The relentless global consumption of energy and the severe environmental pollution caused by the combustion of fossil fuels have intensified the worldwide focus on the development and utilization of renewable energy sources. Among these, photovoltaic (PV) power generation stands as the most widely applied technology today. Within any PV generation system, the power conversion unit—the solar inverter—is the most critical core component. For small-power, distributed PV systems, the microinverter architecture has emerged as a prevailing trend due to its inherent advantages: modularity, independent Maximum Power Point Tracking (MPPT) capability for each panel, and lower installation and maintenance costs. Over recent years, photovoltaic microinverters have seen extensive deployment in grid-connected systems, driven by their benefits of enhanced energy harvest, “plug-and-play” modularity, and system flexibility. Among the numerous topologies considered for microinverters, the flyback converter has consistently been a primary candidate, prized for its structural simplicity, ease of control, and inherent electrical isolation.
However, conventional flyback-based solar inverters often grapple with significant efficiency limitations. The primary switch typically operates under hard-switching conditions, leading to substantial switching losses. Furthermore, the diode on the secondary side suffers from conduction losses and reverse recovery losses. To push the boundaries of efficiency and power density for these crucial components in modern solar arrays, a deeper examination of the flyback converter’s operational modes is essential. The two most common modes for such solar inverters are Discontinuous Conduction Mode (DCM) and Boundary Conduction Mode (BCM). Each presents a distinct trade-off: DCM offers lower, fixed-frequency switching which is efficient at light loads but suffers at higher power due to increased RMS currents, while BCM enables higher power density and conversion ratios but incurs greater switching losses, making it less efficient at light loads.
This continuous pursuit of higher efficiency in power conversion drives innovation. The quest for more efficient, compact, and reliable solar inverters is unending, as it directly translates to improved energy yield and cost-effectiveness of PV installations. In this context, this work proposes a novel flyback-based photovoltaic microinverter topology integrated with a hybrid control strategy. The core innovation lies in the strategic incorporation of a snubber capacitor across the main switch, coupled with an intelligent, load-dependent operational mode that blends DCM and a quasi-resonant BCM. This approach is specifically designed to mitigate key loss mechanisms, enhance soft-switching capabilities, and ultimately elevate the overall weighted efficiency of the microinverter system.
Fundamental Principles of the Low-Loss Flyback Converter
The proposed circuit modifies the traditional flyback topology by introducing a snubber capacitor, \(C_s\), directly across the drain and source of the main power switch, \(S\). An auxiliary switch, \(S_a\), is placed in series with this capacitor. The magnetizing inductance and the leakage inductance of the transformer are denoted as \(L_m\) and \(L_k\), respectively. The total primary inductance is \(L_{tot} = L_m + L_k\). The control philosophy employs a hybrid approach: under light-load conditions, the converter operates in standard DCM with a fixed, low switching frequency and the auxiliary switch \(S_a\) remains off, effectively removing \(C_s\) from the active switching circuit. Under heavy-load conditions, \(S_a\) is turned on and held conducting, permanently connecting \(C_s\) in parallel with the main switch’s output capacitance \(C_{oss}\). The system then operates in a “BCM with Quasi-Resonance” mode. This adaptive strategy leverages the efficiency benefits of each mode across the operating range.
The operation in the high-efficiency “BCM+Quasi-Resonant” mode is detailed through a modal analysis over one high-frequency switching cycle, assuming the input voltage \(V_{in}\) (from the PV panel) and the reflected output voltage \(V_o/N\) are constant within this cycle.
