Final Report

Design of a Compact Vivaldi Antenna Array 2011

Chapter 1

Introduction

1.1 Introduction

The objective of this project has been to design a compact tapered slot Vivaldi antenna

array for UWB see through wall radar.

Vivaldi antennas have received considerable attention due to their high gain, relatively

wide band, simple structure, easy fabrication, and wide use in UWB applications. Their small

lateral dimensions and simple integration make them excellent candidates for array

development [1.1]. Federal Communication Commission (FCC) approved a 1.99 GHz to 11.6

GHz frequency band for use in UWB through-wall imaging systems [1.2]. Yang et.al. [1.3]

designed a Vivaldi antenna array around 10 GHz for UWB see through wall radar utilizing

antipodal Vivaldi antennas with Wilkinson power divider for the binary feed. However, the

size of this 16 element array is relatively too large if the antenna array is duplicated for the

lower band UWB applications; i.e. close to 3 GHz. Therefore, we used here only a 4- element

array and optimized its performance to sustain similar almost constant gain over its operating

band. Similar concepts to that utilized by Abbosh et al. [1.4] to design a compact UWB

antipodal Vivaldi antenna have been utilized here.

In this project, we have developed a Vivaldi antenna array for see through wall UWB

applications. The configuration of the array element was optimized to have a compact size.

Then, a 1 × 2 Vivaldi antenna array is developed using tapered slot antennas (TSA) and a 3

dB power divider. After that, a 1 × 4 Vivaldi antenna array is developed using tapered slot

antennas (TSA) and Wilkinson power divider. Details of the developed Vivaldi antenna,

Wilkinson power divider, Vivaldi antenna arrays, its simulation and experimental results are

presented in this project report.

Department of Electronics, CUSAT Page 1


1.2 Antenna Parameters

The performance of an antenna can be gauged from a number of parameters. Certain

critical parameters are discussed below [1.5].

1.2.1 Gain

Gain is a parameter which measures the degree of directivity of the antenna's radiation

pattern. An antenna with a low gain emits radiation with about the same power in all

directions, whereas a high-gain antenna will preferentially radiate in particular directions.

Specifically, the antenna gain, directive gain, or power gain of an antenna is defined as the

ratio of the intensity (power per unit surface) radiated by the antenna in the direction of its

maximum output, at an arbitrary distance, divided by the intensity radiated at the same

distance by a hypothetical isotropic antenna.

The gain of an antenna is a passive phenomenon - power is not added by the antenna,

but simply redistributed to provide more radiated power in a certain direction than would be

transmitted by an isotropic antenna. An antenna designer must take into account the

application for the antenna when determining the gain. High-gain antennas have the

advantage of longer range and better signal quality, but must be aimed carefully in a

particular direction. Low-gain antennas have shorter range, but the orientation of the antenna

is relatively inconsequential. For example, a dish antenna on a spacecraft is a high-gain

device that must be pointed at the planet to be effective, whereas a typical Wi-Fi antenna in a

laptop computer is low-gain, and as long as the base station is within range, the antenna can

be in any orientation in space. It makes sense to improve horizontal range at the expense of

reception above or below the antenna. Thus most antennas labeled "omnidirectional" really

have some gain.

Power gain is a unit less measure that combines an antenna's efficiency and directivity figures:

(1.1)

If the radiation intensity U in the desired solid angle is known, then power gain for that solid

angle can be calculated:



(1.2)

1.2.2 Radiation pattern

The radiation pattern of an antenna is a plot of the relative field strength of the radio

waves emitted by the antenna at different angles. It is typically represented by a three

dimensional graph, or polar plots of the horizontal and vertical cross sections. The pattern of

an ideal isotropic antenna, which radiates equally in all directions, would look like a sphere.

Many non directional antennas, such as monopoles and dipoles, emit equal power in all

horizontal directions, with the power dropping off at higher and lower angles; this is called an

omnidirectional pattern and when plotted looks like a torus or donut.

The radiation of many antennas shows a pattern of maxima or "lobes" at various

angles, separated by "nulls", angles where the radiation falls to zero. This is because the radio

waves emitted by different parts of the antenna typically interfere, causing maxima at angles

where the radio waves arrive at distant points in phase, and zero radiation at other angles

where the radio waves arrive out of phase. In a directional antenna designed to project radio

waves in a particular direction, the lobe in that direction is designed larger than the others and

is called the "main lobe". The other lobes usually represent unwanted radiation and are called

"sidelobes". The axis through the main lobe is called the "principle axis" or "boresight axis".

Fig. 1.1 Radiation pattern of a directional antenna



1.2.3 Impedance

As an electro-magnetic wave travels through the different parts of the antenna system

(radio, feed line, antenna, free space) it may encounter differences in impedance (E/H, V/I,

etc.). At each interface, depending on the impedance match, some fraction of the wave's

energy will reflect back to the source, forming a standing wave in the feed line. The ratio of

maximum power to minimum power in the wave can be measured and is called the standing

wave ratio (SWR). A SWR of 1:1 is ideal. A SWR of 2.5:1 is considered to be marginally

acceptable in low power applications where power loss is more critical, although an SWR as

high as 6:1 may still be usable with the right equipment. Minimizing impedance differences

at each interface (impedance matching) will reduce SWR and maximize power transfer

through each part of the antenna system.

Complex impedance of an antenna is related to the electrical length of the antenna at

the wavelength in use. The impedance of an antenna can be matched to the feed line and

radio by adjusting the impedance of the feed line, using the feed line as an impedance

transformer. More commonly, the impedance is adjusted at the load with an antenna tuner, a

balun, a matching transformer, matching networks composed of inductors and capacitors, or

matching sections such as the gamma match.

1.2.4 Efficiency

Efficiency is the ratio of power actually radiated to the power put into the antenna

terminals. A dummy load may have an SWR of 1:1 but an efficiency of 0, as it absorbs all

power and radiates heat but very little RF energy, showing that SWR alone is not an effective

measure of an antenna's efficiency. Radiation in an antenna is caused by radiation resistance

which can only be measured as part of total resistance including loss resistance. Loss

resistance usually results in heat generation rather than radiation, and reduces efficiency.

Mathematically, efficiency is calculated as radiation resistance divided by total resistance.

1.2.5 Polarization

The polarization of an antenna is the orientation of the electric field (E-plane) of the

radio wave with respect to the Earth's surface and is determined by the physical structure of

the antenna and by its orientation. It has nothing in common with antenna directionality

terms: "horizontal", "vertical" and "circular". Thus, a simple straight wire antenna will have


http://en.wikipedia.org/wiki/E-plane

http://en.wikipedia.org/wiki/Polarization_(waves)


one polarization when mounted vertically, and a different polarization when mounted

horizontally. "Electromagnetic wave polarization filters are structures which can be employed

to act directly on the electromagnetic wave to filter out wave energy of an undesired

polarization and to pass wave energy of a desired polarization.

Reflections generally affect polarization. For radio waves the most important reflector

is the ionosphere - signals which reflect from it will have their polarization changed

unpredictably. For signals which are reflected by the ionosphere, polarization cannot be relied

upon. For line-of-sight communications for which polarization can be relied upon, it can

make a large difference in signal quality to have the transmitter and receiver using the same

polarization; many tens of dB differences are commonly seen and this is more than enough to

make the difference between reasonable communication and a broken link.

Polarization is largely predictable from antenna construction but, especially in

directional antennas, the polarization of side lobes can be quite different from that of the main

propagation lobe. For radio antennas, polarization corresponds to the orientation of the

radiating element in an antenna. A vertical omnidirectional WiFi antenna will have vertical

polarization (the most common type). An exception is a class of elongated waveguide

antennas in which vertically placed antennas are horizontally polarized. Many commercial

antennas are marked as to the polarization of their emitted signals.

Polarization is the sum of the E-plane orientations over time projected onto an

imaginary plane perpendicular to the direction of motion of the radio wave. In the most

general case, polarization is elliptical, meaning that the polarization of the radio waves varies

over time. Two special cases are linear polarization (the ellipse collapses into a line) and

circular polarization (in which the two axes of the ellipse are equal). In linear polarization the

antenna compels the electric field of the emitted radio wave to a particular orientation.

