Air core transformer results

Hi guys, I'm trying to transfer 12VDC of power by means of using air core transfomer concept taken from various sources, and my goal is to have regulated 9V at the end to power up my arduino nano based circuit (the schematic is attached).
If I remove my nano as the load, the voltage across C4 can reach 20V, however if I put the load I only get around 4.6V.
If I probe the FET Drain-Source it looks like the picture attached. I tried putting snubber across it in attempt to minimise the oscillation (values 2 Ohm, 2.2uF and around that) but it doesn't help.
To note also, my 555 tends to get hot and because of that I have damaged few 555s during my trial and errors.

So far I've tried:

  1. Lowering or increasing the frequency by changing the values of R1 and R2, but it doesn't help.
  2. Lowering or increasing the number of loops of the coils, but it doesn't help. 4.5V is the maximum I can get.

Limitations: I could only increase the input supply to max 15V and max 200mA for the timer555 to operate, based on its datasheet.

Perhaps anybody can give some ideas of what went wrong or any suggestions how to further increase the voltage? :confused:


Air core transformers are extremely inefficient and the coupling between the 2 coils has got to be very tight to get any kind of reasonable energy transfer.
Why are you doing this ?

4.5V is plenty to power up a bare bones Arduino.

Is this for fun, or are you hoping for a practical level of efficiency? If the latter, there are some well understood electrical engineering principles that need to be considered. Plenty of info on the web if you google "wireless power transmission pdf".

Thanks for the quick reply :slight_smile:

Well, I'm doing that because the load is not only bare bone nano but also to light up 24 LEDs through 3xshift registers from the nano + a nrF24L01 module which needs 3.3V supply + another simple 555 circuit to power up an IR led. It is a project of a propeller clock. The load circuit has to be in a spinning circuit and therefore I need wireless transfer from the base circuit. So far I've tested the load circuit offline with a battery/DC supply and unfortunately 9V is what seemed good and stable supply.

@mauried : you are right about the tight coupling between the coils. I modified the secondary coil to be as close as possible to the primary and that gives an increase of voltage. Unfortunately it is not that great, only about 0.5V. Still not enough :frowning:

Obviously you are going to have to learn more about how to design these things. The coils should form a resonant circuit at the driving frequency.

You haven't described the design of your transformer, which is pretty key to things. What inductance
are the windings - is it high enough for the switching frequency you use? Is it bifilar?

@MarkT: I have no LCR meter with me at the moment so I couldn't measure the inductance. I will measure it when I can. Please can you shed some light how to calculate the correct inductance with a pre-determine frequency?

About the coil, I used single 30AWG (0.25mm) insulated copper wire to make the windings. I made each 500 turns for both primary and secondary coil. Yesterday I tried to change the primary to 150 turns and keep the secondary at 500 turns, and also simplify few things on the load - now I could reach around 5V.

But still, I don't quite understand why I only get 5V from 12V (about 42%) with same amount of coil turns (the primary is coiled on a slightly smaller diameter shaft than the secondary so that it would sit inside the secondary's inner diameter hole, but I think that wouldn't make so much different as the total resistance for each sides are not much different).


You still haven't described the detailed design of your transformer. The turns count on its own is not useful information.

Size, geometry, bifilar or not? These details are crucial.

Normally an air-cored transformer would be toroidal and bifilar to have any chance of large mutual inductance.

Hi MarkT,

Here are some pictures to explain the coils.

The white one is primary. 30AWG copper wire is coiled on a 30mm diameter x 15mm shaft.

The yellow one is secondary. 30AWG copper wire is coiled on a 55mm diameter x 15mm shaft.

The picture number 3 is how it is going to be assembled with the PCB mounted with the primary coil (spinning).

Looks like half the flux, approx, from the outer coil doens't cut the inner coil, so you cannot have
an efficient transformer. The two coils need to share the same flux. This is why bifilar toroid is typically
best (although you have strong capacitive coupling from primary to secondary with that, which might not
be desired).

