Talk:Wireless power transfer

Power penetration of different materials

Wireless power transfer can penetrate some materials better than others. The extent to which each type of power transmission can penetrate common materials (eg glass, normal plastics, wood, copper, aluminium, iron, steel, brick, unreinforced concrete) would be a useful addition to this article (ideally in the table that contrasts the different mechanisms of transfer). If anyone knows this information, please add it. Thanks. FreeFlow99 (talk) 14:35, 29 November 2018 (UTC)

This is a fairly complicated subject. A certain amount of the electromagnetic energy reflects from the surface, which varies with angle of incidence and polarization, then some of the transmitted energy is absorbed passing through the material, which depends on thickness, then additional energy reflects from the exit surface. The waves reflected from the front and rear surface of the material superpose, so the reflected energy depends sinusoidally on thickness. When the path length difference equals an even multiple of a half-wavelength you get constructive interference and a strong reflected ray; when it is an odd multiple of a half-wavelength you get destructive interference and weak reflection and more transmission. All of these attenuation coefficients vary with frequency. I'm not sure a single absorption number representing an "average" over a wide frequency band, polarization states and material parameters would tell much. Also, most wireless power transmission is through air. --Chetvorno^TALK 18:46, 22 December 2018 (UTC)

Evanescent wave or classical near field coupling between dipoles ?

Martin Soljacick had suggested that in the witricity patented device their might be some king of evanescent wave coupling. However evanescent coupling only arise at the interface between two mediums. The coupling coefficient between near-by coils as well as for nearby electrical dipoles can be computed in a classical way (through a coupling matrix) it leads to fields decreasing as 1/r^2, 1/r^3, 1/r^n. Besides the amplifying effect of resonances is also well known and described for a wide range of domains (magnetic and electric but also acoustical, mechanical). It is linked to an amplification of the field amplitude near resonance (that can be seen as a cumulative recycling of non-transferred energy in the generator and load side) and has nothing to do with the field geometry, no one as ever observed that the field distribution is affected by frequency as falsely suggested in the corresponding drawing (or you should prove it with actual data). According to my expertise, the Witricity system uses two core-less transformers on both sides of the link for the impedance adaptation function and a classical near-field coupling with a wider distance (improved by the resonance effect on both sides). Following Okkam razor principle there is no need to introduce some kind of evanescent concept that is moreover in contradiction with any accurate measurements on coupled coils(resonant or not), except if you want to patent a century old idea.

I suggest that the dogmatic chapter "inductive resonant coupling" should be removed, that a correct figure showing actual field-line should be used for magnetic coupling and that the resonance effect (increasing fields level without altering the field lines) should be introduced separately in a way to cover in a balanced manner both electric and magnetic near-field coupling.--Henri BONDAR (talk) 11:34, 4 January 2019 (UTC)

I wrote that section and drew the diagram. I appreciate the informed criticism. I basically agree that the diagram should be either replaced or supplemented with a circuit diagram of a more typical resonant inductive wireless power system. At the time I drew it the Soljačić experiment was high profile, so I felt we needed a drawing of it. I believe most modern wireless power systems don't use the "Witricity" self-resonant coils shown in the diagram, but the transmitter and receiver coils are in tank circuits direct-coupled to the oscillator and rectifier circuits, is that right? The magnetic field lines were not meant to be accurate; I was trying to suggest strong coupling. I'm not sure what you mean by "No one has...observed that the field distribution is affected by frequency as falsely suggested in the...drawing". But I'll remove the diagram until I can draw up a more representative one. --Chetvorno^TALK 19:22, 4 January 2019 (UTC)

I included the term "evanescent fields" just because it was used in the Soljačić paper (and a few others). I agree it seems misleading and the fields are just induction fields. If you say it is not used in this application I will remove it. --Chetvorno^TALK 19:22, 4 January 2019 (UTC)

