Wireless power never works on its own inside a levitation product.
The transmit coil sits next to a permanent-magnet ring, a galvanized steel plate, the levitation electromagnet and position sensors. A result measured on an open bench can change completely once those parts are assembled. In this PU500G upgrade, our main job was not to raise the voltage. It was to find where the assembled product was wasting power and remove those losses one by one.
The new transmitter keeps the 12 V platform. At a 35 mm levitation gap, we measured 125 mW at the receiver with approximately 2 W input to the transmitter. At shorter distances, substantially more power is available.
The measured result
| Item | Result |
|---|---|
| Platform | Gleagle PU500G |
| Transmitter supply | 12 V |
| Maximum measured powered gap in this test series | 35 mm |
| Received power at 35 mm | 125 mW |
| Transmitter input at 35 mm | Approximately 2 W |
| Highest measured received power at close range | 315 mW at 15 mm |
| Transmit topology | Single-switch parallel quasi-resonant circuit |
| Tuning method | Frequency sweep in the assembled product, seeking a low-input operating point |
Distance and received power
| Levitation gap | Received power | Relative to 35 mm |
|---|---|---|
| 15 mm | 315 mW | 2.52× |
| 25 mm | 220 mW | 1.76× |
| 30 mm | 160 mW | 1.28× |
| 35 mm | 125 mW | 1.00× |
These numbers come from the assembled PU500G, not from two coils facing each other in open air. Floater weight, radial offset, receiver voltage and normal assembly variation all affect the result.
Two useful calculations
Moving from 35 mm to 15 mm gives 2.52 times the power
The gap falls to 15 ÷ 35 = 42.9% of its original value. Received power rises to 315 ÷ 125 = 2.52 times.
A logarithmic fit to these four measured points gives an effective distance exponent of approximately 1.06. That number describes this data set only; it is not a general wireless-power law. In particular, a large-coil near-field link should not automatically be treated as if it followed a far-field inverse-cube rule.
The 35 mm working point is about 6.25% efficient end to end
At 35 mm, the receiver gets 125 mW while the transmitter takes about 2 W. The system-level efficiency is therefore 0.125 ÷ 2 = 6.25%.
This includes the transmitter circuit, coils, magnet ring and metal structure, and the receiver rectifier. It is not the efficiency of the coil alone. The point of the 35 mm result is a more usable power-transfer distance, not a claim of maximum efficiency. When efficiency matters more, reducing the gap remains the most direct fix.
At a 3.3 V receiver rail, 125 mW corresponds to an ideal current of roughly 38 mA, enough for low-power control and lighting loads on the PU500G. Rectification and regulation losses still have to come out of that budget.
The main loss came from the steel plate, not the Y35 ring
In our earlier 188 kHz no-load comparison, the input power changed as follows:
| Test configuration | Input power | Increase over bare coil |
|---|---|---|
| Bare transmit coil | 0.384 W | — |
| Add Y35 magnet ring | 0.420 W | +0.036 W |
| Add galvanized steel base plate | 0.960 W | +0.540 W |
| Add center coil, core and PCB | 1.056 W | +0.096 W |
The galvanized plate added 0.540 W, about 80.4% of the total extra loss introduced by the surrounding structure. Y35 is a permanent-magnet ferrite with high electrical resistivity and a recoil permeability close to air; in this comparison, it was not the main problem.
That is why this iteration concentrated on the metal base plate rather than spending more time on the magnet ring.
An aluminum isolation structure reduces eddy-current loss
Aluminum has a relative magnetic permeability close to 1, so it does not materially redirect the static magnetic circuit needed for levitation. To the high-frequency alternating field, however, its eddy currents provide reflection and attenuation. The isolation ring reduces the alternating magnetic flux that would otherwise enter the steel plate directly.
We did not use a soft-magnetic ferrite shield here. Such a shield would also alter the static levitation magnetic circuit, which can easily solve one problem while creating another.
We kept the 12 V single-switch platform
Raising the voltage can increase the transmitted field, but it also increases eddy-current loss in the metal structure. With a galvanized plate in the system, part of the expected gain is simply consumed by that extra loss. Higher voltage also puts more pressure on the MOSFET, resonant capacitors and insulation design.
The upgraded system therefore retains the single-switch parallel quasi-resonant topology. We select the operating point by sweeping frequency in the assembled product:
- Install the actual transmit coil and resonant capacitor.
- Sweep frequency while recording input power and switching waveforms.
- Find the quasi-resonant region with relatively low input power.
- Trim slightly toward lower frequency, leaving commutation and thermal margin.
- Install the receiver and repeat the distance-versus-power measurement.
The resonance equation is useful for the first capacitor choice. It does not choose the final working frequency. The magnet ring, steel plate and receiver magnet all contribute to the equivalent load, so the resonance calculated for a bare coil cannot simply be copied into the product.
The receiver and Hall signal have to be checked in the real assembly
Nearby magnets and metal parts change receiver inductance, equivalent loss and resonance. We therefore matched the receiver in the actual PU500G assembly rather than tuning a separate coil on its own.
The Hall-signal path received the same attention: smaller signal-loop area, twisted wiring and low-pass filtering at the sensor. We check the levitation-position signal with wireless power both on and off, because induced high-frequency noise must not become a position error.
What this changes for PU500G designs
The 35 mm / 125 mW operating point gives the PU500G a wider mechanical design window. More material and clearance can sit between the transmitter and the floating receiver while retaining power for low-power loads.
The method is as important as the number. Find the losses in the real metal environment first, then adjust the coils, resonance and receiver match. In this kind of product, that is more useful than simply raising transmit voltage, and it makes temperature and reliability easier to control.
Our next steps are to reduce transmit-side metal loss further and improve receiver rectification and regulation. There is still room in both distance and efficiency, but any further claim has to come from measurements in the assembled product.
Engineering test notes from our PU500G development work. Values are measured results from the stated assembly and test conditions, not universal specifications for every receiver, floater or production configuration.
Related reading: Distance Scaling and Frequency Engineering in WPT · Compensation Topologies and the Efficiency Chain · PU500G Levitation Module.