Depending on the orientation of the antenna mounting, the usual linear cases are horizontal

and vertical polarization. In circular polarization, the antenna continuously varies the electric

field of the radio wave through all possible values of its orientation with regard to the Earth's

surface. Circular polarizations, like elliptical ones, are classified as right-hand polarized or

left-hand polarized using a "thumb in the direction of the propagation" rule. Optical

researchers use the same rule of thumb, but pointing it in the direction of the emitter, not in

the direction of propagation, and so are opposite to radio engineers use.


http://en.wikipedia.org/wiki/Circular_polarization

http://en.wikipedia.org/wiki/Linear_polarization

http://en.wikipedia.org/wiki/Ellipse

http://en.wikipedia.org/wiki/WiFi

http://en.wikipedia.org/wiki/Omnidirectional_antenna

http://en.wikipedia.org/wiki/Line-of-sight_propagation

http://en.wikipedia.org/wiki/Ionosphere


In practice, regardless of confusing terminology, it is important that linearly polarized

antennas be matched, lest the received signal strength be greatly reduced. So horizontal

should be used with horizontal and vertical with vertical. Intermediate matchings will lose

some signal strength, but not as much as a complete mismatch. Transmitters mounted on

vehicles with large motional freedom commonly use circularly polarized antennas so that

there will never be a complete mismatch with signals from other sources.

1.2.6 Return loss

Return loss or reflection loss is the loss of signal power resulting from the reflection

caused at a discontinuity in a transmission line or optical fiber. This discontinuity can be a

mismatch with the terminating load or with a device inserted in the line. It is usually

expressed as a ratio in decibels (dB);

Where RL (dB) is the return loss in dB, Pi is the incident power and Pr is the reflected power.

Properly, loss quantities, when expressed in decibels, should be positive numbers

However, return loss has historically been expressed as a negative number, and this

convention is still widely found in the literature.

Taking the ratio of reflected to incident power results in a negative sign for return loss

Where RL'(dB) is the negative of RL(dB).

Caution is required when discussing increasing or decreasing return loss since these terms

strictly have the opposite meaning when return loss is defined as a negative quantity.

1.3 Project Report Organization

This repot has been organized in six chapters.

The second chapter gives an insight into the tapered slot antenna radiation

characteristics and design.



The third chapter gives a qualitative description of the full wave analysis technique,

Finite Element Method (FEM). Ansoft’s HFSS software based on FEM is used for antenna

simulation in this project. A brief description of HFSS is also given in this chapter.

The fourth chapter gives the design of Tapered Slot Antenna, its simulation and

measurement results.

The fifth chapter gives the design and fabrication of 1×2 Vivaldi antenna array, its

simulation and measurement results.

The sixth chapter discusses the 1×4 Vivaldi antenna array. A comparison between

the results of 2-element and 4-element array is also done.

The seventh chapter is conclusion.

1.4 References

[1.1] Sng-Gyu Kim, Kai Chang, “A low cross polarized antipodal Vivaldi antenna array

for wideband operation”, Antennas and Propagation Society International

Symposium, vol.3, June 2004, pp. 2269-2272.

[1.2] Federal Communications Commission, “Revision of part 15 of the Commission’s

Rules Regarding Ultra Wideband Transmission Systems”, March 12, 2003

[1.3] Yunqiang Yang, Cemin Zhang, Song Lin and Aly E. Fathy, “Development of an

ultra wideband antipodal antenna”, Antenna and Propagation Society International

Symposium, vol. 1A, July 2005, pp. 606-609.

[1.4] A.M. Abbosh, H.K. Kan and M.E. Bialkowski, “Design of compact directive ultra

wideband antipodal antenna”, Microwave and Optical Technology Letters, vol. 48,

no.12, Dec. 2006, pp. 2448-2451.

[1.5] www.wikipedia.org


http://www.wikipedia.org/


Chapter 2

Tapered Slot Antennas

2.1 Introduction

The objective of this chapter is to give an insight into the radiation mechanism and

various topologies possible while designing a Tapered Slot Antenna (TSA). Further,

comparison of their performances is also given based on the papers published by the other

authors.

The terms, notch antenna, tapered notch, flared slot antennas are used synonymously

to represent TSAs. Tapered slot antennas belong to the class of travelling wave antennas.

They are end fire radiators. Tapered slot is formed by gradual widening of a slotline. In 1979

Gibson [2.1] demonstrated an exponentially tapered slot antenna demonstrating a bandwidth

of 8-40 GHz and he called it Vivaldi antenna. In the same year Prasad and Mahapatra [2.2]

first introduced the linearly tapered slot antenna (LTSA). However, the tapered slot antenna

was introduced as an array element by Lewis et al. [2.3] in 1974.

The conventional resonant microstrip antenna size becomes very small as the

operating frequency shifts to millimeter wave frequency band. This increases the cost because

fabrication tolerance level decreases. In addition, the skin effect conductor losses in the

microstrip feed network tends to become excessive at higher frequencies thus lowering

antenna efficiency. Tapered slot antennas can circumvent these problems. The dimensions of

tapered slot antennas are several times the free space wavelength at the frequency of

operation which eases the fabrication tolerance. Many variations of these antennas were

fabricated for frequency of operation of up to about 800 GHz [2.4] and higher within the

required fabrication tolerances using standard printed circuit fabrication techniques.

Furthermore, active circuits like mixers and amplifiers can be integrated with the antenna

using stripline, slotline, microstrip line, finline and coplanar waveguide. In addition to this,

other advantages are:

Multi-octave bandwidth.

Moderate gain.

Symmetric E and H plane radiation patterns.



Some of the disadvantages as compared to the conventional microstrip patch are:

Cannot be designed for dual-frequency operation.

Dual polarization cannot be obtained without structure complexity.

Loses its planar architecture when used in 2D array.

2.2 Taper Profiles

According to the different taper profile they are generally classified into:

Linear tapered (LTSA)

Non-linear tapered (exponential, parabolic)

Constant width (CWSA)

Fig 2.1. Various Tapered slot Antenna Profiles: a) Step constant b) Exponential c)

Linear d) Parabolic e) Linear constant f) Broken linear

Various tapered slot profiles are illustrated in Fig 2.1. A variant of the conventional

TSA is the antipodal tapered slot antenna, see Fig. 2.2. In practice, the conventional planar

TSA is fed by a balanced slotline. One serious drawback of the conventional TSA is in the

fabrication and impedance matching of the slotline. Slotline fabricated on a low dielectric

constant substrate has relatively high impedance which makes matching to a low impedance

microstrip feed very difficult. The antipodal TSA replaces the band limiting microstrip/

slotline transition by a tapered balun section which gives a very wide bandwidth. However, it

has exhibited very poor cross polarization characteristics. The antenna is formed by gradually


(a) (b) (c)

(d) (e) (f)


flaring the strip conductors of the balanced microstrip on opposite sides of the dielectic

substrate with respect to the antenna axis, thus allowing the antenna to be directly fed by a

microstrip feed [2.5]. The antenna suffers from the high cross polarization due to skewing of

the E-field components within the antenna with respect to the physical axis of the antenna. At

the low frequency end of the band this skew is small because the ratio of slot width to

dielectric thickness is large. However as we move to the high frequency end the angle of

skew increases and ultimately tends to 90o. Therefore the antenna has poor cross polarization

(of the order of 5 dB) and also there is severe polarization tilt as the frequency of operation

increases [2.6].

Fig 2.2. Antipodal Vivaldi Antenna

To overcome the high cross polarization problem, Langley et al. [2.6] introduced a

new technique. The idea was to negate the effect of skew by applying another dielectric layer

and metallization layer. This new antenna, known as balanced antipodal Vivaldi which is fed

by stripline, has demonstrated -15 dB lower cross- polarizations across an 18:1 band as

compared to the conventional Vivaldi antenna.

2.3 Radiation mechanism

The Vivaldi antenna is essentially frequency independent, since at a given wavelength

only a section of the exponential curve actually radiates efficiently. As the wavelength varies,

radiation occurs from a different section which is scaled in size in proportion to the

wavelength, and has the same relative shape. This translates into antenna with a large

bandwidth. The main lobe of the antenna is linearly polarized with the electric field parallel

to the aperture.



