question of frequency transfer of inductive link for wireless power transfer system

[bhl777]

Hi all, I am curious about the sine wave transfer characteristic of the inductive links for WPT systems. For example, if I have a near-field inductive links for biomedical applications:
(1) Some people said we can model the inductive links like an ideal transformer, so if I have a sine wave at the primary side, will I have the same frequency sine wave at the secondary side?
(2) Some papers were discussing about misalignment of inductive coils, which may impact the amplitude of the sine wave, but will it also impact the frequency?
(3) If misalignment will not impact frequency at the secondary side, is there anything else can impact it？
(4) If I do not need resonance at the secondary side and do not care about the efficiency, does that mean I can randomly pick the input frequency of the sine wave, to get what I want in the secondary side?
Thank you!

 

[chuckey]

The frequency is usually set by something other then the power transfer coil. This then isolates the frequency from any effects to the power transfer coil. Of course you could design a circuit that uses the power transfer coil as the frequency determining item, but the frequency will then shift all over the place.
If you go for a fixed frequency working, then resonance of the power transfer coil will increase the currents in it and hence its magnetic field and the received power.
Frank

 

[Warpspeed]

(1) Yes, what goes in comes out if its direct magnetic coupling. Definitely the same frequency and same sine waveform.

(2) As the coupling is purely magnetic, what amplitude comes out depends on both spacing and orientation. In fact you can twist the receiving coil about and very likely find an orientation where no energy at all couples, even at short range. The frequency stays the same, what goes in comes out, but how much recoverable energy comes out is the variable.

(3) Nothing in the transmission path can impact the received frequency, only the received amplitude.

(4) The problem with resonance in the secondary is that although it will build up energy in the winding and increase the circulating voltages and currents, you cannot draw that built up energy off in the form of power.
It works in a radio receiver because the magnified voltage created by resonance can be used by a sensitive amplifier to increase receiver sensitivity at the resonant frequency.

But where you are transferring true watts of power into a load, as soon as you try to extract that circulating power from the resonant circuit it damps down the resonance.
Tuned circuits are not over unity devices for power transfer.
What goes in comes out, and no more.

But resonance can have an advantages for increasing sensitivity where there is minimal loading on the tuned resonant circuit.
It would be rather different if you were just trying to transfer some data to a sensitive receiver. Coupling real power into a load is a different game.

 

[FvM]

My general suggestion is to model the link as non-ideal transformer, coupled inductors with coupling factor < 1. A realistic model would also include a series resistance of each coil, comprising DC resistance and proximity/skin effect losses. Eddy current losses generated by metallic parts in the vicinity and losses of a possibly used magnetic core could be modelled by an additional parallel resistor.

You should at least stay below 150 kHz not to collide with general EMC regulations. Permitted stray field strength of inductive power transmission system is still a matter of regulations under development. Several bands are occupied by existing services, e.g. still used DECCA navigation or RFID. Inductive power guys argue that the regulations are only applicable to information transmitting systems.

 

[chuckey]

FvM''s remarks have caused me to reconsider the problem. At one extreme, when the two windings are tightly coupled, then its acting more like a low frequency transformer, i.e. a few percent of the primary inductance as leakage inductance + DC and RF losses, and then the reflected load impedance. At the other end when the coils are far apart its acting likke an inductance + DC and RF losses. Now the DC and RF losses should be minimised, but you end up with a variable "leakage" inductance and a reflected load. Now the "leakage" inductance will present an impedance that will reduce the current and I think should be resonated out with a series capacitor.
Now if you resonate the coil by its self, you will be optimising the impedance for zero power transfer. Not very useful! The amount of coupling depends on the physical arrangement between the two coils, their sizes and their orientation. So I would build a mock up of the two coils and experiment, my feeling is that you will end up with a leakage of 30%, of Lcoil. The same holds for the secondary, it will be reflecting back its leakage inductance and load, to be transformed by the turns ratio. So resonating its leakage inductance will, for a given amount of voltage across the receiving coil allow more current to flow into the load.
Frank

 

[bhl777]

