Analysis of Typical Topologies of Three-Level Soft-Switching DC Converters

1 Introduction

In recent years, people's requirements for the voltage and power levels of power electronic devices have been continuously improved. As a solution to this trend, three-level converters have attracted more and more attention. Three-level converters ^[1] greatly reduce the voltage level of the switch tube, which is conducive to reducing switching losses, improving efficiency and reducing costs. In order to reduce the size and weight of the converter, high frequency has always been the goal pursued by power electronics. With the high frequency, the switching loss problem of power devices has become an increasingly prominent contradiction. Therefore, soft switching technology has emerged as an important means to reduce switching losses, improve system efficiency and improve EMI problems.

The three-level zero-voltage soft-switching DC converter is a new and practical topology that has emerged as a result. It uses phase-shift control technology and the resonance of the junction capacitance of the switch tube and the leakage inductance of the transformer to achieve zero-voltage switching of the switch tube. The output capacitance at both ends of the power switch tube is charged and discharged through the high-frequency transformer leakage inductance energy storage to make the voltage at both ends of the switch tube drop to zero, so that the four switch tubes of the converter are turned on at zero voltage in turn, and turned off at zero voltage under the action of the buffer capacitor, thereby effectively reducing the switching loss and switching noise of the circuit, reducing the electromagnetic interference during the switching process of the device, and providing good conditions for the converter to increase the switching frequency, improve efficiency, and reduce size and weight.

However, in practical applications, three-level zero voltage soft switching (ZVS) converters have several difficult problems to overcome, resulting in a series of improved topologies. To this end, this paper systematically summarizes and analyzes the more practical and typical three-level zero voltage soft switching converter topologies.

2 Advantages and disadvantages of traditional three-level zero-voltage soft-switching DC converter

The traditional three-level ZVS soft switching DC/DC converter (TL-ZVS DC/DC converter) is shown in Figure 1. Its topological feature ^[2] is the introduction of a large-capacity flying capacitor Css _. When the converter is working, its voltage is stabilized at Vin _/ 2, making the conditions for the leading tube and the lagging tube to achieve soft switching independent of each other and non-interfering with each other; and the combination of phase shifting technology and soft switching technology can effectively reduce the loss in the circuit and improve efficiency. Therefore, it is very suitable for high input voltage and high power occasions.

Figure 1 Conventional three-level phase-shifted full-bridge ZVS converter

However, the traditional three-level ZVS soft-switching DC converter also has many problems.

1) It is difficult for the lagging arm to achieve soft switching under light load conditions, making it unsuitable for applications with large load changes;

2) The circulating current energy is greatly increased. The higher the input _Vin , the lower the converter efficiency, because the higher _the Vin , the longer the zero state time. In the zero state, the primary current is in a natural freewheeling state, and no energy is transferred from the primary side to the output stage, but there is conduction loss in the transformer, resonant inductor and switch tube.

3) Due to the existence of resonant inductance, the secondary side of the transformer has duty cycle loss. The larger the transformer leakage inductance L _1k , the larger the duty cycle loss D _loss . D _loss reduces the secondary duty cycle D _sec .

4) The voltage spike of the secondary rectifier diode is large.

3 Improved topology analysis

3.1 Implementation of light-load soft switching of the lagging arm

Referring to Figure 1, in order to improve the zero voltage switching load range of the lagging arm of the traditional three-level FB-ZVS converter, one of the most direct methods is to increase the leakage inductance of the transformer or connect an inductor Lr in series with the primary side of the _transformer to increase the energy storage of the resonant inductor, so that the parallel capacitor of the lagging arm switch tube can be fully charged and discharged under light load, and the zero voltage conduction of the lagging arm switch tube can be achieved. However, this has the following disadvantages.

_{1 ) Circulating current energy further increases Assume that the zero voltage conduction load range of the converter is Io ≥ Iomin, Iomin = 20% Iomax ( Iomax is} the output current value when the converter is fully loaded ₎ . When the converter is running at 20% load, the inductor energy stored when the lagging arm switch is turned off is

E _min =( L _lk ＋ L _r ) I _omin ² /(2 n ² )

When running at full load, the inductor stores energy

E _max =( L _lk ＋ L _r ) I _omax ² /(2 n ² )

Thus there is

=25

This means that when the system is running at full load, the circulating current energy of the system will be 25 times the actual energy required for the zero voltage conduction of the lagging arm switch tube, which will directly lead to a significant increase in the conduction loss of the converter.

2) Further increase the secondary voltage duty cycle ΔD loss The main reason _is that the increase in inductance causes the slope of the primary current changing from one direction to another, _Vin / ( Llk + Lr ₎ , to decrease. During the secondary rectifier commutation process, the two diodes are turned on at the same time, the secondary voltage is clamped at zero, the voltage Vab rises to the power supply voltage _Vin , _and the primary current can be approximately regarded as changing linearly with a slope of _Vin / Llk . _The smaller the slope, the longer the interval of the change time period, and the greater the duty cycle loss. From ΔD = , it can be seen that the duty cycle loss increases ( fs is the switching frequency) _.

3) It aggravates the parasitic oscillation between the leakage inductance and the junction capacitance of the secondary rectifier diode, increasing the withstand voltage of the secondary rectifier tube.

3.2 Introducing filter inductor resonance to expand zero voltage switching load range

As shown in Figure 2, the improved topology uses two saturated inductors S5 _and S6 _as switches, which are connected in series with the anode of the output rectifier diode. The shutdown process of the leading arm is the same as that of the traditional three-level ZVS converter. This topology prevents the secondary rectifier diode from being turned on at the same time during the short period when the lagging arm switch tube is switching state, so the output filter inductor _{n2Lout can} be used to participate in the resonance. Since the output filter inductor is much larger than the transformer leakage inductance , the zero voltage load range of the lagging arm switch tube is greatly expanded. The characteristic of this topology is that since the leakage inductance is no longer an essential component for achieving ZVS, it can be very small, so that the duty cycle loss and the parasitic oscillation of the secondary rectifier diode are also greatly reduced.

Figure 2 Three-level phase-shifted full-bridge ZVS converter with output filter inductor

3.3 Using current doubler rectifier circuit to expand zero voltage switching load range

Reference [6] proposed using a phase-shifted controlled three-level current-doubler rectifier zero-voltage switching converter to expand the zero-voltage switching load range. Current-doubler rectification is derived from the full-wave rectification method, that is, two independent, equal-sized inductors are used to replace a group of rectifier tubes in the full-bridge rectifier topology, while still maintaining the form of "full-wave rectification". In essence, two inductors are staggered in parallel. Therefore, in addition to the advantages of the aforementioned circuits, this topology also avoids voltage spikes and voltage oscillations caused by reverse recovery due to the natural commutation of the secondary rectifier diodes. This topology is shown in Figure 3.

Figure 3 Three-level current doubler rectifier phase shift full-bridge ZVS converter