(a) (b)

Fig 2.4. Vivaldi Antenna

A TSA can be divided into two regions: a non radiating feed region and radiating slot

region [2.7]. With respect to Fig 2.4 these regions can be identified as:

Propagating area defined by Ws < W < Wa

Radiating area defined by Wa < W < Wo

Where

Ws- input slot width

Wa- slot width at radiating area

Wo- output width

The main, non resonant, travelling wave mechanism of radiation is produced by

higher order Hankel function (Ho(n)) mode generated by waves travelling down a curved path

along the antenna [2.3]. The energy in the travelling wave is tightly bound to the conductors

when the separation is very small compared to the free space wavelength and that becomes

progressively weaker and more coupled to the radiation field as the separation is increased.

The guide wavelength and characteristic impedance of a slotline increase with the

increase in width of the slot. For a given dielectric substrate of certain thickness at a specified

frequency, a 20% increase in slot width leads to 1% increase in λ / λ’ and 6% increase in Zo.

If λ’ is less than about 40% of λ the fields will be adequately contained and the slotline

behaves like transmission line. Departure from this condition would result in radiation [2.2].



2.3.1 Antipodal Vivaldi Antenna:

The electric field lines at different cross section along the feed and the antenna are

illustrated in Fig. 2.5. The electric field lines which are spread out in the conventional

microstrip structure, concentrate between the metal strip of the balanced microstrip, and

finally rotate while travelling along the TSA [2.8]. This is the reason for higher cross-

polarizations in these types of antennas. At the higher frequencies (narrower slot width) the

skew angle is almost 90o due to which there can be a severe polarization tilt also.

(a) (b) (c)

Fig 2.5. Electric field distribution at cross sections a) Conventional microstrip b)

Balanced microstrip c) Radiating edge

2.4 Design Considerations

Design of TSA has been primarily based on empirical approach and as a starting

point, one can use the following guidelines [2.8]:

The aperture width of slot: W ≥ λo.

2.05 ≤ c/v ≤ 2.2.

The taper angle, 2α, is typically 5 to 12o

The length of TSA, L, is typically 2 to 12 λo

Where c is the velocity of light in free space; v is guided wave velocity along the slot; λ o is

the operating wavelength in free space.

In general, the design of TSA involves two major tasks:

The design of a broadband transition and feed structure with very wide frequency

range and low return loss, and

Determining the dimensions and shape of the antenna in accordance with the required

half power (3 dB) beam width, side lobe, and back lobe etc. over the operating



frequency range. The geometric parameters such as length, width, dielectric thickness,

ground plane size, and taper- profile have direct impact on the impedance, directivity,

bandwidth, and radiation pattern of the antenna.

2.5 Feeding Mechanism

This is the most critical part of Vivaldi antenna design. The impedance matching of

the low impedance stripline/ microstrip line to the high impedance slotline is crucial. For a

wide bandwidth operation one should design the feed transition with a wide bandwidth. Some

of the feed mechanisms are explained below.

2.5.1 Coaxial- Slotline Transition:

Coaxial line is useful as a feed structure [2.9] because of its compatibility with the

slotline, coplanar microstrip, or balanced microstrip to form wideband transitions; thus, it can

be used to excite all the variants of planar TSA described earlier.

Fig 2.6. Coaxial-Slotline Transition

The coaxial feed can be directly used to excite a TSA by extending the center

conductor over the slotline section of the TSA and anchor the coaxial feed with solder

connection to the ground plane as shown in Fig 2.6. The disadvantage is that it is not planar

and has high losses at higher frequencies. Further, coaxial line is an unbalanced feed line. All

of the currents flow inside the line, i.e. the inner connector and the inside of the shield.

Feeding a balanced antenna with unbalanced coaxial feed may cause currents to flow on the

outside of the shield, which could results in significant power loss and serious distortion in

radiation pattern [2.8].



2.5.2. Microstrip- Slotline Transition

Microstrip line is an unbalanced line and slotline is a balanced line. So feeding a TSA

with a microstrip line requires a balanced-to-unbalanced transition (balun). This is essential

for broadband antenna performance.

Fig. 2.7 4th order Printed Marchand Balun

The most common microstrip/slot transition, the Marchand balun [2.10, 2.11], has

demonstrated a VSWR of 2:1 over an octave bandwidth with an integrated wideband Vivaldi

antenna (DETSA) [2.12]. The balun consists of four quarter-wave sections with the end open-

circuited section extended past the center of the slotline by about one quarter of a guided

wavelength (λm). The Fig.2.7 shows the fourth order Marchand balun.

Another practical microstrip to slot transition consists of a slot, etched on one side of

the substrate, crossing an open circuited microstrip line, located on the opposite side, at a

right angle [2.13]. The slot extends to one quarter of a wavelength (λs) beyond the microstrip

and the microstrip extends one quarter of a wavelength (λm) beyond the slot [2.12]. Further

broadening of the bandwidth can be achieved when the microstrip is terminated by a radial

stub and the slot line is terminated by an elliptical shaped cavity [2.15] as shown in Fig 2.8.



Fig 2.8 Microstrip-Slotline Transition

A microstrip- slotline transition with radial stubs illustrated in Fig 2.8 provides a very

wide bandwidth [2.16-2.17]. This feeding is used in this project.

2.5.3 CPW – Slotline Transition

Another way of exciting the slotline in a Vivaldi antenna is to use a coplanar

waveguide (CPW) feed. Any transmission line with coplanar conductors can be considered a

coplanar waveguide line. The signal line and ground plane are on the same side of a printed

circuit board in a coplanar waveguide (CPW). The normal propagating mode on this

transmission line is the quasi-TEM mode with the electric fields in the two slots oriented in

opposite directions. Fig 2.9 (a) depicts a simple CPW – slotline transition. CPW may be

used to feed a TSA as shown in Fig 2.9 (b): one half of the CPW transitions into a slotline

and feeds a TSA while the other half is terminated in a short circuit [2.18]. This technique

may sometimes yield a bandwidth greater than that obtained with a conventional microstrip

feed [2.19]. The advantage of using a CPW feed is that it can be used in applications where

high circuit density is important as in microwave integrated circuit (MIC) applications.

Another benefit is that it has low radiation loss and the center conductor width can be chosen

independently for the line impedance, which leads to low dispersion and conductive losses

[2.19].



(a) (b)

Fig 2.9 Examples of CPW- Slotline transition

2.6 Factors Affecting the Radiation

Schaubert et al. [2.20 - 2.26] has done an extensive parametric study of the Vivaldi

antenna. The inferences drawn in his work have set major thought process during this project.

The major parameters affecting the performance of the Vivaldi antenna are:

Dielectric constant of the substrate

Diameter of the slotline cavity and the stripline stub

Input stripline and slotline width

Taper profile

Aperture height

The following paragraphs describes briefly about the effect of these parameters.

2.6.1 Effect of Dielectric

The performance of a tapered slot antenna (TSA) is sensitive to the thickness and

dielectric constant of the dielectric substrate. The presence of a dielectric substrate has the

primary effect of narrowing the main beam of the antenna. The substrate thickness primarily

affects H plane beam width. As the thickness increases the H-plane pattern becomes

narrower. The increase in substrate thickness causes increase in cross polarization level.

Increasing the dielectric thickness generally results in increased gain, but with higher side

lobes. For good performance, a TSA should have an effective substrate thickness in the range

of 1.0025λo ≤ teff ≤ 1.028λo, where teff = t(√εr - 1), is the effective thickness of the substrate.

For substrate thickness above the upper bound of 1.028 λo, unwanted substrate modes

develop which degrade the antenna performance resulting in low efficiency and narrow



bandwidth, particularly for dielectrics with high dielectric constants. Further, for millimeter-

wave operations, the upper bound on the effective thickness, constrains to using mechanically

fragile substrates with thickness of only a few hundreds of microns. The methods to

overcome these problems by increasing the effective substrate thickness and suppressing the

excitation of the surface modes are explained in [2.27 – 2.28].