The frequency is usually set by something other then the power transfer coil. This then isolates the frequency from any effects to the power transfer coil. Of course you could design a circuit that uses the power transfer coil as the frequency determining item, but the frequency will then shift all over the place.
If you go for a fixed frequency working, then resonance of the power transfer coil will increase the currents in it and hence its magnetic field and the received power.
Frank

Hi Frank, would you explain "you could design a circuit that uses the power transfer coil as the frequency determining item, but the frequency will then shift all over the place"? For example, if I have a microcontroller, which can be used to set a frequency of sine wave, then if I need some specific frequency at the secondary side, then I can just set it in the primary side and let it transfer to the secondary side, is that true?

- - - Updated - - -

(1) Yes, what goes in comes out if its direct magnetic coupling. Definitely the same frequency and same sine waveform.

(2) As the coupling is purely magnetic, what amplitude comes out depends on both spacing and orientation. In fact you can twist the receiving coil about and very likely find an orientation where no energy at all couples, even at short range. The frequency stays the same, what goes in comes out, but how much recoverable energy comes out is the variable.

(3) Nothing in the transmission path can impact the received frequency, only the received amplitude.

(4) The problem with resonance in the secondary is that although it will build up energy in the winding and increase the circulating voltages and currents, you cannot draw that built up energy off in the form of power.
It works in a radio receiver because the magnified voltage created by resonance can be used by a sensitive amplifier to increase receiver sensitivity at the resonant frequency.

But where you are transferring true watts of power into a load, as soon as you try to extract that circulating power from the resonant circuit it damps down the resonance.
Tuned circuits are not over unity devices for power transfer.
What goes in comes out, and no more.

But resonance can have an advantages for increasing sensitivity where there is minimal loading on the tuned resonant circuit.
It would be rather different if you were just trying to transfer some data to a sensitive receiver. Coupling real power into a load is a different game.

Hi Warpspeed, according to your advise, can I say "in the form of power it is advised to use resonance, but in signal transfer level, it does not matter? " If I just need a specific frequency at the secondary side, I can simply send one signal at the primary side, just be careful the amplitude may get attenuated by some uncertainty?

- - - Updated - - -

My general suggestion is to model the link as non-ideal transformer, coupled inductors with coupling factor < 1. A realistic model would also include a series resistance of each coil, comprising DC resistance and proximity/skin effect losses. Eddy current losses generated by metallic parts in the vicinity and losses of a possibly used magnetic core could be modelled by an additional parallel resistor.

You should at least stay below 150 kHz not to collide with general EMC regulations. Permitted stray field strength of inductive power transmission system is still a matter of regulations under development. Several bands are occupied by existing services, e.g. still used DECCA navigation or RFID. Inductive power guys argue that the regulations are only applicable to information transmitting systems.

Hi FvM, thank you for your advise! I have two questions from your answers:

(1) you mentioned "at least stay below 150 kHz not to collide with general EMC regulations". For example, I have two sets of coils (power coils and signal coils, both sets have primary side and secondary side). For power coil, we will use fixed MHz level input signal at the primary side, and the resonance at the secondary side. For signal coil, should I use frequency below 150kHz in order not to collide with general EMC regulations?
(2) Will the effects within non-ideal transformer impact the frequency transfer in the signal coil? Would you advise some books/papers that I can refer to, to build a practical inductive link models in SPICE?

 

[chuckey]

If you use a microprocessor to set the frequency, thats the operating frequency, but you need some form of power amplification. In a minimal component RF charger, you need two transistors a coil or two and a few extra components and some cheapo power supply. I am not sure what a processor brings to the circuit.
Frank

 

[D.A.(Tony)Stewart]

But the coil will have interwinding capacitance and a Self Resonant frequency, so you will care and what frequency you choose and what losses you have.
Skin Effect losses increase with frequency. Ferrite core losses increase with frequency
But coupling losses reduce with rising frequency where the gap must be < diameter and gap losses reduce with gap size.
A Ferrite Bar is useful for increasing the coupling factor.

