How to use SiC to build a better electric vehicle traction inverter

“In this article, we will investigate the advantages of using SiC technology in electric vehicle traction inverters. We will show how the energy efficiency of the inverter is improved under various load conditions, from light load to full load. Using a higher operating voltage and high-efficiency 1200V SiC FET can help reduce copper loss. The switching frequency of the inverter can also be increased to output a more ideal sinusoidal waveform to the motor windings and reduce the iron loss in the motor. It is expected that under the influence of all these factors, the mileage of a pure electric vehicle on a single charge will increase by 5-10%. At the same time, the reduced loss can also simplify the cooling problem.

“

In this article, we will investigate the advantages of using SiC technology in electric vehicle traction inverters. We will show how the energy efficiency of the inverter is improved under various load conditions, from light load to full load. Using a higher operating voltage and high-efficiency 1200V SiC FET can help reduce copper loss. The switching frequency of the inverter can also be increased to output a more ideal sinusoidal waveform to the motor windings and reduce the iron loss in the motor. It is expected that under the influence of all these factors, the mileage of a pure electric vehicle on a single charge will increase by 5-10%. At the same time, the reduced loss can also simplify the cooling problem.

Introduction

Recent news indicates that the number of pure electric vehicles (BEV) is increasing faster than previously expected. This has prompted automakers (including existing manufacturers and newly joined manufacturers) to reinvest in the development of electric vehicles, trying to find the most effective technology to maximize energy efficiency, reduce volume and weight, and replace as much as possible from expensive battery packs. Benefit from this, thereby extending the mileage on a single charge. This allows SiC transistors to quickly enter the on-board chargers and DC converters of electric vehicles. Given that traction inverters handle 10 times the power level, if SiC transistors can have similar advantages in this environment, it will rewrite the power semiconductor landscape. To this end, SiC technology needs to provide clear cost performance advantages and remove all inevitable obstacles in order to achieve a reliable inverter system design that can be put into manufacturing. The boost stage used in front of the inverter will undoubtedly use SiC, for the same reason we gave when discussing car chargers and DC converters. In this article, we will examine the main advantages of using SiC technology in electric vehicle inverters and discuss several implementation options based on UnitedSiC technology.

Main advantages of SiC technology

The driving conditions of typical electric vehicles, especially those used in cities, will cause the inverter to operate under light or medium load for most of its operating life, but with frequent stops and starts. However, all worst-case stresses must be considered when designing the inverter, such as rapid acceleration, steep slopes, and operation at various ambient temperatures. Figure 1 shows a typical two-level voltage source inverter that can be used to drive an internal permanent magnet motor. This is a common configuration for pure electric vehicles, and its inverter is placed near the motor. Normally, the inverter switch will be under control to apply a 3-phase AC voltage to the motor windings. This goal is achieved by switching the power switch according to the command of the controller, the frequency is 4-10kHz, and the basic AC frequency of up to 1kHz can be generated. The total power level can range up to 50-250kW, which is suitable for electric buses. The DC voltage used depends on the battery system, and due to the use of a boost converter to convert various battery voltages to the fixed DC voltage used by the inverter, this voltage may increase from the current 300-500V to 600- in the near future. 800V, a higher voltage can reduce current and copper loss while providing the same power.

Figure 1: Electric vehicle traction inverter using two-level voltage source converter architecture

The loss of the power switch comes from the conduction loss when the current flows through the switch and the switching loss when the switch is opened and closed. The conduction loss has nothing to do with the switching frequency, but the switching loss is proportional to the switching frequency.

Figure 2 shows the characteristics of SiC FETs and silicon IGBTs. At any given current, I_D*V_DS The product of can all represent a given conduction loss. Therefore, it is easy to see that when unipolar SiC FETs are used, there is no inflection point voltage that occurs when IGBTs are not used. This is beneficial for all current levels up to 200A, and correspondingly lower in light load and medium load operation. It is especially beneficial under current.

Figure 2: Conduction loss characteristics of 200A SiC FET and IGBT

Figure 3 is a comparison of the conduction and switching losses of inverters based on low conduction loss IGBTs and SiC FETs suitable for 750 V devices when operating at a 400 V bus at 8 kHz. IGBT solutions have considerable switching losses even at 8kHz, so they cannot be used effectively at 25kHz. SiC-based solutions not only have low conduction losses at all output levels (at 8kHz, the losses will be greatly reduced), but also can be used at higher inverter frequencies (high energy efficiency).

Figure 3: Comparison of power loss in conduction and switching of inverters based on 1200V IGBT and SiC FET.In all cases there is a loss difference, the difference is very large at 25kHz

Another aspect of an electric vehicle inverter that is different from traditional industrial motor drives is that it requires two-way power transmission. During regenerative braking, the system controls the switch to allow the inverter to act as a rectifier and the motor to act as a generator, allowing electrical energy to flow back into the battery. The SiC FET allows the third quadrant to conduct electricity with the same low conduction loss, which means that synchronous rectification can be used to maintain very low losses in this operating mode. When using IGBTs, this is impossible, and the loss of reverse power flow with reverse parallel freewheeling diodes is relatively high.

