A tapped inductor buck converter steps down a high DC input voltage while allowing a longer switch on-time than a conventional buck at the same conversion ratio.
This article explains tapped inductor operation, voltage conversion, winding and semiconductor stresses, and practical selection checks for design, component and purchasing engineers. The discussion focuses on the non-isolated, series-aiding power-converter arrangement presented by Sam Ben-Yaakov. [1]
Key Takeaways
The topology exchanges some of the conventional buck’s short-duty-cycle and switch-current constraints for different magnetic, semiconductor and filtering requirements.
- A tapped inductor buck can produce the same low output voltage at a larger duty cycle than a conventional buck, provided the winding connection and turns ratio are chosen correctly.
- The turns ratio must be defined explicitly: the total series turns divided by the output-section turns is not the same as the ratio between the two winding sections.
- Lower switch-side current is accompanied by higher off-time current in the output winding section and rectifier. Lower total converter loss is not automatic.
- The basic tapped inductor buck is non-isolated. A low output voltage does not make an output supplied from a hazardous bus safe to touch.
- Leakage inductance, increased semiconductor voltage stress and output-capacitor ripple current must be evaluated with a realistic model and measurements.
- The retained 200 V simulation demonstrates the trade-offs, but its approximately 25 V output must be distinguished from the ideal conversion-law prediction.
Tapped inductor features and benefits
A tapped inductor replaces the single energy-storage winding of a conventional buck with two magnetically coupled sections connected in series. Moving the freewheeling connection to the tap changes the number of active turns between the switch on-time and off-time, modifying both the conversion ratio and the current waveform. [1][2]
This arrangement is useful when a conventional buck’s required on-time becomes inconvenient for the controller or when reducing current in the main switch offers a worthwhile system-level benefit. The advantage depends on the input range, output current, winding resistance, leakage inductance and semiconductor losses; it is not a general guarantee of higher efficiency or smaller magnetics.
The magnetic component can be a purpose-designed tapped winding or a suitable coupled inductor whose separate windings are connected externally. Catalog coupled inductors are available with equal and unequal turns ratios, but their current, insulation and coupling specifications must match the intended connection. [3]
Topology overview
The buck arrangement discussed here uses a high-side power switch, two series-aiding coupled winding sections, a freewheeling diode, an output capacitor and a load. Figure 1 preserves the original overview of tapped-inductor topologies; the equations below apply specifically to the buck connection. [1]

For consistent notation, N1 is the number of turns in the switch-side section and N2 is the number of turns in the output-side section. Their individual self-inductances are L1 and L2, measured with the other winding open. The total-to-output turns ratio N is defined in Eq. (1).
- During the switch on-time, current flows through both winding sections in series and reaches the output.
- During the switch off-time, the freewheeling diode conducts through the output-side section. The switch-side section carries no sustained conduction current after commutation.
- Winding polarity matters. Reversing a winding changes the magnetic connection and invalidates the series-aiding equations used here.
- Tapped windings can also be used in boost and buck-boost arrangements, but their conversion ratios and stresses require separate analysis.
Voltage transfer and duty cycle behavior
The ideal continuous-conduction-mode conversion ratio depends on both duty cycle and turns ratio. Continuous conduction mode, or CCM, means that the energy-storage current does not fall to zero before the next switching cycle; it does not mean that both winding sections conduct continuously. [1][2]
Let Vin and Vout be the DC input and output voltages in volts, D be the switch on-time divided by the switching period, and M be the dimensionless voltage ratio Vout/Vin. For ideal coupling, negligible losses and CCM, the off-time fraction is 1 − D and Eq. (2) applies. [1]
Eq. (2) follows from balancing the winding volt-seconds over one switching period. During the on-time, the input-to-output voltage difference appears across the total series turns; during the off-time, the output voltage acts across the output section with opposite polarity.
- In the limiting case N = 1, Eq. (2) reduces to M = D, the ideal conventional-buck relationship.
- For N > 1, a given duty cycle produces a lower output voltage than a conventional buck.
- For a specified low voltage ratio, a larger N permits a larger operating duty cycle, but also changes current and voltage stresses.
Rearranging Eq. (2) gives the required ideal duty cycle and, alternatively, the total-to-output turns ratio needed for a selected duty cycle in Eq. (3).
