Phase-shifted full-bridge zero-voltage PWM soft switching circuit

    Abstract: This article introduces the composition and working principle of the phase-shifted full-bridge zero-voltage PWM soft switching circuit, analyzes the working process of the soft switching in detail in the time domain, explains the resonance process of the leading arm and the lagging arm, and finally gives the PWM The reason for the loss of duty cycle of the switching circuit and the energy conversion method of the circuit.

    Keywords: phase-shifted soft switching resonance technology

Communication switching power supplies are now developing towards the trend of high power, small size and high efficiency. The current research hotspot in this field is the use of new soft switching circuits to reduce switching losses and increase switching frequency. In recent years, my country's communication switching power supply has experienced four stages: PWM hard-switching converter, resonant converter, zero-voltage quasi-resonant converter and full-bridge zero-voltage PWM soft-switching converter.

1 Soft switching circuit

Soft switching can be divided into three switching forms: zero current switching (ZCS), zero voltage switching (ZVS) and zero voltage zero current switching (ZV-ZCS). There are also two types: soft turn-on and soft turn-off. Ordinary PWM converters adjust the output voltage by changing the pulse width of the drive signal, and there is a large loss during the switching period of the power switch tube. Therefore, this hard switching power supply has large peak interference, poor reliability, and low efficiency. The phase-shift control full-bridge soft-switching power supply adjusts the output voltage by changing the phase shift angle of the upper and lower tube driving voltages on the diagonal lines of the two arms. This method makes the super forearm tube grid voltage lead the lagging arm tube grid voltage by one. Phase, and set different dead time for the two anti-phase driving voltages of the same bridge arm at the IC control end. At the same time, the leakage inductance of the transformer and the junction capacitance and parasitic capacitance of the power tube are cleverly used to complete the resonance process to achieve zero-voltage turn-on. , thus staggering the hard switching state in which the current and voltage of the power device are at high values ​​at the same time, effectively overcoming the shortcomings of inductive turn-off voltage spikes and excessive tube temperature during capacitive turn-on, and reducing switching losses and interference.

The characteristics of this soft switching circuit are as follows:

(1) The phase-shifted full-bridge soft switching circuit can reduce switching losses and improve circuit efficiency.

(2) Since the du/dt of the open pass is reduced, parasitic oscillation is eliminated, thereby reducing the ripple of the power output, which is beneficial to the simplification of the noise filter circuit.

(3) When the load is small, zero-voltage switching cannot be achieved due to insufficient resonant energy, so the efficiency will drop significantly.

(4) This soft switching circuit has a duty cycle loss phenomenon, which is more serious under heavy load. In order to achieve the required maximum output power, the change must be appropriately reduced, which will lead to an increase in the primary current and increase the burden on the switching device. .

(5) Since the resonant inductor and the output rectifier diode junction capacitance form an oscillation, the rectifier diode needs to withstand a higher peak voltage.

2 Working principle

The actual circuit of phase-shifted full-bridge zero-voltage PWM soft switching is shown in Figure 1. It consists of 4 switching power tubes S1, S2, S3, S4 (MOSFET or IGBT), 4 high-speed switching diodes D1, D2, D3, D4 connected in reverse parallel and 4 parallel capacitors C1, C2, C3, C4 ( Including the output junction capacitance of the switching power tube and the external absorption capacitance), compared with the hard-switching PWM circuit, this circuit only has one more resonant inductor Lr, which represents the sum of the leakage inductance of the transformer and the independent inductance. The essence of zero-voltage switching is to use the charging and discharging of parallel capacitors during the resonance process to quickly increase the voltage UA or UB of a certain bridge arm to the power supply voltage or drop to zero, thereby causing the parallel diode to be turned on on the same bridge arm. It is turned on and clamps the terminal voltage of the tube at 0 to create conditions for ZVS. The switching control waveforms of the four switching power tubes in the circuit are shown in Figure 2.

