Cap d24 Wp32 Uoy Pub 01

FP6-IST-2003-506745 CAPANINA
Deliverable Number D24
Report on steerable antenna architectures and critical RF

circuits performance
Document Number
CAP-D24-WP32-UOY-PUB-01
Contractual Date of Delivery to the CEC
29/09/2006
Actual Date of Delivery to the CEC
29/11/2006
Author(s):
Qin Xu, John Thornton
Participant(s) (partner short names):
CSEM, UOY
Editor (Internal reviewer)
John Gerrits, CSEM
Workpackage:
WP 3.2
Estimated person months
38 PM
Security (PUBlic, CONfidential, REStricted)
PUB
Nature
Report
CEC Version
1.1
Total number of pages (including cover):
85
Abstract:
This report covers a major, three-year work package (WP 3.2) for the CAPANINA programme. The remit
for the research was very broad in scope, as it concerned antenna technologies for high altitude platform
communications. In this context, particular challenges were encountered by the high carrier frequencies
involved, the required high data rates for broadband services to users 'on-the-move', and the consequent
need for highly directive antenna beams to track their targets.
Attention was paid to antennas for a HAP payload and antennas for a high speed vehicle. For the HAP
antenna, where multi-beam functionality is particularly important, the spherical lens antenna was
developed. A two-layer dielectric lens produced an efficiency comparable with a conventional reflector
antenna but with a reduced height profile and provided the potential for multiple scanned beams over a
very wide angle. For the train antenna, a mechanically steered small and light weight array antenna was
identified as the most promising and prototypes were constructed and evaluated. The highly directive
Strip-Slot-Foam-Inverted Patch antenna enabled implementation of an array with a reduced number of
elements; thus decreasing feeding network complexity. A 64-element prototype was tested and the
measured results compared well with the theoretical computations.
Keyword list: High Altitude Platforms, Broadband Patch Antennas, High Gain Patch Array, Spherical
Lens, Luneburg,
Report on steerable antenna architectures and critical RF circuits performance CAP-D24-WP32-UOY-PUB-01
DOCUMENT HISTORY
Date
Revisio
n
Comment
Author / Editor
17/11/06
P06
UOY revision after editorial John Thornton

requests
UOY
21/11/06
P07
Revisions to sect. 4.1 & 6
CSEM
23/11/06
P08
YEC revision with CSEMs Graham Long

contributions
UOY
24/11/06
P09
minor
revisions
authors comments
following Graham Long
UOY
28/11/06
01
Graham Long
UOY
Issued
Qin Xu
Affiliation
Document Approval (CEC Deliverables only)
Date of
approva
l
Revisio
n
28/11/06
P09
28/11/06
01
29/11/2006
Role of approver
Approver
Affiliation
Editor (Internal reviewer)
John Gerrits
CSEM
On behalf of Scientific Board
David Grace
UOY
Page 2 of 85
EXECUTIVE SUMMARY
This report covers a major, three-year work package (WP 3.2) for the CAPANINA programme. The remit
for the research was very broad in scope, as it concerned antenna technologies for high altitude platform
communications. In this context, particular challenges were encountered by the high carrier frequencies
involved, the required high data rates for broadband services to users 'on-the-move', and the consequent
need for highly directive antenna beams to track their targets.
Thus the theme of scanning (or steerable) antennas underpins the enquiries which have been reported.
In chapter 2, possible approaches are considered for communicating with the ground from high aerial
vehicles and the other ground-end of the link, particularly on a moving vehicle such as a fast train, is
examined as well. The work builds to some extent on the earlier work carried out by the HeliNet project
under the 5th Framework Programme. Having defined the scenario, Chapter 3 deals with candidate
solutions for mechanically steered antennas, phased array antennas, spherical lens antennas and
describes the prior art. For the train antenna, a mechanically steered small and light weight array
antenna was identified as the most promising. For the HAP antenna, where multi-beam functionality is
particularly important, the spherical lens antenna was selected for further work.
Chapter 4 describes the experimental work carried out on the two preferred approaches selected in
Chapter 3 and covers the work carried out on the Broadband Printed Array Antenna at CSEM and on the
Hemispherical Lens Prototype at UOY. In both cases there is the assumption that beam steering is
primarily a mechanical operation.
In chapter 5 the practical constraints for antenna steering for a train are examined and Chapter 6 lists
the available electrical and mechanical steering technologies currently available.
Having dealt with train aspects in Chapter 5, Chapter 7 concentrates on a HAP cellular network at the
system level and considers some antenna payload issues, for which the spherical lens antenna was
selected as the most appropriate.
To summarize, there have been broadly two areas of investigation:
(i) Antenna(s) for HAP payload
(ii) Antenna(s) for high speed vehicle.
The requirements for both share much common ground.
For the HAP payload antennas for the millimetre-wave bands, near-term solutions would be provided by
dedicated beam antennas. Medium term solutions are offered by multi-beam lens antennas, where a
lens aperture is shared by a number of feeds, thus minimising payload mass. Long-term solutions may
be offered by digital beam forming or "smart antennas". While all techniques have been considered
during the course of WP3.2, and experimental data gathered for aperture and lens antennas, the long
term solution (digital beam forming) has not been explored in practice under WP32: rather, the signal
processing aspects of this technology lie under the remit of WP3.3
For the vehicular antenna, a similar technology road-map has been identified, i.e. near term (single beam
mechatronic antenna), medium term (multi-beam mechatronic lens antenna), and again the long term
solution offered by Digital Beam Forming (DBF), which is considered elsewhere (WP3.3).
29/11/2006
Page 3 of 85
TABLE OF CONTENTS
INTRODUCTION ..............................................................................................................11
POSSIBLE APPROACHES FOR THE RF, EM AND MECHATRONICS ASPECTS OF

THE TRAIN AND HAP AN TENNAS .................................................................................12
2.1
System Requirements and Link budgets ............................................................................ 12
2.2
Geometry ........................................................................................................................ 14
2.3
Required Scan Range....................................................................................................... 14
2.4
Coverage Area ................................................................................................................. 14
CANDIDATE SOLUTIONS AND PRIOR ART.................................................................14

3.1
Scenario.......................................................................................................................... 14
3.2
Candidate technologies..................................................................................................... 15
3.2.1 Mechanically steered antenna......................................................................................... 15

3.2.1.1
Mass of typical antennas ............................................................................................................................. 15
3.2.2 Mechanically steered small high gain array ...................................................................... 16

3.2.3 Phased array/ digital beam forming: circuit level constraints............................................... 18
3.2.3.1
Array factor .................................................................................................................................................... 18
3.2.3.2
Directivity ....................................................................................................................................................... 19
3.2.3.3
Implications of wide element spacing......................................................................................................... 19
3.2.3.4
Effect of antenna element radiation pattern............................................................................................... 20
3.2.3.5
Effect of beam scanning ............................................................................................................................... 21
3.2.3.6
Sub-array interleaving................................................................................................................................... 22
3.2.3.7
Discussion: phased array techniques ........................................................................................................ 24
3.2.4 Array variants and combined electronic and mechanical steering........................................ 24

3.2.5 Recent projects in phased arrays and Digital Beam Forming (DBF) .................................... 25
3.2.5.1
SUITED project.............................................................................................................................................. 25
3.2.5.2
SANTANA ..................................................................................................................................................... 26
3.2.6 The spherical lens antenna ............................................................................................. 27

3.3
Recommendations following the survey of candidate technologies......................................... 30
3.3.1 Phased array and DBF................................................................................................... 30

3.3.2 Mechatronic Solution ..................................................................................................... 30
3.3.3 Recommendations ......................................................................................................... 31
EXPERIMENTAL INVESTIGATIONS OF PREFERRED SOLUTIONS ........................32

4.1
Broadband Printed Array Antenna...................................................................................... 32
4.1.1 Introduction ................................................................................................................... 32

4.1.2 Techniques for second lobe reduction .............................................................................. 32
4.1.3 Non-uniformly spaced array ............................................................................................ 33
4.1.4 Non-uniform power distribution ........................................................................................ 34
4.1.5 Specific single element radiation ..................................................................................... 35
29/11/2006
Page 4 of 85
4.1.6 Single high gain broadband antenna design...................................................................... 36

4.1.7 Array performance ......................................................................................................... 38
4.1.8 Measurement results ..................................................................................................... 40
4.1.9 Polarizer ....................................................................................................................... 45
4.1.10 Loss consideration......................................................................................................... 46
4.1.10.1
Loss in the antenna................................................................................................................................... 46
4.1.10.2
Loss in the polarizer.................................................................................................................................. 47
4.1.11 Conclusion for printed array antenna................................................................................ 47

4.2
Hemispherical lens prototype. ........................................................................................... 48
4.2.1 Constant index lens ....................................................................................................... 48

4.2.2 Primary feeds ................................................................................................................ 50
4.2.3 Development of a multi-layer lens .................................................................................... 51
4.2.3.1
Analysis of single-shell and two-shell lens designs................................................................................ 51
4.2.3.2
Two-layer lens performance......................................................................................................................... 54
4.2.4 Mechanical feed steering concepts.................................................................................. 55

4.2.5 Multi-beam Lens Antenna Conclusions ............................................................................ 56
4.2.5.1
Acknowledgment........................................................................................................................................... 56
MECHANICAL CONSTRAINTS RELATED TO TRAIN APPLICATION........................57
STATE-OF-THE-ART OF ELECTRICAL AN D MECHANICAL STEERING

TECHNOLOGIES .............................................................................................................60
6.1
State-of-the-Art of Mechanical Steering System .................................................................. 60
6.1.1 Ball Aerospace.............................................................................................................. 60

6.1.2 Directed Perceptions Pan-Tilt Unit................................................................................... 61
6.1.3 QuickSet International.................................................................................................... 62
6.1.4 AvL Technologies........................................................................................................... 63
6.1.5 CSEM custom-design antenna gimbal ............................................................................. 64
6.2
Testbed of Mechanical Steering System............................................................................. 65
6.2.1 Antenna gimbal design based on Commercial Off-The-Shelf modules.................................. 65

6.2.2 Signal tracking algorithm implementation ......................................................................... 66
6.2.3 Tracking algorithm testbed.............................................................................................. 67
6.3
Electrical beam steering: possibility of array built with directive elements .............................. 70
6.4
Conclusions on mechatronic array antenna......................................................................... 71
PROPERTIES OF SPHERICAL LENS ANTENNAS FOR HIGH ALTITUDE

PLATFORM PAYLOAD ....................................................................................................72
7.1
Introduction...................................................................................................................... 72
7.2
Dedicated beams for the HAP antenna payload................................................................... 72
7.3
Properties of Spherical Lens Antennas ............................................................................... 73
7.4
Power Contours and Co-channel Interference...................................................................... 75
7.5
Reducing the spherical lens sidelobe level .......................................................................... 77
7.6
Relative payload mass...................................................................................................... 79
7.7
Conclusions for HAP Spherical Lens Antennas ................................................................... 80
29/11/2006
Page 5 of 85
FINAL CONCLUSIONS ....................................................................................................81

