Half bridge
A half-bridge is a basic power electronics circuit topology consisting of two semiconductor switches, such as MOSFETs or IGBTs, connected in series across a DC power supply, with the load attached to the midpoint between the switches and often including freewheeling diodes for current recirculation.[1] This arrangement enables the production of a bipolar output voltage—alternating between positive and negative halves of the supply voltage—by alternately activating the switches, typically with a short dead time to avoid shoot-through currents.[1] In inverter applications, the half-bridge drives an AC load directly or through a filter, generating square-wave or pulse-width modulated (PWM) outputs, while in DC-DC converters, it couples to a transformer with split input capacitors to step up or down voltage levels without DC bias on the transformer core.[2][3] Key advantages of the half-bridge include its structural simplicity, requiring fewer components than a full-bridge counterpart, lower voltage stress on switches (limited to half the input voltage), and cost-effectiveness for medium-power applications.[1][3] However, it faces challenges such as potential voltage imbalance across the input capacitors in symmetric operation modes and higher switching losses in hard-switched designs without resonant techniques.[3] The topology finds extensive use in switched-mode power supplies (SMPS), uninterruptible power supplies (UPS), solar inverters, electric motor drives, and induction heating systems, where its efficiency and compact design support reliable power conversion in the range of tens to hundreds of watts.[1] In modern variants, such as phase-shifted or zero-voltage-switching (ZVS) half-bridges, enhancements like active clamping reduce electromagnetic interference (EMI) and improve performance for high-frequency operation in telecommunications and computing power systems.[3]Overview
Definition and Basic Concept
A half-bridge circuit is a fundamental topology in power electronics, consisting of two switching elements, such as transistors or MOSFETs, connected in series across a DC power supply, with the load attached to the midpoint of this series connection.[4][1] In its basic form, with the load's other end grounded, this arrangement enables unipolar control of the voltage applied to the load, allowing the output to vary between 0 and the full supply voltage without reverse polarity. However, in common configurations for AC applications like inverters, split DC-link capacitors create a virtual midpoint at Vdc/2, enabling bipolar output (±Vdc/2 relative to the midpoint) for a floating load, in contrast to full-bridge configurations that inherently support bidirectional switching across the full supply. The topology is widely used as a building block in converters and inverters due to its simplicity and efficiency in handling power flow.[2] In the basic schematic, the high-side switch connects to the positive rail of the DC bus voltage, while the low-side switch links to the negative rail or ground, forming a totem-pole structure. The output terminal at their junction delivers the switched voltage to the load. Anti-parallel diodes across each switch facilitate freewheeling currents from inductive loads, ensuring safe operation during switching transitions. This setup provides a single output phase, making it ideal for applications needing precise control over power delivery.[4][5] The half-bridge represents half of an H-bridge circuit, where two such legs can be combined to create a full H-bridge for more versatile control, but standalone it supports unipolar drive for single-ended loads or bipolar with split capacitors. A core mathematical relation for the average output voltage in split-capacitor configurations is V_{out} = \frac{D \cdot V_{dc}}{2}, where D is the duty cycle (the fraction of time the high-side switch is on) and V_{dc} is the DC supply voltage; this highlights the topology's role in modulating power transfer.[6]Historical Context
