Contactless and wire-less power coupling?

[treez]

We have “beacons” spread over the countryside which contain microprocessor boards.

These boards (in the beacons) use around 5V, 20mA. -But they have no power source of their own. They get power from a portable handheld “cube” which is fixed onto them when required….but there is no metal contactor….the contact is magnetic….a ferrite core half in the “cube” ends up near a “mating” ferrite core half in the beacon, and thus power can be transferred from “cube” to beacon. It basically forms like a transformer since both ferrite core halves, (in “cube” and in “beacon”) have a little PCB with integrated windings, which couple with the respective core half.

So I guess when cube is fixed to beacon, a transformer is formed, and power can be transferred from “cube” to “beacon”. The core halves don’t actually physically touch when mated…..rather there is a slight gap between them….the reason for this is purely because if they were physically “smashed” touching together then the ferrite might crack.

So what do you think of this “contactless” connection?
Great advantage is its waterproof. And no corrosion of any metal contactors because there are none.

We are thinking of the attached way of coupling power through this “transformer”. That is, using a high frequency (50kHz) sine wave generated in the cube and flowing in the cube "primary" winding. The "secondary" in the beacon will simply full wave rectify it. Do you believe this is the best ploy?

(LTspice sim and pdf schem attached...we will make the L's and C's resonant at 50kHz in the real thing. We will also possibly do pulse skipping so we only send as much power as we need and no more)

I am certain this is already a well-known technique, would you agree?

 

[barry]

This has been used in electric toothbrushes and phone chargers for years. Not going to win you a Nobel prize.

 

[KlausST]

Hi,

... and "wireless chargers" for cell phones.
Thus there are a lot of ICs, design notes and application notes..

Klaus

 

[betwixt]

It's well proven technology, the biggest problem is the time it takes to charge the battery which, depending on usage and battery capacity might be several hours.

Brian.

 

[treez]

Thanks, (might i add that this app isnt for battery charging though), its just to give power to the beacon for as long as possible, and as efficiently as possible.
So i was thinking to promote efficiency, we need to have the circuit operating at resonance, and use a sine wave in the wires. Would you agree?
I am thinking if every time the core halves close together, the gap is slightly different, then the resonance point will be different (due to slightly different inductance) and so this could be an efficiency bad point.

I appreciate the references above to "wireless" charging.....but wireless charging will be more inefficient than what we want to do (due to more straying of the magnetic field) ......we actually just want to create a transformer by closing the two core halves together, and thus making the circuit to power the beacon with. We do have a gap between the core halves, but thats just because it would crack the ferrite if we had them being slammed tightly together. We dont really want the gap.

So really, what we are doing is an efficient, magnetic switch, (with no metal contacts).

 

[KlausST]

Hi,

generally charging or not... it´s just wireless power transfer..

regarding efficiency:
Typical ferrite cores have low core volume and low flux area to reduce cost.
In your case
* a bigger core
* and an increased core area where both halves meet (increase area with a ferrite plate)
may improve efficiency.

gap:
For sure it should be thin for best coupling, but it should be soft for low mechanical stress.
--> I wonder if there is some "rubber-ferrite-composite".

Klaus

 

[treez]

Thanks, also, we were wondering whether to implement the "converter" as a coupled_SEPIC , or like an LLC resonant converter, or some other resonant converter. Efficiency is important as the cube power is from a battery.

 

[betwixt]

Consider something much simpler - a self oscillating push-pull LC circuit. If you make one side a fixed frequency you have to tune the other side to match for best efficiency. If you let the coupling coils (primary-secondary) interact they will find their own resonance.

I'm still puzzled why someone has to take manual action to place the 'cube' to operate the switch. If this device doesn't need charging by this method, why not use a mechanical switch?

Brian.