A step up transformer for sending and step-down transformer winding for receiving improves impedance from very low ESR battery<<1 Ohm loads to the optimum impedance for free-space in the 150 Ohm range.

Therefore plan on using step-up autotransformer windings with a flat ferrite bar just below SRF for optimal power transfer in the license free industrial noise band designated for your area. in the short wave band or below 150kHZ

there is a push for operation in the higher frequency ISM bands of 6.78 MHz and 13.56 MHz where resonant systems allow high spatial freedom at high efficiency. At these high frequencies, traditional MOSFET technology is approaching its capability limit.

Mismatched impedance transformer coupling results in significant reflected power.

For biological power transfer, there is a tradeoff between gap losses , high F absorption and low frequency coupling losses, where something in the 1GHz range is still ideal for charging internal biological battery systems.

 

[bhl777]

But the coil will have interwinding capacitance and a Self Resonant frequency, so you will care and what frequency you choose and what losses you have.
Skin Effect losses increase with frequency. Ferrite core losses increase with frequency
But coupling losses reduce with rising frequency where the gap must be < diameter and gap losses reduce with gap size.
A Ferrite Bar is useful for increasing the coupling factor.

A step up transformer for sending and step-down transformer winding for receiving improves impedance from very low ESR battery<<1 Ohm loads to the optimum impedance for free-space in the 150 Ohm range.

Therefore plan on using step-up autotransformer windings with a flat ferrite bar just below SRF for optimal power transfer in the license free industrial noise band designated for your area. in the short wave band or below 150kHZ

there is a push for operation in the higher frequency ISM bands of 6.78 MHz and 13.56 MHz where resonant systems allow high spatial freedom at high efficiency. At these high frequencies, traditional MOSFET technology is approaching its capability limit.

Mismatched impedance transformer coupling results in significant reflected power.

For biological power transfer, there is a tradeoff between gap losses , high F absorption and low frequency coupling losses, where something in the 1GHz range is still ideal for charging internal biological battery systems.

Hi SunnySkyguy, thank you so much for your advise. As a beginner I cannot fully understand all your input, so can I ask the question in a simpler format?
According to what I learned from all the replies in this thread, if we have a system like the picture below, L3 and L4 are signal coils. Then my questions are
1. will I use microcontroller to send a sine wave to L3, are I expecting the a sine wave at the same frequency at L4?
2. If I want to generate the sine wave with different frequencies across L4, is there any practical limitations? For example, I want to send 100kHz for one case, and send 200kHz for a different case, should I just simply change the input frequencies at L3 then expect the target frequency sine wave at L4?
3. If the power coils have fixed frequency, say, 13.56 MHz, in practical will it impact the received sine wave across L4? For example, can I use microcontroller to generate sine wave from 10kHz to 10MHz to send to L3, and get the sine wave with the same input frequency at L4?
Thank you!

 

[Warpspeed]

The problem can be split into two separate parts.

Firstly we need to generate sufficient power at the required frequency to drive the transmitting coil. Frank has summed up pretty well what is required in #7.

The transmitter coil should be carefully tuned to resonance, and correctly impedance matched to the power amplifier driving it.
That should all be fairly straightforward.

The next consideration is the size and geometry of the transmitter coil.
This depends on the distance and how accurate the placement is going to be.
If this is for a human biomedical implant, presumably the distance will be small and the placement accuracy fairly high.
If its to work with unrestrained animals or a wearable device by a human exercising, the situation could be highly variable.

In the first case, it may be preferable to focus the transmitted magnetic field with a flux concentrator such as a large ferrite U core. This would hugely increase the coupling at very close range, but greatly reduce the far field intensity.

For working at greater ranges the transmitter coil should be large and air cored.

An example might be the dinner plate sized search coils typically used with metal detectors. These can generate a surprisingly large far reaching magnetic field, although the coupling of significant actual useful power with something like that, will still be problematic.

If its to be used within a room, one solution might be to run a multi core cable around the room, and connect all the conductors in series. A 25 core cable then becomes a large area 25 turn coil. Flat ribbon cable may be of interest here too if its to go under a carpet.

