Relevant performance of BIMOSFET used in flyback converter

1 Introduction

A typical application of flyback converter is to provide auxiliary power to IGBT driver in inverter. At this time, the switch tube of flyback converter needs to have relatively high breakdown voltage and fast switching speed. In order to reduce switching loss, the energy of turn-on and turn-off should also be small. One of the main advantages of BIMOSFET is that it has low turn-on loss, and its conduction loss is also relatively small. Comparing MOSFET and BIMOSFET, the loss of BIMOSFET is about 35% smaller.

2 Flyback Operation

The flyback converter is one of the simplest converters. Its circuit only includes a switch tube, a transformer, a diode and two capacitors, as shown in Figure 1. The converter's energy is stored in the air gap of the core. When the switch tube is turned on, the primary current rises ramp, the magnetic core stores energy, and when it is turned off, it is transmitted to the load end through the diode. The maximum power of the flyback converter can reach 300W.

The advantage of this circuit is that it has a very wide input-output voltage ratio and can add auxiliary coils to achieve multiple outputs. In addition, it can achieve good electrical isolation between the primary and secondary sides. Its disadvantages are that the voltage stress of the switch tube is relatively high and the RFI radiation generated by the transformer air gap is relatively high. The flyback converter is not allowed to operate at no load or in open loop, otherwise the output voltage will exceed the allowable limit.

Figure 1 Flyback converter

3 Application of Flyback Converter

One of the main applications of flyback converter is to provide auxiliary power to IGBT driver in inverter. All the requirements in this case can be met by flyback converter.

The shaded part in Figure 2 shows the drive circuit of the inverter, which also includes a startup circuit. Its auxiliary power supply can be composed of very few components and is inexpensive.

Figure 2 Inverter

Since the input voltage of the converter is the DC bus voltage, the voltage variation range is relatively wide. During the charging process of the bus capacitor, the auxiliary power supply must work under the condition of very low DC bus voltage, such as the braking state of the motor. When the DC bus voltage rises to 750V, the output voltage can be easily adjusted by changing the duty cycle of the switch tube.

All isolated DC outputs can be realized by adding independent auxiliary coils, such as 5V to power the microprocessor, positive and negative 15V to the current sensor, positive 15V to drive the three IGBTs below, and three independent positive 15V to drive the upper IGBT.

When the flyback converter is used as an inverter driver, it is important to have a high voltage stress. In the flyback converter, the highest voltage stress of the switch tube is twice the input voltage. Therefore, the minimum withstand voltage of the switch tube should be 2×Vin. As a standard inverter for motor control, its power supply is 400V, and the DC bus voltage is as high as 750V when the motor is in braking state. Therefore, a switch tube with a withstand voltage of 1600V is required.

The switching frequency of the flyback converter is usually 50k to 100KHz. In order to reduce switching losses, the energy required for switching should be as low as possible. To achieve this, the switching speed of the switch tube must be fast. A common trick to avoid turn-on losses is to turn on the transistor until the output diode current drops to zero (discontinuous mode). This requires a certain dead time before the next cycle starts. This method can reduce the commutation loss of the switch tube and the diode, thereby increasing the switching frequency and reducing the size of the transformer.

4 BIMOSFET chip technology

Standard high-voltage IGBTs are too slow for flyback converters. This new high-voltage BIMOSFET is exactly what is needed.

Whether it is MOSFETS or IGBTs, its traditional structure is usually DMOS (double diffused metal oxide silicon), which is to generate a silicon epitaxial layer on a thin and low-resistance silicon substrate, as shown in Figure 3.a.

However, when the voltage exceeds 1200V, the N-silicon layer that bears the blocking voltage tends to be a structure without epitaxial layer as shown in Figure 3.b. This structure is also called "uniform base structure" or NPT.

Referring to Figure 3.b, the pnpn structure in the IGBT is retained, but it should be noted that an N+ collector-short circuit mode is introduced here to reduce the current gain of the PNP transistor and improve its turn-off performance. However, there is a "free" parasitic diode between the emitter and the collector, which is the origin of the acronym BIMOSFET. The turn-off of the BIMOSFET is controlled by the collector. In order to optimize the reverse conduction of the diode and avoid the dv/dt problem caused by commutation, the lifetime of minority carriers should be reduced by irradiation.

There are two types of BIMOSFET, one is called standard type, similar to IGBT, its control voltage is VGE=15/0V; the other is "G" type, its gate voltage is the same as MOSFET, we will introduce it in the next section. In addition, the static and dynamic characteristics of both are the same.