The modern half-bridge developed in the mid-20th century alongside the rise of thyristor and transistor-based switching devices. The invention of the silicon-controlled rectifier (thyristor) in 1958 by General Electric initiated the era of controlled power conversion, enabling early inverter topologies for industrial applications. By the 1960s, switching inverters for variable-speed AC drives in industries like steel mills and railways began to emerge, building on these innovations to provide efficient power flow.[7] A key milestone occurred in the 1970s with the commercialization of power MOSFETs, first introduced by International Rectifier in 1976, which facilitated high-frequency switching and reduced losses in half-bridge circuits. This led to widespread adoption in switched-mode power supplies (SMPS) for improved efficiency over linear supplies. In the 1980s, further integration into consumer electronics, such as TV horizontal deflection circuits and PC power supplies, was driven by advancements like the insulated-gate bipolar transistor (IGBT), invented by Jayant Baliga in the late 1970s and commercialized in the mid-1980s. Seminal publications from this period, including Alesina and Venturini's 1981 paper on generalized multilevel inverter configurations incorporating half-bridge cells, highlighted their potential for harmonic reduction and high-power applications.[8][7][9] By the 2020s, the evolution of half-bridge technology has been propelled by wide-bandgap semiconductors, enabling operation at frequencies up to the MHz range for compact, high-efficiency designs in electric vehicles and renewable energy systems. Silicon carbide (SiC) and gallium nitride (GaN) devices, with commercial half-bridge modules emerging since the 2010s, have reduced switching losses by up to 50% compared to silicon counterparts, marking a significant leap in performance.[10]Circuit Topology
Core Components and Layout
The half-bridge circuit fundamentally consists of two power switches connected in series across a DC voltage source, forming a basic topology for voltage inversion or amplification. The high-side switch, typically an N-channel MOSFET or IGBT, is positioned between the positive DC rail (V+) and the output midpoint, while the low-side switch, also an N-channel MOSFET or IGBT, connects the midpoint to ground (GND).[11][12] Each switch is typically equipped with an antiparallel freewheeling diode—the intrinsic body diode for MOSFETs or an external fast-recovery diode for IGBTs—to facilitate current recirculation through inductive loads when the switch is off.[2] A DC link capacitor is placed in parallel with the DC source to stabilize the voltage supply and minimize ripple during switching operations.[12] In a standard schematic, the high-side switch (Q1) has its drain connected to V+, source to the midpoint node (output), and gate driven by isolated circuitry; the low-side switch (Q2) has its drain at the midpoint, source to GND, and gate driven directly or via level-shifting. The load attaches to the midpoint node, allowing it to swing between V+ and GND based on switch states. Gate drive circuitry, such as a bootstrap driver for the high-side (including a diode, capacitor, and resistor), ensures proper voltage levels above GND for N-channel devices while incorporating dead-time delays to prevent shoot-through, where both switches conduct simultaneously.[11][13] Physical layout emphasizes minimizing parasitic inductance through compact placement: switches are arranged side-by-side on the PCB, with the DC link capacitor and bypass capacitors positioned adjacent to the switch terminals to form short, low-impedance loops. High-side gate driving often requires isolation via optocouplers, transformers, or integrated isolated drivers to separate control signals from the high-voltage power stage, reducing noise and ensuring safe operation. Traces for power paths are widened and routed with ground planes to enclose currents, while gate traces are kept short and shielded to avoid coupling.[13][12]Variants and Configurations
The three-phase half-bridge configuration extends the basic topology by employing three parallel half-bridge legs, each driving one phase of a three-phase AC motor, enabling balanced voltage generation and precise control for applications like induction motor drives. This variant utilizes six switches (two per phase) to produce sinusoidal outputs, improving power handling and fault tolerance compared to single-phase setups.[14] Resonant half-bridge configurations incorporate LC networks to achieve zero-voltage switching (ZVS), minimizing switching losses in high-frequency operation. In series resonant variants, such as the LLC topology, a series inductor-capacitor tank with a magnetizing inductance enables ZVS across wide load ranges by ensuring the resonant current discharges switch capacitances during dead time. Parallel resonant setups add capacitors across switches for similar ZVS benefits, though they are less common due to higher component stress; both reduce EMI and allow efficiencies above 95% in telecom power supplies.