Kasturi and Schaubert [2.23] studied the effect of dielectric permittivity on infinite

array of single polarized Vivaldi antennas. Fig 2.10 illustrates the result from their study.

Fig 2.10 Effect of Dielectric Permittivity on Infinite Vivaldi Array

2.6.2 Effect of length, width and taper profiles

Tapered slot antennas radiate in the end fire direction with symmetric radiation

patterns, and have cross polarization level of -20 dB or lower. But it has significant cross

polarized radiation in the diagonal plane. The lowest cross polarization level in the diagonal

plane is about 10 to 15 dB higher than that of the principle planes which are typically better

than -15 dB. In general, the cross polarization characteristics of planar TSA are superior to

those corresponding to their antipodal counter part.



Being a travelling wave antenna, the phase velocity and guide wavelength λg varies

with any change in the geometrical and material parameters of the antenna, which in turn

impacts the radiation characteristics of the antenna. The gain of a TSA increases with the

length L of the antenna, typically from a few dB to over 10 dB as L increases from 2-5λ g

[2.29]. Maximum measured gain of 16-17 dB with radiation efficiency of 80% has been

reported for long TSA with L greater than 6 λo [2.30]. The beam widths, decrease rapidly as

the length is increased which is obvious. The H-plane beam width varies more slowly in

comparison to the E-plane beam width particularly for L less than 5 λg. Fig 2.11 illustrates the

variation in SWR with variation in tapered slot length at broad side scan for dual polarized

Vivaldi antenna array [2.21].

Fig 2.11 Effect of Tapered Slot length at Broadside Scan onVSWR

The taper profile has been found to have strong effects on both the beam width and

side lobe level (SLL) of the antenna. In general, the beam widths are narrower for CWSA,

followed by the LTSA, and then Vivaldi for antennas with the same length, same aperture

size, and on the same substrate. And the side lobes are highest for the CWSA, followed by

the LTSA, and the Vivaldi. Being a travelling wave antenna, the H-plane beam width

follows1/√L dependence, while the E-plane beam width depends more on the aperture width

or tapered angle [2.31]. Varying the tapered angle will change the phase velocity and hence

λg, which will in turn change the E-plane beam width. Thus, constant beam width in both E

and H plane can be achieved with proper choices of L and tapered angles.



Fig 2.12 Effect of Opening Rate at Broadside Scan on VSWR

Fig 2.12 demonstrates the effect of taper profile on VSWR at broadside scan for a

dual polarized Vivaldi antenna array [2.21]

Schaubert et.al [2.32] have demonstrated a balanced antipodal antenna with elliptical

radiating taper with a constant E and H plane beam width was observed for a LTSA with

tapered angles in the range of 15 to 20 degrees [2.33].

2.6.3 Effect of slot line cavity and stripline stub

The slotline in a Vivaldi antenna is usually terminated with a cavity as can be seen in

Fig 2.4. The stripline extending past slotline on opposite sides of the substrate is terminated

with a stub. The cavity minimizes any reflections at the point of termination of the slotline by

acting as an open circuit. At lower frequencies, the slotline cavity act as a poor open circuit,

so is the large reflections in the VSWR at low frequencies [2.20]. The cavity can be made

either circular or rectangular. The stripline feed is terminated either using a via or a virtual

short using a radial stub.



Fig 2.13 Effect of Slotline Cavity Diameter at Broadside Scan on VSWR

Fig 2.14 illustrates the effect of slotline cavity diameter on the VSWR for a dual

polarized Vivaldi antenna array from Chio and Schubert [2.21].

2.7 Summary

In this chapter Tapered Slot Antennas have been studied and explained in detail. This

chapter discusses the work carried out worldwide on this structure and explains the various

electrical traits of the structure and its variants. Various feeding techniques to obtain a good

impedance match over a wide bandwidth have been explained.

2.8 References

[2.1] P.J.Gibson, “The Vivaldi Aerial”, Proc.,9th European Conf., Brighton, U.K. 1979, pp.

101-105.

[2.2] S.N.Prasad and S. Mahapatra, “A Novel MIC Slotline Aerial”, Proc. 9th European

Microwave Conf., Brighton, U.K. 1979, pp. 120-124

[2.3] L.R.Lewis, M.Fassett and J.Hunt, “A Broadband Stripline Array Element”, IEEE

APS International Symposium, Atlanta, GA, pp-335-337, 1974



[2.4] Pranay Acharya, Hans Ekstrom, Steven S Gearhart, Stellan Jacobson, Jaokim F

Johansson, Erik L Kollberg, Gabriel M Rebeiz, “Tapered Slot Antenna at 802 GHz”,

IEE Transactions on Microwave Theory and Techniques, vol.41, no.10, October

1993, pp. 1715-1719

[2.5] E.Gazit, “Improved design of the Vivaldi antenna”, IEEE Proc., Part H, vol. 135,

No.2, 1988, pp. 89-92.

[2.6] J D S Langely, P S Hall, and P.Newham, “Balanced antipodal Vivaldi antenna for

wide bandwidth phased arrays”, IEEE Proc. Antenna Proag., vol. 143, no. 2, April

1996, pp. 97-102.

[2.7] D.H. Schaubert, “Endfire Tapered Slot Antenna Characteristics”, ICAP89, 4-7

April 1989, vol.1, pp.432-436

[2.8] Richard Q Lee, “Notch Antennas”, http://gltrs.grc.nasa.gov

[2.9] P.Knott and A.Bell, “Coaxially fed Tapered Slot Antenna”, Electronics Letters,

vol.37, no.18, Aug. 2001, pp. 1103-1104.

[2.10] N. Marchand, “Transmission Line Conversion”, Electronics, vol.17, 1944, pp.142-

145

[2.11] Velimir Trifunovic, Branka Jokanovic, “Review of printed Marchand and Double

Y Baluns: Characteristics and application”, IEEE Transactions on Microwave

theory and techniques, vol.42, No.8, August 1994, pp. 1454-1462

[2.12] A.B. Smolders and M.J. Arts, “Wideband Antenna Element with Integrated Balun”,

1998 IEEE AP-S Int. Symposiums Digest, Atlanta, USA, June 1998

[2.13] Knorr J.B, “Slotline Transitions”, IEEE Trans., vol. MTT-22, 1974, pp. 48-554.

[2.14] Schuppert, “Microstrip/ Slotline Transitions: Modeling and experimental

investigations”,IEEE Trans. On Microwave theory and techniques, vol. MTT-36,

1988, pp. 1272-1282.

[2.15] Oraisi and Jam, “Optimum Design of TSA profile”, IEEE Trans. On antennas and

propagation, vol.51, no.8, August 2003, pp. 1987-1995.


http://gltrs.grc.nasa.gov/


[2.16] M. M. Zinieris, R. Sloan and L.E. Davis, “A broadband microstrip to slotline

transition”, Microwave and Optical technology letters, vol.18, no.5, August 5 1988

[2.17] A.H. Atwater, “The design of the radial line stub: A useful microstrip circuit

element”, Microwave J., vol. 28, 1985, pp. 149-156.

[2.18] A.Nesic, “ Endfire slotline antennas excited by a coplanar waveguide”, 1991 IEEE

AP-S International Symposium, vol.2, Ontario, pp.700-702, 1992.

[2.19] R.Q. Lee and R. N. Simons, chapter 9 in “Advances in microstrip and printed

antennas”, John Wiley and Sons, 1997.

[2.20] J. shin and D. H. Schaubert, “A parameter study of stripline fed Vivaldi notch

antenna arrays”, IEEE Trans. On Antennas and Propagation, vol. 47, no.5, May

1999, pp.879-886.

[2.21] D.H. Schaubert and T.H. Chio, “Parameter study and design of wideband, widescan

dual polarized tapered slot antenna arrays”, IEEE Transactions on antennas and

propagation, vol. 48, no. 6, June 2000,pp. 879-886

[2.22] S.Kasturi, A.O. Boryssenko and D.H. Schaubert, “Infinite arrays of tapered slot

antennas with and without dielectric substrate”, proceedings of the 2002 Antenna

Applications Symposium, Monticello, IL., Sept. 2002, pp. 372-390

[2.23] S.Kasturi and D.H.Schaubert, “Effect of dielectric substrate on infinite arrays of

single polarized Vivaldi antennas”, proceedings of the 2003 Antenna Applications

Symposium, 2003.

