Electronics Forum / Basics / August 2008

- snip -

Radiation from an aperture antenna is a matter of diffraction.  At any point
of inspection, the field strength (and therefore the PFD) is the result of
integrating contributions from the distribution of current in the aperture.
If it's a one-dimensional aperture, like a thin wire dipole, then the
integral is over that one dimension; if it's a two-dimensional aperture,
like a dish, the integral is over both dimensions.  This is a little more
complicated than 'high school geometry' but the calculus involved can often
be kept quite simple - it all depends on the form of the current
distribution.

Away from the antenna, simple geometry indicates that the path lengths over
which the different components act are different, so the result of the
integral will depend on the distance between the point of inspection and the
antenna.  However, it has been found for many practical implementations of
dish, etc., that when the difference between path lengths from the centre
and edges of the (normal) aperture is one sixteenth of the wavelength then
the result is essentially independent of the distance.  The distance at
which this condition is achieved is the well-known yardstick 2D^2/lambda,
and, for example, this equates to 360 metres for a 3 metre diameter dish
working at 6 GHz; perhaps the largest size used for terrestrial microwave
links.  Therefore, radiation pattern measurements need to be carried out
using a measurement range at least this long for the patterns to be
generally applicable.

Of course you have the prerogative to apply different interpretations of the
physics to your own experiments and if your interpretation brings theory and
practice together for the general case then it has some validity.  What I've
described above is a small part of the physics that's actually used in the
design of communication links, broadcasting, radar, etc.  There are
well-known simplifications like 'ray optics' that neglect diffraction
phenomena, but these are strictly applicable only to the case of
infinitesimal wavelength, so they might be useable in optics but they cause
approximation when applied to systems with apertures of only 60
wavelengths - the above example.  When the radiation from a dish is
predicted using ray optics, the ray paths from the feed to the reflector all
end up parallel in the aperture so this gives the impression of a
'collimated' parallel-sided beam ... but this is a major approximation and
is certainly not what is measured in practice.

Chris