the motor/generator 16 begins sourcing a current to the converter node 36.

In response to this braking, the bidirectional converter 18 may smoothly change the power-transfer direction to charge the battery 12 by smoothly transitioning from the boost mode of operation to the buck mode of operation along the following path: from the curve 82, to the line 88 via the point 86,126, through the cross-over point 80, 110, through the point 111, and to the curve 112.

Referring to FIGS. 1-6, alternative embodiments of the converter 18 are contemplated. For example, although shown as being equal, R₁ need not equal R₂, |V_C01| need not equal |C_CO2|, and thus the magnitudes of θ₁-θ₄ need not be equal; smooth transitions between portions of the switching cycle and from one mode to another mode may still occur even if one or more of such inequalities exist. Furthermore, these parameters may even be unequal for each half circle. For example, R₁ for the curve 82 may be different than R₁ for the curve 92, such that VC₀₁ for the curve 82 may be different than VC_O1 for the curve 92, and θ₁ may be different than θ₂; similarly, R₂ for the curve 122 may be different than R₂ for the curve 112, such that V_C02 for the curve 122 may be different than V_CO2 for the curve 112, and θ₃ may be different than θ₄. Again, smooth transitioning between portions of the switching cycle and from one one mode to the other mode may still occur even if one or more such inequalities exist.

FIG. 7 is a timing diagram of the signals S₁-S₄ and P₁-P₄ of an embodiment of the converter 18 of FIG. 2 while the duty cycle of the converter is less than 50%, and while the signal timing, for example the duty cycle, may be independent of the direction of power flow. As in the embodiment discussed above in conjunction with FIGS. 2 and 3, the duty cycle is defined as the ratio of the high portion of the S₁ switching period to the total S₁ switching period, although it may be defined differently in other embodiment.

Referring to FIGS. 2 and 7, discussed is an operational mode of an embodiment of the converter 18 where the converter has a steady-state duty cycle of less than 50% and is transferring power from the converter node 34 to the converter node 36 (i.e., from the first converter stage 20 to the second converter stage 22). In this mode of operation, the first converter stage 20 operates as a boost converter, and the second converter stage 22 operates as a synchronous full-wave rectifier. Furthermore, the delay periods dd_x are fixed durations that are independent of the duty cycle, and that may be generated by the controller 30 to allow at least some of the transistors to achieve at least approximately zero-voltage switching (ZVS) or zero-current switching (ZCS) as described below. In contrast, the periods Dx depend on the duty cycle. Moreover, the delay periods ddx, the periods Dx, and the times tx do not necessarily correspond to the delay periods ddx, periods Dx, and the times tx of FIG. 3.

At a time t₁, the signal S₁ is inactive low, the signal S₂ is inactive low, the signal S₃ is transitioning from inactive low to active high, and the signal S₄ is active high; therefore, the transistors 54 and 56 are off, the transistor 58 is turning on, and the transistor 60 is on. Furthermore, the signals P₁ and P₄ are transitioning from active high to inactive low, and the signals P₂ and P₃ are inactive low: therefore, the transistors 64 and 70 are transitioning from on to off, and the transistors 66 and 68 are off. Alternatively, the signals P₁ and P₄ may have transitioned from active high to inactive low at a time t₀.

Because the transistor 54 has been off for at least a delay time dd₁ before the transistor 58 turns on, at least a portion of the boost current flowing from the inductor 50 is flowing through the body diode of the transistor 58.

Therefore, while the transistor 58 is turning on, it achieves at least approximately ZVS.

Also regardless of whether the transistors 64 and 70 turn off at time t₀ or at time t₁, any residual current −I_{secondwinding} flowing through the second-stage winding 46 (e.g., due to leakage inductance L_K2 or L_K1 may dissipate through the body diodes of these transistors.