[15] Integrated variants like the IR2110 monolithic IC consolidate high- and low-side gate drivers into a single package, supporting bootstrap operation up to 500 V with independent channels for half-bridge MOSFETs or IGBTs, simplifying PCB layout and reducing external components. This driver delivers ±2 A peak currents with matched propagation delays of 10 ns, enhancing reliability in compact motor drives and inverters.[16] Hybrid forms, such as synchronous rectification half-bridges, replace output diodes with low-side MOSFETs controlled in phase with the switches, cutting conduction losses by leveraging the MOSFET's low R_DS(on) (typically <10 mΩ) instead of diode forward drops of 0.3–0.7 V. This modification boosts efficiency, particularly at low voltages in DC-DC converters, while requiring dead-time management to avoid body diode reverse recovery issues.[17]Operating Principles
Switching Mechanism
The switching mechanism in a half-bridge circuit employs pulse-width modulation (PWM) to alternately activate the high-side and low-side switches, generating a bipolar AC-like output voltage from a DC input supply by controlling the duty cycle of each switch. This alternating conduction ensures that the output voltage swings between +V_{dc}/2 and -V_{dc}/2, where the DC supply is split equally across two series capacitors providing a midpoint reference (0 V) for the load. The high-side switch connects the load to the positive rail (+ V_{dc}/2) during its on period, and the low-side switch connects it to the negative rail (- V_{dc}/2) during its on period. PWM signals are typically generated by a controller, such as a microcontroller or dedicated IC, operating at frequencies ranging from tens of kHz to several hundred kHz to balance efficiency and output quality.[18] Control signals consist of complementary PWM waveforms applied to the switch gates, where the high-side signal is the inverted version of the low-side signal with a 180-degree phase shift, ensuring non-overlapping conduction. However, to prevent simultaneous conduction (shoot-through), a dead time is inserted between the turn-off of one switch and the turn-on of the other, typically lasting 100-500 ns depending on the switching speed and application requirements. The minimum dead time is calculated as t_{dead} = \frac{t_{rise} + t_{fall}}{2}, where t_{rise} and t_{fall} are the rise and fall times of the power switches, providing a margin to account for propagation delays and ensure both switches are fully off during transitions. This insertion is often programmable in dedicated half-bridge drivers via external resistors or internal timers, minimizing conduction losses while avoiding short-circuits across the DC supply.[19][20] Driver circuits are essential for properly interfacing the low-voltage control signals with the power switches, particularly for the high-side switch which operates at a floating potential relative to ground. A common approach is the bootstrap circuit, where a diode and capacitor form a charge pump: the capacitor charges from the low-side supply through the diode when the low-side switch is on, and this stored charge powers the high-side gate driver during its conduction period, providing a gate voltage typically 10-15 V above the switch source. Level-shifting techniques, such as those using high-speed optocouplers or integrated level-shifters in gate driver ICs, offer alternatives for applications requiring galvanic isolation or precise timing, though bootstrap remains prevalent due to its simplicity and cost-effectiveness in non-isolated topologies.[21][19]Voltage and Current Waveforms
In a half-bridge circuit, the output voltage waveform in basic square-wave operation toggles between +V_{dc}/2 and -V_{dc}/2, resulting in a bipolar square wave with a fundamental component peak amplitude of $2 V_{dc} / \pi and RMS value of V_{dc} / \pi. The square wave itself has an RMS value of V_{dc}/2.[22] When pulse-width modulation (PWM) is applied to the switches, the output voltage becomes a modulated waveform that approximates a sinusoid for applications like motor drives, with the instantaneous value given by V_{out}(t) = V_{dc} (D(t) - 0.5), where D(t) is the PWM duty cycle varying between 0 and 1.