8.1
Bibliography..................................................................................................................... 82
REFERENCES..................................................................................................................83
29/11/2006
Page 6 of 85
LIST OF FIGURES
Figure 1 HAP link geometry ...................................................................................................... 15
Figure 2 Lens-corrected horn antennas..................................................................................... 16
Figure 3 Plastic dual-reflector antenna...................................................................................... 16
Figure 4 Illustration of mechanical steering............................................................................... 17
Figure 5 Pointing angle speed.................................................................................................. 17
Figure 6 Array factor for 4 x 4 rectangular array as a function of element spacing. .................. 19
Figure 7 Directivity on boresight versus element spacing.......................................................... 19
Figure 8 Loss in Directivity associated with grating lobes, 8 x 8 element array......................... 20
Figure 9 Product of array and element patterns for 4 x 4 element array with a = 1.5 (=0 axis). 21
Figure 10 Effect of array scanning.............................................................................................. 21
Figure 11 Array grating lobes..................................................................................................... 22
Figure 12 Array patterns............................................................................................................. 23
Figure 13 Secondary pattern...................................................................................................... 23
Figure 14 Combined pattern, scanned beam ............................................................................. 24
Figure 15 Illustration of combined mechanical and electronic steering. ................................... 25
Figure 16 Possible RF array topology for single axis electronic steering................................... 25
Figure 17 Luneburg lens and hemispherical lens. ..................................................................... 27
Figure 18 Hemispherical lens with elevation scanning feed cluster. ........................................ 28
Figure 19 Hemispherical antenna variants with limited scan..................................................... 29
Figure 20 Sketch of a possible multi-shell lens configuration incorporating radome................. 29
Figure 21 Non-uniform spacing applied to a regular array in which d = 0.5............................ 33
Figure 22 Non-uniform spacing applied to a special array in which d = 2............................... 34
Figure 23 Linear array with uniform, Tschebyscheff and Binominal power distribution (d = 2)35
Figure 24 Combination of AF and special element pattern....................................................... 35
Figure 25 Different layers of a SSFIP antenna .......................................................................... 36
Figure 26 Single element antenna: impedance bandwidth (VSWR) as a function of frequency.36
Figure 27 Co-polarization radiation pattern in E-plane ( = 90) and H-plane ( = 0) at 28.5 GHz37
Figure 28 Measured and simulated impedance matching for adjusted model.......................... 37
Figure 29 16-Element array: impedance bandwidth (S11 magnitude) as a function of frequency38
Figure 30 Simulated gain of the single element compared to a normal patch ......................... 38
Figure 31 Simulated gain of the 22 subarray compared to a normal square patch................ 39
Figure 32 Predicted radiation of 64- and 256-element array based on single element pattern . 40
Figure 33 64-Element prototype ................................................................................................ 41
Figure 34 Measured VSWR of 64-element with respect to frequency........................................ 41
Figure 35 Comparison of normalized co-polarization radiation patterns of the 64-element array
at 27.5 GHz .......................................................................................................................... 42
Figure 36 Measured normalized co- and cross-polarization at 27.5 GHz ................................... 42
Figure 37 Comparison of normalized E-plane co-polarization radiations of the 64-element
array.................................................................................................................................... 43
Figure 38 Measured normalized E-plane co- and cross-polarization at 28.5 and 29.5 GHz........ 43
Figure 39 Enhanced E-plane backward radiation of the 64-element array at 27.5 GHz ............. 44
Figure 40 Four-layer meander line polarizer ............................................................................ 45
Figure 41 Four-layer polarizer layouts...................................................................................... 46
29/11/2006
Page 7 of 85
Figure 42 Measurement jig for proof-of-concept (this one is set up for about 65 elevation
angle). ................................................................................................................................. 48
Figure 43 Radiation patterns for 4 elevation angles. .................................................................. 49
Figure 44 Trends in loss as a function of elevation angle. ......................................................... 50
Figure 45. Measured lens antenna radiation patterns comparing feed types. (Inset: image of
scalar feed.) ........................................................................................................................ 51
Figure 46. Two-shell hemisphere lens geometry ....................................................................... 52
Figure 47. Spherical lens computed directivity D versus feed location for three materials. ....... 53
Figure 48. Components of two-layer lens (left: polyethylene shell, right: Rexolite core.) .......... 54
Figure 49 Two-layer lens radiation patterns at 28 GHz.............................................................. 55
Figure 50. Feed steering concepts (inset: rolling ribbon actuator mechanisms). ...................... 56
Figure 51 Steerable antenna mounted on a TGV roof (http://telecom.esa.int/telecom/www) ... 57
Figure 52 Power spectral Density of lateral carriage body vibration ........................................ 58
Figure 53 Typical additional tilt of up to 8 from new generation tilting trains ......................... 58
Figure 54 Two-axis control gimbal............................................................................................. 60
Figure 55 Nodal version with brackets aligns payload in pan and tilt rotation axes.................. 61
Figure 56 Control interface to a host computer......................................................................... 62
Figure 57 Product features........................................................................................................ 63
Figure 58 Data sheet of AvL Technologies antenna positioner models..................................... 64
Figure 59 CSEM custom-design drive system ............................................................................ 65
Figure 60 CSEM custom-design antenna gimbal solution ......................................................... 65
Figure 61 Directed Perceptions Pan-Tilt Unit in nodal configuration ......................................... 66
Figure 62 Antenna gimbal unit design for CSEM demonstrator.................................................. 66
Figure 63 Spiral trajectory followed by the pan-tilt mechanism ................................................ 67
Figure 64 Maximum finding algorithm...................................................................................... 67
Figure 65 ETS Lindgren model 3116 antenna and interface ..................................................... 68
Figure 66 Block diagram of tracking algorithm testbed ............................................................ 69
Figure 67 Control unit with added trajectory perturbation ........................................................ 69
Figure 68 The 64-element subarray and the polarizer mounted on the Pan-Tilt Unit................ 69
Figure 69 Radiation of a 4-element array using directive element ........................................... 70
Figure 70 Morgan's geometry [12] for generalized Luneburg lens showing aperture plane ...... 74
Figure 71. Radiation patterns for Luneburg lens of 10 wavelength diameter............................. 75
Figure 72. Cellular layout for 121 cells using 3 channels. .......................................................... 75
Figure 73. Concept for antenna payload: multi-beam Luneburg lens for each cell group ......... 75
Figure 74. Power contours for 1 of 3 channels (contour spacing is 1.5 dB)................................. 76
Figure 75. CIR coverage distribution.......................................................................................... 76
Figure 76. Relationship between primary feed and secondary HPBW for Luneburg lens. ......... 77
Figure 77. Increase in secondary beamwidth, compared to that of uniform aperture, as a
function of primary beamwidth. .......................................................................................... 78
Figure 78 CIR distributions for dedicated and multi-beam payloads. ........................................ 78
29/11/2006
Page 8 of 85
LIST OF TABLES
Table 1 Link Budget for 28GHz................................................................................................... 13
Table 2 Pointing angle precision for 2 and 5 beamwidth ....................................................... 18
Table 3 Main simulated characteristics of the single antenna ................................................... 37
Table 4 Simulation results of single element and sub-arrays.................................................... 39
Table 5 Performance of 64-element array................................................................................. 44
Table 6 Estimated efficiency at 28.5 GHz of the single element and array antenna.................. 46
Table 7. Dielectric materials considered for two-shell hemisphere lens. ................................... 52
Table 8. Directivity and focus points of spherical lenses. ........................................................... 53
Table 9 Time in seconds to travel the distance of a train length for estimation of tunnel
obstruction times................................................................................................................. 59
Table 10 Resume of Tip-tilt mechanism requirements.............................................................. 60
Table 11 Ball Aerospace control gimbals product range ........................................................... 61
Table 12 Products available that meet requirements................................................................. 62
Table 13 Control interface to a host computer........................................................................... 63
Table 14. Summary of Luneburg lens beam widening and sidelobe level reduction for various
primary feeds. ..................................................................................................................... 78
Table 15. Comparison of payload properties and CIR values for dedicated beams and multibeam spherical lenses serving 121 cell network ................................................................. 79
29/11/2006
Page 9 of 85
LIST OF ACRONYMS
AF
Array Factor
AR
Axial Ratio
BW
Bandwidth
COTS
Commercial Off-The-Shelf
CP
Circular Polarisation
DBF
Digital Beam Forming
EIRP
Effective Isotropic Radiated Power
FSL
Free Space Loss
HAP
High Altitude Platform
LOS
Line Of Sight
GLL
Grating Lobe Level
MLL
Main Lobe Level
RX
Receiver
SLL
Side Lobe Level
SOTA
State Of The Art
SSFIP
Strip-Slot-Foam-Inverted Patch
TX
Transmitter
29/11/2006
Page 10 of 85
1 Introduction
Todays users of broadband services are generally dependent upon either cable or satellite technologies
for the connection to their home or office. Cable can provide excellent capability, particularly fibre, but is
only commercially viable in high useage areas. Satellite provides widespread geographic coverage but
has limited capability and changing capacity. Communications via High Altitude Platforms (HAPs) will
develop broadband capability from aerial platforms to deliver cost effective solutions providing a viable
alternative to cable and satellite, with the potential to reach rural, urban and travelling users.
High Altitude Platforms (HAPs) are airships or planes, operating in the stratosphere, at altitudes of
typically 17 - 22km (around 75,000 ft). At this altitude (which is well above commercial aircraft height),
they can maintain a quasi-stationary position, and support payloads to deliver a range of services:
principally Communications, and Remote Sensing.
Communications services include broadband, 3G, and emergency communications, as well as
broadcast services. A HAP can provide the best features of both terrestrial masts (which are then not
required) and satellite services (which would be highly expensive). In particular, HAPs permit rapid
deployment, and highly efficient use of the radio spectrum (largely through intensive frequency re-use).
New types of antennas needed to provide these communications, in particular antennas for the HAP
payload and antennas for ground stations of which one of the most difficult to provide will be for a high
speed vehicle such as a train. The technology developed aims to support data rates of up to 120 Mbit/s
to fixed and moving users anywhere within a HAPs 60 km coverage area. In ITU-R [29], the frequency
band allocated for HAPs applications is defined to be 27.5-31.3 GHz.
HAPs offer an effective way of exploiting mm-wave spectrum by supporting multi-cell architectures. The
viability of such systems is largely determined by the overall system data capacity, which is in turn
governed by the properties of the antenna payload which serves the cells on the ground. In previous
studies, ideal antenna beams have been used to model carrier-to-interference ratio. In such cases,
dedicated aperture antennas such as lens antennas may be used to produce the required beam shapes
which are in general asymmetric. This approach leads to one antenna for each cell and thus a bulky
payload. Alternative ways are needed to minimise the HAP payload and to keep the train antenna as
small as possible.
29/11/2006
Page 11 of 85
2 Possible Approaches for the RF, EM and Mechatronics

Aspects of the Train and HAP Antennas
There are elements of the antenna specification that are essentially the same for every antenna type.
2.1
System Requirements and Link budgets
The link budget can be defined to provide differing levels of service, the biggest factor in service level
being the permitted duration of rain outages. A link budget for 28GHz with a HAP at 17km altitude and a
radial distance of 30km from the sub-platform point can be seen in Table 1. This does not make any
assumptions about the transmit power or antenna gains, but does suggest three different modulation
schemes requiring different signal to noise ratios. In addition three different levels of service are defined
99.99%, 99.90% and 99.0% availability, the result at the bottom of the table are 9 different numbers
representing the required sum of transmit power (dBm) plus HAP antenna gain (minus feed losses) plus
customer antenna gain (minus feed losses). It is envisaged that different combinations of antennas and
powers will result in different levels of service provision.
29/11/2006
Page 12 of 85

1
2
3
4
8
9
10
11
12
13
14
15
16
17
18
19
20
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
Transmitter (HAP)
Modulation Scheme
Required Amplifier Input Back-off
Power per backhaul carrier (dBm)
Antenna gain (dBi)
Antenna feed loss (dB)
HAP EIRP (dBm)
Rounded 64QAM
-3
26.0
0.0
0.0
26.0
16QAM
-3
26.0
GMSK
0
29.0
26.0
29.0
Receiver (Ground Station)

The Boltzmann Constant (dBJ/K)
Noise Temperature (K)
Thermal noise density (dBm/Hz)
Receiver noise figure (dB)
Receiver noise density (dBm/Hz)
Receiver interference noise density (dBm/Hz)
Total effective noise density (dBm/Hz)
-228.6
300.0
-173.8
5.0
-168.8
-168.8
-165.8
Antenna gain (dBi)