During a period D₁, the signals S₁ and S₂ are inactive low, and the signals S₃ and S₄ are active high; therefore, the transistors 54 and 56 are off, and the transistors 58 and 60 are on. Furthermore, the signals P₁-P₄ are inactive low; therefore, the transistors 64-70 are off. Referring to FIG. 3, the signals P₁ and P₄ generally have the same level as S₄, and the signals P₂ and P₃ generally have the same level as S₃, when the duty cycle of the bidirectional converter 18 is greater than 50%. But if this were the case when the duty cycle of the converter 18 is less than 50%, then there may be periods during which P₁-P₄ are all active high simultaneously, which would cause all of the transistors 64-70 to be on simultaneously, thus shorting out the capacitor 72. Therefore, to prevent this, the controller 30 may insure that P₁ and P₄ are never active high at the same time as P₂ and P₃ are active high. For example, the controller 30 may force P₁-P₄ inactive low whenever both S₃ and S₄ are active high (an embodiment of overlap-protection circuit for realizing this function is described below in conjunction with FIG. 8A).

Therefore, the boost currents from the inductors 50 and 52 charge the capacitor 62 via the on transistors 58 and 60.

Moreover, because the first-stage winding 44 has its end nodes connected together by the on transistors 58 and 60, the currents I_firstwinding and I_{secondwinding} through the first-stage and second-stage windings 44 and 46 decay to zero.

At a time t₂, the signal S₄ transitions to inactive low, thus turning off the transistor 60. Any boost current that was flowing from the inductor 52 into the capacitor 62 via the transistor 60 now flows through the body diode of the transistor 60. Because there is no more than about 0.7 V across the first-stage winding 44, the currents −I_firstwinding and I_{secondwinding} are relatively small, e.g., zero.

Also at the time t₂, the signals P₂ and P₃ transition to active high, thus turning on the transistors 66 and 68, which achieve at least approximately ZCS. Alternatively, the signals P₂ and P₃ may transition to active high, thus turning on the transistors 66 and 68, at a time t₃ and still achieve at least approximately ZCS.

At the time t₄, the signal S₂ transitions active high, thus turning on the transistor 56, which achieves at least approximately ZCS.

During a period D₂, the on transistors 56 and 58 clamp the voltage across the first-stage winding 44 to V_C1, which thus also clamps the voltage across the second-stage winding 46 to V_C1×the turns ratio TR of the transformer 24 as seen from the first-stage side.

Initially during the period D₂, the boost current from the inductor 50 is greater than I_firstwinding, so the excess portion of the boost current continues to flow through the transistor 58 and into the capacitor 62, thus increasing V_C1.

But because the on transistors 56 and 58 clamp V_C1 across the first-stage winding 44, the current I_firstwinding increases as time passes during the period D₂.

Therefore, at a subsequent time during the period D₂, I_firstwinding exceeds the boost current from the inductor 50. Therefore, the excess portion of I_firstwinding, which equals the difference between the boost current from the inductor 50 and I_firstwinding, is sourced by the capacitor 62, thus causing V_C1 to decrease.

Also during the period D₂, a charging current flows through the inductor 52 and the transistor 56 to ground.

At the time t₄, the signal S₂ transitions from active high to inactive low, thus turning off the transistor 56.

Furthermore, at the time t₄ or at a time t₅, the signals P₂ and P₃ transition from active high to inactive low, thus turning off the transistors 66 and 68.

During a delay period dd₃, the boost current through the inductor 52, which during the period D₂ was flowing through the transistor 56, now flows through the body diode of the transistor 60. The duration of the period dd₃ may be at least long enough to allow the body diode of the transistor 60 to begin to conduct.

At the time t₅, the signal S₄ transitions from inactive low to active high, thus turning on the transistor 60. Because the boost current from the inductor 52 is already flowing through the body diode of the transistor 60, this transistor achieves at least approximately ZVS.

During a period D₃, the boost currents from the inductors 50 and 52 charge the capacitor 62 via the on transistors 58 and 60.

Furthermore, because the first-stage winding 44 has its end nodes connected together by the on transistors 58 and 60, the currents I_firstwinding and I_{secondwinding} through the first- and second-stage windings 44 and 46 decay to approximately zero.

At a time t₆, the signal S₃ transitions to inactive low, thus turning off the transistor 58. Any boost current that was flowing from the inductor 50 into the capacitor 62 via the transistor 58 now flows through the body diode of the transistor 58. Because there is no more than about 0.7 V across the first-stage winding 44, the currents I_firstwinding and I_{secondwinding} are relatively small, or are zero.