[23] For current waveforms, resistive loads produce a waveform mirroring the voltage shape, but inductive loads introduce a phase lag \phi = \tan^{-1}(\omega L / R), where \omega is the fundamental frequency, L the inductance, and R the resistance, causing the current to lag the voltage and continue flowing through freewheeling diodes during switch-off periods to maintain inductor current continuity.[24] This freewheeling path, typically via the anti-parallel diodes of the opposite switch, prevents abrupt current interruption and results in a quasi-sinusoidal current waveform under balanced operation, with peak values determined by the load impedance Z = \sqrt{R^2 + (\omega L)^2}.[2] The unfiltered voltage output contains a fundamental component along with odd harmonics due to the square-wave nature, with amplitudes V_n = \frac{2 V_{dc}}{n \pi} for odd n, leading to a total harmonic distortion (THD) of approximately 48% in the basic case, calculated as THD = \sqrt{\sum_{n=3,5,\dots}^{\infty} (V_n / V_1)^2}.[25] PWM modulation shifts higher harmonics to multiples of the carrier frequency, reducing low-order content and lowering THD to levels like 25% for bipolar schemes at modulation indices around 0.8.[23] Practical implementations exhibit waveform distortions from dead time, the intentional delay between complementary switch transitions to prevent shoot-through, which introduces voltage notches or asymmetry depending on load current direction—causing a positive current to experience a voltage drop toward 0 during dead time, while negative current pulls it toward V_{dc}.[26] This effect reduces the fundamental voltage by up to 18% for dead times of 4 μs at 50 Hz and amplifies low-order harmonics, increasing THD and necessitating compensation techniques for precision control.[27]Applications
In Motor Control Systems
Half-bridges are commonly employed in motor control systems to drive DC motors, where a single half-bridge configuration enables unidirectional speed control through pulse-width modulation (PWM). In this setup, the high-side and low-side switches of the half-bridge alternately connect the motor to the supply voltage or ground, with the PWM duty cycle determining the average voltage and thus the motor speed.[28] For bidirectional control, two half-bridges are paired to form an H-bridge, allowing reversal of the voltage polarity across the motor by activating the appropriate switch pairs, which facilitates direction changes in addition to speed regulation via PWM.[29] In stepper motor applications, half-bridges are integrated into bipolar driver integrated circuits, such as the A4988, to manage phase excitation by controlling current flow through the motor coils. Each phase of the bipolar stepper requires an H-bridge composed of two half-bridges, enabling precise bidirectional current steering for microstepping modes that enhance resolution and smoothness during operation.[30] This configuration supports full-step, half-step, and finer microstepping, with the half-bridges handling up to 2 A per coil to achieve accurate positioning without external translators in many designs.[31] Half-bridge-based PWM control in these systems contributes to torque ripple reduction by minimizing current fluctuations in the motor windings, as higher switching frequencies smooth the average current delivery. Typical PWM frequencies range from 10 to 20 kHz, selected to avoid audible noise generation in the human hearing range (20 Hz to 20 kHz) while balancing switching losses.[32] For example, in brushless DC (BLDC) motors, three half-bridges form a three-phase inverter for six-step commutation, where sequential activation of the bridges energizes two phases at a time based on rotor position feedback from Hall-effect sensors, producing a trapezoidal back-EMF waveform for efficient torque production.[33]In Switched-Mode Power Supplies
In switched-mode power supplies (SMPS), the half-bridge topology serves as a fundamental configuration for both isolated DC-DC converters and resonant designs, enabling efficient power conversion in applications requiring galvanic isolation and high power density. In forward converters, the half-bridge arrangement utilizes two switches and a pair of capacitors to apply half the input voltage across the transformer primary, facilitating isolated output voltages suitable for supplies up to 500 W, particularly in universal input ranges of 90-264 V AC.[34] This topology derives from the buck converter principle, where energy transfer occurs during the switch-on period, making it ideal for isolated supplies in distributed power architectures.