Application of SiC in Electric Vehicle Power Conversion

Electric vehicle (EV) shipments are growing rapidly and are expected to accelerate in the 2020s. Major automakers have either launched EVs or have plans to do so, and they are working with partners to find the best power electronics to maximize range and reduce costs. The main applications for SiC devices are shown in Figure 1, and forecasts indicate that SiC shipments are expected to reach $10 billion by 2030. The most important power element of an EV is the EV traction inverter, which we will discuss in a later article. Other important converters are the on-board charger and the DC converter. These increasingly involve bidirectional power flow and benefit greatly from fast switching and excellent parasitic diode behavior. SiC FET products are now qualified to AEC-Q101 to meet these needs. We will discuss the main topologies and look at the advantages of using SiC devices, especially in higher voltage (500-800V) battery systems.

Figure 1: UnitedSiC Advantages

On-board charger topology

The on-board charger (OBC) is located inside the vehicle, so it must be able to use a converter topology with the highest possible power density and energy efficiency in order to reduce its size and weight. The choice of topology depends on the power range, which can be 6.6kW, 11kW, or 22kW (electric buses). In some cases, the on-board charger can be bidirectional, which means that the circuit not only allows the battery to draw current from the grid, but the electric vehicle can also act as a distributed power source to feed power into the grid. In this mode, the power flows in the reverse direction, but the peak operating power can be lower (half) than the battery charging rating.

Figure 2 shows two configurations of on-board chargers designed for unidirectional power flow. The circuit has two main sections, the front-end rectifier stage and the DC converter stage. The rectifier stage rectifies the AC mains voltage to provide a DC rail with unity power factor. The DC-to-DC full-bridge phase-shifted stage then provides a precisely controlled output to charge the battery pack. During Li-ion battery charging, the circuit first operates in controlled current mode, then in constant power mode for fast charging, and finally in constant voltage mode until the battery is fully charged.

Figure 2: Two configurations in an on-board charger designed for unidirectional power flow

To maximize the efficiency of the rectifier stage, bridgeless topologies are becoming increasingly popular as they avoid the conduction losses of the diode rectifier bridge. Figure 2 shows a totem pole (TPPFC) circuit that can be used for lower power levels. The circuit has a fast switching phase leg, while the rest of the circuit switches at line frequency. The fast switching leg can operate in continuous conduction mode (CCM) and critical conduction mode (CRM). In continuous conduction mode, the switch is difficult to turn on, and the best option is to use a wide bandgap switch with an excellent low QRR parasitic diode. Figure 3 compares device parameters, including the QRR of UnitedSiC FETs compared to advanced superjunction devices. If the switching frequency exceeds 20kHz, a wide bandgap switch must be used, and the standard gate drive provided by UnitedSiC FETs will make it easy to insert UnitedSiC FETs and upgrade from superjunction devices to UnitedSiC FETs.

Figure 3: Device parameter comparison, including QRR of UnitedSiC FETs versus leading edge super devices

If critical conduction mode is used, the peak current becomes higher, placing additional constraints on the inductor and requiring a switch with lower on-resistance. In the absence of hard turn-on, silicon-based superjunction FETs can be used, at least at lower bus voltages. Even in this case, using SiC FETs makes sense because there are now options with very low on-resistance, and SiC FETs are becoming increasingly competitively priced compared to superjunction FETs with similar on-resistance. In addition, with 1200V SiC FETs, the topology can be extended to higher DC rail voltages, increasing power output with minimal switching times.

For higher power levels such as 11-22kW, a 3-phase active front-end rectifier is an excellent option. The bus voltage is typically 600-800V, which requires 1200V devices. In addition, the dual-level 3-phase circuit in Figure 2 requires switches with low switching losses and low QRR, making SiC FETs a better choice instead of IGBTs. Figure 4 shows the turn-on and turn-off characteristics of the 35mohm, 1200V, TO247-4L (UF3C120040K4S) UnitedSiC FAST FET. Given the very low turn-on and turn-off losses of the devices, these devices are used in parallel to achieve a highly efficient active front-end rectifier. When using a 4-pin Kelvin package, the user can switch faster with lower losses and cleaner gate waveforms.