Also at the time t₆, the signals P₁ and P₄ transition to active high, thus turning on the transistors 64 and 70, which achieve at least approximately ZCS. Alternatively, the signals P₁ and P₄ may transition to active high, thus turning on the transistors 64 and 70, at a time t₇, and still achieve at least approximately ZCS.

At the time t₇, the signal S₁ transitions to active high, thus turning on the switch 54, which achieves at least approximately ZCS.

During a period D₄, the on transistors 54 and 60 clamp the voltage across the first-stage winding 44 to −V_C1, which thus also clamp the voltage across the second-stage winding 46 to −V_C1×the turns ratio TR of the transformer 24 as seen from the first-stage side.

Initially during D₄, the boost current from the inductor 52 is greater than −I_firstwinding, so the excess portion of the boost current continues to flow through the transistor 60 and into the capacitor 62, thus increasing V_C1.

But because the on transistors 54 and 60 clamp −V_C1 across the first-stage winding 44, the magnitude of the current −I_firstwinding increases as time passes during the period D₄.

Therefore, at a subsequent time during D₄, the magnitude of −I_firstwinding exceeds the boost current from the inductor 52. Consequently, the excess portion of −I_firstwinding, which equals the difference between the boost current from the inductor 52 and the magnitude of −I_firstwinding, is sourced by the capacitor 62, thus causing V_C1 to decrease.

Also during the period D₄, a charging current flows through the inductor 50 and transistor 54 to ground.

At a time t₈, the signal S₁ transitions from active high to inactive low, thus turning off the transistor 54.

Then the above-described cycle repeats.

Referring again to FIGS. 2 and 7, now described is an operational mode of an embodiment of the converter 18 where the converter has a steady-state duty cycle of less than 50% and is transferring power from the converter node 36 to the converter node 34 (i.e., from the second converter stage 22 to the first converter stage 20). In this mode of operation, the first converter stage 20 operates as a buck converter, and the second converter stage 22 operates as a DC-AC converter. As discussed below, the switching sequence of FIG. 7 allows at least some of the transistors of the bidirectional converter 18 to achieve at least approximately ZVS or ZCS, which may improve the efficiency of the converter as compared to a conventional converter.

At the time t₁, the signal S_i is inactive low, the signal S₂ is inactive low, the signal S₃ is transitioning from inactive low to active high, and the signal S₄ is active high; therefore, the transistors 54 and 56 are off, the transistor 58 is turning on, and the transistor 60 is on. Furthermore, the signals P₁ and P₄ are transitioning from active high to inactive low, and the signals P₂ and P₃ are inactive low: therefore, the transistors 64 and 70 are transitioning from on to off, and the transistors 66 and 68 are off. Alternatively, the signals P₁ and P₄ may have transitioned from active high to inactive low at the time t₀.

Because there is no current flowing through it or its body diode, the transistor 58 achieves at least approximately ZCS.

Also regardless of whether the transistors 64 and 70 turn off at t₀ or t₁, any current I_{secondwinding} still flowing through the second-stage winding 46 may dissipate through the body diodes of the transistors 66 and 68.

During the period D₁, the signals S₁ and S₂ are inactive low, and the signals S₃ and S₄ are active high; therefore, the transistors 54 and 56 are off, and the transistors 58 and 60 are on. Furthermore, the signals P₁-P₄ are inactive low therefore, the transistors 64-70 are off. Referring to FIG. 3, the signals P₁ and P₄ generally have the same level as S₄, and the signals P₂ and P₃ generally have the same level as S₃, when the duty cycle of the bidirectional converter 18 is greater than 50%. But if this were the case when the duty cycle of the converter 18 is less than 50%, then there may be periods during which P₁-P₄ are all active high simultaneously, which would cause all of the transistors 64-70 to be on simultaneously, thus shorting out the capacitor 72. Therefore, to prevent this, the controller 30 may insure that P₁ and P₄ are never active high at the same time as P₂ and P₃ are active high. For example, the controller 30 may force P₁-P₄ inactive low whenever both S₃ and S₄ are active high (an embodiment of a circuit for realizing this overlap-protection function is described below in conjunction with FIG. 8B).