As for removing the "Resonant inductive coupling" section, I think it is clearly needed to explain this important technology. The text is thoroughly sourced. Although I am just a BSEE and not expert in wireless power, I don't see any technical errors (beyond use of the word "evanescent"); maybe you can point them out. I also don't see that it is "dogmatic"; if you mean it overemphasizes the "Witricity" approach, I disagree; it seems to me the text in the section applies to all resonant inductive systems. Keep in mind that this is not a technical explanation for engineers but an encyclopedia article that attempts to be comprehensible to general readers. --Chetvorno^TALK 19:22, 4 January 2019 (UTC)

As you probably know I have published a few articles concerning non-radiating near-field situations. Beside I am the inventor of the longitudinal configuration for the wireless capacitive resonant coupling. The first subject I'd like to discuss with you is the use of the unipolar/dipolar terms for capacitive coupling. According to me they suggest a difference in essence with the magnetic coupling case instead of a clear duality between the two. The idea that two coils as well as two electrical dipoles can be arranged in a longitudinal way (along the same axis in the direction of the energy transfer) or in a transverse way (perpendicular to the direction of the energy transfer) is much more in agreement with the duality principle (an with the transverse character of the far field). Beside in one of our articles we have demonstrated that for near-field conditions the coupling in both cases (magnetic or electrical) are twice higher for the longitudinal configuration than for the transversal one. In the electrical case the coupling can be further increased using electrode asymmetry. The dual case for magnetic coupling is the use of conical coils with the smaller diameter sides facing.

The second aspect that worries me much more is the incorrect suggestion that resonance has something to do with the coupling whereas it has not. The coupling coefficient can be obtained using the ideas of self and mutual inductance or equivalently self and mutual capacitance. The problem in the current page is that you suggest that these quantities are modified according to frequency whereas they are not at all frequency dependent (if the medium in between in not dispersive). All the quantities involved in the coupling computation depend only on the distribution of charges and currents then only on the geometries of electrodes or coils and the distance between them (with an extremely small influence of the skin effect in the magnetic case). If you want to make an appropriate picture of the field lines you should represent the field lines of a source coil nearly unaffected by the distant load coil with only a very small amount of field lines intercepted by the distant load coil.

The practical coupling coefficients for WPT systems are usually extremely small (often below one percent) as it can be easily measured for instance using the effective coupling coefficient obtained as the difference of the square of the anti-resonance and resonance frequencies divided by the square of the anti-resonance frequency (there is a very nice wiki page on this topic). Of course resonances improve the situation to a very large extent, however it doesn't affect the link but are only internal to the source and load devices. Resonance works as a cumulative effect inside the devices. For instance in the well known case of the Tacoma bridge, the small energy transfer due to the wind is progressively amplified with time in the bridge structure. A simple demonstration is to visualize the oscillation levels for resonant situations. It is very easy to show that the energy increases linearly with time until destruction arises. The amplitude levels are very large if the Q-factor of the circuit is large enough leading to both a larger energy transfer for the same coupling and smaller losses. Outside resonance frequencies energy is dissipated in the fastening structure (in the generator or load circuit in the electrical case, more accurately in the switching transistors or the rectifying diodes). So the only cause of power increase for resonant coupling (whatever electric magnetic, mechanic...) is not the change in coupling coefficient but the increase of amplitudes in the devices themselves. The critical quantity for coupled resonant circuits is the quantity kQ that was called coupling index in some old articles (somewhere in early 1900). When the coupling index is above unity, the efficiency is high. Said otherwise a small coupling can be compensated by large Q-factors (a larger possibility to recycle non-transferred or received energy to increase the amplitudes). All this is well explained for instance if you look in the technical pages of the Q-alliance (the non dogmatic approach of the resonant magnetic coupling and happily also much more successful on the market). The key aspect is to understand that resonance is a general internal process not related to the coupling link whatever its nature.

