S-Band Antenna Array

(1)

S-Band Antenna Array

Mathias Dalevi

(2)

UTH-enheten Besöksadress:

Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0

Postadress:

Box 536 751 21 Uppsala Telefon:

018 – 471 30 03 Telefax:

018 – 471 30 00 Hemsida:

http://www.teknat.uu.se/student

This report presents concepts for a planar active electronically scanned

antenna(AESA). The goal of the project was to devlop a low-weight, low profile, thin, S-band antenna with wide-scan angle capabilities. In the final concept the service aspects of the T/R-modules was also taken into acount in order to allow easy and fast replacements of these components. The antenna was designed and optimised using the commercial software Ansoft HFSS. A prototype of the antenna was constructed and later measured and verified. The final concept is a 2m×2m antenna with an estimated weight of around 320 kg, around 11 cm thick (where the thickness of the antenna element is 1.76 cm) and has a maximum scan angle range of more than 45 degrees (with <–10dB active reflection) in the frequency band 3–3.5 GHz.

Sponsor: Saab Electronic Defense Systems

ISSN: 1401-5757, UPTEC F10 020

Examinator: Thomas Nyberg

Ämnesgranskare: Roger Karlsson

Handledare: Hanna Isaksson

(3)

This document and the information contained herein is the property of Saab AB and must not be used, disclosed or altered without Saab AB prior written consent.

S-Band Antenna Array

Master Thesis By Mathias Dalevi

The work has been carried out at Saab Electronic Defense Systems

Mölndal

Mentor: Hanna Isaksson

Examiner: Thomas Nyberg

Reviewer: Roger Karlsson

(4)

(5)

Abstract

This report presents concepts for a planar active electronically scanned

antenna(AESA). The goal of the project was to devlop a low-weight, low

profile, thin, S-band antenna with wide-scan angle capabilities. In the final

concept the service aspects of the T/R-modules was also taken into acount in

order to allow easy and fast replacements of these components. The antenna

was designed and optimised using the commercial software Ansoft HFSS. A

prototype of the antenna was constructed and later measured and verified. The

final concept is a 2m×2m antenna with an estimated weight of around 320 kg,

around 11 cm thick (where the thickness of the antenna element is 1.76 cm)

and has a maximum scan angle range of more than 45 degrees (with <–10dB

active reflection) in the frequency band 3–3.5 GHz.

(6)

Populärvetenskaplig Sammanfattning

Det här arbetet presenterar koncept för en aktiv elektronisk styrd antenn(AESA). Arbetet är ett sammarbete mellan två examenarbeten, elektriska och mekaniska koncept, där de elektriska koncepten innefattar framtagning och optimering av antennelementen. Målet med projektet var att ta fram en plan, lätt och tunn AESA med stora utstyrningsmöjligheter i S- bandet mellan 3–3.5 GHz.

En AESA har många fördelar jämfört med en konventionell mekaniskt styrd antenn eftersom den kan rikta in loben elektriskt och därmed snabbare scanna av ett område. Detta är möjligt eftersom varje antennelement styrs av sin egen T/R-modul (sändar/mottagar-modul). Loben styrs genom att individuellt kontrollera fasen av strömmen på varje antennelement. Det går även att kontrollera amplituden av strömmen vid varje antennelement vilket ger möjigheter som att motverka störningar, få mindre sidlober eller en smalare respektive bredare lob mm. Att varje element styrs av sin egen modul gör även systemet mer pålitligt eftersom det fortfarande skulle fungera trots att en sändar/mottagar-modul gått sönder.

Arbetet inleddes med en litteraturstudie där olika koncept undersöktes och utvärderades för att sedan gå vidare med det mest lovande konceptet som var en aperture kopplad stackad patch. Antennelementet designades och

optimerades med EM-simulatorprogrammet Ansoft HFSS v11.2 där elementet realiserades med periodiska randvillkor och optimerades med en olinjär programmeringsmetod. Efter optimeringen konstruerades en prototyp av antennen bestående av 10×10 element som sedan verifierades och testades.

Resultatet av det simulerade antennelementet visar utstyrningsmöjligheter på

mer än 45 grader i varje plan med reflektioner på mindre än –10dB i bandet

mellan 3–3.5 GHz. De uppmätta resultaten på prototypen skiljer sig något från

de simulerade resultaten på prototypen och visar en något bättre prestanda.

(7)

Preface

This master thesis project has been carried out at Saab Electronic Defense

Systems at Lackarebäck, Gothenburg, in the period September-February. It is

the concluding part of the masterprogram of Engineering Physics at Uppsala

University. I like to thank all the co-workers at Saab Electronic Defence

Systems for their help and support throughout the project and especially my

mentor Hanna Isaksson and Jonas Wingård (co-worker). Finally I would like

to thank Lovisa Björklund for making this project possible and giving me the

opportunity to realize it at Saab Electronic Defence Systems.

(8)

Abbreviations

HFSS – HIGH FREQUENCY STRUCTURE SIMULATOR AESA – ACTIVE ELECTRONICALLY SCANNED ANTENNA PCB – PRINTED CIRCUIT BOARD

VSWR – VOLTAGE STANDING WAVE RATIO

TRM – TRANSMIT RECEIVE MODULE

(9)

1 Introduction ... 1

1.1 Background ... 1

1.2 Goal specification ... 1

2 Basic Radar Theory ... 2

3 Antenna theory ... 4

3.1 Antenna Array ... 4

3.2 Microstrip patch element ... 5

3.3 Aperture coupled patch ... 7

3.4 Surface-wave coupling ... 8

3.5 Grating lobes ... 9

3.6 Wide angle impedance match ... 9

4 Concept and design ... 10

4.1 Overall antenna geometry ... 10

4.2 Antenna concept ... 11

4.3 Aperture coupled stacked patch design ... 12

4.4 Antenna feed ... 15

4.5 Aperture coupled stacked patch final version ... 17

4.6 Quarter wave patches ... 21

4.7 Meander patch ... 22

5 Prototype ... 22

5.1 Antenna parts ... 22

5.2 Mechanical parts ... 25

6 Result ... 26

6.1 Simulated Results of the prototype ... 28

6.2 Simulated results of the optimized antenna ... 30

6.3 General measurement theory ... 32

6.4 Measured Results ... 33

7 Conclusion and discussion ... 40

8 Future Recommendations ... 40

References ... 41

Appendix ... 43

A.1 Material ... 43

A.2 Connector ... 43

A.3 Radar bands ... 44

(10)

1 Introduction

1.1 Background

This work is done as a master thesis at Saab Electronic Defense Systems in Gothenburg and the goal is to develop new electronic concepts for Active Electronically Scanned Antennas (AESA). The work is done together with Christian Norinder and Fredrik Övgård who are responsible for the mechanical concept of the antenna. Since the development of electronical and mechanical solutions of an antenna system such as an AESA are equally important and highly depend on each other, me, Fredrik and Christian worked closely toghether in order to optimize with regard to both aspects.

