Documente Academic
Documente Profesional
Documente Cultură
php
Introduction to
Antennas
In the 1890s, there were only a few antennas in the world. These rudimentary
devices were primarly a part of experiments that demonstrated the transmission
of electromagnetic waves. By World War II, antennas had become so ubiquitous
that their use had transformed the lives of the average person via radio and
television reception. The number of antennas in the United States was on the
order of one per household, representing growth rivaling the auto industry during
the same period.
By the early 21st century, thanks in large part to mobile phones, the average
person now carries one or more antennas on them wherever they go (cell phones
can have multiple antennas, if GPS is used, for instance). This significant rate of
growth is not likely to slow, as wireless communication systems become a larger
part of everyday life. In addition, the strong growth in RFID devices suggests
that the number of antennas in use may increase to one antenna per object in the
world (product, container, pet, banana, toy, cd, etc.). This number would dwarf
the number of antennas in use today. Hence, learning a little (or a large amount)
about of antennas couldn't hurt, and will contribute to one's overall understanding
of the modern world.
A 70-meter dish antenna. Part of NASA's Jet Propulsion
Laboratory (JPL) communication system.
Antenna Theory
History
What is the origin of the antenna? I'm ruling out such early devices as compasses,
because while they in some sense receive a magnetic field, it is not an
electromagnetic field. Ben Franklin's kite experiment wasn't quite an antenna, as
that captured lightning discharge, which is a direct current path where the energy
is not transferred independent of the medium it travels. The human eye of course
receives high frequency electromagnetic waves (light, to the layman).
Technically the eye could be classified as an antenna; however since it can't
transmit waves, it is really a sensor, so I'll exclude that as well.
The first experiments that involved the coupling of electricity and magnetism and
showed a definitive relationship was that done by Faraday somewhere around the
1830s. He slid a magnetic around the coils of a wire attached to a galvanometer.
In moving the magnet, he was in effect creating a time-varying magnetic field,
which as a result (from Maxwell's Equations), must have had a time-varying
electric field. The coil acted as a loop antenna and received the electromagnetic
radiation, which was received (detected) by the galvanometer - the work of an
antenna. Interestingly, the concept of electromagnetic waves had not even been
thought up at this point.
A rough outline of some major antennas and their discovery/fabrication dates are
listed:
References
[2] W2AEE Antenna History. Arthur M. Kay (?), scanned by Alan Crosswell.
http://www.w2aee.columbia.edu/history/antenna-history.html
Antenna Basics
Antenna Fundamentals
Lets get right down to the study of antennas and Antenna Basics. Suppose one
day you're walking down the street and a kind but impatient person runs up and
asks you to design an antenna for them. "Sure", you quickly reply, adding "what
is the desired frequency, gain, bandwidth, impedance, and polarization?"
Or perhaps you have never heard of (or are a little rusty) on the above
parameters. Well then, you've come to the right place. Before we can design an
antenna or discuss antenna types, we must understand the basics of antennas,
which are the fundamental parameters that characterize an antenna.
So let us learn something. We'll start with frequency and step through radiation
patterns, directivity and gain, and ultimately close with an explanation on why
antennas radiate. Jump ahead if this is already familiar to you.
Frequency
Frequency is one of the most important concepts in the universe and to antenna
theory, which we will see. But fortunately, it isn't too complicated.
All electromagnetic waves propagate at the same speed in air or in space. This
speed (the speed of light) is roughly 671 million miles per hour (1 billion
kilometers per hour). This is roughly a million times faster than the speed of
sound (which is about 761 miles per hour at sea level). The speed of light will be
denoted as c in the equations that follow. We like to use "SI" units in science
(length measured in meters,time in seconds,mass in kilograms), so we will
forever remember that:
Before defining frequency, we must define what a "electromagnetic wave" is.
This is an electric field that travels away from some source (an antenna, the sun,
a radio tower, whatever). A traveling electric field has an associated magnetic
field with it, and the two make up an electromagnetic wave.
The universe allows these waves to take any shape. The most important shape
though is the sinusoidal wave, which is plotted in Figure 1. EM waves vary with
space (position) and time. The spatial variation is given in Figure 1, and the the
temporal (time) variation is given in Figure 2.
How fast someone walks depends on the size of the steps they take (the
wavelength) multipled by the rate at which they take steps (the frequency). The
speed that the waves travel is how fast the waves are oscillating in time (f )
multiplied by the size of the step the waves are taken per period ( ). The
equation that relates frequency, wavelength and the speed of light can be tattooed
on your forehead:
Basically, the frequency is just a measure of how fast the wave is oscillating. And
since all EM waves travel at the same speed, the faster it oscillates the shorter the
wavelength. And a longer wavelength implies a slower frequency.
This may sound stupid, and actually it probably should. When I was young I
remember scientists discussing frequency and I could never see why it mattered.
But it is of fundamental importance, as will be explained in the "more advanced"
section on frequency.
Let's start with a question. What is the frequency of the following waveform?
Figure 1. A simple waveform.
Well, you'd look for what the period is and realize that it isn't periodic over the plotted region. The
you'd tell me the question was stupid. But here we go:
Hence, for all of antenna theory, we will frequency be discussing wavelength of frequency. Actual
antennas radiate real world signals - data from the internet over WIFI, speech signals, etc etc etc.
However, since every piece of information in the universe can be decomposed into sine and cosine
components of varying frequencies, we always discuss antennas in terms of the wavelength it
operates at or the frequency we are using.
As a further consequence of this, the power an antenna can transmit is divided into frequency
regions, or frequency bands. In the next section, we'll look at what we can say about these
frequency bands.
Frequency Bands
How can your cell phone and your television work at the same time? Both use antennas to receive
information from electromagnetic waves, so why isn't there a problem?
The answer goes back to the fundamental secret of the universe. No matter what information you
want to send, that waveform can be represented as the sum of a range of frequencies. By the use o
modulation (which in a nutshell shifts the frequency range of the waveform to be sent to a higher
frequency band), the waveforms can be relocated to separate frequency bands.
As an example, cell phones that use the PCS (Personal Communications Service) band have their
signals shifted to 1850-1900 MHz. Television is broadcast primarily at 54-216 MHz. FM radio
operates between 87.5-108 MHz.
The set of all frequencies is referred to as "the spectrum". Cell phone companies have to pay big
money to get access to part of the spectrum. For instance, AT&T has to bid on a slice of the
spectrum with the FCC, for the "right" to transmit information within that band. The transmission
of EM energy is greatly regulated. When AT&T is sold a slice of the spectrum, they can not
transmit energy at any other band (technically, the amount transmitted must be below some
threshold in adjacent bands).
The Bandwidth of a signal is the difference between the signals high and low frequencies. For
instance, a signal transmitting between 40 and 50 MHz has a bandwidth of 10 MHz. This means
that the energy of the signal is contained between 40 and 50 MHz (and the energy in any other
frequency range is negligible).
We'll wrap up with a table of frequency bands along with the corresponding wavelengths. From th
table, we see that VHF is in the range 30-300 MHz (30 Million-300 Million cycles per second). A
the very least then, if someone says they need a "VHF antenna", you should now understand that
the antenna should transmit or receive electromagnetic waves that have a frequency of 30-300
MHz.
Radiation Pattern
A radiation pattern defines the variation of the power radiated by an antenna as a function of the
direction away from the antenna. This power variation as a function of the arrival angle is observe
in the antenna's far field.
As an example, consider the 3-dimensional radiation pattern in Figure 1, plotted in decibels (dB) .
Figure 1. Example radiation pattern for an Antenna (generated with FEKO software).
This is an example of a donut shaped or toroidal radiation pattern. In this case, along the z-axis,
which would correspond to the radiation directly overhead the antenna, there is very little power
transmitted. In the x-y plane (perpendicular to the z-axis), the radiation is maximum. These plots
are useful for visualizing which directions the antenna radiates.
Typically, because it is simpler, the radiation patterns are plotted in 2-d. In this case, the patterns
are given as "slices" through the 3d plane. The same pattern in Figure 1 is plotted in Figure 2.
Standard spherical coordinates are used, where is the angle measured off the z-axis, and is
the angle measured counterclockwise off the x-axis.
If you're unfamiliar with radiation patterns or spherical coordinates, it may take a while to see
that Figure 2 represents the same radiation pattern as shown in Figure 1. The radiation pattern on
the left in Figure 2 is the elevation pattern, which represents the plot of the radiation pattern as a
function of the angle measured off the z-axis (for a fixed azimuth angle). Observing Figure 1,
we see that the radiation pattern is minimum at 0 and 180 degrees and becomes maximum
broadside to the antenna (90 degrees off the z-axis). This corresponds to the plot on the left in
Figure 2.
The radiation pattern on the right in Figure 2 is the azimuthal plot. It is a function of the
azimuthal angle for a fixed polar angle (90 degrees off the z-axis in this case). Since the
radiation pattern in Figure 1 is symmetrical around the z-axis, this plot appears as a constant in
Figure 2.
A pattern is "isotropic" if the radiation pattern is the same in all directions. Antennas with
isotropic radiation patterns don't exist in practice, but are sometimes discussed as a means of
comparison with real antennas.
Some antennas may also be described as "omnidirectional", which for an actual antenna means
that the radiation pattern is isotropic in a single plane (as in Figure 1 above for the x-y plane, or
the radiation pattern on the right in Figure 2). Examples of omnidirectional antennas include the
dipole antenna and the slot antenna.
The third category of antennas are "directional", which do not have a symmetry in the radiation
pattern. These antennas typically have a single peak direction in the radiation pattern; this is the
direction where the bulk of the radiated power travels. These antennas are very common;
examples of antennas with highly directional radiation patterns include the dish antenna and the
slotted waveguide antenna. An example of a highly directional radiation pattern (from a dish
antenna) is shown in Figure 3:
Field Regions
The fields surrounding an antenna are divided into 3 principle regions:
Reactive Near Field
Radiating Near Field or Fresnel Region
Far Field or Fraunhofer Region
The far field region is the most important, as this determines the antenna's
radiation pattern. Also, antennas are used to communicate wirelessly from long
distances, so this is the region of operation for most antennas. We will start with
this region.
[Equation 1]
[Equation 2]
[Equation 3]
The first and second equation above ensure that the power radiated in a given
direction from distinct parts of the antenna are approximately parallel (see Figure
1). This helps ensure the fields in the far-field region behave like plane waves.
Note that >> means "much much greater than" and is typically assumed satisfied
if the left side is 10 times larger than the right side.
Figure 1. The Rays from any Point on the Antenna are Approximately Parallel in
the Far Field.
Finally, where does the third far-field equation come from? Near a radiating
antenna, there are reactive fields (see reactive near field region, below), that
typically have the E-fields and H-fields die off with distance as and . The
third equation above ensures that these near fields are gone, and we are left with
Note that depending on the values of R and the wavelength, this field may or may
not exist.
Directivity
Directivity is a fundamental antenna parameter. It is a measure of how 'directional'
an antenna's radiation pattern is. An antenna that radiates equally in all directions
would have effectively zero directionality, and the directivity of this type of
antenna would be 1 (or 0 dB).
[Silly side note: When a directivity is specified for an antenna, what is meant
is 'peak directivity'. Directivity is technically a function of angle, but the
angular variation is described by its radiation pattern. Hence, directivity
throughout this page will mean peak directivity, because it is rarely used in
another context.]
[Equation 1]
As an example consider two antennas, one with radiation patterns given by:
Antenna 1
Antenna 2
Antennas for cell phones should have a low directivity because the signal can come from any
direction, and the antenna should pick it up. In contrast, satellite dish antennas have a very high
directivity, because they are to receive signals from a fixed direction. As an example, if you get a
directTV dish, they will tell you where to point it such that the antenna will receive the signal.
Finally, we'll conclude with a list of antenna types and their directivities, to give you an idea of
what is seen in practice.
Typical Directivity
Antenna Type Typical Directivity
(dB)
As you can see from the above table, the directivity of an antenna can vary over several order of
magnitude. Hence, it is important to understand directivity in choosing the best antenna for your
specific application. If you need to transmit or receive energy from a wide variety of directions
(example: car radio, mobile phones, computer wifi), then you should design an antenna with a low
directivity. Conversely, if you are doing remote sensing, or targetted power transfer (example:
received signal from a mountain top), you want a high directivity antenna, to maximize power
transfer and reduce signal from unwanted directions.
The general rule in Antenna Theory is that you need an electrically small antenna to produce low
directivity. That is, if you use an antenna with a total size of 0.25 - 0.5 (a quarter- to a half-
wavelength in size), then you will minimize directivity. That is, half-wave dipole antennas or half-
wavelength slot antennas typically have directivities less than 3 dB, which is about as low of a
directivity as you can obtain in practice. Ultimately, we can't make antennas much smaller than a
quarter-wavelength without sacrificing antenna efficiency (the next topic) and antenna bandwidth.
Conversely, for antennas with a high directivity, we'll need antennas that are many wavelengths in
size. That is, antennas such as dish (or satellite) antennas and horn antennas have high directivity,
in part because they are many wavelengths long.
Why is this? [I don't know how to explain this without getting mathematical, sorry!]
Ultimately, it has to do with the properties of the Fourier Transform. When you take the
Fourier Transform of a short pulse, you get a broad frequency spectrum. The analogy is
present in determining the radiation pattern of an antenna: the pattern can be thought of as
the Fourier Transform of the antenna's current or voltage distribution. As a result, small
antennas have broad radiation patterns (low directivity), and antennas with large uniform
voltage or current distributions have very directional patterns (and thus, a high directivity).
Now that we know what directivity is, we can move on to the next antenna concept, gain.
Antenna Efficiency and Antenna
Gain
On this page, we'll introduce two fundamental antenna parameters: antenna
efficiency and antenna gain.
Antenna Efficiency
The efficiency of an antenna relates the power delivered to the antenna and the
power radiated or dissipated within the antenna. A high efficiency antenna has
most of the power present at the antenna's input radiated away. A low efficiency
antenna has most of the power absorbed as losses within the antenna, or reflected
away due to impedance mismatch.
The losses associated within an antenna are typically the conduction losses (due
to finite conductivity of the antenna) and dielectric losses (due to conduction
within a dielectric which may be present within an antenna).
The antenna efficiency (or radiation efficiency) can be written as the ratio of the
radiated power to the input power of the antenna:
[Equation 1]
[Equation 2]
Efficiency is one of the most important antenna parameters. It can be very close
to 100% (or 0 dB) for dish, horn antennas, or half-wavelength dipoles with no
lossy materials around them. Mobile phone antennas, or wifi antennas in
consumer electronics products, typically have efficiencies from 20%-70% (-7 to
-1.5 dB). The losses are often due to the electronics and materials that surround
the antennas; these tend to absorb some of the radiated power (converting the
energy to heat), which lowers the efficiency of the antenna. Car radio antennas
can have a total antenna efficiency of -20 dB (1% efficiency) at the AM radio
frequencies; this is because the antennas are much smaller than a half-wavelength
at the operational frequency, which greatly lowers antenna efficiency. The radio
link is maintained because the AM Broadcast tower uses a very high transmit
power.
Antenna Gain
The term Antenna Gain describes how much power is transmitted in the
direction of peak radiation to that of an isotropic source. Antenna gain is more
commonly quoted in a real antenna's specification sheet because it takes into
account the actual losses that occur.
An antenna with a gain of 3 dB means that the power received far from the
antenna will be 3 dB higher (twice as much) than what would be received from a
lossless isotropic antenna with the same input power.
Antenna Gain is sometimes discussed as a function of angle, but when a single
number is quoted the gain is the 'peak gain' over all directions. Antenna Gain (G)
can be related to directivity (D) by:
[Equation 3]
The gain of a real antenna can be as high as 40-50 dB for very large dish
antennas (although this is rare). Directivity can be as low as 1.76 dB for a real
antenna (example: short dipole antenna), but can never theoretically be less than
0 dB. However, the peak gain of an antenna can be arbitrarily low because of
losses or low efficiency. Electrically small antennas (small relative to the
wavelength of the frequency that the antenna operates at) can be very inefficient,
with antenna gains lower than -10 dB (even without accounting for impedance
mismatch loss).
These concepts can be easily illustrated. Consider the radiation pattern given by:
This pattern is actually fairly easy to generate using Antenna Arrays, as will be
seen in that section. The 3-dimensional view of this radiation pattern is given in
Figure 1.
Figure 1. 3D Radiation Pattern.
The polar (polar angle measured off of z-axis) plot is given by:
The main beam is the region around the direction of maximum radiation (usually
the region that is within 3 dB of the peak of the main beam). The main beam in
Figure 2 is centered at 90 degrees.
The sidelobes are smaller beams that are away from the main beam. These
sidelobes are usually radiation in undesired directions which can never be
completely eliminated. The sidelobes in Figure 2 occur at roughly 45 and 135
degrees.
The Half Power Beamwidth (HPBW) is the angular separation in which the
magnitude of the radiation pattern decrease by 50% (or -3 dB) from the peak of
the main beam. From Figure 2, the pattern decreases to -3 dB at 77.7 and 102.3
degrees. Hence the HPBW is 102.3-77.7 = 24.6 degrees.
Antenna Impedance
Antenna impedance relates the voltage to the current at the input to the antenna.
This is extremely important as we will see.
Note that "j" is the square root of -1. Imaginary numbers are there to give phase
information. If the impedance is entirely real [Z=50 + j*0], then the voltage and
current are exactly in time-phase. If the impedance is entirely imaginary [Z=0 +
j*50], then the voltage leads the current by 90 degrees in phase.
This means the phase of the current will lag the voltage by 45 degrees. That is,
the current waveform is delayed relative to the voltage waveform. To spell it out,
if the voltage (with frequency f) at the antenna terminals is given by
Hence, antenna impedance is a simple concept. Impedance relates the voltage and
current at the input to the antenna. The real part of the antenna impedance
represents power that is either radiated away or absorbed within the antenna. The
imaginary part of the impedance represents power that is stored in the near field
of the antenna. This is non-radiated power. An antenna with a real input
impedance (zero imaginary part) is said to be resonant. Note that the impedance
of an antenna will vary with frequency.
While simple, we will now explain why this is important, considering both the
low frequency and high frequency cases.
Low Frequency
When we are dealing with low frequencies, the transmission line that connects
the transmitter or receiver to the antenna is short. Short in antenna theory always
means "relative to a wavelength". Hence, 5 meters could be short or very long,
depending on what frequency we are operating at. At 60 Hz, the wavelength is
about 3100 miles, so the transmission line can almost always be neglected.
However, at 2 GHz, the wavelength is 15 cm, so the little length of line within
your cell phone can often be considered a 'long line'. Basically, if the line length
is less than a tenth of a wavelength, it is reasonably considered a short line.
From circuit theory, we know that P=I*V. The power that is delivered to the
antenna is:
For maximum power to be transferred from the generator to the antenna, the ideal
value for the antenna impedance is given by:
Hence, we now know that for an antenna to work properly, its impedance must
not be too large or too small. It turns out that this is one of the fundamental
design parameters for an antenna, and it isn't always easy to design an antenna
with the right impedance - particularly over a wide frequency range.
High Frequency
This section will be a little more advanced. In low-frequency circuit theory, the
wires that connect things don't matter. Once the wires become a significant
fraction of a wavelength, they make things very different. For instance, a short
circuit has an impedance of zero ohms. However, if the impedance is measured at
the end of a quarter wavelength transmission line, the impedance appears to be
infinite, even though there is a dc conduction path.
It turns out (after studying transmission line theory for a while), that the input
impedance Zin is given by:
This makes things much simpler. If the antenna is not matched, the input
impedance will vary widely with the length of the transmission line. And if the
input impedance isn't well matched to the source impedance, not very much
power will be delivered to the antenna. This power ends up being reflected back
to the generator, which can be a problem in itself (especially if high power is
transmitted). This loss of power is known as impedance mismatch. Hence, we see
that having a tuned impedance for an antenna is extremely important. For more
information on transmission lines, see the transmission line tutorial.
VSWR
We see that an antenna's impedance is important for minimizing impedance-
mismatch loss. A poorly matched antenna will not radiate power. This can be
somewhat alleviated via impedance matching, although this doesn't always work
over a sufficient bandwidth (bandwidth is the next topic).
A common measure of how well matched the antenna is to the transmission line
or receiver is known as the Voltage Standing Wave Ratio (VSWR). VSWR is a
real number that is always greater than or equal to 1. A VSWR of 1 indicates no
mismatch loss (the antenna is perfectly matched to the tx line). Higher values of
VSWR indicate more mismatch loss.
In the next section on antenna basics, we'll look at the very important antenna
parameter known as bandwidth.
Bandwidth
Bandwidth is another fundamental antenna parameter. Bandwidth describes the
range of frequencies over which the antenna can properly radiate or receive
energy. Often, the desired bandwidth is one of the determining parameters used
to decide upon an antenna. For instance, many antenna types have very narrow
bandwidths and cannot be used for wideband operation.
Note that the above does not imply that 96% of the power delivered to the
antenna is transmitted in the form of EM radiation; losses must still be taken into
account.
Also, the radiation pattern will vary with frequency. In general, the shape of the radiation pattern
does not change radically.
There are also other criteria which may be used to characterize bandwidth. This may be the
polarization over a certain range, for instance, an antenna may be described as having circular
polarization with an axial ratio < 3dB (less than 3 dB) from 1.4-1.6 GHz. This polarization
bandwidth sets the range over which the antenna's operation is approximately circularly polarized.
