To size a high-side buck converter MOSFET for a 100W load stepping 48V down to 12V, you must convert output watts to RMS current: the MOSFET requires a minimum continuous drain current rating of 8.33A (to cover transient spikes and thermal derating) and an RDS(on) below 28mΩ to keep conduction losses under 0.5W. This assumes a 92% conversion efficiency, Continuous Conduction Mode (CCM), and a 500kHz switching frequency. If you are stepping down from a rectified AC mains source rather than a 48V DC battery, the required voltage blocking rating ($V_{DS}$) and switching loss profile shift drastically.
The Core Conversion: Output Watts to MOSFET RMS Current
Sizing a MOSFET for a DC-DC buck stage requires converting the DC output load into the pulsed RMS current the high-side switch actually experiences. The high-side MOSFET only conducts during the "on" time of the switching cycle, defined by the duty cycle ($D$).
Here is the step-by-step formula with values substituted for a 100W, 48V-to-12V buck converter:
- Calculate Output Current ($I_{OUT}$):
$I_{OUT} = P_{OUT} / V_{OUT} = 100W / 12V = 8.33A$ - Calculate Duty Cycle ($D$):
$D = V_{OUT} / V_{IN} = 12V / 48V = 0.25$ (The MOSFET is ON 25% of the time) - Convert to High-Side RMS Current ($I_{RMS}$):
$I_{RMS(HS)} = I_{OUT} \times \sqrt{D} = 8.33A \times \sqrt{0.25} = 4.16A$ - Calculate Maximum $R_{DS(on)}$ for a 0.5W Conduction Loss Target:
$P_{COND} = I_{RMS}^2 \times R_{DS(on)}$
$R_{DS(on)} = 0.5W / (4.16A)^2 = 0.028\Omega$ (28mΩ)
According to Texas Instruments' guidelines on MOSFET power losses, you should select a part with an $R_{DS(on)}$ at least 20% lower than your calculated maximum to account for the positive temperature coefficient of silicon as the die heats up. A part like the CSD18540Q5B (typical $R_{DS(on)}$ of 3.2mΩ) would easily handle this thermal load.
How Input Voltage Scaling Shifts the Specs (120V vs 230V vs 3-Phase)
A common point of confusion is applying AC mains metrics directly to a buck converter. A buck converter is strictly a DC-DC topology. However, if your buck converter is the second stage of an AC-DC power supply, the AC input voltage dictates the rectified DC bus voltage, which fundamentally shifts your MOSFET selection criteria.
- 120V AC Input: Rectifies to a ~170V DC bus. Your buck converter MOSFET must have a $V_{DS}$ rating of at least 250V. The duty cycle for a 12V output drops to ~7%, increasing switching losses.
- 230V AC Input: Rectifies to a ~325V DC bus. You must step up to a 400V or 500V Superjunction MOSFET (e.g., Infineon CoolMOS). At this voltage, switching losses ($P_{SW}$) begin to dominate conduction losses, requiring a faster gate driver.
- 3-Phase AC Input: Rectifies to a ~600V DC bus. Standard silicon MOSFETs suffer from massive switching penalties here. The design must shift to 800V/900V SiC (Silicon Carbide) or GaN (Gallium Nitride) devices to prevent thermal runaway, as the duty cycle shrinks below 2%.
Neighboring Values: ±20% Load Scaling Table
Designs rarely sit at exactly 100W. Below is a spec-sheet-table showing how the required MOSFET parameters shift across a ±20% load range for the same 48V-to-12V topology. This assumes the same 0.5W maximum conduction loss target.