An Active Electronically Scanned Antenna (AESA) consist of a number of antenna elements aligned in an array. By controlling the phase in each antenna element it is possible to steer the electromagnetic field that propagates from the antenna. In other words this could be explained as scanning the beam or steering the beam [1]. By controlling the amplitude in each element it is possible to shape the beam, supress sidelobes, supress jamming signals etc.

The ability to control the phase and the amplitude is made possible by having a transmit/receive-module (T/R-module) behind every element of the array.

Since each element is controlled by a T/R-module, the system will have a graceful degradation, i.e. if one T/R-module would stop functioning the system would not shut down.

Important electrical performance parameters for an AESA are bandwith, scan angle, standing wave ratio (SWR), antenna gain, polarization, etc. Important mechanical properties are heat generation (cooling), stiffness, weight, accessibility to the T/R-modules and thickness.

When AESA antennas first were introduced to the market it was an expensive technology and only considered as a solution when there was a big financial support and no other solutions were good enough, e.g. fighter air-plane radars.

During the 1990’s the technology had matured to the extent that AESAs were competitive to projects with limited budgets [2]. This has resulted in great improvement of the performance of these systems due to the AESA’s many advantages.

1.2 Goal specification

The goal of this project was to develop concepts for a planar, low-weight, low

profile, S-band, active electronically scanned antenna. Much freedom was

given to the designer, however properties that were required was specified at

(11)

the beginning of the project as well as some goals. The specified goals and properties of the antenna can be seen in Table 1.1 and the properties and their consequence are explained below.

There are many different types of antenna elements that can be used in an AESA but for this project only planar elements were considered. Advantages with planar elements are that they have lower profile compared to other elements such as noches and this is an advantage especially when there is a demand for high structural integration. One type of planar elements are microstrip patch elements which will be brought up and explained in this report. The ambition in this project was to develop an AESA antenna operating in the S-band (see Appendix A3 for definition) capable of scan angles greater than 45 degrees in every direction and maximum antenna element thickness of 3cm. It was also desirable to develop mechanical solutions that can simplify the service procedure of the T/R-modules, to develop an antenna with low profile and at the same time limit the total weight of the antennasystem to around 500 kg, see Table 1.1.

Table 1.1: The requirements for the project.

2 Basic Radar Theory

Radar uses electromagnetic radiation in order to detect and locate reflecting objects [1]. The technique is basically to send a signal and if the signal returns, the comparing of the echo signal with the original signal can give the location, speed, and size of the object. The maximum range of the radar can be

determined by the radar range Equation (2.1) ANTENNA ELEMENT

(ELECTRICAL)

ANTENNA SYSTEM (MECHANICAL)

OTHER

<3 cm thick ~500 kg Design that allows for

an easy way to switch T/R-modules

>45 degrees scan angle

~2x2 meter ² Scalable

0.5 GHz bandwidth in S-band with active reflection < −10dB

Planar Cost-effective

(12)

 

²

^{ }

² ₂ ² ²

2

4 4

1 ,  









   



 



 





_ff _w

t r

R e D

P

P . (2.1)

Where P r is the received power of the antenna, P t is the transmitted power from the antenna, σ is radar cross section, e _ff is the efficiency of the antenna, Г is the reflection coefficient, D(θ,ϕ) is the directivity of the antenna, λ is the wavelength, R is the distance between the antenna and the target, ρ w is the polarisation unit vector of the scattered wave and ρ is the polarization of the antenna.

An AESA steers the antenna beam electronically which means that it can direct its beam much faster than the conventional mechanically steered antenna. An AESA consists of different subsystems in which each subsystem contributes to the performance of the antenna. The subsystems in an AESA are radiating elements (antenna elements), T/R modules, exciter, distribution network and receiver, see Figure 2.1. The properties of each subsystem and the interface between the subsystems are important and affect the overall antenna performance.

The distribution network in an AESA is a multilayer PCB which task is to distribute the signals to the right locations. The T/R-modules transmits and receives the signals and are directly connected to the antenna elements. Since the received signals are weak when they return to the antenna it is important that the T/R-modules are located very close to the antennas in order to minimize losses.

In a conventional radar (without taking into account improvements which can be made with signal-processing) the resolution is determined by the

beamwidth of the antenna [1]. When there is a need to track fast moving objects, such as missiles or airplanes, the resolution of a conventional radar is not good enough. In order to improve the resolution a sum and a difference channel can be used to perform a mono pulse measurement.

(13)

Figure 2.1: A simple illustration of the subsystem of a radar.

3 Antenna theory

3.1 Antenna Array

The radiation characteristics of an array are determined by different factors. To make the theory easier, an ideal case is assumed where every element in the array have the same radiation characteristics.

The radiation of the array can then be controlled by a number of design parameters: the geometrical configuration of the overall array, the relative displacement between the elements, the excitation amplitude and phase of the different elements and the relative pattern of the individual elements [3]. The ability to control these parameters independently will yield good control of the radiation pattern of the antenna. The total field for an array is given by

AF EF

E

_total

  . (3.1)

Where EF is the element factor which is the radiation characteristic for an

individual element and AF is the array factor which depends on the relative

distance between the elements, the excitation phase and the excitation

amplitude. For a rectangular array with the elements spaced a distance d x in

the x-direction and spaced a distance d y in the y-direction (see Figure 3.1 for

geometry) the array factor is given by

(14)

  ⁽ ¹ ⁾⁽ ^sin( ^)cos( ⁾ ⁾

1 1

) ) )cos(

sin(

)(

1 ( 1 1

y x y

x

j n kd

N n

M m

kd m j m

n I e e

I

AF ^  _  _ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^ ^. ^(3.2)

Where β x and β y are the progressive phase shift between elements in the x- direction and the y-direction, I m1 is the excitation coefficient in the x-direction and I _1n is the excitation coefficient in the y-direction.

