Yagi-Uda Antenna

Richard Laugesen

Phys327 – Electromagnetic Applications

Abstract. A Yagi-Uda antenna was designed and simulated using a software package called Yagimax then
experimentally tested using an AMS506 antenna modeling system. The characteristics of the antenna and a
comparison between the simulated and the experimental results were discussed. The design process was
complicated, with a large number of interacting parameters governing the antenna response it was hard to
simultaneously optimize the many characteristics of the antenna, The simulated 4 element Yagi performed
best at 1410 MHz where it had a forward gain of 9.95 dBi, a front to back ratio of 25 dB, and a beamwidth
of 55o. The constructed antenna had a forward gain of 7.95 dBi, a front to back ratio of 20 dB, and a
beamwidth of 65o. The experimental antenna performs similar to the simulated antenna, but it is clear that
the simplifications made in the computer simulation do not accurately model an antenna in the real world.

Introduction Element Length (λ) Separation (λ)

Reflector > 0.50 0.15 – 0.25
A Yagi-Uda antenna is a widely used antenna
design due to its high forward gain capability, low Driver 0.45 1 -
cost and ease of construction, it is commonly used as
a roof top television receiver. It consists of a number Directors 0.40 – 0.45 0.30 – 0.40
of metal rods called elements arranged on a central
support beam. The elements are a dipole - which is
the only driven element - arranged with a number of A polar plot of the gain verses orientation
parasitic elements, of which there are two types; (radiation pattern) is useful when characterizing
antennas. Some important features that appear on
• A reflector plot are;
• One or more directors • Forward gain – expressed in dB relative to an
isotropic source or a standard dipole (in
The function of the parasitic elements is to
direction of maximum gain) represents the
improve the radiation pattern in the forward
improvement in signal level to a reference
direction. The reflector is placed behind the dipole
antenna.
and is slightly longer, it provides 3dB of additional
forward gain, but having more than one reflector has • Front to back ratio – ratio of signal level in
little benefit. The directors are placed in front and the forward direction to the signal level in the
slightly shorter in length than the dipole provide an back direction (rotated 180o) also expressed in
additional 3dB of forward gain each. dB.
The parasitic elements provide forward gain by • Beamwidth – angle between directions where
redistributing the energy of the EM signal generated the power is half the value at direction of
by the dipole, since they are not driven they can only maximum gain, hence -3dB. It gives a
redistribute energy in one direction at the expense of measure of the directivity of the antenna.
other directions.
• Side lobes – these are unwanted peaks in the
For a particular operating frequency a typical
gain at angles other than in the forward
design would be;

1
Since an EM wave has a reduced velocity in the
conductor a factor of 0.95 is usually applied, so in practice
the driver is generally 0.475λ long.

Richard Laugesen Page 1 3/11/2001

 direction, they reduce the amount of useful the others, for example a large forward gain but
energy contained in the forward direction. equally large side lobes. Other designs had excellent
characteristics but the dimensions made them unable
Other characteristics that do not appear on the
to be built on the experimental equipment.
polar plot but which are equally important are;
Eventually a final design satisfying our criteria was
• Bandwidth – the range of frequency over found;
which the antenna exhibits acceptable
Element Length (λ) Separation (λ)
characteristics.
Reflector 0.546 -
• Radiative Resistance – want the impedance of
the antenna to match the impedance of the Driver 0.475 0.12
transmission cable used to drive it otherwise
Director 1 0.455 0.15
signal loss and high voltages in the cable may
occur. Director 2 0.421 0.25
It is difficult to optimize all of these
characteristics simultaneously so the aim when
designing an antenna depends on the requirements of
the situation in which the antenna will be used.
Optimization is achieved by simulating the
radiation pattern of the antenna while varying the
lengths and separations of the elements.