Therefore, the capacitor 62 discharges via the on transistors 58 and 60 to source the buck currents flowing through the inductors 50 and 52 to the node 34. In this state, the first converter stage 20 operates as a current multiplier (a current doubler in this embodiment).

Moreover, because the winding 44 has its end nodes connected together by the on transistors 58 and 60, the currents I_firstwinding and I_{secondwinding} through the windings 44 and 46 decay to substantially zero.

At the time t₂, the signal S₄ transitions to inactive low, thus turning off the transistor 60. Any buck current that was flowing into the inductor 52 from the capacitor 62 via the transistor 60 now flows through the body diode of the transistor 56.

Also at the time t₂, the signals P₂ and P₃ transition to active high, thus turning on the transistors 66 and 68, which achieve at least approximately ZCS. Alternatively, the signals P₂ and P₃ may transition to active high, thus turning on the transistors 66 and 68, at the time t₃. In this alternative scenario, because approximately the voltage V_C1 is across the first-stage winding 44 (via the transistor 58 and the body diode of the transistor 54), a current −I_firstwinding may begin to flow through the first-stage winding, and thus a current I_{secondwinding} may begin to flow through the second-stage winding 46. However, I_{secondwinding} will flow through the body diodes of the transistors 66 and 68 such that when they turn on, they will achieve at least approximately ZVS.

At the time t₃, the signal S₂ transitions active high, thus turning on the switch 56, which achieves at least approximately ZVS due to the inductor 52 buck current flowing through its body diode per above.

During the period D₂, the on transistors 66 and 68 clamp the voltage across the second-stage winding 46 to V₂, which thus also clamps the voltage across the first-stage winding 44 to V₂×÷ the turns ratio TR of the transformer 24 as seen from the second-stage side.

Initially during the period D₂, the buck current through the inductor 50 is greater than −I_firstwinding, so the excess portion of the buck current continues to flow through the on transistor 58 and discharge the capacitor 62, thus decreasing V_C1.

But because the on transistors 66 and 68 clamp V₂ across the first second-stage winding 46, the currents −I_{secondwinding} and −I_firstwinding increase as time passes during the period D₂.

Therefore, at a subsequent time during the period D₂, −I_firstwinding exceeds the buck current into the inductor 50. Therefore, the excess portion of −I_firstwinding, which equals the difference between the buck current into the inductor 50 and the magnitude of −I_firstwinding, flows into the capacitor 62, thus causing V_CS to increase.

Also during the period D₂, a discharge current flows from ground through the transistor 56 and the inductor 52.

At the time t₄, the signal S₂ transitions from active high to inactive low, thus turning off the transistor 56.

Furthermore, at the time t₄ or t₅, the signals P₂ and P₃ transition from active high to inactive low, thus turning off the transistors 66 and 68.

During the delay period dd₃, the buck current through the inductor 52, which during the period D₂ was flowing through the transistor 56, now flows through the body diode of the transistor 56. Alternatively, the turning off of the transistor 56 may be delayed until the time t₅ to reduce the amount of time during which the buck current flows through the body diode of this transistor, which may improve the efficiency of the converter 18.

At the time t₅, the signal S₄ transitions from inactive low to active high, thus turning on the transistor 60. Because no current is flowing through the transistor 60 or its body diode, this transistor achieves at least approximately ZCS

During the period D₃, the buck currents into the inductors 50 and 52 discharge the capacitor 62 via the on transistors 58 and 60. Therefore, in this state, the first converter stage 20 operates as a current multiplier.

Furthermore, because the first-stage winding 44 has its end nodes coupled together by the on transistors 58 and 60, the currents −I_firstwinding and −I_{secondwinding} decay to approximately zero (the current −I_{secondwinding} may decay through the body diodes of the transistors 64 and 70).

At the time t₆, the signal S₃ transitions to inactive low, thus turning off the transistor 58. Any buck current that was flowing into the inductor 50 from the capacitor 62 via the transistor 58 now flows through the body diode of the transistor 54.

