Transformer-Based Power-Line Harvester Magnetic Design

Power-line energy harvesting using current transformers offers an elegant way to power distributed electronics directly from existing AC conductors. This tutorial by Sam Ben‑Yaakov focuses on the magnetic design aspects of such harvesters, showing how core material, geometry, turns ratio and loading determine the maximum usable power and how designers can use simulation to optimize their design.

Key features and benefits

• Use of a current transformer architecture (single‑turn primary, multi‑turn secondary) for both current sensing and power harvesting from AC lines.
• Clear separation between the measurement function (current transformer as sensor) and harvesting function (rectified and conditioned output to a load).
• Emphasis on core saturation as the dominant limiting factor for harvested power in transformer‑based systems.
• Use of nonlinear inductor modeling in LTspice to represent real B–H behavior and saturation of ferromagnetic cores.
• Demonstration of how core cross‑section, path length and relative permeability influence magnetizing inductance and maximum voltage.
• Identification of an optimal reflected load resistance RPR_PRP that maximizes harvested power for given line current and core.
• Illustration of operation in linear, near‑saturation and deep saturation regimes, including typical power levels in each case.
• Discussion of rectification options (diode bridge vs active rectifier) and their impact on flux waveforms and saturation.
• Practical guidance on using power‑factor‑correction or DC‑DC stages to emulate an adjustable input resistance and implement maximum power point tracking (MPPT).
• Design‑level guidance suitable for engineers specifying cores, turns ratios and loads in energy‑harvesting current transformers.

Aspect	Benefit for designers
Current transformer usage	Non-intrusive harvesting from existing AC lines via clamp-on cores
Focus on magnetic design	Clear understanding of saturation limits and power capability
Nonlinear LTspice modeling	Realistic prediction of core behavior and flux waveforms
Core parameter tuning	Ability to trade off size, material and power level
Load optimization	Identification of an optimal reflected resistance for max power
Rectifier topology choice	Control over efficiency and saturation behavior
MPPT/control integration	Stable operation over varying line currents and loads

Typical applications

Although the video does not enumerate application markets, the presented approach is directly relevant to:

• Clamp‑on sensor nodes on LV or MV AC distribution lines.
• Powering condition‑monitoring or metering electronics from the current in busbars, feeders or industrial mains.
• Retrofitted sensing on existing wiring where adding a dedicated low‑voltage supply is impractical.
• Non‑intrusive power for data loggers or communication modules mounted along overhead or buried cables.
• Long‑term autonomous sensor operation where battery replacement is costly and grid access is nearby.

In all of these use cases, the magnetic element is a key passive component, and its proper selection and modeling are crucial for reliable long‑term operation.

Technical highlights

From current transformer to harvester

The starting point is the well‑known current transformer. A single‑turn primary is formed by the AC line passing through a toroidal or split core, and the secondary has nn turns feeding a burden resistor RLR_L. The secondary current is approximately IP/nI_P / n, with IPI_P the primary (line) current, and the voltage across RLR_L is Vout=(IP/n)⋅RLV_{\text{out}} = (I_P / n) \cdot R_L. This configuration is used to measure AC currents, with appropriate selection of core, turns and burden for the frequency range of interest.

To turn this into a harvester, the burden resistor is replaced by a rectifier and typically a DC‑DC converter that stabilizes the DC voltage and optimizes power extraction. The secondary then feeds a real load, such as a sensor system or battery charger. The tutorial explicitly focuses on the magnetic part of this structure, not on detailed converter electronics.

Core saturation and voltage as the limiting factor

The core material has a B–H characteristic with saturation: as the magnetic field HH increases, the magnetic flux density BB eventually reaches a maximum BmaxB_{\text{max}} beyond which further increases in HH produce little change in BB. This is true for all practical ferromagnetic cores; only air cores are exempt, but their coupling is too weak for this application.

A central point of the tutorial is that it is the applied voltage, not directly the current, that drives the core into saturation. Under normal transformer operation, the primary and secondary currents are arranged so that their magnetizing effects largely cancel, according to Lenz’s law. Even with high load current, the net magnetizing current can remain small. The flux in the core is related to the applied voltage and frequency via Faraday’s law:v(t)=nAdB(t)dtv(t) = n A \frac{dB(t)}{dt}where nnn is the number of turns, AA the core cross‑section and B(t)B(t) the flux density. Integrating over time shows that for a given BmaxB_{\text{max}}, there is an upper limit to the volt‑seconds that may be applied each half‑cycle. Once this limit is exceeded, the core saturates; at that point the voltage collapses and current maintains the saturated state.