There is an other dogmatic aspect originated from Witricity and Marin Soljacik reinvention of the wheel that now pervades many wiki pages and articles: the idea that nearby coils or dipoles are coupled through evanescent waves but this is another story.Henri BONDAR (talk) 14:25, 7 January 2019 (UTC)

For the students, to explain the difference between coupling and resonance I often use the example of guitar cords: Two nearby guitar cords are weakly coupled (coupling coefficient is about 10^-4) moreover they are tuned to different frequencies, then when you move one cord, the other one stays nearly at rest. But if you tune two cords at the same frequency you will see with bare eyes that nearly all the energy given to one cord is progressively transferred to the other one and back and forth several times before the energy is totally dissipated. This highly efficient transfer through a weak coupling takes of course quite a long time (only an extremely small amount of energy is transferred at each alternation) and is due to the very high Q-factor involved (around 10^5) and then to the large value of the coupling index kQ=10. When the Q are different in the two coupled devices (the two cords here), the coupling index is k.sqr(Q1Q2),then better results are observed if the two sides have large Q factors (this gives a simple answer to the idea of second resonance).

Another clear explanation of the effect of resonance is to use the impedance adaptation theorem that states that the power transfer is optimized when the reactance of the link is compensated by the appropriate conjugate reactance; leading to a resonant circuit. One illustration of this idea is given on one of my didactic videos: https://www.youtube.com/watch?v=YegIW-1hbvQ&t=3s. Henri BONDAR (talk) 21:35, 7 January 2019 (UTC)

What specific text do you think should be changed in the article? And what do you think it should be changed to? In the capacitive coupling section I changed the terms "bipolar" and "unipolar" to "transverse" and "longitudinal" to address your concern. --Chetvorno^TALK 23:19, 7 January 2019 (UTC)

I think the page is well structured in particular the far-field (radiative) and the non-radiative near-field are well described (good job indeed). I suggest, if not already done, that in non-radiative near-field chapter, you should explain first that there is no directive effect (no antenna gain) and that the field decreases more quickly for quadripoles or more complex structures than for dipoles. That in contrary to far-field were dipoles are always set transverse to the propagation direction, in non-radiative near-field dipoles could be arranged in longitudinal and transverse manner (electric and magnetic coupling treated on equal footing a reference to the electromagnetism duality theorem welcomed). An illustration of the two possible configurations is also welcomed (at least for coils). You may also explain that the longitudinal arrangement leads to higher coupling coefficients at least for intermediate distances and add a reference to support the idea (our article on this topic can provide other ref. if you don't use it directly) https://www.sciencedirect.com/science/article/abs/pii/S0304388613000314. The picture for the magnetic field lines in the longitudinal configuration that is already on the page is OK for me. Maybe a similar drawing for two electrical dipoles should be inserted, the picture should show a few shared field lines instead of all of them (in your two pictures for capacitive coupling some fields lines not linked to the load are missing to illustrate the idea that the coupling is usually weak (I will seek the web for a neutral illustration if any).

According to me the general resonance mechanism should follow in a separate paragraph for instance using the impedance tuning theorem (a conjugate reactance to cancel the natural reactance of the link). The idea that resistance should also be optimized on both sides could also be introduced. Ideally a picture showing up and down transformers to adjust the resistance on both sides, down in the generator side and up in the load side for inductive coupling because the link impedance is very small, and up-down for capacitive coupling because the link impedance is very large (see the examples in my didactic videos). An introduction of the key importance of the coupling index kQ is also welcomed (unfortunately the Qi consortium didactic pages have been removed, I found this https://phys.org/news/2014-09-versatile-pilotage-wireless-power.html but it point to a MIT introduction in 2007 whereas the introduction of the concept was done a century ago at least. I will search for another neutral and clearer source).

Then you may introduce some implementations. Among them the Witricity patent with its specific embedded core-less impedance adapters (the only original input according to me), some other implementations and references to the Qi WPT consortium if not done, some examples of transverse capacitive coupling; Tesla with its transverse vertical dipoles arrangements and his fortuitous discovery of standing waves (Telsa was working in intermediate distances between near and far field). More recently (around 1950 if I am not wrong) the New Zealand team and their first powered vehicle using transverse capacitive coupling, followed by a lot of other implementations (I remember of a Philips toothbrush in the years 1970). Finally, if you like, a reference to our more recent introduction of the asymetric longitudinal capacitive coupling with the only common root (expectedly) our 2006 patent, the following ones and Murata work on the subject.