The waveform is divided into 8 intervals according to the time domain within one cycle, and each interval represents a process of circuit operation. Except for the dead time, there are always two switches in the circuit that are turned on at the same time; there are four configurations: S1 and S4, S1 and S3, S2 and S3, S2 and S4, and the cycle starts again and again. It can be seen from Figure 2 that when S1 and S4, S2 and S3 are combined, that is, the T0-T1, T4-T5 time periods are the output power state of the working circuit, and when S1 and S3, S2 and S4 are combined, that is, T2-T3, The T6-T7 time period is the freewheeling state of the circuit; the T3-T4 and T7-T8 time periods are the resonance process from the freewheeling state to the output power; the T1-T2 and T5-T4 time periods are the resonant process from the output power state to the output power state. The last four intervals of the resonance process of freewheeling state transition are called dead zones. The resonance processes all occur in the dead zone, and the dead zone time is set by the controller.

The following is a detailed analysis of the working principles of each interval.

2.1 Output power state 1 (T0-T1)

If the initial state is in the T0-T1 interval, then at this moment, the power switches S1 and S4 are in the on state, the voltage between points A and B is U, the primary current rises linearly from the initial Ip point, and the secondary induction of the transformer The voltage will turn on DR2 and turn off DR1. The output current will flow to the output inductor through DR2 and provide current to the load after the capacitor stores energy. When the time T1 is reached, the output power state 1 process ends.

2.2 Super forearm resonance process 1 (T1-T2)

When the T1 moment arrives, the switch S4 changes from on to off. The energy stored in the inductor charges C4. At the same time, C3 discharges so that the voltage at point B gradually increases. When the voltage of C4 is charged to U, D3 conducts When the switch is turned on, the source-drain voltage of the switching power transistor S3 is 0, thus preparing the conditions for the zero-voltage turn-on of the switching power transistor S3. Because the secondary output inductor participates in resonance and the equivalent inductance is k2L, the inductor has sufficient energy storage and can easily make point B reach the U value, so the super forearm can easily achieve zero-voltage turn-on.

In this process, the capacitance participating in the resonance is the parallel connection of C3 and C4, and the inductance is the series inductance of Lr and secondary induction. Right now:

C=C3+C4,L=Lr+k2L

The differential equation of the super forearm resonance process is as follows:

LC（d2Uc/dt2）+Uc=kU0

Among them, Uc(0)=U, iLr(0)=I0/k in the initial state.

2.3 Freewheeling state 1 (T2-T3)

Since the switching power transistors S1 and S3 are both turned on, the potentials of point A and point B are both U at this time, and the transformer is initially in a short-circuit state and does not output power. From the T2 moment, the polarity of the voltage at both ends of the output inductor L changes, the output inductor changes from the energy storage state to the energy discharge state, the load is provided with current by the output inductor and the output capacitor, and the corresponding primary current of the transformer still flows in the original direction. , after entering the freewheeling state, the current drops slightly. The initial current of the transformer passes through the switching power tube and diode to reduce the loss of the switching power tube.

2.4 Lagging arm resonance process 1 (T3-T4)

When T3 arrives, switch S1 changes from on to off, and the energy storage inductor begins to charge C1. At the same time, capacitor C2 begins to discharge, causing the voltage at point A to gradually decrease until the voltage of C2 reaches 0, causing D2 to turn on. This prepares the conditions for the zero-voltage conduction of the switching power transistor S2. In this process, the capacitance participating in the resonance is the parallel connection of C1 and C2, and the inductance is only Lr, that is, C=C1+C2, L=Lr

The differential equation of the lagging arm resonance process is:

LC（d2Uc/dt2）+Uc=0

Among them, Uc(0)=0, iLr(0)=I0/k in the initial state.

In this process, since only Lr participates in resonance, and if the current Ilr of Lr is small at the beginning of resonance and the energy storage of Lr is not enough, the peak value of the resonance voltage Uc of the capacitor C may not reach U, so the diode will not be able to If it is turned on, its corresponding switch cannot achieve zero-voltage turn-on. In order for the peak resonance voltage of the capacitor to reach U, the energy storage of the inductor must be high enough, so at the beginning of resonance, the current Ilr of the inductor Lr must satisfy:

1/2(Li2Lr)=1/2(CU2)

This equation is the basis for designing the resonant inductor Lr.