Figure 3.1: Geometry of a planar array configuration with the elements positioned in a rectangular lattice. Adapted from Balanis, 2005 [3].

3.2 Microstrip patch element

Microstrip patch elements have advantages such as low profile, low cost, easy

manufacturing, and low weight, see Balanis, 2005 [3] for details. However,

there are disadvantages such as a narrow bandwidth. The most simple

configuration of a microstrip patch antenna consist of a thin metallic strip

placed on top of a dielectric layer placed above a ground plane (thin metallic

layer), see Figure 3.2.

(15)

Figure 3.2: The basic configuration of a microstrip patch with a microstrip feeding arrangement. After Balanis, 2005 [3].

The microstrip patch in Figure 3.2 has a bandwidth of around 1-3% which is quite narrow. There are some ways to enhance the bandwidth at the expense of a more complicated structure and it is feasible to obtain a bandwidth of about 90% when scanning the antenna at broadside [4]. A stacked patch has higher bandwidth and consists of two or more radiating patch elements, the original patch and the parasitic stacked patch. The stacked patch is placed on top of a new dielectric layer above the original patch. This will enhance the bandwidth to about 14% because a new resonance frequency is introduced. By using an electromagnetic simulation program, for example ansoft HFSS, it is possible to optimise the dimension of the patch and the stacked patch and in this way achieve a greater bandwidth.

There are some guidelines when designing a microstrip element and the procedure according to the transmission line model is explained below [3].

First determine the shape of the microstrip patch such as rectangular or circular etc. The rectangular shape is the most widely used geometry and depending on the shape of the patch, the design procedure varies.

Here a rectangular patch is assumed and the first step of the design is to specify the desired frequency (f), the height (h) of the substrate and the dielectric constant of the substrate (ε r ). The choice of ε r and h is important for the performance of the antenna patch. ε _r normally lies in the interval 2.2<

ε _r <12 and h normally lies in the interval 0.003λ ₀ < h<0.05λ ₀ where λ ₀ is the wavelength of the electromagnetic wave propagating from the patch at its resonance frequency f.

As a rule of thumb the initial value for the width of the patch that normally

leads to efficient radiating characteristics can be calculated from

(16)

12 0

1 2 2



 

 





 

r

f

r

W C

 ^. ^(3.3)

Because of the finite dimension of the patch, the electromagnetic fields along the edges undergo fringing which basically means that the antenna looks longer from an electromagnetic point of view. This affects the resonance frequency of the patch. The extent of the fringing depends on the ratio W/h together with the dielectric constant of the substrate ε r , where W is the width of the patch and h is the height of the substrate.

Since W/h is large for microstrip patches, the fringing depends mainly of the dielectric constant ε r . When W/h>1 the effective electric constant is

2 1

,

1 12

2 ) 1 ( 2

) 1

(

^

 

  

 

 

W

r

h

r eff r



  . (3.4)

When the effective dielectric constant is known it is possible to calculate the effective length of the patch according to

 

  



 



 



 

 



 



 

8 . 0 )

258 . 0

264 . 0 3

. 0 412

. 0

, ,

h W h W h

L

eff r



(3.5)

L L L

_eff

  2 

. (3.6)

It should be noted that the transmission line model is not very accurate and by calculating the dimension of the patch according to this model, the resonance frequency will probably deviate from the desired one. This model is therefore best used in order to calculate initial values which are later optimised in an electromagnetic simulation program.

3.3 Aperture coupled patch

Another way to enhance the bandwidth is to use a more complicated feeding

structure, for example, aperture coupling or proximity coupling. The aperture

coupled feeding structure illustrated in Figure 3.3 consists of five layers.

(17)

The 1 ^st layer is a ground plane, the 2 ^nd layer is a dielectric, the 3 ^rd layer is prepreg dielectric (thin dielectric that is used to bond the 2 ^nd and 4 ^th layers together), the 4 ^th layer is another dielectric and the 5 ^th layer is another ground plane. The 2 ^nd layer has a microstrip line on top of it which is used as a feeding point for the structure.

The 5th layer is a ground plane with a slot which is excited by the microstrip line in the 2 ^nd layer. This feeding structure could then be used to feed a patch which would be placed on top of a third dielectric layer placed above the ground plane with the slot. This configuration introduces a quite complex feeding structure but the gain will be a bandwidth of around 14% for a single patch

Figure 3.3: Aperture coupled configuration.

To further enhance the bandwidth it is possible to use an aperture coupled stacked patch configuration. By using this configuration it is possible to obtain a bandwidth of around 90%. The common denominator, between all

bandwidth-enhancement techniques is a more complex, less compact structure.

3.4 Surface-wave coupling

A problem with patch antennas is that they excite surface-waves, which are

guided by the substrate and the ground plane. In array applications, the

surface-wave coupling between antenna elements could severely degrade the

performance of the antenna. According to Nikolic, 2005 [5], this coupling

becomes important when the normalized electrical thickness h/ λ ₀ of the

substrate has a value that fulfill the relation

(18)

r

h



 2  3 . 0

0

 . (3.7)

When Equation.(3.7) is satisfied it is important to suppress these surface- waves in order to improve the performance of the antenna. This can be done by enclosing each antenna in a cavity which in the simplest way is done by surrounding each antenna with metalized via-holes that are in contact with a ground plane.

3.5 Grating lobes

A grating lobe is defined as a maxima other than the principal maxima of radiation that occurs when the spacing between elements are large enough (larger than λ/2) to permit in-phase addition of radiated fields in more than one direction [3]. Assume a one dimensional array with N isotropic radiating elements, uniform excitation and a spacing of distance d y (see Figure 3.1). The the criterion for no grating lobes is then given by

) sin(

1 1



0



^y

 

d . (3.8)

Where θ ₀ is the scan angle and λ is the wavelength of the highest frequency of the operating band.

3.6 Wide angle impedance match

When scanning an array in different directions the reflection coefficient changes.

