If you need 5V at 3A (15W output) from a 12V battery using a standard synchronous buck converter operating at 90% efficiency, the converter will draw 1.39A from the 12V source. The governing power conversion formula is I_in = (V_out × I_out) / (V_in × η). Substituting the exact values: I_in = (5V × 3A) / (12V × 0.90) = 15W / 10.8W = 1.388A. This direct answer assumes steady-state continuous conduction mode (CCM) and ignores transient startup spikes. Understanding how a DC-DC converter works requires moving beyond ideal theory and calculating the real-world thermal and topological losses that dictate your actual input current.
The Core Conversion Math and Neighboring Values
A DC-DC switching regulator does not create power; it transforms voltage and current while conserving energy, minus the losses dissipated as heat. The efficiency factor (η) is the critical multiplier. A cheap $2 LM2596 clone module might claim 3A output but will thermally throttle and drop efficiency to 75% under load. A genuine Texas Instruments TPS5430 or MPS MP2315 synchronous buck will reliably hit that 90% to 95% efficiency mark.
Below is a reference table showing how the required input current shifts across a ±20% range of output currents for a 12V-to-5V step-down (buck) conversion. This is critical for sizing your upstream wiring and fuses.
| Output Current (A) | Output Power (W) | Input Current @ 92% Eff (A) | Input Current @ 80% Eff (A) | Power Lost as Heat (W) |
|---|---|---|---|---|
| 2.4A (-20%) | 12.0W | 1.09A | 1.25A | 1.04W / 3.00W |
| 2.7A (-10%) | 13.5W | 1.22A | 1.41A | 1.17W / 3.38W |
| 3.0A (Nominal) | 15.0W | 1.36A | 1.56A | 1.30W / 3.75W |
| 3.3A (+10%) | 16.5W | 1.50A | 1.72A | 1.43W / 4.13W |
| 3.6A (+20%) | 18.0W | 1.63A | 1.88A | 1.57W / 4.50W |
Note: The 'Power Lost as Heat' column shows the thermal penalty of using an inefficient non-synchronous converter (80%) versus a modern synchronous design (92%). At 3.6A, that 2.93W difference is enough to require a dedicated heatsink on a TO-263 package.
What Assumptions Fix the Answer (and When It Shifts)
The 1.39A calculation is locked in by three assumptions: Continuous Conduction Mode (CCM), a fixed DC input voltage, and a known steady-state efficiency. If any of these shift, the math changes.
Topology Shifts (Buck vs. Boost):
If you reverse the scenario and use a boost converter (like the MT3608) to step up a 3.7V Li-ion cell to 12V at 1A (12W output), the input current math shifts dramatically. Assuming 85% efficiency, I_in = 12W / (3.7V × 0.85) = 3.81A. The input current is now higher than the output current, and your battery's internal resistance (ESR) will cause significant voltage sag if it isn't rated for 4A+ continuous discharge.
Upstream AC Shifts (120V vs 230V vs 3-Phase):
While DC-DC converters only see DC, the source of that DC bus shifts based on your AC mains. If your DC bus is derived from a standard 120V AC line (rectified to ~160V DC), a 400W isolated DC-DC stage stepping down to 48V will draw roughly 2.7A on the primary side. If you switch to a 230V AC input (rectified to ~320V DC) or a 3-phase system (yielding a ~600V DC bus), the primary-side RMS current drops proportionally, but the secondary-side DC-DC output math remains governed by the exact same P_out / η rule. The front-end Power Factor Correction (PFC) stage handles the AC-DC variance; the downstream DC-DC converter only cares about the resulting DC bus voltage.
When the Conversion Math Becomes Meaningless
There are three specific bench scenarios where plugging numbers into I_in = P_out / (V_in × η) will give you dangerously wrong expectations:
- Light-Load Pulse Skipping (PFM): At very light loads (e.g., drawing 10mA from a 3A-rated converter), the controller switches from Pulse Width Modulation (PWM) to Pulse Frequency Modulation (PFM) to save switching losses. Efficiency doesn't stay at 90%; it becomes highly non-linear, and quiescent current (
I_q) dominates the input draw. - Thermal Derating and Inductor Saturation: If you use an inductor with a saturation current (
I_sat) lower than your peak ripple current, the inductance collapses. The converter will spike in current, trigger over-current protection (OCP), or melt the MOSFETs. The math assumes the magnetic components are properly sized for the peak current, not just the RMS average. - Transient Response Sags: The formula calculates steady-state averages. If a microcontroller wakes from sleep and instantly demands 2A, the output capacitors must supply that current before the inductor current can ramp up (
V = L × di/dt). If the capacitor ESR is too high, the voltage will dip below the brownout threshold regardless of the input power available.
FAQ: Long-Tail Questions on DC-DC Operation
How does a DC-DC converter work without losing power as heat like a linear regulator?
A linear regulator (like the LM7805) acts as a variable resistor, burning off excess voltage as heat. Dropping 12V to 5V at 1A with a linear regulator wastes 7W ((12V - 5V) × 1A). A DC-DC switching converter uses a MOSFET to rapidly chop the input voltage into a high-frequency square wave, passes it through an inductor (which stores energy in a magnetic field), and smooths it with a capacitor. Because the MOSFET is either fully ON (low resistance) or fully OFF (zero current), it dissipates minimal power, achieving 85-95% efficiency.
How does a DC-DC converter work at high frequencies to shrink the inductor?
The required inductance is inversely proportional to the switching frequency. The fundamental inductor equation is V = L × (di/dt). If you increase the switching frequency (decreasing the time dt for each cycle), you can use a proportionally smaller inductance L to maintain the same current ripple di. This is why modern converters switching at 1.5 MHz to 3 MHz can use tiny, surface-mount shielded ferrite inductors, whereas older 50 kHz designs required large, heavy toroidal cores. The trade-off is that higher frequencies increase switching losses and electromagnetic interference (EMI).
How does a DC-DC converter work when the input voltage drops below the output?
A standard buck converter cannot output a voltage higher than its input. If your input sags below the target output, the converter will enter 100% duty cycle (pass-through mode), and the output will drop. To maintain a fixed output when the input crosses above and below the target (e.g., a 12V nominal battery sagging to 10V while you need a steady 12V output), you must use a SEPIC (Single-Ended Primary-Inductor Converter) or a four-switch buck-boost topology. These circuits dynamically reconfigure their switching sequences to either step-down or step-up the voltage seamlessly, though they require more complex magnetics and suffer slightly lower efficiency (typically 85-88%) due to the extra switching nodes.