Cable loss at ground station
Maximum C/(Io+No) (dBHz)
0.0
0.0
191.8
191.8
194.8
Modulation Scheme
Required Eb/No (BER 10-5)
Bit/symbol
Rounded 64QAM
10.4
6
16QAM
6.7
4
GMSK
2.7
1
25.0
0.6912
83
79.2
89.6
25.0
0.6912
55
77.4
84.1
25.0
0.6912
23
73.6
76.3
102.2
107.7
118.5
-50.1
76.1
-39.3
68.3
Bandwidth (MHz)
Code Rate
Data Rate (Mbit/s) (25% rolloff)
Data Rate (dBbit/s)
Required C/(Io+No) (dBHz)
Maximum allowed losses (dB)
Link Parameters
Frequency (GHz)
Wavelength (m)
Ground Distance (km)
Platform Height (km)
28.0
0.011
30.0
17.0
LOS Distance (km)

FSPL (dB)
Misc Atmospheric Losses (dB)
Edge of cell and antenna beam losses
Clear air losses (dB)
34.48
152.1
0.7
5.0
157.8
Received margin clear air (dB)

Minimum required transmit power clear air (dBm)
-55.6
81.6
Rain Attenuation Zone K 99.99% (dB)

32.5
12.3
3.3
Received margin 99.99% (dB)

-88.0
-67.8
-58.8
-82.6
-62.4
-53.4
-71.8
-51.6
-42.6
Minimum required transmit power (dBm) + Antenna Gains 99.99%

114.0
93.8
84.8
108.6
88.4
79.4
100.8
80.6
71.6
Table 1 Link Budget for 28GHz
29/11/2006
Page 13 of 85
2.2
Geometry
The geometry of the train antenna can be flexible with the one constraint that it needs to be
commensurate with installation on top of a high speed train and hence should have a reasonably low
profile.
2.3
Required Scan Range
The sweep angle for the train antenna should be 150. The figure is based on a HAP elevation angle is
30 ( i.e. above the horizon), combined with the attitude of the train (in terms of maximum levels of track
camber etc.) which has been estimated at 15, leading to an effective elevation angle of 15. It is
possible that there might have to be a reduced link specification at these extremes of the coverage area
for example a reduced data rate or more coding.
2.4
Coverage Area
Within the HeliNet project the coverage area was defined as a 60 km diameter circle. This was based on
a maximum elevation angle of the HAP from the ground user of approximately 30
3 Candidate Solutions and prior art

It is intended that this section will contain many subsections each dealing with a different alternative
solution exploring solutions previously reported in the literature as well a new or hybrid solutions.
The antennas could be anything from a full phased array through mixed electrically and mechanically
steered to a single aperture antenna that is entirely mechanically steered, the aim is to explore the
solution space and set out:
advantages and disadvantages
flexibility
complexity
cost
power requirements
performance
geometry
steering rate
3.1
Scenario
The scenario of interest is that of a terminal mounted on a moving vehicle (i.e. a high speed train) which
communicates with a terminal / transponder on an aerial platform i.e. a HAP. This scenario is not too
dissimilar with satellite communications to vehicles. It is thus of much commercial interest and a mainstream topic in communications research and technology.
It is worth listing some of the characteristics and requirements of this scenario:
Scanning to low elevation angles (a figure of 15 has recently been mooted).
thus, about 150 conical scan - almost a full hemisphere.
any loss in the antenna gain associated with scanning to a low elevation angle is particularly
disadvantageous - it is here where maximum gain is need to counter the free-space and rain losses
associated with the path length.
29/11/2006
Page 14 of 85
HAP
~ 20 km
~ 80 km
high elevation
low elevation
rain layer
terminal 2
terminal 1
Figure 1 HAP link geometry
The speed of scanning over large angles is relatively slow - this is associated of the change in
location of the vehicle from place to place.
The speed of scanning over smaller angles may need to be rapid - this is associated with
changes in attitude of the vehicle.
Both transmit and receive functionalities are required.
A multi-beam function would allow for communication with multiple platforms.
Constraints on the terminal: The terminal carried by the vehicle (train) is probably not too constrained by
mass, or area, as we assume it would be roof mounted. It is probably not too constrained by power. It
may have some constraint on physical height, and probably a medium constraint on cost.
3.2
Candidate technologies
The scope here is very large. A few candidates are listed.
3.2.1
Mechanically steered antenna
The antenna might be a reflector, lens-horn, or array type. This would be one of the simplest solutions,
and would have good RF performance. The flexibility of the RF feed, and a possible requirement for a
rotary RF joint, could be an issue. It could be advantageous to include a frequency-converter at the feed
(as with domestic satellite TV receiver) - this would have to be powered, and would add to the steered
mass. The antenna would most likely require housing with an RF transparent cover (radome) to isolate it
from wind stresses, rain etc.
This is not a very elegant solution, although it may be a relatively cheap one. Disadvantages would
include the inability to produce independently steered, multiple beams over a wide angle. High-speed
scanning to maintain track as the vehicle attitude changes may be problematic. Some isolation from
pitch, roll and vibration may be needed. A narrow-angle multi-beam facility should be feasible for reflector
with multiple feeds, and this could assist with high-speed tracking, in a similar way to monopulse radar.
This multi-beam facility would require multiple feeds, further adding to the mass which is steered.
There may be commercially available products which a close to meeting our requirements, e.g. ship
mounted satellite terminals, although the majority of these are likely to be for Ku band.
3.2.1.1
Mass of typical antennas
In this section some typical antennas for 28 GHz are shown as these can provide a first estimate of
typical dimensions and mass.
29/11/2006
Page 15 of 85
Lens-corrected horn.
A group of such items are illustrated in Figure 2, which are manufactured by Flann in the UK.
Figure 2 Lens-corrected horn antennas

The 150 mm lens aperture yields two orthogonal half power beamwidths of approximately 5 and 6.
The gain is about 30 dBi. The aperture diameter is 160 mm and the horn length is about 250 mm. The
mass is 740 g. A reflector antenna would give comparable performance for similar mass and volume.
Plastic dual-reflector antenna.
We have (on loan) a compact and lightweight reflector antenna manufactured by ERA Ltd. to meet the
demands of a mass-producible and cost-effective antenna for broadband applications at 28 GHz.
Figure 3 Plastic dual-reflector antenna.

Measurements of the antenna shown indicate a gain of around 27 dBi. The aperture measures 140 mm x
100 mm, and its weight is approximately 80 g. It has been developed for a steerable base station
antenna for terrestrial applications. Due to its compactness and small weight, it would present a modest
load to mechanical actuators. This item may not be commercially available, or at least difficult to obtain.
3.2.2
Mechanically steered small high gain array
This solution consists of a relatively small high gain array and a high precision fully mechanical beam
steering system.
The advantages are the antennas low profile, low cost, small size (~2020cm), light weight (<1Kg) and
construction with fully passive components. According to the results reported by Japanese researchers,
a gain of more than 29 dBi has been achieved using a three-layer 1616-element array [25]. The
potential disadvantage of this approach is that there is no fast wide angle electrical beam steering. This
can be compensated by a rapid and high precision mechanical steering system.
29/11/2006
Page 16 of 85
One scenario of the mechanical system that can be imagined is a two-axis steering system driven by
the desired elevation and azimuth angles. The train speed, pitch/roll/vibration and banking angle in the
turns should be considered as important constraints in the mechanical design. The approximate
dimensions of the mechanical part should fill in a box of dimensions 252522 cm (Figure 4).
Antenna
Altitude rotation
22 cm
Azimuth rotation
Figure 4 Illustration of mechanical steering

Appending to Figure 5, the scan angle speed can be computed as
= tan 1 (
300 km
1
) 0.24 / s
3600 s 20 km
20 km
Baloon
300 km/h
Figure 5 Pointing angle speed

The required pointing angle precision can be estimated by taking an example of a square horn whose
2
sin( sin )
beam shape is approximated using function
, where d is side length and is a half of
d sin

beamwidth angle. Assuming a figure of 5 dB for antenna pointing losses relative to the bore sight based
on the link budget in section 2.1, one can easily evaluate the pointing angle precision for the two cases
listed in Table 2.
29/11/2006
Page 17 of 85
PBBW (-3 dB beam width)
-5 dB beam width
Pointing angle precision
6.6
0.8
2.64
0.32
Table 2 Pointing angle precision for 2 and 5 beamwidth
3.2.3
Phased array/ digital beam forming: circuit level constraints
The term "smart antenna" is often used in this context. In terms of complexity and ambition, this
solution is at the opposite extreme compared to the mechanically steered dish. Some recent
programmes in this field are reviewed.
The main advantages include: conformality of the array, electronic beam scanning, rapid scanning over
wide angles, multiple beam forming, free space combining of transmitted power. The main disadvantage
is probably the very large number of antenna and active elements which are required to achieve a high
gain aperture. Added to this, a conventional approach requires sub-half wavelength element spacing.
This leads to very complex RF circuitry which may be close to being physically non-implementable.
Less conventional approaches could include the use of widely spaced element, sub-arrays etc. In any
case, a very large number of elements are required, which will be extremely costly. A planar array also
exhibits scanning loss which inevitably reduces the aperture gain where it is most needed.
These effects, along with the constraints imposed by the physical circuit and the inter-element spacing,
are discussed below along with some examples and derivations from first principles.
3.2.3.1
Array factor
The array factor F is the interference pattern for the electric field. It is a function of element spacing a (in
the simplest case this is the same in both cartesian axes i.e. a = ax = ay). The maximum is given by
the number of elements.
The normalized two dimensional array factor is:
M .u x Sin N .u y
Sin
2
1
2 1
F ( , ) =
M
u N
uy
Sin x
Sin
2
2
u x = k .a x .Sin .Cos + x
where
and
u y = k .a y .Sin .Sin + y
k is the wave number
2 and a and a are the element spacings along the x and y axes respectively.
x
y
Similarly, x and y are the relative phase shifts between elements
Below is the array factor for 4 x 4 elements where a = ax = ay (in wavelengths).
29/11/2006
Page 18 of 85
F
16
14
element spacing
12
a=0.25
10
a=0.75
a=2
6
4
2
-75
-50
-25
25
50
75
degrees
Figure 6 Array factor for 4 x 4 rectangular array as a function of element spacing.

Clearly for increasing a (e.g. from a<0.1 to a=1) the main lobe becomes more narrow. For a>1 grating
lobes appear. We have not yet considered scaling the pattern for absolute directivity, nor the effect of
beam scanning.
3.2.3.2
Directivity
On scaling for absolute directivity D we use
D( , ) =
4 P( , )
Prad
where Prad is the total power radiated over all space:
Prad =
P( , ).Sin .d .d
=0 =0
Directivity is shown in Fig. 7. It is here assumed that the elements radiate into half space only. Evidently
the directivity is maximized for a spacing of about 0.75 . As the spacing increases D tends to oscillate
above and below the value associated with the number of elements (= M.N). Thus, for 16 elements, this
mean value for D is about 12 dB, plus 3 dB from the half-omni elements, leads to 15 dB.
dBi
26
directivity
dBi
directivity
19
18
17
16
15
a
0.5
1.0
1.5
2.5
3.5
(a) 4 x 4 elements
24
22
20
a
0.25
0.5
0.75
1.25
1.5
1.75
(b) 8 x 8 elements
Figure 7 Directivity on boresight versus element spacing