1. Procedure
Designed a Yagi-Uda antenna using a software
package called Yagimax, attempted to
simultaneously maximize the forward gain while
minimizing the side lobes by varying element lengths Figure 1.Diagram of the antenna
and separations.
This antenna produced the following simulated
Once an ideal antenna had been designed that was
radiation pattern at 1.4 GHz, which was the intended
within the physical constraints of the antenna testing
operating frequency;
equipment it was carefully constructed. The radiation
pattern of this antenna was then generated using the
AMS506 antenna modeling system and compared to
those generated by the simulation software.

2. Results and Discussion

Designing and simulating the antenna

The aim of the design process was to design an

antenna with the following properties; Forward gain
greater than 10 dB over an isotropic source, front to
back ratio greater than 20 dB, and side lobes 20 dB
less than forward gain. The antenna testing
equipment imposed constraints on the length of the Figure 2.Simulated radiation pattern at 1400 MHz
elements, separation and operating frequency, thus
these were taken into account during design process. Notice the large forward gain, 9.86 dB, the side
lobes are relatively small, the front back ratio is quite
Began with a basic design of a reflector, dipole large at 69 dB, and it has an impedance of 11.29 –
and one director, then proceeded to change element j0.63 all close to our design goals.
separations and lengths, adding additional directors
and using the built in optimization functions of the The real part of the impedance is relatively low
software. Several designs were created which which is not ideal when trying to match the antenna
maximized one characteristic but failed to maximize to a transmission line, but this is not that critical as

Richard Laugesen Page 2 3/11/2001

an impedance may be transformed to another more
acceptable value if required.
As the frequency changes the characteristics of
the antenna also change, the range over which these
characteristics are acceptable is termed the
bandwidth. The interesting thing about this design
was that the following characteristics were all
maximized (or minimized) right at 1.4 GHz, see
figure 13; the front to back ratio, the side lobes, and
the reactive component of the impedance. The
forward gain increases slightly with frequency.

Figure 4.Simulated radiation pattern of H-plane

operating at 1410 MHz

It is clear that improving one characteristic of the

antenna may have a detrimental effect on another and
that small changes in element length and separation
can cause large changes in the characteristics of the
antenna.

Constructing and testing the real antenna

The antenna designed using the software was

Figure 3.Simulated radiation pattern at 1410 MHz constructed using AMS506 antenna modeling
system, vernier calipers were used to accurately set
At 1410MHz the antenna had the highest forward the lengths and separations of the elements. This
gain for any frequency, 9.95 dB over an isotropic antenna was then experimentally tested to compare
source, see figure 3. Notice that the side lobes have the simulated radiation patterns with experimental
decreased in size slightly and the large dip in the radiation patterns.
reverse direction has been replaced by a small peak.
Therefore the front to back ratio is not as high, 25
dB, but the energy from the side lobes has now been
used in the forward direction giving a slightly higher
forward gain at the expense of the front to back ratio.
The beamwidth was estimated from the plot to be
approximately 55o, and does not change noticeably
with frequency.
This operating frequency was considered to be
the best compromise as it gave the highest forward
gain. Even though it did have a detrimental effect on
the front to back ratio it should be noted that the plot
is on a logarithmic scale. Therefore anything less
than 20 dB would be negligible when compared to
the forward gain on a linear scale, such as the entire
back and section of the radiation pattern.
Figure 5.Experimental radiation pattern at 1400
MHz

The radiation patterns of the experimental

antenna were similar to the simulated patterns, but
were shifted in frequency. The optimum frequency
for the experimental antenna was 1355 MHz whereas
for the simulated antenna it was 1410 MHz.

Richard Laugesen Page 3 3/11/2001

 At 1400 MHz the experimental antenna exhibited
an undesirable large back lobe which did not occur in
the simulated pattern, compare figures 2 & 5.
However at 1360 MHz the experimental radiation
pattern appeared very similar to the pattern generated
by the simulated antenna operating at 1400 MHz, see
figure 6, with the characteristic large dip in the
backwards gain appearing. Radiation patterns seen
either side of 1400 MHz in the simulated tests were
also seen located either side of 1360 MHz in the
experimental antenna tests.