Figure 4: Turn-on and turn-off characteristics of a 35mohm, 1200V, TO247-4L (UF3C120040K4S) UnitedSiC FAST FET

An alternative to the front-end rectifier is the Vienna rectifier, shown in Figure 5, which allows the use of 650V silicon superjunction devices in conjunction with SiC Schottky diodes to reduce cost. In this circuit, the switches are not hard switched. However, the number of semiconductors required is higher, and the diode voltage drop limits the best achievable efficiency.

Figure 5: Vienna rectifier, which allows 650V silicon superjunction devices to be used with SiC Schottky diodes to reduce costs

DC Converter

As shown in Figure 2, both the battery charger and the main DC converter that provides 12V/24V power are phase-shifted full-bridge converters. At full load, the circuit uses FETs that turn on in a zero-voltage switching (ZVS) manner, and uses buffer capacitors to minimize the turn-off losses of the entire device. The circuit can operate at high frequency (100-300kHz) and is highly efficient. SiC FETs are ideal due to their low conduction and turn-off losses and simple gate drive requirements. The same is true for UnitedSiC FETs that can be driven at 0 to 12V or driven by a simple pulse transformer that outputs -12/0/12V voltages. At light load conditions, hard switching can occur, which can cause problems for superjunction FETs, which are also more susceptible to diode recovery induction failures, and IGBT circuits are prone to higher losses.

The LLC topology shown in Figure 6 is an excellent choice, especially when the output voltage is fixed. This topology is most common in DC converter stages with fixed outputs, while the phase-shifted full-bridge topology is better suited to handle variable output voltages. At low bus voltages, superjunction FETs and fast diodes are used in LLC circuits. At higher voltages, IGBT power losses become too high, making SiC FETs a better choice.

Figure 6: At low bus voltages, superjunction FETs and fast diodes are used in LLC circuits. At higher voltages, IGBT power losses become too high, making SiC FETs a good choice.

Even at low bus voltages, the UnitedSiC 650V SiC FETs achieve very low gate charge, very short output capacitor charge time and very low parasitic diode conduction losses, which can be used to increase LLC operating frequency from 100kHz to 500kHz. On-resistance is now as low as 7mohm, 650V in a TO247-4L package. For applications with low profile space constraints, 27mohm, 650V devices are available in the industry standard DFN8x8 package.

For bidirectional DC conversion, Figure 7 shows a dual active bridge (DAB) and CLLC circuit with active switches on the output side. For battery charging, the DAB can be controlled from a fixed DC bus by varying the gate PWM waveform because the output voltage varies over a wide range. If a CLLC topology is used, the bus voltage must be varied by varying the control scheme of the active rectifier stage (either totem pole PFC or 3-phase active front end) in order to maintain near resonance operation of the DC-DC stage. In both cases, SiC FETs must be used on the secondary side to effectively hard switch in reverse mode. These FETs can be 650V to 1200V FETs for battery charging or lower voltage silicon FETs in the 100-150V class for 12V/24V output.

Figure 7: Dual Active Bridge (DAB) and CLLC circuits with active switches on the output side

Easy transition

Wide-bandgap SiC FETs enable the use of more sophisticated topologies and higher frequencies to achieve high power density and efficiency targets. The increased space in electric vehicles at the system level can easily offset the higher switching costs. UnitedSiC FETs have the important advantage of being compatible with all types of gate drive voltages, so they can be plugged into both silicon-based and SiC MOSFET designs. As designers around the world increasingly adopt SiC device deployments, this advantage allows them to easily make the transition, sometimes simply by upgrading existing silicon-based designs.

Recent Developments

The next stage will most likely involve an integrated driver and FET stage, such as the SIP half-bridge with driver shown in Figure 8, which uses 35mohm, 1200V stacked cascode switches. The switching waveforms show that such devices allow very fast, very clean switching and can be used as building blocks for all of the circuit options described in this article.

Figure 8: Integrated driver and FET stages, such as the SIP half-bridge with driver shown, using 35mohm, 1200V stacked cascode switches

SiC FET technology is advancing rapidly, with the 2020 version of the switch currently under development capable of 2X better performance characterization. Combined with discrete packaging improvements and the introduction of SiC-based intelligent power modules, these three advances will further increase power density as EV deployment continues to increase.