[2.24] D.H. Schaubert, A.O. Boryssenko and T.H Chio, “Analysis of finite arrays of

wideband tapered slot antennas”, Proceedings of the 2002 URSI General Assembly,

Maastricht, The Netherlands, 2002.

[2.25] D.H. Schaubert and T.H. Chio, “wideband Vivaldi arrays for large aperture antennas”, NFRA International Conference on Perspectives in radio astronomy: Technologies for large antenna arrays, Dwindeloo, Netherlands, pp. 49-57, Apr 1999.

[2.26] D.H Schaubert, S.Kasturi, A.O Borryssenko and W.M Elsallal, “Vivaldi antenna

arrays for wide bandwidth and electronic scanning”.



[2.27] J.S. Colbum and Y. Rahmat Samii, “Printed antenna pattern improvement through

substrate perforation for high dielectric constant material: An FDTD Evaluation”,

Microwave and optical technology letters, vol. 18, no.1, May 1998, pp. 27-32.

[2.28] Thomas J. Ellis and Gabriel M. Rebiz, “MM-wave tapered slot antennas on

micromachined Photonic Band gap dielectrics”, 1996 IEEE MTT-S International

Symposium Digest, pp. 1157-1161.

[2.29] Kai Fong Lee and Wei Chen, “Advances in microstrip and Printed antennas”, Wiley

Interscience, New York 1997. Chapter 9, p. 443.

[2.30] K. Sigfrid Yngvesson, T.L. Korzeniowski, Young-Sik Kim, Erik L. Kollberg, and Jaokim F. Johansson, “ The tapered slot antenna- A new integrated element for millimeter wave application”, IEEE Trans. Microwave and Techniques, vol. 37, no. 2, Feb. 1989, pp. 365-374.

[2.31] T.Thungren, E.L. Kollberg and K.S. Yngvesson, “Vivaldi antennas for single beam integrated receiver”, Proceedings of the 12th European Microwave Conference, 1982, pp. 474 - 481.

[2.32] T.L. Korzeniowski, D.M. Pozar, D.H. Schaubert, and K.S. Yngvesson, “Imaging

system at 94 GHz using tapered slot antenna elements”, Proceedings of the 8th

International Conference on Infrared and Millimeter waves, 1983.

[2.33] P.S.Kooi, T.S.Yeo, and M.S.Leong, “Parametric studies of the Linearly Tapered

Slot Antenna (LTSA)”, Microwave and Optical Tech. Lett., vol.4, no.5, Apr 1991,

pp. 200-206



Chapter 3

Analysis Methods

3.1 Introduction

The numerical electromagnetic simulators solve the Maxwell’s equations. The

Maxwell’s equations are:

∇× E⃗=−∂∂ t

B⃗ (3.1)

∇× H⃗= J⃗+ ∂∂ t

D⃗ (3.2)

∇ . D⃗=ρ (3.3)

∇ . B⃗=0 (3.4)

With associated consecutive equations

B⃗=μ H⃗ (3.5)

D⃗=ϵ E⃗ (3.6)

The actual solution of the Maxwell equations is complex, and for realistic problems,

approximations are usually required. The numerical approximation of Maxwell’s equations is

known as Computational Electromagnetics (CEM)

3.2 Numerical Methods

The differences between various numerical techniques reside essentially in the

following aspects [3.1]:

The electromagnetic quantity that is being approximated;

The expansion functions that are used to approximate the unknown solution;

The strategy employed to determine the coefficients of expansion functions.



The solution of an electromagnetic problem may require finding the electric or

magnetic field, a potential function, or a distribution of charges and/or currents. While these

quantities are related, they have different properties; hence, problem formulations for field,

potential, and charge or current solutions are different. Finding fields or potentials will

require expansion functions in the field space (domain methods), while unknown charge or

current distributions are expanded into functions defined mostly on boundaries (boundary

methods). Finally, there exists a variety of strategies for computing the unknown coefficients,

which involve the inversion of large matrices, implicit and explicit iteration schemes,

evolutionary algorithms etc. The various existing numerical methods employ different

combinations of these aspects.

The widely used full-wave techniques are:

Method of Moments (MoM)

Finite Difference Time Domain Method (FDTD)

Finite Element Method (FEM)

The electromagnetic simulation tool HFSS which is used in this project, is based on

FEM. In the following sections, a brief description of FEM and HFSS are given.

3.3 Finite Element Method (FEM)

The Finite Element Method (FEM) is one of the best-known methods for the solution

of partial differential equations. It is a method for solving a differential equation subject to

certain boundary values.

The FEM may be derived on two view points: one uses variational analysis, the other

weighted residuals. Both start with the partial differential equation (PDE) form of Maxwell’s

equations. The former finds a variational functional whose minimum corresponds with the

solution of the PDE, subject to certain boundary conditions. The latter also starts with the

PDE form of Maxwell’s equations, and then introduces a “weighted” residual (error); using

Green’s theorem, one of the differentials in the PDE is “shifted” to the weighing functions

[3.2, 3.3-3.5]. For most applications, these procedures result in identical equations. In both

cases, the unknown field is discretized using a finite element mesh; typically, triangular

elements are used for surface meshes and tetrahedrons for volumetric meshes, although many

other types of elements are available.



This represents a very large class of electromagnetic engineering applications of the

FEM, including antenna, radar cross-section, microwave circuit and periodic structure

analysis. As with the FDTD method, the FEM does not include the radiation condition. For

closed regions (e.g. waveguide devices or cavities) this is of no concern. However, for open

regions (e.g. radiation or scattering problems), this requires special treatment, and this must

be incorporated using either an artificial absorbing region within the mesh (the numerical

analogy of an anecholic chamber) or using a hybridization with the MoM to terminate the

mesh.

Traditionally, the FEM has been formulated in the frequency domain, although time

domain formulations have also been used for specialized applications.

Ansoft’s HFSS package is widely regarded as the market leader among the

commercially available packages based on FEM. A fairly recent entry, FEMLAB, has also

attracted users.

The strong points of the FEM are the following [3.2]:

Very straightforward treatment of complex geometries and material in-homogeneities.

Very simple handling of dispersive materials (i.e. materials with frequency dependant

properties).

Ability to handle eigen problems.

Potentially better frequency scaling than the MoM – although the requirement to mesh

a volume rather than a surface means that the number of unknowns in the problem is

usually much larger.

Straightforward extension to higher-order basis functions. It is also possible to use

conformal elements to better approximate curved geometries.

The weak points of the FEM include the following [3.2]:

Inefficient treatment of highly conducting radiators when compared to the MoM (due

to the requirement to have some mesh between the radiator and the absorber).

The FEM meshes can become very complex to implement than the FDTD method.

In conclusion, the FEM is the preferred method for microwave device simulation and

eigen problem analysis. Using FEM / MoM hybrids, scattering problems involving



electromagnetically penetrable media and specialized antenna problems can be accurately and

efficiently solved.

3.4 High Frequency Structure Simulator (HFSS)

The basic approach of FEM is to divide a complex structure into smaller sections of

finite dimensions known as elements. These elements are connected to each other via joints

called nodes. Each unique element is then solved independently of the others thereby

drastically reducing the solution complexity. The final solution is then computed by

reconnecting all the elements and combining their solutions. These processes are named

assembly and solution respectively in the FEM [3.3].

Fig 3.1 Tetrahedral Element

FEM is the basis of the simulation in HFSS [3.6]. HFSS divides the geometric model

into a large number of tetrahedral elements, see Fig 3.1. Each tetrahedron is composed of

four equilateral triangles and the collection of tetrahedra forms what is known as the finite

element mesh. At each vertex of the tetrahedron, components of the field tangential to the

three edges meeting at that vertex are stored. The other stored component is the vector field at

the midpoint of selected edges, which is also tangential to a face and normal to the edge.

Using these stored values, the vector field quantity such as the H-field or the E-field inside

each tetrahedron is estimated. A first-order tangential element basis function is used for

performing the interpolation. Maxwell’s equations are then formulated from the field



quantities and are later transformed into matrix equations that can be solved using traditional

numerical techniques. Fig 3.2 illustrates the meshing of a Vivaldi antenna in HFSS.