This phenomenon results in degradation of the performance of the antenna

array. One approach to solve this problem is to use a Wide Angle Impedance

Match (WAIM) layer spaced in front of the array [6]. A WAIM layer consist

of one or several dielectric layers that match the reflection coefficient to wider

scan angles which improves the scanning capabilities of the antenna. A WAIM

layer could be the radome of the antenna.

(19)

4 Concept and design

4.1 Overall antenna geometry

When designing an AESA there are basically two different architectures that can be chosen, the tile-architecture or the brick-architecture. The different architectures describe the orientation of the T/R-modules relative the antenna elements. When using the tile-architecture, the T/R-modules are parallel to the antenna elements and when using the brick-architecture the T/R-modules are positioned perpendicular to the antenna elements. In this project several concepts involving both brick and tile architectures have been considered but the concept that was most promising uses tile-architecture

The most promising tile-concept has T/R-modules connected to the antenna elements through double-sided T/R-modules (see Figure 4.1) which makes this concept very thin and the main reason why this concept was chosen.

Figure 4.1: Illustration of how the T/R-modules would be connected to the distribution network when normal SMP-connectors are used.

Normally a radar-system is designed so that behind the antenna elements there are T/R-modules and behind the T/R-modules there is a distribution network.

This design makes it difficult to reach the T/R-modules in order to perform maintenance work. Therefore, the decision was made to place the distribution network in front of the T/R-modules which makes the service procedure of the T/R-modules easier. It also makes is possible to combine the antenna elements and the distribution network into one solid piece which will contribute to the stability of the antenna and simplify the manufacture. Another advantage is that the contact connecting the T/R-modules with the antenna elements could be removed (see Section 4.4).

Antennaelement Distributionnetwork

SMP connector cccocontact

TRM

(20)

4.2 Antenna concept

At the beginning of the project, literature was studied in order to investigate what kind of antenna elements that would be suitable for this project. After carefully considering the three most interesting elements, the aperture coupled stacked patch was chosen to be the element on which simulations was to be performed.

Important design issues were the array lattice geometry and the inter element spacing. The choice of the inter element distance depends on the highest operating frequency but should also be chosen so that there are no grating lobes, see Equation.(3.8). After some research and discussions with the project members the decision was made to use a triangular lattice, see Figure 4.2. A triangular lattice was chosen because it minimizes the amount of T/R-modules in the antenna array [1].

This means that less T/R-modules are required for an analogous performance and therefore a more cost-effective solution is obtained. The downside of choosing a triangular lattice is a more complicated geometry. In this project, the inter-element spacing was chosen to be around half of the wavelength λ corresponding to the highest frequency in the operating band.

Figure 4.2: Illustration of the geometry of the triangular lattice used in this project. From Skolnik, 1990 [7].

The three antenna elements chosen after the literature study were an aperture

coupled stacked patch, a quarter-wave patch and a meander patch element.

(21)

A quarter-wave patch is smaller than a λ/2 patch. Smaller elements means that the mutual coupling between the elements could be reduced which enables a wider scan angle capability in one of the planes. Another good property is the wide beamwidth. The reason not to choose this element was that because of the connection of the element to the ground plane there will be an anti- symmetrical current distribution which will contribute to higher cross- polarization of the field. For some applications it may be desirable to have very low cross-polarization. Another reason not to choose the quarter-wave patch was that there has been few published report on this element in array applications.

The meander element was interesting mainly because of a published report from a group in India that illustrated properties necessary in this project, i.e. S- band, wide scan angle up to 60 degrees and low profile [9]. After careful consideration, the decision was made to leave this element due to doubts concerning the feasibility of obtaining these properties with the given element.

The aperture coupled stacked patch was developed by FOI [8] and good result had been obtained, both in simulations and in practice. Another group from Italy [4] had also obtained good result from simulation with a similar element.

These two observations were the major reasons of choosing this type of element.

4.3 Aperture coupled stacked patch design

The aperture coupled stacked patch was the chosen element after the literature study and the design of the antenna element has been carried out in the

commercial software Ansoft HFSS v11.2. The element has been simulated with periodic boundary conditions (infinite array) and has been optimised with a non linear programming method. The design procedure for this element is explained below.

The first step in the design procedure for this antenna element was to begin with an element that already was designed before that had properties that were required for this project. In this case the antenna that FOI had developed [8]

was used as a starting point when recreating and simulating the model in

HFSS, see Figure 4.3 and Figure 4.4. This element has been designed for both

vertical and horizontal polarization but in this project only one polarization has

been used.

(22)

Figure 4.3: Original antenna element that was used as starting point in the design [8]. The black bolded lines represents layers of copper (patch and ground plane), the red bolded lines represents microstrip lines (fork).

Figure 4.4: Illustration of the feeding arrangement from the element which originates from [8].The structure is constructed to radiate two polarizations, vertical and horizontal. It consists of two perpendicular microstrip-forks and two perpendicular H-slots.

Rogers 4350 ( _r = 3.48) Rogers 4403 ( _r = 3.17) Rogers 4450B ( _r = 3.58) Rogers RT/duroid ( _r = 2.2) Rohacell HF 71 ( _r = 1.09)

0.76 0.25 1.14 1.2 0.51

[mm]

(23)

The element was first recreated for the frequencies that the element originally had been designed for (X-band and K _u -band). The second step was to rescale the element to work in the S-band and this was done by multiplying all the dimensions of the element with a scaling factor S f which was calculated with Equation.(4.1)

) _

(

) _

(

element desired

f

element original

S

_f

 f . (4.1)

Where f(original_element) is the highest frequency of the operating band for the original antenna element and f(desired_element) is the highest frequency of the operating band for the desired antenna element.

It was found that there was a complex dependence between the feeding arrangement and the lower patch. This means that the element behaved differently after the scaling-adjustment of the patch dimensions and optimization of the element was required in order to make it function in a similar way as before. The element was then placed into a triangular lattice.

This was followed by optimizing the different dimensions of the element: the dimension of the feeding fork, the thickness of the dielectric layers, the placement of the via-holes, the dimension of the H-slot and the dimensions of the patches (see Figure 4.9 and Figure 4.10 for illustration of the element).