3.2.3.3
Implications of wide element spacing
29/11/2006
Page 19 of 85
Wide spacing (a > ) leads to grating lobes. It has been argued that grating lobes do not contribute to
interference because the area on ground and hence the number of users within the beam does not
change. This argument is only approximately true, because the beam directivity (a measure of total
radiated power - i.e. the spatial integral of the radiation pattern) is seen to vary quite a lot with element
spacing.
D (dB)
a=0.75
25
a=1.5
20
grating lobe
15
10
5
0
-5
20
40
60
80
()
Figure 8 Loss in Directivity associated with grating lobes, 8 x 8 element array
3.2.3.4
Effect of antenna element radiation pattern
Thus far the antenna elements have been assumed to be omni-directional (or half-omni for half space).
For practical, directive elements, the radiation pattern is the product of the array factor and the elemental
pattern. The directivity of the element can be exploited to suppress grating lobes. The maximum
possible element directivity is that associated with the maximum physical size. The radiation pattern for
the element may be approximated by the "sine" function for the element aperture:
Sin( a Sin )
P _ element =
a Sin
(The element radius could be smaller than dimension a, but not larger.)
The product of such an element pattern and the array factor is shown in Fig 9, where the suppression of
the grating lobes is evident. In practice, the element pattern might have a broader beam (it cannot have a
narrower beam) and so the sidelobe suppression may not be as great an effect as that illustrated.
29/11/2006
Page 20 of 85
F
16
array pattern
14
element pattern
12
product pattern
10
8
6
4
2
-75
-50
-25
25
50
75
angle (degrees)
Figure 9 Product of array and element patterns for 4 x 4 element array with a = 1.5 (=0 axis)
The patterns in Figure 9 are not scaled for Directivity.
3.2.3.5
Effect of beam scanning
For the 4 x 4 array with a = 0.5 wavelength spacing, with directive elements, the effect of scanning is
illustrated below. For small scan angles e.g. < 5 the beam shape is not much affected. For increased
scan angles the grating lobes of the array factor move into the element pattern and hence become
visible.
D (dB)
30
non-scanned beam
scanned beam
20
10
-10
-75
-50
-25
25
50
75
angle (degrees)
Figure 10 Effect of array scanning
29/11/2006
Page 21 of 85
D (dB)
non-scanned beam
30
scanned beam
20
grating lobe
10
-10
-75
-50
-25
25
angle (degrees)
50
75
Figure 11 Array grating lobes

The effect of grating lobes can be suppressed by using interleaved arrays because the effective subarray spacing is reduced.
3.2.3.6
Sub-array interleaving
The effect of grating lobes can be suppressed by using interleaved arrays [51].
4-element primary array coefficients: 1, 1.74, 1.74, 1
primary array spacing: 0.68
4-element secondary array coefficients: 1, 1.66, 1.66, 1
secondary array spacing: 1.71
The secondary array spacing is less than the primary array length of 2.73 which is only physically
possible if the primary arrays are interleaved.
29/11/2006
Page 22 of 85
The arrays reported are one-dimensional (linear) arrays. The following array patterns are not scaled for
directivity.
dB
dB
deg
deg
-75
-50
-25
25
50
-75
75
-50
-25
25
50
75
-10
-10
-20
-20
-30
-30
-40
-40
-50
-50
(a) Primary array pattern
(b) Secondary array pattern

dB
dB
deg
-75
deg
-75
-50
-25
25
50
-50
-25
25
50
75
75
-10
-10
-20
-20
-30
-30
-40
-40
-50
-50
-60
-60
(c) combined patterns (boresight conditions)
(d) combined array patterns with scanned beam
Figure 12 Array patterns
The following patterns use a secondary array spacing of 2.73 i.e. the primary array length. This shows
the case for a non-interleaved array, hence the grating lobes are increased compared to the above
(interleaved) case.
dB
deg
-75
-50
-25
25
50
75
-10
-20
-30
-40
-50
Figure 13 Secondary pattern
29/11/2006
Page 23 of 85
dB
deg
-75
-50
-25
25
50
75
-10
-20
-30
-40
-50
-60
Figure 14 Combined pattern, scanned beam
The literature reports on an experimental array prototype. The title's reference to "millimetre-wave"
appears somewhat spurious, as the prototype operates at X-band, claiming to be a scale model of a 60
GHz system. The prototype uses printed type antennas. A scanning beam was not implemented, but
two fixed-beam versions, for a non-scanned and scanned beam respectively. The scan angle of the latter
is about 17.
While the inter-leaving principle is interesting, this paper doesn't appear to shed much light on
implementation issues, particularly for two-axis scanning or operation at frequencies above X-band.
3.2.3.7
Discussion: phased array techniques
Having reviewed a number of techniques which attempt to reduce the complexities of array topologies for
smart antennas, it was felt that none offer the required functionality in terms of efficiency, scan
performance and cost. Often, when a technique is put forward to mitigate some disadvantage (e.g. use
of sub-arrays to reduce number of active components), some other disadvantage is made apparent (e.g.
reduced scan angle or appearance of grating lobes).
An array antenna variant was later developed, successfully, by CSEM, but as this was mechanically
steered we do not refer to it as a 'smart' (or beamforming) antenna.
3.2.4
Array variants and combined electronic and mechanical steering
The disadvantages of a conventional planar array can be mitigated in various ways if we allow one
antenna axis to be mechanically steered, as illustrated in Figure 15. Most likely this would be the
azimuth axis. If the array boresight is then tilted to say 37.5, it is require to scan in elevation over a 75
range to meet the requirement (15 elevation to vertical.) This approach mitigates against scanning loss
and, if the array does not scan in azimuth, far fewer active elements (phase shifters) are required. This
solution however does not permit electronic (rapid) scan or multiple-beam forming across azimuth
planes.
29/11/2006
Page 24 of 85
electronic elevation
scan
mechanical
azimuth steer
field of view ~90

antenna within
housing
Figure 15 Illustration of combined mechanical and electronic steering.
A type of array which could implement the single-axis electrically steered concept of Figure 15 is shown
in Figure 16.
antenna elements
(e.g. series fed linear array)
active circuits
(amplify, frequency convert)
N rows
IF outputs
xN
LO distribution
xN
LO
ADC
DSP
Figure 16 Possible RF array topology for single axis electronic steering.
3.2.5
Recent projects in phased arrays and Digital Beam Forming (DBF)
In this section we list some of the features of recent and ongoing European projects in phased array
antennas for communications. Their objectives are 2-way satellite communications for moving terminals,
using 20/30 GHz band. These projects have similar aims to the antennas research work packages of
CAPANINA and so it is particularly useful to investigate their approaches, outputs and costs.
3.2.5.1
SUITED project
This IST Framework 5 project ran from 2000 for 30 months. There were 11 partners. The value was 9 M
euro, the EU contribution was 4 M euro. TTI of Spain were the lead partners in the array design and
fabrication.
For more information see Microwave Journal, January 2004 Vol.47 No. 1.
29/11/2006
Page 25 of 85
Technical summary, SUITED antenna terminal:

frequencies
20 GHz receive
30 GHz transmit
array design
sub-arrays
(may be inclined)
8 element sub-sub
array.
microstrip
4 x 8 linear sub-array
12 rows
phase shifter loss
20 dB
scanning
mechanical azimuth
scan.
electronic elevation
scan
no. phase shifters = 12
elevation scan
specification 32 - 48
grating lobes
approx - 10 dB
receive gain, 20 GHz
40 dB claimed
G/T
for
aperture
estimated at 5 dB/K
dimensions
receive
transmit
(mm)
300 x 200 x 80
240 x 120 x 80
3.2.5.2
total no. radiating

elements = 384
SANTANA
This all-German consortium of 4 organisations including DLR and Astrium, funded by the German
government, have developed digital beam forming arrays for 20/30 GHz. An application they have
emphasized is satellite communications to aircraft during flight, and hence very rapid wide-angle
scanning is required.
The project ran for about 3 years, and finished at the end of 2003. The project delivered a 16 element
module which was intended as a demonstrator for a follow-on phase SANTANA II. The whole budget
allocated until the end of SANTANA II is approx. 4.5 MEuro and comprises around 350 person months
(this info. courtesy of DLR).
Technical summary, SANTANA antenna terminal:
frequencies
20 GHz receive
30 GHz transmit
array design
4 x 4 element module
scanning
DBF
intended as a plug-in
module for part of a
much bigger array
(no mechanical scan)

grating lobes
gain
suppressed by subwavelength spacing

receiver specification:
for 5 Mbps: 26 dBi
for 155 Mbps: 41 dBi
this clearly cannot be

met by the 4 x 4
prototype, which would
offer approx 16 dBi
Separate transmit and receive arrays were fabricated. The approach was to use microstrip radiating
elements fed by dielectric-loaded circular waveguide. MMIC amplification and frequency conversion
stages were employed at each element. The MMIC components were mounted on a water-cooled
substrate. For both arrays the physical size of the antenna was dominated by the IF circuits and
associated electronics.
29/11/2006
Page 26 of 85
3.2.6
The spherical lens antenna
Due to its symmetry, a spherical lens can employ multiple feeds to produce multiple beams without any
scanning loss. For beam steering, the feed position must be changed. This could be achieved either by
mechanically moving a single feed, or by switching between multiple feeds which are spaced around the
outside of the lens.
A classic type of spherical lens is the Luneburg lens, which has its focus exactly on the lens surface.
The dielectric constant varies radially between 1 at the surface and 2 at the centre. This type of lens is
rather difficult to fabricate, and is usually approximated by a series of concentric shells.
A constant-dielectric spherical lens is obviously much easier to fabricate, but suffers from non-exact
focussing, reflection loss, and therefore a somewhat limited aperture efficiency. A lens comprising two
dielectric shells can be a good compromise.
A very useful technique is to use a combination of a hemispherical lens with a reflective ground plane.
This has the same effective aperture as a full spherical lens, but occupies half the height (see Figure
17).
beam 1
beam 2
Luneburg lens
feed 2
feed 1
beam 1
feed 1
hemispherical lens
beam 2
feed 2
ground plane
plane wave
feed
reflective plane
virtual lens
effective
height
Figure 17 Luneburg lens and hemispherical lens.
29/11/2006
Page 27 of 85
The technique illustrated has many advantages for a scanning, multi-beam antenna placed on a train:
The lens does not have to be steered. The moving part (the feed) has much less mass than the
lens.
The physical height of the antenna is approximately half that of a circular dish of equivalent gain.
Multiple feeds can be employed for multiple beams
There is no scan loss.
It should be within our capabilities and budget to build a prototype.
These advantages have lead to the Luneburg lens approach being used for a satellite communications
antenna for use on aircraft. The latter point (no scanning loss) is not strictly true for the hemispherical
lens. For a zenith-directed beam, the feed blocks the aperture and results in reduced gain. This,
however, is less important for the zenith beam because here some loss can be tolerated (Figure 1).
There are of course some apparent disadvantages:
Mechanical scanning over a spherical surface is not trivial.
Each beam requires its own independent steering mechanism. This looks quite complicated and
difficult to achieve in practice. In some cases a feed could block the field of view of the other feed /
beam.
The routing of a flexible RF cable, waveguide, or use of a rotary RF joint is needed (for each feed).
Rapid scanning, to cope with pitch/roll/vibration, may be difficult.
Many of these disadvantages are also associated with any mechanically steered solution, but are here
ameliorated by the fact that it is only the feed - a relatively small and light weight item - which is steered,
as opposed to the antenna in its entirety. There is scope to consider some novel approaches which
could improve the functionality and performance of a hemispherical multi-beam lens antenna:
Use a multiple element feed array for electronic elevation scan. Then, only azimuth scan is
mechanical. A disadvantage is increased aperture blockage for high elevation angles
Use of a feed cluster for a multiple-beam electronic scan over a narrow angular range. This could
accommodate rapid beam tracking to cope with vehicle pitch/roll/vibration.
feed array
Figure 18 Hemispherical lens with elevation scanning feed cluster.
29/11/2006
Page 28 of 85
Hemispherical lens for customer premises equipment antenna.

It is worth digressing slightly to point out that variants of the hemispherical lens antenna could be an
economically viable solution for the customer antenna. In this case a relatively narrow range of steering
angles is needed (~ 20 ) and this could either be achieved with mechanical steering or with a multibeam feed.
ground plane
/reflector
multiple beams
scanned beam
lens /
multi-shell lens
feed
frequency convert,
amplify
feed cluster
mechanism
to customer terminal
cluster:
(a) mechanical scan
(b) multi-feed
Figure 19 Hemispherical antenna variants with limited scan.