Fig 3.2 Meshing of Vivaldi antenna in HFSS

3.4.1 Size of Mesh Vs. Accuracy

There is a trade-off among the size of the mesh, the desired level of accuracy, and the

amount of available computing resources.

The accuracy of the solution depends on the size of each of the individual elements

(tetrhedron). Generally speaking, solutions based on meshes using thousands of elements are

more accurate than solutions based on course meshes using relatively few elements. To

generate a precise description of a field quantity, each element must occupy a region that is

small enough for the field to be adequately interpolated from the nodal values.

However, generating a field solution involves inverting a matrix with approximately

as many elements as there are tetrahedral nodes. For meshes with a large number of elements,

such an inversion requires a significant amount of computing power and memory. Therefore,

it is desirable to use a mesh fine enough to obtain an accurate field solution but not so fine

that it overwhelms the available computer memory and processing power.

To produce the optimal mesh, HFSS uses an iterative process, called an adaptive

analysis, in which the mesh is automatically refined in critical regions. First, it generates a

solution based on a course initial mesh. Then, it refines the mesh in areas of high error

density and generates a new solution. When selected parameters converge within a desired

limit, HFSS breaks out of the loop.



3.4.2 The HFSS Solution Process

To calculate the S-matrix associated with a structure with ports, HFSS does the

following:

Divides the structure into a finite element mesh.

Computes the modes on each port of the structure that are supported by a

transmission line having the same cross-section as the port.

Computes the full electromagnetic field pattern inside the structure, assuming that

one mode is excited at a time.

Computes the generalized S-matrix from the amount of reflection and transmission

that occurs.

The resulting S-matrix allows the magnitude of transmitted and reflected

signals to be computed directly from a given set of input signals, reducing the full 3D

electromagnetic behavior of a structure to a set of high frequency circuit parameters.

3.4.3 HFSS Antenna Design Kit

HFSS Antenna Design Kit is a guide that creates parametric HFSS models for a

variety of common antenna types [3.5]. It can be used to easily generate antenna models and

assist in learning proper usage of HFSS for antenna design. The parameters of the initial

model can be easily modified.

The taper profile of the antenna is drawn with the help of Antenna Design kit.

3.5 Summary

In this chapter, at first, the various full wave techniques widely used for

computational electromagnetics is specified. The FEM method is explained briefly. The

simulation tool used in this project, which is Ansoft’s HFSS, has been explained. Next

chapter will describe about the design of Tapered slot antenna, and its simulated results.

3.6 References

[3.1] Daniel G. Swanson Jr., Wolfgang J. R. Hoefer, “Microwave Circuit Modeling Using

Electromagnetic Field Simulation”, Artech house, Inc., 2003



[3.2] David B. Davidson, “Computational Electromagnetics for RF and Microwave

Engineering”, Cambridge University Press, 2005

[3.3] Anastasis C. Polycarpou, “Introduction to the Finite Element Method in

Electromagnetics”, Morgan & Claypool Publishers, 2006

[3.4] Susanne C. Brenner, L. Ridgway Scott, “The Mathematical Theory of Finite

Element Methods”, Third Edition, Springer, 2008

[3.5] John L.Volakis, Arindam Chatterjee, Leo C. Kempal, “Finite Element Method for

Electromagnetics”, IEEE Press, 1998

[3.6] http://www.ansoft.com


http://www.ansoft.com/


Chapter 4

Design of Tapered Slot Vivaldi Antenna

4.1 Introduction

Computer Aided Design (CAD) of the Vivaldi antenna can be described in the

following hierarchical manner:

1. Choice of substrate material

2. Choice of primitive antenna dimensions to start with [4.1].

3. Modeling and Optimization of antenna using CAD software. In this project,

the Finite Element Method (FEM) based software by Ansoft Corporation;

HFSS v12 [4.2] has been used.

4.2 Choice of substrate

Performance of tapered slot antennas is very sensitive to thickness and dielectric

constant of the substrate. The acceptable range of dielectric thickness for good antenna

operation was found to be

0.0025λ0 ≤ teff ≤ 0.028λo, where teff = t(√εr - 1)

Bandwidth and efficiency are generally having an inverse relationship with the

substrate dielectric constant. The line width becomes smaller with increase in dielectric

constant. So, it becomes difficult to realize microstrip lines especially at higher frequencies.

Substrate thickness and dielectric constant are chosen such that microstrip / stripline trace is

realizable. Line width increases with substrate thickness. Hence, an optimum thickness for

the substrate has to be chosen for the correct design. FR4 epoxy substrate with dielectric

constant, εr = 4.4, thickness of 1.6mm and dielectric loss tangent of 0.02 has been chosen in

this project.



4.3 Tapered Slot Antenna Configurations

The design parameters of the proposed TSA are shown in Fig. 4.1. The top layer

shows the microstrip line and the series radial stub used for feeding the tapered slot antenna.

The bottom layer indicates the exponential taper profile [4.3], which is drawn with the help of

an HFSS tool called “Antenna Design Kit” [4.2].

Fig. 4.1 Configurations of the proposed tapered slot antenna

Given the highest frequency of operation (fH), the width, Wtaper of the tapered slot

antenna should satisfy equation given below to circumvent the grating lobes of Vivaldi array.

Wtaper < c

fh√εe

Where εe is the effective relative dielectric constant.

In addition, the TSA has been designed to match at 100Ω instead of 50Ω. Therefore, the

width of the microstrip line feeder Wm should be defined to give the characteristic impedance



(Z0) of 100Ω [4.4]. The relationship between characteristic impedance and width Wm is given

by the equation:

Shin et al. [4.5] demonstrated that the wideband performance of the Vivaldi notch antenna

arrays fed by microstrip line could be improved systematically. After defining the parameters

cited above, all other parameters are optimized with HFSS to get both the compact size and

good performance at the operating band.

The structural design had undergone many optimizations before reaching the final

design specifications. The values of all the parameters shown in Fig 4.1 are given below:

Wslot = 0.6866455mm

Lfeed = 5mm

Ltaper = 55mm

Wtotal = 70mm

Ltotal = 110mm

Dcavity = 10mm

Rstub = 8mm

Wm = 0.686645mm

Wtaper = 35mm



4.4 HFSS simulation setup

Fig. 4.2 Tapered Slot Vivaldi Antenna

The designed Vivaldi antenna has been modeled in HFSS based on primitive design

[4.1] as shown in Fig 4.2. The conductors are assigned perfect electric boundary condition

[4.6]. An air box is defined surrounding the antenna which defines the boundary condition.

The six faces of the air box have been selected and assigned radiation boundary. The

microstrip port excitation is shown in Fig 4.3.



Fig 4.3 Port Excitation

4.5 Simulation Results

4.5.1 Return loss

1.00 2.00 3.00 4.00 5.00 6.00 7.00Freq [GHz]

-50.00

-40.00

-30.00

-20.00

-10.00

0.00

dB(S

(1,1

))

Ansoft LLC Vivaldi_Antenna_ADKv1XY Plot 2Curve Info

dB(S(1,1))Setup1 : Sw eep1

Fig 4.4 Simulated Return loss of TSA

Fig 4.4 shows that TSA can operate between 2.6 to 5.24 GHz with a return loss of

lower than -10dB.



4.6 Vivaldi Antenna Array

The TSA antenna designed above is not fabricated because the width Wm of the

antenna, (Fig.4.1), is designed to give a characteristic impedance of 100Ω. The network

analyser will have a perfect match with the antenna only if the impedance of the antenna is

50Ω.

So, as the next step we have designed and fabricated a 1×2 Vivaldi antenna array and

a 3dB power divider was designed and fabricated to compose the feed network for the 1×2

Vivaldi antenna array. And the impedance at the input port of this array is 50Ω. The design

and fabrication of 1×2 Vivaldi antenna array is explained in the next chapter.

4.7 Summary

The Tapered Slot Vivaldi antenna was designed. The simulation resuts are plotted and

discussed. This antenna is expected to be useful for wide bandwidth arrays.