The bandwidth is often specified in terms of its Fractional Bandwidth (FBW). The FBW is the rati
of the frequecny range (highest frequency minus lowest frequency) divided by the center
frequency. The antenna Q also relates to bandwidth (higher Q is lower bandwidth, and vice versa)
To give some concrete examples of bandwidth, here is a table of the bandwidths for common
antenna types. This will answer such questions as "what is the bandwidth of a dipole antenna?" an
"which antenna has a higher bandwidth - a patch or a spiral antenna?". For a fare comparison, we
set the center frequency for each antenna to 1 GHz (1000 MHz).
Linear Polarization
Let's start by understanding the polarization of a plane electromagnetic wave.
In equation (1), the symbol is a unit vector (a vector with a length of one),
which says that the E-field "points" in the x-direction.
Polarization is the figure that the E-field traces out while propagating. As an
example, consider the E-field observed at (x,y,z)=(0,0,0) as a function of time
for the plane wave described by equation (1) above. The amplitude of this fiel
is plotted in Figure 2 at several instances of time. The field is oscillating at
frequency f.
Figure 2. Observation of E-field at (x,y,z)=(0,0,0) at different times.
Observed at the origin, the E-field oscillates back and forth in magnitude,
always directed along the x-axis. Because the E-field stays along a single line,
this field would be said to be linearly polarized. In addition, if the x-axis was
parallel to the ground, this field could also be described as "horizontally
polarized" (or sometimes h-pole in the industry). If the field was oriented along
the y-axis, this wave would be said to be "vertically polarized" (or v-pole).
A linearly polarized wave does not need to be along the horizontal or vertical axis. For instance, a
wave with an E-field constrained to lie along the line shown in Figure 3 would also be linearly
polarized.
Figure 3. Locus of E-field amplitudes for a linearly polarized wave at an angle.
The E-field in Figure 3 could be described by equation (2). The E-field now has
an x- and y- component, equal in magnitude.
One thing to notice about equation (2) is that the x- and y-components of the
E-field are in phase - they both have the same magnitude and vary at the
same rate.
Circular Polarization
Suppose now that the E-field of a plane wave was given by equation (3):
In this case, the x- and y- components are 90 degrees out of phase. If the field
is observed at (x,y,z)=(0,0,0) again as before, the plot of the E-field versus
time would appear as shown in Figure 4.
Figure 4. E-field strength at (x,y,z)=(0,0,0) for field of Eq. (3).
The E-field in Figure 4 rotates in a circle. This type of field is described as a circularly polarized
wave. To have circular polarization, the following criteria must be met:
If the wave in Figure 4 is travelling out of the screen, the field is rotating in the
counter-clockwise direction and is said to be Right Hand Circularly
Polarized (RHCP). If the fields were rotating in the clockwise direction, the
field would be Left Hand Circularly Polarized (LHCP).
Elliptical Polarization
If the E-field has two perpendicular components that are out of phase by 90
degrees but are not equal in magnitude, the field will end up Elliptically
Polarized. Consider the plane wave travelling in the +z-direction, with E-field
described by equation (4):
The locus of points that the tip of the E-field vector would assume is given in Figure 5.
In addition, elliptical polarization is defined by its eccentricity, which is the ratio of the major and
minor axis amplitudes. For instance, the eccentricity of the wave given by equation (4) is 1/0.3 =
3.33. Elliptically polarized waves are further described by the direction of the major axis. The wav
of equation (4) has a major axis given by the x-axis. Note that the major axis can be at any angle in
the plane, it does not need to coincide with the x-, y-, or z-axis. Finally, note that circular
polarization and linear polarization are both special cases of elliptical polarization. An elliptically
polarized wave with an eccentricity of 1.0 is a circularly polarized wave; an elliptically polarized
wave with an infinite eccentricity is a linearly polarized wave.
In the next section, we will use the knowledge of plane-wave polarization to characterize and
understand antennas.
Polarization of Antennas
Now that we are aware of the polarization of plane-wave EM
fields, antenna polarization is straightforward to define.
In general, for two linearly polarized antennas that are rotated from each other by
an angle , the power loss due to this polarization mismatch will be described
by the Polarization Loss Factor (PLF):
Hence, if both antennas have the same polarization, the angle between their
radiated E-fields is zero and there is no power loss due to polarization mismatch.
If one antenna is vertically polarized and the other is horizontally polarized, the
angle is 90 degrees and no power will be transferred.
As a side note, this explains why moving the cell phone on your head to a
different angle can sometimes increase reception. Cell phone antennas are often
linearly polarized, so rotating the phone can often match the polarization of the
phone and thus increase reception.
Recall that circular polarization is really two orthongal linear polarized waves 90
degrees out of phase. Hence, a linearly polarized (LP) antenna will simply pick
up the in-phase component of the circularly polarized (CP) wave. As a result, the
LP antenna will have a polarization mismatch loss of 0.5 (-3dB), no matter what
the angle the LP antenna is rotated to. Therefore:
Then the effective aperture parameter describes how much power is captured
from a given plane wave. Let p be the power density of the plane wave (in
W/m^2). If P_t represents the power (in Watts) at the antennas terminals
available to the antenna's receiver, then:
Hence, the effective area simply represents how much power is captured from the
plane wave and delivered by the antenna. This area factors in the losses intrinsic
to the antenna (ohmic losses, dielectric losses, etc.).
A general relation for the effective aperture in terms of the peak antenna gain (G)
of any antenna is given by:
Effective aperture will be a useful concept for calculating received power from a
plane wave. To see this in action, go to the next section on the Friis transmission
formula.
Assume that Watts of total power are delivered to the transmit antenna. For
the moment, assume that the transmit antenna is omnidirectional, lossless, and
that the receive antenna is in the far field of the transmit antenna. Then the power
density p (in Watts per square meter) of the plane wave incident on the receive
antenna a distance R from the transmit antenna is given by:
If the transmit antenna has an antenna gain in the direction of the receive antenna
given by , then the power density equation above becomes:
The gain term factors in the directionality and losses of a real antenna. Assume
now that the receive antenna has an effective aperture given by . Then the
power received by this antenna ( ) is given by:
Since the effective aperture for any antenna can also be expressed as:
The resulting received power can be written as:
[Equation 1]
This is known as the Friis Transmission Formula. It relates the free space path
loss, antenna gains and wavelength to the received and transmit powers. This is
one of the fundamental equations in antenna theory, and should be remembered
(as well as the derivation above).
Another useful form of the Friis Transmission Equation is given in Equation [2].
Since wavelength and frequency f are related by the speed of light c (see intro to
frequency page), we have the Friis Transmission Formula in terms of frequency:
[Equation 2]
Equation [2] shows that more power is lost at higher frequencies. This is a
fundamental result of the Friis Transmission Equation. This means that for
antennas with specified gains, the energy transfer will be highest at lower
frequencies. The difference between the power received and the power
transmitted is known as path loss. Said in a different way, Friis Transmission
Equation says that the path loss is higher for higher frequencies.
The importance of this result from the Friis Transmission Formula cannot be
overstated. This is why mobile phones generally operate at less than 2 GHz.
There may be more frequency spectrum available at higher frequencies, but the
associated path loss will not enable quality reception. As a further consequence
of Friss Transmission Equation, suppose you are asked about 60 GHz antennas.
Noting that this frequency is very high, you might state that the path loss will be
too high for long range communication - and you are absolutely correct. At very
high frequencies (60 GHz is sometimes referred to as the mm (millimeter wave)
region), the path loss is very high, so only point-to-point communication is
possible. This occurs when the receiver and transmitter are in the same room, and
facing each other.
Finally, if the antennas are not polarization matched, the above received power
could be multiplied by the Polarization Loss Factor (PLF) to properly account for
this mismatch. Equation [2] above can be altered to produce a generalized Friis
Transmission Formula, which includes polarization mismatch:
[Equation 3]
See also decibel math, which can greatly simplify the calculation of the Friis
Transmission Equation.
Antenna Temperature
Antenna Temperature ( ) is a parameter that describes how much noise an
antenna produces in a given environment. This temperature is not the physical
temperature of the antenna. Moreover, an antenna does not have an intrinsic
"antenna temperature" associated with it; rather the temperature depends on its
gain pattern and the thermal environment that it is placed in. Antenna temperature
is also sometimes referred to as Antenna Noise Temperature.
This states that the temperature surrounding the antenna is integrated over the
entire sphere, and weighted by the antenna's radiation pattern. Hence, an isotropic
antenna would have a noise temperature that is the average of all temperatures
around the antenna; for a perfectly directional antenna (with a pencil beam), the
antenna temperature will only depend on the temperature in which the antenna is
"looking".
Let us get back to charge for a moment. Suppose that for some reason, there is a
negatively charged particle sitting somewhere in space. The universe has decided,
for unknown reasons, that all charged particles will have an associated electric
field with them. This is illustrated in Figure 1.
Figure 1. A negative charge has an associated Electric Field with it, everywhere
in space.
Now, suppose someone came up and punched the charge with their fist, for the
fun of it. The charge would accelerate and travel away at a constant velocity.
How would the universe react in this situation?
The universe has also decided (again, for no apparent reason) that disturbances
due to moving (or accelerating) charges will propagate away from the charge at
the speed of light - c0 = 300,000,000 meters/second. This means the electric
fields around the charge will be disturbed, and this disturbance propagates away
from the charge. This is illustrated in Figure 2.
Once the charge is accelerated, the fields need to re-align themselves. Remember,
the fields want to surround the charge exactly as they did in Figure 1. However,
the fields can only respond to events at the speed of light. Hence, if a point is
very far away from the charge, it will take time for the disturbance (or change in
electric fields) to propagate to the point. This is illustrated in Figure 2.
In Figure 2, we have 3 regions. In the light blue (inner) region, the fields close to
the charge have readapted themselves and now line up as they do in Figure 1. In
the white region (outermost), the fields are still undisturbed and have the same
magnitude and direction as they would if the charge had not moved. In the pink
region, the fields are changing - from their old magnitude and direction to their
new magnitude and direction.
Hence, we have arrived at the fundamental reason for radiation - the fields
change because charges are accelerated. The fields always try to align themselves
as in Figure 1 around charges. If we can produce a moving set of charges (this is
simply electric current), then we will have radiation.
Now, you may have some questions. First - if all accelerating electric charges
radiate, then the wires that connect my computer to the wall should be antennas,
correct? The charges on them are oscillating at 60 Hertz as the current travels so
this should yield radiation, correct?
Answer: Yes. Your wires do act as antennas. However, they are very poor
antennas. The reason (among other things), is that the wires that carry power to
your computer are a transmission line - they carry current to your computer
(which travels to one of your battery's terminals and out the other terminal) and
then they carry the current away from your computer (all current travels in a
circuit or loop). Hence, the radiation from one wire is cancelled by the current
flowing in the adjacent wire (that is travelling the opposite direction).
Another question that will arise is - if its so simple, then everything could be an
antenna. Why don't I just use a metal paper clip as an antenna, hook it up to my
receiver and then forget all about antenna theory?
Answer: A paper clip could definitely act as an antenna if you get current
flowing on the antenna. However, it is not so simple to do this. The impedance of
the paper clip will control how much power your receiver or transmitter could
deliver to the paper clip (i.e. whether or not you could get any current flowing on
the paper clip at all). The impedance will depend on what frequency you are
operating at. Hence, the paper clip will work at certain frequencies as an antenna.
However, you will have to know much more about antennas before you can say
when and it may work in a given situation.
Wire Antennas
Short Dipole Antenna
Dipole Antenna
Half-Wave Dipole
Broadband Dipoles
Monopole Antenna
Microstrip Antennas
Rectangular Microstrip (Patch) Antennas
Reflector Antennas
Corner Reflector
Yagi-Uda Antennas
Spiral Antennas
Aperture Antennas
Slot Antenna
Inverted-F Antenna
Horn Antenna
Vivaldi Antenna
Other Antennas
NFC Antennas
The Short Dipole Antenna
The short dipole antenna is the simplest of all antennas. It is simply an open-
circuited wire, fed at its center as shown in Figure 1.
If the short dipole antenna is oriented along the z-axis with the center of the
dipole at z=0, then the current distribution on a thin, short dipole is given by:
The current distribution is plotted in Figure 2. Note that this is the amplitude of
the current distribution; it is oscillating in time sinusoidally at frequency f.
The fields radiated from the short dipole antenna in the far field are given by:
The above equations can be broken down and understood somewhat intuitively.
First, note that in the far-field, only the and fields are nonzero. Further,
these fields are orthogonal and in-phase. Further, the fields are perpendicular to
the direction of propagation, which is always in the direction (away from the
antenna). Also, the ratio of the E-field to the H-field is given by (the intrinsic
impedance of free space).
This indicates that in the far-field region the fields are propagating like a plane-
wave.
Second, the fields die off as 1/r, which indicates the power falls of as
Third, the fields are proportional to L, indicated a longer dipole will radiate more
power. This is true as long as increasing the length does not cause the short
dipole assumption to become invalid. Also, the fields are proportional to the
current amplitude , which should make sense (more current, more power).
Finally, the spatial variation of the fields as a function of direction from the
antenna are given by . For a vertical antenna oriented along the z-axis,
the radiation will be maximum in the x-y plane. Theoretically, there is no
radiation along the z-axis far from the antenna.
In the next section further properties of the short dipole will be discussed.
The polarization of this antenna is linear. When evaluated in the x-y plane, this
antenna would be described as vertically polarized, because the E-field would be
vertically oriented (along the z-axis).
We now turn to the input impedance of the short dipole, which depends on the
radius a of the dipole. Recall that the impedance Z is made up of three
components, the radiation resistance, the loss resistance, and the reactive
(imaginary) component which represents stored energy in the fields:
In the above equation represents the conductivity of the dipole (usually very
high, if made of metal). The frequency f come into the above equation because of
the skin effect. The reactance or imaginary part of the impedance of a dipole is
roughly equal to:
As an example, assume that the radius is 0.001 and the length is 0.05 .
Suppose further that this antenna is to operate at f=3 MHz, and that the metal is
copper, so that the conductivity is 59,600,000 S/m.
For short dipole antennas that are smaller fractions of a wavelength, the radiation
resistance becomes smaller than the loss resistance, and consequently this
antenna can be very inefficient.
The bandwidth for short dipoles is difficult to define. The input impedance varies
wildly with frequency because of the reactance component of the input
impedance. Hence, these antennas are typically used in narrowband applications.
For a dipole antenna of length L oriented along the z-axis and centered at z=0, the
current flows in the z-direction with amplitude which closely follows the
following function:
Note that this current is also oscillating in time sinusoidally at frequency f. The
current distributions for the quarter-wavelength (left) and full-wavelength (right)
dipole antennas are given in Figure 1. Note that the peak value of the current
is not reached along the dipole unless the length is greater than half a
wavelength.
Figure 1. Current distributions on finite-length dipole antennas.
Before examining the fields radiated by a dipole antenna, consider the input
impedance of a dipole as a function of its length, plotted in Figure 2 below. Note
that the input impedance is specified as Z=R + jX, where R is the resistance and
X is the reactance.
If the dipole antenna's length becomes close to one wavelength, the input
impedance becomes infinite. This wild change in input impedance can be
understood by studying high frequency transmission line theory. As a simpler
explanation, consider the one wavelength dipole shown in Figure 1. If a voltage
is applied to the terminals on the right antenna in Figure 1, the current
distribution will be as shown. Since the current at the terminals is zero, the input
impedance (given by Z=V/I) will necessarily be infinite. Consequently, infinite
impedance occurs whenever the dipole antenna is an integer multiple of a
wavelength.
In the next section, we'll consider the radiation pattern of dipole antennas.
The normalized radiation patterns for dipole antennas of various lengths are
shown in Figure 3.
Figure 3. Normalized radiation patterns for dipole antennas of specified length.
The full-wavelength dipole antenna is more directional than the shorter quarter-
wavelength dipole antenna. This is a typical result in antenna theory: it takes a
larger antenna in general to increase directivity. However, the results are not
always obvious. The 1.5-wavelength dipole pattern is also plotted in Figure 3.
Note that this pattern is maximum at approximately +45 and -45 degrees.
The 3D pattern for the 1-wavelength dipole antenna is shown in Figure 4. This
pattern is similar to the pattern for the quarter- and half-wave dipole antenna.
Figure 4. Normalized 3d radiation pattern for the 1-wavelength dipole antenna.
In the next section, we'll look at the most common dipole antenna, the half-wave
dipole antenna.
To make it crystal clear, if the antenna is to radiate at 600 MHz, what size should
the half-wavelength dipole be?
One wavelength at 600 MHz is = c / f = 0.5 meters. Hence, the half-
wavelength dipole antenna's length is 0.25 meters.
The directivity of a half-wave dipole antenna is 1.64 (2.15 dB). The HPBW is 78
degrees.
The above length is valid if the dipole is very thin. In practice, dipoles are often
made with fatter or thicker material, which tends to increase the bandwidth of the
antenna. When this is the case, the resonant length reduces slightly depending on
the thickness of the dipole, but will often be close to 0.47 .
The first thing apparent from Figure 1 is that the fatter the dipole is made, the
larger the bandwidth becomes. For instance, if the bandwidth is measured as the
frequency range over which |S11|<-9 dB, then the bandwidths are 6.5 MHz, 14
MHz, and 24 MHz, for the blue, green and red curves, respectively.
Secondly, the fatter the dipole gets the lower the resonant frequency becomes. In
other words, if an antenna is to resonate at 100 MHz, the resonant length
decreases as the dipole gets fatter.
Using image theory, the fields above the ground plane can be found by using the
equivalent source (antenna) in free space as shown in Figure 1(b). This is simply
a dipole antenna of twice the length. The fields above the ground plane in Figure
1(a) are identical to the fields in Figure 1(b), which are known and presented in
the dipole antenna section. The monopole antenna fields below the ground plane
in Figure 1(a) are zero.
The radiation pattern of monopole antennas above a ground plane are also known
from the dipole result. The only change that needs to be noted is that the
impedance of a monopole antenna is one half of that of a full dipole antenna. For
a quarter-wave monopole (L=0.25* ), the impedance is half of that of a half-
wave dipole, so Zin = 36.5 + j21.25 Ohms. This can be understood since only
half the voltage is required to drive a monopole antenna to the same current as a
dipole (think of a dipole as having +V/2 and -V/2 applied to its ends, whereas a
monopole antenna only needs to apply +V/2 between the monopole antenna and
the ground to drive the same current). Since Zin = V/I, the impedance of the
monopole antenna is halved.
Note that the resulting radiation pattern for this monopole antenna is still
omnidirectional. However, the direction of peak-radiation has changed from the
x-y plane to an angle elevated from that plane. In general, the large the ground
plane is, the lower this direction of maximum radiation; as the ground plane
approaches infinite size, the radiation pattern approaches a maximum in the x-y
plane.
Typically, the width d of the folded dipole antenna is much smaller than the
length L.
Because the folded dipole forms a closed loop, one might expect the input
impedance to depend on the input impedance of a short-circuited transmission
line of length L. However, you can imagine the folded dipole antenna as two
parallel short-circuited transmission lines of length L/2 (separated at the midpoint
by the feed in Figure 1). It turns out the impedance of the folded dipole antenna
will be a function of the impedance of a transmission line of length L/2.
Also, because the folded dipole is "folded" back on itself, the currents can
reinforce each other instead of cancelling each other out, so the input impedance
will also depend on the impedance of a dipole antenna of length L.
Folded Dipol
Impedanc
The folded dipole antenna is resonant and radiates well at odd integer multiples
of a half-wavelength (0.5 , 1.5 , ...), when the antenna is fed in the center as
shown in Figure 1. The input impedance of the folded dipole is higher than that
for a regular dipole, as will be shown in the next section.
The folded dipole antenna can be made resonant at even multiples of a half-
wavelength ( 1.0 , 2.0 ,...) by offsetting the feed of the folded dipole in
Figure 1 (closer to the top or bottom edge of the folded dipole).
Since the loop is electrically small, the current within the loop can be
approximated as being constant along the loop, so that I= .
The small loop is often referred to as the dual of the dipole antenna, because if a
small dipole had magnetic current flowing (as opposed to electric current as in a
regular dipole), the fields would resemble that of a small loop.
While the short dipole has a capacitive impedance (imaginary part of impedance
is negative), the impedance of a small loop is inductive (positive imaginary part).
The radiation resistance (and ohmic loss resistance) can be increased by adding
more turns to the loop. If there are N turns of a small loop antenna, each with a
surface area S (we don't require the loop to be circular at this point), the radiation
resistance for small loops can be approximated (in Ohms) by:
For a small loop, the reactive component of the impedance can be determined by
finding the inductance of the loop, which depends on its shape (then
X=2*pi*f*L). For a circular loop with radius a and wire radius p, the reactive
component of the impedance is given by:
Small loops often have a low radiation resistance and a highly inductive
component to their reactance. Hence, they are most often used as receive
antennas. Exaples of their use include in pagers, and as field strength probes used
in wireless measurements.
Microstrip Antennas
Rectangular Microstrip (Patch) Antennas
Microstrip or patch antennas are becoming increasingly useful because they can
be printed directly onto a circuit board. Microstrip antennas are becoming very
widespread within the mobile phone market. Patch antennas are low cost, have a
low profile and are easily fabricated.