A disadvantage of the geometries illustrated above could be the physical height required to house the
primary feed and steering jig for high elevation angle operation. A cover (radome) will be required. It
would be ideal if this cover could contribute constructively to the focussing effect of a multi-shell lens
system, where the focus lies inside the outer shell.
outer shell
feed cluster
steering mechanism
Figure 20 Sketch of a possible multi-shell lens configuration incorporating radome.
29/11/2006
Page 29 of 85
3.3
Recommendations following the survey of candidate technologies
A number of concepts for steerable antennas have been discussed. While emphasis has been placed
on an antenna for a ground vehicle the arguments are generally applicable also to an antenna for the
HAP.
The aim of Work Package 3.2 was to develop a prototype antenna which shows the key features
required for a steerable antenna terminal for CAPANINA.
While many techniques and options could be further investigated, two contrasting solution families
emerged following this study phase of the project, these being:
direct radiating phased array with digital beam forming.
mechanically steered antenna
3.3.1
Phased array and DBF
This offers:
advantages:
Ultimate beam agility and smart capability, in theory.

Free space power combining for transmit array.
disadvantages. Cost: Certainly beyond the financial scope of CAPANINA.

Complexity: very difficult to implement a full scale prototype.
Scan loss: increases at low elevation angles, where loss can least be tolerated.
Variant with mechanical steering largely negates the advantages of DBF.
Recommendation:
In terms of fabricating and performing measurements, the phased array solution was considered too
ambitious compared to the resource available to the project, and is similarly a very expensive solution
when we consider that CAPANINAs aims are cost effective broadband. The experiences of other
projects (e.g. SANTANA) corroborate this conclusion, where much more resource has been deployed to
develop a 16 element phased array prototype. The intention is not in any way to criticize these other EU
projects in smart antennas, rather, it is to question the value of attempting to replicate or advance on
their work with a small fraction of their resource.
Nevertheless, the DBF solution has remained an area of research interest, particularly under WP3.3
where beamforming methods have been studied in great detail. In terms of pursuing an experimental
programme of research under WP32 however, other more pragmatic solutions were found preferable.
3.3.2
Mechatronic Solution
This offers:
advantages:
Cost: full scale prototype within the financial scope of CAPANINA.

Complexity: moderate
Scan loss: negligible.
disadvantages. Complex mechanical drives and RF feed routing

Non-proven technology
29/11/2006
Page 30 of 85
3.3.3
Recommendations
Two approaches were identified as being particularly suitable.

For the train antenna, a mechanically steered small array antenna was identified as the most promising
for the following design and experimental phase. Its inherent light weight and relative ease of fabrication
makes it a very cost effective solution and one which is easily integrated with practical mechatronic and
control systems.
For the HAP antenna the spherical lens antenna was considered the most promising solution to take
forward into a phase of further research, including that of design, fabrication and measurement.
29/11/2006
Page 31 of 85
4 Experimental Investigations of Preferred Solutions

4.1
Broadband Printed Array Antenna
The main objective of the antenna design in this section is targetted at a ground mobile-user
requirement.
4.1.1
Introduction
Microstrip antennas are used in a wide range of applications thanks to their thin profile, light weight, low
cost and ease of integration with other RF devices. In array configuration, they can achieve high gain and
become an alternative to parabolic reflectors for millimetre wave applications, such as communications
via HAPs.
However, microstrip antennas operate only over a narrow bandwidth due to their resonant nature. Future
HAPs subscriber antennas require typically 30 dBi gain and a frequency bandwidth of 13% around
29 GHz. Two techniques commonly used to improve the bandwidth are to add parasitic patches which
are gap-coupled with the main resonator [30][31][32][33] or to use strip slot coupled multilayer structure
[34]. The former reported an achieved operating bandwidth (Voltage Standing Wave Ratio VSWR = 2) up
to 14%, while the latter increased the bandwidth to 19.5% with a stripline-fed multilayer antenna.
When using microstrip configuration at high frequency, some important issues need to be addressed
with care. The first issue is microstrip loss which becomes important at high frequency due to dielectric
materials. Special low-loss substrate should be considered. In addition, attention should also be paid
when designing a microstrip feeding network of arrays having a large number of elements. Strip lines
and/or multilayer structures can help to minimize the length of microstrip feeding lines thus to reduce
loss. The second issue is the required high gain with respect to the small size of the antenna.
Techniques already known for the gain enhancement are such as multiple superstrates [22], high
permittivity superstrate [23] [24] as well as coupling and shielding [25]. The gain G and directivity D of
an antenna is related by radiation efficiency as G = D , where includes the conduction and
dielectric efficiency which depend on the geometry of the antenna and on the substrate material. It is
observed that directivity increases with increase in substrate thickness while efficiency has an inverse
trend. So care needs to be taken to ensure the antenna efficiency is maintained when the gain is used
as a global parameter to be optimised on the antenna performance.
The third issue comes with the question how to provide large bandwidth with a printed antenna. To
address this problem, solutions found from literatures mentioned stacked patch [26] and slot excitation.
The purity of the circular polarisation (CP) within very large bandwidth is another issue. Till now,
sequential rotation feeding network [27] or polarisation transformer [28] has been employed to get good
AR of CP within a relatively large bandwidth. Last but not least, from the point-of-view of fabrication,
substrate planarity should be guaranteed because phase error very sensitive at high frequency
applications.
In this work, a strip slot coupled solution has been proposed because the feeding network is located on
a different layer than the patches enabling the routing to be optimized separately. In addition, we
considered an array composed of a small number of elements to further reduce feeding network
complexity while maintaining the performance required by HAPs application.
4.1.2
Techniques for second lobe reduction
In the design procedure of any array, array factor (AF) is an important element which is, in general,
related to the number of elements, the array geometry, the excitation phase and magnitude, as well as
the element spacing. For a N-element linear array of uniform spacing and uniform power distribution, the
AF can be written as
29/11/2006
Page 32 of 85
sin (k d cos + )
2
AF
k
d
cos
Where
N = number of elements in the array
d = element spacing in the array
= angle between observing direction and array direction
= phase excitation of the elements in the array
k = wave number defined as equal to
In a uniform amplitude and phase array, the element spacing should be less than to prevent unwanted
grating lobes in broadside radiation region [39]. With the high-order mode SSFIP of size , it is
impossible to satisfy this spacing condition. Here we will discuss three possibilities of reducing the
sidelobes level.
4.1.3
Non-uniformly spaced array
In general, the sidelobes of array antennas with uniform amplitude excitation can be reduced by using
non-uniform spacing between elements. In reference [42], an iteration method based on the solution of a
set of linear simultaneous equations at each iteration is proposed to obtain the non-uniform array
element positions. The position of the nth element form the origin is written as:
n
dn = ( + n ) d
2
where d is the spacing between elements of the referential uniform array, and
with respect to its uniform array location.
n the shifted distance
In Figure 21, the red curve shows resulted side lobe level (SLL) of a 24-element linear array by using one
set of element positions proposed in [42]. The SLL of the same array with uniform spacing is also
plotted (in blue) for comparison. This technique results in 11 dB SLL reduction when assuming the
element spacing of the referential uniform array d = 0.5.
N= 24 elements and d = 0.5*lambda
0
Uniform spacing
case B
-10
Normalized AF (dB)
-20
-30
-40
-50
-60
-70
-80
0
20
40
60
80
100
theta ()
120
140
160
180
Figure 21 Non-uniform spacing applied to a regular array in which d = 0.5
29/11/2006
Page 33 of 85
In our case, the spacing of array elements is much bigger due to the large size of the highly directive
array element. Consequently, grating lobes related to the array geometry are unavoidable, as can be
observed on the blue curve in Figure 22. The application of the same set of nonuniform positions reduces
the grating lobe level (GLL) but arises at the same time the SLL associated to the grating lobes. This
result is shown by the red curve in Figure 22. For more satisfactory SLL/GLL reduction, new codes need
to be implemented by taking into account the large element spacing.
N= 10 elements and d = 2*lambda
0
-10
Normalized AF (dB)
-20
-30
-40
-50
-60
Uniform spacing
case B
-70
-80
0
20
40
60
80
100
theta ()
120
140
160
180
Figure 22 Non-uniform spacing applied to a special array in which d = 2
4.1.4
Non-uniform power distribution
Non uniform power distribution can lower SLL (Dolph-Tschebyscheff power distribution) or even totally kill
the second lobes (Binominal power distribution). Under the condition that element spacing is equal or
less than 0.5., Dolph-Tschebyscheff power distribution can be designed using Tschebyscheff
polynomials to provide any desired SLL. When required SLL = - dB, the Tschebyscheff array becomes
a Binominal array which has no side lobes.
Applying the condition of element spacing d = 2, Figure 23 shows comparison of uniform, DolphTschebyscheff and binomial power distribution: with SLL = -25 dB for Dolph-Tschebyscheff. It is apparent
that the grating lobes are not modified by power distribution. In addition, the main beam width and the
side lobe level are compromised.
29/11/2006
Page 34 of 85

N= 8elements and d = 2*lambda
0
-5
Normalized Array Factor (dB)
-10
-15
-20
-25
-30
-35
Uniform
Dolph-Tschebyscheff
Binomial
-40
-45
-50
20
40
60
80
100
theta ()
120
140
160
180
Figure 23 Linear array with uniform, Tschebyscheff and Binominal power distribution (d = 2)
4.1.5
Specific single element radiation
For any arrays of identical elements, its far field is expressed as the product of the field of a single
element and the array factor (AF). The directional characteristics of the radiating elements themselves
give an additional degree of freedom for working towards the desired radiation pattern. In Figure 24 is an
example showing how a special radiation pattern of the array element, which has nulls at the place
where the grating lobes occur in the AF function, can contribute to lower the side lobe level. The special
radiation pattern can be worked out mainly through changing patch geometry.
Non-uniform spacing, N= 8 elements and d = 2*lambda
Simulated single element radiation pattern
10
0
-10
-10
-20
-30
Gain [dB]
Field intensity [dB]
-20
-40
-40
-50
-50
-60
-70
-200
-30
phi=90
phi=0
-60
-150
-100
-50
0
theta []
50
100
150
-70
-200
200
AF: theoretical computation
-150
-100
-50
0
theta []
50
100
150
200
Element radiation: EM simulation

Array radiation pattern, N= 8 elements and d = 2*lambda
10
phi=90
0
-10
Gain [dB]
-20
-30
-40
-50
-60
-70
-200
-150
-100
-50
0
theta []
50
100
150
200
Figure 24 Combination of AF and special element pattern