[35] A prominent variant is the LLC resonant half-bridge converter, which incorporates a resonant tank circuit to achieve zero-voltage switching (ZVS) and zero-current switching (ZCS), significantly reducing switching losses and electromagnetic interference (EMI). This design excels in high-efficiency telecom power supply units (PSUs), where it supports front-end DC-DC conversion from high input voltages (e.g., 400 V post-PFC) to intermediate rails, operating at frequencies around or above resonance for optimal load regulation across wide ranges, including zero-load conditions.[15] In telecom applications, the LLC half-bridge enables compact, high-density PSUs with efficiencies exceeding 94%, prioritizing soft-switching to minimize heat and component stress in noise-sensitive environments.[36] The conversion process in half-bridge SMPS relies on transformer coupling to step down voltages while maintaining isolation, as seen in 48 V to 12 V DC-DC applications common in data centers and automotive systems. During operation, the half-bridge generates an alternating voltage across the primary winding, inducing a proportional output on the secondary, which is rectified and filtered to produce the desired DC rail; this setup achieves step-down ratios of 4:1 with peak efficiencies over 95% through resonant operation and synchronous rectification.[15] For instance, in LLC-based designs, the transformer's magnetizing inductance integrates with the resonant elements to ensure efficient energy transfer, supporting outputs up to 60 A at 12 V while keeping ripple currents low.[37] For AC output generation, half-bridge configurations find use in single-phase inverters within uninterruptible power supply (UPS) systems, where sinusoidal pulse-width modulation (SPWM) modulates the switches to produce a 50/60 Hz sine wave from a DC source. The half-bridge drives a transformer or filter to shape the output, enabling reliable backup power for critical loads; waveform generation aligns with principles of pulse modulation to approximate sinusoidal voltage with minimal harmonics.[38] A practical example is in ATX PC power supplies, where the half-bridge topology converts the rectified AC input to stable 5 V and 12 V rails for motherboard and peripheral powering, often employing resonant variants for compliance with 80 PLUS efficiency standards. This setup handles outputs up to 750 W, balancing core utilization and ripple reduction in compact desktop designs.[39]Design and Implementation
Component Selection Criteria
In half-bridge designs, switch selection begins with ensuring the voltage rating exceeds 1.5 times the DC bus voltage (V_dc) to accommodate voltage spikes from parasitic inductances and overshoot during switching transients.[40] The current rating must surpass the peak load current, including margins for inrush and fault conditions, to prevent thermal runaway and ensure reliable operation under dynamic loads.[41] For applications below 1 kW, MOSFETs with low on-state resistance (R_ds(on)) are preferred to minimize conduction losses and enable high-frequency switching, while for power levels exceeding 1 kW, insulated-gate bipolar transistors (IGBTs) offer advantages in handling higher voltages and currents with reduced saturation voltage drop.[42] The DC-link capacitor is sized to limit voltage ripple (ΔV) while managing RMS ripple current, using the formula C = \frac{I_{C,\mathrm{rms}}}{\omega \Delta V}, where I_{C,\mathrm{rms}} is the RMS capacitor current, \omega = 2 \times 2\pi f_{\mathrm{sw}} (angular ripple frequency, twice the switching frequency for half-bridge operation), and ΔV is the maximum allowable peak-to-peak ripple.[43] The RMS capacitor current I_{C,\mathrm{rms}} is typically approximated from the load RMS current (e.g., I_{C,\mathrm{rms}} \approx 1.3 \times I_{M,\mathrm{rms}} / 2 for certain inverter configurations). Capacitors should feature low equivalent series resistance (ESR) and inductance (ESL) to handle high-frequency currents, with voltage ratings at least 1.1 times the nominal DC bus voltage and operation up to 105°C to maintain longevity in automotive or industrial environments.[44] Gate drivers for half-bridges must support high-side floating operation up to the bootstrap voltage limit (e.g., 120 V), providing sufficient peak source and sink currents (typically 4 A) to charge/discharge MOSFET gate capacitances rapidly.