Dielectric Rod Antenna: Can It Radiate?

[sophiecentaur]

You cannot put the 2 conductors of the coax cable onto one end of the dielectric rod and hope that RF is generated

Yes, of course that's right. The waves have to be matched into the dielectric just as with any antenna. If it's a dielectric resonator antenna, the impedance to match to could be very relevant.
Edit: PS just connecting a coax to any old part of any antenna doesn't guarantee it will radiate significant power, whether it's made of metal or a dielectric. It has to be done 'appropriately'. Which part of a regular antenna do you call the 'radiator' and which part do you call the beam-former?

 

[Baluncore]

Which part of a regular antenna do you call the 'radiator' and which part do you call the beam-former?

It might pay to consider if there is a transducer, something that transforms one form of energy into another form. A transmit antenna is a transducer that adapts one end of a non-radiating feedline into an EM wavefront. A receive antenna is a transducer that performs the reciprocal conversion.

So where is the transducer in an antenna? Where does energy change form?
A transmission line is not a transducer.
A waveguide is not a transducer, it is a hollow transmission line.
A dielectric lens is not a transducer.
A horn on a waveguide is not a transducer.

A parabolic reflector is not a transducer, even though it changes the hand. Or is it? The propeller on an aircraft or boat is often considered to be a disc transducer that converts torque into thrust. How much change of energy form is required? Is the transducer the coupling loop into a waveguide, or is it the horn at the radiating end of the waveguide? It is all EM waves, even in a twisted pair. Where do we draw the line?

 

[sophiecentaur]

A receive antenna is a transducer that performs the reciprocal conversion.

That's one way of looking at it. To my mind, the only transducers involved are the transmitting output device and the receiver front end. They actually do change the 'form' of energy. Wires, wave-guides, antennae are only constraining and directing EM Power that's already been generated elsewhere.
But there's no end to this subject because we all have our own model in our heads.
Why can't everyone see things my way?

 

[tech99]

I would like to buy an argument. Consider a resonant half-wave dipole in free space. Model it as an inductive wire in the middle with capacitance between the ends. The current at the centre will be a maximum as the voltage between the ends passes zero. V and I are therefore in quadrature. But since the intrinsic impedance of free space is real, with a value of about 377 ohms, an EM wave in space has the E and H components in phase.

What gives?

The current in the element seems to generate the magnetic component. That H then creates the perpendicular E electric field which permits the wave to proceed through space. It seems the voltage on the dipole is irrelevant.

My take on this is based on experiments which I undertook as an amateur scientist. There are some imperfections still in understanding. The inductance and capacitance are not part of the radiation mechanism. They store reactive energy and create induction fields which are in quadrature. At switch-on, these fields charge up over a few cycles and at switch-off they similarly decay. The L-C resonance makes it easy to drive a current into the antenna, only the Radiation Resistance being involved if losses are small.
The radiation is caused when the voltage at the feedpoint accelerates electrons in the wire. This causes the static fields of the electrons to be tilted, creating a transverse radiated E-field. When this electric wave passes a magnetic sensor, as a consequence of Relativity we see a magnetic field.
The acceleration also causes the electrons to have a velocity, and this equates to a current. The magnetic field of this current is initially indistinguishable from the B radiation, as it is in-phase and falls off with 1/r. However, at a distance of lambda/2 pi, it is found that this field begins to decay more rapidly, whereas the radiated B field continues outwards as radiation, falling with 1/r for ever.
The radiated E-field remains constant out to lambda/2pi, but retards in phase. It then falls with 1/r for ever. The induction E-field does not retard in phase with distance - I believe Hertz noticed this.
Very close to the feedpoint we see a local E-field caused by the feeder voltage. This voltage is equal to the volt drop in the Radiation Resistance, and may be equated to the transverse E-field generated by the accelerating electrons. The induction E-field from the dipole ends is strong near the ends but small near the equatorial plane of the antenna.
In a general case, if we bring probes near a radiating dipole, we see that the E-field initially increases at 6dB per octave of distance as we approach, but remains about constant from lambda/2 pi until very close to the antenna. Very close we see a 6dB increase due to the feeder voltage.
With a magnetic probe the field increases 6dB per octave all the way until we touch the centre of the antenna, so may be typically 20dB or so greater than the E-field when touching the conductor.
If we use a terminated dipole as a probe, the received power remains constant from lambda/2 pi until we physicaly touch the antenna. When we touch the antenna there is no jump in received power. Closer than lambda/2 pi the received power is always half the transmit power, so the path loss between two dipoles never falls below 3dB. This is because when they touch, they form one antenna and half the power is radiated and the other half delivered to the receiver.
.

 

[Baluncore]

The inductance and capacitance are not part of the radiation mechanism. They store reactive energy and create induction fields which are in quadrature.

