MHz power electronics and their application to wireless power for drones

[Schottky]

[minion]

Very cool!

What is the approximate efficiency at 12cm?

[raptor1956]

I worked for a Japanese company, Murata Machinery, that provides automation equipment to the Semiconductor industry and much of the automation is powered by noncontact power delivery using approximately 10KHz induction.  The power flows through two parallel wires using Litz cable and the vehicles, using a couple core pickups receive that power and condition it to power the 100kg vehicles at speeds of 3+ meters/sec -- more than twice the walking speed of humans.  The system can power upwards of 15 vehicles per loop and a full sized FAB might have many dozens of such loops and hundreds of vehicles.  Another system I installed had a double loop power a crane of about 1000 pounds at similar speeds for a material storage system.

With what efficiency is the power transferred to the drone?


Brian

[Schottky]

It is a bit difficult to measure the power consumption of the quadrotor when it is hovering. For this specific quadrotor that I am using it consumes anywhere between 5-10W when it hovers with a shield on. The input DC voltage to the inverter was 77V and the input DC current was 0.2A when the quadrotor was not near the inverter (not within the magnetic field generated by the transmitting coil), the input current peaked at 0.33A when the quadrotor was brought closer and began to hover. A rough estimation of the efficiency DC-to-quad is between 20%-40%.
The majority of the losses are due to the ESR of the transmitting coil which was implemented using 1oz copper on a PCB. The thermal camera shows the transmitting coil heating up. I have measured the ESR of the coil which is about 0.3ohm, I know the current in it is about 8A, so the power lost in the coil is about 10W. So the efficiency of the power transmitted by the coil to the quad is well above 70%.
Using thicker copper will definitely decrease the losses in the coil and consequently increase the overall efficiency

This setup was not optimised to achieve high efficiency, it was just a quick demo that I did using boards that were designed for a different project. I guess what I wanted to show is that by going to MHz frequencies, the passives become much smaller and lighter. The receiving coil on the quadrotor was just a single turn of copper tape which had an inductance of 300nH. At kHz frequencies you will certainly need several turns which increases size and weight.

Sam

[T3sl4co1l]

  The coupling coefficient of those two coils is about 0.2, or better -- are they even trying?

12cm, that's great, sure, but can it do 13cm?

Tim

[raptor1956]

Sam, thanks for the feedback...

And to think Nikola Tesla was working on an intercontinental wireless power delivery systems nearly a century ago...

The other area of interest in long range wireless power transmission was/is the idea of putting solar collectors in Earth orbit and beaming the energy back to Earth.  Not sure I'd want to get between the sats and the ground station...

The wireless power system I describe for our automation systems have a clearance between the core and Litz cable of about 1cm as I recall.  The important point of this is you need highish frequency AC in order to speed the magnetic field and just as the inductors/transformers for power supplies can be made smaller if the frequency is higher the ability to transfer meaningful power is greater at higher frequency.  The power supplies we used we called SCPS's for "Supper Clean Power Supplies" -- a bit of marketing BS but also an important requirement when installing these systems in a $6B semi FAB.


Brian

[MagicSmoker]

...
The majority of the losses are due to the ESR of the transmitting coil which was implemented using 1oz copper on a PCB. The thermal camera shows the transmitting coil heating up. I have measured the ESR of the coil which is about 0.3ohm, I know the current in it is about 8A, so the power lost in the coil is about 10W. So the efficiency of the power transmitted by the coil to the quad is well above 70%.
Using thicker copper will definitely decrease the losses in the coil and consequently increase the overall efficiency.

Using thicker copper on the pcb won't help nearly as much as you think - the skin depth at 13.56MHz is about 18um so 1oz Cu (35um) is about as thick as you can go before the ratio of R_AC to R_DC is too high. The only practical solution is to make the traces wider (or lower the frequency, but 13.56MHz is a special frequency reserved for noisy/high power industrial processes). That said...

...
I guess what I wanted to show is that by going to MHz frequencies, the passives become much smaller and lighter.
...

But, as operating frequency goes up the transfer efficiency at a given distance goes down (or to put it another way, the distance allowed for a given throughput shrinks) unless you are trying to do truly resonant power transfer outside of the near-field boundary, but that requires sophisticated PLL-based means of ensuring the resonant frequency of the receiver LC network is precisely matched to that of the transmitter. At least so far as I understand how such is supposed to work, anyway.

[tatus1969]

impressive :-) But - do you have any sort of application for this in mind?

[Mechatrommer]

1) bee noise generator to the already quiet house
2) energy discharger
3) to prove that "i can"