29/11/2006
Page 35 of 85
4.1.6
Single high gain broadband antenna design
Features of SSFIP antennas operating at fundamental resonance have been well documented in [34],
[35] and [36]. A full-wavelength SSFIP providing two off-axis main lobes at 30 angles with respect to
broadside has also been reported [37]. Here we present a novel design of a higher order mode SSFIP
with etched patch which exhibits high directivity in broadside direction.
As presented in Figure 25, the patch (top-layer) of the high-order mode SSFIP is coupled with the
microstrip feed line (bottom-layer) through the slot (mid-layer). The patch is of size at operating
frequency. The patch has been etched in a fractal-like shape to provide a specific radiation pattern as
well as broadband impedance matching. The antenna in Figure 25 has been simulated using commercial
electromagnetic Method of Moment (MoM) software [38]. The substrate is Rogers RT/Duroid 5880 with
thickness h = 0.254 mm and dielectric constant r = 2.2. The overall size of the antenna is
13.8 mm 16.5 mm 1.4 mm. The radiation pattern was studied to enhance grating lobe suppression
when integrated as an element into an array.
Top view of
the patch
Substrate
Inverted patch
Spacer
Slot
Substrate
Strip
Figure 25 Different layers of a SSFIP antenna
Figure 26 depicts impedance matching bandwidth obtained by simulation. As it can be seen in Figure
26, the operating frequency for a VSWR = 2 ranges from 27.0 to 33.1 GHz, meaning 20.3% of
impedance bandwidth. An example of radiation pattern in the E-plane ( = 90) and H-plane ( = 0) is
presented in Figure 27.
Table 3 summarizes the gain and half-power beamwidth at four simulated frequencies. The gain flatness
remains less than 1 dB in the whole frequency range. The simulated cross-polarization is about 75 dB
lower than the co-polarization meaning that the high order mode is properly excited.
3.5
VSWR [-]
2.5
1.5
1
26
27
28
29
30
31
Frequency [GHz]
32
33
34
Figure 26 Single element antenna: impedance bandwidth (VSWR) as a function of frequency.
29/11/2006
Page 36 of 85
Operation frequency
[GHz]
Gain
[dBi]
E-plane HPBW
[]
H-plane HPBW
[]
27.5
10.1
32.8
30.4
28.5
10.8
30.8
29.6
29.5
10.4
29.2
30.0
30.5
9.9
27.8
30.8
Table 3 Main simulated characteristics of the single antenna
90
120
[dB]
10
60
0
-10
150
30
-20
-30
-40
180
210
330
phi = 0
phi = 90
240
300
270
Figure 27 Co-polarization radiation pattern in E-plane ( = 90) and H-plane ( = 0) at 28.5 GHz
The prototype has been fabricated by two PCB manufacturers. It is important to mention that the
performance of the antenna is sensitive to the thickness of the spacer placed in between the patch and
the ground plane containing the slot (see Figure 25). To be able to compare the simulated and the
measured results, the spacer thickness in the simulation was adjusted to be the same as that of the
prototype. The result of the comparison is shown in Figure 28. The difference between the measured
and simulated results is due to the uncertainty of dielectric loss at high frequency.
Comparaison of simulation and measurement
0
S11 magnitude [dB]
-5
-10
-15
-20
simulated s11
measured s11
-25
-30
26
27
28
29
Frequency [GHz]
30
31
32
Figure 28 Measured and simulated impedance matching for adjusted model
29/11/2006
Page 37 of 85
4.1.7
Array performance
In a uniform amplitude and phase array, the element spacing should be less than to prevent unwanted
grating lobes in broadside radiation region [39]. With the high-order mode SSFIP of size , it is
impossible to satisfy this spacing condition. To minimize the grating lobes, the inter-element spacing of
1.8 is adjusted to place the H- and E-plane grating lobes at the radiation direction where the single
element antenna has a null in the radiation pattern.
Due to hardware limitations, only 4- and 16-element arrays were simulated using the commercial EM
software[38]. In Figure 29, impedance matching bandwidth versus frequency range of the 16-element
array is presented. The broadband behaviour of single element remains stable in the array configuration.
Simulated cross-polarization is about 23 dB lower than the co-polarization. The improved radiation gain
of the single element and the 22 subarray compared with a normal square patch are shown in Figure
30 and Figure 31. It can be observed that with our special shaped single element radiation pattern, 6 dB
of energy in the grating lobes are moved into the main lobe. Other simulated performance of the single
element and the sub-arrays are summarized as well in Table 4.
5
4.5
4
VSWR [-]
3.5
3
2.5
2
1.5
1
0.5
0
26
27
28
29
30
31
Frequency [GHz]
32
33
34
Figure 29 16-Element array: impedance bandwidth (S11 magnitude) as a function of frequency

Simulated gain of single element
10
Gain [dBi]
-10
-20
-30
-40
Square patch, phi=90
-50
Square patch, phi=0

FSP, phi=90
FSP, phi=0
-80
-60
-40
-20
0
theta []
20
40
60
80
Figure 30 Simulated gain of the single element compared to a normal patch
29/11/2006
Page 38 of 85
Gain of 8x8array, d = 1.8lambda, phi = 90

30
20
20
10
10
0
Gain [dBi]
Gain [dBi]
Gain of 8x8array, d = 1.8lambda, phi = 0

30
-10
-10
-20
-20
-30
-30
-40
-40
Square patch, phi=0

FSP, phi=0
-50
-100
-80
-60
-40
-20
0
theta []
20
40
60
80
100
-50
-100
Square patch, phi=90

FSP, phi=90
-80
-60
-40
H-plane
-20
0
theta []
20
40
60
80
100
E-plane
Figure 31 Simulated gain of the 22 subarray compared to a normal square patch
Simulation results
Single element
4-element array
16-element array
Frequency range
27.1 33 GHz
26.3 32.9 GHz
26.7 32.4 GHz
19.6%
21.8%
19.3%
Bandwidth
More than 10x of a normal patch array
HPBW
gain
Gain flatness
MLL/SLL
31
15.4
7.6
10.7 dBi
(5 dB better than a
rectangular square patch)
17.5 dBi
(>5.5 dB better than
omnidir.-element array
of same apertur)
23.2 dBi
(>5.1 dB better than
omnidir.-element array
of same aperture)
~ 2 dB
< 1 dB
< 1 dB
~ 11 dB
Table 4 Simulation results of single element and sub-arrays
29/11/2006
Page 39 of 85

Gain of 16x16array, d = 1.8lambda
Gain of 8x8array, d = 1.8lambda
40
30
30
20
20
10
10
0
Gain [dBi]
Gain [dBi]
0
-10
-20
-10
-20
-30
-30
-40
phi=0
phi=90
-40
-50
-100
-80
-60
-40
-20
0
theta []
20
phi=0
phi=90
-50
40
60
80
100
-60
-100
-80
-60
88 element
-40
-20
0
theta []
20
40
60
80
100
1616
Figure 32 Predicted radiation of 64- and 256-element array based on single element pattern
4.1.8
Measurement results
A 64-element prototype (Figure 33) has been fabricated for verification. The measured impedance
matching presented in Figure 34 exhibits relatively good broadband performance. The theoretical
radiation pattern of a general case, e.g. array of x-element spaced by distance d, has been computed by
applying the planar array theory on a simulated single element. The theoretical results of a 64-element
array are compared with the measured curves in Figure 35, Figure 36, Figure 37 and Figure 38.
Measured radiation patterns show excellent agreement with theoretical prediction in both E- and Hplane. Electrical features of the array remain stable through the whole frequency range. In Figure 35 it is
shown that the E-plane grating lobes are minimized to 15.7 dB lower compared to the main lobe at
27.5 GHz. The grating lobes level for other frequencies can be found both in Figure 37 and in Table 5.
The E- and H-plane cross-polarizations are at about -23 dB compared to the co-polarization within the
whole operation frequency range. Measured cross-polarizations are presented in Figure 36 and Figure
38. The second lobes of value -11.7 dB at 30.5 GHz are due to the second lobes of the single element at
that frequency. The front-to-back ratio is 13 dB for the worst case. This ratio can be further enhanced to
30 dB by absorbing material placed on the backside of the array (Figure 39).
The gain estimation of the 64-element array has been made by comparing with a commercial broadband
Horn antenna of +/-1 dB uncertainty [40]. The values and comparison with the simulations are
summarized in Table 5. The measured gains are 3.5 to 5.8 dB lower than computed values depending on
operating frequencies. A way that could improve the theoretical prediction is to include the loss of the
feeding network into MATLAB simulation. At high frequency, any uncertainties concerning the dielectric
material characteristics or metallic surface conductivity impact the gain predicted by the simulations.
29/11/2006
Page 40 of 85
Top layer
Bottom layer
Figure 33 64-Element prototype
10
9
8
VSWR [-]
7
6
5
4
3
2
1
0
26
27
28
29
30
31
Frequency [GHz]
32
33
34
Figure 34 Measured VSWR of 64-element with respect to frequency
29/11/2006
Page 41 of 85
120
90
90
[dB] 0
[dB] 0
120
60
-20
-20
150
150
30
-30
-40
-50
-50
180
210
30
-30
-40
180
210
330
330
Numeric
Measurement
Numeric
Measurement
240
60
-10
-10
240
300
300
270
270
(a) E-plane
(b) H-plane
Figure 35 Comparison of normalized co-polarization radiation patterns of the 64-element array

at 27.5 GHz
90
120
[dB] 0
90
60
120
-10
-20
30
-30
150
-40
-50
-50
180
210
330
180
210
330
E co-polarization
H co-polarization
E cross-polarization
Hcross-polarization
300
30
-30
-40
240
60
-10
-20
150
[dB] 0
240
300
270
270
(a) E-plane
(b) H-plane
Figure 36 Measured normalized co- and cross-polarization at 27.5 GHz
29/11/2006
Page 42 of 85
90
120
[dB] 0
90
60
120
-10
-20
30
-30
150
-40
-50
-50
180
210
30
-30
-40
180
330
210
330
Numeric
Numeric
Measurement
Measurement
240
60
-10
-20
150
[dB] 0
300
240
300
270
270
(a) 28.5 GHz
(b) 29.5 GHz
Figure 37 Comparison of normalized E-plane co-polarization radiations of the 64-element

array
120
90
90
[dB] 0
[dB] 0
60
120
-10
-20
150
-20
30
-30
150
-40
-50
-50
180
210
330
180
210
330
E co-polarization
E co-polarization
300
30
-30
-40
240
60
-10
240
300
270
270
(a) 28.5 GHz
(b) 29.5 GHz
Figure 38 Measured normalized E-plane co- and cross-polarization at 28.5 and 29.5 GHz
29/11/2006
Page 43 of 85
Frequency
[GHz]
Theor. gain
[dBi]
Meas. gain
[dBi]
Theor. HPBW
[]
Meas. HPBW
[]
Meas. E-plane
grating lobes [dB]
27.5
28.2
24.7
3.4
3.6
-15.7
28.5
28.8
24.9
3.6
3.7
-14.2
29.5
28.4
22.6
3.6
3.6
-12.3
30.5
28.1
22.4
3.0
3.4
-11.7
Table 5 Performance of 64-element array
90
120
[dB] 0
60
-10
-20
150
30
-30
-40
-50
180
210
330
Without Absorbant
With absorbant
240
300
270
Figure 39 Enhanced E-plane backward radiation of the 64-element array at 27.5 GHz
29/11/2006
Page 44 of 85
4.1.9
Polarizer
The present polarizer has the function of converting a wave from linear polarization to circular
polarization. It is also called polarization transformer. The basic approach of a polarization transformer
is to make an array of structures which appear to be predominantly inductive to one polarization and
predominantly capacitive to the orthogonal polarization [41]. The polarizer is built with four-layer meander
line arrays which produce two field components with differential phase shift of 90 form the incident wave
and generate circular polarization. Using a polarizer to get the circular polarization from the linear
polarized array has several advantages:
High polarization purity (low axial ratio) over a large frequency bandwidth
Low dissipation loss
Good VSWR for impedance matching
Compact size at Ka band
Capability of handling high power (15.5 W/cm2 at 9 GHz reported in [41])
Low complexity
low cost
A polarizer, to be placed in front of the array antenna, has been formed from multiplayer foam with a thin
flexible dielectric substrate for the etched conductors and has been rescaled for 30 GHz application. A
side view of the polarizer section and the layouts are shown in Figure 40 and Figure 41.
Layer 1
Layer 2
Spacer
Layer 3
Flexible
substrate
Layer 4
Figure 40 Four-layer meander line polarizer
29/11/2006
Page 45 of 85
Zoom
Layer 1 & 4
Layer 2 & 3
Figure 41 Four-layer polarizer layouts
4.1.10 Loss consideration

4.1.10.1 Loss in the antenna
Loss in a printed antenna is distributed in the form of dissipation in the conductors and dielectric, and
surface wave. Antenna radiation efficiency r is primarily related to the substrate thickness and
permittivity. Though, cautions need to be made through the design of array at high frequency. For
example, low-permittivity low-loss substrate was employed to widen the impedance bandwidth and
reduce the surface wave excitation. Metal thickness of 25 m has been chosen for compromising
between loss and spurious radiation. For the design considered, loss in the microstrip feed line was
estimated to be from 10 to 10.5 dB/m at 28.5 GHz.
The radiation efficiency r can be estimated using the maximum directivity D0 and the gain G0 of the
antenna as: G0 = rD0. Theoretically, the maximum directivity of any antenna can be computed from its
maximum effective aperture [39]:
Aeff =
2
D0
4
Based upon the aperture size of the antennas and the simulated gain, the radiation efficiency at
28.5 GHz of the single and array antennas was estimated and summarized in Table 6.
Efficiency
Single element
22 array
44 array
88 array
1616 array
88.6%
59.2%
36.6%
23.2%
10.2%
Table 6 Estimated efficiency at 28.5 GHz of the single element and array antenna
The single element efficiency is lower than a microstrip line fed patch. When using a strip-slot coupled
feeding, as the substrate thickness increases, the surface wave power increases, and consequently the
efficiency decreases. It can be observed that the radiation efficiency decreases rapidly as array element
number increases due to increasing loss in the microstrip feed lines.
29/11/2006
Page 46 of 85
4.1.10.2 Loss in the polarizer

