Faster Charging Could Mean Quicker Electric Vehicle AdoptionApril 30, 2020 by Anup Bhalla, UnitedSiC
This article discusses how new silicon carbide semiconductor technology could be a solution to faster and more efficient EV charging, enabling quicker adoption of electric vehicles.
Consumer demand for reliable EVs with comparable range to fossil-fuel cars is currently outrunning the technology itself — certainly at a widely accessible price point. Charge time also remains an issue, especially for those who need to travel far distances and ‘top-up’ their fuel as they go.
EV adoption is accelerating despite continuing reservations about the initial cost, range, charging infrastructure, and options. Cost and range are all about economy of scale and technology improvements which can only advance as time passes and available options are being addressed with a projected 400+ models set to be available by 2023. Charging is a different matter, however.
The Challenge of Charging EVs
The complexities surrounding EV chargers include different providers and a mix of options for level 1 and 2 (120/240VAC) up to around 20kW and DCFC, fast DC charging at up to 350kW. AC charging requires an on-board charger to convert to the appropriate DC voltage for the battery while a static DC charger must convert utility AC, either single- or three-phase to DC.
Consumer demand is territory-dependent with single-home users with garages in suburban areas happy to charge overnight from their domestic supply whereas city apartment dwellers forced to seek out public charge points. Those in rural areas suffer from a need to commute further but with fewer fast-charge stations available. Countries vary widely as well, with users in market-driver China much less likely to have access to personal garage charging than in the US and UK, where there is a tradition of single-home ownership. Some projections of how the mix might change from 2020 to 2030 are given in Figure 1 adapted from data by McKinsey.
Figure 1. Projected change of charging methods 2020-2030 USA, EU, and China. Image adapted from McKinsey data
Charger technology development must, therefore, cover multiple and changing scenarios varying from mass-produced, on-board types which must be efficient, small and light, so as not to affect range unduly, to industrial-grade roadside types costing from $20,000 up to $200,000 for the highest power DC fast chargers.
In all, AC and DC charging will represent a projected 280-terawatt hours of energy consumption globally by 2030 so high efficiency in the power conversion process is critical to reducing wasted energy, dollar cost to the consumer, and global emissions from the utility power generators.
Addressing Charger Efficiency
Chargers of all types have a similar makeup: utility AC is converted to high voltage DC with active power factor correction (PFC), the DC is switched to generate high-frequency AC which is passed through an isolation transformer, then rectified to provide DC to charge the battery. Feedback controls switching duty cycle to keep the output accurate.
Figure 2 shows the outline of a level 1 or 2 charger with a ‘totem pole’ bridgeless AC rectification and PFC stage, a ‘phase-shifted full-bridge’ DC-DC converter, and synchronous output rectification.
Figure 2. A Level 1 or 2 EV charger outline
The topology suits home charging as it is efficient and can be configured for bi-directional power flow, so the EV battery can generate AC to feed back into the grid. Large DCFC chargers typically have 480V three-phase AC as their supply but can have similar DC-DC conversion and rectification stages (although these may be several in parallel with switching frequencies phased to share the load and reduce ripple currents). Three-phase input rectification and PFC are often implemented with a Vienna rectifier, which is not bi-directional but doesn’t need to be in this application.
Semiconductor switches are used through the different conversion stages in all charger types and are a significant source of power loss. Until a few years ago, IGBTs and Si-MOSFETs were the options for switches, and epitaxial fast recovery diodes were used for high voltage output rectification. Wide band-gap semiconductors — both SiC and GaN — have now changed this with their lower conduction and switching losses, high-voltage operation, and high-temperature characteristics.
SiC MOSFETs and GaN HEMT cells are not perfect for all applications, however. Their gate drive requirements are quite particular and temperature-sensitive, the SiC MOSFET’s body diode has a high forward voltage drop, and GaN has limited voltage rating and no inherent avalanche over-voltage protection.
