Breakthrough AC/DC Converter Eliminates Electrolytic Capacitors in EV Charging—Boosts Battery Life and Reliability
By Jacobin
In the race to electrify transportation, the spotlight often shines on batteries, motors, and charging networks—yet quietly, a critical bottleneck persists beneath the chassis: the power electronics that bridge the grid and the battery pack. A new converter architecture, unveiled by researchers in China, promises to quietly but profoundly reshape how electric vehicles (EVs) charge—by eliminating one of the oldest, most failure-prone components in power systems: the electrolytic capacitor.
This isn’t just a component swap. It’s a systems-level rethink—one that tackles heat, reliability, and longevity in a single stroke. And at a time when automakers are scrambling to extend battery warranties, reduce thermal management complexity, and cut costs without compromising safety, such a development couldn’t be more timely.
The innovation, detailed in a recent paper published in Chinese Journal of Electron Devices, offers a single-stage, bidirectional AC/DC converter that operates without electrolytic capacitors—while simultaneously suppressing harmful low-frequency power ripple that degrades lithium-ion cells over time. Led by Wang Wenlong and Chen Lihui, the team from Xingtai Vocational and Technical College and Hebei Normal University’s College of Career Technology has engineered not just a new topology, but a tightly integrated, four-phase control strategy that ensures zero-voltage switching (ZVS) across all power transistors—a feat that improves efficiency and enables higher switching frequencies, shrinking magnetic components and overall system size.
To appreciate the significance, one must understand the hidden toll of conventional EV chargers.
Today’s onboard chargers (OBCs)—the units that convert AC grid power to DC for the battery—typically rely on an electrolytic capacitor to smooth out the 100 Hz or 120 Hz power ripple inherent in single-stage AC/DC conversion. This ripple, a legacy of rectifying sinusoidal mains voltage, doesn’t just sit idly in the circuit. When it flows into the battery, it causes repeated, small-scale charge/discharge cycles at low frequency—cycles that generate excess heat, accelerate electrode degradation, and—critically—shorten usable battery life.
Studies have long linked ripple-induced heating to increased solid-electrolyte interphase (SEI) growth and lithium plating, especially under high-state-of-charge conditions. In other words, the very act of plugging in your EV can, over years, subtly chip away at its range and resale value—not because of how you drive, but because of how the vehicle charges.
Worse still, the electrolytic capacitors deployed to absorb this ripple are themselves a weak link. They’re bulky. They dry out over time. Their lifespan is temperature-sensitive—often degrading faster in the hot, vibration-prone environment of an EV powertrain bay. When one fails, the entire charger may shut down—or worse, cause unpredictable behavior in high-voltage systems. From a reliability engineering standpoint, they’re a known liability.
For years, engineers have debated trade-offs: keep the capacitor and accept the lifespan hit? Or remove it and deal with ripple—and potential instability?
Previous attempts to go capacitor-free often added complexity elsewhere—extra active filtering stages, additional switching legs, or costly auxiliary circuits. A 2019 design cited in the paper, for instance, introduced a standalone active filter on the DC side—but at the cost of higher bill-of-materials (BOM), control instability under dynamic loads, and questionable scalability for mass production.
What sets the new converter apart is elegance through integration.
Rather than bolting on an external ripple suppressor, the team embeds the solution directly into the DC-side architecture: a buck-type active filter is co-designed with the main resonant bridge, sharing key power pathways and control signals. This integration slashes component count—one fewer branch than prior art—while preserving high power density.
The topology hinges on a modified bidirectional full-bridge resonant converter, with three critical switching legs (referred to as Branch I, II, and III in the paper) working in carefully orchestrated phase. Branch I and III handle primary power transfer between grid and battery via a series resonant tank (L_res and C_res), while Branch II—interfacing with a small, film-type decoupling capacitor (C_d)—acts as the ripple sink. Crucially, C_d is not an electrolytic. It’s a compact, long-life film capacitor—capable of withstanding high ripple current and surviving well beyond the vehicle’s service life.
But hardware alone isn’t enough. The real magic lies in the control layer.
Here, the researchers introduce a novel four-phase controller—a real-time, recursive optimization engine that continuously computes four key parameters: duty cycles d₁, d₂, d₃ and phase shifts Φ₁₂, Φ₂₃. These aren’t static settings. They’re dynamically tuned to meet three simultaneous objectives:
- Accurate power transfer—whether drawing 1 kW from the grid (G2V: Grid-to-Vehicle) or feeding power back (V2G: Vehicle-to-Grid).
- Zero-voltage switching (ZVS) for all switches—minimizing switching losses and electromagnetic interference (EMI).
- Minimized RMS current in the resonant network—reducing conduction losses and copper heating.
Achieving all three—especially ZVS across bidirectional operation and varying power factors—is notoriously difficult. Most resonant converters sacrifice one for the others, or rely on precomputed lookup tables that demand large memory and lack adaptability.
The team’s solution? A lean, iterative search algorithm running in fixed time per control cycle. Instead of brute-force computation, it uses the previous optimal solution as a warm start—scaling it intelligently with gain factors (k_I, k_d1, k_d3) to jump close to the new optimum. The outer loop sweeps resonant current magnitude (|I_res|) from its theoretical minimum upward—ensuring that the first feasible ZVS point it finds is also the one with lowest conduction loss.
It’s a clever hybrid: the precision of numerical optimization, married to the speed of model-based prediction.
To validate, the team built both a high-fidelity PSIM simulation and a 1 kW hardware prototype. Operating at switching frequencies between 60 kHz and 400 kHz—far beyond traditional 20–50 kHz designs—the converter demonstrated seamless transitions between G2V and V2G modes, with near-unity power factor in both directions.
Most compelling were the ripple results.