For the retained 200 V to 25 V example, M = 0.125. With N1/N2 = 5, Eq. (1) gives N = 6, and Eq. (3) gives D = 0.4615. At D = 0.5, the ideal prediction is instead 200/7 = 28.57 V. An ideal 25 V output at D = 0.5 would require N = 7, corresponding to N1/N2 = 6.
A conventional ideal buck requires D = 25/200 = 0.125. The presentation’s settings of approximately 0.5 for the tapped converter and 0.14 for the conventional buck are simulation operating points, not exact ideal-law values. The demonstrated output is approximately 25 V in both circuits. [1]
Do not extend Eq. (2) unchanged into discontinuous conduction mode, where a zero-current interval introduces additional dependence on load, inductance and switching frequency. Practical CCM predictions also require allowance for winding resistance, semiconductor voltage drops and commutation effects.
Current stress and energy flow
The tapped connection redistributes current rather than reducing current everywhere. The switch and switch-side section can carry less current than their conventional-buck counterparts, while the output section and rectifier carry higher off-time current. [1][2]
In the ideal transition, continuity of the main magnetic flux requires the output-section current immediately after turn-off to rise in proportion to the change in active turns. Let ion be the series-winding current immediately before turn-off and ioff the output-section current immediately afterwards. Neglecting leakage and finite commutation time, Eq. (4) applies.
Eq. (4) describes a transfer of current between winding paths, not an instantaneous jump in current through an isolated, uncoupled inductor. In hardware, leakage inductance, device capacitance and the clamp network shape the transition.
- In a conventional buck operating in CCM, the inductor carries the output current throughout the cycle, and the main switch carries that current during its on-time.
- In the tapped converter, both sections carry the lower series current during the on-time. Only the output section carries the larger freewheeling current during the off-time.
- The presentation shows approximately 12 A peak in the switch-side section, approximately 27 A in the conventional-buck inductor, and output-section peaks around 80 A. These are waveform readings from the simulation, not interchangeable average or RMS ratings. [1]
With N = 6, an ideal 12 A turn-off current corresponds to 72 A immediately after redistribution. The approximately 80 A displayed peak belongs to the non-ideal simulated waveform; it should not be treated as an exact verification of the ideal current-step equation.
Assess conduction loss from the RMS current in each actual path, switching loss from the device waveforms, and magnetic loss from the flux excursion and winding construction. Lower switch current alone does not establish lower total loss, because output-winding copper loss, rectifier loss, capacitor heating and clamp dissipation can increase.
Typical applications
This topology is a candidate for non-isolated power conversion where the required step-down ratio makes a conventional buck difficult to control or burdens its main switch with high current. Suitability depends on the complete voltage, thermal, filtering and safety requirements. [1][2]
- High-voltage DC buses supplying lower-voltage internal rails, including the presentation’s approximately 200 V to tens-of-volts example.
- Auxiliary or intermediate conversion stages where a longer switch on-time is useful and galvanic isolation is provided elsewhere or is not required.
- Specialized pulse-current circuits in which the winding-current transformation is deliberately incorporated into the load interface.
- Educational and experimental platforms for studying coupled magnetics, current redistribution, leakage energy and converter modeling.
High current in the output winding does not automatically appear as the same pulse at the load. In a regulated supply, the output capacitor absorbs much of the difference between winding current and load current. A pulsed-load application therefore needs its own timing, impedance and capacitor analysis.
The basic series-connected arrangement does not provide galvanic isolation between input and output. When supplied from a rectified mains or another hazardous bus, its output remains part of that hazardous circuit unless a separate, appropriately designed isolation barrier is present. It must not be substituted for an isolated flyback solely because the output voltage is low.
Technical highlights
A coupled magnetic model is necessary to predict current commutation and device stress. A single lumped inductance can reproduce some low-frequency behavior but cannot represent the leakage-energy paths that dominate switching transients. [1][3]
Magnetic structure and equivalent circuit
For the same approximately linear magnetic circuit, winding self-inductance scales with turns squared. Let AL be the inductance factor in henries per turn squared. Eq. (5) relates the section inductances to their turns and prevents confusion between turns ratio and inductance ratio.
For two coupled sections, let k be the dimensionless coupling coefficient and M12 the mutual inductance in henries. The series-aiding inductance Lseries is given by Eq. (6). This symbol M12 is distinct from the voltage ratio M used earlier.
