A buck converter (also called a step-down converter) is a switching DC-DC power supply topology that produces an output voltage lower than its input voltage by rapidly switching a transistor on and off, storing energy in an inductor, and filtering the result with a capacitor. It is one of the most common power conversion topologies in embedded systems, used wherever an efficient voltage step-down is needed.
In practice
Buck converters appear throughout embedded hardware wherever a supply rail must be stepped down efficiently. Common examples include converting a 12 V or 5 V board supply to a 3.3 V or 1.8 V core rail for an MCU or SoC, powering a Cortex-A application processor core from a multi-cell Li-ion battery, or generating the 1.0 V supply that many DDR memory interfaces require. Efficiency in well-designed discrete buck stages typically ranges from 85% to 95%, making them far preferable to linear regulators (LDOs) when the input-to-output voltage difference is large or the load current is high.
The core operating principle is duty-cycle control. A high-side switch (usually a MOSFET) connects the input to the inductor for a fraction of each switching period; a low-side switch or diode then allows inductor current to recirculate to ground. The ratio of on-time to total period sets the conversion ratio: in an ideal buck converter operating in continuous conduction mode (CCM), V_out ≈ D × V_in, where D is the duty cycle. Real converters deviate from this due to switch voltage drops, inductor resistance, and other losses. Most integrated buck converter ICs handle gate driving, oscillator, and feedback internally, exposing only a feedback resistor divider and a few passives to the designer.
Inductor and capacitor selection are the main design variables a firmware or hardware engineer will encounter when working with discrete or semi-integrated buck regulators. Choosing too small an inductor increases ripple current, which stresses the output capacitor and can push the converter into discontinuous conduction mode (DCM) at light loads, changing its transfer function. Capacitor ESR (equivalent series resistance) directly sets output voltage ripple and affects control-loop stability. The EmbeddedRelated article "Lost Secrets of the H-Bridge, Part III: Practical Issues of Inductor and Capacitor Ripple Current" discusses ripple current considerations that apply directly to buck converter passive design.
A common pitfall is ignoring switching-frequency-related EMI. Buck converters switch at frequencies typically ranging from around 100 kHz in lower-frequency discrete designs to 2-4 MHz in modern integrated ICs targeting small passive sizes. The fast switching edges (with rise times that can reach low nanoseconds in advanced MOSFET-based designs) generate conducted and radiated noise that can couple into ADC references, RF front ends, or sensitive analog circuitry. Proper layout, grounding, and input bypass capacitors placed close to the IC's input and GND pins are essential and are often underestimated, especially when PCB layout is done after schematic capture.
Frequently asked
When should I use a buck converter instead of an LDO?
Prefer a buck converter when the input-to-output voltage difference is large (say, dropping from 12 V to 3.3 V) or when load current is high enough that the power dissipated in an
LDO (P = (V_in - V_out) × I_load) would require a heatsink or cause thermal problems. LDOs are simpler, quieter, and cheaper for small voltage drops or very low current loads where efficiency matters less.
What sets the output voltage of a buck converter?
In most integrated buck ICs, a resistor divider from the output to a feedback (FB) pin sets the output voltage. The IC's internal reference (commonly 0.6 V to 1.2 V, depending on the device) is compared to the divided-down output, and the duty cycle is adjusted to keep them equal. The output voltage is then V_out = V_ref × (1 + R_top / R_bot). Some ICs instead accept a digital interface (
I2C or
SPI) for programmatic voltage adjustment.
What is the significance of continuous vs. discontinuous conduction mode (CCM vs. DCM)?
In CCM,
inductor current never reaches zero during a switching cycle; in DCM it does. CCM gives a simpler, load-independent conversion ratio (V_out ≈ D × V_in for an ideal converter) and is typical at moderate to heavy loads. DCM occurs at light loads with a given inductor value, and the conversion ratio becomes load-dependent, which can complicate control-loop design. Many modern integrated converters include a pulse-skipping or burst mode to maintain efficiency in DCM at light loads.
How does switching frequency affect passive component size and noise?
Higher switching frequency allows smaller
inductors and output
capacitors for the same ripple specification, which is why many
MCU-targeted integrated bucks run at 1-4 MHz. The trade-off is increased switching losses in the FETs and gate drivers, and more high-frequency
EMI. Lower switching frequencies (100-500 kHz) are common in higher-power or efficiency-critical designs where larger passives are acceptable.
What layout rules matter most for a buck converter PCB?
The most critical loop is the high-frequency switched loop: input
bypass capacitor, high-side
MOSFET, and low-side MOSFET (or
diode). Keep this loop as small as possible to minimize stray inductance and radiated
EMI. The switch node (the junction of high-side switch, low-side switch, and
inductor) is the noisiest point on the board; keep it compact and away from sensitive analog nets. Place the input bypass
capacitor as close as physically possible to the IC's input and GND pins.
Differentiators vs similar concepts
A buck converter is often contrasted with a
boost converter (step-up: output voltage higher than input) and a
buck-boost converter (output can be higher or lower than input, at the cost of added complexity and typically lower efficiency). It is also frequently compared to a
low-dropout regulator (LDO): both produce a lower output voltage from a higher input, but an LDO is a
linear regulator that dissipates the excess voltage as heat, while a buck converter switches to transfer energy efficiently. A charge pump is another switching alternative, but charge pumps use
capacitors rather than
inductors and are generally limited to low currents and fixed or simple conversion ratios.