Be careful, most recent articles (IEEE) do not cite our original patents and the following ones by Murata and more generally all our work. For instance there is a team in Detroit Michigan (now in San Diego) that we visited four years ago under a non disclosure agreement that published two years later several thesis and papers on the longitudinal capacitive coupling for car charging without citing their original source. More generally they have published in IEEE papers all the content of our patents and articles (including very technical aspects patented in Japan) without ever citing their sources (Their bad excuse was that they only cite IEEE sources). They won some innovation prices and earn several founding from Ford and the automotive market. Their main researcher pretend to be the word specialist in capacitive coupling whereas he doesn't even grasp the notions of self and mutual capacitance. Its how many structures works now; personal interest above moral rules and good practice, leading somehow to the evanescence of the structured knowledge !.Henri BONDAR (talk) 08:48, 8 January 2019 (UTC)

For the coupling index kQ, I found the proper link: https://en.wikipedia.org/wiki/Double-tuned_amplifier. The name coupling index is not explicitely written but you find it easily in french documents ("indice de couplage") http://philipperoux.nexgate.ch/Resources/Circuits_couples_VP1.pdf . I am certain that it originated before the years 1940 when the first vacuum tube HF amplifiers were made.

For the electric field lines in case of two coupled dipoles (a quadripole field in the strong coupling case) I found: http://xaktly.com/ElectricField.html but it is probably not the best illustration as the shared field lines corresponding to the mutual capacitance are not clearly separated from the field lines associated to the self capacitances of the two dipoles (the field lines starting on one charge/electrode and ending on the second charge/electrode of the same dipole).

Beside I have a problem with the following page that is more advertising for Witricity than anything else: https://en.wikipedia.org/wiki/Resonant_inductive_coupling. The corresponding pictures are also the first one that you see when you make a WPT search on Google.Henri BONDAR (talk) 11:25, 8 January 2019 (UTC).

Note that the Resonant inductive coupling page was perfectly free of any dogmatic content up to the 23 January 2017. Before the progressive migration of the content since this date; the addition of the incorrect field line pictures, the introduction of the evanescent wave coupling...., the resonance mecanism was correctly explained and the "coupling index" called "factor of merit" was correctly introduced (since 2012).Henri BONDAR (talk) 12:02, 8 January 2019 (UTC)

For a first accurate description of both capacitive and inductive resonant couplings see . Older work (originating around 1932) concerning double tuned resonant circuits is summarized in the chapter 5 (pp 201 to 226), the coupling coefficient and coupling index are introduced in the bottom of page 202, for a direct access see: https://www.jlab.org/ir/MITSeries/V18.PDF.Henri BONDAR (talk) 15:53, 8 January 2019 (UTC)

Let me see if I understand your argument about frequency. The coupling coefficient k for a given geometry is not dependent on frequency. Resonant coupling can transmit more power at greater range because the coupling device (coil or capacitor) is in a resonant circuit. When driven at its resonant frequency, the current in the transmitter coil of a resonant inductive system will be Q times the current in a nonresonant system. Similarly in a resonant capacitive system the voltage on the capacitor plates will be Q times the voltage in a nonresonant system. Therefore the field strength in the resonant systems will be Q times the field strength in a corresponding nonresonant system. Therefore the resonant system will have the same power transfer efficiency as a nonresonant system with coupling coefficient kQ. Is that basically right? --Chetvorno^TALK 21:13, 8 January 2019 (UTC)

Yes, you can also say that; when reactance tuning is made, the impedance seen by the generator is resistive instead of being inductive/capacitive, the energy doesn't oscillate back and forth in the generator/load structure leading to large induced losses and you have much less losses in the device. Quantitatively in the worst case current/voltage are increased Q times for the same amount of losses. The key point is to understand that this process is only device specific and has nothing to do with the link itself. So calling the product kQ a 'factor of merit' instead of a 'coupling index' as historically introduced is awkward but not totally inappropriate. This mechanism is known since at least a century (and perhaps more in acoustics) and has been formulated correctly since 1935 at least in electronics. Tesla was perfectly aware of it when he tried to destroy is hosting building using a mechanical resonance around 1890. Please note that reactance tuning is only the first step in the general impedance matching problem (see https://en.wikipedia.org/wiki/Impedance_matching and go to Maximum power transfer matching subsection)."They are no great discoveries, only a great ignorance of the past" and we are here to avoid that as far as possible.Henri BONDAR (talk) 05:21, 9 January 2019 (UTC)