4.8 References

[4.1]. Yazhou Wang, Aly E. Fathy, “Design of a Compact Tapered Slot Vivaldi

Antenna Array for See Through Concrete Wall UWB Applications”, EECS

Department, University of Tennessee, Knoxville, USA.

[4.2]. http://www.ansoft.com

[4.3]. P.J. Gibson, “The Vivaldi aerial”, in Proc. 9th European Microwave Conf.,

Brighton, U.K., 1979, pp. 101-105.

[4.4]. Constantine A. Balanis, “Antenna theory”, 2nd ed., Wiley, Newyork, 1997.

[4.5]. Joon Shin and Daniel H. Schaubert, “A parameter study of stripline-fed

Vivaldi notch-antenna arrays”, IEEE Trans. On Antenna and Propagation.

Vol. 47, no.5, May 1999, pp. 879-886.

[4.6]. “The 2000 CAD Benchmark”, Microwave Engineering Europe,

http://img.cmpnet.com/edtn/europe/mwee/pdf/CAD.pdf


http://img.cmpnet.com/edtn/europe/mwee/pdf/CAD.pdf

http://www.ansoft.com/


Chapter 5

1×2 Vivaldi Antenna Array

5.1 Introduction

As mentioned earlier, a 1×2 Vivaldi antenna array is designed and fabricated because

the Tapered slot antenna, discussed in previous chapter, is not manufactured due to its

matching problem with the Network analyzer.

The two element array uses the tapered slot antenna as the array element and a 3dB

power divider is used for feeding the array elements. A brief description of 3 dB power

divider is given in the next section followed by the design and fabrication of 1×2 Vivaldi

antenna array and its simulated and measured results.

5.2 3dB power divider

Basically, a 3dB power divider is a passive device which accepts an input signal and

delivers multiple output signals with 0o phase relationship and equal amplitude [5.2]. There is

high isolation between the two output signals. And the insertion loss is 3dB.

Fig 5.1 indicates the 3dB power divider used for the 2-element array.

Fig 5.1 3dB power divider

The widths of the microstrip lines of port 2 and port 3 are designed to match at 100Ω,

and are connected at the ends of microstrip line feeders of the 2 TSA elements. The width of



the microstrip line at port 1 is designed to match at 50Ω impedance and it acts as the input

port of the 1×2 Vivaldi antenna array.

5.3 Design of the 1×2 Vivaldi Antenna Array

The design parameters of the 2-element Vivaldi antenna array is shown in Fig. 5.2

Fig. 5.2 1×2 Vivaldi antenna array

The dimensions of the TSA are remained as such. Also, the TSA elements are not

separated by any distance. So, the Ltotal × Wtotal of the array becomes 122.5mm × 140mm.

The width of the 3dB power divider is 12.5 mm. Therefore Ltotal becomes 110 + 12.5 =

122.5mm.

5.4 Fabrication process

Following steps are involved in the fabrication process:

A precision mask of the top and bottom layers of the antenna are first drawn using the

Corel Draw software and its printout is taken. Remember that the area in which there



is no conductor parts are given black shades and the area in which there is conductor

parts are given white shades. The top and bottom layers drawn is shown below:

(a) (b)

Fig 5.3 Top and bottom layer masks

A 2-sided conductor coated pcb board is taken whose substrate is FR4 Epoxy.

Specified dimension is cut out from it.

First we are going to etch the top layer. So the other side is coated with cello-tape to

protect the conductor layer of that side while etching.

Clean the layer which we are going to etch, using wire-mesh and acetone.

Coat the layer with photosensitive material (generally a negative photo-resist) and dry

it in a dark room to improve adhesion.

Place the mask shown in Fig. 5.3 (a) above the photo-resist.

Irradiate with ultraviolet light. The layer below the black shield is unaffected and the

exposed area gets polymerized.

Develop the photoresist layer by using a suitable chemical to remove the unexposed

part of the photoresist layer.

Rinse in water.

Chemical etching of unprotected metal layer leaving behind the desired conductor

pattern. The chemical used is ferric chloride.

The same steps are repeated to develop the conductor pattern on the bottom layer.



Fig 5.4 shows the top and bottom layers of the fabricated array.

(a)

(b)

Fig 5.4 Top and bottom view of the fabricated 1×2 Vivaldi antenna array

5.5 Experimental setup

Experimental setup consists of network analyzer, interfacing computer and the

antenna array under test [5.5]. The feed line of the antenna is connected to the output port of

the network analyzer which is controlled by the computer. Fig 5.5 shows the measurement

setup.



Fig 5.5 Measurement setup

The return loss can be measured directly by connecting the Vivaldi array to the probe

of the analyzer as shown in the above figure. The antenna should be held in air.

To plot the far field radiation pattern, the Vivaldi array is aligned in the horizontal

axis of a horn antenna and is kept at a distance of 2D2/ λ, where D represents the length of the

diagonal of the horn. Then the Vivaldi array is rotated around its own axis and the power

delivered to the horn at each position is noted. After one complete rotation, what results is the

radiation pattern.

The gain is also measured by aligning the antenna in horizontal axis of the horn at a

distance of 2D2/ λ. There is no need of rotation.



5.6 Measured and simulated results

5.6.1 Return loss

Fig 5.6 Measured and simulated return loss of the 1×2 Vivaldi antenna array

Measured result in Fig. 5.6 indicate that the array can operate from 1.225 to 7.75 GHz

with a return loss of lower than -10dB.



5.6.2 Gain (dB)

Fig 5.7 Measured and simulated Gain Vs. Frequency of the Vivaldi antenna array.

A gain of more than 6dB is sustained over the frequency range of 1.98 to 7.56 GHz,

according to the measured result. The maximum measured gain obtained is 9.32dB at 3.5

GHz.




Let the 1×2 Vivaldi antenna array be aligned in the 3 dimensional space as shown in Fig 5.8.

Fig 5.8 Vivaldi Antenna Array in 3 dimensional space

5.6.3.1 Radiation pattern in XY plane

(a) (b)



(c)

Fig 5.9 Radiation pattern Vs. Frequency in XY plane

It is seen that when frequency increases the main lobe beam width decreases. It can be

explained as follows [5.3]:- The directivity of an antenna according to Friss transmission

formula is given by:

D= 4πA/ λ2

Therefore, the directivity decreases with increase in wavelength. As the wavelength and

frequency are inversely proportional, we can say that directivity increases with increase in frequency.

And hence beam width decreases with increase in frequency.

5.6.3.2 Radiation pattern in XZ plane

(a) (b)



(c)

Fig. 5.10 Radiation pattern Vs Frequency in XZ plane

5.6.3.3 3D polar plot

Fig 5.11 Simulated Far field radiation pattern at 3.8667GHz



5.7 Summary

A 1×2 Vivaldi antenna array was designed and fabricated. The measured and

simulated results are plotted. The return loss indicates that the array can operate in a very

wide bandwidth. A moderate gain is sustained over a certain range of bandwidth. A very

good radiation pattern is also obtained at different frequencies.

5.8 References

[5.1] Yazhou Wang, Aly E Fathy, “Design of a Compact Tapered Slot Vivaldi

Antenna Array for See Through Concrete wall UWB Applications”, EECS

Department, University of Tennessee, Knoxville, USA.

[5.2] http://www.minicircuits.com

[5.3] Constantine A. Balanis, “Antenna theory”, 2nd ed., Wiley, New York, 1997

[5.4] Chao-Tang Chuang, Shyh-Jong Chung, “Synthesis and Design of new printed

filtering antenna”, IEEE Trans, Antennas and Propagation, vol. 59, issue-3, 2011,

pp. 1036-1042.

[5.5] Martens J, Judge D, Bigelow J., “Multiport vector network analyzer

measurements”, Microwave Magazine, IEEE, VOL. 6, issue-4, 2005, pp: 72-81


http://www.minicircuits.com/


Chapter 6

1×4 Vivaldi Antenna Array

6.1 Introduction

An optimized 1×4 Vivaldi antenna array designed for the lower-band UWB (2-4.5 GHz)

is explained in this chapter. Here, the single Vivaldi antennas are fed by cascaded Wilkinson

power divider network. In this chapter, the Wilkinson power divider is explained first and

then the explanation of the design and fabrication of 4-element array is given. And it is

followed by the presentation of measured and simulated results.