| Mode | Interval | Conducting Devices | Key Energy Transfers |
|---|---|---|---|
| 1. Magnetizing | \(t_0\) – \(t_1\) | S (ON), D (OFF) | Energy from input source stored in \(L_{tot}\). |
| 2. Capacitive Charging / Turn-off | \(t_1\) – \(t_2\) | S (OFF), D (OFF) | Energy from \(L_k\) charges \(C_{tot}\) (\(C_s//C_{oss}\)), slowing \(v_{ds}\) rise. |
| 3. Demagnetizing / Power Transfer | \(t_2\) – \(t_3\) | S (OFF), D (ON) | Energy from \(L_m\) and \(L_k\) transferred to output. |
| 4. Quasi-Resonant Reset | \(t_3\) – \(t_4\) | S (OFF), D (OFF) | \(L_m\) and \(C_{tot}\) resonate, recycling energy to input. |
| 5. Zero-Voltage Switching (ZVS) Window | \(t_4\) – \(t_5\) | Body diode of S (ON), D (OFF) | \(v_{ds}\) clamped near zero, enabling ZVS turn-on of S. |
Mode 1 [\(t_0 – t_1\)]: At \(t_0\), the main switch \(S\) is turned on. The input voltage \(V_{in}\) is applied across \(L_{tot}\), causing the primary current \(i_{Lm}\) to increase linearly from zero. The secondary diode \(D\) is reverse-biased. The load is supplied by the output capacitor \(C_o\). This mode ends when \(i_{Lm}\) reaches the sinusoidal reference current \(i_{ref}\) at \(t_1\), and \(S\) is turned off. The duration of this interval is controlled to shape the input current.
$$ i_{Lm}(t) = \frac{V_{in}}{L_{tot}} t $$
$$ \Delta t_{01} = t_1 – t_0 = \frac{i_{ref} L_{tot}}{V_{in}} $$
Mode 2 [\(t_1 – t_2\)]: At \(t_1\), \(S\) turns off. The energy stored in the leakage inductance \(L_k\) initiates a resonant transition with the total capacitance \(C_{tot} = C_s + C_{oss}\). This capacitance is charged by the current \(i_{Lm}(t_1) = i_{ref}\), causing the drain-source voltage \(v_{ds}\) to rise. The presence of \(C_s\) significantly increases \(C_{tot}\), which slows the rate of voltage rise (\(dv/dt\)) across \(S\), thereby reducing its turn-off switching loss. This is a primary function of the snubber in these solar inverters. The resonant frequency is:
$$ \omega_r = \frac{1}{\sqrt{L_k C_{tot}}} $$
Assuming a short resonant period, the voltage rise can be approximated as linear. The time for \(v_{ds}\) to rise from zero to the clamp level \((V_{in} + V_o/N)\) is:
$$ \Delta t_{12} \approx \frac{(V_{in} + V_o/N) C_{tot}}{i_{ref}} $$
Mode 3 [\(t_2 – t_3\)]: Once \(v_{ds}\) reaches \(V_{in} + V_o/N\) at \(t_2\), the secondary diode \(D\) becomes forward-biased and conducts. The voltage across the primary winding is clamped to \(V_o/N\). The energy stored in the magnetizing inductance \(L_m\) is now transferred to the output. The secondary current \(i_D\) and the primary current \(i_{Lm}\) decay linearly to zero. The duration of this demagnetization period is:
$$ \Delta t_{23} = \frac{i_{ref} L_m}{V_o / N} $$
Mode 4 [\(t_3 – t_4\)]: At \(t_3\), the secondary current reaches zero. With no secondary clamping, the magnetizing inductance \(L_m\) and the capacitance \(C_{tot}\) begin to resonate. The energy remaining in \(C_{tot}\) (which was charged during Mode 2) is now fed back to the input source through this resonance, resetting the snubber capacitor. This energy recovery is a key efficiency-boosting feature. The resonant behavior causes \(v_{ds}\) to sinusoidally decrease. If the parameters are designed such that:
$$ \frac{1}{2} C_{tot} (V_{in}+V_o/N)^2 > \frac{1}{2} L_m i_{Lm}(t_3)^2 $$
then \(v_{ds}\) can resonate back to zero within a half resonant period. The time for this voltage fall is governed by:
$$ \Delta t_{34} = \pi \sqrt{L_m C_{tot}} $$
This is the quasi-resonant phase that enables soft switching.