2.5 Output power state 2 (T4-T5)

During this process, the switching power tubes S2 and S3 are turned on, the initial current of the transformer flows from B to A, and the voltage at two points AB is -U. The secondary induced voltage of the transformer makes DR1 conductive, and stores the output inductor and capacitor through DR1. able.

2.6 Super forearm resonance state 2 (T5-T6)

During this process, the switching power transistor S3 changes from on to off, the capacitor C3 begins to charge, the capacitor C4 begins to discharge, and the voltage at point B gradually drops to 0, preparing the conditions for the zero-voltage turn-on of the switching power transistor S4.

2.7 Freewheeling state 2 (T6-T7)

At this time, the voltage across A and B is 0, the primary current flows in the original direction, and the current intensity gradually decreases. DR2 on the secondary side of the transformer is still in a conductive state to maintain the current provided by the inductor to the load.

2.8 Lagging arm harmonic process 2 (T7-T8)

At time T7, the switching power transistor S2 changes from on to off, the capacitor C2 begins to charge, and the capacitor C1 begins to discharge, causing the voltage at point A to gradually rise to U, so that the diode D1 conducts, turning on the zero voltage of the switching power transistor S1. Conditions are prepared. At this point, a cycle ends.

3 Circuit analysis

3.1 Comparison of two resonance processes

During the resonance process when the output power state transitions to the freewheeling state, due to its large inductance (L=Lr+k2L) and large energy storage, the load current can cause the capacitor voltage to resonate to zero at a very small time. Therefore, the phase-advanced The two bridge arm switches S3 and S4 can easily achieve zero-voltage turn-on.

In the resonance process of converting from freewheeling state to output power state, the inductance is small, and only Lr participates in the resonance. Therefore, the energy storage is small. Only when the load current zero reaches a certain value can the capacitor voltage resonate to U. Therefore, it is not easy for the two bridge arms S1 and S2 with phase lag to achieve zero-voltage turn-on.

In order to make the latter easily realize zero-voltage turn-on, when designing the switching power tube control signal, the dead time of the lagging arm should be greater than the dead time of the super forearm, and the values ​​of C1 and C2 should be smaller than C3 and C4.

3.2 Duty cycle loss phenomenon

A special phenomenon of the phase-shifted full-bridge zero-voltage PWM soft switching circuit is the loss of duty cycle. It always occurs at the end of the transition from freewheeling state to output power state. At time T4, the switching power transistor S2 has just been turned on, and the current of the resonant inductor Lr has just attenuated to zero or has not yet decayed to zero. The primary of the transformer is in a freewheeling state, and the voltage at both ends is zero. The voltage that the resonant inductor Lr withstands is U, and its The current gradually increases in the opposite direction. Only when its current increases to I0/k, the transformer exits the freewheeling state, the voltage at both ends rises to U, and the current in the inductor Lr stops increasing. In this way, from the time S2 is turned on until the transformer exits the freewheeling state, the transformer does not output voltage. This period of time is the lost duty cycle, and its duty cycle is:

ΔD=2LrI0/ (kUT)

It can be seen from the formula that the larger the resonant inductance Lr is, the larger the load current I0 is and the more serious the duty cycle loss is. The loss of duty cycle will directly lead to an increase in the loss of the switching power tube, so measures must be taken to overcome it. Currently, this is usually achieved by reducing the transformation ratio.

3.3 Energy conversion

This phase-shifted full-bridge zero-voltage PWM soft switching circuit adds a resonant inductor in series with the primary side of the main transformer (primary side), thereby promoting the realization of ZVS in the lagging arm of the circuit. Because the charging and discharging energy of two parallel capacitors on the same bridge arm during switching will reach Wc=1/2 (CU2), that is, the capacitor energy storage change after one charge and one discharge reaches CU2. Such a large electric field energy needs to be used in the inductor. of magnetic energy to convert. In order to successfully complete the charging and discharging of the parallel capacitor, the parallel diode is conductive and clamped. A large enough inductance is designed in the circuit to help the charge in the capacitor change. This is the role of Lr and L in the circuit.

4 Conclusion

    Compared with the ordinary full-bridge, the phase-shifted full-bridge soft-switching circuit has the advantages of low switching loss and high circuit efficiency, but it also has shortcomings. How to improve its shortcomings while retaining its advantages requires further research.