The loss in the polarizer comes from three sources: loss in the meander lines, loss in the spacers and
loss in the substrate. For the particular polarizer design here, the loss in the passband is approximately
[43]:
12 .5
1
1
1
+
+
Qmeander
Qspacer Qsubstrate
dB
where Q is the unloaded quality factor related to different sources and is the bandwidth.
The loss in copper meander lines is a function of the skin depth, the free-space wavelength, the metal
resistivity and the meander lines dimensions. The loss at 10 GHz has been calculated in [41] as to be
approximately 0.02 dB. It can be easily scaled to the frequency of 30 GHz by considering Q 1/(f)1/2
and we get the loss due to the meander lines at 30 GHz is approximately 0.035 dB.
The loss in the spacer and the substrate related to the dielectric properties is estimated as [43]
total loss in dB
27 .3 r tan( )
hi
0
where 0 i s the wavelength in air, r and tan( ) are respectively the relative permittivity and the
dissipation factor of the material and h i is the thickness of the i th substrate. The loss tan( ) of the
spacer becomes more important at 30 GHz, in the order of 0.01, while the r remains stable at 1.04.
Thus the loss dissipated in the three quarter wavelength spacers is about 0.2 dB.
The relative permittivity of flexible substrate is about 3.1 and its dissipation factor ( tan( ) ) is as low as
0.0009 (tested at 10 GHz). With the thickness of 50m, the loss in the substrate is negligible.
(The test result of loss tangent is available at up to 10 GHz from the substrate manufacturer. However
in the cost estimation, a margin has been taken in considering higher frequency application)
In summary, the total energy dissipated in the polarizer at 30 GHz is estimated to be 0.24 dB.
4.1.11 Conclusion for printed array antenna

A broadband high-order mode SSFIP antenna with high directivity in the broadside direction has been
developed. The highly directive SSFIP antenna enables implementation of an array with a reduced
number of elements; thus decreasing feeding network complexity. Subarrays of 4, 16 and 64 elements
were studied by simulation. A 64-element prototype has been tested and the measured results have
been compared with the theoretical computations and shown to be in good agreement.
29/11/2006
Page 47 of 85
4.2
4.2.1
Hemispherical lens prototype.

Constant index lens
A basic hemispherical lens prototype was been constructed and investigated experimentally at the
University of York.
This first prototype consisted of a single layer of polyethylene, of dielectric constant 2.3
The antenna is illustrated in Figure 42. The feedhorn illustrated is not optimized for this application. The
lens diameter is only 160 mm which is somewhat less than anticipated for the final requirement.
The purpose of this first prototype was to investigate scan performance.
Figure 42 Measurement jig for proof-of-concept (this one is set up for about 65 elevation
angle).
The measurement facility was an anechoic chamber where the antenna under test was placed on a
motorized turntable and acted as the receiver. The carrier frequency used was 28 GHz. A harmonic
mixer was used to downconvert the carrier to 1.7 GHz so that a coaxial cable could be used to connect
to a spectrum analyzer, external to the chamber, without incurring excessive loss. WR75 waveguide
was routed into the chamber to provide the local oscillator signal at around 13 GHz. Data capture was
automated under control of a personal computer running LabView which interfaced with the turntable
controller via a serial port, and with the spectrum analyzer via a GPIB interface.
It should be noted that, due to the limited length of the chamber (2.7 m between transmit and receive
antennas) the measurement was not that of the far field. Nevertheless the modal expansion analysis can
readily accommodate near-field effects and thus the presented results compare measurement with
theory at the 2.7 m distance (not far field). This has the effect of sidelobe growth compared to the far
field patterns.
29/11/2006
Page 48 of 85
reference antenna
90 elevation
-100
45 elevation
-50
50
100
150
degrees from zenith
65 elevation
-10
15 elevation
-20
-30
-40
-50
dB
Figure 43 Radiation patterns for 4 elevation angles.
Figure 43 shows measured radiation patterns for the reference antenna (a lens-corrected horn type) and
4 cases of elevation angle for the single-shell hemispherical lens. The main features of this
hemispherical lens prototype are:
inferior gain compared to the reference antenna (between 1.1 and 2.9 dB)
good wide angle scan properties
a significant pattern degradation for 90 elevation (when the aperture is blocked)
The feature of aperture blockage is due to the hemisphere lenss use of a ground plane reflector. Thus,
for a 90 (or similar) elevation angle, the primary feed blocks the aperture. This leads to gain reduction
and increased sidelobe levels.
As has been discussed, the zenith gain reduction is not particularly problematic for the HAP geometry,
since here the link length is at a minimum. Of greater importance is that the scan loss at low elevation
angles should be minimized. Here, the advantages of the hemispherical lens are apparent, as
highlighted in Figure 44. The normalized measured scan loss of the hemispherical lens is shown (4
points) compared with the free space loss for a HAP height of 20 km. Also shown is an approximation of
scan loss for a planar antenna (e.g. a planar array) based on the effective aperture area.
29/11/2006
Page 49 of 85
measured hemispherical lens scan loss

12
10
8
approx. scan loss of planar aperture
6
attenuation (dB)
free space loss
4
2
20
30
40
50
60
70
80
elevation angle ()
Figure 44 Trends in loss as a function of elevation angle.
Figure 44 shows how the maximum scan loss of the hemispherical lens coincides with the minimum
free space loss and improves at lower elevation angles. In contrast, the trend for a planar array is that
the scan loss increases at lower elevation angles.
(For a spherical lens there would be no reflector plane and no aperture blockage and a plot of the type
in Figure 44 would, theoretically, show all points at 0 dB scan loss.)
4.2.2
Primary feeds
A better primary feed was then developed, being a scalar circular waveguide feed of the design reported
by [15] whereby three concentric grooves of quarter-wave depth were machined into the waveguide
flange. This yields a broader primary feed beamwidth than that of the pyramid horn, thus better
illuminating the lens, and also offers better cross-polar performance. Using this feed, the measured gain
was approximately 30 dBi, representing an aperture efficiency of approximately 40 %. The measured
radiation patterns are compared in
Figure 45. The scalar feed can also be used in combination with a dielectric waveguide polarizer so as to
yield circular polarisation.
29/11/2006
Page 50 of 85
angle ()
-40
-20
20
-10
40
feed type:
scalar
pyramid
-20
-40
-50
relative gain (dB)
-30
Figure 45. Measured lens antenna radiation patterns comparing feed types. (Inset: image of
scalar feed.)
Having established that the desired wide scan properties are achievable, and investigating
experimentally the effect of the primary feed on inherent efficiency of single-layer hemisphere lenses, the
advantages of multi-shell lenses are next discussed
4.2.3
Development of a multi-layer lens
The rationale of this section of the research programme was to investigate the performance gains
associated with a multi-layer lens compared to a single layer lens, so as to quantify trade-offs between
lens antenna cost and performance. A key objective was to find a solution to a major problem of the
Luneburg lens antenna i.e. the complexity of the fabrication problem and the associated performance
degradation which can easily be introduced.
4.2.3.1
Analysis of single-shell and two-shell lens designs
A mathematical analysis of the radiation properties of concentric dielectric spherical shells is given by
[16] where a modal expansion method is used to derive the field scattered when the source is a short
dipole. This technique has been used to investigate the radiation patterns of various combinations of
shell radius and dielectric constant. Taking these as input parameters, the scattering matrices and
resulting far field patterns, which are the sum of the source and scattered fields, are computed using
approximately 250 lines of Mathematica code. The maximum directivity is then computed from the far
field pattern. The number of terms needed in the modal expansions is proportional to the maximum
dimension of the antenna system. For an 8 wavelength outer radius 60 terms are adequate. While [18]
reported on an optimisation of dielectric constant, of particular interest to the present study is to choose
from a small list of possible materials which have low loss, known r, are machinable and obtainable.
The focal distance is important as this should be minimized (within limits dictated by the phase centres
of practical primary feeds) so as to minimize the physical height of the antenna. The geometry is shown
in Figure 46, where f is the distance of the primary feed from the lens centre. A second feed in phase
quadrature with the first can be arranged to synthesize a directive primary feed pattern. The materials
considered are listed in [20].
29/11/2006
Page 51 of 85

0.25 (two-element
end-fire array)
f
r2
r 2
r 1
r1
ground plane
Figure 46. Two-shell hemisphere lens geometry
material
loss factor
(tan )
polyurethane
1.2
< 10-4
polyethylene
2.25
8 x 10-4
Rexolite
2.53
7 x 10-4
fused silica
3.8
7.5 x 10-4
foam
Table 7. Dielectric materials considered for two-shell hemisphere lens.
As a reference point against which to compare a two-shell design, the theoretical properties of singleshell spherical lenses are firstly analysed.
Figure 47 shows the sensitivity of an 8 wavelength radius spherical lens to the location of the feed (the
theoretical directivity of a uniform 8 aperture is 34.7 dBi). The trend that lower r materials exhibit a
greater focal distance is expected, as reported by [19], although, to digress, the "paraxial focus" of [19]
which is given by F =
leads to a somewhat larger value of the focal length than that which
2 r 1
maximizes directivity as derived by the modal expansion method.

The need to move the feed away from the lens edge leads to an increase in the required height for the
antenna installation. A minimum feed-lens separation of 0.2 has been used so as to accommodate a
practical primary feed, where the phase centre lies, typically, up to this distance inside the feed aperture
[15]. The maximum directivity D for the three materials is shown in
Table 8 along with the feed position at which this was achieved. The last column shows the case for a
fixed 8.2 feed location and where the lens diameter is reduced accordingly so that the relative focal
position is maintained. This allows a comparison of performance where the total height is fixed.
29/11/2006
Page 52 of 85
Rexolite
computed directivity (dBi)
32
31
polyethylene
30
29
28
fused silica
27
26
8.2
8.4
8.6
8.8
9.2
9.4
feed location (wavelengths)
Figure 47. Spherical lens computed directivity D versus feed location for three materials.
D (dBi) for 8
lens
radius
feed
location
( )
D for fixed
8.2 feed
location
polyethylene
31.9
9.2
31.3
Rexolite
32.7
9.0
32.0
fused silica
29.1
8.2
29.1
Table 8. Directivity and focus points of spherical lenses.
A two-shell lens with feed fixed at 8.2 was next investigated, where the outer radius r2 was fixed at 8
and r1 allowed to vary. Materials were chosen from Table 8 and always with r1 > r2. (Dielectric loss is
readily andled in the modal analysis but was found to be small for the materials chosen.) To summarize
the effects observed:
there is no advantage adding an outer layer to a fused silica core.
there is minimal advantage in using the low-constant foam as an outer layer. The layer would need
to be thin (~ 1 ) and hence difficult to fabricate.
The best results were obtained using a Rexolite inner core of radius 4.2 and a polyethylene outer
core. This yielded 33.46 dBi which represents 76 % aperture efficiency (albeit with an idealized feed
model).
To sum up, given a fixed antenna height, a 2-shell lens of 8 outer diameter offers a 1.46 dB gain
improvement over a single shell lens. Whether this gain is justified would depend on manufacturing costs
and the constraints on terminal size.
Lenses of greater diameter were next investigated: for a feed at 11.5 the directivity of a single Rexolite
sphere is maximized at 34.4 dBi for a radius of 10.35 while a 2-shell Rexolite/polyethylene lens of
outer radius 11.1 offers 36.1 dBi, being a 1.7 dB improvement.
These dimensions would represent an outer lens diameter of 236 mm at 28 GHz - about the largest size
which was considered suitable for machining at the University of York Physics and Electronics
Mechanical Workshops. This antenna prototype was then fabricated.
29/11/2006
Page 53 of 85
4.2.3.2
Two-layer lens performance
The above discussed two-layer hemisphere lens was fabricated. The core radius was 56.7 mm while the
outer diameter was 235.5 mm. The inner, concave surface of the outer polyethylene layer was machined
first, but this later experienced some mechanical stress relief on machining the outer profile and which
initially resulted in the inner Rexolite core protruding very slightly. When placed on the 350 mm square
ground plane, a small air gap (about 0.3 mm) was initially present. Measured gain was 34.2 dBi. (This
includes, being mindful of the near field distance, a correction factor of +0.57 dBi which was computed
by evaluating (1) as a function of r.) The components are shown in Figure 48
Figure 48. Components of two-layer lens (left: polyethylene shell, right: Rexolite core.)
To eliminate the air gap at the ground plane, the flat (lower) surfaces of both layers were lapped using a
fine abrasive. A small residual air gap remained at the lower edge of the interface between the two
dielectric layers, due to the stress distortion of the polyethylene. Nevertheless, following this
modification the measured gain was 35.1 dBi 0.4 dB which represents an aperture efficiency in the
region 68 %. This efficiency is comparable with a conventional reflector antenna and suggests that
dielectric loss is not too significant a factor. The measured radiation pattern is compared to the theory,
taking into account the 2.7 m measurement distance, in Figure 49 (a), while the theoretical far field
patterns are shown in Figure 49 (b).
29/11/2006
Page 54 of 85
angle ()
-40
-20
20
40
measurement
theory
-20
-30
-40
relative gain (dB)
-10
(a) H-plane patterns at 2.7 m measurement distance