Also at the time t₆, the signals P₁ and P₄ transition to active high, thus turning on the transistors 64 and 70, which achieve at least approximately ZCS. Alternatively, the signals P₁ and P₄ may transition to active high, thus turning on the transistors 64 and 70, at the time t₇. In this alternative, because the voltage V_C1 is across the first-stage winding 44 (via the transistor 60 and the body diode of the transistor 54), a current −I_firstwinding may begin to flow through the first-stage winding, and thus a current −I_{secondwinding} may begin to flow through the second-stage winding 46. However, −I_{secondwinding} will flow through the body diodes of the transistors 64 and 70 such that when they turn on at the time t₇, they will achieve at least approximately ZVS.

At the time t₇, the signal S₁ transitions active high, thus turning on the switch 54, which achieves at least approximately ZVS due to the buck current into the inductor 50 flowing through its body diode per above.

During the period D₄, the on transistors 64 and 70 clamp the voltage across the second-stage winding 46 to V₂, which thus also clamps the voltage across the first-stage winding 44 to V₂ the turns ratio TR of the transformer 24 as seen from the second-stage side.

Initially during the period D₄, the buck current into the inductor 52 is greater than I_firstwinding, so the excess portion of the buck current continues to flow through the transistor 60 from the capacitor 62, thus decreasing V_c1.

But because the on transistors 64 and 70 clamp V₂ across the second-stage winding 46, the magnitude of the current I_{secondwinding}, and thus the magnitude of the current I_firstwinding, increase as time passes during the period D₄.

Therefore, at a subsequent time during the period D₄, the magnitude of I_firstwinding exceeds the buck current into the inductor 52. Consequently, the excess portion of I_firstwinding, which equals the difference between the buck current into the inductor 52 and the magnitude of I_firstwinding, flows into (i.e., charges) the capacitor 62, thus causing V_C1 to increase.

At the time t₈, the signal S₁ transitions from active high to inactive low, thus turning off the transistor 54.

Furthermore, the transistors 64 and 70 turn off at either time t₈ or t₉.

Alternatively, the controller 30 may transition the signal S₁ to inactive low at the time t₉ to reduce the time during which the inductor 50 buck current flows through the body diode of the transistor 54.

Then the above-described cycle repeats.

Still referring to FIGS. 2 and 7, the bidirectional converter 18, when operating with a duty cycle of less than 50%, may provide advantages similar to those described above in conjunction with FIG. 3 for operation with a duty cycle of greater than 50%, such as each transistor achieving at least approximately ZVS or ZCS when the controller 30 switches it, clamping of the windings 44 and 46 to V_C1 and V₂, respectively, and smooth transition between power-flow directions in a manner similar to that described above in conjunction with FIG. 6.

Still referring to FIGS. 2 and 7, alternate embodiments of the operation with duty cycle of less than 50% are contemplated. For example, any of the delay periods dd_x may be eliminated, for example, if an eliminated delay period is not needed to allow one or more of the transistors to achieve at least approximately ZVS or ZCS. Furthermore, the timing of one of the signals S₁-S₄ and P₁-P₄ may be adjusted, for example, to improve the efficiency of the converter 18.

Referring again to FIGS. 2, 3, and 7, the bidirectional converter 18 may also operate with a duty cycle of approximately 50%. In such an operating mode, the overlap between the signals S₁ and S₂ would be small or zero, as would the overlap between the signals S₃ and S_q. Consequently, one or more of the transistors 54-60 may be unable to achieve at least approximately ZVS, and the currents I_firstwinding and I_{secondwinding} through the windings 44 and 46 may have insufficient time to decay to approximately zero, in which case one or more of the transistors 64-70 may be unable to achieve at least approximately ZCS. But in at least most cases, the current I_{secondwinding} through the winding 46 that would have otherwise decayed to approximately zero allows the transistors 64-70 to achieve at least ZVS. It is believed, however, that in most applications, the amount of time that the converter 18 will operate at approximately 50% duty cycle is relatively small compared to the total operating time; therefore, it is believed that the converter 18 may operate with an overall higher efficiency than a conventional bidirectional converter, even if it operates with an approximately 50% duty cycle during some periods.

FIGS. 8A and 8B are a schematic diagram of the converter stage 20 and 22 and transformer 24 of FIG. 2, an embodiment of the controller 30 of FIG. 1, a current-sense circuit 150, and the source/loads 12 and 14 of FIG. 1 where these source loads are respective batteries.

