How Antennas Work
antenna-theory.com
antenna-theory.com
1. Accelerate an electron, get a photon. A good transmitting antenna is something that is an efficient structure for accelerating electrons.
2. Antennas are reciprocal. They receive as well as they transmit.
3. Resonant structures are often used because you can keep more electrons accelerating with less energy -- the damp finger on the rim of a wine glass effect.
4. Power can be directed by appropriately phasing the radiating sub-structures to create constructive and destructive interference in the radiated energy.
The rest is modeled simply with a set of simultaneous three dimensional second-order partial differential equations.
While true, this could be really misleading in practical application. A really good transmitting antenna doesn't always (or even usually) make a really good receiving antenna.
Easy answer is, for a fixed load (I.e. a resistor) and frequency, one varies amplitude with power. (perhaps absolute value of area under the curve might be a better description)
For more information see here: http://hyperphysics.phy-astr.gsu.edu/hbase/Waves/powstr.html This is for a physical string, which I think is easier to wrap your head around. The math is very similar for electrical waves.
The places where there is a benefit to having different antennas for receive and transmit depends on the characteristics of the channel and on the application. The classic example being where a receiving antenna that reduces reception of local noise can give a better signal-to-noise ratio than an antenna that has been optimized for the best transmitted signal footprint at the location of the other station.
So, while I agree that in practice there are plenty of times where separate receive and transmit antennas have a benefit, it isn't because of antenna physics.
I have an active magnetic loop for HF RX. This allows me to null out near field interference (a few dB from the transformer in the back yard) and provides very broad bandwidth (VLF thru HF). It is 75 feet from the house, away from that near field interference.
TX is an inverted L, up the side of the house, with a tuner at the base. I don’t have electrical length, but I do have power, so I can trade off efficiency for a power amp, and the bandwidth with the tuner. Interference is irrelevant, but I can see some 60 Hz cross modulation on a monitoring receiver, due to coupling to the house wiring.
Do not confuse the analysis of an antenna in free space with real-world deployment with real ground, noise sources of various kinds, and asymmetrical propagation.
Receiving Antennas for the Radio Amateur Receiving Antennas for the Radio Amateur Transmitting and receiving antennas have different jobs to do. Although the fundamental characteristics of antennas apply to both transmission and reception, the requirements and priorities of receiving antennas can be vastly different from those of transmitting antennas. Receiving Antennas for the Radio Amateur focuses entirely on active and passive receiving antennas and their associated circuits. There are relatively few cases where a radio amateur cannot benefit from a separate, well-designed receiving antenna or antenna system. On the low bands, including our new allocations at 630 and 2,200 meters, heavy emphasis on the receiving end of these radio paths is essential for success.
There is a long, fruitfull discussion of the reciprocity properties of receiving and transmitting antennas.
He discusses at some length how, even at HF, variations in polarization of arriving waves can vastly influence the received signal legibility.
The cheap-skate's antenna modelling package is NEC2, which is public domain. It is a Method-of-Moment solver, not FDTD which is much more useful at microwaves. But for UHF and below, NEC2 is great as long as you know how to dodge the (reasonably well understood) deficiencies. The downside is that it is from the age of FORTRAN, so the native interface is column-sensitive punch card images. People have done free or low-cost wrappers, though, to take the misery out of it.
NEC4 fixes a number of the model bugs in NEC2, but the last I heard NEC4 fell under ITAR, so requires a license.
I don't know of a freebie FDTD EM package, though.
Also, see my comment about reciprocal: https://news.ycombinator.com/item?id=22790001
Now the thing to keep in mind is that an antenna is a passive device. It does amplify the signal but it does not add any power to it, it just collects the power to specific points. This may be easier to understand with a receiving antenna (which collects the signal).
For example, consider the satellite dish which is of course an antenna. Due to its design it should be conceptually easy to understand that the power of the transmitted field is all gathered in a very small area in the front of the dish. The largest the dish, the smaller the point where all the transmitted power is pointed, so less power would be needed to cover largest distances (and more difficult finding where the dish needs to point).
This is why electrically small antennas have an effective aperature much, much larger than their physical size. That ferrite rod antenna in an AM radio can have a massive electrical aperature (antenna gain) since it is narrow band (high Q). The resonance is amplifying the signal.
You can also transmit through it if you can keep it cool. It may have 0.1% efficiency, radiating 1W for 1 kW of input power, if you have a nuclear reactor (say a VLF antenna on an aircraft carrier where you can’t have wire antennas), you can trade power for size.
On the transmitter side keeping the signal linear means you generally don't see great efficiency in terms of power in -> power radiated.
In all electric circuits to change the power a component will consume you either change the voltage you apply to it or change its resistance, based on the P=VI & R=V/I equations.
The analogy of the waves that another comment mentioned is a nice one (since we talk about EM waves).
Another thing that may help you understand that more power means that the electrical charge to the protons/electrons in the air where the EM field is transmitted will also be more.
Of course I know that an antenna is most efficient at radiating power when its length is some fraction/multiple of the emitted wavelength. But I cannot for the life of me intuit how the electrons are being excited and behaving.
If I use the bathtub analogy of sloshing water, it cannot be (I believe) that the electrons are sloshing in bulk up and down the antenna and "accumulating" at one end at the speed of light.
On the other hand, if each electron along the length of the antenna is oscillating in its own relatively stable position, what then does the length matter to the electron at one end versus the other?
Or should I understand it as, energy is being transmitted out of the antenna, like it is in a flute being played, and the electrons are most constructively-interference being reinforced to resonate at the frequency desired (by standing waves in the conductor) if the length of the material matches the wavelength?
This has always been hard to visualize.