Rated voltage

At present, many electric vehicle inverters are based on 750V IGBT, and the inverter bus voltage is 300-500V. In order to handle high power more efficiently, the 1200V switch allows the use of batteries with a voltage of 600-800V.

Table 1 shows some calculation data of the 450A, 750V half-bridge module used in the traction drive of a 200KW pure electric vehicle. The drive is based on the low conduction loss IGBT and UnitedSiC FET with the same rating of 750V. Each switch position uses 3 IGBTs and 3 diodes. They were replaced with 6 stacked SiC FETs, each with a resistance of 5.4mohm and no more than half of the original volume. Cases 1 and 2 show the difference in total conduction loss, switching loss, and total loss at 8kHz. At 200kW, the total loss will be halved, and at 50kW, the total loss will be close to a quarter of the original. Given that the inverter runs under light load most of the time, this feature is very beneficial. Please note that when SiC FET is used, the conduction loss and switching loss are relatively low, but at 200KW, the difference in switching loss is nearly 8 times. The table also shows that the same module can also be used up to 300KW, keeping all FETs below Tj=150C, thus allowing the same inverter hardware to be used in a 300KW system. Case 3 shows a better way to handle 300KW, which is to use 8 SiC FETs for each switch to reduce the peak loss from 3425W to 2666W.

Table 1: Comparison of operating power loss between 750V IGBT and 750V SiC FET-based 450A, 750V 3-phase inverter modules used in 200kW electric vehicle inverters. The lower part of the table compares the 400A, 1200V IGBT module used in a 200kW inverter with the corresponding 1200V SiC FET module. In all cases, we consider the use of nail-fin heat sink type 3-phase modules at a cooling temperature of 90°C. In all cases, the maximum junction temperature is kept below 150°C, even if the SiC FET is rated at 175°C and can withstand a short time of 200°C. The lower switching loss can be used to run the inverter at 25kHz, thereby improving the waveform quality and reducing the iron loss. Even in this case, it can be seen that the SiC FET solution (Table 1 Case 4) has lower losses than the IGBT solution under all load conditions. When the output power is 200KW, the IGBT solution will dissipate 3580W at 8kHz, while the SiC FET solution will dissipate 2061W at 25kHz.The module can achieve 250KW output in 6 parallel SiC FETs

If only 4 SiC FETs are used per switch, lower costs can be achieved at the expense of higher losses. This situation is shown in Case 5 of Table 1. At this time, the loss is still much lower than the IGBT-based solution.

The lower part of the table compares the losses when using 1200V transistors and operating bus voltage of 800V. It compares the case of 4 IGBTs and 4 diodes per switch (Case 6) with 4 SiC FETs per switch (Case 7, 9). If SiC FETs are used, the loss at 8kHz is less than half of the full power, and at 50kW it is a quarter of the full power. Case 8 shows how this module can be easily expanded to 300KW operating power with 6 SiC FETs per switch. Although these IGBTs cannot be switched at 25kHz due to high switching losses, Case 9 shows how to use SiC FETs to achieve this frequency while maintaining high energy efficiency. The loss is still much lower than the loss when the IGBT operates at 8kHz, and likewise, a smoother waveform can also help reduce the iron loss of the motor, and the inverter switching frequency greatly exceeds the audible frequency range. Please note that in all cases, the power output of this module is higher under the same board area.

Extreme conditions

For all motor drives, including electric vehicle inverters, an important safety requirement is to be able to withstand short circuits during maintenance or operation. This type of short circuit may occur everywhere in the DC bus, from the motor winding to the ground on the entire circuit or between the windings. For semiconductor switches, this means that the switch must be able to withstand when it opens and a short circuit occurs, until the gate drive detects the short circuit within 3-5 µs and closes the switch. In addition, when the switch is already conducting, a short circuit may also occur. In any case, the switch must be able to withstand such a short circuit, no matter how high the initial temperature of the chip is when such a short circuit occurs, and the device characteristics cannot be changed, so that the service life is degraded.

Figure 4 compares the short circuit withstand time (SCWT) difference of IGBT, SiC MOSFET and SiC FET. During a short circuit, the SiC MOSFET experiences extremely high peak currents, which may damage the MOSFET gate diode. This can be managed by using a lower gate voltage drive, which can reduce short-circuit current at the expense of very high conduction losses. SiC FETs (including Si MOSFETs stacked on SiC JFETs) perform much better in this regard. The peak saturation current can be adjusted to provide the required short-circuit withstand time, and the change in conduction loss caused by this adjustment is very small. The saturation current is set by the JFET, so it is different from the V applied to the MOSFET_GS Irrelevant. Experiments have shown that SiC FETs can safely handle such stresses caused by more than 100 repetitive events. In addition, even if the starting chip temperature reaches 200°C, the device can handle this type of short circuit.