The controller 30 includes control circuitry 152 for generating the switching signals S₁-S₄ and P₁-P₄, and the controller may also include the current-sense circuit 150.

In operation of the current-sense circuit 150, a transformer 154 couples the current I_firstwinding through the first-stage winding 44 of the transformer 24 to a bridge 156 and voltage divider 158, which generates a voltage VIsense that is proportional to a current flowing through the first-stage winding 44, and that has twice the switching frequency (i.e., twice the frequency of any of the switching signals S₁-S₄ and P₁-P₄).

In operation of the control circuitry 152 during a mode of operation where the motor/generator 16 (FIG. 1) is operating as a motor, or is generating a relatively small output current, a PID circuit 160 conventionally provides compensation by generating an error signal V_error from a voltage V_feedback from a voltage-control loop, wherein V_feedback is generated by a voltage divider 162 to be proportional to V₂, and a capacitor 164 and resistors 166 and 168 set the compensation of a current control loop. A comparator 170 compares V_error to an externally provided signal Vext-ramp (Vext-ramp may be a PWM ramp from which the controller 30 of FIG. 1 may generate the signals S₁-S₄ and P₁-P₄ per below), and generates a PWM Control signal in response to this comparison. In response to the PWM Control signal, a first logic circuit 172 generates the switching signals S₁-S₄, and in response to the signals S₃ and S₄, a second logic circuit 174 generates the switching signals P₁-P₄ such that P₁ and P₄ have little or no overlap with P₂ and P₃ (preventing such overlap may prevent V₂ from be connected directly to e.g., ground).

In operation of the control circuitry 152 during a boost mode of operation when the motor/generator 16 (FIG. 1) is generating an intermediate amount of current that causes the “−” input of the amplifier 176 to become higher than a reference voltage Vref1, then an amplifier 176 pulls down the inverting input node of the comparator 170 to decrease the duty cycle of the PWM control signal, and to thus decrease the duty cycle of S₁ and S₂. This action allows the bidirectional converter 18 to react relatively quickly to a relatively sudden, but relatively modest, increase in V₂ so as to maintain V₂ in regulation by transferring excess charging current (current not needed to charge the battery 14) from the motor/generator 16 to the battery 12.

In operation of the control circuitry 152 during a boost mode of operation when the motor/generator 16 (FIG. 1) is generating a relatively high amount of voltage that causes V_feedback to become higher than a reference voltage Vref2, an amplifier 178 pulls down the inverting input node of the comparator 170 to decrease the duty cycle of the PWM control signal, and to thus decrease the duty cycle of S_i and S₂. This action allows the bidirectional converter 18 to react even more quickly to a relatively sudden and significant increase in the voltage V₂ so as to maintain V₂ in regulation by transferring excess charging current (current not needed to charge the battery 14) from the motor/generator 16 to the battery 12.

Furthermore, a current-limit circuit 182 may prevent the current into or out from the converter node 36 from exceeding a maximum safe value. The current-limit circuit 182 may also limit the current charging the battery 14 to a maximum safe value.

During a buck mode of operation, the control circuit 152 may operate similarly, but to increase the duty cycle of S₁ and S₂.

Alternate embodiments of the controller 30 (and current-sense circuit r 150 if not part of the controller) are contemplated. For example, any number of the components of the controller 30 may be disposed on a same or different integrated circuit (IC), and at least one of these ICs may also include at least one of the other components (e.g., one or more of the transistors 54-60 and 64-70) of the converter 18. For example, at least the comparator 170, logic circuit 172, and amplifier 176 may be disposed on an Intersil® ISL6742 power-supply-controller IC.

FIG. 9 is a schematic diagram of an embodiment of a bidirectional converter 190, where like numbers refer to components common to the bidirectional converter 18 of FIG. 2. The converter 190 is similar to the converter 18 except that the second converter stage 22 of FIG. 2 is replaced with a voltage-doubling stage 192, which effectively doubles the voltage-boosting and voltage-dividing capability of the converter 190 as compared to the converter 18.

The voltage-doubling stage 192 is similar to the second converter stage 22 of FIG. 2 except that the transistors 66 and 70 of the second converter stage 22 are replaced with capacitors 194 and 196 (alternatively, the transistors 66 and 70 may remain, and the transistors 64 and 68 may be replaced with the capacitors). A potential benefit of the stage 192 is that a higher boost or buck ratio may be achieved without increasing the turns ratio of the transformer 24.