For perfect coupling, Eq. (6) becomes the square of the sum of the square roots of the section inductances. The simple sum L1 + L2 is appropriate only when mutual coupling is negligible, not for the intended tightly coupled, series-aiding structure.
The presentation uses an equivalent circuit containing an ideal transformer, a magnetizing inductance referred to the output side, and leakage inductances. Its associated reference is the 1994 APEC paper on a SPICE-compatible model of tapped-inductor PWM converters. [1]
When implementing a model, preserve winding polarity, turns definitions and the side to which each inductance is referred. Avoid adding separate leakage elements to a coupled-inductor model that already includes the same leakage through its coupling coefficient.
Leakage inductance and snubber behavior
Leakage inductance stores energy that does not follow the ideal transformer coupling. At switch turn-off, that energy interacts with device capacitance and the commutation network, producing overshoot and ringing. Both the main switch and the rectifier circuit require voltage-stress checks. [1][2]
For an initial energy budget, let Lleak be the effective commutating leakage inductance referred to the current path being analyzed, Ioff the current in that inductance at turn-off, and fs the switching frequency in hertz. Eq. (7) gives leakage energy Eleak in joules and the corresponding power Pleak in watts if that energy is dissipated once per cycle.
Eq. (7) is a first-order budget, not a complete snubber-loss calculation. A practical network may also dissipate energy associated with capacitance, magnetizing current or other circuit paths. The effective leakage value and current must use the same reference side.
- An RC network primarily damps a resonance and modifies transition speed; it does not establish a precise clamp voltage.
- An RCD, avalanche or other clamp network limits a transient through a defined energy path. Its threshold, dynamic behavior, tolerances and repetitive dissipation must be evaluated.
- Energy-recovery or active-clamp arrangements can reduce dissipative loss but add circuit and control requirements.
The presentation uses a 1200 V switch and shows a peak near 700 V with its snubber. This is an example waveform, not a general derating rule. Raising a clamp threshold can reduce loss in some dissipative clamp arrangements, but the relationship is circuit-dependent and must not be generalized to every RC or RCD network. [1]
Example simulation parameters
The original comparison is retained as a time-domain simulation example. Table 1 separates the stated operating settings and approximate waveform readings from ideal calculations, so the example is useful without implying that its numerical values form a lossless design. [1]
| Parameter | Tapped inductor buck | Conventional buck or interpretation |
|---|---|---|
| Input voltage | 200 V DC | 200 V DC |
| Displayed output voltage | Approximately 25 V | Approximately 25 V |
| Switching frequency | 50 kHz | 50 kHz |
| Switching period | 20 µs | 20 µs |
| Demonstrated duty cycle | 0.5 | 0.14 |
| Demonstrated on-time | 10 µs | 2.8 µs |
| Section turns ratio | N1/N2 = 5 | Not applicable |
| Total-to-output turns ratio | N = 6 | Not applicable |
| Section self-inductances | L1 = 125 µH; L2 = 5 µH | The larger section is on the switch side |
| Coupling coefficient | k = 0.95 | A simulation assumption, not a universal target |
| Output capacitor | 100 µF | 100 µF |
| Approximate current readings | 12 A switch-side peak; 80 A output-section peak | Approximately 27 A inductor current |
| Switch voltage example | Approximately 700 V peak with a 1200 V device | Not a general device-selection margin |
| Ideal output at demonstrated duty cycle | 28.57 V from Eq. (2) | 28 V from the ideal buck law |
| Ideal duty cycle for 25 V output | 0.4615 from Eq. (3) | 0.125 |
The section values satisfy Eq. (5): 125/5 = 25, so the turns ratio is the square root of 25, or 5. With k = 0.95, Eq. (6) gives Lseries = 177.5 µH. At perfect coupling, the corresponding value is 180 µH, rather than the uncoupled sum of 130 µH.
The comparison keeps the coupled structure in its conventional-buck reference arrangement. It therefore should not be interpreted as a complete optimization study between separately designed converters with equal loss, cost, volume or ripple performance. In particular, the same nominal output capacitance does not impose the same capacitor RMS current or voltage ripple.
The difference between the ideal and simulated output values cannot be assigned to a single loss mechanism from rounded waveform readings alone. A reproducible design comparison must include the actual semiconductor models, winding resistances, clamp circuit, load and steady-state operating conditions.