6.2 Wilkinson power divider

The three-port hybrid is useful both as a power divider and combiner [6.1, 6.2]. In the

divider application, power entering the input port is split equally and with zero phase

difference between the output ports. All ports are well matched and the output ports are

highly isolated. The generalized form of the hybrid circuit is a T junction followed by a

multiplicity of cascaded pairs of TEM line lengths and interconnecting resistors. Due to

symmetry, the resistors are decoupled from the input port, but they serve an essential function

in providing output-port match and isolation. Each pair of lines and its associated resistor are

referred to as a section. The divider is often made in microstrip or stripline form, as depicted

in Fig 6.1(a); the corresponding transmission line circuit is given in Fig. 6.1(b). Z0 represents

the characteristic impedance. The widths of the two quarter wave sections are designed to

give an impedance of √2 Z0. The resistor had a resistance of 2Z0, and is the reason that it is

matched at ports 2 and 3, and the reason that ports 2 and 3 are isolated. Additional sections

can provide large increase in bandwidth.



Fig 6.1 The Wilkinson power divider. (a) An equal split Wilkinson power divider in

microstrip form. (b) Equivalent transmission line circuit.

A 3-section Wilkinson power divider is used in this project so that a high operative

bandwidth is obtained. The design of this wide band power divider is given below.

6.2.1 Design of wide band Wilkinson power divider

Cohn et al. [6.1], had designed Wilkinson power dividers having one, two, three and

seven sections. He showed that additional sections can provide a large increase in bandwidth.

The exact design configurations given by him for 3-section power divider is given below:-

If Z0 =1, then

Z1 = 1.1124 R1 = 10.0000

Z2 = 1.4142 R2 = 3.7460

Z3 = 1.7979 R3 = 1.9048

Where, Z0, Z1, Z2, Z3, R1, R2 and R3 are shown in Fig 6.2. λ1, λ2 and λ3 indicate the wave

lengths corresponding to the three frequencies which is taken in-between 2 and 5 GHz. The

frequencies taken here are f1 = 2.4GHz, f2 = 3.5GHz, f3 = 4.8GHz. The widths of the quarter

wave sections are designed in order to give the characteristic impedance Z1, Z2 and Z3

respectively. The input port and two output ports are matched at a characteristic impedance of

100Ω, i.e., Z0 = 100Ω .



Fig 6.2 Circuit of a 3-section power divider

A fabricated 3-section Wilkinson power divider is shown in Fig 6.3 [6.3].

Fig 6.3 Fabricated 3-section Wilkinson power divider

Fig. 6.4 indicates the simulated return loss and insertion loss of the power divider. In

the operating band from 2 to 4.5 GHz, the return loss is lower than -15dB and output ports

have almost equal power level with insertion loss around -3.5dB.



Fig 6.4 Simulated return loss and insertion loss of Wilkinson power divider

6.3 Design of 1×4 Vivaldi Antenna Array

A small variation has been made to the array element, i.e. the TSA discussed in

chapter 4. Wtotal in Fig. 4.1 has been reduced from 70mm to 45mm, so that the array become

more compact. All other parameters are remained the same. Fig. 6.5 demonstrates the 1×4

Vivaldi array designed in HFSS.



Fig 6.5 Design of 1×4 Vivaldi antenna array

The TSA and binary Wilkinson power divider network are both matched at 100Ω. At

the feed port the 4-element array, two Wilkinson power dividers were shunted together to

make the array matched to 50Ω. The overall size of the array is 180×150mm.

Fig 6.6 demonstrates the top layer and bottom layer of the fabricated array.



(a)

(b)

Fig 6.6 Top and bottom view of the fabricated 1×8 Vivaldi antenna array



6.4 Simulated and Measured Results

6.4.1 Return loss

Fig 6.7 Measured and simulated return loss of the 1×2 Vivaldi antenna array

Measured results in Fig 6.7 indicate that the array can operate from 1.75 to 4.8 GHz

with a return loss of lower than -10 dB.



6.4.2 Gain

Fig 6.8 Measured and simulated Gain Vs. Frequency of the 1×2 Vivaldi antenna array

A gain of more than 7dB is sustained over the operating range 2 to 4.5 GHz. The

maximum gain obtained is 12.4dB at 4.5 GHz.




Let the 1×4 Vivaldi antenna array be aligned in 3 dimensional space as shown in Fig 6.9.

Fig 6.9 Alignment of 1×4 Vivaldi antenna array in 3-dimensional space

6.4.3.1 Radiation pattern in XY plane

(a) (b)



(c) (d)

Fig 6.10 Radiation pattern Vs. Frequency in XY plane

Fig 6.10 demonstrates a good radiation pattern of the 1×4 Vivaldi antenna array,

which shows that the side lobes 15 dB lower than the main beam in XY plane.

6.4.3.2 Radiation pattern in XZ plane

(a) (b)



(c) (d)

Fig 6.11 Radiation pattern Vs Frequency in XZ plane

6.4.3.3 3D polar plot

Fig 6.12 Simulated farfield radiation pattern at 4.5 GHz



6.4.4 Field plot

2GHz

(a)3GHz

(b)

4GHz

(c)

Fig 6.13 E-field pattern Vs. Frequency



From Fig 6.13 we can observe the sinusoidal field variation along the edges of the

curve. The flared region contributes to most of the radiation. The Figure indicates that the

magnitude of the E-field near the flared region increases with the increase in frequency.

6.5 Comparison between the results of 1×2 and 1×4 Vivaldi antenna arrays

6.5.1 Return loss

Fig 6.14 Measured return loss of 1×2 and 1×4 Vivaldi antenna arrays

Fig 6.14 indicates that the 1×2 Vivaldi antenna array can operate in a wider operating

range with a return loss of lower than -10dB. Above figure shows that the return loss of

2-element array is lower than -10dB in the range of 1.225 to 7.75 GHz, while for 4-element

array the return loss is lower than -10dB in the range 1.75 to 4.8 GHz.



6.5.2 Gain

Fig 6.15 Measured Gain Vs Frequency

Fig 6.15 shows that the gain of 4-element array is far more better than the 2-element

array.



6.5.3 Radiation pattern in XY plane

(a) (b)

(c) (d)

Fig 6.12 Radiation pattern Vs Frequency in XY plane of 2-element and 4-element array



Fig 6.12 shows that the pattern of 1×4 Vivaldi antenna array is more directive than

that of 1×2 Vivaldi antenna array.

6.6 Summary

In this chapter, the design of optimized 1×4 Vivaldi antenna array is discussed. Then,

the measured and simulated results are plotted. The comparison between 2-element array and

4-element array show that 4-element array perform more better than the 2-element array.

6.7 References

[6.1] S.B. Cohn, “A class of broadband three-port TEM-mode hybrids”, IEEE Trans. on

Microwave Theory and Techniques, Vol. MTT-16, Feb 1968, pp. 110-116.

[6.2] D.M. Pozar, “Microwave Engineering”, 3rd ed., Wiley, New York, 2005.

[6.3] Yazhou Wang, Aly E.Fathy, “ Design of a compact Tapered Slot Antenna Array for

See Through Concrete Wall UWB Applications”, EECS Department, University of

Tennessee, Knoxville, USA.



Chapter 7

Conclusion

In this project, a tapered slot Vivaldi antenna array was developed to be a part of the

see through concrete wall detection. Tapered slot antennas and 3dB power divider were

utilized to compose the 1×2 Vivaldi antenna array. Tapered slot antennas and Wilkinson

power divider were utilized to compose the 1×4 Vivaldi antenna array. The configurations of

the tapered slot antennas were optimized to get a compact size. The measured results of the

return loss, gain and radiation patterns demonstrated the good performance of the Vivaldi

antenna array. A moderate gain of about 7dB is sustained over the operating range 2-4.5 GHz

for the 4-element array. At certain frequencies the gain is more than 12dB.


Final Report

Documents

gain gain

dish antenna

antenna designer

developed vivaldi antenna

power gain

vivaldi antenna arrays

hypothetical isotropic

ideal isotropic antenna