A parameter that greatly influences the performance of the element is the width of the two parallel striplines, which build up the fork. This parameter influences the impedance of the fork and since it is directly connected to the feeding structure it is important that they are matched, i.e. they have the same impedance. Another example are the heights of the first two dielectric layers, which build up the feeding structure (see Figure 3.3). This influence can be explained by the importance of the positioning of the fork relative to the ground plane with the H-slot and the patches, this is important for the coupling of the different parts of the element. Other important parameters are the

dimensions of the lower and upper patches and their positioning relative one another. It was important that these parameters were optimized since these dimensions greatly would influence the performance of the antenna. The next step in the design was to take the mechanical aspects under more careful consideration since this project has intended to produce a low-weight solution.

To reduce the weight of the antenna element, the thickness of the dielectric

layers were reduced and replaced by a low-weight distance material called

Rohacell, which from an electromagnetic point of view behaves as air. Since

Rohacell is electrically shorter than the dielectric material it replaced, the new

Rohacell layers had to be thicker.

(24)

The feeding arrangement had to be re-optimized with the rohacell layer because it turned out that there was a dependence of the positioning of the feeding arrangement as well. After the procedure of reducing the weight of the element it was time to replace some of the dielectric layers that was used in the element.

RT/duroid 5880 which was used initially is a Teflon based material and it is difficult to attach it to other materials with glue. This means that if this material would have been used the antenna would have been difficult to assemble. The materials Rogers 4350B and FR-4 were chosen instead because they function well in the operating frequency range used in this project, they are easy to assemble and they are used widely in industry which makes it fast and easy to order and receive the materials. The next step was to introduce a more practical arrangement to feed the microstrip fork. Until this point, the feed had been a coaxial probe which is impractical to manufacture.

4.4 Antenna feed

The antenna feed consist of a metalized via hole, called a signal-via, which connect the microstrip fork with a SMP connector. The signal-via goes

through the Rogers 4350B layers to a etched pattern at the lower ground plane, see Figure 4.5. This pattern is connected to the SMP-connector which enables the signal to go from the SMP-conncector to the microstrip fork.

Figure 4.5: The figure shows the etched pattern in the ground plane which is used in the prototype (the vertical lines doesn’t represent an etched pattern and should be interpreted as part of the ground plane. The signal-via is indicated by the grey dot. The distance between the pad and the ground plane is 0.250 mm in the prototype but has been adjusted to around 0.110 mm for the optimized element.

In order to prevent parallel plate modes from propagating it was necessary to

enclose each antenna element in a cavity. This was done by surrounding each

antenna element with metallised via-holes which connect the lower ground

plane and the ground plane with the H-slot. In the simulations, the vias that

build up the cavity were simulated as perfect conducting sheets in order to

reduce the complexity of the system.

(25)

The connection transition used in the prototype is different from the one that would be used when designing the whole antenna system, i.e. with distribution network and TR-modules. The distribution network and the antenna elements will be made as one solid piece which will contribute to the stiffness of the antenna and make the manufacture simpler. A stiffer antenna will need less supporting materials and the antenna will therefore be lighter.

The SMP-connectors, see Figure 4.1, will be replaced with another kind of RF-connector, see Figure 4.6, which will reduce the number of RF-connectors used in the system. In the prototype, a SMP-connector is used to feed the antenna element through a signal-via which is connected to the microstrip fork. When the distribution network is added the idea is that the signal-via will go all the way from the upper Rogers4350B layer (the dielectric layer right underneath the ground plane with the H-slot) down to the other side of the distribution network. The signal-via is then connected to a RF-connector which is mounted on the distribution network and finally the RF-connector is connected to a T/R-module.

The RF-connector is a kind of computer connector (see Figure 4.6) which is able to process both power and several RF-signals in one unit which means that instead of using several SMP-connectors and a power supplier it would only be necessary to use one computer connector. By using the computer connector the complexity of the system could be reduced and a more cost- effective solution would be provided. The explanation for this is that 4 SMP- connectors costs much more than 1 computer contact. The computer contacts also serve the purpose of holding the T/R-modules which means that no additional structure is needed for this.

When designing this system it will be necessary to widen the signal-via which goes through the antenna and the distribution network. This is necessary in order to guaranty a fully metalized via-hole because its diameter d has to fulfil the relation

h

d  7 . (4.2)

Where h is the height of the signal-via and d is the diameter.

(26)

Figure 4.6: Illustration of the computer-contact which is used instead of the 4 SMP-connectors in Figure 4.1.

4.5 Aperture coupled stacked patch final version

The final version of the aperture coupled stacked patch consists of several layers as shown in Figure 4.10 where the first layer is a Rogers 4350B dielectric layer placed above a ground plane with a microstrip fork etched on top of it, (see Figure 4.7 for optimized dimensions of the fork).

The next layer is another Rogers 4350B layer which has a ground plane with an H-slot on top of it, see Figure 4.8 for optimized dimensions of the H-slot.

From the ground plane with the H-slot to the other ground plane, near the edges there are metallised via holes that connect the two ground planes to each other. In this way the lower structure (red in Figure 4.9) becomes a cavity that isolates the individual elements from unwanted radiation from neighbouring elements. This cavity also prevents that unwanted parallel modes will occur.

The next layer is a layer of rohacell which is followed by a Rogers 4350B dielectric layer with a microstrip patch on top of it. The structure then

continues with a layer of rohacell and above the rohacell there is a layer of FR- 4 with a stacked patch at its bottom side, see Figure 4.9 and Figure 4.10 for geometry.

In addition of holding the stacked patch, the FR-4 could work as a radome, i.e.

a protecting layer against weather conditions. However in most cases it will be

necessary to place a conventional radome on top of the element because the

antenna will not be stiff enough in order to allow for building the complete

2×2 meter ² system. A conventional radome means that there is a quarter

wavelength of rohacell with a protective layer on top of it such as FR-4.

(27)

The quarter-wavelength of the distance-material is required so that the

radiation not will be exposed of humidity and other weather related conditions because it could affect the radiation characteristics of the antenna. The radome for this antenna will apart from making the antenna more solid and protecting it from weather conditions, also serve as a WAIM-layer, see Section 3.6. This means that the radome will help the antenna to be matched for high scan angles and boost its performance. Note that the dielectric constant of the rohacell in the radome has another value than the other rohacell, see Figure 4.10. The ultimate 2 mm layer of FR-4 can with advantages be switched to another material called cyanate ester which has both better mechanical and electrical properties than FR-4.