Relationship to flyback converters
The tapped inductor buck shares energy-storage and leakage-inductance concepts with a flyback converter, but its winding connection changes both the conduction paths and the isolation boundary. The similarity is useful for modeling, not a reason to treat the two circuits as interchangeable. [1]
- In the basic flyback, the output rectifier is blocked during the main switch on-time. Energy is stored in the magnetizing inductance and transferred through the output winding during the off-time.
- In the tapped inductor buck, current already reaches the output during the on-time because the two winding sections are connected in series.
- During the off-time, magnetic energy supplies the output through the output winding section and freewheeling rectifier.
- In CCM, stored magnetic energy remains at the end of the off-time. Complete demagnetization is not required every cycle.
- A flyback can provide galvanic isolation when designed with an appropriate transformer and insulation system. The basic tapped inductor buck described here does not.
Both arrangements require attention to leakage energy, switch overshoot, rectifier stress and magnetic loss. Their controller models, capacitor currents and safety requirements must nevertheless be evaluated for the actual topology.
Design-in notes for engineers
Start from the required operating envelope and calculate stresses before selecting a magnetic component. Then refine the design using realistic device models, winding data, thermal constraints and measurements rather than selecting turns ratio from the voltage equation alone.
Selecting turns ratio and duty cycle range
Use Eq. (3) to calculate the duty cycle at minimum and maximum input voltage, then check the controller’s minimum on-time, maximum duty cycle, current-sense blanking and gate-drive requirements. Include output tolerance and semiconductor drops when refining the operating range.
A higher total-to-output ratio can ease a short-on-time constraint, but it also increases the ideal current redistribution factor. Evaluate the resulting output-section peak current, rectifier current and device voltage stress before accepting the ratio. [1][2]
Magnetic component and coupling
Specify the section turns ratio, self-inductances, mutual coupling and winding polarity explicitly. High coupling is generally useful in this topology for limiting unwanted leakage effects, but no single coupling coefficient guarantees acceptable commutation or efficiency. Coupling requirements remain application-dependent. [3]
- Check inductance under operating bias and temperature, not only the zero-bias catalog value.
- Evaluate saturation using the combined winding excitation and actual current waveforms. Two separate winding-current ratings do not fully describe the magnetic limit.
- Size the core and gap for the required energy and flux excursion, while checking core loss at the switching waveform.
- Use conductor geometry and winding arrangements that balance DCR, AC winding loss, leakage inductance, parasitic capacitance, insulation and manufacturability.
Interleaving or placing a low-turn, high-current section close to the other winding can improve coupling, but may increase interwinding capacitance or complicate insulation. Compare the complete trade-off rather than maximizing coupling in isolation.
Handling high pulse currents
The output-side section, rectifier, capacitor connections and PCB return path must withstand the high off-time current. Use peak current for instantaneous stress and RMS current for heating; use the actual waveform for frequency-dependent winding and capacitor losses.
The approximately 80 A peak shown in the example is a warning about local current stress, not a general output-current capability. Verify winding terminations, rectifier surge and repetitive ratings, conductor heating and current-probe bandwidth before drawing conclusions from a short pulse.
Snubber and switch selection
The switch can experience a blocking voltage above the input voltage even before leakage overshoot is added. For the ideal buck connection analyzed here, let VSW,off be the main-switch off-state voltage in volts. The nominal relationship in Eq. (8) follows from the reflected output voltage during freewheeling.
For 200 V input, 25 V output and N = 6, Eq. (8) gives 325 V before diode drop, leakage overshoot and other parasitic effects. This nominal value must not be confused with the presentation’s approximately 700 V clamped transient peak.
- Select switch voltage rating from the worst-case blocking voltage and measured or simulated overshoot, with the required project-specific margin.
- Check switch current, switching energy, safe operating area, gate-drive conditions and thermal impedance.
- Check rectifier reverse voltage, repetitive peak current, RMS heating and recovery or capacitive commutation behavior.
- Size clamp components for repetitive energy, pulse current, voltage tolerance and temperature, then verify them at operating extremes.