Mode 5 [\(t_4 – t_5\)]: When \(v_{ds}\) resonates to zero (or a minimum value) and the resonant current through \(L_m\) goes negative, it flows through the anti-parallel body diode of the main switch \(S\). The main switch can then be turned on during this interval with virtually zero voltage across it, achieving Zero-Voltage Switching (ZVS). This eliminates the turn-on switching loss, which is a major advantage for high-frequency solar inverters. A fixed delay \(T_d\), set to the maximum possible \(\Delta t_{34}\), ensures the turn-on signal is issued when \(v_{ds}\) is at its minimum.
Control Strategy for the Hybrid Mode Microinverter
The overall control system for this advanced flyback microinverter is designed to manage power flow, ensure grid synchronization, and implement the hybrid DCM/BCM mode selection seamlessly. The strategy integrates several key functional blocks commonly found in high-performance solar inverters.
The process begins with sampling the PV array’s output current \(i_{PV}\) and voltage \(v_{PV}\). These values are fed into an MPPT algorithm (e.g., Perturb and Observe or Incremental Conductance) which continuously computes the reference for maximum available power, \(P_{o\_mppt}\). Simultaneously, the grid voltage \(v_g\) is sampled and processed by a Phase-Locked Loop (PLL) to accurately extract the grid phase angle \(\theta\) and frequency. This angle is crucial for synchronization.
Using \(P_{o\_mppt}\) and \(\theta\), the controller generates two sinusoidal half-wave reference current envelopes: \(i_{ref\_DCM}\) for DCM operation and \(i_{ref\_BCM}\) for BCM operation. Their peak values are determined by the power level and the characteristic equations of each mode. A mode selection block then compares the instantaneous power demand against a predefined threshold. Below this threshold, the DCM reference is selected, the auxiliary switch \(S_a\) is kept off, and the main switch \(S\) is controlled with fixed-frequency peak-current-mode control. Above the threshold, the system switches to the BCM+Quasi-resonant mode: \(S_a\) is turned on and held on, and the controller uses the BCM reference \(i_{ref\_BCM}\) to govern \(S\). The turn-off of \(S\) is triggered when the sensed primary current \(i_p\) reaches \(i_{ref\_BCM}\). The turn-on is triggered after a fixed delay \(T_d\) following the detection of zero secondary current \(i_d\), exploiting the quasi-resonant valley for ZVS.
Finally, the sinusoidal half-wave output from the flyback stage is unfolded into a full sine wave using a grid-frequency unfolding H-bridge. The switching signals for this H-bridge are directly derived from the PLL’s phase angle \(\theta\) to ensure the output current is perfectly in phase with the grid voltage. An LC filter at the output attenuates the high-frequency switching ripple to meet grid interconnection standards.
| Parameter | Symbol | Design Equation / Consideration |
|---|---|---|
| Peak Primary Current (BCM) | \(I_{p\_peak}\) | $$ I_{p\_peak}(\theta) = \frac{2 \sqrt{2} P_o}{\eta V_{in\_min} \sin(\theta)} $$ Where \(\eta\) is estimated efficiency, \(P_o\) is output power. |
| Magnetizing Inductance | \(L_m\) | Chosen based on power level and desired BCM/DCM boundary frequency. A smaller \(L_m\) allows higher power but increases current ripple. |
| Snubber Capacitance | \(C_s\) | $$ C_s = \frac{i_{ref} \cdot t_{fall}}{V_{clamp}} – C_{oss} $$ Designed to achieve desired \(dv/dt\) at turn-off and ensure complete resonant reset within available time. Typically tens of nanofarads. |
| Resonant Reset Time | \(T_{reset}\) | $$ T_{reset} = \pi \sqrt{L_m (C_s + C_{oss})} $$ Must be less than the minimum off-time available at peak power to ensure ZVS condition. |
| Mode Change Threshold | \(P_{threshold}\) | Determined empirically or analytically by comparing loss models for DCM and BCM+QR operations to maximize weighted efficiency. |
Experimental Verification and Performance Analysis
The operational principles of the proposed flyback-based solar inverter were validated through a constructed prototype. The experimental waveforms vividly demonstrate the benefits introduced by the hybrid mode and the snubber capacitor.