Think of a tube stuffed tightly with marbles. Push the marbles at one end, and they move near instantly at the other end. The force transfers instantly, but the marbles may hardly move.
Now imagine the marbles connected with stiff, little springs. Push on one end, and the compression wave moves through the charges quickly (the speed depends on the spring constant). It hits the other end, yielding a little more charge accumulation, then bounces back, yielding a little less charge. Do that a 2.5 billion times a second and you have a WiFi antenna. The charges don’t move much all; it’s the charge displacement that moves, which is the E and H field. The key is also that the displacement must have acceleration (harmonic motion) to have a derivative. Accelerating charges radiate; constant velocity charges (direct current) don’t.
The jist of it is that all the energy is contained in the fields. That metal rod has charge that is easily displaced.
Now as to how an accelerating charge radiates; I can’t remember, but Feynman covers it in his 3 volume lectures.
If you mean a perfect, isotopic antenna; I don’t believe it exists. Though I suppose it could, if you consider an antenna pattern to be a wave function. A spherical mass should be emitting black-body radiation with an isotropic pattern. Maybe only noise can be isotropic. If you try to encode information, it can’t.
(It's almost 100 years old but the physics have not changed. If anything, these early books often have far more detailed explanations because they assumed far less about the reader.)
Thats one of the things that made the ARRL Antenna handbook so valuable to me; they have some solid advice on weatherproofing and dealing with things like wind and snow loads.
Its easy to make an antenna; its just a bent length of wire at the end of the day. Making it be the same length of bent wire today, tomorrow and beyond, when its mounted outside and /or in harsh conditions, thats difficult.
http://www.antenna-theory.com/basics/whyantennasradiate.php
That answers the fundamental question of how antennas work.
Edit: looks like kawfey’s comment answered the issue already, I didn’t scroll down far enough: https://news.ycombinator.com/item?id=22787249
http://www.antenna-theory.com/basics/whyantennasradiate.php is as deep as it gets.
For DEEP antenna theory, I can't recommend the Balanis book high enough. https://www.amazon.com/Antenna-Theory-Analysis-Constantine-B...
They don’t mean that.
I've heard people talk about tv remotes as ``remote controls'' several orders of magnitude more times than RC gear.
I've never seen someone have to reorient their RC transmitter to get a better signal. Modern RC airplane/car/helicopter/boat transmitters have multiple fixed antennas (as do the receivers) and have for many years (since people switched from 50MHz radios to 2.4GHz spread spectrum radios). The range is long enough that it you'll generally lose sight of your vehicle before you lose the radio connection.
The reason you usually put them vertical is so that you can use the Earth as a ground plane, leading to less wasted power. A horizontally positioned whip antenna will lose a fair bit of power to the ground.
"Specifically, consider this statement: Complexity is not a sign of intelligence; simplify. I have found this to a priceless amount of wisdom."
https://www.technologyreview.com/s/611977/get-ready-for-atom...
This topic is uncommon enough that we won't call this a dupe (this came up yesterday: https://news.ycombinator.com/item?id=22781498).
While this kind of approach allows for the proper engineering of antenna systems it is at least 50% wrong regarding the underlying physics.
If you have to guide an EM wave without dispersion (e.g. TEM propagation), it takes at least two conductors; those have loss. The smaller they get (think a thin coax) the more lossy they become.
If you need a wide bandwidth, you need a smaller conductor arrangement to keep it from “over-moding” (becoming non TEM). Once it’s non-TEM, you get dispersion and corrupt your signal.
So there is a fundamental trade-off of bandwidth versus loss. Free space propagation is always TEM, so plenty of bandwidth, but now you must direct it with antennas as opposed to guiding it with conductors.
You can get 110 GHz of bandwidth on a 1 mm coax, but it is very lossy, so much so that the microwave industry/research is looking into non-contact wafer probing for mmWave and THz applications.
Fiber has similar issues. It’s extremely low-loss, and non-TEM, but that dispersion is small enough you can multiplex in multiple channels without much dispersion across a single channel. Over long runs, still the dispersion is large enough that it needs to be compensated with various tricks.
A 10G Ethernet copper cable is at most 3 meters, and hard-wired to the SFP modules. But you can buy a 10G mmWave radio and get 10 km. Now several km of that copper cable would be hundreds dB loss.
https://en.wikipedia.org/wiki/Radio_propagation
There are some ray-tracing and wave theory solvers out there that will readily accept a 2D floor plan, but 3D full wave solutions are intensely complicated computational problems that probably won't even resolve to real-life just because of the complexities of the real world. The biggest problem for AP (router) placement is multipath interference, which either creates a null or node at any given location, compromising of basically infinite paths between the AP, bouncing off of walls, furniture, people, animals, and everything,
The network engineer's rule of thumb is to ignore multipathing completely, and start at 0dB a foot from the router, and subtract 10dB for passing through drywall, 15dB for brick walls, 8dB for glass, and 6dB for every doubling of distance between the wireless AP and the desired client locations. As long as you stay above around -20- -30dB, you should have good signal. I've come up with these numbers in my own experience as an RF engineer. Typical software usually uses n-bounce ray tracing to determine deadzones or optimum placements, but that stuff is expensive and only as accurate as the 2D or 3D model of your space.
An even less intense rule of thumb is place as close to common client locations then move them by trial and error until it works best. I place a single AP on the ground in the center of my home, and another mesh node for my back yard, and my whole house is covered. (I use ubiquiti unifi gear, which are far more powerful than a router/AP-in-one, and it gives you a lot of insight on how well the clients are connected, interference, and other useful data).
This is also why mesh wireless is getting popular, so you can dot mesh APs around the home in just about every room, and ignore the whole problem.