There are people who spend their careers on the subject.
John Kraus probably wrote the most definitive work on
the subject, but there are hundreds of others. And there is
no simple answer. An antenna that works great for long
distance communication on a given frequency may work
lousy for short range communication, and vice versa.
An antenna that won't "play" well at a given height may
play real well when it is LOWER! Usually, as a very
general rule, the higher the better, but not always.
A directional antenna pointed say due north may not
work to pick up signals originating in that direction,
but may work to pick them up if aimed east or west,
as the signal may be reflected.
There's all kinds of shapes and sizes for antennas
for a given frequency. And when you consider
the whole RF spectrum, the range of antenna
configurations boggles the mind.
There are definite relationships between size
and frequency. F=L/C where F is the frequency,
L is Lambda (the wavelength) and C is the speed
of light. There are also relationships between size
and resonance and impedance. These vary according
to the antenna configuration. Just as an example,
a beam antenna with 3 elements has some characteristics
of interest - gain, front to back ratio, front to side ratio,
stuff like that. The size of it plays a direct role in those
characteristics. Changing the spacing between the elements
affects them. A higher gain 3 element beam will have greater
spacing between the elements than a lower gain 3 element
beam, when the elements are the same size.
You have asked a question that thousands of people ask,
and experiment to find the best solution for their interests.
Getting down to your reference to TV antennas:
a coat hanger will work just fine if it is close enough
to a typical TV transmitting antenna and if there is
not a lot of reflected signal to cause ghosting. But
when faurther away and/or when there is a lot of
reflection, the need for a typical roof mounted TV
antenna arises. These are usually in a configuration
called "log periodic", and exhibit gain over your
coat hanger antenna, as well as a greater degree
of rejection of reflected signals coming it to the antenna
from the side. If you look in catalogs, you'll see that
TV antennas with greater gain have more elements and
longer booms than those with lesser elements.
Finally, the halfwave dipole antenna, mentioned earlier.
That is an antenna which has, according to the
texts, a characteristic impedance in free space
of about 72 ohms. It is made with 2 "poles".
Each "pole" is an element about 1/4 wavelength
long. The two "poles" are run in a (roughly) straight
line, end to end for a total distance of about 1/2
wave - thus the term halfwave dipole. A transmission
line (coax for example) is connected in the center
of the antenna - the center conductor connects to
one pole and the braid connects to the other.
This antenna's impedance will vary depending
on how close it is to other objects and the
ground. (The impedance is discussed at the
frequency where the antenna is resonant. Resonance
is the point at which the inductive and capacitive
reactances cancel each other out, and is frequency
This long winded reply is actually an over simplification!
You can spend a lot of years studying antennas and just
scratch the surface.
> What is it about an antenna that makes it a good antenna for a given
> frequency range?
> Now, I know the basics about electromagnetic waves inducing a
> current in a conductor but what shape and how big is the most
> appropriate antenna for a given job? I'm thinking mainly in terms of
> television signals but I guess radio and cell-phones are all covered by
> the same principles.
> Maybe I'm asking a lot (there are probably whole text-books covering
> this stuff ??) but are there some relationships between the dimensions
> of the antenna (say for simple-geometry antennae), the frequency of the
> EM wave, the orientation of the antenna, the input impedance of it(which
> I can't say I understand very well) etc.
> eg Why aren't coat-hangers (or "cat's ears" \_/ ??) a good solution?
> Any advice or pointers would be very welcome,
> PS: Please feel free to send a copy of your reply directly to me at
> [without the ANTI SPAM of course :-)]