Figure 4.7: The optimised dimensions of the microstrip fork. Both the red lines and the parts which are enclosed by the red lines build up the microstrip fork.

14mm

1.46 mm mm

1.66mm

2.09mm

(28)

Figure 4.8: The optimized dimensions for the H-slot. Both the red lines and the parts that are enclosed by the red lines are part of the slot.

Figure 4.9: A transparent view of the aperture coupled stacked patch configuration.

16.054mm

2.45 mm

21.25mm

(29)

Figure 4.10: Illustration of the layers of the Aperture coupled stacked patch configuration and the distribution network (light-green bottom plate).

Apparaturkopplad stackad patch Thickness (without distribution network and radome) =17.589 mm

Ground-plane of copper (black)

Thickness=0,052 mm Rogers 4350B: (red) Thickness=2,388 mm Rogers4450B: (green) Thickness=0,102 mm Rogers 4350B: (red) Thickness=0.946 mm Ground-plane of copper (black)

Thickness=0,035 mm Rohacell: ( εr=1.09, white)

Thickness=4,7 mm Patch of copper Thickness=0.017 mm Rogers4350B: (red) Thickness=0,762 mm Rohacell: ( εr=1.09, white)

Thickness=8,17 mm FR-4: (purple) Thickness=0,4 mm Patch of copper Thickness=0.017 mm

Distribution-network 12 lager Rogers 4350B Thickness=3mm, (light- green)

Radome:

FR-4(or cyanate ester):

Thickness= 2mm.

Radome:

FR-4: (purple) (or cyanate ester) Thickness= 2mm

Rohacell ( ε r=1.5, white):

Thickness =20 mm

(30)

4.6 Quarter wave patches

A quarter-wave patch, also called a planar inverted-F antenna (PIFA-antenna) is illustrated in Figure 4.11. The quarter-wave patch is put together in the same way as a regular patch with the difference that a shortening pin is used to terminate the patch to the ground plane at a point where the electrical field of the resonant mode is zero [10]. By shortening the patch it is possible to reduce the length of the patch to a quarter-wavelength. Holub and Milan, 2008 [11]

has shows that with menaderly folded shorted-patches it is possible to reduce the length of the patch with up to λ/16. However, these configurations are too complex and not practical for array applications because these elements have a narrower bandwidth and the degree of freedom when designing these elements are reduced. Conclusively, it would not be possible, for example, to put

together an aperture coupled stacked λ/16 patch. It should be possible to put together an aperture coupled stacked quarter wave patch without complicating the element too much and this could be done by shortening the patch with a metalized via that connect the ground-plane with the H-slot to the patch. This configuration is more complicated than the regular configuration which has been investigated during this project and one interesting thing to investigate is how much the metalized via would change, if at all, the performance of the antenna. Another downside of using a quarter-wave patch is that the material between the ground-plane and the patch has to be a dielectric material compatible to PCB-technology.

This means that it is not possible to use the low-weight distance material rohacell which has been used in this project. This will increase the weight of the antenna and if this parameter of the antenna needs to be minimized this is probably not a good solution. Another good property which has been

published by Chair el al, 1999 [12] was big bandwidth improvement

compared to the half-wave patch and another bad property was the high cross- polarisation of the field, especially in the H-plane [12].

Figure 4.11: A simple quarter-wave patch configuration. From Waterhouse,

1995 [10].

(31)

4.7 Meander patch

As mentioned earlier, a meander-patch was considered in the literature-study, see Figure 4.12. The specific element considered by Beenamole et al., 2007 [9] doesn’t have the qualities required for this project but the possibility to use this element in a more complicated configuration with broader bandwidth could be very interesting. An example of a configuration which would be suitable is the aperture coupled stacked meander patch configuration which there haven’t, to the author’s knowledge, been any published reports of. The aperture coupled stacked meander-patch is like the regular aperture coupled stacked patch configuration with the difference that the lower or upper patch is switched to a meander-patch instead of the regular patch. However, there is a problem with the dimensions of the meander-patch considered by Beenamole et al., 2007 [9] because it is too large to fit in the available space in the antenna geometry.

Figure 4.12: Dimension and geometry of the meander-patch considered in the literature-study From Beenamole et al., 2007 [9] .

5 Prototype

5.1 Antenna parts

The patch, the stacked patch and the feeding-substrate are all manufactured

with printed circuit board technology (PCB-technology), see Figure 4.10 for

illustration of the different layers of the element. To place an order of the

PCBs, a blueprint of the three boards has been made in the commercial

software Allegro, Figure 5.1 and Figure 5.2 show the blueprint of the PCB

which contain the feeding arrangement of the antenna.

(32)

The company Teltex has been selected to manufacture the PCBs. The materials that have been chosen for this project are Rogers 4350B and FR-4.

They function well in the operating frequency range and they are used widely in industry which makes it fast and easy to order and receive the material from Teltex.

The thickness of the material comes with standardised thicknesses which limit the design freedom of the antenna. One standard thickness of the Rogers 4350B material had been mixed up with another material called Rogers 4003C when it was delivered to Teltex and in order to speed-up the delivery of the product it was necessary to switch this specific PCB to the new material. This switch has caused changes to both the mechanical and electrical properties of the antenna, however since the material has similar properties as the original Rogers 4350B material, the changes are small enough to be acceptable. When the PCBs were obtained from Teltex, the SMP-connectors were surface- mounted on the feeding-substrate in the production-facility at Saab Electronic Defense Systems. Some problems were encountered here due to the large dimension of the PCB and due to the fact that the PCB had an error in its design but these problems were solved.

The rohacell-sheets that were used in this project are called HF71 and the

company Hagema was selected to slice the rohacell into pieces according to

Figure 4.10. Hagema was also chosen to manufacture the aluminium sheets

which support the antenna structure, see Section 5.2. When the RF-connectors

had been mounted it was time for the rohacell, the mechanics and the PCBs to

be attached together with very strong glue at the workshop of Saab Electronic

Defense Systems. In order to glue all pieces together in an accurate way, two

reference-holes have been made at the edges of every piece of the antenna, see

Figure 5.2. By placing a pin in each hole it is possible to glue the pieces of the

antenna in a precise way.