Keep the commutation and clamp loops compact. The switch technology used in the presentation is one implementation choice; the topology itself does not require that semiconductor technology. [1]
Output filtering and ripple
The output capacitor must absorb the difference between the winding-delivered current and load current. This waveform includes the current redistribution at switching transitions, so a conventional-buck triangular-ripple approximation is not generally sufficient.
Let idelivered(t) be the instantaneous current entering the output node from the power stage, iload(t) the load current, iC(t) the capacitor current, and Cout the effective output capacitance in farads. The ideal capacitive voltage change ΔvC between times t1 and t2 follows Eq. (9).
Eq. (9) covers charge-related ripple. Add ESR-related voltage steps and ESL-related high-frequency spikes separately, and verify the capacitor’s RMS ripple-current capability and temperature rise. Use effective capacitance at the operating voltage and temperature.
The presentation uses 100 µF in both converters and shows greater output ripple in the tapped version. More capacitance, lower ESR or a suitably designed additional filter may be needed, but the correct solution depends on whether charge ripple, resistive steps or switching spikes dominate. [1]
Control and stability considerations
Use a model derived for the tapped topology rather than copying a conventional-buck compensation network. Include the chosen current-sense location, output-capacitor characteristics and relevant parasitic effects. The presentation’s referenced SPICE modeling work addresses both large-signal and AC analysis. [1]
- Check loop stability and transient response across input voltage, load and temperature.
- Evaluate entry into discontinuous conduction, pulse skipping or other light-load controller modes.
- Verify startup, shutdown, overload, short-circuit and recovery behavior. A current limit sensing the switch-side path must be related to the larger current in the output section.
- Include any added output filter in the control analysis and choose the feedback-sensing point deliberately.
Design and selection checklist
Compare candidate parts using equivalent definitions and test conditions. A matching nominal inductance, turns ratio or footprint does not establish interchangeability.
- Magnetics: section turns ratio and polarity, self-inductance and tolerance, leakage measurement conditions, inductance under bias, DCR at temperature, winding RMS current, magnetic saturation and total loss.
- Semiconductors: worst-case blocking and reverse voltage, peak and RMS current, switching behavior, recovery effects, safe operating area and thermal limits.
- Capacitors and clamp parts: effective capacitance, ESR, ESL, RMS ripple current, repetitive pulse energy, voltage rating and temperature dependence.
- Catalog versus custom: start with catalog coupled inductors when their documented characteristics fit. Consider custom magnetics when ratio, insulation, leakage, winding current or mechanical constraints cannot be met adequately. [3]
- Qualification and sourcing: confirm the exact part number, required reliability evidence, traceability, soldering compatibility, environmental limits, change notification and production availability.
- Safety: define working voltage, transient stress, insulation system, creepage, clearance and required dielectric tests for the finished equipment. A winding withstand specification does not turn a series-connected converter into an isolated supply.
Simulation and measurement validation
Measure both winding currents, switch voltage, rectifier stress, output ripple and component temperatures. Use voltage and current probes rated for the circuit’s common-mode voltage, transient amplitude and bandwidth; do not connect an earth-referenced oscilloscope ground to an unisolated hazardous node.
Compare steady-state averages and RMS values over complete switching cycles after startup has settled. Record probe location and bandwidth, and distinguish switching spikes from lower-frequency ripple. Validate the design at maximum input, minimum input, load extremes, startup and relevant fault conditions before release.
Conclusion
A tapped inductor buck can achieve a high step-down ratio without the very short on-time required by a conventional buck. Its main benefit is a different allocation of duty cycle and current stress: lower switch-side current is exchanged for higher output-section pulse current, additional voltage stress and more demanding commutation and filtering.
Define the turns ratio explicitly, separate ideal calculations from non-ideal simulation results, and select the magnetic component from its actual winding waveforms. For production designs, combine realistic modeling with verification of leakage overshoot, capacitor RMS current, thermal performance, protection and the non-isolated safety boundary.
FAQ
These answers summarize the operating principles and selection questions most relevant to engineers and component buyers.
A tapped inductor buck converter is a non-isolated step-down converter that uses two coupled winding sections connected in series. Current reaches the output through both sections during switch on-time and through the output section during off-time. The changing active turns modify the voltage conversion ratio and redistribute current stress.
A tapped inductor can achieve a low output voltage with a longer switch on-time and lower switch-side current than a conventional buck. These benefits can help when minimum on-time or switch current is restrictive. They must be weighed against output-winding pulse current, semiconductor voltage stress, leakage effects and capacitor heating.
