Power Tips #149: Boosting EV charger efficiency and density with single-stage matrix converters

An onboard charger converts power between the power grid and electric vehicles or hybrid electric vehicles. Traditional systems use two stages of power conversion: a boost converter to implement unity power factor, and an isolated DC/DC converter to charge the batteries with isolation. Obviously, these two stages require additional components that decrease power density and increase costs.

Matrix converters use a single stage of conversion without a boost inductor and bulky electrolytic capacitors. When using bidirectional gallium nitride (GaN) power switches, the converters further reduce component count and increase power density.

Comparing two-stage power converters with single-stage matrix converters

A two-stage power converter, as shown in Figure 1, requires a boost inductor (L_B) and a DC-link electrolytic capacitor (C_B), as well as four metal-oxide semiconductors (MOSFETs) for totem-pole power factor correction (PFC). 

Figure 1 Two-stage power converter diagram with L_B, C_B, and four MOSFETs for totem-pole PFC. Source: Texas Instruments

A single-stage matrix converter, as shown in Figure 2, does not require a boost inductor nor a DC-link capacitor but does require bidirectional switches (S₁₁ and S₁₂). Connecting common drains or common sources of two individual MOSFETs forms the bidirectional switches. Alternatively, when adopting bidirectional GaN devices in matrix converters, the number of switches decreases. Table 1 compares the two types of converters.

Figure 2 Single-stage matrix converter diagram that does not require L_B or C_B, but necessitates the use of two bidirectional switches: S₁₁ and S₁₂ . Source: Texas Instruments 

Table 1 A two-stage AC/DC and single-stage matrix converter comparison. Source: Texas Instruments

Single-stage matrix converter topologies

There are three major topologies applied to EV onboard charger applications.

Topology No. 1: The LLC topology

Figure 3 shows the inductor-inductor-capacitor (LLC) topology. The LLC converter regulates current or voltage by modulating switching frequencies. L_r and C_r form a resonant tank to shape the resonant current. Selecting the proper control algorithms will achieve a unity power factor.

With a three-phase AC input, the voltage ripple on the primary side is much smaller compared to a single-phase AC input. Therefore, the LLC topology is more suitable for three-phase applications. LLC converters operate at a higher frequency and realize a wider range of zero voltage switching (ZVS) than other topologies.

Figure 3 An LLC-based matrix converter with a three-phase AC input. Source: Texas Instruments

Topology No. 2: The DAB topology

Figure 4 shows a dual active bridge (DAB)-based matrix converter. The DAB topology can apply to a three-phase or single-phase AC input. Controlling the inductor current will realize unity power factor naturally. The goal of a control algorithm is to realize a wide ZVS range to reduce switching losses, reduce root-mean-square (RMS) current to reduce conduction losses, and achieve low current total harmonic distortion and unity power factor.

Triple-phase shift is necessary to achieve these goals, including primary-side internal phase shift, secondary-side internal phase shift, and external phase shift between the primary side and secondary side. Additionally, modulating the switching frequency will extend the ZVS range.

Figure 4 A DAB-based matrix converter with a single-phase AC input. Source: Texas Instruments

Topology No. 3: The SR-based topology

Figure 5 shows a series resonant (SR) matrix converter. The resonant tank formed by L_r and C_r shapes the transformer current to reduce turnoff current and turnoff losses. Meanwhile, the reactive power is reduced, as are conduction and switching losses. Compared to the LLC topology, the switching frequency of SR matrix converters is fixed, but higher than the resonant frequency.

Figure 5 An SR-based matrix converter with a single-phase AC input. Source: Texas Instruments

The control algorithm of single-stage matrix converters

In an LLC topology-based onboard charger with a three-phase AC input, switching frequency modulation regulates the charging current or voltage and uses space vector control based on grid polarity. The voltage ripple applied to the resonant tank is small. The resonant tank determines gain variations and affects the converter’s operation.

A DAB or SR DAB-based onboard charger usually adopts triple-phase shift (TPS) control to naturally achieve unity power factor, a wide ZVS range, and low RMS current. Optimizing switching frequencies further reduces both conduction and switching losses.

Figure 6 illustrates pulse width modulation (PWM) waveforms of TPS control of matrix converters for a half AC cycle (for example, Vac > 0). Figure 4 shows where PWMs connect to the power switches: d1 denotes the internal phase shift between PWM1A and PWM4A, d2 denotes the internal phase shift between PWM5A and PWM6A, and d3 denotes the external phase shift between the middle point of d1 and d2. PWM1B and PWM4B are gate drives for the second pair of bidirectional switches.

Figure 6 TPS PWM waveforms for a single-stage matrix converter for a half AC cycle. Source: Texas Instruments

Regardless of the topology selected, matrix converters require bidirectional switches, formed by connecting two GaN or silicon carbide (SiC) switches with a common drain or common source. Bidirectional GaN switches are emerging devices, integrating two GaN devices with common drains and providing bidirectional control with a single device.

Matrix converters

Matrix converters use single-stage power conversion to achieve a unity power factor and DC/DC power conversion. They provide two major advantages in onboard charger applications:

• High power density through the use of single-stage conversion, while eliminating large boost inductors and bulky DC-link electrolytic capacitors.
• High power efficiency through reduced switching and conduction losses, and a single power-conversion stage.

There are still many challenges to overcome to expand the use of single-stage matrix converters to other applications. High ripple current is a concern for batteries that require a low ripple charging current. Matrix converters are also more susceptible to surge conditions given the lack of DC-link capacitors. Overall, however, matrix converters are gaining popularity, especially with the emergence of wide-band-gap switches and advanced control algorithms.

Sean Xu currently works as a system engineer in Texas Instruments’ Power Design Services team to develop power solutions using advanced technologies for automotive applications. Previously, he was a system and application engineer working on digital control solutions for enterprise, data center, and telecom power. He earned a Ph.D. degree from North Dakota State University and a Master’s degree from Beijing University of Technology, respectively.

Related Content

• Power Tips #122: Overview of a planar transformer used in a 1-kW high-density LLC power module
• Power Tips #145: EIS applications for EV batteries
• Power Tips #102: CLLLC vs. DAB for EV onboard chargers
• Extreme Fast Charging Architectures From 350 kW to 3.75 MW
• The Power of Bidirectional Bipolar Junction Technology

The post Power Tips #149: Boosting EV charger efficiency and density with single-stage matrix converters appeared first on EDN.