In operation, an embodiment of the converter 190 operates similarly to an embodiment of the converter 18 of FIG. 2 as described above in conjunction with FIGS. 2-7 except for the below-described differences, where it is assumed for example purposes that the capacitors 194 and 196 have approximately the same capacitances.

During a boost operating mode while the converter 190 is transferring power from the converter node 34 to the converter node 36 (e.g., charging the battery 14 of FIG. 1 from the battery 12), in a first portion of the switching cycle (e.g., period D₃ of FIG. 3), the controller 30 generates P₁ active high and P₃ inactive low to turn on the transistor 64 and turn off the transistor 68 such that the current −I_{secondwinding} charges the capacitor 194 to a voltage approximately equal to V_C1×the turns ratio TR of the transformer 24 as seen from the first-stage side. Similarly, in a second portion of the switching cycle (e.g., period D₁ of FIG. 3), the controller 30 generates P₁ inactive low and P₃ active high to turn off the transistor 64 and turn on the transistor 68 such that the current I_{secondwinding} charges the capacitor 196 to a voltage approximately equal to V_C1×the turns ratio of the transformer 24.

Therefore, during a boost mode of operation, the converter 190 generates V₂≈2V_C1×(turns ratio of transformer 24), which, for a same value of V_C1, is double the value that the converter 18 of FIG. 2 generates for V₂. Therefore, for a same value of V₂, the converter 190 may allow the turns ratio of the transformer 24 to be reduced by up to ½ as compared to the turns ratio of the transformer 24 of the converter 18.

During a buck operating mode while the converter 190 is transferring power from the converter node 36 to the converter node 34 (e.g., charging the battery 12 of FIG. 1 from the battery 14 or with the motor/generator 16), in a first portion of the switching cycle (e.g., period D₃ of FIG. 3), the controller 30 generates P₁ active high and P₃ inactive low to turn on the transistor 64 and turn off the transistor 68 such that the on transistor 64 couples the capacitor 194 across the second winding 46 of the transformer. Therefore, the voltage across the second-stage winding 46 is clamped to approximately V₂/2, and the voltage across the first-stage winding 44 of the transformer is clamped to approximately (V₂/2)×(turns ratio of the transformer 24). Similarly, in a second portion of the switching cycle (e.g., period D₁ of FIG. 3), the controller 30 generates P₁ inactive low and P₃ active high to turn off the transistor 64 and turn on the transistor 68 such that the on transistor 68 couples the capacitor 196 across the second-stage winding 46 of the transformer. Therefore, the voltage across the second-stage winding 46 is again clamped to approximately V₂/2, and the voltage across the first-stage winding 44 is again clamped to approximately V₂/(2×(turns ratio of the transformer 24 as seen from the second-stage side)).

Therefore, during a buck mode of operation, the converter 190 generates V_C1≈V₂/(2×(turns ratio of transformer 24)), which, for a same value of V₂, is half the value that the converter 18 of FIG. 2 generates for V_C1. Therefore, for a same value of V₂, the converter 190 may allow the turns ratio of the transformer 24 to be reduced by up to ½ as compared to the turns ratio of the transformer 24 of the converter 18.

Still referring to FIG. 9, alternate embodiments of the bidirectional converter 190 are contemplated. For example, one or more of the embodiments discussed above for the bidirectional converter 18 of FIGS. 2 and 8A may be applicable to the converter 190.

FIG. 10 is a schematic diagram of an embodiment of a bidirectional converter 200 having N phases, where N may be greater than two, and where like numbers reference components common to the bidirectional converters 18 and 190 of FIGS. 2 and 9, respectively. As compared to the bidirectional converters 18 and 190, the converter 200 may produce less ripple on the voltages V₁ and V₂, and may have a smaller current per phase for a given output/input current. Such an N-phase structure may be a promising candidate for high-power applications.

The converter 200 includes a first converter stage 202, a second converter stage 204, and transformers 24₁-24_N/2.