Figure 8.Experimental H-plane radiation pattern at

1355 MHz.

To measure the forward gain relative to an

isotropic source, the parasitic elements were
removed and radiation pattern for the dipole alone
was recorded. This is the radiation pattern of a
reference dipole. The forward gain relative to an
isotropic source is the difference in forward gain
between the reference dipole and the antenna plus an
additional 2.15 dB, which is the gain of the dipole
relative to an isotropic source, [2] (pp 11-2). Thus at
Figure 6.Experimental radiation pattern at 1360 1355 MHz the forward gain was 7.95 dBi, had a
MHz front to back ratio of 20 dB, side lobes approximately
-23 dB, and a beamwidth of approximately 65o.
The experimental frequency with the largest
These values are for only one measurement
forward gain was 1355 MHz, figure 7. Visual
process at this frequency, therefore the significance
inspection of the radiation pattern shows that it has
of the values is limited. A method to obtain more
similar features to the simulated radiation pattern for
accurate results would be to repeat the measurement
maximum forward gain at 1410 MHz, figure 3.
of antenna radiation pattern many times and average
the characteristics of the antenna, the standard
deviation of this data set could be used as
uncertainties for the values. To determine the
forward gain an averaged value of the dipole strength
would also be required as it varied significantly from
one measurement to the next2. Unfortunately
problems with the equipment and time constraints
did not allow this to be performed.
The numbers are relatively close to those
predicted by the simulation but the forward gain is
significantly lower, this could be because of the
following reasons;
• Reflections – reflections from objects in the
room, the walls, floor and ceiling caused
significant distortions in the radiation patterns.
Figure 7.Experimental radiation pattern at 1355 The plots that were reasonable and which were
MHz used for this analysis must still have been
degraded by these reflections and since they
The radiation pattern for the antenna in the H- were not taken into account in the simulation
plane was highly distorted due to the reflections from they could account for the discrepancy.
the ground, compare figures 8 & 4.

2
Could have been sourcing problems to the dipole.

Richard Laugesen Page 4 3/11/2001

• Impedance matching – the impedance of the The experimental antenna performs relatively
antenna and the transmission line feeding the similar to the simulated one, the shape of the
antenna were not investigated for mismatching, simulated radiation patterns are of the same form as
so this may have caused a loss in input signal the experimental, but there are a number of
feeding the antenna. distortions and the frequency shift give the
simulation limited real world value. It is clear that
• Construction – During the design process it was the approximations made in the computer model do
found that small changes in the element length not allow accurate modeling a real antenna.
or separation can cause large changes in the
characteristics of the antenna, therefore any On the other hand it would also be extremely
measurement errors introduced during the difficult to design a Yagi-Uda antenna without
construction of the antenna could cause the real simulation software.
antenna to perform quite differently to the
predicted.
4. References

3. Conclusion 1. D. K. Cheng, Field and Wave Electromagnetics, 2nd

Edition, Addison-Wesley, 1989.
Found that the design process is extremely 2. The ARRL Antenna Book, 19th ed
complicated, with many interacting parameters it is
3. Yagimax 3.11;
hard to optimize the many characteristics of the http://www.qrz.com/ftp/antennas/yagim311.zip
antenna simultaneously.

Richard Laugesen Page 5 3/11/2001

5. Appendix

Figure 9.Experimental radiation pattern of constructed dipole operating at 1400 MHz. The dipole was verified
to have the same forward gain at any frequency.

Figure 10.Experimental radiation pattern of dipole overlaid with the pattern of antenna operating at 1355 MHz
to estimate the forward gain over an isotropic source.

Richard Laugesen Page 6 3/11/2001

 Figure 11.Three dimensional view of the simulated radiation pattern of the antenna operating 1410 MHz.

Figure 12.Simulated radiation pattern of dipole at 1400 MHz.

Richard Laugesen Page 7 3/11/2001

 Figure 13.Frequency response of simulated antenna – note quite narrow bandwidth if front to back ratio is
important to role the antenna is designed for. Bandwidth is alright when concerned only with the forward gain,
~200 MHz.

Richard Laugesen Page 8 3/11/2001