For harvesting, this means that the maximum available power is fundamentally limited by the maximum allowable voltage at a given frequency and core geometry.

Core/design parameter	Change direction	Typical effect on magnetizing inductance and power capability
Relative permeability μr\mu_r	Ferrite → nanocrystalline	Higher μr\mu_r: higher inductance, higher possible harvesting voltage
Saturation flux BmaxB_{\text{max}}	0.3 T → 1 T	Higher BmaxB_{\text{max}} more volt‑seconds before saturation
Cross‑section area AA	Small → large	Larger AA: lower flux density for same voltage, higher power headroom
Magnetic path length ll	Long → short	Shorter ll: higher inductance for same material and cross‑section
Turns ratio nn	1:100 → 1:200	Higher nn: higher secondary voltage, lower current
Line frequency ff	50 Hz → 60 Hz	Higher fff higher reactance, more voltage for same current

Magnetizing inductance and core parameters

The transformer is modeled as an ideal transformer with a magnetizing inductance in parallel on the primary side. The magnetizing inductance LL is determined by:

• Core cross‑section area: larger AA yields higher LL.
• Magnetic path length: shorter path ll yields higher LLL
• Relative permeability: higher μr\mu_r yields higher LL.

A higher inductance increases the reactance XL=ωLX_L = \omega L at the line frequency, which in turn allows higher voltage for a given current. The power delivered to the load (reflected to the primary as RP=RL/n2R_P = R_L / n^2) is:P=V2RPP = \frac{V^2}{R_P}so higher voltage directly increases power for a given RPR_P.

The tutorial highlights that for high‑power harvesting, materials with high relative permeability and high BmaxB_{\text{max}}, such as suitable nanocrystalline alloys, are beneficial. For lower power levels, standard ferrite cores can be sufficient.

Equivalent circuit for harvesting

On the primary side, the line current is treated as a current source, since the harvester extracts only a small portion of the energy and the additional voltage drop is small. The equivalent circuit comprises:

• An AC current source representing the line current.
• A magnetizing inductance representing the core.
• A reflected load resistance RP=RL/n2R_P = R_L / n^2 in parallel with the magnetizing inductance.

The primary voltage is then:vprimary(t)=iline(t)⋅Zeqv_{\text{primary}}(t) = i_{\text{line}}(t) \cdot Z_{\text{eq}}where ZeqZ_{\text{eq}} is the parallel combination of the magnetizing reactance and RPR_P. This voltage produces flux in the core according to Faraday’s law. The power delivered to the load depends on how the current source splits between the magnetizing branch and the reflected load branch.

The tutorial points out the “play” between RPR_P and the magnetizing impedance:

• If RPR_P is very small, most current flows through RPR_PRP but the voltage is low, so the power is low.
• If RPR_P is very large (ignoring saturation for the moment), most current flows through the inductance; the voltage can be high but the current through RPR_P is small, so again power is low.

Therefore, an optimum RPR_P exists where the product V2/RPV^2 / R_P is maximized without driving excessive saturation.

Nonlinear inductor modeling in LTspice

To analyze this behavior in detail, the tutorial uses a nonlinear inductor model in LTspice. The flux is defined as a function of current using a hyperbolic tangent, which mimics the saturating behavior of ferromagnetic cores:

• The tanh⁡\tanhtanh function is symmetric and has a linear region around zero, followed by a gradual saturation.
• Two constants K1K_1 and K2K_2 are used such that their ratio K1/K2K_1 / K_2 equals the inductance in the linear region.
• In the example, the resulting inductance is 3 μH3\,\mu\text{H} for the single‑turn primary.

The flux density is then obtained by integrating the voltage across the inductor and scaling by turn count and core area:B(t)=1nA∫v(t) dtB(t) = \frac{1}{n A} \int v(t)\,dtAssuming a cross‑section of 1 cm21\,\text{cm}^2 (i.e. 10−4 m210^{-4}\,\text{m}^2) gives B(t)=104∫v(t) dtB(t) = 10^4 \int v(t)\,dt in Tesla. This integration is implemented as a behavioral voltage source using LTspice’s integration operator, allowing B(t)B(t) to be plotted during simulation.

A current source of 100 A at 50 Hz is used to drive the inductor as a stand‑alone element. Simulation shows that at 100 A peak, the core saturates, producing a distorted voltage waveform and a flux density waveform limited to about 1 T. Plotting BB versus current yields the B–I curve, showing saturation at around 50–60 A for this particular model.

