To step down a 12V DC input to a 5V DC output using a buck converter, the required duty cycle is 41.67%. The fundamental Continuous Conduction Mode (CCM) formula is D = Vout / Vin. Substituting our exact query values: D = 5V / 12V = 0.4167 (or 41.67%). This means the internal MOSFET must remain ON for 41.67% of every switching cycle to maintain a steady 5V at the output inductor.
Below is the immediate conversion table for a 12V input system, covering the ±20% neighborhood around the standard 5V logic rail. This is critical when trimming feedback resistors for non-standard voltages (e.g., 4.2V for Li-ion charging or 5.5V for specific sensor arrays).
| Target Vout (DC) | Ideal Duty Cycle (D) | Real-World Duty Cycle (with 0.5V Diode Drop) | Common Application |
|---|---|---|---|
| 4.0V | 33.33% | 36.36% | Li-ion single-cell charge limit |
| 4.5V | 37.50% | 40.54% | USB PD variable supply |
| 5.0V | 41.67% | 44.71% | Standard 5V logic / Arduino VCC |
| 5.5V | 45.83% | 48.85% | High-tolerance servo power |
| 6.0V | 50.00% | 52.94% | 6V lead-acid battery charging |
Assumptions That Fix the Conversion (and Real-World Losses)
The simple D = Vout / Vin formula relies on two massive assumptions: Continuous Conduction Mode (CCM) and ideal, lossless components. In reality, every component in the power path drops voltage, forcing the controller to increase the duty cycle to compensate.
When selecting a high-current converter, buck regulator ICs with integrated FETs (like the Texas Instruments LM2596 or TPS5430) introduce specific parasitic losses. The MOSFET has an on-resistance (RDS(on)) causing a voltage drop (VDS), and the catch diode has a forward voltage drop (Vf). The corrected, real-world duty cycle formula is:
D = (Vout + Vf) / (Vin - VDS + Vf)
Worked Example: If your 12V-to-5V buck converter uses a Schottky diode with a Vf of 0.5V, and the internal MOSFET drops 0.2V (VDS) under your specific load current, the math shifts: D = (5 + 0.5) / (12 - 0.2 + 0.5) = 5.5 / 12.3 = 44.71%. The controller must push the duty cycle up by over 3% just to overcome the diode and switch losses. If you are using a synchronous buck converter (which replaces the diode with a second low-side MOSFET), Vf is replaced by the low-side VDS, bringing the real-world duty cycle much closer to the ideal 41.67%.
How the Conversion Shifts for 120V, 230V, and 3-Phase AC
A frequent and dangerous point of confusion on the workbench is attempting to apply DC-DC buck formulas directly to AC mains. A buck converter cannot natively process 120V AC, 230V AC, or 3-phase AC. It is strictly a DC-DC topology. If your source is AC mains, you must first rectify and filter it to create a high-voltage DC bus before any buck conversion can occur.
Here is how the duty cycle shifts when stepping down from rectified AC mains DC buses to a standard 24V DC industrial rail:
| AC Source | Rectified DC Bus (Nominal) | Required Duty Cycle for 24V Out | Topology Verdict |
|---|---|---|---|
| 120V AC (Single Phase) | ~170V DC | 14.1% | Borderline. High step-down ratio causes severe switch-node ringing. Consider a 2-stage buck or flyback. |
| 230V AC (Single Phase) | ~325V DC | 7.3% | Impractical for standard buck. Minimum on-time limits will cause pulse-skipping. Use isolated flyback. |
| 480V AC (3-Phase) | ~678V DC | 3.5% | Meaningless for buck. Requires isolated LLC resonant or forward converter with high-voltage SiC/GaN FETs. |
As the input DC bus voltage climbs from rectified AC, the required duty cycle plummets. At 7.3% duty cycle (from 230V AC to 24V DC), a 500kHz switching frequency yields an ON-time of just 146 nanoseconds. Most standard buck controllers cannot physically switch that fast, rendering the basic buck conversion formula practically useless without a high-frequency GaN driver or an isolated topology.
When Buck Converter Voltage Conversion Becomes Meaningless
The D = Vout / Vin formula breaks down entirely in three specific edge cases. If your design falls into these categories, the calculated duty cycle will not match reality:
When the load current drops below the inductor's ripple current threshold, the inductor current falls to zero before the switching cycle ends. In DCM, the output voltage is no longer strictly tied to the duty cycle and input voltage. The conversion ratio becomes dependent on the load current, inductance value, and switching frequency. Modern ICs mitigate this by entering "pulse-skipping" or "burst mode," but the basic formula is mathematically invalid here.
Every buck controller has a maximum duty cycle limit, typically between 90% and 99%. For example, the classic LM2596 maxes out around 95%. If you are powering a 12V relay from a 12V car battery, and the battery sags to 12.2V under cranking, the required duty cycle is 98.3%. The controller will hit its 95% limit, and the output will drop to ~11.6V (minus switch losses). The formula says 100%, but the hardware physically cannot deliver it.
Conversely, if you step down 48V to 3.3V, the ideal duty cycle is 6.8%. At a 2MHz switching frequency (period = 500ns), the required ON-time is 34ns. If your chosen IC (like the LMR33630) has a minimum on-time (ton(min)) of 50ns, it physically cannot output a 34ns pulse. The controller will skip pulses, resulting in high output voltage ripple and potential acoustic noise from the inductor. Always verify ton(min) against your calculated ON-time before finalizing part selection.
Frequently Asked Questions
Can I use a buck converter to step up voltage if I invert the duty cycle?
No. A buck converter topology is strictly unidirectional for step-down conversion. The inductor and diode/FET arrangement physically prevents the output from exceeding the input. If you need step-up capability, you must use a boost converter or a buck-boost (SEPIC/Cuk) topology.
Why does my measured duty cycle on the oscilloscope differ from the calculated value?
Oscilloscope measurements at the switch node (SW pin) will show the real-world duty cycle, which includes the compensation for parasitic losses (diode Vf, inductor DCR, and MOSFET RDS(on)). Additionally, if your scope ground lead is too long, switch-node ringing can distort the PWM waveform edges, making the measured ON-time appear longer than it actually is. Always use a coaxial pigtail or a short ground spring for high-frequency buck measurements.