Under BCM+Quasi-resonant operation, the measured drain-source voltage \(v_{ds}\) and gate drive signal for the main switch reveal critical improvements. At the instant the gate drive falls (turn-off command), the \(v_{ds}\) voltage begins to rise. Notably, the slope of this rise is deliberately slowed compared to a hard-switching case, a direct consequence of the added snubber capacitor \(C_s\) increasing the effective \(C_{tot}\). This reduced \(dv/dt\) directly correlates to lower turn-off loss. Furthermore, the voltage overshoot beyond the theoretical clamp voltage \((V_{in} + V_o/N)\) is significantly suppressed, allowing for the use of a main switch with a lower voltage rating—a key factor in reducing cost and conduction loss in solar inverters.
Following the demagnetization phase, the characteristic quasi-resonant dip in \(v_{ds}\) is clearly observed. The voltage resonates downwards, and after reaching its minimum point, the gate drive signal is reapplied to turn on the main switch. This turn-on occurs when \(v_{ds}\) is at or near zero, confirming the achievement of Zero-Voltage Switching (ZVS). The elimination of the capacitive turn-on loss associated with \(C_{tot}\) during this event is a major contributor to the system’s high efficiency in the high-power mode.
Measurements of the grid voltage \(v_g\) and the secondary-side diode current \(i_d\) showcase the quality of the power processing. Thanks to the primary-side peak current control, the envelope of \(i_d\) faithfully follows a rectified sinusoidal shape. After processing by the grid-synchronized unfolding H-bridge, the final AC output current is a clean, low-distortion sine wave that maintains precise phase alignment with the grid voltage, satisfying the fundamental requirements for safe and efficient grid interconnection for photovoltaic systems.
| Feature | Mechanism | Benefit for Solar Inverters |
|---|---|---|
| Hybrid DCM/BCM Control | Load-adaptive mode switching. | Optimizes efficiency across the entire power range, improving weighted EURE and energy yield. |
| Snubber Capacitor (\(C_s\)) | Slows \(dv/dt\) at main switch turn-off. | Reduces turn-off switching loss; limits voltage overshoot enabling lower-voltage, lower-Rds(on) switches. |
| Quasi-Resonant Reset | Resonant energy recovery from \(C_{tot}\) to input. | Recycles trapped leakage energy and enables ZVS turn-on, drastically cutting switching losses in BCM mode. |
| Peak Current Control & PLL | Shapes input current, synchronizes to grid. | Ensures high power factor, low THD in grid current, and stable grid-tied operation. |
Conclusion
This work has presented a novel and efficient topology for flyback-based photovoltaic microinverters. The core innovation rests on two pillars: a hybrid DCM/BCM control strategy that intelligently selects the most efficient operating mode based on load conditions, and the introduction of a strategically placed snubber capacitor that is activated during the high-power BCM mode. This capacitor is not merely a passive snubber; it is an integral part of a quasi-resonant circuit that actively shapes the switching trajectory. Its functions are multifaceted: it limits voltage overshoot and reduces turn-off loss by controlling the \(dv/dt\), it facilitates the recycling of leakage inductance energy back to the source, and most importantly, it creates the resonant conditions necessary for achieving Zero-Voltage Switching (ZVS) of the main power switch.
The proposed approach addresses several classic limitations of conventional flyback solar inverters. By mitigating switching losses through soft-switching and reducing voltage stress on the semiconductor devices, it paves the way for higher switching frequencies, which can lead to increased power density—a highly desirable trait for modular microinverters. The hybrid control ensures that these benefits are harnessed where they matter most (at high load) while maintaining good light-load efficiency through DCM operation. Experimental results have validated the theoretical modal analysis, demonstrating the effective voltage clamping, the quasi-resonant behavior, and the successful ZVS turn-on. Therefore, this topology and control method represent a significant step forward in enhancing the performance, efficiency, and practicality of flyback converters for next-generation, distributed photovoltaic energy systems.