Transformer model and resistive load

Building on the nonlinear inductor, the tutorial constructs a full transformer harvester model:

• The magnetizing inductance is the nonlinear inductor.
• An ideal transformer is implemented with two large inductances, 1 H for the primary and 40 kH for the secondary, giving an effective turns ratio of 1:200 since inductance scales as n2n^2 (2002=40 000200^2 = 40\,000).
• A resistive load is connected across the secondary; its value is parameterized (e.g. 30 Ω, 10 Ω, higher values).
• A behavioral expression multiplies load current and voltage to obtain instantaneous power, whose average is the harvested power.

The line current is modeled as a 10 A or 50 A AC source.

Operating points and power levels

Several operating points are demonstrated:

• 10 A line current, 30 Ω secondary load
  The core operates in the linear region, with flux density around 200 mT. Waveforms are sinusoidal. The primary current is slightly higher than the magnetizing current; the 90° phase shift between them means the vector sum equals the 10 A line current. The instantaneous power peaks around 50 mW, with an average of about 25 mW.
• 50 A line current, 30 Ω secondary load
  The core operates close to saturation, with flux density reaching about 1 T. Some waveform distortion appears. The peak instantaneous power is about 1 W, and the average harvested power is about 0.5 W. This is a useful power level for many sensor applications where a few hundred milliwatts suffice.
• 50 A line current, 10 Ω secondary load
  The core returns to deep linear operation, with flux density around 40 mT. The voltage across the load is lower, and the average power is about 300 mW, less than at the 30 Ω load, despite remaining comfortably away from saturation.
• 50 A line current, high secondary resistance
  With much higher load resistance, the system becomes highly saturated. The voltage attempts to rise but is limited by BmaxB_{\text{max}}Bmax. Most current flows through the nonlinear magnetizing branch; current through the high resistance is small. The power is again low, on the order of a few hundred milliwatts or less, and the conduction interval is narrow.

These examples illustrate that:

• There is an optimal load resistance for maximum harvested power.
• Operating near, but not deeply into, saturation can be beneficial for maximizing power.
• Both too low and too high load resistances reduce power, by different mechanisms.

Rectifier‑based harvester and flux waveforms

The tutorial then moves from pure AC resistive loading to a more realistic rectified harvester:

• The secondary is connected to a diode bridge feeding a DC load modeled as a constant voltage source (e.g. 2 V load plus diode drops for about 3 V total).
• The voltage at the secondary becomes a square wave corresponding to the rectifier conduction, and so does the primary voltage.
• With a square‑wave voltage, flux density becomes a triangular waveform: it ramps linearly up and down between minimum and maximum values each half‑cycle.

In the case of a 3 V total DC load, the flux density remains around 300 mT, indicating safe operation within the linear region. The current splits between the nonlinear magnetizing branch and the ideal transformer branch, and useful power is delivered to the DC load.

Raising the DC load voltage (for example, to 4 V at the load, about 5 V at the bridge output) increases the square‑wave amplitude and pushes the core into saturation. The flux density reaches 1 T and holds there for intervals of the cycle, with additional oscillations attributed to numerical artifacts. Physically, this represents the transformer entering a regime where the secondary effectively opens when the voltage collapses, so the secondary cannot conduct, and the magnetizing current dominates.

This is a problematic state for hardware: it can cause noise, long‑term core magnetization and uncontrolled behavior.

Protection by shorting the secondary

To mitigate saturation problems in practical designs, the tutorial suggests shorting the secondary during saturation intervals. If an active rectifier is used (e.g. a MOSFET bridge), the control circuitry can:

• Detect the onset of saturation (via flux‑related signals or other indicators).
• Turn on a pair of MOSFETs (either both low‑side or both high‑side devices) to short the secondary.
• Release the short once the core has returned to a safe operating region.

Shorting the secondary during saturation prevents the core from remaining magnetized, reduces noise and protects the magnetic element without harming the transformer. It is presented as an important system‑level design measure for high‑power harvesters.

Maximum power point and emulated load resistance

Because harvested power is highly sensitive to the effective load resistance, the tutorial notes that maximum power point tracking is desirable in real applications. An active power factor correction front end or a DC‑DC converter can emulate a tunable input resistance:

• By adjusting duty cycle or control setpoints, the converter changes the effective input resistance seen by the transformer secondary.
• A control algorithm can scan or track the load resistance that maximizes measured output power.
• Once the optimal reflected RPR_PRP is found, the system can maintain operation at this point as line conditions vary.