In G2V mode (θ_g = 0), battery-side current showed no discernible 120 Hz component. Oscilloscope traces of the decoupling inductor current (i_d) revealed clear AC ripple absorption, while the battery filter output remained smooth. Switching waveforms confirmed ZVS: each MOSFET turned on only after its drain-source voltage had fully collapsed to zero—verified by clean, overlap-free gate-drive and voltage signals.
In V2G mode (θ_g = π), the system reversed cleanly—pushing 1 kW back into a simulated 230 V grid, with voltage and current perfectly anti-phase, and the active filter continuing to suppress reverse-flow ripple.
Efficiency wasn’t explicitly tabulated in the paper, but the implications are clear: lower losses across switching (ZVS), conduction (minimized RMS current), and thermal (no capacitor self-heating) add up. Even a 1–2% net gain in OBC efficiency translates to meaningful range extension over a vehicle’s lifetime—especially for fleet operators charging thousands of times.
From a manufacturing perspective, the benefits compound.
Removing the electrolytic capacitor eliminates a major sourcing headache. Film capacitors, while slightly more expensive per microfarad, offer far better volumetric efficiency at high frequencies and require no derating for ripple current. Their flat, low-profile packages also lend themselves to planar PCB layouts and automated assembly—key for high-volume automotive production.
Moreover, with no liquid electrolyte to leak or evaporate, the converter becomes inherently more tolerant of under-hood thermal cycling. Consider a vehicle parked in Arizona summer heat: conventional electrolytics may lose 50% of their capacitance after two years. This design? Its weakest capacitor is now rated for 100,000 hours at 105°C—easily covering 15+ years of operation.
There’s also a subtle systems advantage: bidirectionality by default.
Unlike many unidirectional OBCs retrofitted for V2G (which often require hardware modifications), this topology is symmetric from the ground up. That opens the door to true vehicle-to-everything (V2X) use cases—not just feeding homes during outages, but participating in grid frequency regulation, peak shaving for commercial depots, or even peer-to-peer energy sharing in microgrids.
Imagine a future where your EV doesn’t just consume power—but stabilizes the local grid during evening demand spikes, earning you credits on your utility bill. The economics only work if the hardware is robust enough to cycle thousands of times without degradation. This converter is built for that mission.
Of course, challenges remain before this reaches production.
The control algorithm, while real-time feasible on modern DSPs (e.g., TI C2000 or ST STM32G4), demands precise current and voltage sensing—and tight synchronization between PWM modules. Calibration of dead-time and parasitic parameters becomes critical at 100+ kHz switching.
Thermal design also shifts: while bulk capacitor heating vanishes, localized hot spots may emerge around high-frequency magnetics and switching nodes. Advanced thermal vias, embedded copper coins, or even direct liquid-cooled substrates may be needed for >3 kW versions.
And scaling to 11 kW or 22 kW (common for European and premium EVs) will require careful re-optimization of the resonant tank—balancing L_res and C_res to maintain soft-switching across wider input/output ranges.
But these aren’t roadblocks—they’re engineering refinements. The core concept is sound, validated, and scalable.
Industry watchers note that several Tier-1 suppliers—including Bosch, Valeo, and Delta Electronics—have filed related patents on capacitor-less OBCs in the past 18 months. While details remain confidential, the trend is unmistakable: the industry is converging on active ripple cancellation as the next frontier in power electronics miniaturization.
What’s remarkable about this work is its academic origin. Often, university prototypes prioritize novelty over practicality—adding switches, sensors, or exotic materials that never see a production line. Here, the opposite is true: the design reduces part count, uses off-the-shelf semiconductors (Si MOSFETs suffice at 1 kW), and relies on control ingenuity rather than hardware extravagance.
That’s the hallmark of mature engineering: solving hard problems with elegant simplicity.
For EV owners, the downstream impacts could be significant.
Longer battery warranties—perhaps even lifetime coverage—become more financially viable for automakers when one major degradation vector is neutralized. Thermal management systems can be downsized, freeing up packaging space and reducing weight. And with fewer failure-prone parts, service intervals could extend—lowering total cost of ownership.
For utilities and grid operators, the ripple suppression has another benefit: cleaner power injection during V2G. Harmonic distortion and low-frequency interharmonics—often flagged by grid codes—become easier to meet without bulky EMI filters.
And for sustainability advocates, there’s a lifecycle win: eliminating electrolytic capacitors reduces reliance on aluminum etching and boric acid-based electrolytes—both environmentally sensitive in manufacturing and end-of-life handling.
In many ways, this converter embodies the quiet revolution happening in EV tech: not flashy, not headline-grabbing—but deeply consequential. It’s the kind of innovation that doesn’t change how a car looks or feels to drive, but fundamentally alters its long-term economics, reliability, and environmental footprint.
As charging speeds climb toward 800 V architectures and megawatt-scale depot charging, the pressure on power electronics will only intensify. Solutions that marry high density, high reliability, and bidirectional flexibility aren’t just desirable—they’re essential.
This single-stage, capacitor-free AC/DC converter may well be a stepping stone to that future. Not because it’s the final answer—but because it proves, decisively, that some of the oldest assumptions in power conversion can be rewritten.
And in an industry racing against time and physics, that’s no small feat.
—
Wang Wenlong¹,³, Chen Lihui², Hou Chenguang¹,³
¹Department of Automotive Engineering, Xingtai Vocational and Technical College, Xingtai, Hebei 054000, China
²College of Career Technology, Hebei Normal University, Shijiazhuang, Hebei 050024, China
³Hebei Special Vehicle Modification Technology Innovation Center, Xingtai, Hebei 054000, China
Chinese Journal of Electron Devices, Vol. 46, No. 4, Aug. 2023
doi:10.3969/j.issn.1005-9490.2023.04.020