3 New Topologies Challenge Legacy EV Power Systems
In the race to electrify transportation, the silent battle beneath the hood is no longer about horsepower—but about how efficiently electrons move from battery to motor. While automakers tout range and charging speed, a quieter revolution is unfolding in the architecture of onboard power electronics. At the heart of this shift lies a new generation of DC-DC converters, engineered not just for efficiency, but for resilience across wildly varying voltage demands. Among the most promising breakthroughs is a novel fixed-frequency, wide-gain LLC resonant converter developed by researchers at Hubei University of Technology and Southeast University—a design that could redefine how electric vehicles manage energy conversion under dynamic load conditions.
This innovation arrives at a pivotal moment. As China pushes toward its 2030 carbon peak and 2060 carbon neutrality goals, the national strategy increasingly hinges on high-efficiency power electronics that support everything from fast-charging infrastructure to next-generation battery systems. Traditional LLC converters, long favored for their soft-switching capabilities and low electromagnetic interference, have struggled to deliver both wide voltage gain and consistent performance without sacrificing reliability or complexity. The problem is especially acute in electric vehicles, where battery voltages can swing from under 200V in depleted states to over 800V in high-performance architectures. Conventional frequency-modulated LLC topologies respond by drastically altering switching frequency—sometimes by a factor of three or more—leading to increased circulating currents, suboptimal transformer utilization, and compromised power density.
The newly proposed topology sidesteps these trade-offs entirely. By integrating a shared resonant branch with a dual-bridge primary structure and a reconfigurable secondary rectifier, the design enables output voltage regulation from 0.5x to 3x the nominal input—all while locking the switching frequency precisely at the resonant point. This fixed-frequency operation is more than a technical nuance; it’s a strategic advantage. It eliminates the need for wide-range frequency sweeps, stabilizes magnetic component design, and dramatically reduces circulating energy losses that plague variable-frequency alternatives.
At the core of the architecture is an elegant duality. The primary side features a full-bridge inverter formed by switches S1–S4, with S3 and S4 also serving a secondary half-bridge leg that shares the same resonant inductor (Lr) and capacitor (Cr). This co-resonant arrangement cuts magnetic component count while ensuring current balance—addressing a chronic reliability issue in split-resonance designs where mismatched inductances cause uneven stress. On the secondary side, two transformers with identical turns ratios are connected in parallel on the primary and in anti-series on the secondary, feeding into a hybrid rectifier network. A single auxiliary switch, S5, toggles the rectifier between full-bridge and voltage-doubler modes, effectively modulating the effective turns ratio without altering physical windings.
Control is equally refined. Instead of chasing frequency, the system employs pulse-width modulation (PWM) at a fixed 100 kHz—exactly matching the LC resonant frequency. The duty cycle of S3 and S5, varied from 0 to 50%, dictates how much energy flows through the auxiliary half-bridge and whether the secondary operates in standard or doubled-voltage mode. At 0% duty, the converter behaves like a conventional half-bridge LLC with minimal gain (0.5). At 50%, both primary legs are fully engaged, and the secondary rectifier doubles the output, achieving a gain of 3. Crucially, this transition is smooth, continuous, and free of mode-switching transients that plague hybrid control schemes.
What sets this approach apart isn’t just its range—it’s its consistency. Unlike prior art that sacrifices efficiency at voltage extremes, this topology maintains near-ideal soft-switching across the entire operating envelope. Primary-side MOSFETs achieve zero-voltage switching (ZVS) in all conditions, thanks to sufficient resonant current during dead time. Secondary diodes naturally reach zero-current switching (ZCS) due to the discontinuous conduction inherent in the rectifier transitions. Most notably, circulating currents—those wasteful loops of energy that slosh back and forth without delivering useful power—are minimized because the system never operates far from resonance.
Experimental validation on a 1.5 kW prototype confirms the theory. With an input of 100 V and output tunable from 60 V to 360 V, the converter demonstrated stable waveforms, low voltage overshoot during load transients, and seamless transitions between operating points. Even under abrupt load changes—from full 1.5 kW down to 680 W—the output voltage settled within milliseconds, with negligible ringing or instability. Perhaps most compelling is the stress profile on components: the auxiliary switch S5 sees its voltage clamped by output capacitors, reducing its blocking requirement to half the full output voltage at high gain settings—a significant boon for cost and reliability.