Figure 4: Comparison of short-circuit withstand capability of IGBT, SiC MOSFET and SiC FET, and ranking of ability to handle repetitive shocks

The semiconductor in Figure 4 is drawn to scale, where the SiC FET is the smallest 100A device with the chip volume. An important advantage of SiC JFET is that it can withstand the large amount of heat generated during a short circuit, which forms the basis of SiC FET stacked cascode. The different chip sizes also explain why the use of SiC FETs can reduce on-resistance for a given module footprint.

The technical method of SiC-based inverters

The most commonly used inverter topology is the well-known two-level voltage source converter in Figure 1. The type of switch used with this type of inverter is called hard switching, which causes the high voltage across the switch to overlap with the current passing through the switch during the conversion period. Based on the results in Figure 3, one method designers can use is to use fast switching SiC devices to reduce switching losses and conduction losses, even at frequencies up to 25kHz. In this case, the switching occurs at high dV/dts. In pure electric vehicles, as in standard industrial drives, the length of the wire between the inverter and the motor is not a problem. However, directly applying high dV/dt waveforms to the motor windings may cause large displacement currents in the isolation range. You can use a filter at the output of the inverter to rectify this, and only make the high dV/dt part smoother, like the so-called dV/dt filter, or use a set of sinus filters to smooth the waveform, providing almost Perfect sine curve output. Obviously, if the switching frequency is higher, filtering will be easier. It is expected that reducing the ripple in the current waveform will increase the overall energy efficiency of the motor by 1-3% and extend the life of the motor. This energy efficiency benefit can be translated into longer mileage on a single charge or reduced battery size.

Another method is to keep the switch at a low frequency of 5-8kHz and run a device with a very low dV/dt rating, such as below 8V/ns. In this case, the switching overlap loss per cycle may be very high, but the low frequency can make the total power loss controllable. Figure 5 shows the preferred technology for using SiC FETs in this situation. The stacked low-voltage MOSFET is only used as an enable switch to ensure long-off operation under startup and short-circuit fault conditions, but the SiC JFET gate switches directly. This enables very low dV/dts and lowest loss. This solution can achieve excellent short-circuit handling capability. If the JFET gate reaches +2.5V instead of 0V, it can further reduce the conduction loss by 15-20%. In order to manage the third quadrant conduction, JFET can be used with low dead time, or only a small JBS diode can be added during the dead time to carry the freewheeling current. The graph on the right of Figure 5 shows the third quadrant behavior of a SiC JFET.

Figure 5: Directly drive the gate of the JFET and use a stacked N-channel MOSFET as a switching scheme to start the switch.It makes it easier to implement low dV/dt switches

There is also a more complex method that can achieve the highest energy efficiency is to use a fully resonant switch, like the auxiliary resonant converter method. For this purpose, Pre-Switch Inc. has developed a new controller. Figure 6 shows the circuit topology and typical switching waveforms, which can completely eliminate turn-on and turn-off switching losses while maintaining low dV/dts. Although this circuit helps to reduce IGBT switching losses and improve energy efficiency, the IGBT still has to withstand losses due to the need to remove the stored charge in each cycle. In addition, the influence of the conduction loss of the inflection point voltage in the IV curve still exists. Therefore, SiC FET can obtain the best peak energy efficiency under all load conditions. It is a unipolar device without tail current and knee voltage. The converter can also operate at a very high frequency such as 50-100kHz, resulting in a smoother sinusoidal output waveform. This can further improve the efficiency of the motor by reducing the iron loss, and combined with the minimum power loss of the inverter, it can increase the maximum single-charge driving range of pure electric vehicles. Figure 6 is an example of a compact 200kW inverter using this model and SiC FET.

Figure 6: ARCP topology that can eliminate all switching losses in the inverter. The combination of this structure and SiC FET can achieve very high power density without high dV/dt switching problems.This will bring very high motor operating efficiency and very high inverter energy efficiency

in conclusion

Many industrial and academic groups have conducted in-depth investigations and concluded that SiC MOSFETs have significant advantages in improving the energy efficiency of traction inverters and extending the mileage of pure electric vehicles on a single charge. In this article, we discussed the reasons for this evaluation result and considered the robustness requirements of the power transistors used in SiC inverters. We introduced three implementation topologies suitable for pure electric vehicle inverters, allowing users to choose the method that best suits their overall system constraints. In order to obtain the highest energy efficiency, the ARCP solution eliminates all switching losses and can maximize the use of the ultra-low conduction loss characteristics of SiC FETs.

The Links:   SKD115-16   LLG-VLGE1486-01A   IGBTS