 

[wwfeldman]

Great advantage is its waterproof. And no corrosion of any metal contactors because there are none.

waterproof connectors cost a little more than non-waterproof connectors
you can add a cover that connects the same way and protects the contacts
since a person has to be there to put the source in place, and remove it,
why not just use a regular connector?
you can certainly find waterproof covers for the beacon side when not powered

what is the cost difference between a waterproof connector and your magnetic coupling gadget?

 

[treez]

***Explanation of why this magnetic method is being used***
..i know what you mean....but in fact, another different device uses this method and its more suited there....but they want to re-use it here because its kind of "on trial" and we want to see how it performs.
The actual "other" device that uses this method involves the contacts being pushed together, and often the mating pieces are not true due to wear and tear...and the whole product gets thrown around.......so all the problems of contact wear etc are prevalent, and this magnetic coupling method is used to circumvent all the problems......they have already tried it with contacts and its not workable...keeps getting damaged and worn out....so the magnetic method is coming to the fore.

..but i am a little surprised that this technique is not already "oven ready" with chipsets off-the-shelf.

Betwixt your "self-oscillating" method which automatically finds the resonant frequency sounds of great interest. I assume there is some kind of iterative feedback loop which varies the frequency bit by bit till max output is gotten?

Then again i am wondering about "coupled_SEPIC", because that can tolerate a wide range of leakage inductance.

 

[betwixt]

I assume there is some kind of iterative feedback loop which varies the frequency bit by bit till max output is gotten?

FAR simpler than that. The tuning is obviously set to be fairly close to optimum but there is nothing to control it except the coupling to the secondary. You can do it with two transistors (one if you are careful!). Just use a center tapped primary as the collector load and cross couple the collector to the other base, rather like a multivibrator. The frequency will be determined by the inductance of the primary and the value of cross-coupling capacitors but it will 'pull' somewhat according to the load (secondary) characteristics. It isn't perhaps the most accurate method but it costs virtually nothing and can't be electrically much simpler.

Brian.

 

[Easy peasy]

most of these transformers with gap - have a resonant series cap on each side - driven by a (usually ) sine wave source and set up for good resonance matching when the transmitter in is the right place - you can in fact transmit up to 50W pretty easily with this method for air gaps up to 5mm - typ freq is 250kHz

 

[treez]

Thanks,

The resonant cap sounds good for maximising power throughput. I must confess in this case, the power transfer needed is just for a secondary side micro consuming say 1-10mA from 3v3. Sorry i didnt spec it out fully in top post.

The simple Royer type oscillator also sounds good, but I confess to wanting something to which I can vary power throughout …and this can be done here by varying the switching frequency to the LLC stage. Sorry i didnt spec it out fully in top post.

Hence I have come up with the attached LLC converter based solution. (pdf schem and LTspice sim attached)

The LLC transformer will use an ETD29 with N87 ferrite. When the core halves are “closed together”, there will be an overall gap of some 0.5mm or so…but I expect with wear and tear etc this wont always be the case, so some tolerance there. We will wind 32 turns directly on the primary side core half, and 82 turns on the secondary core half (no former used). We will then fix them each in place, so that they can be closed together in the product.

The idea will be to iteratively adjust the secondary voltage by adjusting the frequency of the fets on the primary side. Since vin is only 5V, ive noticed that it wont even matter if we temporarily accidentally “go the other (LHS) side of the gain peak” of the LLC converter….since reverse recovery isn’t an issue with 5V.

So what we will do, is use the LED transmitter to iteratively tweak the switching frequency until the overhead voltage on the secondary 3v3 regulator is at a minimum…..that way we will use as little power as possible, and the primary side battery will last as long as possible.

..There may be two leds….if one is ON, then it means the output voltage is too high…..if the other one is on, then it means the output voltage is too low….and subsequently the primary side micro can suitably tweak the switching frequency.

I think this is a good way, since it means we don’t have to worry about getting the gap exactly right…and can have a cheaper “core assembly” with a wider tolerance…..and just adjust Fsw to suit.

I must admit i cant help thinking there's already an offtheshelf module that does this kind of thing?