The width W of the microstrip antenna controls the input impedance. Larger
widths also can increase the bandwidth. For a square patch antenna fed in the
manner above, the input impedance will be on the order of 300 Ohms. By
increasing the width, the impedance can be reduced. However, to decrease the
input impedance to 50 Ohms often requires a very wide patch antenna, which
takes up a lot of valuable space. The width further controls the radiation pattern.
The normalized radiation pattern is approximately given by:
The fields of the microstrip antenna are plotted in Figure 2 for W=L=0.5 .
Figure 2. Normalized Radiation Pattern for Microstrip (Patch) Antenna.
The fringing fields around the antenna can help explain why the microstrip
antenna radiates. Consider the side view of a patch antenna, shown in Figure 4.
Note that since the current at the end of the patch is zero (open circuit end), the
current is maximum at the center of the half-wave patch and (theoretically) zero
at the beginning of the patch. This low current value at the feed explains in part
why the impedance is high when fed at the end (we'll address this again later).
Since the patch antenna can be viewed as an open circuited transmission line, the
voltage reflection coefficient will be -1 (see the transmission line tutorial for
more information). When this occurs, the voltage and current are out of phase.
Hence, at the end of the patch the voltage is at a maximum (say +V volts). At the
start of the patch antenna (a half-wavelength away), the voltage must be at
minimum (-V Volts). Hence, the fields underneath the patch will resemble that of
Figure 4, which roughly displays the fringing of the fields around the edges.
It is the fringing fields that are responsible for the radiation. Note that the
fringing fields near the surface of the patch antenna are both in the +y direction.
Hence, the fringing E-fields on the edge of the microstrip antenna add up in
phase and produce the radiation of the microstrip antenna. This paragraph is
critical to understanding the patch antenna. The current adds up in phase on the
patch antenna as well; however, an equal current but with opposite direction is on
the ground plane, which cancels the radiation. This also explains why the
microstrip antenna radiates but the microstrip transmission line does not. The
microstrip antenna's radiation arises from the fringing fields, which are due to the
advantageous voltage distribution; hence the radiation arises due to the voltage
and not the current. The patch antenna is therefore a "voltage radiator", as
opposed to the wire antennas, which radiate because the currents add up in phase
and are therefore "current radiators".
As a side note, the smaller is, the more "bowed" the fringing fields become;
they extend farther away from the patch. Therefore, using a smaller permittivity
for the substrate yields better radiation. In contrast, when making a microstrip
Since the current has a sinusoidal distribution, moving in a distance R from the
end will increase the current by cos(pi*R/L) - this is just noting that the
wavelength is 2*L, and so the phase difference is 2*pi*R/(2*L) = pi*R/L.
The voltage also decreases in magnitude by the same amount that the current
increases. Hence, using Z=V/I, the input impedance scales as:
In the above equation, Zin(0) is the input impedance if the patch was fed at the
end. Hence, by feeding the patch antenna as shown, the input impedance can be
decreased. As an example, if R=L/4, then cos(pi*R/L) = cos(pi/4), so that
[cos(pi/4)]^2 = 1/2. Hence, a (1/8)-wavelength inset would decrease the input
impedance by 50%. This method can be used to tune the input impedance to the
desired value.
The goal is to match the input impedance (Zin) to the transmission line (Z0). If
the impedance of the antenna is ZA, then the input impedance viewed from the
beginning of the quarter-wavelength line becomes
This input impedance Zin can be altered by selection of the Z1, so that Zin=Z0
and the antenna is impedance matched. The parameter Z1 can be altered by
changing the width of the quarter-wavelength strip. The wider the strip is, the
lower the characteristic impedance (Z0) is for that section of line.
The position of the feed can be altered as before (in the same way as the inset
feed, above) to control the input impedance.
The coaxial feed introduces an inductance into the feed that may need to be taken
into account if the height h gets large (an appreciable fraction of a wavelength).
In addition, the probe will also radiate, which can lead to radiation in undesirable
directions.
The advantage of the coupled feed is that it adds an extra degree of freedom to
the design. The gap introduces a capacitance into the feed that can cancel out the
inductance added by the probe feed.
Aperture Feeds
Another method of feeding microstrip antennas is the aperture feed. In this
technique, the feed circuitry (transmission line) is shielded from the antenna by a
conducting plane with a hole (aperture) to transmit energy to the antenna, as
shown in Figure 5.
Figure 5. Aperture coupled feed.
The upper substrate can be made with a lower permittivity to produce loosely
bound fringing fields, yielding better radiation. The lower substrate can be
independently made with a high value of permittivity for tightly coupled fields
that don't produce spurious radiation. The disadvantage of this method is
increased difficulty in fabrication.
First, the length of the patch L controls the resonant frequency as seen here. This
is true in general, even for more complicated microstrip antennas that weave
around - the length of the longest path on the microstrip controls the lowest
frequency of operation. Equation (1) below gives the relationship between the
resonant frequency and the patch length:
(1)
Second, the width W controls the input impedance and the radiation pattern (see
the radiation equations here). The wider the patch becomes the lower the input
impedance is.
The permittivity of the substrate controls the fringing fields - lower
permittivities have wider fringes and therefore better radiation. Decreasing the
permittivity also increases the antenna's bandwidth. The efficiency is also
increased with a lower value for the permittivity. The impedance of the antenna
increases with higher permittivities.
The height of the substrate h also controls the bandwidth - increasing the height
increases the bandwidth. The fact that increasing the height of a patch antenna
increases its bandwidth can be understood by recalling the general rule that "an
antenna occupying more space in a spherical volume will have a wider
bandwidth". This is the same principle that applies when noting that increasing
the thickness of a dipole antenna increases its bandwidth. Increasing the height
also increases the efficiency of the antenna. Increasing the height does induce
surface waves that travel within the substrate (which is undesired radiation and
may couple to other components).
The following equation roughly describes how the bandwidth scales with these
parameters:
The transient pulse will be of the form given by exp(-(t-T0)/T )^2, where T0 is
the time delay and will be 45 pS (picoseconds, 10^-12), and T is a parameter that
controls the rate of rise and fall, which is 15 pS. This function is plotted in Figure
2.
Figure 2. Incident (transient) pulse fed to a Patch Antenna.
In the following video, we will view the z-directed electric field, immediately
below the patch antenna. Note that the surface of the patch is normal to the z-
axis. We can clearly see the incident pulse propagate down the microstrip line, be
disturbed by the microstrip antenna, then some of the fields are reflected, some
radiate away, and some stay resonant below the patch and eventually radiate
away or reflect back down the microstrip line.
This video shows a Gaussian pulse travel down the microstrip. Some energy is
reflected back. Incidentally, taking the Fourier transform of the incident pulse and
the returned signal and taking the ratio would give S11 (return loss) as a function
of frequency for this antenna.
If you would like to see how this numerical electromagnetics simulation was
developed, see the patch antenna numerical example page.
PIFA - The Planar Inverted-F
Antenna
Antenna designers are always looking for creative ways to improve performance.
One method used in patch antenna design is to introduce shorting pins (from the
patch to the ground plane) at various locations. To illustrate how this may help,
two instances will be illustrated, the quarter-wavelength Patch Antenna, which
leads into the Planar Inverted-F Antenna (PIFA).
Quarter-Wavelength Patch
Because the patch is shorted at the end, the current at the end of the patch
antenna is no longer forced to be zero. As a result, this antenna actually has the
same current-voltage distribution as a half-wave patch antenna. However, the
fringing fields which are responsible for radiation are shorted on the far end, so
only the fields nearest the transmission line radiate. Consequently, the gain is
reduced, but the patch antenna maintains the same basic properties as a half-
wavelength patch, but is reduced in size 50%.
You may be tempted to think that the shorting pin would zero out any power
delivered to the antenna. However, because patches are high frequency devices
(typically used at >1 GHz), the shorting pin actually introduces a parallel
inductance to the antenna impedance. The equivalent circuit of the above antenna
is shown in Figure 3. The antenna impedance is given by ZA, and the shorting
pin introduces a reactance equal to jX.
The affect of the parallel inductance shifts the resonant frequency of the antenna.
In particular, the two components in parallel would result in their admittances
(Y=1/Z) adding. Hence, the admittance of the patch has a 1/(jX) added to it. In
this manner, the resonant frequency can be altered.
In addition, the shorting pin can become capacitive if instead of extending all the
way to the ground plane, it is left floating a small amount above. This introduces
another design parameter to optimize performance.
Planar Inverted F-Antenna (PIFA)
The Planar Inverted-F antenna (PIFA) is increasingly used in the mobile phone
market. The antenna is resonant at a quarter-wavelength (thus reducing the
required space needed on the phone), and also typically has good SAR properties.
This antenna resembles an inverted F, which explains the PIFA name. The Planar
Inverted-F Antenna is popular because it has a low profile and an omnidirectional
pattern. The PIFA is shown from a side view in Figure 4.
The PIFA is resonant at a quarter-wavelength due to the shorting pin at the end.
We'll see how the resonant length is defined exactly in a minute. The feed is
placed between the open and shorted end, and the position controls the input
impedance.
In Figure 5, we have a PIFA of length L1, of width L2. The shorting pin (or
shorting post) is of width W, and begins at one edge of the PIFA as shown in
Figure 5. The feed point is along the same edge as shown. The feed is a distance
D from the shorting pin. The PIFA is at a height h from the ground plane. The
PIFA sits on top of a dielectric with permittivity as with the patch antenna.
The impedance of the PIFA can be controlled via the distance of the feed to the
short pin (D). The closer the feed is to the shorting pin, the impedance will
decrease; the impedance can be increased by moving it farther from the short
edge. The PIFA can have it's impedance tuned with this parameter.
The resonant frequency of the PIFA depends on W. If W=L2, then the shorting
pin runs the entire width of the patch. In this case, the PIFA is resonant (has
maximum radiation efficiency) when:
[Equation 1]
Suppose that W=0, so that the short is just a pin (or assume W << L2). Then the
PIFA is resonant at:
[Equation 2]
Why does the resonant length of the PIFA depend on the shorting pin length W?
Intuitively, think about how a quarter-wavelength patch antenna radiates. It needs
a quarter-wavelength of space between the edge and the shorting area. If W=L2,
then the distance from one edge to the short is simply L1, which gives us
Equation [1].
What about when W=0? Since it is the fringing fields along the edge that give
rise to radiation in microstrip antennas, we see that the length from the open-
circuited radiating edge (the far edge in Figure 5) to the shorting pin is on
average equal to L1+L2. You can convince yourself of this by measuring the
distance from any point on the far edge of the PIFA to the shorting pin. The
clockwise and counter-clockwise paths always add up to 2*(L1+L2), so on
average, resonance will occur when the path length (L1+L2) for a single path is a
quarter-wavelength.
[Equation 3]
To make things concrete, suppose L1=0.1 meters (10cm), L2=0.05 meters (5 cm),
W=0.02 meters (2cm), and that =4. Then what is the resonant frequency? The
solution can be found in Equation [4]:
[Equation 4]
In Equation [4], note that we used one of the fundamental antenna equations,
relating wavelength, speed of light and permittivity:
[Equation 5]
Why does this work? Well, to the right of the feed in Figure 6, we have a short
circuit to ground. Short circuits with a small fraction of a wavelength can be
viewed as a parallel inductance to ground, as far as the impedance is concerned.
Similary, the open circuit and arm to the left of the feed in Figure 6 can be
viewed as a capacitor (if this isn't too clear, you might want to check out the
transmission line tutorial). The distances from the feed to the shorting pin, or the
feed to the open edge of the PIFA determine the inductance and capacitance,
respectively. In some sense, the lengths are required such that the inductance and
capacitance can be balanced out.
Hence, if we shorten up the length of the PIFA, we lose some of the capacitance
to the left of the feed in Figure 6. To compensate for this, we add a parallel
capacitance, and (from an impedance perspective), everything remains balanced
and the PIFA radiates.
This technique works, but be careful: you lose radiation efficiency by using this
technique (and the bandwidth of your PIFA will decrease as well). You can't just
decrease the size of your PIFA, replace it with capacitance and expect everything
to be the same: you can't get something for nothing; antenna engineering is all
about trade-offs.
The phone has 6 antennas, as shown in Figure 7. The Tx/Rx cellular antenna is
the blue square at the bottom and the diversity cellular antenna is in the upper left
region. The GPS antenna (1.575 GHz) is on the top, and the WIFI antenna (which
is dual band according to the FCC report, operating at 2.4 GHz and 5 GHz) is the
green square on the lower right side. This phone also has WiMax antennas
(operating at 2.6 GHz), one for Tx/Rx and another as a diversity (Rx only)
antenna.
These antennas are PIFAs. There is a single large ground plane that supports the
circuit board and touch screen, and this is the ground plane for all the antennas. It
is important to note that even though the FCC report labels the specific regions as
the antennas, the entire ground plane (i.e. the entire phone) makes up the antenna.
That is - if you cut away the ground plane the phone would not radiate well at the
lowband 850 MHz (where the half-wavelength is 6" or 17 cm).
In addition, the SAR report for this phone gives a very low value for peak SAR,
equal to 0.402 W/kg averaged over 1g of tissue (the FCC limit to sell a phone in
the US is 1.6 W/kg). This is an advantageous property of PIFAs: since the
radiation is away from the ground plane (towards the rear of the phone), the
energy is directed away from the head, giving a low value for SAR.
Reflector Antennas
Corner Reflector
The radiation pattern of this antenna can be understood by using image theory,
and then calculating the result via array theory. For ease of analysis, we'll assume
the reflecting plates are infinite in extent. Figure 2 below shows the equivalent
source distribution, valid for the region in front of the plates.
The dotted circles indicate antennas that are in-phase with the actual antenna; the
x'd out antennas are 180 degrees out of phase to the actual antenna.
The above directly follows from Figure 2 and array theory (k is the wave number.
The resulting pattern will have the same polarization as the original vertically
polarized antenna. The directivity will be increased by 9-12 dB. The above
equation gives the radiated fields in the region in front of the plates. Since we
assumed the plates were infinite, the fields behind the plates are zero.
The radiation pattern, impedance and gain of the antenna will be influenced by
the distance d of Figure 1. The input impedance is increased by the reflector
when the spacing is one half wavelength; it can be reduced by moving the
antenna closer to the reflector. The length L of the reflectors in Figure 1 are
typically 2*d. However, if tracing a ray travelling along the y-axis from the
antenna, this will be reflected if the length is at least . The height of the
plates should be taller than the radiating element; however since linear antennas
do not radiate well along the z-axis, this parameter is not critically important.
Parabolic reflectors typically have a very high gain (30-40 dB is common) and
low cross polarization. They also have a reasonable bandwidth, with the
fractional bandwidth being at least 5% on commercially available models, and
can be very wideband in the case of huge dishes (like the Stanford "big dish"
above, which can operate from 150 MHz to 1.5 GHz).
The smaller dish antennas typically operate somewhere between 2 and 28 GHz.
The large dishes can operate in the VHF region (30-300 MHz), but typically need
to be extremely large at this operating band.
Unlike resonant antennas like the dipole antenna which are typically
approximately a half-wavelength long at the frequency of operation, the
reflecting dish must be much larger than a wavelength in size. The dish is at least
several wavelengths in diameter, but the diameter can be on the order of 100
wavelengths for very high gain dishes (>50 dB gain). The distance between the
feed antenna and the reflector is typically several wavelenghts as well. This is in
contrast to the corner reflector, where the antenna is roughly a half-wavelength
from the reflector.
In the next section, we'll look at the parabolic dish geometry in detail and why a
parabola is a desired shape.
The parabola is completely described by two parameters, the diameter D and the
focal length F. We also define two auxilliary parameters, the vertical height of
the reflector (H) and the max angle between the focal point and the edge of the
dish ( ). These parameters are related to each other by the following equations:
There are two observations that can be made from Figure 5. The first is that both
rays end up travelling in the downward direction (which can be determined
because the incident and reflected angles relative to the normal of the surface
must be equal).
The rays are said to be collimated. The second important observation is that the
path lengths ADE and ABC are equal. This can be proved with a little bit of
geometry, which I won't reproduce here. These facts can be proved for any set of
angles chosen. Hence, it follows that:
All rays emanating from the focal point (the source or feed
antenna) will be reflected towards the same direction.
The distance each ray travels from the focal point to the
reflector and then to the focal plane is constant.
In the next section, we'll look at gain calculations for a parabolic reflector
antenna.
The actual gain is in terms of the effective aperture, which is related to the
physical area by the efficiency term ( ). This efficiency term will often be on
the order of 0.6-0.7 for a well designed dish antenna:
Understanding this efficiency will also aid in understanding the trade-offs
involved in the design of a parabolic reflector. The efficiency can be written as
the product of a series of terms:
Radiation Efficiency
The radiation efficiency is the usual efficiency that deals with ohmic losses, as
discussed on the efficiency page. Since horn antennas are often used as feeds, and
these have very little loss, and because the parabolic reflector is typically metallic
with a very high conductivity, this efficiency is typically close to 1 and can be
neglected.
This efficiency can be improved by increasing the F/D ratio, which also lowers
the cross-polarization of the radiated fields. However, as with all things in
engineering, there is a tradeoff: increasing the F/D ratio reduces the spillover
efficiency, discussed next.
Spillover Efficiency
The spillover efficiency is simple to understand. This measures the amount of
radiation from the feed antenna that is reflected by the reflector. Due to the finite
size of the reflector, some of the radiation from the feed antenna will travel away
from the main axis at an angle greater than , thus not being reflected. This
efficiency can be improved by moving the feed closer to the reflector, or by
increasing the size of the reflector.
Other Efficiencies
There are many other efficiencies that I've lumped into the parameter . This is
a major of all other "real-world effects" that degrades the antenna's gain and
consists of effects such as:
Calculating Efficiency
The efficiency is a function of where the feed antenna is placed (in terms of F
and D) and the feed antenna's radiation pattern. Instead of introducing complex
formulas for some of these terms, we'll make use of some results by S. Silver
back in 1949. He calculated the aperture efficiency for a class of radiation
patterns given as:
TYpically, the feed antenna (horn) will not have a pattern exactly like the above,
but can be approximated well using the function above for some value of n.
Using the above pattern, the aperture efficiency of a parabolic reflector can be
calculated. This is displayed in Figure 1 for varying values of and the F/D
ratio.
Figure 1. Aperture Efficiency of a Parabolic Reflector as a function of F/D or the
angle , for varying feed antenna radiation patterns.
In the next section, we'll look at the radiation pattern of a parabolic antenna.
Yagi-Uda Antennas
Spiral Antennas
The most popular helical antenna (helix) is a travelling wave antenna in the shape
of a corkscrew that produces radiation along the axis of the helix antenna. These
helix antennas are referred to as axial-mode helical antennas. The benefits of this
helix antenna is it has a wide bandwidth, is easily constructed, has a real input
impedance, and can produce circularly polarized fields. The basic geometry of
the helix antenna shown in Figure 1.
Figure 1. Geometry of Helical Antenna.
The antenna in Figure 1 is a left handed helix antenna, because if you curl your
fingers on your left hand around the helix your thumb would point up (also, the
waves emitted from this helix antenna are Left Hand Circularly Polarized). If the
helix antenna was wound the other way, it would be a right handed helical
antenna.
The radiation pattern will be maximum in the +z direction (along the helical axis
in Figure 1). The design of helical antennas is primarily based on empirical
results, and the fundamental equations will be presented here.
Helix antennas of at least 3 turns will have close to circular polarization in the +z
direction when the circumference C is close to a wavelength:
Once the circumference C is chosen, the inequalites above roughly determine the
operating bandwidth of the helix antenna. For instance, if C=19.68 inches (0.5
meters), then the highest frequency of operation will be given by the smallest
wavelength that fits into the above equation, or =0.75C=0.375 meters, which
corresponds to a frequency of 800 MHz. The lowest frequency of operation will
be given by the largest wavelength that fits into the above equation, or
=1.333C=0.667 meters, which corresponds to a frequency of 450 MHz. Hence,
the fractional BW is 56%, which is true of axial helical antennas in general.
The helix antenna is a travelling wave antenna, which means the current travels
along the antenna and the phase varies continuously. In addition, the input
impedance is primarly real and can be approximated in Ohms by:
The helix antenna functions well for pitch angles ( ) between 12 and 14
degrees. Typically, the pitch angle is taken as 13 degrees.
The normalized radiation pattern for the E-field components are given by:
For circular polarization, the orthogonal components of the E-field must be 90
degrees out of phase. This occurs in directions near the axis (z-axis in Figure 1)
of the helix. The axial ratio for helix antennas decreases as the number of loops N
is added, and can be approximated by:
In the above, c is the speed of light. Note that for a given helix geometry
(specified in terms of C, S, N), the gain increases with frequency. For an N=10
turn helix, that has a 0.5 meter circumference as above, and an pitch angle of 13
degrees (giving S=0.13 meters), the gain is 8.3 (9.2 dB).
For the same example helix antenna, the pattern is shown in Figure 2.
Yagi-Uda Antenna
The Yagi-Uda antenna or Yagi Antenna is one of the most brilliant antenna
designs. It is simple to construct and has a high gain, typically greater than 10
dB. The Yagi-Uda antennas typically operate in the HF to UHF bands (about 3
MHz to 3 GHz), although their bandwidth is typically small, on the order of a
few percent of the center frequency. You are probably familiar with this antenna,
as they sit on top of roofs everywhere. An example of a Yagi-Uda antenna is
shown below.