| Output Power ($P_{OUT}$) | Output Current ($I_{OUT}$) | High-Side RMS Current ($I_{RMS}$) | Max $R_{DS(on)}$ (@ 25°C) | Recommended Die Size / Package |
|---|---|---|---|---|
| 80W | 6.67A | 3.33A | 45mΩ | 3x3mm SON / SOT-23 |
| 90W | 7.50A | 3.75A | 35mΩ | 5x6mm PowerPAK |
| 100W (Baseline) | 8.33A | 4.16A | 28mΩ | 5x6mm PowerPAK |
| 110W | 9.17A | 4.58A | 23mΩ | LFPAK56 / D2PAK |
| 120W | 10.00A | 5.00A | 20mΩ | LFPAK56 / D2PAK-7 |
When This Conversion is Meaningless
Converting output watts to RMS current and $R_{DS(on)}$ is only half the thermal equation. This conversion becomes meaningless under two specific conditions:
- When Switching Frequency ($f_{sw}$) and Gate Charge ($Q_g$) are Unknown: If you are designing for a high-frequency point-of-load regulator (e.g., 2MHz), switching losses ($P_{SW}$) will entirely dwarf conduction losses. Selecting a MOSFET purely based on a low $R_{DS(on)}$ calculation will lead you to choose a physically large die with high parasitic capacitance. The resulting gate drive losses and switching overlap losses will overheat the part. As noted in All About Circuits' MOSFET selection guides, you must optimize the Figure of Merit ($R_{DS(on)} \times Q_g$) rather than just $R_{DS(on)}$.
- When Applying Power Factor (PF): Power Factor is strictly an AC metric relating real power to apparent power. Because the input to the buck converter stage is a DC bus (maintained by bulk capacitors), PF is entirely irrelevant to the buck MOSFET's internal dissipation calculations. Do not attempt to derate MOSFET current based on AC power factor.
Frequently Asked Questions
Do I need a P-channel or N-channel MOSFET for a buck converter?
Almost all modern buck converters use N-channel MOSFETs for both the high-side and low-side switches. N-channel silicon offers significantly higher electron mobility than P-channel, resulting in lower $R_{DS(on)}$ and lower gate charge for the same die size. While a P-channel MOSFET avoids the need for a high-side bootstrap gate driver, the conduction and switching penalties make it unsuitable for anything above ~2A. For high-side N-channel driving, a dedicated bootstrap circuit or charge pump is mandatory.
How does inductor ripple current affect the MOSFET RMS calculation?
The formula $I_{RMS} = I_{OUT} \times \sqrt{D}$ assumes an ideal, infinitely large inductor with zero ripple. In reality, inductor ripple current ($\Delta I_L$) adds a triangular AC component to the RMS value. The corrected formula is $I_{RMS} = \sqrt{D \times (I_{OUT}^2 + \frac{\Delta I_L^2}{12})}$. If your ripple current is kept below 30% of the DC load current (the industry standard design rule), the ripple adds less than 1% to the total RMS current and can be safely ignored for initial MOSFET sizing.
Why do buck converter MOSFETs fail even when the datasheet current rating is much higher than my load?
Datasheet continuous drain current ($I_D$) ratings are often calculated assuming the case temperature ($T_C$) is held perfectly at 25°C—an impossibility on a real PCB. A MOSFET rated for "100A" might only handle 15A at a realistic 100°C board temperature. Furthermore, failures are usually caused by avalanche breakdown from parasitic inductance ringing during turn-off, or shoot-through caused by Miller capacitance ($C_{gd}$) coupling the switching node spike back into the gate. Always check the $V_{DS}$ spike margin and use a negative turn-off gate voltage or a strong pull-down resistor.
Can I use an IGBT instead of a MOSFET in a high-voltage buck converter?
Generally, no. While IGBTs handle high voltages (600V+) and high currents easily, they suffer from a fixed collector-emitter saturation voltage ($V_{CE(sat)}$) of roughly 1.5V to 2.5V. In a 12V output buck converter, a 2V drop across the switch represents massive conduction loss. IGBTs also have a "current tail" during turn-off that severely limits their maximum switching frequency to around 20kHz-50kHz. For high-voltage DC-DC buck stages, Superjunction MOSFETs or SiC devices are the correct choice due to their resistive conduction profile and fast switching capabilities.