(33)

Figure 5.1: The PCB containing the pattern of the fork and via-holes (green).

The two boards in Figure 5.1 and Figure 5.2 were from the beginning two separate circuit-boards but in a later stage they have been bonded together with a prepreg material called Rogers 4450B that has similar properties as the two boards. This will create a transition region in the material which is anisotropic with varying thickness and dielectric constant. Since this antenna is constructed for the S-band, the wavelength should be long enough so that these irregularities will not have a major impact of the performance of the antenna but nonetheless it will have a measurable effect. Another comment regarding the antenna element in the middle of the first row and the antenna element in the middle of the 10 ^th row is that these elements could be

functioning differently in comparison to the other elements. The reason for

this is that when the PCBs were manufactured, the machine required that there

were two holes in the PCBs and since the size of the antenna array was close

to the actual panel-size available, the drilled holes will be very close to these

two antenna elements and they might be affected by it.

(34)

Figure 5.2: The board containing the pattern of the upper ground plane with H-slots.

5.2 Mechanical parts

Behind the antenna elements there is a 2 mm thick aluminium-sheet in order to

make the prototype more solid. The aluminium sheet was attached to the

antenna elements with glue. Attached at the corners of the sheet there are

supporting structures (see Figure 5.3) which are compatible with the interface

at Saabs measurement room A15 but they also provide the possibility to attach

the antenna to a table for demonstration purposes or for other measurement

purposes. In addition to these there are some “dummy” T/R-modules mounted

on the antenna, see Figure 5.3, which serve the purpose of demonstrating the

manner in which the real T/R modules would be attached to the antenna.

(35)

Figure 5.3: Illustration of the fully assembled prototype as seen from the back/rear

6 Result

The goals of the project were to design a low-weight, thin, S-band antenna

with 0.5 GHz bandwidth with scan angle capabilities of at least 45 degrees at

every plane with less than −10dB reflections see Table 1.1. The actual result of

the project can be seen in Table 6.1.

(36)

Table 6.1: The table illustrates the result of the project.

Thickness of antenna element

Bandwidth (S

₁₁

=−10dB) when scanning up to 40 degrees

Weight

Antenna Properties

1.76 cm 0.5 GHz 317 kg

By comparing Table 6.1 with Table 1.1, the conclusion that the goal of less than 3 cm thick antenna element has been fulfilled with good marginal which also is the case of the total weight of the antenna. The simulated results of scan angle capabilities of 45 degrees at every plane has not been reached for the prototype; however they have been fulfilled with some adjustments of the element, as is shown in Section 8. For the prototype the antenna is only capable of scan angles up to 40 degrees in the H-plane and more than 45 degrees in the E-plane. This result could have been better but since this project was carried out as a master thesis, there wasn’t enough time to finish the optimisation of the antenna element in time before the deadline for prototype manufacturing had been reached. In Table 6.2 the calculations for the weight of the total system is presented.

Table 6.2: The table shows the estimated masses for different parts of the antenna.

Mass [kg]

Calculated values

Contacts 2.4

Antenna layer structure 112

Layer structure support frame 6.6 TRM:s 39

Estimated values TRM covers 2

Cooling structure 10

Exciter & Receivers 70

Cabling 20

Fans 15

Power distribution box 5

Heat exchanger 35

Total mass 317

Required max mass 500

Allowed maximum mass for remaining parts 183

(37)

6.1 Simulated Results of the prototype

Figure 6.1: Active reflection coefficient (S ₁₁ ) for different scan angles for H- plane.

As illustrated in Figure 6.1, the reflection coefficient S 11 for the E-plane in the

frequency range of interest is below −10dB for scan angles up to 40 degrees

and around −8.5dB when scanning 45 degrees. This result is below the goal of

scanning-capabilities of at least 45 degrees.

(38)

Figure 6.2: Active reflection coefficient (S 11 ) for different scan angles for D- plane (diagonal plane).

The reason to show the diagonal-plane is that the antenna elements are

positioned in a triangular lattice and sometimes when this is the case,

performance is degraded in this plane. This is not the result in this case as

illustrated by Figure 6.2. As illustrated in Figure 6.2, the reflection coefficient

S 11 for the D-plane is below −10dB in the frequency range of interest when

scanning the array 45 degrees and below −8dB when scanning 60 degrees.

(39)

Figure 6.3: Active reflection coefficient (S ₁₁ ) for different scan angles for E- plane.

As illustrated in Figure 6.3, the reflection coefficient S 11 for the H-plane is below −10dB in the frequency range of interest when scanning the array 45 degrees and below −8.9dB when scanning 60 degrees.

6.2 Simulated results of the optimized antenna

The simulated results of the optimised antenna have better performance than

the results of the prototype and it is possible to scan the antenna 48 degrees or

more in all the planes.

(40)

Figure 6.4: Simulated results of the active reflection coefficient for the optimised antenna in the H-plane shows scan angle capabilities of up to 48 degrees.

Figure 6.5: Simulated results of the active reflection coefficient for the

optimised antenna in the D-plane illustrate scan angle capabilities of more

than 45 degrees.

(41)

Figure 6.6: Simulated results of the active reflection coefficient for the optimised antenna in the E-plane illustrate scan angle capabilities of more than 45 degrees.

6.3 General measurement theory

The driving impedance of an antenna element in an array has two parts, the self impedance and the mutual impedance (which is caused by other elements in the array or by interfering obstacles), see [3]. The sum of all contributions of an individual antenna elements impedance is called the active reflection coefficient, Γ â . To measure Γ â one must measure the coupling between all the antenna elements in the array, see [13]. In order to measure the coupling between 2 elements in an array it is necessary to terminate all the other elements to 50 ohm and then measure the S-parameters from the two ports of the network analyser, S ₁₁ , S ₁₂ , S ₂₁ and S ₂₂ . To measure the active reflection coefficient for one element one must measure and sum the coupling between this element and all other elements of the array. So, for example, to measure Γ â for one antenna element in an array of 100 elements it is necessary to perform 99 measurements.

The results which HFSS illustrates are of the active reflection coefficient.