You topology can be improved if the coils are much closer to each other.

Transformers with a magnetic core are easy, the core does all the work and any coil geometry will
have pretty much 100% flux-sharing. Air-cored and you have to do that with coil geometry alone.

And use schottky diodes for the rectifier, less voltage loss there.

Coil inductance can be estimated from various approximations, and there are on line calculators, like this one:

The voltage would be increased for a resonant circuit.

The primary coil is already resonant by default, as the primary coil is in series with the parallel capacitance of the FET, which from the datasheet is 310 pf.
Thats whats causing all the ringing on the scope display.
I cant see how this arrangement can work, unless the coupling coefficient can be increased, and that means extremely close coil spacings.
As an example, Ive got an electric toothbrush that works on this principle, and the coil spacings are approx 2mm, is running at 56khz , and only achieves a coupling coefficient of around 0.2 , which is woeful, but works as its only charging a small AA battery.

If both primary and secondary are resonant I think you can get higher output voltages - in such situations
you are seeking to minimize losses to allow higher Q in the resonance. It is no longer a simple transformer
then, but a pair of coupled LC tank circuits. This is how power transfer can be practical with low coupling.

If you can wind a bifilar coil I think you can get good response without resonance, this is the principle of
the wideband balun transformer used at RF frequencies, for instance. But the impedance range is more

Of course if you have a bifilar transformer the primary and secondary are intimately physically connected,
which isn't useful for wireless charging for instance!

Using ferrite half-cores is one way to up your efficiency - in effect you slice a pot-core ferrite transformer
across the middle.
This sort of thing (look for Ferroxcube Pot Core): Ferrite: Pot Cores - Slugs & Miscellaneous
Just have primary in one half, secondary in the other half, bring together...

Or just buy a slip ring? There's some neat little ones available. Depending on the expected lifetime of the device, you may need to step up to something a little more expensive than the average hobby-grade components. If you only need it to run for a short time, just a pair of copper rings and hand-bend contacts will work.

Thanks for the responses guys!

Just to update, I found a reference (Ridley Engineering | - [011] Flyback Converter RCD Clamp Design) and after adding C and a diode, I managed to suppress the FET oscillation and the output voltage is now around 5.2V which is just nice to power up the load circuit. I can live with that for now but of course I would like to increase it for more reliable supply. I did not add any resistor in the RCD because by calculation, I found that I need really small R so I neglected it. With this addition it is able to suppress from >100V oscillation peak to 28V. However I still hear ringing sound from the FET :frowning:

I couldn't make the primary coil nearer to the secondary coil further, as I need to spin the primary and any mechanical friction is not desirable. And for the same reason I chose air core transformer concept, as using a mechanical contact such as slip ring would create another friction problem. The clock also intended to run continuously so I would like the parts to last long.

From the link from jremington, I calculate that the inductance for my primary (150turns now) = 2.6 nH and my secondary (500 turns) = 60 nH.
From the formula it shows also that if the wire diameter is reduced, I can get higher inductance. So maybe I can try to find 0.1mm copper wire and do another winding for the primary.

Any suggestions?

Try again, 2.6nH is the inductance of an inch of plain wire or something like that! You inductance will be in the uH or
mH range.

How much power are you feeding into the primary coil, ie whats the current thru the coil and fet , and how much resistance is there in the coil?
You should be able to work out the inductance of the primary coil by measuring the frequency of the ringing when the fet turns off.

How much power are you feeding into the primary coil, ie whats the current thru the coil and fet , and how much resistance is there in the coil?
You should be able to work out the inductance of the primary coil by measuring the frequency of the ringing when the fet turns off.

Care to explain how you can measure inductance from just the self-resonant frequency?

Its resonating because of the Drain Source capacitance of the Fet which is in series with the coil, and thats 310 pf from the data sheet for that fet.
The resonance occurs when the fet is turning off.
The resonant frequency F = 1/ 2 X PI X SQroot(L X C) so L can be obtained from F and C .