The Yagi antenna was invented in Japan, with results first published in 1926. The
work was originally done by Shintaro Uda, but published in Japanese. The work
was presented for the first time in English by Yagi (who was either Uda's
professor or colleague, my sources are conflicting), who went to America and
gave the first English talks on the antenna, which led to its widespread use.
Hence, even though the antenna is often called a Yagi antenna, Uda probably
invented it. A picture of Professor Yagi with a Yagi-Uda antenna is shown below.
In the next section, we'll explain the principles of the Yagi-Uda antenna.
The element to the left of the feed element in Figure 1 is the reflector. The length
of this element is given as R and the distance between the feed and the reflector is
SR. The reflector element is typically slightly longer than the feed element. There
is typically only one reflector; adding more reflectors improves performance very
slightly. This element is important in determining the front-to-back ratio of the
antenna.
Having the reflector slightly longer than resonant serves two purposes. The first
is that the larger the element is, the better of a physical reflector it becomes.
Secondly, if the reflector is longer than its resonant length, the impedance of the
reflector will be inductive. Hence, the current on the reflector lags the voltage
induced on the reflector. The director elements (those to the right of the feed in
Figure 1) will be shorter than resonant, making them capacitive, so that the
current leads the voltage. This will cause a phase distribution to occur across the
elements, simulating the phase progression of a plane wave across the array of
elements. This leads to the array being designated as a travelling wave antenna.
By choosing the lengths in this manner, the Yagi-Uda antenna becomes an end-
fire array - the radiation is along the +y-axis as shown in Figure 1.
The rest of the elements (those to the right of the feed antenna as shown in Figure
1) are known as director elements. There can be any number of directors N,
which is typically anywhere from N=1 to N=20 directors. Each element is of
length Di, and separated from the adjacent director by a length SDi. As alluded to
in the previous paragraph, the lengths of the directors are typically less than the
resonant length, which encourages wave propagation in the direction of the
directors.
The above description is the basic idea of what is going on with the Yagi-Uda
antenna. Yagi antenna design is done most often via measurements, and
sometimes computer simulations. For instance, let's look at a two-element Yagi
antenna (1 reflector, 1 feed element, 0 directors). The feed element is a half-
wavelength dipole, shortened to be resonant (gain = 2.15 dB). The gain as a
function of the separation is shown in Figure 2.
The above graph shows that the gain is increased by about 2.5 dB if the
separation SD is between 0.15 and 0.3 wavelengths. Similarly, the gain for this
Yagi antenna can be plotted as a function of director spacings, or as a function of
the number of directors used. Typically, the first director will add approximately
3 dB of overall gain (if designed well), the second will add about 2 dB, the third
about 1.5 dB. Adding an additional director always increases the gain; however,
the gain in directivity decreases as the number of elements gets larger. For
instance, if there are 8 directors, and another director is added, the increases in
gain will be less than 0.5 dB.
In the next section on Yagis, I'll go further into the design of Yagi-Uda antennas.
As an example, consider the table published in "Yagi Antenna Design" by P Viezbicke from the
National Bureau of Standards, 1968, given in Table I. Note that the "boom" is the long element tha
the directors, reflectors and feed elements are physically attached to, and dictates the lenght of the
antenna.
Table I. Optimal Lengths for Yagi-Uda Elements, for Distinct Boom Lengths
d=0.0085
Boom Length of Yagi-Uda Array (in )
SR=0.2
There's no real rocket science going on in the above table. I believe the authors of
the above document did experimental measurements until they found an
optimized set of spacings and published it. The spacing between the directors is
uniform and given in the second-to-last row of the table. The diameter of the
elements is given by d=0.0085 . The above table gives a good starting point to
estimate the required length of the antenna (the boom length), and a set of lengths
and spacings that achieves the specified gain. In general, all the spacings, lengths,
diamters (including the boom diameter) are design variables and can be
continuously optimized to alter performance. There are thousands of tables that
further give results, such as how the diamter of the boom affects the results, and
the optimal diamters of the elements.
Spiral Antennas
Spiral antennas belong to the class of "frequency independent" antennas; those
with a very large bandwidth. The fractional Bandwidth can be as high as 30:1.
This means that if the lower frequency is 1 GHz, the antenna could still be in
band at 30 GHz, and every frequency in between.
Spiral antennas are usually circularly polarized. The spiral antenna's radiation
pattern typically has a peak radiation direction perpendicular to the plane of the
spiral (broadside radiation). The Half-Power Beamwidth (HPBW) is
approximately 70-90 degrees.
Spiral antennas are widely used in the defense industry for sensing applications,
where very wideband antennas that do not take up much space are needed. Spiral
antenna arrays are used in military aircraft in the 1-18 GHz range. Other
applications of spiral antennas include GPS, where it is advantageous to have
RHCP (right hand circularly polarized) antennas.
[Equation 1]
In Equation [1], is a constant that controls the initial radius of the spiral
antenna. The parameter a controls the rate at which the spiral antenna flares or
grows as it turns. Equation [1], in English, states that the spiral antenna radius
grows exponentially as it turns. In Figure 1, a plot of a planar Log-Periodic Spiral
Antenna is shown.
The planar spiral antenna of Figure 1 will have peak radiation directions into and
out of the screen (broadside to the plane of the spiral, in both the front and the
back). The spiral antenna of Figure 1 will radiate Right Hand Circularly
Polarized (RHCP) fields out of the screen, and Left Hand Circularly Polarized
(LHCP) fields into the screen. The sense of the circularly polarized fields can be
determined by placing your thumb in the direction of the fields, and curling your
fingers in the direction of the spiral antenna (If your fingers curl the right way
using your right hand, then it is RHCP. Otherwise, it is LHCP).
The parameters that effect the radiation of the spiral antenna include:
2. The Flare Rate (a) - The rate at which the spiral grows with angle is the flare
rate. If it is too small, the spiral is tightly wrapped around itself. In this case, it
will behave more like a capacitor, with closely coupled conductors, giving poor
radiation. If the flare rate is too small, the spiral acts more like a dipole as it
doesn't wrap around itself. A commonly used value is a = 0.22.
3. Feed Structure - The feed must be controlled with a balun so that the spiral has
balanced currents on either arm. A commonly used balun for spiral antennas is
the infinite balun. More importantly, the feed structure determines the high end of
the operating band. How tightly you can wrap the spiral in on itself determines
how small the wavelength can be that will fit on your spiral and still maintain
spiral antenna operation. The highest frequency in the spiral antenna's operating
band occurs when the innermost radius of the spiral (i.e. where the spiral starts
after the feed structure) is equal to lambda/4 (one quarter wavelength). That is,
the highest frequency can be determined from the inner radius ( <="" i=""> in
Equation [1]):
[Equation 3]
4. Number of Turns (N) - The number of turns of the spiral is also a design
parameter. Experimentally it is found that spirals with at least one-half turn up to
3 turns work well, with 1.5 turns being a good number.
Radiation occurs from the spiral antenna when the currents on the spiral's arms
are in phase. As the spiral winds outward from the center, there will exist some
region for each frequency (wavelength) where the currents add constructively
and produce radiation. This radiation removes energy from the electric current on
the spiral antenna; as a result, the magnitude of the current dies off with distance
from the spiral antenna. Hence, there is little reflection of current from the end of
the spiral antenna. How quickly the current decreases in magnitude away from
the center of the spiral is a function of the geometry of the spiral antenna.
To furthur reduce reflection from the end of the spiral (which will decrease
overall antenna efficiency and bandwidth), sometimes resistive loads are applied
to the end of spirals to keep current from reflecting at the end of the spiral arms.
[Eq
That is, the Log-Periodic Spiral Antenna has a theoretical impedance of about
188 Ohms. In actual realizations of Spiral antennas the impedance tends to be
less than this, in the 100-150 Ohm range.
Radiation Patterns
The radiation pattern of the Log-Periodic Spiral Antenna is approximately given
by:
[Eq
This pattern has two equal radiation peaks, both broadside to the plane of the
spiral antenna (which lies in the z=0 plane, or x-y plane). One peak is above the
planar spiral antenna and the other is below. The Spiral Antenna has circular
polarization over a wide beamwdith, often for angular regions as wide as
[Equation 6]
Equation [6] states that the radius r of the antenna increases linearly with the
angle . The parameter a is simply a constant that controls the rate at which the
spiral flares out. The second arm of the Archimedean spiral the same as the first,
but rotated 180 degrees. A plot of the Archimedean spiral given in Equation [6] is
shown in Figure 2:
In Figure 2, we have two arms of the Archimedean spiral antenna flaring away
from the center, as defined by Equation [6]. The feed of the antenna (the voltage
source), is placed directly across between the two arms of the spiral - the positive
end to one arm and the negative end of the feed to the second spiral arm.
Cavity-Backed Slot Antennas
The slot antennas described previously have been planar slot antennas. As
discussed, they are commonly used on aircraft or other metallic backed objects.
As such, it is desirable to design the slot antenna to be cavity-backed by some
metal. This isolates the spiral antenna from what is behind it, so that it can be
mounted on objects without worrying about retuning the antenna.
Cavity-backed slots come in two types. The first is a simple metallic backing
separated from the spiral by some distance or depth, d. This metallic backing will
causes reflections of the radiated fields that enter the cavity. As such, they can
often cancel the fields travelling broadside to the plane of the spiral antenna. If
the depth d is small relative to a wavelength, or an integer multiple of a half-
wavelength, the reflected field will tend to cancel the forward travelling field of
the spiral antenna, thus leading to poor radiation. Hence, doing a metallic cavity-
backed spiral will work, but will decrease the wideband characteristics of the
spiral antenna.
The VSWR in Figure 3 isn't particularly well matched (a little less than 50%
mismatch loss across the band). However, the VSWR is farely constant, which is
a desirable property of broadband antennas. The VSWR for this antenna could be
improved via optimization of the feed structure, increasing the number of turns
and optimizing the shape (how quickly the spiral winds, or using the Log-
Periodic Spiral Shape). However, this is pretty good for a 5-minute antenna. The
antenna efficiency is shown in Figure 4:
Figure 4. Efficiency of Constructed Spiral Antenna.
The efficiency remains fairly constant, at about -5 dB across the band. With the
mismatch loss being about -3 dB, there appears to be about 2 dB of loss
associated with the structure itself. Hence, there is a good amount of optimization
that can occur to improve the performance of the spiral antenna. However, Figure
4 shows that this antenna is a decent radiator.
Aperture Antennas
Slot Antenna
Inverted-F Antenna
Slot Antennas
Slot antennas are used typically at frequencies between 300 MHz and 24 GHz.
The slot antenna is popular because they can be cut out of whatever surface they
are to be mounted on, and have radiation patterns that are roughly
omnidirectional (similar to a linear wire antenna, as we'll see). The polarization
of the slot antenna is linear. The slot size, shape and what is behind it (the cavity)
offer design variables that can be used to tune performance.
To gain an intuition about slot antennas, first we'll learn Babinet's principle (put
into antenna terms by H. G. Booker in 1946). This principle relates the radiated
fields and impedance of an aperture or slot antenna to that of the field of its dual
antenna. The dual of a slot antenna would be if the conductive material and air
were interchanged - that is, the slot antenna became a metal slab in space. An
example of dual antennas is shown in Figure 2:
Figure 2. Dual antennas - (left) the slot antenna, (right) the dipole antenna.
Note that a voltage source is applied across the short end of the slot antenna. This
induces an E-field distribution within the slot, and currents that travel around the
slot perimeter, both contributed to radiation. The dual antenna is similar to a
dipole antenna. The voltage source is applied at the center of the dipole, so that
the voltage source is rotated.
Babinet's principle relates these two antennas. The first result states that the
impedance of the slot antenna ( ) is related to the impedance of its dual antenna
( ) by the relation:
In the above, is the intrinsic impedance of free space. The second major result
of Babinet's/Booker's principle is that the fields of the dual antenna are almost the
same as the slot antenna (the fields components are interchanged, and called
"duals"). That is, the fields of the slot antenna (given with a subscript S) are
related to the fields of it's complement (given with a subscript C) by:
Hence, if we know the fields from one antenna we know the fields of the other
antenna. Hence, since it is easy to visualize the fields from a dipole antenna, the
fields and impedance from a slot antenna can become intuitive if Babinet's
principle is understood.
Note that the polarization of the two antennas are reversed. That is, since the
dipole antenna on the right in Figure 2 is vertically polarized, the slot antenna on
the left will be horizontally polarized.
Duality Example
What are the fields from a slot at 1 GHz, with the same dimensions as the dipole?
The impedance of the slot for this case is much larger, and while the dipole's
impedance is inductive (positive imaginary part), the slot's impedance is
capacitive (negative imaginary part). The E-fields for the slot can be easily
found:
We see that the E-fields only contain a phi (azimuth) component; the slot antenna
is therefore horizontally polarized.
The basic cavity-backed slot antenna is shown in Figure 1 (in a rectangular cube
of size A*B*C). The walls are metallic (electrically conducting), and the inside is
hollow. On one end, a slot is cut out. The cavity is typically excited by a probe
antenna in the intererior of the cavity, which typically is modelled as a monopole
antenna. The exciting monopole antenna is shown in green.
Figure 1. Cavity-backed slot antenna.
I'll give some experimental results for this antenna. Let the height of the cavity
C=36mm, the length be A=87mm and the height B=36mm. The height of the
monopole antenna will be 29.5mm, so that the monopole is a quarter-wavelength
long at 2.55 GHz. The monopole will be centered about the cavity in the y-
direction, and 61.5mm behind the slot in the x-direction. The slot is 58mm long
(in the y-direction) and 3 mm high (in the z-direction).
S11 is measured for this antenna (relative to a 50 Ohm source), and is plotted
versus frequency in Figure 2.
Figure 2. S11 as a function of Frequency for the Cavity-backed Slot Antenna.
This cavity-backed slot antenna has a first resonance at about 2.45 GHz. At this
frequency, the cavity backed slot antenna is roughly 0.474 wavelengths long -
which is roughly the length of a resonant dipole antenna. S11 drops to below -20
dB at this frequency, indicating that most of the power is radiated away. The
bandwidth, measured (somewhat arbitrarily) as the frequency span that S11 is
less than -6 dB is roughly from 2.35 GHz to 2.55 GHz, giving a fractional
bandwidth of slightly over 8%.
Note that two other dips ('resonances') in the S11 curve occur, at approximately 3
GHz and 4.18 GHz. At these frequencies, the slot length is 0.58 and 0.81
wavelengths, respectively.
The volume of the cavity typically influences the bandwidth - a larger volume
often yields a higher bandwidth. The material within the cavity (so far I have
assumed it was filled with air), can be replaced with a dielectric medium. This
reduces the resonant length of slot, allowing for a smaller antenna. The tradeoff is
that the bandwidth and efficiency typically decrease with a dielectric cavity
medium.
The radiation pattern at 2.45 GHz is now presented. The H-plane (xy plane) is
shown on the in Figure 3, and the E-plane (xz plane) is shown in Figure 4.
Figure 3. H-plane (xy plane). Angle measured off x-axis towards y-axis.
Figure 4. E-plane (xz plane). Angle measured off z-axis (to x-axis).
The radiation pattern of the cavity backed slot antenna somewhat resembles that
of a dipole antena in the forward H-plane. The 3-dB beamwidth is roughly 60
degrees in this plane. The radiation pattern is diminished in the rear H-plane, with
a significant back lobe about 6 dB down from the peak of the main beam. In the
E-plane, the pattern is fairly broad, with a 3-dB beamwidht of about 120 degrees.
The broad pattern of these antennas make them well suited for use in antenna
arrays. The peak gain of a thin slot is usually around 2-3 dB.
The feed is placed from the ground plane to the upper arm of the IFA. The upper
arm of the IFA has a length that is roughly a quarter of a wavelength. To the left
of the feed (as shown in Figure 1), the upper arm is shorted to the ground plane.
The feed is closer to the shorting pin than to the open end of the upper arm. The
polarization of this antenna is vertical, and the radiation pattern is roughly donut
shaped, with the axis of the donut in the vertical direction. The ground plane
should be at least as wide as the IFA length (L), and the ground plane should be
at least lambda/4 in height. If the height of the ground plane is smaller, the
bandwidth and efficiency will decrease. The height of the IFA (H), should be a
small fraction of a wavelength. The radiation properties and impedance are not a
strong function of this parameter (H).
Because the structure somewhat resembles an inverted F, this antenna takes the
name "Inverted F Antenna".
Analysis
Why does this structure radiate? Lets go back and look at the slot antenna, shown
in Figure 2.
The slot antenna should be a half-wavelength long for proper radiation (more
generally, the perimeter of the slot antenna should be roughly one wavelength
long). The way this antenna is fed (or excited by the voltage source), the voltage
at the ends of the slot (across the aperture) must be zero because of the shorting
posts on either side. If the voltage is zero at the edges of the slot, then the voltage
will be at a maximum a quarter-wavelength away (at the center of the slot).
Now, where is the current at a maximum? Since this antenna can also be viewed
as a transmission line, the source basically "sees" a short circuited transmission
line in either direction. We know from transmission line theory that when a
transmission line is short-circuited, the voltage and current are 90 degrees out of
phase. As a result, the current will be zero at the center of the slot antenna, and
will be maximum on the edges. The voltage and current distributions are shown
in Figure 3 (note the peak voltage is assumed to be P volts, and the peak current
is A Amps).
The slot antenna radiates because the voltage is in-phase across the entire
aperture, so that the E-field is vertical and lines up everywhere along the slot.
This also gives rise to the vertical polarization.
How does this relate to the IFA? Here is the key point: If the current at the center
of the slot is zero (as shown in Figure 3), then the slot antenna can be thought of
as having an open circuit at the center of the slot. Hence, if we break the slot in
half, and get rid of the right side, we are left with the IFA antenna as shown in
Figure 1.
Note that the IFA can support the exact same mode of radiation. That is, since the
IFA has an open circuit on the right side of the feed (Figure 1), the current will be
zero at that point and the voltage will be a maximum - exactly as in the slot
antenna case. Hence, the IFA can be viewed as "half a slot antenna". And indeed,
this is a valid model for the antenna. Hence, the IFA is classified as an aperture
antenna, even though the aperture is not "closed".
Note that for the slot antenna of Figure 2, the short-circuit to the right of the feed
still creates a capacitance, because the length of the slot to the right of the feed is
larger than a quarter-wavelength.
In regards to the real part of the impedance for an IFA or slot antenna, note that if
the slot antenna was fed at the center of the slot (where the voltage is maximum
and the current is zero --- Z=V/I), the impedance would be practically infinite, so
that the antenna would not radiate. By moving the feed away from the center for
a slot, this also allows the current to decrease from zero, which allows the
impedance to drop to a more desirable value. The same is true on the IFA
antenna. Hence, the feed location is a critical factor in designing an IFA or slot. A
good location for the feed can easily be obtained experimentally during an
antenna design.
The yellow strip on the left side of the Palm Antenna is the GPS antenna, which
is an IFA. Since the GPS frequency is 1.575 GHz, a quarter wavelength is about
1.87 inches (4.75 cm). This is roughly the length of the IFA in the actual product
(it is shortened for proper tuning).
Two other antennas are visible in Figure 5. In the upper right side, there is the
diversity cell antenna, which is a receive only antenna (does not transmit). At the
bottom of the device is a dual band IFA. This antenna is the transmit/receive cell
antenna, which should operate at the 900 MHz and 1800 MHz bands. To do this,
the antenna engineers made an IFA for the high band (1800 MHz), which is the
shorter arm shown in Figure 5. Using the same feed and shorting pin, they
branched off another arm for the low band (the longer arm, that is wrapped
around itself somewhat). Because the designers had limited space (which is a big
challenge in mobile phone antenna development), they wrapped the IFA around
itself on the edges. By doing this, they were able to obtain the required length for
a 900 MHz IFA. However, by wrapping the antenna around itself, the bandwidth
and radiation efficiency decrease.
In summary, the IFA antenna is useful because of its small size and easy
construction. In the next section, we'll look at slotted waveguide antennas.
The slots on the waveguide will assumed to have a narrow width. Increasing the
width increases the Bandwidth (recall that a fatter antenna often has an increased
bandwidth); the expense of a larger width is a higher degree of cross-polarization.
The Fractional Bandwidth for thin slots can be as low as 3-5%; wide slots can
have a FBW on the order of 75%. An example of a slotted waveguide array is
shown in Figure 1 (dimensions given by length a and width b)
Figure 1. Basic geometry of a slotted waveguide antenna.
As in the cavity-backed slot antenna, each slot could be independently fed with a
voltage source across the slot. However, (especially for large arrays) this would
be very difficult to construct, and I've never seen this done in practice. Instead,
the waveguide is used as the transmission line to feed the elements.
The position, shape and orientation of the slots will determine how (or if) they
radiate. In addition, the shape of the waveguide and frequency of operation will
play a major role. To understand what is going on, we'll need to understand the
fields within the waveguide first. For a primer on waveguides, see here:
waveguide primer.
The dominant TE10 mode will be assumed to exist within the waveguide. Using
the geometry of Figure 1, the fields that exist within the waveguide are given by:
In the above, f is the frequency of interest, k is the wavenumber and is a
constant that specifies how much power is added to the waveguide.
