To convert a 12V nominal battery source (which sags to 9V and spikes to 14.4V) into a stable 12V @ 2A output, you need a buck and boost converter circuit operating in Continuous Conduction Mode (CCM). For this exact conversion, the duty cycle shifts from 0.25 (in boost mode at 9V input) to 0.83 (in buck mode at 14.4V input). This requires a 2.2µH shielded inductor rated for at least 3.5A saturation current. For loads under 2A, the concrete pick is the Texas Instruments TPS63020; for loads up to 5A, use the Analog Devices LT8705.
The Core Conversion Formulas (With Substituted Values)
A non-inverting buck-boost topology dynamically switches between buck and boost modes depending on the input-to-output voltage ratio. The math assumes a 90% efficiency baseline and CCM operation.
1. Boost Mode (when Vin < Vout)
When the battery sags to 9V, the circuit must boost to 12V. The duty cycle (D) formula is:
D = 1 - (V_in / V_out)
Substituted: D = 1 - (9V / 12V) = 1 - 0.75 = 0.25
2. Buck Mode (when Vin > Vout)
When the alternator spikes the battery to 14.4V, the circuit must buck down to 12V:
D = V_out / V_in
Substituted: D = 12V / 14.4V = 0.833
3. Inductor Sizing
To prevent inductor saturation and maintain CCM, the inductance (L) is calculated using the switching frequency (fsw) and target ripple current (ΔIL, typically 30% of Iout). For the TPS63020 (fsw = 2.4MHz) targeting 0.6A ripple:
L = (V_in × D) / (f_sw × ΔI_L)
Substituted (Worst case boost): L = (9V × 0.25) / (2,400,000 × 0.6A) = 1.56µH.
We round up to the nearest standard E12 value: 2.2µH (e.g., Würth Elektronik 74438336022).
Neighboring Values: ±20% Input/Output Variance
Real-world sources rarely sit exactly at nominal. Below is a spec-sheet-table showing how the duty cycle and minimum inductor requirements shift across a ±20% variance in a 12V nominal system (assuming 2A load, 1.5MHz switching frequency for a generic controller).
| Target Vout | Vin Range (±20%) | Operating Mode | Duty Cycle Range | Min Inductor (CCM) |
|---|---|---|---|---|
| 10.0V | 8.0V – 12.0V | Boost / Buck | 0.20 to 0.83 | 2.2µH |
| 12.0V | 9.6V – 14.4V | Boost / Buck | 0.20 to 0.80 | 2.7µH |
| 14.0V | 11.2V – 16.8V | Boost / Buck | 0.20 to 0.76 | 3.3µH |
What Fixes the Answer (and When the Math Breaks)
The calculations above are fixed by three core assumptions: Continuous Conduction Mode (CCM), 90% efficiency, and a fixed switching frequency. If any of these shift, the component values must change.
A buck-boost converter is strictly a DC-DC stage. However, if your upstream source is 120V AC, 230V AC, or 3-phase AC, the DC bus voltage after the rectifier shifts drastically (170V DC, 325V DC, and ~600V DC respectively). At 170V DC, a standard TPS63020 will instantly vaporize. You must shift to a high-voltage 4-switch buck-boost controller like the Analog Devices LT8705 (rated to 80V, requiring a front-end buck stage for 120VAC) or an isolated flyback topology for 230VAC/3-phase systems to handle the MOSFET VDS stress and meet safety isolation creepage requirements.
When the conversion is meaningless:
The CCM formulas become invalid in two specific scenarios. First, at light loads (typically <10% of max current), the inductor current hits zero before the switching cycle ends, entering Discontinuous Conduction Mode (DCM). In DCM, the duty cycle becomes load-dependent, and the simple Vin/Vout ratio no longer dictates D. Second, when Vin is exactly equal to Vout (the "deadband"), 4-switch topologies can experience control loop chatter, rapidly toggling between buck and boost states and destroying efficiency. Modern ICs solve this with a dedicated "buck-boost transition mode" that forces a specific duty cycle overlap.
Decision Tree: Picking Your Buck-Boost IC
Use this decision-tree-table to terminate your part selection process. Do not over-spec the IC; higher current controllers require larger, more expensive inductors and suffer from lower light-load efficiency.
| Condition (If...) | Then Select... | Concrete Part Number |
|---|---|---|
| Iout < 2A, Vin/Vout < 16V, space-constrained | Integrated single-inductor buck-boost | TI TPS63020 |
| Iout 2A–5A, Vin/Vout < 80V, high efficiency needed | 4-switch synchronous controller (external FETs) | ADI LT8705 |
| Iout > 5A, Vin up to 80V, high power density | Multi-phase or high-current 4-switch controller | TI LM5170-Q1 |
| Vin > 80V (e.g., 48V nominal with 100V transients) | High-voltage SEPIC or isolated topology | ADI LT3795 (SEPIC/Boost) |
FAQ: Buck and Boost Converter Circuit Edge Cases
Why not just use a SEPIC converter instead of a 4-switch buck-boost?
A SEPIC (Single-Ended Primary-Inductor Converter) is great for low-cost, low-current (<1A) applications because it uses a standard boost controller and a coupling capacitor. However, the coupling capacitor must handle the full RMS load current, leading to massive thermal issues at 2A+. For anything above 1A, a 4-switch non-inverting buck-boost is vastly superior in efficiency and thermal management.
How do I handle inductor saturation during startup?
During a dead-short or startup into a massive capacitive load, the inductor current can spike beyond the IC's cycle-by-cycle current limit. Always select an inductor with a saturation current (Isat) rating at least 20% higher than the IC's peak current limit. For the TPS63020 (peak limit ~3.8A), your inductor must have an Isat of at least 4.5A to prevent the core from saturating, dropping inductance to near-zero, and blowing the internal MOSFETs.
Does the PCB layout change between buck and boost modes?
The physical layout doesn't change, but the critical current loops do. In buck mode, the high di/dt loop flows from Vin through the high-side FET, inductor, and output cap. In boost mode, the high di/dt loop flows from the inductor, through the low-side FET, and back to the input cap. Your input and output ceramic bypass capacitors must be placed as close to the IC's PGND pins as physically possible to minimize parasitic inductance in both loops.
