Project Pharos - Wirelessly Powered LED Lamp.

OVERVIEW

This resonant wireless-power-transfer LED lamp was the final project for MIT's 6.131 Power Electronics Laboratory. It transfers approximately one watt across an air gap through high-frequency magnetic coupling, then powers LED filaments with no electrical connection between transmitter and receiver.

The project combines resonant power electronics, magnetics design, analog control, and power conversion in one physical system. An early objective was full magnetic levitation; the final implementation instead concentrates on a robust, characterized WPT chain that can support more advanced mechanical and control work later.

SYSTEM ARCHITECTURE

The transmitter is a MOSFET half-bridge inverter driving a resonant LC tank. A phase-locked loop continuously tracks the tank's resonant frequency. The matched receiver tank couples magnetically to the transmitter, and a Schottky rectifier plus discontinuous-conduction-mode boost converter condition the received power for the LED load.

There is no digital communication between TX and RX. The system instead depends on analog behavior, resonance, and the impedance reflected through the coupled tanks.

COIL DESIGN

The TX and RX coils were designed using published WPT coil-optimization guidance relating diameter, turn count, wire gauge, frequency, and attainable Q. The intent was to maximize coupling and Q while remaining mechanically compact and manufacturable.

Each coil uses a custom 3D-printed form with channels that set turn spacing and geometry. Wire is pressed directly into the channels, producing repeatable geometry and consistent electrical characteristics.

• Measured at 200 kHz: TX inductance ~17.4 μH; RX inductance ~17.3 μH.
• Series resistance: ~0.25 Ω per coil; measured Q: ~80–85.
• Coupling coefficient k: ~0.35–0.50, measured with both shorted-coil and driven-current methods as spacing changed.

RESONANT TANKS

Both transmitter and receiver tanks were designed to resonate in the 230–270 kHz range: high enough for useful coil Q and manageable component size, but low enough to keep switching losses and component availability practical.

With L approximately 17.4 μH, each tank uses 22 nF capacitance for a nominal resonance near 260 kHz. The exact natural frequency changes with coupling and load, making frequency tracking necessary.

TRANSMITTER DESIGN

The transmitter uses a MOSFET half-bridge inverter to excite the resonant tank. Although the inverter produces a square wave, the LC tank filters the drive into high-amplitude sinusoidal current.

Even with a 3–5 V supply, resonance provides substantial voltage gain when the system is unloaded. Once the receiver is coupled and power is drawn, the tank voltage naturally collapses to a safe operating range.

PLL CONTROL

The transmitter's PLL keeps the drive at resonance as coupling and load conditions change. Resonant-tank voltage is reduced and converted to a logic-level square wave with a voltage divider and comparator. The PLL compares that phase with the inverter switching signal and adjusts switching frequency toward zero phase error—the LC tank's resonant condition.

This provides continuous resonance tracking without explicit frequency measurement or digital control.

RECEIVER FRONT END

The receiver resonant tank mirrors the transmitter tuning. Its received AC is passed through a Schottky bridge rectifier and smoothing capacitor. At roughly 260 kHz, modest capacitance is sufficient to produce low ripple at the rectifier output.

LOAD MATCHING

Directly powering the LED string from the rectified receiver voltage was ineffective. The LEDs need roughly 10–11 V at 50–100 mA, presenting a high effective resistance and poor reflected impedance; the result was weak power transfer.

Adding a DCM boost converter on the RX side allowed the receiver to present a lower apparent input impedance to the resonant tank while producing the voltage required by the LEDs. That load-matching change was critical: it moved the coupled system nearer its useful operating point and dramatically improved delivered power.

MECHANICAL DESIGN

The coil forms, lamp structure, and enclosure have been fully designed in CAD and the major mechanical components have been 3D printed. The printed parts establish repeatable coil spacing, provide electrical isolation, and give the receiver a defined physical architecture rather than leaving alignment to the bench setup.

What is not complete yet is the final compact presentation. The current transmitter remains a breadboard prototype, and the printed components have not yet been integrated into the small, polished lamp assembly shown in the CAD. The next mechanical pass is therefore packaging: miniaturizing the transmitter onto perfboard, routing the remaining electronics cleanly, and assembling the printed parts into the finished object.

RESULTS

The completed prototype delivered approximately one watt, with LED output around 10.5 V at 50–100 mA. Efficiency ranged from roughly 25% to 60% with alignment and spacing. It reliably transfers power across several centimeters with no electrical contact.

Maximum performance occurred slightly off center, showing that impedance interaction in the coupled system—not geometric centering alone—determines the best operating point.

FUTURE WORK

The next steps are a constant-impedance RX front end for consistent dimming, closed-loop TX amplitude control, RX-to-TX communication, and a return to the original magnetic-levitation goal.

CONCLUSION

This project demonstrates a complete resonant wireless-power system integrating magnetics design, resonant power electronics, analog control, and impedance matching. PLL-based resonance tracking and the DCM boost converter for load matching proved critical to achieving useful transfer, making the lamp both a functional demonstration and a platform for future contactless-power and dynamic-control work.