The stored energy typically feeds a charging circuit for a battery or supercapacitor that powers the actual load (sensor, radio, etc.). In simpler applications, designers may omit the intermediate converter and connect a rectifier directly to the line, but the video stresses that in such cases the operating point is fixed and must be carefully chosen with respect to saturation and power needs.

Design‑in notes for engineers

For passive component and magnetics engineers specifying cores and windings for transformer‑based line harvesters, the tutorial suggests several practical guidelines:

• Core material selection
  • Prefer high‑permeability materials with high BmaxB_{\text{max}} when higher harvested power is required. Nanocrystalline cores are highlighted as suitable candidates for such applications, combining high relative permeability with high saturation flux density.
  • For lower power requirements, ferrite cores remain viable, provided designers respect the lower BmaxB_{\text{max}} and frequency constraints.
• Core geometry
  • Increase cross‑section area AA to raise magnetizing inductance and reduce flux density for a given voltage.
  • Minimize magnetic path length lll within mechanical and insulation constraints to further increase inductance.
• Turns ratio
  • Use the relationship L∝n2L \propto n^2 to relate inductance to turns.
  • For current transformer‑like harvesters, a single‑turn primary (the conductor) and large‑turn secondary (e.g. 200 turns) is typical; in the example, a 1:200 ratio is realized by a 1 H primary and 40 kH secondary in the ideal transformer model.
• Saturation margin
  • Derive a target BmaxB_{\text{max}} from the material’s datasheet.
  • Use simulation (with nonlinear inductance) to ensure that, at the highest anticipated line current and operating load, the flux density waveform remains within acceptable limits most of the time. Near‑saturation operation can be acceptable for maximizing power, but prolonged deep saturation should be avoided.
• Load resistance selection
  • Recognize that the effective RPR_P seen by the primary critically affects harvested power.
  • For fixed‑resistor designs, choose RLR_L (and thus RPR_P) based on simulation or calculation to approximate the maximum‑power operating point for the expected line current range.
  • For converter‑based designs, incorporate a control function that can adjust the effective input resistance dynamically.
• Rectifier and control topology
  • Use diode bridges for simplicity at the cost of conduction losses and potential exacerbation of saturation.
  • For higher power and better control, employ MOSFET‑based active rectifiers, which reduce losses and offer the ability to short the secondary during saturation.
  • Consider integrating MPPT logic into the DC‑DC stage or PFC front end to maintain optimum power extraction over varying conditions.
• Simulation methodology
  • Start with a nonlinear inductor model calibrated to the intended core and winding (using material data and desired inductance).
  • Implement an ideal transformer with realistic turns ratio and large inductances to focus on magnetizing behavior.
  • Include realistic line current levels and sweep both current and load resistance.
  • Plot not only voltages and currents but also flux density and instantaneous power to visualize saturation and harvested power.

These design‑in considerations connect directly to the passive components that readers of a passive‑components‑focused portal select and specify: core material, core shape, winding strategy and burden / load components.

Line current (A)	Secondary load (Ω)	Flux region (qualitative)	Approx. average harvested power	Notes
10	30	Linear (~200 mT)	Tens of milliwatts	Clean sinusoidal waveforms, ample saturation margin
50	30	Near saturation (~1 T)	Roughly half a watt	High power, moderate distortion, practical target
50	10	Deep linear (~40 mT)	Few hundred milliwatts	Lower voltage, reduced power compared to optimum
50	High (>>30 Ω)	Deep saturation (~1 T)	Reduced, few hundred milliwatts	Narrow conduction, noisy behavior, not recommended

Source

The information in this article is based on a technical tutorial video by Sam Ben‑Yaakov, focusing on the fundamentals of transformer‑based power‑line harvesting and the magnetic design considerations behind it. Exact numeric limits and material properties should be verified against the manufacturer datasheet of the specific core and components selected in a design.

Frequently Asked Questions: Transformer-Based Power-Line Harvesting

References

1. The fundamentals of transformer-based power-line harvesting: A tutorial, Sam Ben-Yaakov, YouTube. https://www.youtube.com/watch?v=PjWEvqRt4sg
2. Simulating a physical inductor using LTspice, Sam Ben-Yaakov, YouTube. https://www.youtube.com/watch?v=rWPFbiu5o2s