SiC FETs Could Offer the Solution
The implementation of a cascode using a SiC JFET driven by a Si-MOSFET is closer to the solution for all functional positions. Called a SiC FET, it has easy IGBT and Si-MOSFET compatible gate drive, lowest specific on-resistance of all switch types, robust avalanche and short-circuit current ratings, and a low-drop, fast body diode effect. Being a wide band-gap device, it also switches fast with the natural SiC advantages of high-temperature operation and thermal conductivity.
As an example, in Figure 2, Q1 and Q2 perform the PFC function and the circuit can operate in continuous (CCM) or discontinuous conduction (DCM) mode. Using Si-MOSFETs, CCM is not as efficient as the lossy body diodes that must conduct part of the switching cycle. In DCM, the diodes do not conduct but peak currents are much higher, making conduction losses unacceptable. SiC MOSFETs perform much better and in practice make the topology viable but SiC FETs have lower losses still. Q3 and Q4 switch at line frequency and perform the AC line rectification function along with Q1 and Q2 and only conduct in reverse so could be discrete diodes, but the low channel resistance of synchronously driven SiC FETs reduces losses further.
Q5, Q6, Q7, and Q8 form a ‘phase-shifted full-bridge’ (PSFB) converter which features soft switching at zero voltage for low loss and EMI. Body diodes do not conduct but the low channel resistance of SiC FETs keeps efficiency high. The gates of switches in the PSFB topology are driven at 50% duty cycle, the fixed frequency with phase variation achieving regulation. Two high-side drives are required which must be very carefully controlled with SiC MOSFETS and GaN devices to avoid over-voltages, particularly when transformer coupling. The easy gate drive of SiC FETs with wide +/-25V clamped maximum rating is a distinct advantage, making for a very simple gate drive circuit.
SiC FETs Enable High-Voltage Synchronous Rectification
Synchronous rectification is normally reserved for low voltage applications where high currents cause unacceptable losses in diodes. At high voltage, the benefits are less pronounced with lower currents for the same power. High voltage Si-MOSFETS are also not appropriate as their channel resistance is relatively high and body diodes slow, with a high forward drop. This is significant during ‘dead’ time when the channel is synchronously switched off and current ‘commutates’ through the body diode, causing losses. For this reason, high voltage outputs often use SiC diodes which have a relatively high forward voltage drop but no reverse recovery current, making them an improvement on the use of fast recovery epitaxial diodes.
SiC FETS, however, can be used here to advantage as well; they have fast, low-drop body diodes and new versions such as the UF3SC FET series from UnitedSiC achieve less than 7 milliohms channel resistance at 650V rating and less than 9 milliohms at 1200V rating.
Figure 3 shows a rectifier arrangement using a diode full-bridge for a fast battery charger output.
Figure 3. Replacing 4x SiC diodes D5 – D8 with 4x SiC FETs, more than halves dissipation and enables bi-directional power conversion
At 100A load for example, if the diodes were SiC types and the duty cycle is 50%, they would drop 2V at 125°C, dissipating 100W per diode. If each were replaced with a UnitedSiC UF3SC065007K4S SiC FET, the voltage drop would be 0.9V and power dissipation 45W per SiC FET. The use of synchronous SiC FETs also enables bi-directional power conversion, not possible with the SiC diodes.
Using SiC FETs to Scale Up High Power Rapid Chargers
SiC FET solutions are scalable to very high power in the largest of rapid chargers. Devices can be used in the Vienna rectifier for three-phase AC inputs to achieve efficiency savings and bridge DC-DC converters of all types can use parallel SiC FETs for higher power. SiC FETs naturally share current when paralleled due to the positive temperature coefficient of channel resistance and do not suffer from the variation of gate threshold voltage with temperature which is problematic with Si- and SiC-MOSFETs.
Gate charge is extremely low, so low-power gate drivers can be used for multiple paralleled SiC FETs with no problem. The United SiC 1200V SiC FETs with sub-9 milliohm resistance are also ideal for 480VAC systems.
Low acquisition and running costs are the drivers for EV battery charger development and better efficiency helps in both respects. SiC FETs enable a lower loss solution that drives down hardware costs and size with less cooling required. Less energy wasted helps the user recognize the value of their EV and reduces the environmental impact of energy generation; if this results in quicker adoption of EVs, it presents a solution to the EV rapid charging challenge.