Something like that might offer your best chance if the receiving coil is moving around unpredictably within an enclosed space.
A rule of thumb might be to make the transmitting coil diameter at least as large as the expected distance to the receiver coil, and preferably surrounding the receiver coil completely if you can manage to do that.

At the receive end, the problem is quite different. As we are trying to transfer power, not just detect a weak signal, the current induced in the receive coil will produce its own magnetic field which opposes the field that produced it.

What happens is the surrounding exciting field will try to flow AROUND the receiving coil. It will always take the easy path, rather than try to exert itself driving power into your load.

For that reason a flux concentrator at the receiver coil will be of very great benefit.
A ferrite rod is recommended, the physically larger the better.
This is common practice with transistor radios that work at lower frequencies.
The magnetic field will find a much easier path through the ferrite rod than through the surrounding air, even if the receiver coil is heavily loaded, which it will be.

Trying to tune a highly resonant peak in the receiver coil is not likely to happen because of the very heavy loading and resultant low Q.

But balancing out the inductive reactance with capacitive reactance will bring the voltage and current into phase which should give a broad peak in output power, which is a requirement for transferring maximum true power.

 

[bhl777]

The problem can be split into two separate parts.

Firstly we need to generate sufficient power at the required frequency to drive the transmitting coil. Frank has summed up pretty well what is required in #7.

The transmitter coil should be carefully tuned to resonance, and correctly impedance matched to the power amplifier driving it.
That should all be fairly straightforward.

The next consideration is the size and geometry of the transmitter coil.
This depends on the distance and how accurate the placement is going to be.
If this is for a human biomedical implant, presumably the distance will be small and the placement accuracy fairly high.
If its to work with unrestrained animals or a wearable device by a human exercising, the situation could be highly variable.

In the first case, it may be preferable to focus the transmitted magnetic field with a flux concentrator such as a large ferrite U core. This would hugely increase the coupling at very close range, but greatly reduce the far field intensity.

For working at greater ranges the transmitter coil should be large and air cored.

An example might be the dinner plate sized search coils typically used with metal detectors. These can generate a surprisingly large far reaching magnetic field, although the coupling of significant actual useful power with something like that, will still be problematic.

If its to be used within a room, one solution might be to run a multi core cable around the room, and connect all the conductors in series. A 25 core cable then becomes a large area 25 turn coil. Flat ribbon cable may be of interest here too if its to go under a carpet.

Something like that might offer your best chance if the receiving coil is moving around unpredictably within an enclosed space.
A rule of thumb might be to make the transmitting coil diameter at least as large as the expected distance to the receiver coil, and preferably surrounding the receiver coil completely if you can manage to do that.

At the receive end, the problem is quite different. As we are trying to transfer power, not just detect a weak signal, the current induced in the receive coil will produce its own magnetic field which opposes the field that produced it.

What happens is the surrounding exciting field will try to flow AROUND the receiving coil. It will always take the easy path, rather than try to exert itself driving power into your load.

For that reason a flux concentrator at the receiver coil will be of very great benefit.
A ferrite rod is recommended, the physically larger the better.
This is common practice with transistor radios that work at lower frequencies.
The magnetic field will find a much easier path through the ferrite rod than through the surrounding air, even if the receiver coil is heavily loaded, which it will be.

Trying to tune a highly resonant peak in the receiver coil is not likely to happen because of the very heavy loading and resultant low Q.

But balancing out the inductive reactance with capacitive reactance will bring the voltage and current into phase which should give a broad peak in output power, which is a requirement for transferring maximum true power.

Hi Warpspeed, thank you for pointing out the main concern in designing power coils. My questions now will focus on detecting the weak signal at the receiver side, assuming the power transferring is OK. Would you advise me the three questions in #9? Thank you!

 

[Warpspeed]

Its rather difficult unless you can give us some much better specifics about what you are actually trying to do.
Required amount of power to transfer, distances involved and the application and any limitations we can only guess about ?
Are we talking about hundreds of milliwatts, or microwatts of power transmission ?
Millimeters of distance through skin, tens of meters of range outdoors, or what ?