On the top wall of the waveguide (where the slots are), the induced currents will
be:
Radiation occurs when the currents must "go around" the slots in order to
continue on their desired direction. As an example, consider a narrow slot in the
center of the waveguide, as shown in Figure 2.
Figure 2. Waveguide with a thin slot centered about its width.
In this case, the z-component of the current will not be disturbed, because the slot
is thin and the z-current would not need to travel around the slot. Hence, the x-
component of the current will be responsible for the radiation. However, at this
location (x=a/2), the x-component of the current density is zero - i.e. no current
and therefore no radiation. As a result, slots can not be placed in the center of the
waveguide as shown in Figure 2.
If the slots are displaced from the centerline as shown in Figure 1, the x-directed
current will not be zero and will need to travel around the slot. Hence, radiation
will occur. Note that the distance from the edge will determine the magnitude of
the current. As a result, the power that the slot radiates can be altered by moving
the slots closer or farther from the edge. In this manner, a phased array can be
designed with varying excitation to each element.
If the slot is oriented as shown in Figure 3, the slot will disturb the z-component
of the current density. This slot will then radiate. If this slot is displaced away
from the center line, the amount of power that it radiates can be adjusted.
Figure 3. Horizontal slot in a waveguide.
If the slot is rotated at an angle about the centerline as shown in Figure 4, it will
radiate. The power it radiates will be a function of the angle (phi) that it is rotated
- specifically given by . Note that the z-component of the current is still
responsible for radiation in this case. The x-component is disturbed; however the
currents will have opposite magnitudes on either side of the centerline and will
thus tend to cancel out the radiation.
The front end (the open face at the y=0 in the x-z plane) is where the antenna is
fed. The far end is usually shorted (enclosed in metal). The waveguide may be
excited by a short dipole (as seen on the cavity-backed slot antenna) page, or by
another waveguide.
To begin to analyze the antenna of Figure 1, lets view the circuit model. The
waveguide itself acts as a transmission line, and the slots in the waveguide can be
viewed as parallel (shunt) admittances. The end of the waveguide is short
circuited, so a rough circuit model of Figure 1 is:
Before we discuss choosing the sizes, they will be given in terms of the guide-
wavelength, which is the wavelength within the waveguide. The guide
wavelength ( ) is a function of the width of the waveguide (a) and the free
space wavelength. For the dominant TE01 mode, the guide wavelength is given
by:
The distance between the last slot and the end d is often chosen to be a quarter-
wavelength. Transmission line theory states that the impedance of a short circuit
a quarter-wavelength down a transmission line is an open circuit. Hence, Figure 2
then reduces to:
The 3D radiation pattern for the slotted waveguide is shown in the following
figure (it was calculated using a numerical electromagnetics package called
FEKO). That gain is approximately 17 dB.
Note that in the x-z plane (or h-plane), the beamwidth is very narrow (2-5
degrees). In the y-z plane (or e-plane), the beamwidth is much larger. Obtaining a
more pencil-type beam using slotted waveguides is discussed in the next section.
Antenna arrays made up of hundreds or even thousand elements are often slotted
waveguides similar to those described above. These are typically narrowband (a
small deviation away from the design frequency often change the impedance of
the individual slots, and the many slots add up producing a highly reactive
impedance associated with the waveguide away from the resonant frequency). As
an example, consider Boeing's wedgetail:
While I don't know for certain, I would guess that the walls of the odd structure
on top of the above airplane contain a large slotted waveguide array, for scanning
in the horizontal (azimuth) plane around the airplane.
The maximum power is shown in Table I. The max power is a function of the
altitude, and will be further divided as CW or "continuous wave" and a 1
microsecond pulse. The CW column gives the maximum power handling
capability of a slotted waveguide when the array is continuosly transmitting in
MegaWatts [MW]. The pulse column gives the maximum power when the
waveguide radiates a brief pulse and then shuts off. Note that the power handling
capabilities of a waveguide decrease with altitude (given in feet). The results are
given for a typical X-band (10 GHz) waveguide; note that the power capabilities
increases as the frequency decreases and vice-versa.
Table I. Power versus altitude for a typical X-band (10 GHz) waveguide.
Altitude [ft] Max Power (CW) [MW] Max Power (1 uS Pulse) [MW]
0 0.95 1.05
10,000 0.55 0.62
20,000 0.25 0.30
30,000 0.15 0.20
40,000 0.06 0.10
50,000 0.03 0.05
Note that the power decreases rapidly with altitude; this is a necessary constraint
to keep in mind in designing radars for aircraft.
The gain of horn antennas often increases (and the beamwidth decreases) as the
frequency of operation is increased. This is because the size of the horn aperture
is always measured in wavelengths; at higher frequencies the horn antenna is
"electrically larger"; this is because a higher frequency has a smaller wavelength.
Since the horn antenna has a fixed physical size (say a square aperture of 20 cm
across, for instance), the aperture is more wavelengths across at higher
frequencies. And, a recurring theme in antenna theory is that larger antennas (in
terms of wavelengths in size) have higher directivities.
Horn antennas have very little loss, so the directivity of a horn is roughly equal to
its gain.
Another example of a horn antenna is the H-plane horn, shown in Figure 2. This
horn is flared in the H-plane, with a constant height for the waveguide and horn
of h.
The most popular horn antenna is flared in both planes as shown in Figure 3. This
is a pyramidal horn, and has a width B and height A at the end of the horn.
Figure 3. Pyramidal horn antenna.
Instead of the traditional academic derivation approach, I'll state some results for
the horn antenna and show some typical radiation patterns, and attempt to
provide a feel for the design parameters of horn antennas. Since the pyramidal
horn antenna is the most popular, we'll analyze that. The E-field distribution
across the aperture of the horn antenna is what is responsible for the radiation.
The radiation pattern of a horn antenna will depend on B and A (the dimensions
of the horn at the opening) and R (the length of the horn, which also affects the
flare angles of the horn), along with b and a (the dimensions of the waveguide).
These parameters are optimized in order to taylor the performance of the horn
antenna, and are illustrated in the following Figures.
Observe that the flare angles ( and ) depend on the height, width and
length of the horn antenna.
Given the coordinate system of Figure 6 (which is centered at the opening of the
horn), the radiation will be maximum in the +z-direction (out of the screen).
The E-field distribution across the opening of the horn antenna can be
approximated by:
The E-field in the far-field will be linearly polarized, and the magnitude will be
given by:
The above equation states that the far-fields of the horn antenna is the Fourier
Transform of the fields at the opening of the horn. Many textbooks evaluate this
integral, and end up with supremely complicated functions, that I don't feel sheds
a whole lot of light on the patterns.
In the next section, we'll look at the radiation patterns for horn antennas.
This antenna is simulated using a commercial solver, FEKO (which runs method
of moments). The radiation pattern at 2 GHz is shown in Figure 2.
Figure 2. Horn radiation pattern at 2 GHz.
The gain of the horn is 18.1 dB in the +z-direction. The half-power beamwidth is
15 degrees in the xz-plane (H-plane) and 11 degrees in the yz-plane (E-plane).
The gain at 1.5 GHz is -2.54 dB, approximately 20 dB lower than at 2 GHz. The
waveguide feed acts as a high-pass filter; it blocks energy below its 'cutoff'
frequency and passes energy above this level. At 2.5 GHz, the gain increases
slightly to 18.8 dB.
The horn antenna geometry affects its antenna gain. For a desired antenna gain,
there are tables and graphs that can be consulted in antenna handbooks that
describe the optimal geometry in terms of the length and aperture size of the
horn. However, this optimal geometry is only valid at a single frequency. Since
horn antennas are to operate over a wide frequency band, they are often designed
to have optimal gain at the lowest frequency in the band. At higher frequencies,
the geometry is no longer optimal, so the E-field across the aperture is not
optimal. However, the horn's aperture becomes electrically larger at higher
frequencies (the aperture is more wavelengths long as the frequency increases or
the wavelength decreases). Consequently, the loss of an optimal aperture field is
offset by an electrically larger horn, and the antenn again actually increases as the
frequency increases.
See this link for calculating the radiation pattern of horn antennas from the
Fourier Transform.
Vivaldi Antennas
Vivaldi antennas are simple planar antennas that are very broadband. The
polarization is linear, and the basic antenna structure is shown in Figure 1:
In Figure 1, we have the antenna feed connecting two symmetric sides of a planar
metallic antenna. To the left of the feed is is a short-circuit. However, antennas
are RF-type devices and therefore the short-circuit acts more like a parallel
inductor. To the right of the feed is the radiating element. It can be considered a
tapered slot antenna or an aperture antenna.
To start, let's just take a square area of copper tape, as shown in Figure 2. The
length (horizontal dimension) is 84 mm, and the height (vertical dimension) is 54
mm. This is just a flat sheet of copper tape sitting on a piece of FR4:
Now, to start, I'm going to cut a slot out of the slab in Figure 2. The slot will be
about 80mm long, and about 5-10mm wide, as shown in Figure 3:
Figure 3. Cutting a Slot out of a Rectangular Slab of Copper.
This slot is not an antenna yet, because there is no feed. So I grab a standard
coaxial cable with an SMA connection and solder it about 38mm from the end of
the slot, as shown in Figure 4:
Figure 4. Adding the Feed to Our Antenna.
In Figure 4, I've soldered the center conductor of the coaxial cable to one side of
the slot, and the ground (shield or outside) of the cable to the other side. I also
solder the cable along the length to the antenna structure. This keeps the cable
itself from being a separate radiator - since it is part of the antenna structure the
electric currents don't care if they are flowing on the cable or the antenna. This is
similar to a balun.
In Figure 5, we see that our antenna, which is pretty simple, already has a few
resonances. When I say resonance here, I mean a region where the VSWR dips
and then goes back up. The reason I call this is a resonance, is that there is no
loss in my circuit - no matching components, resistive devices, loss materials,
etc. Hence, if the VSWR drops, then energy is probably being radiated away (it
is, but I'll get to that later). From Figure 5, we see that we have 3 frequency bands
where our antenna acts somewhat like an antenna:
around 1 GHz
From 2.5-4 GHz
At about 5.8 GHz
This is interesting in my opinion. All we really did was feed a metallic structure,
and we get a bunch of radiation. This is pretty cool, and it shows that nature
wants things to radiate. From Maxwell's Equations, we know that if we can just
get electric currents or voltage to add in phase, we will have radiation. And that's
cool.
Our antenna is a little bit like an IFA (Inverted-F Antenna) at this point, a little bit
like a slot antenna, and also a bit like a dipole antenna. But never mind too much
analysis right now. Let's say we shortened the slot of our antenna, so that the feed
is now 18mm from the left edge, as shown in Figure 6:
The 1 GHz resonance did not shift - however, it did get deeper, which indicates
the antenna has a better impedance match with a shorter slot. Shortening the slot
has the effect of decreasing the shunt inductance that is the short circuit to the left
of the feed. In Figure 6, the antenna resembles a dipole antenna that has the short
circuit to the left (the inductive path). Hence, by shortening the slot, I basically
improved the impedance matching of the dipole antenna mode - and I also now
know that the dipole antenna mode occurs at 1 GHz, and the slot/ifa antenna
mode occurs at about 3.5 GHz.
The higher resonance had a slight downshift, but really it appears that the
resonance got broader as well. Hence I suspect this is also a slot antenna mode,
but this relies more on the slot to the right of the feed, as the very large
inductance to the left of the feed basically doesn't matter at 6 GHz (this is
because the impedance of an inductor is very high and basically an open circuit
for high frequencies).
The preceeding 3 paragraphs is exactly how antenna engineers think. Now, the
morons in the university would spend all year trying to think up a crappy
equation - and they would get it wrong. Not only that, these people have never
put an antenna in a product, so don't have a clue. If you can make a change to an
antenna, and observe the changes in resonances and explain them, then you
understand the antenna. If not you might as well work in the defense industry
doing nothing with your life.
Now, back to the design. We know that more volume for antennas generally
means more bandwidth. However, we have a little too much ground plane here -
and that means a lot of capacitance. And capacitance kills your bandwidth and
isn't good for radiation. So right now, we'll taper the slot, or flare the aperture as
shown in Figure 8:
The VSWR of this antenna is measured and plotted with the other curves in
Figure 9:
Figure 9. VSWR of the Tapered Slot Antenna (in Fig 8).
In Figure 9, the green curve is the VSWR of the tapered slot antenna of Figure 8.
We see some nice things happened - the two resonances at 3.5 GHz and 5.8 GHz
start to blend together, giving a very large bandwidth. In addition, the lowband (1
GHz) resonance also became broader band and better matched, as you can see
from the lower and broader VSWR. This means our tapering of the slot made
things much better from a radiation perspective.
Since this was a pleasant change, let's taper both sides as shown in Figure 10:
Figure 10. Dual Tapered Slot - The Vivaldi Antenna.
The VSWR of this antenna is plotted in Figure 11, with the others:
Figure 11. VSWR of the Vivaldi Antenna (in red).
Figure 11 shows that the overall bandwidth of the Vivaldi antenna increased (the
impedance matching was better over a wider range). We also see the lowband
resonance increase in frequency slightly (i.e. the 1 GHz resonance shifts up). This
is due to less capacitive loading (capacitive loading, or capacitance in the antenna
structure tends to shift resonances down in frequency).
We have now constructed a Vivaldi Antenna. We see that is has a low resonance
and a very broad higher frequency resonance. We suspect very strongly that this
radiates (since there is no other losses in the structure). However, to be sure, we
should measure the antenna efficiency in an anechoic chamber. This is done and
is plotted in Figure 12:
Figure 12. Antenna Efficiency for the Vivaldi.
From Figure 12, we see that the Vivaldi antenna I threw together in Figure 10 has
very broad bandwidth and very high efficiency. This shows that our resonances in
the VSWR plots are indeed due to radiation resistance - and not loss.
The direction of peak radiation is frequency dependent. For the low frequency
band of radiation, I stated that the antenna acted as a dipole type antenna. In this
case, the direction of peak radiation is directly into or out of the page as viewed
in Figure 10. For the higher frequency band of radiation (3.5-6 GHz), the antenna
is using the aperture to radiate. In this case, the direction of peak radiation is
towards the right as seen in Figure 10.
Finally, you may also see Vivaldi antenna with odd shaped slots cut out of the left
side of the feed as shown in Figure 13:
Figure 13. Vivaldi with a Loop to the Left.
The primary purpose of altering the short-circuit path to the left of the feed is for
antenna tuning purposes (i.e. impedance matching). There is no real radiation that
is added via adding the circular cut shown in Figure 13, but it does give the
antenna designer more freedom in optimizing the VSWR or tuning of the
antenna.
NFC Antennas
How do NFC antennas work? This page will discuss that in a simple manner.
We'll try to explain why NFC antennas aren't really antennas at all (like the
dipole antenna or the Horn Antenna). But first, let's give some background on
NFC.
NFC Background
NFC stands for Near Field Communications and goes by the acronym NFC. NFC
is simply a set of standards for smartphones or whatever to establish
communication with each other by bringing them into close together (typically 0-
5 centimeters). This set of standards is just like 802.11b or 802.11n for WIFI - it
sets the protocols to send and receive information. The application of NFC
include swiped proximity payments (such as google wallet for paying at
Starbucks), information exchange at small distances (for instance, touching
smartphones to share contact information), and simplified setup of devices such
as Wi-Fi or Bluetooth. Communication is also possible between an NFC device
and an unpowered NFC chip, called a tag (as in RFID tag).
NFC Antennas
So NFC is just a method of communication between two devices at short
distance. What makes the NFC antenna design simple or easy? If you know much
about antennas, the first question you might ask is what is the operating
frequency of NFC. It turns out that these devices operate around 13.56 MHz. The
corresponding wavelength is 22 meters long - this means to get a nice half-wave
dipole antenna (that radiates well) we would need a device about 11 meters in
length.
To sum up the discussion so far: NFC antennas operate at low frequency (large
wavelengths) on small devices. A consequence of this is that the radiation
efficiency of an NFC antenna will be about 0.
Therefore, NFC antennas are not really antennas, in that no one cares about
typical antenna parameters such as the radiation pattern. So how then do they
work?
You may recall from your electric circuits classes that inductors can be made to
couple together - that is, there exists mutual coupling. If the magnetic fields from
one inductor pass near another inductor, an induced current will exist within the
second inductor. This is contactless energy transfer - exactly what NFC requires.
Hence, an NFC antenna isn't an antenna at all - it is really just a big inductor. In
general, the larger the inductance of the antenna can be made, the better it will
perform. Note that this doesn't mean you can place a very small chip inductor as
your NFC antenna - the magnetic fields on these inductors are tightly wound and
don't extend much beyond the chips themselves. Rather, a good NFC antenna is
as large of a wrapped coil of wire as possible. Recall that a loop of wire around a
material gives a strong magnetic field within the loop (and generally the more
turns you have the more inductance you create). Hence, NFC antennas are often
simply loops of wire, occupying as much surface area as the device allows.
In Figures 1 and 2, the NFC reader excites current at 13.56 MHz through the
reader NFC antenna (which is really an inductor). This induces a magnetic field,
which furthur (via mutual coupling) induces an electric current in the
smartphone's NFC antenna when they are closely placed. This induced electric
current can be read, and we have communication. This is exactly what is
happening when it comes to NFC communication in practice, with the focus on
the antennas.
One final note related to NFC antenna design. We've already stated that the larger
the surface area, the better performing your NFC antenna will be. We haven't
mentioned anything about volume. Can the NFC antenna be infinitely thin? The
answer is yes - if there is no metal or conductive material around the NFC
antenna. However, the NFC antenna illustrated in Figure 1 is on the back of a
smartphone, which is metallic. If you know much about electromagnetics, you
will know that a ground plane directly beneath magnetic or electric fields will
very much degrade them. Hence, performance degrades for the NFC inductor
style antenna when placed near a metallic surface. As a result, for performance
the height of the NFC loop will need maximized. If the space is near zero, the
performance will suffer.
One small trick to get around this is to use high permeability sheets (ferrite, iron
based, whatever) between the NFC antenna and the metallic ground plane. This
serves to concentrate the magnetic fields, effectively making them think the
distance is larger between the NFC antenna and the ground plane. This helps to
alleviate the height problem, but does not eliminate it.
This concludes our discussion of NFC antennas and some of the basics in NFC
antenna design. As you can see, this is a very atypical antenna, and really I don't
classify it as an antenna at all. Peace!
To understand antenna arrays and phased arrays, navigate through the following
pages:
The positions of the elements in the phased array are illustrated in the following
Figure.
Figure 3. Geometry of an arbitrary N element antenna array.
The antennas in the phased array are spaced one-half wavelength apart (centered
at z=0). The E-field of the plane wave (assumed to have a constant amplitude
everywhere) can be written as:
In the above, k is the wave vector, which specifies the variation of the phase as a
function of position.
The (x,y) coordinates of each antenna is (0,0); only the z-coordinate changes for
each antenna. Further, assuming that the antennas are isotropic sensors, the signal
received from each antenna is proportional to the E-field at the antenna location.
Hence, for antenna i, the received signal is:
The received signals are distinct by a complex phase factor, which depends on
the antenna separations and the angle of arrival on the plane wave. If the signals
are summed together, the result is:
Suppose (as in Figure 4 here) that the signals from the elements in the antenna
array are each multiplied by a complex weight ( ) and then summed together to
form the phased array output, Y.
The output of the antenna array will vary based on the angle of arrival of an
incident plane wave (as described here). In this manner, the array itself is a
spatial filter - it filters incoming signals based on their angle of arrival. The
output Y is a function of , the arrival angle of a wave relative to the array.
In addition, if the array is transmitting, the radiation pattern will be identical in
shape to the receive pattern, due to reciprocity.
where k is the wave vector of the incident wave. The above equation can be
factor simply as:
The quantity AF is the Array Factor. The Array Factor is a function of the
positions of the antennas in the array and the weights used. By tayloring these
parameters the antenna array's performance may be optimized to achieve
desirable properties. For instance, the antenna array can be steered (change the
direction of maximum radiation or reception) by changing the weights.
Using the steering vector, the Array Factor can be written compactly as:
In the above Array Factor equation, T is the transpose operator. We'll now move
on to weighting methods (selection of the weights) used in antenna arrays, where
some of the versatility and power of antenna arrays will be shown.
Side Note: If the elements are identical (antenna array made up of all the
same type of antennas), and have the same physical orientation (all point or
face the same direction), then the radiation (or reception) pattern for an
antenna array is simply the Array Factor multiplied by the radiation pattern
. This concept is known as pattern multiplication.
Antenna Arrays
A weighting method is a means of selecting the weights that multiply the signals
from the antennas in an antenna array:
The weights used in the antenna array are fundamental in controlling the
behavior of the array. Some methods are now presented, which also serve to
explain the versatility of antenna arrays.
Phased Arrays
Dolph-Tschebysheff Weights
MMSE Weights
Antenna Arrays
A weighting method is a means of selecting the weights that multiply the signals
from the antennas in an antenna array:
The weights used in the antenna array are fundamental in controlling the
behavior of the array. Some methods are now presented, which also serve to
explain the versatility of antenna arrays.