Sorry I talk too much, it is one of my bad habits (quantity instead of quality). I have read the article again, after your last corrections there are no more urgent issues. However, I suggest that instead of using the inductive/resonant inductive, capacitive/resonant capacitive separation it would be much clearer to use a strong coupling/ weak coupling separation. In strong coupling situations (very close devices) the impedances are mostly resistive, the only reactive contributions are due to leakage inductance and leakage capacitance (more often called stray capacitance) so that impedance matching reduces to resistance tuning (choice of appropriate voltage ratio according to conditions). When the coupling gets smaller (devices set farther apart) the impedance matching process becomes critical to reach high power transfer or good efficiencies. In this case the link impedance is dominantly reactive and using an appropriate matching in the devices is critical. In this section the coupling index can be introduced as well as the importance of the Q-factors. I can send you my pictures for reactance tuning and resistance tuning process illustration if you like (down-up or up-down scheme according to situations). For the Witricity patent, according to me, it should remain only the idea that using subsidiary coils to provide the resistance tuning function provides an elegant way to reduce the number of windings (compared for instance to separate transformers used in my didactic videos) and that large coils lead generally to large Q-factors. Another aspect that we may introduce later is to define the relative range (separation divided by dipole size), in our article we have demonstrated that the maximum practical range for such near-field systems is limited to relative distances of about 10 due to maximum Q-factor considerations.Henri BONDAR (talk) 08:52, 10 January 2019 (UTC)

I appreciate all the expert technical information! I don't know the best sources for learning about this, so I'd really appreciate any links or sources you can give me. Here are my feelings about the points you raise above:

• The point you made about the coupling not being frequency-dependent, and the improvement in efficiency of resonant systems being due solely to the increased amplitude of sources due to resonance, and the "coupling index" k(Q₁Q₂)^1/2 should be in the article.
• On changing the current section names from Inductive coupling / Resonant inductive coupling and Capacitive coupling / Resonant capacitive coupling to Strong coupling / Weak coupling, I oppose this. I think the current organization is good for introducing nontechnical readers to resonant coupling. General readers are likely to be familiar with nonresonant inductive coupling in the form of transformers and cordless toothbrushes, and this type of wireless power is easy to explain. With this background the article introduces resonant coupling as a modification which increases the currents and voltages in the coils and thus increases the strength and range of the fields. The differences you mention between strong resonant coupling and weak resonant coupling are kind of technical and might be better left to a separate mathematical section.
• Resonant inductive coupling article: I absolutely agree with your above criticisms; it is a very poor article. As you mentioned, in Jan 2017 it was rewritten by an editor who not only introduced a lot of WP:UNDUE WEIGHT Witricity material, but made other unfortunate changes:
  • The editor for some reason emphasized single-tuned systems over double-tuned, in which only the receiver coil is a tuned circuit, while the transmitter coil is untuned. He seemed to think that these circuits were better and that industrial wireless power systems were increasingly using them (although all the supporting references he gave were in Japanese). Is this true? What is the advantage of a single-resonant system? It would seem to me that a system in which both coils are resonant circuits is always going to have higher power transfer efficiency than one with only one resonant circuit.
  • More generally, the article overemphasizes wireless power applications. Resonant inductive coupling is widely used in all kinds of areas besides wireless power: impedance matching in radio transmitters and antenna tuners, IF transformer bandpass filters in receivers, inverters and power supplies, ferroresonant transformers. These are the main applications and should be given the most emphasis in the article and listed first.

