How a Buck Converter Reduces DC Voltage
A buck converter is a power circuit that converts a higher DC voltage into a lower, regulated DC voltage. It does this by rapidly switching the input, storing energy in an inductor (a component that resists sudden changes in current), and using feedback to adjust the switching pattern.
The problem it solves
Suppose a system has a 12 V supply but a processor needs 5 V. A resistor divider can produce 5 V only for a particular load current; when the load changes, its output changes too. A linear regulator can hold the voltage steady, but the excess voltage becomes heat. At 12 V in and 5 V out, a 1 A load wastes 7 W in the regulator.
A buck converter avoids most of that loss by making its switching device either mostly on or mostly off. In those states, the device ideally has either a small voltage across it or a small current through it. The converter transfers energy in controlled packets instead of continuously burning the voltage difference.
The three essential parts
A basic buck converter contains:
- A switching transistor, usually a MOSFET, that connects and disconnects the input from the rest of the circuit at high speed.
- An inductor that receives energy while the switch is on and releases it while the switch is off.
- A feedback controller that measures the output voltage and changes the switching pattern to keep that voltage near its target.
There is also an output capacitor. It stores charge and supplies the load between switching events, smoothing the inductor's changing current. A conventional converter uses a diode to provide the inductor's current path when the transistor is off. A synchronous buck replaces that diode with a second transistor to reduce conduction loss.
One switching cycle
Consider an ideal buck converter with input voltage Vin, output voltage Vout, and a switching period T. During each period, the main transistor is on for a fraction called the duty cycle, written D.
Switch on
When the transistor turns on, the input is connected to the inductor and output network. The inductor sees approximately Vin - Vout across it. Its current therefore rises, with a slope of approximately:
(Vin - Vout) / L
where L is the inductance.
The inductor current does not jump instantly. That is the defining behavior of an inductor: changing its current requires a voltage over time. While the current rises, the inductor stores magnetic energy.
Switch off
When the transistor turns off, the inductor still has current and cannot simply stop conducting. The diode, or the synchronous low-side transistor, provides a path for that current into the output. The inductor now sees approximately -Vout, so its current decreases at roughly:
-Vout / L
The current continues feeding the load and recharging the output capacitor until the main transistor turns on again.
In steady operation, the inductor current must return to the same value at the end of each cycle. For the common case where it never reaches zero, called continuous conduction mode (CCM), this balance gives the ideal relationship:
Vout ≈ D × Vin
Thus, a 12 V input needs a duty cycle of about 0.42 to produce 5 V. Real converters need a somewhat different duty cycle because of transistor resistance, diode voltage drop, inductor resistance, and other losses.
Where inductor-current ripple comes from
The inductor current is triangular or sawtooth-like rather than perfectly constant. Its ripple current is the difference between its peak and valley current during one switching cycle.
During the on-time, D × T, the current rises by approximately:
ΔI = (Vin - Vout) × D × T / L
Using switching frequency fs = 1/T, the same result is:
ΔI = (Vin - Vout) × D / (L × fs)
During the off-time, the current falls by the same amount in steady state. Higher inductance or higher switching frequency reduces ripple. A larger voltage difference across the inductor increases it.
The output capacitor absorbs much of this changing current, so the load sees a smoother voltage. It cannot remove everything: capacitor properties, inductor ripple, wiring resistance, and control behavior all contribute to output-voltage ripple. Ripple is therefore a normal operating behavior, not automatically a fault.
A useful mental model is that the transistor controls when energy enters the inductor, while the inductor and capacitor smooth how that energy reaches the load.
Feedback closes the loop
The converter measures a scaled version of its output and compares it with an internal reference voltage. If the output falls below the target because the load increased, the controller generally increases the duty cycle or otherwise sends more energy per cycle. If the output rises, it reduces the duty cycle.
The controller may use a fixed switching frequency and vary the on-time, a method called pulse-width modulation (PWM). It may also skip pulses or vary frequency at light load. The feedback loop is deliberately designed to respond quickly without becoming unstable; compensation components set how aggressively it reacts.
This is why a buck converter appears in system logs or hardware configuration as a regulator, rail controller, PWM source, switching frequency, or duty-cycle limit. A voltage reading that sags under load can indicate that the converter has reached its current limit, lost regulation, or entered a different operating mode.
Discontinuous conduction
At a light load, the inductor may start a cycle with so little current that its falling current reaches zero before the next on-time begins. This is discontinuous conduction mode (DCM): there is an interval during which the inductor current is zero.
The cycle then has three intervals:
- The switch is on and inductor current rises.
- The switch is off and inductor current falls to zero.
- Both normal inductor-current paths are off, so the output capacitor alone supplies the load until the next cycle.
DCM changes the voltage relationship. The simple Vout ≈ D × Vin rule assumes CCM and no losses; in DCM, the result also depends on inductance, switching frequency, input voltage, duty cycle, and load current. Controllers often enter DCM intentionally because switching less or transferring smaller energy packets improves light-load efficiency.
The boundary between CCM and DCM occurs when the inductor-current valley just reaches zero. Below that load, the converter is discontinuous; above it, the current remains positive throughout the cycle. Understanding that boundary explains why measured duty cycle, ripple, and efficiency can change noticeably as a system moves from normal load to standby.