Phased Arrays
Dolph-Tschebysheff Weights
MMSE Weights
Let's take an example. Suppose that we had a N=5 element linear array with one-
half wavelength spacing between the antennas (assumed to be isotropic for
simplicity). The positions are each given by:
To be absolutely explicit about it, the weight vector can also be written as:
In the above, k is the wave vector, v(k) is the steering vector, and N=5.
The array response is the magnitude of the output, as a function of the incident
angle of the plane wave. The array response is identical in form to the radiation
pattern of this array (when the same weights are used). This response is shown in
Figure 2.
Figure 2. Response of array versus angle of incidence of a plane wave.
Figure 2 captures the fundamental properties of the designed array. First, the
array has maximum radiation (or reception) in the direction of (img
src="thetad.jpg">), as designed. Second, the array has nulls at roughly 72, 94,
120, and 153 degrees. The null-directions are those in which the array completely
blocks signals. This array has an associated HPBW of roughly 30 degrees.
Finally, note that the sidelobes are 12 dB below the peak of the main beam.
It turns out that this isn't real hard to do, either. In general, an N element array can
place N-1 independent nulls in its radiation pattern. This just requires a little math
to work through, and will be illustrated via an example. Let's assume we have an
N element linear array with uniform spacing equal to a half-wavelength and lying
on the z-axis.
To start, the array factor of a uniformly spaced linear array with half-wavelength
spacing can be rewritten using a variable substitution as:
We've introduced variables, and have gotten rid of the weights. Hence, we can
choose the zeros to be whatever we want, and then figure out what the weights
should be to give us the same pattern.
To make the example concrete, let N=3. Suppose we want the array's radiation
pattern to have zeros at 45 and 120 degrees. We can simply use the equation for z
above, and substitute these values for the angle. We then obtain the
corresponding zeros, . The z values corresponding to the zeros are:
For simplicity, we'll let . The AF then becomes:
We already know what the are, so we automatically have the weights. Using
these weights to plot the magnitude of the array factor, we obtain the result in
Figure 1.
This method can easily be performed on linear arrays with many more elements.
This method can be The Schelkunoff Polynomial Method easily extends to planar
and multi-dimensional arrays. The simplicity of placing nulls in the radiation
pattern adds a powerful advantage for using arrays in practice.
Dolph-Chebyshev Weights
You may have noticed that the antenna array factors for arrays with uniform
weights have unequal sidelobe levels, as seen here. Often it is desirable to lower
the highest sidelobes, at the expense of raising the lower sidelobes. The optimal
sidelobe level (for a given beamwidth) will occur when the sidelobes are all
equal in magnitude.
This problem was solved by Dolph in 1946. He derives a method for obtaining
weights for uniformly spaced linear arrays steered to broadside ( =90 degrees).
This is a popular weighting method because the sidelobe level can be specified,
and the minimum possible null-null beamwidth is obtained.
Observe that the oscillations within the range [-1, 1] are all equal in magnitude.
The idea is to use these polynomials (with known coefficients) and match them
somehow to the array factor (the unknown coefficients being the weights).
To begin to see how this is achieved, lets assume we have a symmetric antenna
array - for every antenna element at location dn there is an antenna element at
location -dn, both multiplied by the same weight wn. We'll further assume the
array lies along the z-axis, is centered at z=0, and has a uniform spacing equal to
d. Then the array factor will be of the form given by:
The array is even if there are an even number of elements (no element at the
origin), or odd if there are an odd number of elements (an element at the origin).
Using the complex-exponential formula for the cosine function:
The array factors can be rewritten as:
The parameter is used to determine the sidelobe level. Suppose there are N
elements in the array, and the sidelobes are to be a level of S below the peak of
the main beam in linear units (note, that if S is given in dB (decibels), it should
be converted back to linear units SdB=20*log(S), where the log is base-10). The
the parameter can be determined simply from:
The resulting Array Factor (AF) will have the minimum null-null beamwidth for
the specified sidelobe level, and the sidelobes will all be equal in magnitude.
Dolph-Chebyshev Example
In the previous page on Dolph Weights, the Dolph-Tschebysheff method was
introduced. On this page, we'll run through an example. Consider a N=6 element
array, with a sidelobe level to be 30 dB down from the main beam (S=31.6223).
We'll assume the array has half-wavelength spacing, and recall that the Dolph-
Chebyshev method requires uniform spacing and the array to be steered towards
broadside (and yes, everyone spells Chebyshef in a different way, which is why I
keep changing the spelling).
The array has an even number of elements, so we'll write the array factor as:
Using the trigonometric identities for the cosine function, the above equation can
be rewritten as:
The above equation is valid for all values of t. Hence, the terms that multiply t
must equal, the terms that multiply t cubed must be equal, and the terms that
multiply t to the 5th power must be equal. As a result, we have 3 equations and 3
unknowns, and we can easily solve for the weights:
Note that as desired, the sidelobes are equal in magnitude and 30 dB down from
the peak of the main beam. The beamwidth obtained here (approximately 60
degrees) is the minimum possible beamwidth obtainable for the specified
sidelobe level using any weighting scheme.
This is the Dolph-Tschebysheff method. The basic math is all used in designing
digital filters that have equal-ripple filter characteristics in the pass and stop
bands. Note that if you understand digital signal processing, weight selection in
antenna arrays is simple. And if you've learned something about weighting
methods in antenna arrays, you've learned something about designing filters for
digital systems. The underlying mathematics is largely the same.
Lets start with an arbitrary array of N elements, which can be a 1-, 2-, or 3-
dimensional array of arbitrary geometry (the array does not need to be uniformly
spaced). We'll assume the desired signal is specified as s(t), with a direction and
frequency specified by its wavevector, . We'll now assume there is noise at
each antenna, specified as . The noise does not need to follow a particular
statistical distribution (i.e. Gaussian). We will write the noise at each antenna as a
vector:
In addition to noise, we'll assume that there are G interfering signals, which we
will write as , each with a corresponding wavevector , for a=1,2,...,G.
The resulting signals at each antenna can be written in vector notation using
steering vectors as:
Note that each term in the above equation is also oscillating at frequency f, but
this is not needed in the analysis. We've now defined the signals that are present
at the terminals of the N antennas in the array. We now want to determine a set of
weights that will yield the desired signal s(t) as closely as possible.
The actual output is the antenna signals X multiplied by the weights and added
together:
In the above, H is the Hermitian operator (conjugate transpose). The error in the
actual output and the desired output is simply:
In the above, the * operator represents complex conjugate. The E[] operator
represents the expected value operator (this basically returns the average value of
what is inside - with the averages weighted by which values are statistically more
likely to occur). Our goal is to find weights that minimizes the above equation.
First we'll substitute in the error definition above along with the specified terms:
YOu may have noticed we are in the middle of a somewhat long derivation. To
recap, we have an arbitrary N element array, and we're trying to find weights
minimize the MSE between the desired output of the array and the actual output
of the array. We'll continue this on the next page.
(1)
Because the weights are constant (don't vary with time), they can be pulled out of
the expectation operator, and the first term in equation (1) simplifies to:
We'll now declare some definitions using equation (1). We'll define a commonly
used matrix in signal processing - the Autocorrelation Matrix, written as ,
which is defined to be:
The second term in equation (1) is the average power in the desired signal, :
We'll define the correlation between the signals on the antennas and the desired
signal as:
The third term can be written using the above definition as:
Finally, the fourth term in eq. (1) is the conjugate of the third term:
All these defintions are at first a bit cumbersome, but equation (1) can now be
rewritten as:
We're pretty close to finding the desired weights now. Note that the above
equation is actually a quadratic function of the weight vector W. Hence, to
minimize the mean-squared error, we can take the gradient (or vector-derivative)
with respect to W, set the result equal to zero, and we have the optimal weights.
The gradient of the above equation is:
Setting this to zero and solving for W, the optimal MMSE-weights are:
These are the optimal weights we have been looking for. The result is statistical -
in terms of expected values of the desired signals. In the next section, we'll look
at this in a little more depth and present an example that illustrates this weighting
method.
The above equation requires knowledge of the autocorrelation matrix and the
vector , both defined here. The inverse of the autorcorrelation matrix is often
estimated using the Sample Matrix Inverse (SMI) method. The estimate is
denoted with a bar overhead. It uses K snapshots (or samples) of the input vector
X to determine the estimate:
Assuming the desired signal is uncorrelated in time with the noise and
interference (a reasonable assumption), the vector can be simplified to:
The above equation states the the vector can be determined if the direction of
the signal (given by ) and the signal power ( ) are known.
These parameters can be estimated using a training sequence to calibrate the
array (training or known bits are often sent to calibrate a digital system to
optimize transmission). Using the above equation, the optimal weights can be
rewritten as:
The optimal value of the Mean-Squared Error (MSE) can be evaluated using the
formula for the mse and substituting in the optimal weights:
The fundamentals of this method have been presented. In the next section, we'll
look at an example that illustrates the utility of this method.
Two cases will be considered. Case 1 will have a noise power given by
(SNR = 20 dB), and Case 2 will have a noise power given by (SNR = 0
dB). The optimal weights can be determined from the equation at the top of this
page.
The required steering vector corresponding to the desired wave vector ( ) can
be easily calculated by substituting in the definitions. Assuming the noise,
desired signal and interference are uncorrelated, the autocorrelation matrix can be
expanded and written as:
In the above, I is an NxN identitiy matrix. The weights can then be determined
for both cases. The resulting normalized AFs are plotted in Figure 1.
Figure 1. Normalized AFs for both cases using MMSE weights.
The interferers are located at 40 and 90 degrees. The desired signal is located at
110 degrees. In both cases, the array tries to maximize power reception in the
desired direction, and to minimize it near the interference locations. Note
however that when the Signal-to-Noise Ratio (SNR) is high (implying low
noise), as in Case 1 (SNR=20 dB), the array does a better job at placing nulls in
the direction of the interference. In Case 2 (SNR=0 dB), when the noise is more
dominant, the array places less of an emphasis on nulling out the interference,
and the nulls are not exactly in the direction of the interference. The weights end
up being more uniform in amplitude for the second case, which tends to increase
the SNR for the array output. In this manner, the array is willing to sacrifice
quality on blocking interference in order to minimize the noise affect. This result
shows that the optimal weights vary as a function of the noise power.
In this section, the LMS algorithm is introduced. LMS stands for Least-Mean-
Square. This algortihm was developed by Bernard Widrow in the 1960s, and is
the first widely used adaptive algorithm. It is still widely used in adaptive digital
signal processing and adaptive antenna arrays, primarily because of its simplicity,
ease of implementation and good convergence properties.
The goal of the LMS algorithm is to produce the MMSE weights for the given
environment. The definitions of all the terms used on this page follows that from
the MMSE page, which should be understood before reading this page. The goal
of the LMS algorithm is to adaptively produce weights that minimize the mean-
squared error between a desired signal and the arrays output; loosely speaking, it
tries to maximize reception in the direction of the desired signal (who or what the
array is trying to communicate with) and minimize reception from the interfering
or undesirable signals.
Just as in the MMSE case, some information is needed before optimal weights
can be determined. And just as in the MMSE weighting case, the required
information is the desired signal's direction and power. The direction is specified
via the desired signal's steering vector ( ) and the signal power is written as
. Note that these parameters can vary with time, as the environment is assumed to
be changing. The directions and power can be determined using various direction
finding algorithms, which analyze the received signals at each antenna in order to
estimate the directions and power.
Recall that the Mean-Squared Error between the desired signal and the array's
output can be written as:
The gradient (vector derivative with respect to the weight vector) can be written
as:
The LMS algorithm then approximates the gradient of the MSE by substituting in
the above simple approximation for the autocorrelation matrix:
The adaptive weights will be written as W(k), where k is an index that specifies
time. The LMS weighting algorithm simply updates the weights by a small
amount in the direction of the negative gradient of the MSE function. By moving
in the direction of the negative gradient, the overall MSE is decreased at each
time step. In this manner, the weights iteratively approach the optimal values that
minimize the MSE. Moreover, since the adaptive algorithm is continuously
updating, as the environment changes the weights adapt as well.
The weights are updated at regular intervals, and the weight at time k+1 is related
to time k by:
The parameter controls the size of the steps the weights make, and affects the
speed of convergence of the algorithm. To guarantee convergence, it should be
less than 2 divided by the largest eigenvalue of the autocorrelation matrix.
Substituting in the estimate for the gradient above, the LMS update algorithm can
be written as a simple iterative equation:
The algorithm simplicity is the primary reason for its widespread use. The above
update equation does not require any complex math, it just uses the current
samples of the received signal at each antenna (X).
The algorithm is starting assuming a weight vector of all ones (the starting
weight vector ideally has no impact on the end results):
Using random noise at every step, the algorithm is stepped forward from the
initial weight. The resulting MSE at each time step is shown in the following
figure, relative to the optimal MSE.
The LMS algorithm is fairly efficient in moving towards the optimal weights for
this case. Since the algorithm uses an approximateion of the autocorrelation
matrix at each time step, some of the steps actually increase the MSE. However,
on average, the MSE decreases. This algorithm is also fairly robust to changing
environments.
Several adaptive algorithms have expanded upon ideas used in the original LMS
algorithm. Most of these algorithms seek to produce improved convergence
properties at the expense of increased computational complexity. For instance,
the recursive least-square (RLS) algorithm seeks to minimize the MSE just as in
the LMS algorithm. However, it uses a more sophisticated update to find the
optimal weights that is based on the matrix inversion lemma. Both of these
algorithms (and all others based on the LMS algorithm) have the same optimal
weights the algorithms attempt to converge to.
Antenna Measurements
Testing of real antennas is fundamental to antenna theory. All the antenna theory
in the world doesn't add up to a hill of beans if the antennas under test don't
perform as desired. Antenna Measurements is a science unto itself; as a very
good antenna measurer once said to me "good antenna measurements don't just
happen".
3. Phase Measurements
The third antenna measurements section focuses on determining phase
information from an antenna's radiation pattern. The phase is more important in
terms of 'relative phase' (phase relative to other positions on the radiation
pattern), not 'absolute phase'.
4. Polarization Measurements
The fourth antenna measurements section discusses techniques for determining
the polarization of the antenna under test. These techniques are used to classify
an antenna as linearly, circularly or elliptically polarized.
5. Impedance Measurements
The fifth antenna measurement section illustrates how to determine an antenna's
impedance as a function of frequency. Here the focus is on the use of a Vector
Network Analyzer (VNA).
The Receiving System simply needs to determine how much power is received
from the test antenna. This can be done via a simple bolometer, which is a device
for measuring the energy of incident electromagnetic waves. The receiving
system can be more complex, with high quality amplifiers for low power
measurements and more accurate detection devices.
The Positioning System controls the orientation of the test antenna. Since we
want to measure the radiation pattern of the test antenna as a function of angle
(typically in spherical coordinates), we need to rotate the test antenna so that the
source antenna illuminates the test antenna from different angles. The positioning
system is used for this purpose.
Once we have all the equipment we need (and an antenna we want to test), we'll
need to place the equipment and perform the test in an antenna range, the subject
of the next section.
The drawback to anechoic chambers is that they often need to be quite large.
Often antennas need to be several wavelenghts away from each other at a
minimum to simulate far-field conditions. Hence, it is desired to have anechoic
chambers as large as possible, but cost and practical constraints often limit their
size. Some defense contracting companies that measure the Radar Cross Section
of large airplanes or other objects are known to have anechoic chambers the size
of basketball courts, although this is not ordinary. universities with anechoic
chambers typically have chambers that are 3-5 meters in length, width and
height. Because of the size constraint, and because RF absorbing material
typically works best at UHF and higher, anechoic chambers are most often used
for frequencies above 300 MHz. Finally, the chamber should also be large
enough that the source antenna's main lobe is not in view of the side walls,
ceiling or floor.
Elevated Ranges
Elevated Ranges are outdoor ranges. In this setup, the source and antenna under
test are mounted above the ground. These antennas can be on mountains, towers,
buildings, or wherever one finds that is suitable. This is often done for very large
antennas or at low frequencies (VHF and below, <100 MHz) where indoor
measurements would be intractable. The basic diagram of an elevated range is
shown in Figure 2.
The source antenna is not necessarily at a higher elevation than the test antenna, I
just showed it that way here. The line of sight (LOS) between the two antennas
(illustrated by the black ray in Figure 2) must be unobstructed. All other
reflections (such as the red ray reflected from the ground) are undesirable. For
elevated ranges, once a source and test antenna location are determined, the test
operators then determine where the significant reflections will occur, and attempt
to minimize the reflections from these surfaces. Often rf absorbing material is
used for this purpose, or other material that deflects the rays away from the test
antenna.
Compact Ranges
The source antenna must be placed in the far field of the test antenna. The reason
is that the wave received by the test antenna should be a plane wave for
maximum accuracy. Since antennas radiate spherical waves, the antenna needs to
be sufficiently far such that the wave radiated from the source antenna is
approximately a plane wave - see Figure 3.
However, for indoor chambers there is often not enough separation to achieve
this. One method to fix this problem is via a compact range. In this method, a
source antenna is oriented towards a reflector, whose shape is designed to reflect
the spherical wave in an approximately planar manner. This is very similar to the
principle upon which a dish antenna operates. The basic operation is shown in
Figure 4.
Figure 4. Compact Range - the spherical waves from the source antenna are
reflected to be planar (collimated).
Due to reciprocity, the radiation pattern from the test antenna is the same for both
the receive and transmit modes. Consequently, we can measure the radiation
pattern in the receive or transmit mode for the test antenna. We will describe the
receive case for the antenna under test.
The test antenna is rotated using the test antenna's positioning system. The
received power is recorded at each position. In this manner, the magnitude of the
radiation pattern of the test antenna can be determined. We will discuss phase
measurements and polarization measurements later.
The coordinate system of choice for the radiation pattern is spherical coordinates.
Measurement Example
An example should make the process reasonably clear. Suppose the radiation
pattern of a microstrip antenna is to be obtained. As is usual, lets let the direction
the patch faces ('normal' to the surface of the patch) be towards the z-axis.
Suppose the source antenna illuminates the test antenna from +y-direction, as
shown in Figure 1.
Figure 1. A patch antenna oriented towards the z-axis with a Source illumination
from the +y-direction.
In Figure 1, the received power for this case represents the power from the angle:
. We record this power, change the position and record again.
Recall that we only rotate the test antenna, hence it is at the same distance from
the source antenna. The source power again comes from the same direction.
Suppose we want to measure the radiation pattern normal to the patch's surface
(straight above the patch). Then the measurement would look as shown in Figure
2.
In Figure 2, the positioning system rotating the antenna such that it faces the
source of illumination. In this case, the received power comes from direction
. So by rotating the antenna, we can obtain "cuts" of the radiation
pattern - for instance the E-plane cut or the H-plane cut. A "great circle" cut is
when =0 and is allowed to vary from 0 to 360 degrees. Another common
radiation pattern cut (a cut is a 2d 'slice' of a 3d radiation pattern) is when is
fixed and varies from 0 to 180 degrees. By measuring the radiation pattern
along certain slices or cuts, the 3d radiation pattern can be determined.
It must be stressed that the resulting radiation pattern is correct for a given
polarization of the source antenna. For instance, if the source is horizontally
polarized (see polarization of plane waves), and the test antenna is vertically
polarized, the resulting radiation pattern will be zero everywhere. Hence, the
radiation patterns are sometimes classified as H-pol (horizontal polarization) or
V-pol (vertical polarization). See also cross-polarization.
Measuring Gain
On the previous page on measuring radiation patterns, we saw how the radiation
pattern of an antenna can be measured. This is actually the "relative" radiation
pattern, in that we don't know what the peak value of the gain actually is (we're
just measuring the received power, so in a sense can figure out how directive an
antenna is and the shape of the radiation pattern). In this page, we will focus on
measuring the peak gain of an antenna - this information tells us how much
power we can hope to receive from a given plane wave.
We can measure the peak gain using the Friis Transmission Equation and a "gain
standard" antenna. A gain standard antenna is a test antenna with an accurately
known gain and polarization (typically linear). The most popular types of gain
standard antennas are the thin half-wave dipole antenna (peak gain of 2.15 dB)
and the pyramidal horn antenna (where the peak gain can be accurately
calculated and is typically in the range of 15-25 dB). Consider the test setup
shown in Figure 1. In this scenario, a gain standard antenna is used in the place of
the test antenna, with the source antenna transmitting a fixed amount of power
(PT). The gains of both of these antennas are accurately known.
Figure 1. Record the received power from a gain standard antenna.
From the Friis transmission equation, we know that the power received (PR) is
given by:
If we replace the gain standard antenna with our test antenna (as shown in Figure
2), then the only thing that changes in the above equation is GR - the gain of the
receive antenna. The separation between the source and test antennas is fixed,
and the frequency will be held constant as well.
Figure 2. Record the received power with the test antenna (same source antenna).
Let the received power from the test antenna be PR2. If the gain of the test
antenna is higher than the gain of the "gain standard" antenna, then the received
power will increase. Using our measurements, we can easily calculate the gain of
the test antenna. Let Gg be the gain of the "gain standard" antenna, PR be the
power received with the gain antenna under test, and PR2 be the power received
with the test antenna. Then the gain of the test antenna (GT) is (in linear units):
The above equation uses linear units (non-dB). If the gain is to be specified in
decibels, (power received still in Watts), then the equation becomes:
And that is all that needs done to determine the gain for an antenna in a particular
direction.