The first stage 202 includes the capacitor 62 and a number of two-phase substages 206 that may each have a topology and operation that are respectively similar to the topology and operation of the first converter stage 20 of FIG. 2.

The second converter stage 204 includes N/2 half-bridges each formed from a respective pair of transistors 64₁-64_N/2 and 68₁-68_N/2, and the capacitor 72.

The transformers 24₁-24_N/2 may each have a respective core, or may share a common core.

Alternate embodiments of the bidirectional converter 200 are contemplated. For example, the second stage 204 may include a voltage multiplier, such as, for example, a voltage doubler similar to that formed by the capacitors 194 and 196 of FIG. 9. Furthermore, one or more of the transistors 64 and 68 may be replaced with a diode.

From the foregoing it will be appreciated that, although specific embodiments have been described herein for purposes of illustration, various modifications may be made without deviating from the spirit and scope of the disclosure. Furthermore, where an alternative is disclosed for a particular embodiment, this alternative may also apply to other embodiments even if not specifically stated.

Claims

1. A controller, comprising: circuitry operable to cause a bidirectional signal converter: to regulate a first signal at a first converter node; andto have a switch timing that is independent of a direction of power transfer between the first converter node and a second converter node.

2. The controller of claim 1 wherein the circuitry is further operable to cause the bidirectional signal converter to regulate the first signal to have a first level that is different than a second level of a second signal at the second converter node.

3. The controller of claim 1 wherein the circuitry is further operable to cause the bidirectional signal converter to regulate the first signal to have a first level that is higher than a second level of a second signal at the second converter node.

4. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as a signal rectifier while the direction of power transfer is from the second converter node to the first converter node.

5. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to be synchronized with a second stage of the bidirectional signal converter that is coupled to the second converter node.

6. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as an alternating-signal source while the direction of power transfer is from the first converter node to the second converter node.

7. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as a multiphase alternating-signal source while the direction of power transfer is from the first converter node to the second converter node.

8. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as an alternating-signal source with fewer than all of multiple phases active while the direction of power transfer is from the first converter node to the second converter node.

9. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as a signal divider while the direction of power transfer is from the first converter node to the second converter node.

10. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as a multiphase signal divider while the direction of power transfer is from the first converter node to the second converter node.

11. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as a signal divider with fewer than all of multiple phases active while the direction of power transfer is from the first converter node to the second converter node.

12. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as a signal multiplier while the direction of power transfer is from the second converter node to the first converter node.

13. The controller of claim 1 wherein the circuitry is further operable to cause a first stage of the bidirectional signal converter that is coupled to the first converter node to operate as a signal multiplier with fewer than all of multiple phases active while the direction of power transfer is from the second converter node to the first converter node.

14. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a buck converter while the direction of power transfer is from the first converter node to the second converter node.

15. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a boost converter while the direction of power transfer is from the second converter node to the first converter node.

16. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a multiphase buck converter while the direction of power transfer is from the first converter node to the second converter node.

17. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a multiphase boost converter while the direction of power transfer is from the second converter node to the first converter node.

18. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a buck converter with fewer than all of multiple phases active while the direction of power transfer is from the first converter node to the second converter node.

19. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a boost converter with fewer than all of multiple phases active while the direction of power transfer is from the second converter node to the first converter node.

20. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a current multiplier while the direction of power transfer is from the first converter node to the second converter node.

21. The controller of claim 1 wherein the circuitry is further operable to cause a second stage of the bidirectional signal converter that is coupled to the second converter node to operate as a current multiplier with fewer than all of multiple phases active while the direction of power transfer is from the first converter node to the second converter node.

22. The controller of claim 1 wherein the circuitry is further operable to cause at least one switch circuit of the bidirectional signal converter to transition between switch states while there is approximately zero volts across the switch circuit.

23. The controller of claim 1 wherein the circuitry is further operable to cause at least one switch circuit of the bidirectional signal converter to transition between switch states while there is no more than approximately a diode voltage drop across the switch circuit.

24. The controller of claim 1 wherein the circuitry is further operable to cause at least one switch circuit of the bidirectional signal converter to transition between switch states while there is approximately zero current flowing through the switch circuit.

25. The controller of claim 1 wherein the circuitry is inoperable to receive a power-transfer-direction signal.