This matters deeply for automotive applications. Modern EVs demand DC-DC converters that can interface between high-voltage traction batteries (400–800 V) and low-voltage auxiliary systems (12–48 V), often while supporting bidirectional energy flow for vehicle-to-grid (V2G) services. Legacy solutions either stack multiple converters or accept compromised efficiency at partial loads. The co-resonant LLC design offers a single-stage alternative that scales elegantly across this spectrum. Moreover, its fixed-frequency nature simplifies EMI filtering and enables tighter integration with digital control platforms—key enablers for software-defined powertrains.
The implications extend beyond passenger vehicles. In commercial electric trucks, where battery packs exceed 1 MWh and voltage swings are even more extreme, such a converter could reduce system weight and cooling demands. In renewable energy microgrids—another strategic priority for China—this topology could serve as a robust interface between variable solar/wind sources and stable DC buses. Even in fast-charging stations, where efficiency directly impacts operational cost and thermal management, the reduction in circulating current translates to smaller heatsinks and higher power density.
Critically, the design avoids the pitfalls of earlier wide-gain attempts. Some prior architectures introduced extra switches or magnetic components to widen the gain window, inadvertently increasing cost and failure points. Others relied on hybrid modulation—mixing pulse-frequency and pulse-width control—which created discontinuities during mode transitions and complicated controller design. By contrast, this approach uses only five primary switches and four diodes, fewer than many competing topologies, and relies solely on PWM—a mature, digitally friendly control method already embedded in most automotive microcontrollers.
From a manufacturing standpoint, the shared resonant branch reduces bill-of-materials complexity. Instead of two independent resonant tanks, the system uses one Lr and one Cr for both power paths, cutting component count and improving thermal symmetry. The transformers, though dual, share identical specifications, simplifying procurement and assembly. And because gain is decoupled from load-dependent parameters like quality factor (Q) and inductance ratio (k), designers gain unprecedented freedom in selecting magnetic components—enabling larger magnetizing inductances that further suppress circulating current without sacrificing regulation range.
This decoupling is no small feat. In traditional LLC designs, voltage gain is tightly coupled to load and component tolerances, forcing conservative design margins that limit performance. Here, theoretical analysis and experimental data confirm that gain remains stable across varying loads and k-values—a rare trait that enhances robustness in mass production, where component variance is inevitable.
Looking ahead, the architecture opens doors to further innovation. The fixed-frequency, PWM-based control is inherently compatible with gallium nitride (GaN) and silicon carbide (SiC) devices, which thrive under stable switching conditions. Integrating this topology with wide-bandgap semiconductors could push efficiency beyond 98% while shrinking the converter’s footprint—critical for space-constrained EV platforms. Additionally, the modular nature of the dual-transformer secondary suggests potential for fault-tolerant designs, where one leg can be disabled without total system failure.
China’s push for technological self-reliance in core components makes such advances strategically significant. Power electronics—especially high-efficiency converters—are designated as key technologies in the nation’s 14th Five-Year Plan, with explicit support for innovations that enhance energy conversion in EVs and renewable systems. This co-resonant LLC converter aligns precisely with that vision: it is domestically developed, patentable, and addresses a genuine bottleneck in the electrification chain.
Industry observers note that while laboratory prototypes abound, few make the leap to automotive-grade deployment. Yet this design’s simplicity, component count, and compatibility with existing control paradigms improve its odds. With China’s EV market now accounting for over 60% of global sales, domestic innovations that solve real-world engineering constraints are increasingly likely to set global benchmarks.
In a sector where incremental gains are celebrated, a threefold voltage range with fixed-frequency operation and minimal loss represents a leap. It’s not just another converter—it’s a rethinking of how power should flow in the electric age. As automakers race to differentiate on efficiency, range, and charging speed, the unsung heroes may well be the circuits that quietly, reliably, and elegantly manage every volt.
Hongzhan Guo¹, Jian Pan¹, Jiaxin Xiong²
¹School of Electrical and Electronic Engineering, Hubei University of Technology, Wuhan 430068, China
²School of Electrical Engineering, Southeast University, Nanjing 210096, China
Journal of Power System Technology, DOI: 10.13335/j.1000-3673.pst.2023.0138