Recall that the efficiency of an antenna is simply the ratio of the peak gain to the
peak directivity:
Hence, once we have measured the radiation pattern and the gain, the efficiency
follows directly from these.
In the next section, we'll look at measuring the phase of an antenna's radiation
pattern.
Measuring Phase
For an antenna's radiation pattern to be completely specified, we need the
magnitude of the power received or transmitted, AND the phase. These
measurements should be specified in two orthogonal directions in order to
capture all the components (polarizations) of the antenna.
For instance, suppose an antenna transmits at frequency f and the fields travelling
in the +y direction at a particular point are seen to be:
The E-fields are orthogonal to the direction of travel in the far field region. The
magnitude of the x-component of the E-field is A and the magnitude of the z-
component of the E-field is B. The phase of the x-component is D and the phase
of the z-component is F (relative to the oscillation at frequency f). If D=F, the
components are in phase and the polarization is linear. If D and F are separated
by pi/2 radians (90 degrees) and the amplitudes are equal, the E-field is circularly
polarized.
The phase is a relative quantity - that is, it must be measured relative to some
fixed reference. On this page, we will discuss determining the phase of the fields
radiated from an antenna.
The easiest way to measure phase is the method shown in Figure 1. In this
method the test antenna is used as the source antenna, and another antenna is
used to receive the fields. For this method to work, the observation point is not
too far from the test antenna, so that the source waveform feeding the test
antenna can also be run into a phase measurement box. This box compares the
locations of the peaks and valleys of the received signals and determines the
relative phase from this information. The receive antenna is moved and then the
process is repeated.
Figure 1. Measuring the Phase of a Test Antenna When the Observation is 'Near'
the source.
The probe antenna should have good polarization purity, so that it can pick up
one component of the received field. To get the other orthogonal component, it
could simply be rotated or another probe antenna used.
If the test antennas are very far from each other and the reference (source)
waveform can not be fed directly into the phase measurement circuit (this
happens at low frequencies and large outdoor ranges where many wavelengths
becomes a large distance), then a standard antenna with known phase
characteristics is used to transmit a wave, which is used to compare with the
received signal from the test antenna.
Polarization Measurements
Fundamental to an antenna's radiation pattern is its polarization. On this page,
we'll discuss methods and techniques for measuring the polarization of an
antenna. Note that the polarization varies depending on the direction of radiation
from an antenna. For instance, a circularly polarized antenna may be
approximately circular only over a narrow beamwidth, and linearly polarized
away from the antenna's main beam (this is often the case for circularly polarized
patch antennas).
To perform the measurement, we will use our test antenna as the source. Then we
will use a linearly polarized antenna (typically a half-wave dipole antenna) as the
receive antenna. The linearly polarized receive antenna will be rotated, and the
received power recorded as a function of the angle of the receive antenna. In this
manner, we can gain information on the polarization of the test antenna. This
received information only applies to the polarization of the test antenna for the
direction in which the power is received. For a complete description of the
polarization of the test antenna, the test antenna must be rotated so that the
polarization can be determined for each direction of interest.
The power is recorded for the a fixed position (orientation) of the receive
antenna, then it is rotated about the x-axis as shown in Figure 1, and the power is
recorded again. This is done for a complete rotation of the linearly polarized
receive antenna.
From this information, a lot can be determined about the polarization of the test
antenna. Lets look at a couple of cases. Suppose that the test antenna is vertically
linearly polarized, and that the receive antenna is also vertically linearly
polarized, and that the rotation angle zero has both antennas polarization
matched. Then the output of our experiment, as a function of the rotation angle of
the receive antenna, would look something like the graph shown in Figure 2.
Figure 2. Output of Measurement when the Test Antenna is Linearly Polarized.
Left: Rectangular Plot. Right. Polar Plot.
The plots in Figure 2 give two views of the output. The left side gives an x-y plot
of the output. The right side gives a polar plot, which may be helpful in
visualizing the results. Note that the result is periodic - when the receive antenna
is rotated 180 degrees, it is again vertically polarized so that the received power
is identical.
Suppose now that the test antenna was horizontally polarized - again linearly
polarized, but initially not polarization matched to the receive antenna. Then the
resulting received power plots would resemble that shown in Figure 3.
Figure 3. Output measurement of Linearly Polarized Test Antenna (Horizontal
Pol).
In this case, we see that the shape of the resulting measurements are the same, but
that the peaks of the received power occur for different angles. As a result, we
know that when the test antenna is linearly polarized, the received power will
resemble the shapes shown in Figures 2 and 3, and by determining the angle in
which the received power is at a peak, we can determine the angle of the linear
polarization.
Suppose now that the test antenna was radiating a RHCP (Right Hand Circularly
Polarized) wave. If the test antenna was subject to the same measurement as
above, the normalized (make the peak output power equal to one for simplicity)
output power would resemble that of Figure 4.
Figure 4. Output of Measurement when the Test Antenna is Circularly Polarized.
In the meantime, lets look at another example. Suppose the test antenna is
elliptically polarized, with a tilt angle of 45 degrees and an axial ratio of 3 dB.
The E-field for this antenna might be described using the coordinate system of
Figure 1 by the equation:
Now its getting interesting. Isn't Figure 5 cool? I think so. First, we can tell the
tilt angle of the elliptically polarized wave by where the received power is at a
peak - in this case at 45 degrees, which can be seen from either figure. The other
parameter of elliptical polarization - the axial ratio, can be determined as well
from these plots. The ratio of the peak output (in this case 1.0 when the angle is
45 degrees) to the minimum power output (in this case 0.5, when the angle 8s
135 degrees) gives the axial ratio (1/0.5 = 2 = 3dB). Hence, by simply observing
the plots and being aware of the types of polarization, we can quickly determine
the type of polarization for this direction of the test antenna's radiation pattern.
Again, note that we don't know the direction of rotation of the E-field (left or
right). Finally, one more example. Suppose the axial ratio is 9 dB and the major
axis of the ellipse is the z-axis. The result would resemble that of Figure 6.
Figure 6. Output of Measurement when the Test Antenna is Elliptically Polarized
(axial ratio = 9 dB, major axis along 0 degrees).
Finally, we would like to determine the sense of polarization for the antenna
under test. Suppose that the test antenna is RHCP, so that we get the output
shown in Figure 4. We can't tell from this whether the result is RHCP or LHCP. A
simple approach to determine the sense of rotation in this case is to use an
antenna that is known to be RHCP as the receive antenna. The result is recorded,
and then the measurement is performed again with a receive antenna that is
LHCP. The result for the first case should be much larger than the output for the
second case if the test antenna is RHCP. By selecting the polarization sense based
on the output power that is larger, the sense of rotation for the polarization can be
determined. The difficulty in this is that it requires two antennas at the frequency
of interest that are closely RHCP and LHCP, which isn't always easy.
Before we begin, I'd like to point out that object placed around the antenna will
alter its radiation pattern. As a result, its input impedance will be influenced by
what is around it - i.e. the environment in which the antenna is tested.
Consequently, for the best accuracy the impedance should be measured in an
environment that will most closely resemble where it is intended to operate. For
instance, if a blade antenna (which is basically a dipole shaped like a paddle) is to
be utilized on the top of a fuselage of an aiprlane, the test measurement should be
performed on top of a cylinder type metallic object for maximum accuracy. The
term driving point impedance is the input impedance measured in a particular
environment, and self-impedance is the impedance of an antenna in free space,
with no objects around to alter its radiation pattern.
Fortunately, impedance measurements are pretty easy if you have the right
equipment. In this case, the right equipment is a Vector Network Analyzer
(VNA). This is a measuring tool that can be used to measure the input impedance
as a function of frequency. Alternatively, it can plot S11 (return loss), and the
VSWR, both of which are frequency-dependent functions of the antenna
impedance. The Agilent 8510 Vector Network Analyzer is shown in Figure 1.
Figure 1. The popular Agilent (HP) 8510 VNA.
Next, we need to calibrate the VNA. This is much simpler than it sounds. We will
take the cables that we are using for probes (that connect the VNA to the antenna)
and follow a simple procedure so that the effect of the cables (which act as
transmission lines) is calibrated out. To do this, typically your VNA will be
supplied with a "cal kit" which contains a matched load (50 Ohms), an open
circuit load and a short circuit load. We look on our VNA and scroll through the
menus till we find a calibration button, and then do what it says. It will ask you to
apply the supplied loads to the end of your cables, and it will record data so that
it knows what to expect with your cables. You will apply the 3 loads as it tells
you, and then your done. Its pretty simple, you don't even need to know what
you're doing, just follow the VNA's instructions, and it will handle all the
calculations.
Now, connect the VNA to the antenna under test. Set the frequency range you are
interested in on the VNA. If you don't know how, just mess around with it till you
figure it out, there are only so many buttons and you can't really screw anything
up.
If you request output as an S-parameter (S11), then you are measuring the return
loss. In this case, the VNA transmits a small amount of power to your antenna
and measures how much power is reflected back to the VNA. A sample result
(from the slotted waveguides page) might look something like:
Note that the S-parameter is basically the magnitude of the reflection coefficient,
which depends on the antenna impedance as well as the impedance of the VNA,
which is typically 50 Ohms. So this measurement typically measures how close
to 50 Ohms the antenna impedance is.
Another popular output is for the impedance to be measured on a Smith Chart. A
Smith Chart is basically a graphical way of viewing input impedance (or
reflection coefficient) that is easy to read. The center of the Smith Chart
represented zero reflection coefficient, so that the antenna is perfectly matched to
the VNA. The perimeter of the Smith Chart represents a reflection coefficient
with a magnitude of 1 (all power reflected), indicating that the antenna is very
poorly matched to the VNA. The magnitude of the reflection coefficient (which
should be small for an antenna to receive or transmit properly) depends on how
far from the center of the Smith Chart you are. As an example, consider Figure 3.
The reflection coefficient is measured across a frequency range and plotted on a
Smith Chart.
In Figure 3, the black circular graph is the Smith Chart. The black dot at the
center of the Smith Chart is the point at which there would be zero reflection
coefficient, so that the antenna's impedance is perfectly matched to the generator
or receiver. The red curved line is the measurement. This is the impedance of the
antenna, as the frequency is scanned from 2.7 GHz to 4.5 GHz. Each point on the
line represents the impedance at a particular frequency. Points above the equator
of the Smith Chart represent impedances that are inductive - they have a positive
reactance (imaginary part). Points below the equator of the Smith Chart represent
impedances that are capacitive - they have a negative reactance (for instance, the
impedance would be something like Z = R - jX).
To further explain Figure 3, the blue dot below the equator in Figure 3 represents
the impedance at f=4.5 GHz. The distance from the origin is the reflection
coefficient, which can be estimated to have a magnitude of about 0.25 since the
dot is 25% of the way from the origin to the outer perimeter.
In summary, the Smith Chart is a useful tool for viewing impedance over a
frequency range in a concise, clear form.
In summary, there are a bunch of ways to measure impedance, and a lot are a
function of reflected power from the antenna. We care about the impedance of an
antenna so that we can properly transfer the power to the antenna.
One method we can use in this case is that of Scale Model Measurements. In this
technique, a scaled model (typically a much smaller physical model of the actual
structure) is used to represent the platform on which the antenna operates. Now
the question is - is this valid? Can we get proper measurements from a smaller
model?
The answer is yes, if we know what we are doing. If you look at the page on
frequency, I mentioned that frequency is one of the fundamental secrets of the
universe - every signal can be represented as a combination of frequencies.
Further, all of electromagnetics and antenna theory are completely described by
Maxwell's Equations. If we are riding along a monochromatic (single frequency)
electric field wave (stay with me here) and you encountered an obstacle, your
behavior would only depend on the size of the obstacle in wavelengths. That is,
plane waves respond the same wave to a perfectly conducting circular plate that
is 3 wavelengths in diameter, no matter what the frequency.
The quality of the results depends on the quality of the model. This method is a
valid and often practiced method of antenna measurements in the aerospace and
defense world.
The following table is provided for understanding proper scaling. In the table, the
model is assumed to be scaled down by a factor of n. For instance, in the
previous paragraph, the airplane example had a scaled down model by a factor
n=10. As another example, if n=100, then an L=100 meter real-world object
would be represented by a scaled model that is L/n = 1 meter in size.
Table I is the result of working through the math in Maxwell's Equations. The
good news is that some properties do not need scaled at all, particularly the
permittivity and permeability (if you are modeling dielectrics or magnetic
materials). In addition, the antenna impedance and gain do not need scaled either,
which is a good thing. However, some things like conducitivity need to be
increased by a factor of n in the scaled model. One way to understand this is to
note that the resistance of the scaled model should be constant, and resistance is
proportional to the conductivity of the object times the length divided by the
cross sectional area. If the length goes down by a factor of n, and the cross
sectional area goes down by a factor of n^2, then for constant resistance we
require that the conductivity increases by a factor of n on the scaled model.
Typically, the real world airplane has a shell that has high conductivity (metals).
Hence, care needs exercised that the conductivity of the model is as high as
possible. This can be done with polished high quality metals for the scaled
model.
[1]
The units of SAR are W/kg, or equivalently, mW/g. The SAR limit in the US for
mobile phones is 1.6 W/kg, averaged over 1 gram of tissue. In Europe, the SAR
limit is 2.0 W/kg averaged over 10 grams of tissue. It is typically harder to
achieve the US specification than the Europe spec, so if the phone meets the US
spec it will typically also meet the European spec.
Measuring SAR
To obtain official SAR measurements, the DASY system by SPEAG is used. This
is shown in Figure 1:
In Figure 1, there is a hollow tub that the yellow robot moves the measuring
probe into. The tub is formed in standardized shapes, which replicate the shape of
the human head. The tub that the probe is measuring for Figure 1 simulates the
left side of the human head. The DUT is the device-under-test (the mobile phone)
is placed directly on the edge of the tub, and transmits at maximum power
continuously. The probe is moved through the head region by the yellow robot
arm, performing the averaging in Equation [1].
To simulate the conductivity and density correctly, the tub is filled with a fluid
that has similar properties to human tissue. The fluids are frequency-dependent,
so a standardized fluid for a measurement at 1800 MHz will be different from the
correct fluid for measuring SAR at 900 MHz.
The SAR must be measured on both the left and right side of the head as shown
in Figure 1. Even though geometrically the measurements are fairly similar, the
results can be very different due to the chaotic nature of the near field. The SAR
values quoted for a mobile phone are the highest value of SAR measured for any
frequency the phone operates in, from both the left and right side of the head.
Lowering SAR
SAR is critical to antenna design, because if the SAR is too high the antenna
must be changed. Typically, if the SAR is too high the transmit power is lowered,
which directly yields lower SAR. However, since there are minimum transmit
power specifications for mobile devices, the SAR cannot be dropped indefinitely.
As a result, the antenna positioning is critical. The antennas for mobile phones
are typically on the bottom of the phone, to keep the radiating part of the phone
as far as possible from the brain region.
Other methods for dropping the SAR include impedance matching changes and
parasitic resonators which will disturb the antenna's radiation pattern (hopefully
lowering SAR).
Impedance Transformations
Intro to Impedance Matching: Series L and C
Immittance Charts
Smith Charts were originally developed around 1940 by Phillip Smith as a useful
tool for making the equations involved in transmission lines easier to manipulate.
See, for instance, the input impedance equation for a load attached to a
transmission line of length L and characteristic impedance Z0. With modern
computers, the Smith Chart is no longer used to the simplify the calculation of
transmission line equatons; however, their value in visualizing the impedance of
an antenna or a transmission line has not decreased.
This section of the antenna theory site will present an introduction to Smith Chart
basics.
Smith Chart Tutorial
We'll now begin to explain the fundamentals of the Smith Chart. The Smith Chart
displays the complex reflection coefficient [Equation 1, below], in polar form, for
an arbitrary impedance (we'll call the impedance ZL or the load impedance). The
reflection coefficient is completely determined by the impedance ZL and the
"reference" impedance Z0. Note that Z0 can be viewed as the impedance of the
transmitter, or what is trying to deliver power to the antenna. Hence, the Smith
Chart is a graphical method of displaying the impedance of an antenna, which
can be a single point or a range of points to display the impedance as a function
of frequency. For a primer on complex math, click here.
[Equation 1]
For this Smith Chart tutorial, we will assume Z0 is 50 Ohms, which is often, but
not always the case. Note that the Smith Chart can be used with any value of Z0.
1. As such, the set of all possible values for must lie within the unit circle:
Figure 2. The Complex Reflection Coefficient must lie somewhere within the
unit circle.
In Figure 2, we are plotting the set of all values for the complex reflection
coefficient, along the real and imaginary axis. The center of the Smith Chart is
the point where the reflection coefficient is zero. That is, this is the only point on
the Smith Chart where no power is reflected by the load impedance.
The outter ring of the Smith Chart is where the magnitude of is equal to 1.
This is the black circle in Figure 1. Along this curve, all of the power is reflected
by the load impedance.
[Equation 2]
From equation [2], with Z0=50 Ohms, a reflection coefficient of 0.5 corresponds
to a load impedance ZL=150 Ohms. We can plot gamma_1 on the smith chart:
Since is entirely real, the point lies along the real gamma axis (x-axis) in
Figure 3, and the imaginary axis value (y-axis) location is 0.
From Equation [2] and using Z0=50, we note that corresponds to a load
impedance
Since VSWR is only a function of the absolute value of , we can get the
VSWR for a load from the Smith Chart as well. That is, a VSWR = 1 would be
the center of the Smith Chart, and VSWR=3 would be a circle centered around
the center of the Smith Chart, with magnitude =0.5. Circles centered at the
origin of the Smith Chart are constant-VSWR circles. Note that the outer
[1]
[2]
In equation [2], Y is any real number. What would the curve corresponding to
equation [2] look like if we plotted it on the Smith Chart for all values of Y?
That is, if we plotted z1 = 1 + 0*i, and z1 = 1 + 10*i, z1 = 1 - 5*i, z1 = 1 - .333*i,
.... and any possible value for Y that you could think of, what is the resulting
curve? The answer is shown in Figure 1:
In Figure 1, the outer blue ring represents the boundary of the smith chart. The
black curve is a constant resistance circle: this is where all values of z1 = 1 + i*Y
will lie on. Several points are plotted along this curve, z1 = 1, z1 = 1 + i*2, and
zL = 1 - i*4.
Suppose we want to know what the curve z2 = 0.3 + i*Y looks like on the Smith
Chart. The result is shown in Figure 2:
Figure 2. Constant Resistance Circle for zL=0.3 on Smith Chart.
In Figure 2, the black ring represents the set of all impedances where the real part
of z2 equals 0.3. A few points along the circle are plotted. We've left the
resistance circle of 1.0 in red on the Smith Chart.
These circles are called constant resistance curves. The real part of the load
impedance is constant along each of these curves. We'll now add several values
for the constant resistance, as shown in Figure 3:
Figure 3. Constant Resistance Circles on Smith Chart.
In Figure 3, the zL=0.1 resistance circle has been added in purple. The zL=6
resistance circle has been added in green, and zL=2 resistance circle is in black.
On this page, we start to see where the all the curves and lines on the smith chart
come from. In the next section, we'll look at constant reactance curves.
On this page, we will look at the set of curves defined by zL = R + iY, where Y is
held constant and R varies from 0 to infinity. Since R cannot be negative for
antennas or passive devices, we will restrict R to be greater than or equal to zero.
As a first example, let zL = R + i. The curve defined by this set of impedances is
shown in Figure 1:
Figure 1. Constant Reactance Curve for zL = R + i*1.
The resulting curve for Im[zL]=-1 is plotted in green in Figure 2. The point
zL=1-i is placed on the Smith Chart, which is at the intersection of the Re[zL]=1
circle and the Im[zL]=-1 curve.
An important curve is given by Im[zL]=0. That is, the set of all impedances given
by zL = R, where the imaginary part is zero and the real part (the resistance) is
greater than or equal to zero. The result is shown in Figure 3:
Figure 3. Constant Reactance Curve for zL=R.
The reactance curve given by Im[zL]=0 is a straight line across the Smith Chart.
There are 3 special points along this curve. On the far left, where zL = 0 + i0, this
is the point where the load is a short circuit, and thus the magnitude of is 1, so
all power is reflected.
In the center of the Smith Chart, we have the point given by zL = 1. At this
location, is 0, so the load is exactly matched to the transmission line. No
power is reflected at this point.
The point on the far right in Figure 3 is given by zL = infinity. This is the open
circuit location. Again, the magnitude of is 1, so all power is reflected at this
point, as expected.
Finally, we'll add a bunch of constant reactance curves on the Smith Chart, as
shown in Figure 4.
Figure 4. Smith Chart with Reactance Curves and Resistance Circles.
Figure 4 shows the fundamental curves of the Smith Chart. If you look back on
Figure 1 on the intro to Smith Charts page, it should now make sense where all
the chaotic curves come from.
In the next section, we'll begin to look at some of the things that make Smith
Charts so useful.
[1]
Now, the question is: If we have a load impedance ZL, what is the input
impedance Zin a distance L down the transmission line, using the Smith Chart?
To figure this out, let's just take an example. Let ZL = 100 Ohms, so that the
normalized load is zL=100/50 = 2.0. Let's plot, on the Smith Chart, a few values
for zin=Zin/Z0, which are given by:
[2]
[3]
Let's take an example. Let zL = 2.0 again as above. If we draw a circle centered
at the center of the Smith Chart and travelling through zL, then we get the curve
given in Figure 3 by the black x's:
Figure 3. Constant VSWR circle.