With data transmission it would be really helpful to know the required data rate is it a few hertz or megahertz of bandwidth ?

 

[bhl777]

Its rather difficult unless you can give us some much better specifics about what you are actually trying to do.
Required amount of power to transfer, distances involved and the application and any limitations we can only guess about ?
Are we talking about hundreds of milliwatts, or microwatts of power transmission ?
Millimeters of distance through skin, tens of meters of range outdoors, or what ?

With data transmission it would be really helpful to know the required data rate is it a few hertz or megahertz of bandwidth ?

Thank you Warpspeed! What I am trying to do is to use the detected signal to implement some internal signal processing. We can think the distance is very close to the skin, the circuit designed to process the detected sine signal would be in microwatts level.
For a simplest case, my internal circuit is a comparator (compare input sine wave with GND), the power supply of this comparator is from the rectified voltage from the power coil (say resonance frequency 13.86MHz).
So can I send a sine signal with what ever frequency I want in the primary side of the data coil, to get the same freuqncy sine wave at the secondary side data coil, then use comparator to process and generate a square wave at the same frequency?
Thank you!

 

[Warpspeed]

That helps a lot !!

Close to skin means very short range, so a flux concentrator at the transmitter coil will help a very great deal. A ferrite U core would be my first choice.

The magnetic field strength bridging the end poles of the U core could be made quite intense so both power and data transmission becomes very easy.

The receiver coil could be quite small and very few turns should be required.

Your best bet might be to generate a sinewave frequency that can shift back and forth between two very close frequencies.
This is called frequency shift keying and is a very effective way to transmit data.

That would provide continuous uninterrupted power transmission, and at the receiver the change in frequency would be very easy to detect and turn back into a digtal data stream.

It all then becomes fairly simple to implement with minimum parts and complication.

So can I send a sine signal with what ever frequency I want in the primary side of the data coil, to get the same freuqncy sine wave at the secondary side data coil, then use comparator to process and generate a square wave at the same frequency?

Yes exactly.
In fact you could use a simple FM receiver chip which would have all the necessary signal processing circuitry inside do what you require, and it would work very well over a wide variation in received signal strength.

 

[bhl777]

That helps a lot !!

Close to skin means very short range, so a flux concentrator at the transmitter coil will help a very great deal. A ferrite U core would be my first choice.

The magnetic field strength bridging the end poles of the U core could be made quite intense so both power and data transmissin becomes very easy.

The receiver coil could be quite small and very few turns should be required.

Your best bet might be to generate a sinewave frequency that can shift back and forth between two very close frequencies.
This is called frequency shift keying and is a very effective way to transmit data.

That would provide continuous uninterrupted power transmission, and at the receiver the change in frequency would be very easy to detect and turn back into a digtal data stream.

It all then becomes fairly simple to implement with minimum parts and complication.

Thank you so much Warpspeed! The last question from me is the range I can use. If I did exactly as your advise in power coils and data coils, can I choose the frequency of the input sine wave what ever I want? For example, can I use any frequency between 1kHz and 10MHz? Or there is some practical limitation, either from the 13.86MHz power line, or any other regulation? Thank you!

 

[Warpspeed]

At that close proximity pretty much any frequency within the range you suggest will transfer power, but at 1Khz the data transmission rate would need to be very slow, but that may not be a limitation for you ?

Also at the low frequency end the receive coil would likely need more turns, so its probably better to keep the frequency reasonably high.
Physical size of the receiver will need to be minimised, so it might be best to start with that, there will be far fewer constraints at the transmit end.

I know that there are programmable cardiac pacemakers in common use, but I have no idea what frequencies they operate on. Some research in that direction may be productive.

Interference should not be a problem for you.
Others here may be able to better comment on EMC issues and legality.
There are the ISM radio bands (industrial, scientific, medical) where its pretty much free go with regards building and using experimental radio equipment.

There, its every man for himself, and the authorities do not want to know about interference issues you may be causing, or be a victim of.