Figure 3

5 Driver Requirements

a) Standard BIMOSFET

Experiments show that gate resistance and gate voltage have a great impact on losses. We found that when the gate resistance is less than 30 ohms, the drive waveform will oscillate; and when the resistance is greater than 50 ohms, the conduction loss will increase. Therefore, the best working condition for IXBH9N160 BIMOSFET is a drive voltage of 15V and a gate resistance between 30-50 ohms. In order to obtain full conduction, a gate voltage of 15V is necessary because the threshold voltage of 6V is relatively high compared to MOSFET.

b) G-type BIMOSFET

The threshold voltage of G-type BIMOSFET is usually around 4V, slightly lower than the standard type. Therefore, the gate drive voltage can be 10V. BIMOSFET can be used as a 1000V MOSFET in a flyback converter. Since its blocking voltage is as high as 1400/1600V, the absorption capacitor can be reduced or even omitted. However, the drive voltage should be at least 15V to reduce the turn-on loss.

The G type is indicated by the letter G at the end. The first batch of devices produced so far are IXBF9N140G and IXBF9N160G.

6 Static characteristics

By comparing the output curves, we can see the linear characteristics of the MOSFET (Figure 4a) and the bipolar characteristics of the BIMOSFET (Figure 4b).

Figure 4a tells us that when the drive voltage is only 6V, the MOSFET can flow 2A of current. Comparing the output characteristics of the BIMOSFET in Figure 4b, we see that when the drive voltage is 7V, no current flows. This is the biggest difference between BIMOSFET. When the current is less than 5A, we need at least 11V of drive voltage to turn it on. In situations where the current peak is relatively high, we need a drive voltage of 15V. The losses during conduction are also different. When the drive voltage is 15V and the current flowing is 2A, the MOSFET has an 18V voltage drop, while the BIMOSFET has only a 4V voltage drop, which reduces the loss by 4.5 times. In addition, the current rating of the BIMOSFET is also relatively high. Ordinary MOSFETs can only flow 3A, while BIMOSFET can reach more than 10A.

Figure 4 Output characteristics

7 Switching Characteristics

In order to quantify the performance of MOSFET and BIMOSFET, we conducted a series of comparative tests. Figure 5a and Figure 5b show the waveforms of a complete switching cycle and calculate the losses. The drain current, drain voltage, and gate voltage were also measured. The power dissipation and total energy were also calculated from these data.

The test device is a double pulse tester. When the MOSFET is turned on, the freewheeling diode is still turned on. Therefore, the waveform at turn-on will be slightly affected by the reverse recovery of the diode. However, since the effect of the diode on MOSFET and BIMOSFET is the same, the two can still be compared.

Figure 5 Switching waveform

The conditions are as follows:

Shutdown current amplitude = 4A

Voltage = 800V

Gate drive = 15V, 40 ohms

Junction temperature = 125 degrees Celsius

t0 to t1 is the end of the on-state. At the end of this state, we can see that the energy curve rises (shown by the solid line), which is caused by the relatively high loss when the MOSFET is turned on.

The next stage (t1 to t2) is the off state. The dotted lines show that there is basically no difference between the two, with the BIMOSFET being slightly less.

After the shutdown is completed (t2 to t3), there is no tail current in the BIMOSFET. The energy curve rises slightly because the result we get is the same as that of the MOSFET, and the MOSFET has no tail current, so there may be a little error in the measurement during the shutdown state.

The next stage is the turn-on stage, from t3 to t4. We can see that the turn-on stage generates relatively large losses. The solid line above shows the current spike, which is caused by the diode commutation. The turn-on time of the MOSFET is longer than that of the BIMOSFET. The peak power of the MOSFET is 250nS, 4KW; while the peak power of the BIMOSFET is 130nS, 5KW. The total switching loss of the MOSFET is about 0.5mJ, while that of the BIMOSFET is only 0.4mJ, which is about 20% less.

The last 500nS, from t4 to t5, is the beginning of the on-state. The energy curve of the MOSFET rises due to the higher on-resistance. The curve of the BIMOSFET is flatter because of its lower saturation voltage drop.

8 Conclusion

The advantages of BIMOSFET are first its low turn-on loss and second its low conduction loss. At t5, the total energy consumption per cycle is 0.95mJ for MOSFET and only 0.62mJ for BIMOSFET. The total loss of BIMOSFET is reduced by about 35%.