A few points are plotted along this circle. If you travel lambda/8 (one eighth of a
wavelength) down the transmission line in Figure 1, the resulting input
impedance can be found by rotating 90 degrees in the clockwise direction on the
Smith Chart. Similarly, if you want the input impedance lambda/4 (one quarter of
a wavelength) from the load impedance, the resulting input impedance can be
found by rotatin 180 degrees in the clockwise direction around the Smith Chart.
Hence, the input impedance (from equation [1] or the Smith Chart) repeats itself
every half-wavelength. That is, a half-wavelength along the transmission line
corresponds to a complete rotation on the Smith Chart (back to where you
started).
Solution. This problem is very simple thanks to the Smith Chart. First, plot the
known value of z1, along with the constant VSWR curve (the circle centered in
the Smith Chart going through z1):
Figure 5. The First Step: Plotting z1 and the Constant VSWR Circle.
To determine ZL, we want to move on the Smith Chart towards the load.
Remember:
Hence, to find ZL, we want to move in the counter-clockwise direction. Note that
L1 = 5*lambda/8. Recall that a distance of lambda/2 on a transmission line
corresponds to a complete rotation on the Smith Chart. Hence, this is equivalent
to moving 5*lambda/8 - lambda/2 = lambda/8. Hence, we can simply rotate in
the counter-clockwise direction by 90 degrees (one quarter turn on the Smith
Chart). We can then read off the value for this impedance, and the result is:
zL = 0.198 - i*0.9802 ==> ZL = 9.901 - i*49.0099
Similarly, to find z2, the impedance at the generator, we simply move lambda/8
in the clockwise direction (since we are traveling away from the antenna/load).
The result is a 90 degree rotation, plotted in Figure 6. The result is:
The correspond Smith Chart is shown, with only the needed constant resistance
circles and constant reactance curves:
Figure 6. The Second Step: Using the Constant SWR Circle to Get zL and z2.
In Figure 6, the circle corresponds to the magnitude of being 0.8182. This can
be found from using the equation for (equation [1] here). In Figure 6, I only
plotted the reactance curves for Im[Z]=0.9802 and Im[Z]=-0.9802. The reason is
that all the others were irrelevant to this analysis. On a real smith chart, you
would simply interpolate between the closest reactance curves. Similarly, I only
plotted the resistance circles given by Re[Z]=0.1 (for the z1 impedance) and
Re[Z]=0.198 (for zL and z2). All other curves are irrelevant on the Smith Chart
for this problem.
Finally, on complicated, detailed Smith Charts, you will see a scale along the
outer perimeter as shown below:
We'll now introduce some of the basic building blocks to make this happen.
Series Inductor
An inductor has a normalized impedance given by:
[1]
The question now is: what does a series inductor do to a load impedance ZL? The
block diagram is shown in Figure 1:
Figure 1. Series Inductor and load impedance ZL.
[2]
From equation [2], we see that the series inductor will move the impedance zL
along the constant resistance circles of the Smith Charts. Hence, if the reactance
(X) of the load impedance ZL is negative, then we can use a series inductor to
cancel out this reactance, making the input impedance purely real.
As an example, let zL = 1 - i2 when f=1 GHz. Then we can cancel out the
reactance with a series inductor, determined by:
[3]
That is, equation [3] states that we can cancel out the reactance of the load with a
15.9 nH series inductor. This move is illustrated on the Smith Chart in Figure 2:
Series Capacitor
A capacitor has normalized impedance given by:
[4]
In equation [4], f is frequency, and C is the capacitance in Farads. Note that the
capacitor gives rise to a negative reactance.
The question now is: what does a series capacitor do to a load impedance ZL?
The block diagram is shown in Figure 3:
From equation [5], we see that the series capacitor will move the impedance zL
along the constant resistance circles of the Smith Charts, but in the opposite
direction that the inductor moves it. If the reactance (X) of the load impedance
ZL is positive, then we can use a series capacitor to cancel out this reactance,
making the input impedance purely real.
As an example, let zL = 0.3 + i when f=500 MHz. Then we can cancel out the
reactance with a series capacitor, determined by:
[6]
That is, equation [6] states that we can cancel out the reactance of the load with a
6.4 pF series capacitor. This move is illustrated on the Smith Chart in Figure 4:
Figure 4. Series Capacitor cancelling reactance of load impedance zL.
Note that the impedance zL is translated along the constant resistance circle
(Re[z]=0.3) to the equator of the Smith Chart. By lowering the capacitance C, we
can move the impedance zL farther along the constant resistance circle (note that
the capacitor's reactance is inversely proportional to the capacitance - equation
[4]). In this case, we have not exactly matched the load impedance zL to the
center of the Smith Chart. The reflection coefficient is reduced, but not ideally.
Summary
On this page, we've learned that the series inductor moves the load impedance
along the constant resistance circles in a clockwise fashion, as shown in Figure 2.
We've also seen that the series capacitor moves the load impedance along the
constant resistance circles in a counter-clockwise fashion, as shown in Figure 4.
These translations should be remembered, it will help with impedance matching
in the future.
Smith Chart Tutorial -
Impedance Matching with Tx
Lines, Series L and C
On the previous page, we showed how to use series inductors and capacitors to
cancel out the reactance of a load. However, as seen in the second example on the
previous page, this method won't work unless the real part of the load impedance
is 1 (that is, Re[ZL]=1). On this page, we'll show how to use transmission line
sections and a series component to exactly match any load impedance.
Let's look at example 2 from the previous page again. That is, let zL = 0.3 + i.
Instead of using a component, suppose we use a transmission line section to
move the impedance. We know from the page on impedance transformations that
a transmission line section will enable the load impedance to move in a circle
about the center of the Smith Chart. That is, given the load impedance zL, a
transmission line section can relocate the impedance to any location in the black
circular ring of Figure 1:
Figure 1. Impedance zL on the Smith Chart along with its constant VSWR circle.
Recall that if an impedance is of the form z1 = 1 + iX, then we can exactly match
it using a series inductor or capacitor. Now, if we use a transmission line section
to rotate the impedance on the Smith Chart until it intersects the Re[z]=1 circle,
then we can exactly match any impedance that does not have a reflection
coefficient magnitude equal to 1 (that is, we can match anything on the interior of
the Smith Chart).
To figure out the rotation needed, draw a line from the center of the Smith Chart
through zL, and then draw another line from the center of the Smith Chart to the
location where the constant SWR circle intersects the Re[Z]=1 circle. Then find
the angle between the two lines, as shown in Figure 2.
In Figure 2, we have found the angle between the two lines to be 45.5 degrees.
This can be done in different ways; for example, by drawing the lines and then
using a protracter to determine the angles, or reading off the angle difference
using the degrees scale on the outer rim of the smith chart.
Since one wavelength corresponds to two complete rotations around the Smith
Chart, we have:
[1]
[2]
Hence, we can rotate the impedance zL to the Re[zL]=1 circle
using a transmission line of length 0.0632 wavelengths. The
resultant rotation on the Smith Chart is illustrated via the brown
curve in Figure 3:
The resultant impedance is z1 (shown in Figure 3), and the value for this can be
read directly off the smith chart yielding z1 = 1 + i*2.228. For clarity, note that
z1 is the impedance shown in Figure 4:
Since z1 = 1 + i*2.228, we can exactly match this impedance to the center of the
Smith Chart uses a series capacitor. We simply choose C such that z_cap =
-i*2.228, and the impedance z2 will be 1.0. For details on how this is done, see
previous page. The path taken by zL to reach the center of the Smith Chart is
shown via the brown curve in Figure 5.
Figure 5. Moving zL to z1 and then z2 with a tx line and series capacitor.
The problem with the above method is that it works for a single band (not ok if
you want dual band matching) and it requires a tx line which can take up too
much space. We'll look at other methods, involving parallel components in the
coming sections.
To generalize the Smith Chart and to make things easier for parallel impedance
matching, we will now introduce the Admittance Smith Chart.
In equation [1], Y_L is the admittance of the load, which is the inverse of the
impedance of the load (Z_L). The real part of Y_L is written as C, and is known
as the conductance. The imaginary part of Y_L is written as S, and is known as
the susceptance.
Note that the admittance Smith Chart will still plot the reflection coefficient
exactly as before. The admittance Y can be related back to the reflection
coefficient by rewritting equation 1 on the Smith Chart tutorial page:
[2]
[3]
A few examples for y_L values are plotted along the constant conductance circle
shown in Figure 1. Again, the outer blue circle in Figure 1 represents the
boundary of the Smith Chart; this is where the magnitude of the reflection
coefficient is unity.
The smallest circle on the left in Figure 2 is the Re[y]=6 circle. The largest purple
circle is Re[y]=0.1. Several other constant conductance circles are shown in
Figure 2. These will be very useful for impedance matching with parallel
components.
Note that positive values of susceptance lie on the southern hemisphere of the
Smith Chart, whereas positive values of reactance lie on the northern hemisphere
of the Smith Chart. To complete the admittance Smith Chart, we will sketch
several more constant susceptance curves, as shown in Figure 4:
Figure 4. Constant Susceptance Curves on Admittance Smith Chart.
On the next page, we'll introduce the immitance Smith Chart, and look at parallel
components for impedance matching.
Recall that if an admittance y_ind is placed in parallel with the load admittance
y_L, the combination of the two admittances add. That is:
If we want to match the load (i.e., bring it to the center of the Smith Chart), then
we want the parallel combination of the load (antenna) and the inductor to equal
1.0. Assuming Z0=50 Ohms and we are impedance matching at f=850 MHz, we
can cancel out the susceptance of the load with a parallel inductor:
[2]
The admittances y_1, y_L and y_IND are shown in Figure 2. Note the brown path
shown. By using a parallel inductor, the antenna y_L can be moved counter-
clockwise along the constant conductance circle Re[y]=1.0.
Parallel Capacitors
The normalized admittance of a capacitor y_C is given by:
[3]
[4]
The admittances y_1, y_L and y_C are shown in Figure 4. Note the brown path
shown. By using a parallel capacitor, the antenna y_L can be moved clockwise
along the constant conductance circle Re[y]=0.3.
Figure 4. Load Admittance Matched with using Parallel Capacitor.
Again, we can't perfectly match y_1 in this example with a parallel capacitor or
inductor, because they only allow us to move along the constant susceptance
curves. The point of this example is to understand how a parallel capacitor moves
an admittance (load) on the Smith Chart.
In the next section, we'll look at impedance matching with parallel transmission
line stubs.
Impedance Matching with
Parallel Tx Line Stubs
On the previous page we introduced impedance matching with
parallel inductors and capacitors. On this page, we'll illustrate
how transmission lines can be placed in parallel to perform
impedance matching.
Short-Circuited Tx Lines
If you haven't already, you will need to review the short and open circuited
transmission lines page. From that page, the impedance and admittance of a short
circuited tx line is shown to be:
[1]
Equation [2] states that we can use an open-circuited line (with a length less than
a quarter-wavelength) to create a capacitive impedance. That is, the tx line will
have Re[Zin]=0 and Im[Zin]<0 for 0 < L < lambda/4.
of is less than 1. This method is very similar to the techinique given on the
impedance matching with Tx Lines, Series L, and Series C page.
Consider the load admittance yL = 0.3 + i*0.5 (then zL = 0.882 - i*1.471). The
absolute value of is 0.6176, which can be found from equation [2] on the
admittance Smith Chart page. The load yL and the constant VSWR circle (in
black) is plotted in Figure 1:
Figure 1. Load Admittance and the Corresponding Constant SWR Circle.
To match the load of Figure 1 with a parallel (shunt) component, we will need to
swing the load admittance around on the Smith Chart until it intersect the
Re[y]=1 constant conductance circle. That is, if we add a length of transmission
line in front of the load admittance, we can move the load yL along the black
constant SWR circle of Figure 1 until it intersects the Re[y]=1 circle. Then a
parallel transmission line section can be used to cancel out the susceptance (the
imaginary part of the admittance), such that the load will be exactly matched.
To do this, we first find the point y1, which is located on the constant SWR
circle, where it intersects the Re[y]=1 constant conductance circle. This value,
y1, is found to be y1=1+i*1.571 (which can be read off a more detailed Smith
Chart). Next, we measure the angles on the Smith Chart between y1 and yL,
which will enable us to determine the required transmission line length. The
angle can be determined using a protractor or by using the angle scales on the
outer rim of the Smith Chart. This overall process is illustrated in Figure 2:
Figure 2. Determining Transmission Line Length to Intersect the Re[y]=1 Circle.
Using equation [1] on the impedance matching with tx lines page, we can
calculate that the angle 71.57 degrees corresponds to a line length of roughly
0.0994 wavelengths.
From the point y1 shown in Figure 2, we bring the admittance to the center of the
Smith Chart with a parallel inductance, of y_sc = -i*1.571. On this page, we will
implement this with a short-circuited shunt tx line. Using equation [1] from
above, we can get the line length:
[3]
The key to understanding this section is realizing that a parallel susceptance can
be used to match any load/admittance/impedance that intersects the Re[y]=1
constant conductance circle.
On this page, we'll look at impedance matching with both series and parallel
components. It turns out, that if we combine the impedance Smith Chart with the
admittance Smith Chart, we will have no trouble doing this. The combination is
known as the Immittance Smith Chart.
The immittance Smith Chart of Figure 1 looks insanely complicated. But it is just
the admittance Smith Chart in red overlayed on top of the impedance Smith
Chart in green. This Smith Chart, while it looks like chaos, is extraordinarily
helpful.
The reason Figure 1 is awesome: whether we want to know how a parallel
component or a series component affects the impedance, we can visualize it using
the immittance chart. Parallel capacitors, inductors and transmission lines will
affect the load impedance by movement along the constant conductance circles
(in red). Series capacitors and inductors affect the load or antenna impedance by
translation along the constant resistance circles (in green). Finally, even though
we haven't talked about it and it isn't used much for impedance matching with
antennas, series resistors will move an impedance along the constant reactance
curves, and parallel resistors will move an impedance along the constant
susceptance curves.
Using the Immittance Smith Chart, we can match any impedance (with a
reflection coefficient magnitude less than 1) using a capacitor and an inductor.
Here's a fundamental observation:
Example 1
Suppose an antenna has an impedance of z_A=0.1-i0.2. We know that a series
inductor will move this impedance in a clockwise direction along the Re[z]=0.1
circle. We would like to move this impedance along this circle until it intersects
the Re[y]=1 circle.
Using the immittance Smith Chart of Figure 1, we will need to move the antenna
admittance
y_1 is equal to 1 - 3i, which places us on the Re[y]=1 circle, which is where we
want to be.
Now we can use a parallel component to move the impedance z_1 to the center
of the Smith Chart. Since y_1 = 1 - 3i, we can use a capacitor with a normalized
susceptance of 3 (recall from admittance charts page that the susceptance is the
imaginary part of the admittance, and this is positive for a capacitor, see equation
[3] on the parallel L and C page). The exact value for the capacitance depends on
what frequency we are interested in. The parallel capacitor moves the impedance
z_1 along the constant conductance circle as shown in Figure 3:
Figure 3. Completing the Impedance Matching.
Hence, we have matched the impedance z_A with a series inductor and a parallel
capacitor. This matching network is shown in Figure 4.
Recall that we moved the impedance z_A to intersect the Re[y]=1 circle with a
series capacitor. By observing the immittance Smith Chart, we could also
accomplish this with a series capacitor. Using a capacitor with a series reactance
of -i*0.1, the impedance will be transformed to z1=0.1-i*0.3, which is equivalent
to y1=1+i*3. This is illustrated in Figure 1:
Figure 1. Using a Series Capacitor to Move zA to the Re[y]=1 circle.
To complete the impedance matching, we just need to cancel out the susceptance.
That is, if we add a parallel inductor with a susceptance of -i*3, the impedance
will be translated to the center of the Smith Chart, shown in Figure 2:
Figure 2. A parallel inductor cancels out the susceptance.
Which is no big deal, since we can simply use a parallel component to move the
impedance to intersect the Re[z]=1 circle. In fact, we have two options here: we
can use a parallel capacitor or a parallel inductor to move the impedance to
intersect the Re[z]=1 circle. I will use a parallel capacitor.
The impedance z1 can then be easily matched with a series inductor with a
reactance of zL = i*1.7321. This impedance matching process is shown in Figure
4. Note that for clarity, the circle Re[y]=0.25 and the curve Im[z]=-1.732 is
shown. The points and path in Figure 4 should be understood. If not, work
through Example 2 and the preceding pages until it is clear.
Figure 4. Impedance Matching for z_A = 2+i*2.
Getting the exact values for the inductors and capacitors is not important. The
idea of impedance matching is to understand how the components move you on
the Smith Chart, and then get approximate components to perform the matching.
Exact values of capacitors and inductors do not exist (i.e. you can find a 0.8 pF
capacitor but you can't find a 0.832 pF cap), and the components are non-ideal
(some resistance, non-ideal frequency characteristics, etc.), so impedance
matching comes down to understanding how components move an impedance on
the Smith Chart. Once that is understood, impedance matching is fairly simple.
How do you know which one to use? For single frequency impedance matching,
it doesn't matter so much. Large value of inductors are often avoided because
they have higher loss, which reduces antenna efficiency. The impedance
matching networks are sometimes chosen for what components are readily
available.
However, for dual band matching, the type of impedance matching network used
will be very important. This is the topic of the next section.
The Problem
The blue curve in Figure 1 represents the antenna's impedance, plotted on the
Smith Chart continuously from 800 Mhz to 2 GHz.
The goal is to match the above antenna at the lower frequency fLow=900
MHz and at the upper frequency fHi = 1900 MHz. These are common
frequencies used in mobile phones.
The Solution
In Figure 1, the black circles indicate the location on the blue curve
corresponding to 900 MHz and 1900 MHz. We would like to bring both of these
impedances to the center of the Smith Chart with a single matching network.
To do this, we will first attempt to match the low band. In matching the low band,
we will only make use of series capacitors and parallel inductors. This is very
important: series capacitors will affect lower frequencies more than higher
frequencies. Similarly, parallel inductors will affect low frequencies more than
high frequencies. As a result, if we use only these components, we can match the
low band (900 MHz) with minimal effect on the high band (1900 MHz).
The second step will be to match the high band without messing up the low band
impedance match. This can be done with the use of parallel capacitors and series
inductors (both of which affect high frequencies more than low frequencies).
Step 1 - Matching the Lowband (900 MHz)
For matching the lowband, we can use series capacitors or parallel inductors.
Note the location of the 900MHz impedance in Figure 1. A parallel inductor
would not move the impedance to intersect the Re[z]=1 or the Re[y]=1 circles.
However, a series capacitor will move the impedance to intersect the Re[y]=1
circle. By going through the math on the series L and C page, we can find that a
3.2 pF series capacitor will move the 900MHz impedance to intersect the
Re[y]=1 circle.
Side Note. Using a series capacitor, we could intersect the Re[y]=1 circle in two
locations (either at roughly y1 = 1 - i*2.9 (on the TOP half of the Smith Chart)
or y2 = 1 + i*2.9 (on the BOTTOM half of the Smith Chart). If we had chosen
the first location, y1, then we would have needed to complete the match with a
parallel capacitor - which is not good because it will affect the high band MORE
than the low band. Hence, we use the point y2, where the match can be
completed with a parallel inductor.
Note that in Figure 3, the lower frequencies are shifted more by the series
capacitor. This is expected, since the impedance of the capacitor decreases in
magnitude as frequency gets higher.
Now, the 900 MHz frequency point is approximately located on the Re[y]=1
circle. We can complete the lowband match with a parallel inductor. By using the
math on the parallel L and C page, we find that the 900 MHz point can be
matched with approximately a 3nH parallel inductor.
The resulting impedance curve (the antenna, the 3.2 pF series capacitor and then
the parallel 3nH inductor) is shown in Figure 4:
Figure 4. Frequency Response of Antenna after Series Capacitor and Parallel
Inductor.
From Figure 4, we see that the low band (900 MHz) is moved to the center of the
Smith Chart - i.e., we have impedance matched this region. However, the high
band (1900 MHz) region has moved from it's original location, even though that
was not our intent. The goal now is to try to impedance match the 1900 MHz
point without undoing the 900 MHz match.
Note that in Figure 5, the 900 MHz point is moved slightly away from the center.
We can live with that though. The highband is now set up to be matched with a
series inductor. The required value can be found to be approximately 5 nH of
series inductance. The addition of this component to the overall matching
network produces the final impedance curve, shown in Figure 6:
Figure 6. Frequency Response of Antenna after Lowband and Highband
Matching.
Observe that in Figure 6, the high and lowbands are not exactly matched.
However, Figure 7 presents the VSWR corresponding to the impedance plot of
Figure 6. At 900 MHz and at 1900 MHz, the VSWR is less than 2.0. This is
typically considered a good match. If a better match was desired, the 4
component values could be optimized to improve the response (I would not
recommend adding more components to the matching network, as the bandwidth
of the match tends to decrease as more components are added, in addition to
more losses associated with more components).
Figure 7. VSWR of Antenna after Lowband and Highband Matching.
This concludes the Smith Chart tutorial. I hope you had as much fun as I did.
Remember: ostrich burger has less fat, but you eat more of it.
Antenna Definitions
Antenna Theory (Home)