I plan to rewrite that article when I get a chance. If you could leave a note on the Resonant inductive coupling Talk page with your criticisms that would help improve the article.

• Your point that to maximize power throughput the impedances must be conjugate-matched in the transmitter and receiver (more generally through the entire power chain) is important and should be in the article.

I view the existing sections on inductive and capacitive coupling as introductions for general (non technically educated) readers, and as such I think there is a limit to the amount of detailed technical/mathematical information which should be included in them. It seems to me that there should be an additional section for the mathematical details, something like Theory of near-field power transfer, which could include a derivation of the power transfer equations for resonant inductive and capacitive wireless systems, and a lot of the great points you mention above could be made in that section. --Chetvorno^TALK 01:09, 15 January 2019 (UTC)

I agree with all your suggestions, will contribute actively for the theoretical side and will try to provide more historical references.

• For the general presentation of non-radiating near-field couplings, I continue to think that introducing the idea that dipoles could be arranged either longitudinally or transversely, with up a too time larger coupling for the first case, should be great. A picture with core-less coils will be fine as the introduction of a dual capacitive situation will not be obvious for many readers. Beside it appears later in the capacitive coupling sub-section. Most non-radiative near-field devices are arranged longitudinally (two Tesla coils structures arranged longitudinally in case of the Witricity patent) by the way we should be careful that a correct balance is made between all participants (Witricity, Qi consortium and others).
• Because Witricity has, according to me, a tendency to present their technology as brand new and well above any standard, they have a lot of aggressive followers leading to a form of proselytism over the web. One example is the case of asymmetric resonance, more generally the resonant induction page but also some other pages (see evanescence below). You are right when you think that the Q factors have similar weights on both side of the link as indicated by the coupling index. However, if for some reasons one coil is going to be better than the other one; using the best one on the load side has an advantage in term of field level. When the high-Q coil is used on the generator side, for the same power transfer you have a higher field level and lower energy extraction on the load side, whereas in the second case the field level is lower and the extraction of energy on the load side higher. This is important when you want to fulfill some field regulations while still hiding it under a quasi-magical frame.
• I have some difficulties in the evanescent field page: https://en.wikipedia.org/wiki/Evanescent_field where Witricity dipole-dipole interaction is suggested as an example of evanescent field as well as the capacitor and transformers internal fields (Already removed). Soljacik non-neutral position is according to me not sustained by any other reliable secondary sources. Your opinion welcomed.
• If you have some remaining time, your opinion is also welcomed in the Q-factor page: https://en.wikipedia.org/wiki/Q_factor where I suggest a reorganization of the presentation and an improved initial picture. Henri BONDAR (talk) 06:45, 15 January 2019 (UTC)

References

Doesn't seem especially neutral

I see no mention of any problems with wireless power transfer, except in the titles of references. I also see no mention of alternatives, such as using solar panels and batteries to receive wireless power from the most powerful wireless transmitter in the solar system (the Sun :-). The article seems way too gung-ho; some down-to-earth comments about the challenges would greatly help. For example, an section which explicitly states why analog and digital wireless data transfer is so widespread, but wireless power is not would help. Sanpitch (talk) 22:08, 5 June 2021 (UTC)

Thanks for your feedback! It really helps to improve the article. Read the 3rd paragraph of the Overview section. Wireless (radio) communication is a lot easier than wireless power because a radio receiver only has to receive a tiny amount of the transmitted power to get the information (an FM radio can receive a station clearly with less than one picowatt, 10⁻¹² watt, of the station's power absorbed by the antenna). Receiving a significant percentage of the transmitted power is much harder because the radio waves spread out. I agree that the article does not adequately explain this. --Chetvorno^TALK 08:17, 6 June 2021 (UTC)

Crystal radios

I'm not quite sure where this should go, but the crystal radio is an example of power transfer by radio. Although the power is tiny it can be enough to drive a pair of earphones. Martin of Sheffield (talk) 11:47, 29 November 2021 (UTC)

Nikola Tesla

What happened to Nikola Tesla free electricity system 107.77.169.23 (talk) 04:18, 8 January 2022 (UTC)