The current in radiation sources has not necessarily to be a conduction current but can be a polarization current or a "magnetic current" due to magnetization. And you don't have to mix dynamic (in the sense of changing) fields because a changing electric field is always accompanied by a magnetic field and vice versa (as predicted by two of Maxwell's equations).

On the other hand, the mixing of changing electric and magnetic fields is a linear superposition: You'll just get the sum of the fields without any new effect apart from additional power flows - as long as the polarization resp. magnetization of the sources is not affected.

So in principle, you can create an EM wave by pushing or pulling a piece of piezoelectric crystal.

But in terms of applications, this might be less effective than a usual antenna driven by an electromagnetic signal.

 Do you know any experimental evidence that changing E or M field produces EM wave?

I checked it at low frequencies and found nothing in the near-field zone.

Article Magnetodynamic waves in the air

• Yes, theoretically possible!!!!
• But, it is better to create it with correct phase so that it will be exactly similar to traditional EM waves

Hello Alexander,

thanks for providing me with the full text of your article! Before I can comment, I have to ask at least one question. If I understand correctly you claim that certain details or consequences of Maxwell's equations can be refuted by experiment; but since you do not provide corrected equations, I have to ask: Do you think that "curl H = J" should still be part of the equations, i. e. "each electric current is accompanied by a vortex of the H field"?

Dear Joerg,

I suppose that useful equations describe experimentally obtained data and predicte new ones. The Bio-Savart law (not "curl H = J") is presented by such an equation. It has many experimental verifications.

Dear Alexander,

to be exact, there is not a causal relation between the changing of one field and the curl of the other. Both phenomena appear just simultaneously, as W. G. V. Rosser pointed out:

http://aapt.scitation.org/doi/abs/10.1119/1.10273?journalCode=ajp

But I do not understand how you intend to produce a changing electric field not accompanied by a magnetic field: If the charge is resting, we have a static field. In order to get a changing field, the charge has to move. The Biot-Savart law says that moving charge (= current) is accompanied by a magnetic field. So, a changing electric field is generally accompanied by a magnetic field.

In your experiments with changing electric fields, the amount of charge on the plate or on the ball is changing. Consequently, there must be an electric current and a nonzero magnetic field. Your result is that your Hall sensor did not register any magnetic field. But this proves only that the sensor is too insensitive or too slow or both to register the magnetic field which arises according to Biot-Savart.

There is one case of moving charge without magnetic field, and you kind of hint at it in connection with the ball: If the charge is evenly distributed on a spherical shell, and if the radius of the shell changes which is equivalent to evenly distributed radial currents, there appear longitudinal waves. However, these waves are not waves in the electric or magnetic field but in the vector and scalar potential fields, A and Phi! Since the vector field A remains without curl, and B = curl A, there is no B field. Furthermore, E = -Laplaceop.(Phi) - d A/d t, but the effects of the waves on these terms cancel each other exactly. So, there is just a static E field.

By the way, I wonder whether anyone has ever tried to verify the existence of these longitudinal potential waves in an experiment similar to the Aharonov–Bohm setup? The outer static E field can be easily cancelled by a charge of opposite sign in the center of the shell.

Anyway, all experiments I know of (including that of Monstein, with nice details), intended to produce longitudinal waves, suffer necessarily from the fact that tangential currents are flowing on the surface of the spherical shell (or something similar) thus generating ordinary radiation fields.

The only way to produce a changing electric field not accompanied by a magnetic field would be to create charge of only one sign, and that seems to be impossible.

As to your experiment with a rotating permanent magnet: In every coil resting in the neighborhood of the magnet, and being so oriented that the magnetic flux through the coil changes, a voltage will be induced according to V = E * length of wire. This principle is employed in millions of bicycle dynamos. In my opinion, your experiment just shows that a flat rectangular piece of metall used as a kind of a monopole antenna isn't a suitable sensor for curl E.

I hope I did not misunderstand important points in your article; for the moment, my view as stated in my first answer isn't modified.

 1. I agree with W. G. V. Rosser in general. E and H intensities are measurable separate parameters of the main phenomenon- EM wave. There may be other parameters, Aharonov–Bohm effect mentioned above confirms it.

2. The most simple method to obtain changing electric field without magnetic one is displacement of a charged wand. For such changing fields special term exists- "quasistatic". In my experiment with plate and ball there was weak unregisterable magnetic field due to the current of course. But energy densities of electric and magnetic fields in EM wave should be the same- according to classical theory. I registered no magnetic field density of the same order of magnitude. It means there was no EM wave.

3. Moving charges may be without magnetic field at the absence of conduction current. The article for bound charges current is in attachment.

4. Magnetic fields produced by the tangential currents are weak and may be 0 if we rotate a piece of electret for example.

5. The scale of "curl" E field in EM wave has the order of its wavelength. It was a lot of kilometers for used low frequencies. So, any metallic piece may be used. And I've written about energy densities above.

I didn't use a coil to avoid the effect of EMI. And simple counterexample to what changing magnetic field always produces electric field in a conductor: We can take closed conductive ring and place it symmetrical to the axe of a magnetic coil, switch on alternating current in the coil and obtain induced current without electrical field inside the ring due to uniform movement of free electrons. In more detail about EMI you can read in the article below.

P. S. More to the point of topic: You didn't answer on my question above.

Article On electromagnetic induction in electric conductors

Article On the magnetic field of bound charges current

If your P. S. refers to your question: "Do you know any experimental evidence that changing E or M field produces EM wave?": My reference to Rosser was meant as an answer.

If you mean by "EM wave" "radiation field": Obviously, there are cases with changing E and B fields but no or very weak radiation; example: short Gaussian current pulse on a straight wire. In all the other cases: For example, have a look at the 10 volumes of "Ultra-wideband, short pulse electromagnetics"; at least one third of the articles are reports on laboratory work.

If you mean by "EM wave" "simultaneously changing and propagating E and B field": How can anyone measure two quantities at the same time and at the same location, and then claim that one quantity is the cause of the other? (That's Rosser's point in short.)

Rosser wrote about EM wave only. I mean by "EM wave" the wave phenomenon characterized by E and H intensities together. Its internal structure as well as the process of its forming are the puzzle up to this time.

Long and short conduction current pulses are off topic.

Thank you Joerg for attention to my question.

Dear Alexander,

after reading your article "On electromagnetic induction ..." I believe I'm beginning to understand how your view differs from the textbooks: When the B field penetrating a wire loop increases, for example, you do not assume that this change is accompanied by an E field but that the B field moves from the outside into the loop. I. e., of F = q * E + q * (v x B), not the first but the second term is responsible for EMI, in your view. Is that correct?