[45] Controlled slew rates, such as 7-10 V/ns, balance electromagnetic interference (EMI) reduction—by mitigating dV/dt-induced noise—against switching losses, with integrated features like UVLO and delay matching enhancing robustness. An example is the UCC27211 IC, which offers high-side capability with 20 ns propagation delay and ESD protection for noisy power electronics environments.[45] Thermal management is critical, with junction temperatures limited to 125-150°C depending on the device datasheet to avoid degradation.[46] Power loss (P_loss) from conduction and switching is calculated per device, and heatsink sizing ensures the junction-to-ambient thermal resistance (θ_ja) satisfies T_j = T_a + \theta_{ja} \cdot P_{loss}, where T_a is ambient temperature; for natural convection, θ_ja values of 20-50°C/W are targeted by selecting heatsink volumes of 500-800 cm³°C/W.[46] Environmental factors like airflow and enclosure temperature influence final sizing, prioritizing devices with low θ_jc for efficient heat spreading.[46]Protection and Efficiency Features
Half-bridge circuits incorporate several protection mechanisms to safeguard against faults and ensure reliable operation. Overcurrent sensing is commonly implemented using shunt resistors placed in the low-side path to monitor load current; when the voltage drop across the resistor exceeds a threshold, the driver IC triggers shutdown to prevent damage from excessive currents. Desaturation detection (DESAT) provides short-circuit protection by sensing the collector-emitter or drain-source voltage of the power switch during on-state; if the voltage fails to drop below a set level (indicating a fault like a short-circuit), the gate drive is interrupted, often with a soft turn-off to limit fault current—typically protecting against currents up to five times the rated value. Undervoltage lockout (UVLO) monitors the supply voltage and disables the driver if it falls below a safe threshold (e.g., 8-10 V for high-side and 4-5 V for low-side), preventing partial conduction that could lead to shoot-through or overheating.[47][48][49] Efficiency in half-bridge designs is enhanced through techniques that minimize switching and conduction losses. Soft-switching methods, such as zero-voltage switching (ZVS) and zero-current switching (ZCS), are employed in resonant topologies like the LLC half-bridge converter; ZVS turns on switches when their voltage is zero, discharging parasitic capacitances via resonant current during dead time, while ZCS applies to secondary rectifiers by ensuring zero current at turn-off, collectively reducing switching losses by up to 20% and enabling operation at higher frequencies without excessive heat. Synchronous rectification replaces output diodes with actively driven MOSFETs, lowering forward voltage drop from ~0.7 V (diode) to ~100 mV (MOSFET R_DS(on)), which cuts conduction losses in the rectification stage—particularly beneficial in high-output-current applications like DC-DC converters, improving overall efficiency by 2-5%.[15][17] Losses in half-bridge circuits are quantified to optimize thermal management and efficiency, with total power loss given byP_{\text{total}} = P_{\text{cond}} + P_{\text{sw}} + P_{\text{driver}},
where P_{\text{cond}} represents conduction losses from on-state resistance (R_{\text{DS(on)}} or V_{\text{CE(sat)}}) multiplied by RMS current squared, P_{\text{sw}} accounts for dynamic switching events, and P_{\text{driver}} includes quiescent current and gate drive overhead. Switching losses are approximated as
P_{\text{sw}} = \frac{V_i I_o t_{\text{sw}} f_{\text{sw}}}{2},
per switch, where V_i is input voltage, I_o is output current, t_{\text{sw}} is transition time, and f_{\text{sw}} is switching frequency; this formula captures the energy dissipated during voltage-current overlap, emphasizing the need to minimize t_{\text{sw}} through fast drivers or soft switching.[50][51] Integrated diagnostics in half-bridge driver ICs provide fault flags for real-time monitoring and fault-tolerant operation, especially in automotive applications. These ICs output status signals via pins like nFAULT, indicating conditions such as overcurrent, overtemperature, or open-load; for instance, in devices like the DRV8243-Q1, configurable registers report specific faults (e.g., OCP on high-side FET) through SPI, enabling automatic retry or latch-off with recovery after fault clearance. In automotive half-bridges, such features support ASIL compliance by including self-diagnostics for supply integrity and thermal shutdown, ensuring safe operation in motor drives under harsh conditions.[52][53]