For the buck arrangement discussed here, the conversion equation uses total series turns divided by output-section turns. This equals one plus the switch-side section turns divided by output-section turns. The distinction matters because a section turns ratio of five gives a total-to-output ratio of six, not five.
The basic tapped inductor buck does not provide galvanic isolation because its winding sections are electrically connected. A flyback can provide isolation when its transformer and insulation system are designed accordingly. A low-voltage output from a tapped converter connected to a hazardous bus must not be assumed safe to touch.
Compare winding ratio and polarity, inductance under bias, leakage inductance, winding resistance, current limits, loss and temperature data. Also check insulation, dimensions, qualification and supply continuity for the exact part number. Catalog parts can be suitable, but demanding leakage, current or construction requirements may justify a custom magnetic component.
Leakage inductance produces switching overshoot that must be controlled by a suitable commutation and damping or clamp network. Meanwhile, current redistribution creates substantial output-capacitor ripple current. Check semiconductor voltage, clamp energy, effective capacitance, ESR and capacitor heating together, then validate the chosen components using actual waveforms and operating temperatures.
How to design and validate a tapped inductor buck converter
Use the following procedure to move from electrical requirements to a verified implementation. Treat the ideal equations as a starting point and confirm the final design with non-ideal simulation and appropriately rated measurements.
Design and validate the non-isolated tapped inductor buck using consistent turns definitions and realistic current, voltage and thermal checks.
- Define input, output and safety requirements
Record the input range, output voltage, load range, switching frequency, temperature limits and ripple target. Establish whether galvanic isolation is required before selecting this non-isolated topology.
- Choose a feasible duty-cycle range
Check controller minimum on-time, maximum duty cycle, gate drive and current sensing. Select an operating range that remains feasible across input voltage and load conditions.
- Calculate and define the turns ratio
Use the ideal CCM conversion equation to choose the total-to-output turns ratio. Convert it to the section turns ratio and document winding polarity and notation explicitly.
- Design the energy-storage component
Determine winding inductances, peak excitation and flux excursion. Check saturation, winding RMS current, copper loss, core loss, insulation and temperature using the actual operating waveforms.
- Model coupling and commutation
Include leakage inductance, device capacitance, winding resistance and realistic semiconductor models. Avoid double-counting leakage and inspect the current redistribution at switching transitions.
- Select semiconductors and the clamp network
Check main-switch and rectifier voltage and current stresses. Size the damping or clamp network for repetitive energy, tolerances and thermal limits, then verify startup and fault behavior.
- Dimension filtering and verify control
Calculate output-capacitor charge ripple and RMS current from the delivered-current waveform. Include ESR, ESL and effective capacitance, then check loop stability and transient response.
- Validate hardware and release criteria
Measure winding currents, device stress, output ripple and temperatures with appropriately rated probes. Confirm operating extremes, protection, qualification, production tolerances and component availability before release.
Further Reading
These related articles provide background on conventional converter design, magnetic components and inductance measurements.
- Buck Converter Design and Calculation: compare conventional-buck current paths, component selection and layout considerations.
- Flyback Converter Design and Calculation: review the related energy-storage mechanism and the different isolated output-current path.
- Inductor Basics: Inductors, Chokes and Selection: extend the discussion of bias-dependent inductance, parasitics, losses and component qualification.
- Saturation Current of Inductors and its Measurement: see the inductance-versus-DC-bias measurement concept used when interpreting magnetic current limits.
References
The numbered sources support the topology explanation, retained simulation example and component-selection principles. Undated web resources are identified rather than assigned an assumed publication year.
- Sam Ben-Yaakov, Power Electronics fundamentals: Tapped inductor converters, educational presentation, YouTube, 2026.
- Carl Nelson, AN-44: LT1074/LT1076 Design Manual, Linear Technology / Analog Devices, undated online edition, section on tapped-inductor buck conversion.
- Coilcraft, A Guide to Coupled Inductors, Coilcraft educational resource, undated online edition.
Latest technical review and revision: 19 November 2026. Winding notation and ideal calculations were corrected, and the guidance on current redistribution, leakage stress, filtering, selection and non-isolated safety was expanded.



