What is all this sinusoidal stuff with λ and 2π ? Ever since Sir Oliver Lodge, back in 1897, wandering the gas-lit streets of London, to invent frequency tuning, we have been fighting an uphill battle to spread the spectrum. Only the designers of cheap λ/2 dipole antennas need sinusoidal waves. The few of us who are over tuned circuits, are content to get it right by launching the radiant energy at the first attempt. We do not have to tune everything to resonance, then hope that maybe a little bit more will leak out next time, with the inevitable progressive loss of signal coherence. Who in their right mind would stick with inefficient radiators, sinewaves and limited bandwidth.

All an antenna needs to do is to match the impedance of the feedline to that of free space. There is no requirement that the signal be sinusoidal, nor that the antenna be a tuned quarter wave transformer.

OK, I did say I wanted to buy an argument. I am a discerning buyer and believe the one I have here may be of a premium quality, at lower cost. It is as easy to understand as is falling off a log.

When you nudge an electron one way, on the surface near the midpoint of a conductive dipole, there are two equal but opposite electromagnetic disturbances that travel away towards their respective ends. Those disturbances are EM waves guided by their own Narcissian reflection, seen in electrons on the surface of the conductor. The E and M fields of each disturbance are in phase since they are traveling in opposite directions along a non-reactive line. At some point the disturbance must reach a point where the impedance on the dipole surface passes the intrinsic impedance of free space, Zo = Uo * c ≡ 4π * 29.97924580 ≈ 376.73 ohms. At that point the EM wave will be unable to differentiate between the dipole surface and free space. It is not surprising that, once it encounters that perfect match, the wave falls away from the dipole, to radiate outwards into space. The two disturbances traveling away from each other, sum in phase broadside to the dipole, to create the typical dipole radiation pattern. There will be some radiation from all parts of the dipole, but it will be limited by the impedance mismatch between the conductive line and free space.

 

[tech99]

What is all this sinusoidal stuff with λ and 2π ? Ever since Sir Oliver Lodge, back in 1897, wandering the gas-lit streets of London, to invent frequency tuning, we have been fighting an uphill battle to spread the spectrum.

To run me must first walk. As any waveform may be decomposed into sinusoidal components; this seems good starting point.
Prof. John Hughes made tests with a mobile receiver in 1879 in Portland Place, London. I am not aware he insisted on sine waves for his test as he was using damped waves from a spark gap and induction coil. I am not aware of Oliver Lodge doing the same.
The concept of frequency tuning was proposed by Marconi in 1901 in his famous UK 7777 patent.

 

[Baluncore]

To run me must first walk. As any waveform may be decomposed into sinusoidal components; this seems good starting point.

That is OK, until you try to make a broadband quarter wave transformer for many different sinewaves.

The concept of frequency tuning was proposed by Marconi in 1901 in his famous UK 7777 patent.

Marconi stole everything from everywhere to make a system that worked. He took out patents on inventor's techniques to stop others commercialising them.

Prof. John Hughes made tests with a mobile receiver in 1879 in Portland Place, London.

I think you are referring not to John, but to David Edward Hughes, Professor of Music, discoverer of the bad electrical contact, and inventor of the microphone, & etc.

Prof. Oliver Lodge improved the sensitivity of the coherer by tuning the receiving circuit to the natural frequency of the spark transmitter. Wikipedia says; “In 1898 he was awarded the "syntonic" (or tuning) patent by the United States Patent Office”.
Edit:

No. 11,575 A.D. 1897
Date of Application, 10th May, 1897
Complete Specification Left, 5th Feb., 1898 Accepted, 10th Aug., 1898
PROVISIONAL SPECIFICATION.
Improvements in Syntonised Telegraphy without Line Wires.
I, Oliver Joseph Lodge, D.Sc., F.R.S.; of 2, Grove Park, Liverpool, in the County of Lancaster, Professor of Physics, do hereby declare the nature of this invention to be as follows :–
The object of my invention is to enable an operator to transmit messages across space to anyone or more of a number of different individuals in various localities, each of whom is provided with a suitably arranged receiver.
The method consists in utilising certain processes and apparatus for the purpose of producing and detecting rapid electric oscillations, and in so arranging them that the excitation of a particular frequency of oscillation at the sending station may cause a Morse or any other telegraphic instrument to respond at a distant station, by reason of being associated, through a relay or otherwise, with a subsidiary circuit actuated by electric oscillations of that same particular frequency, or by some multiple or sub-multiple of that frequency. Another distant station will similarly be made to receive messages by exciting at the sending station alternations of a different frequency, and so on: and thus individual messages can be transmitted to individual stations without disturbing the receiving appliances at other stations which are tuned or timed or syntonised to a different frequency. Each station will usually be provided with both sending and receiving apparatus, and messages can travel simultaneously in opposite or in cross directions without the least confusion or interference. ...

https://worldwide.espacenet.com/pub...=GB&NR=189711575A&KC=A&rnd=1530989643124&FT=D#