HFSS uses periodic boundary condition which means that an infinite number

of antenna elements are assumed in the model. So how well the measured

results coincide with the simulated results depends on the accuracy of the

simulated model and on how many elements there are in the array.

(42)

The active reflection coefficient is given by

m n n mn m

a m

m

V

S V V

V

^





^ 





_,

. (6.1)

Where V _m ⁻ is the backward going wave in element m, V _m ⁺ is the complex voltage feeding element m and S m,n is the coupling between element m and n.

6.4 Measured Results

Figure 6.7 shows the coupling between the 45:th element and all the other elements in the array. It illustrates that the coupling is very symmetrical and lower than −23dB. The figures 6.11−6.19 show that there is a difference between the simulated results and the measured results. The explanations for the differences are mainly due to two effects. First, the simulated results are based on a on a simplified model which is a little different from reality.

Second, the model assumes an infinite array while the prototype consists of a finite number of elements.

Figure 6.7: Measured results of the coupling between the 45:th element and

all other elements in the array, the coupling is given in dB.

(43)

Figure 6.8: The figure shows the reflection of the different scan angles θ in the H-plane. The figure indicates scan angle capabilities of more than 45 degrees.

Figure 6.9: The figure shows the reflection of the different scan angles θ in the

D-plane. The figure indicates scan angle capabilities of more than 45 degrees.

(44)

Figure 6.10: The figure shows the reflection of the different scan angles θ in the E-plane. The figure indicates scan angle capabilities of more than 45 degrees.

Figure 6.11: Comparison of the simulated and measured results for the

prototype.

(45)

Figure 6.12: Comparison of the simulated and measured results for the prototype.

Figure 6.13: Comparison of the simulated and measured results for the

prototype.

(46)

Figure 6.14: Comparison of the simulated and measured results for the prototype.

Figure 6.15: Comparison of the simulated and measured results for the

prototype.

(47)

Figure 6.16: Comparison of the simulated and measured results for the prototype.

S-Band Antenna Array

S-Band Antenna Array

Mathias Dalevi

This report presents concepts for a planar active electronically scanned

Sponsor: Saab Electronic Defense Systems

ISSN: 1401-5757, UPTEC F10 020

Examinator: Thomas Nyberg

Ämnesgranskare: Roger Karlsson

Handledare: Hanna Isaksson

S-Band Antenna Array

Master Thesis By Mathias Dalevi

The work has been carried out at Saab Electronic Defense Systems

Mölndal

Mentor: Hanna Isaksson

Examiner: Thomas Nyberg

Reviewer: Roger Karlsson

Abstract

This report presents concepts for a planar active electronically scanned

antenna(AESA). The goal of the project was to devlop a low-weight, low

profile, thin, S-band antenna with wide-scan angle capabilities. In the final

concept the service aspects of the T/R-modules was also taken into acount in

order to allow easy and fast replacements of these components. The antenna

was designed and optimised using the commercial software Ansoft HFSS. A

prototype of the antenna was constructed and later measured and verified. The

final concept is a 2m×2m antenna with an estimated weight of around 320 kg,

around 11 cm thick (where the thickness of the antenna element is 1.76 cm)

and has a maximum scan angle range of more than 45 degrees (with <–10dB

active reflection) in the frequency band 3–3.5 GHz.

Populärvetenskaplig Sammanfattning

Arbetet inleddes med en litteraturstudie där olika koncept undersöktes och utvärderades för att sedan gå vidare med det mest lovande konceptet som var en aperture kopplad stackad patch. Antennelementet designades och

optimerades med EM-simulatorprogrammet Ansoft HFSS v11.2 där elementet realiserades med periodiska randvillkor och optimerades med en olinjär programmeringsmetod. Efter optimeringen konstruerades en prototyp av antennen bestående av 10×10 element som sedan verifierades och testades.

Resultatet av det simulerade antennelementet visar utstyrningsmöjligheter på

mer än 45 grader i varje plan med reflektioner på mindre än –10dB i bandet

mellan 3–3.5 GHz. De uppmätta resultaten på prototypen skiljer sig något från

de simulerade resultaten på prototypen och visar en något bättre prestanda.

Preface

This master thesis project has been carried out at Saab Electronic Defense

Systems at Lackarebäck, Gothenburg, in the period September-February. It is

the concluding part of the masterprogram of Engineering Physics at Uppsala

University. I like to thank all the co-workers at Saab Electronic Defence

Systems for their help and support throughout the project and especially my

mentor Hanna Isaksson and Jonas Wingård (co-worker). Finally I would like

to thank Lovisa Björklund for making this project possible and giving me the

opportunity to realize it at Saab Electronic Defence Systems.

Abbreviations

HFSS – HIGH FREQUENCY STRUCTURE SIMULATOR AESA – ACTIVE ELECTRONICALLY SCANNED ANTENNA PCB – PRINTED CIRCUIT BOARD

VSWR – VOLTAGE STANDING WAVE RATIO

TRM – TRANSMIT RECEIVE MODULE

Contents

1 Introduction ... 1

1.1 Background ... 1

1.2 Goal specification ... 1

2 Basic Radar Theory ... 2

3 Antenna theory ... 4

3.1 Antenna Array ... 4

3.2 Microstrip patch element ... 5

3.3 Aperture coupled patch ... 7

3.4 Surface-wave coupling ... 8

3.5 Grating lobes ... 9

3.6 Wide angle impedance match ... 9

4 Concept and design ... 10

4.1 Overall antenna geometry ... 10

4.2 Antenna concept ... 11

4.3 Aperture coupled stacked patch design ... 12

4.4 Antenna feed ... 15

4.5 Aperture coupled stacked patch final version ... 17

4.6 Quarter wave patches ... 21

4.7 Meander patch ... 22

5 Prototype ... 22

5.1 Antenna parts ... 22

5.2 Mechanical parts ... 25

6 Result ... 26

6.1 Simulated Results of the prototype ... 28

6.2 Simulated results of the optimized antenna ... 30

6.3 General measurement theory ... 32

6.4 Measured Results ... 33

7 Conclusion and discussion ... 40

8 Future Recommendations ... 40

References ... 41

Appendix ... 43

~2x2 meter ² Scalable

^{ }