angle ()
10
20
-10
30
40
50
H-plane
relative gain (dB)
E-plane
-20
-30
-40
-50
(b) Theoretical E- and H-plane far field patterns
Figure 49 Two-layer lens radiation patterns at 28 GHz.
4.2.4
Mechanical feed steering concepts
A number of layouts were considered. One tentative concept is illustrated in Figure 50 which shows two
independently steerable feeds. This arrangement would allow the antenna to communicate
simultaneously with two different vehicles or user groups, which would facilitate handover between cells
and also allow for diversity techniques. The sizes of the moving components should be minimized so as
to minimize both the aperture blockage and the height of the installation.
29/11/2006
Page 55 of 85
Bearings
RF unit /
feeder
Motor and
encoder for
feeder
elevation
rolling ribbon
Motor and
encoder for
azimuth rotation
illustration by Lorenzo Zago, CSEM, Switzerland
Figure 50. Feed steering concepts (inset: rolling ribbon actuator mechanisms).
A design of a steering mechanism of the type illustrated in Figure 50 was carried out by CSEM. However
construction of a prototype at UOY could not be carried out within the projects resource constraints.
The seeking of external funds to develop a prototype has been tackled as a project exploitation task.
4.2.5
Multi-beam Lens Antenna Conclusions
Hemisphere lens antennas have been identified as a promising solution for a multi-beam, wide-scanning
antenna for communications to moving vehicles via satellite or High Altitude Platform and using Ka band.
The use of a hemisphere with a ground plane yields an effective aperture size the equivalent of a full
sphere. The radiation properties of such structures (which may use multiple concentric dielectric layers)
has been described as a superposition of two regions of that of a spherical lens, namely those arising
from the real and virtual sources in the plane of observation. The spherical lens patterns may be
computed using an established modal expansion technique, which accounts for the boundaries at each
concentric layer. The hemisphere radiation theory agrees well with measurements of a single layer
polyethylene lens antenna at 28 GHz. This antenna also exhibits the desired wide scanning properties,
but a mediocre aperture efficiency of 40 %.
To improve upon this efficiency, the properties of two-layer lenses have been investigated, with an
emphasis on a constrained antenna total height and therefore limited by a fixed primary feed location.
Also, a few readily available dielectric materials have been chosen, rather than attempting to optimize a
Luneburg-like radial variation of dielectric constant. Within this framework, two-layer lenses based on a
Rexolite (r = 2.53) inner core and polyethylene (r = 2.28) outer layer have been put forward as a good
solution. Measurements of a two-layer prototype of 236 mm outer diameter indicate a gain of 35.1 dBi at
28 GHz, which represents an aperture efficiency of 68 %. This efficiency is comparable with a
conventional reflector antenna while the 2-layer lens structure offers a reduced height profile and the
potential to use multiple scanned beams over a very wide angle.
4.2.5.1
Acknowledgment
Thanks to Brent Wilkinson and Rob Easton at the Electronics and Physics mechanical workshop at The
University of York for their meticulous construction of the antenna components.
29/11/2006
Page 56 of 85
5 Mechanical Constraints Related To Train Application

Establishing and maintaining a continuous RF link from a moving train to a stationary balloon, or satellite
for that sake, is a demanding task. The main reasons come from the loss of line of sight (LoS) due to
the alternating landscape and swinging train tracks as is found in much of Europe. With the high speed
tilting train, this feature adds another difficulty for the antenna steering algorithms.
However, why shouldnt such a link still be feasible and reliable?
Adding a steering antenna on a train is feasible and has been tested by different companies as the
example shown hereby mounted on a French TGV train (Figure 51).
One can establish a preliminary list of difficulties to overcome such as:
*
Lateral and vertical vibration due to train bogy oscillations and rail swing
Rotations due to the rail curves
Tilt motion due to new generation tilt trains and their COG compensation.
Local obstructions from tunnels, mountains, narrow valleys, high buildings etc
From the literature one can get some information to help settle the requirements for the LOS control
algorithms of a mobile (azimuth / elevation type) antenna.
Typical measurements (from a Swedish X2000 high-speed train) show oscillations up to some 10 Hz,
which also interact with the human body attitude control, something every train commuter has
experienced (Figure 52).
Track curve rotations for a high-speed train are up to 4/s for curve radii of 1 to 1.25 km at 185 km/h or
440 to 800 metres at 125 to 145 km/h velocities.
Figure 51 Steerable antenna mounted on a TGV roof (http://telecom.esa.int/telecom/www)

29/11/2006
Page 57 of 85
Figure 52 Power spectral Density of lateral carriage body vibration

Furthermore the high-speed tilting trains, allowing a significant increase in travel speed on the old
sinuous rail tracks of the last century, add additional difficulties by their extra tilt angles and higher tilt
velocities, as illustrated in Figure 53.
Figure 53 Typical additional tilt of up to 8 from new generation tilting trains

However, on the plus side, the number of unknown degrees of freedom (DoF) for a direct LoS link are
limited because, to communicate in relay to a sequence of high-flying balloons, trains on ground only
move in a two-dimensional plane (approximately with respect to the HAPs. Furthermore, the train paths
are well known and can be learned with fuzzy logic type algorithms; this is also valid for most
obstacles. Similarly, there will be a gradual change in travelling speed of the trains because of their
limited acceleration and deceleration capabilities. This allows the dynamics of beam steering control
and maintenance to be separated into two categories:
29/11/2006
Page 58 of 85
Low speed (seconds or tens of seconds timeframe) for track and relay position tracking
algorithms, including relay hand-over due to known obstacles
High speed (fraction of a second to some 10-20 Hz timeframe) tilt / rotation motions due to train
oscillations, track swings etc
If the low speed components are highly predictable, so are also, at least partly, the high-speed
components.
Most regular train commuters have noticed that the train does similar swings and lurches at the same
location day after day. This is mainly due to imperfections of the rail tracks, and this behaviour could be
learned by the appropriate algorithms and stored locally. As most trains travel over the same path
week after week this accumulated knowledge can be efficiently reused.
To overcome the loss of communication due to tunnels and major obstacles, one obvious solution is to
use two or more terminals on a given carriage composition. Today most train compositions are made of
predetermined groups formed typically by 3 to 9 wagons of typically of 25-metres length each. Hence
the two receiver antennas could be situated some 75 to 225 metres apart allowing for a minimum
obstruction length without full link loss. For longer tunnels, as the tunnel walls are close to the train
structure, other types of antenna and WiFi transmission could be considered, possibly with one or two
fixed stations at each tunnel entrance. Similar systems are readily used today for the mobile phone
network in the tunnels in several countries.
Train velocity
Time in seconds to travel a train length in

metres
[km/h]
[m/s]
75
150
225
50
14
5.4
10.7
16.1
100
28
2.7
5.4
8.0
150
42
1.8
3.6
5.4
200
56
1.3
2.7
4.0
250
69
1.1
2.2
3.3
300
83
0.9
1.8
2.7
350
97
0.8
1.5
2.3
400
111
0.7
1.4
2.0
Table 9 Time in seconds to travel the distance of a train length for estimation of tunnel
obstruction times
29/11/2006
Page 59 of 85
6 State-of-the-Art of Electrical and Mechanical Steering

Technologies
6.1
State-of-the-Art of Mechanical Steering System
A review of potential tip-tilt steering systems was made of products available on the market that fulfil the
requirements for the steerable antenna mechanism. The principle technical requirements used to
evaluate the mechanism are given in the table below (Table 10). The payload considered had a total
mobile mass of 1kg, including interfaces.
Item
Requirements
Scan angles
Azimuth
180
Elevation
0-90 (15-90)
Angular speed
0.24/sec
Table 10 Resume of Tip-tilt mechanism requirements
6.1.1
Ball Aerospace
The Ball Aerospace two-axis control gimbals [44] are destined for a space application market and are
designed for the high performance market. The cost issue is considered as an expensive solution with
respect to a low-cost train application that is envisioned in the scope of this project.
Figure 54 Two-axis control gimbal
29/11/2006
Page 60 of 85
Table 11 Ball Aerospace control gimbals product range
6.1.2
Directed Perceptions Pan-Tilt Unit
The Directed Perceptions product range is destined primarily for use on closed circuit surveillance
systems to mount a video camera. This solution is considered as low-cost yet meets all technical
requirements for a train antenna application. It can be used with a direct mount or with nodal brackets if
the centre of rotation of the payload needs to be aligned with the rotation axes (see Figure 55). The pantilt unit is delivered with an integrated controller which easily interfaces to a PC host computer (see
Figure 56).
Figure 55 Nodal version with brackets aligns payload in pan and tilt rotation axes
29/11/2006
Page 61 of 85
Figure 56 Control interface to a host computer

Pan-Tilt Performance:
*
Maximum Rated Payload: 2.7kg
Maximum Speed: over 300 degrees/second
Acceleration/Deceleration: Trapezoidal, on-the-fly speed and position changes
Resolution: 3.086 arc minutes (0.051428)
Tilt Range (approx): 78 range (31 minimum up and 47 down ) with option of 80 down
(111 range)
Pan Range (approx): 159 (318 range) with option of 180 (360 range)
6.1.3
QuickSet International
Quickset [45] is the leading supplier of rugged all-weather antenna positioning systems for mobile or
fixed mount applications. Their products are used in space, marine, military and surveillance
applications. Their smallest range products are positioners to handle payloads of up to 10 kg. Even
though this particular product range is very versatile, it is considered as oversized for a steerable
antenna application on a train.
Medium
Positioners to handle payloads
up to 20 lbs. (10 kg)
QPT-15XD
QPT -20XD
QPT-20IC
15 lb capacity, 24 VAC fixed

speed models, 217.5 pan,
90 tilt.
20 lb capacity, 12 & 24 VDC

variable speed, 217.5 pan,
90 tilt.
With integrated microprocessor

control.
Table 12 Products available that meet requirements
29/11/2006
Page 62 of 85
Controllers
QuickSet offers a full suite of pan and tilt controllers for precise and smooth
control of pan and tilts and can easily be desktop or rack mounted.
Controllers range from basic pushbutton/joystick to full RS 232 & RS 422 PC
Computer Controlled models.
QuickTrac
QuickComm
Analog Controls
Models for 12 & 24 VDC output,

12 & 24 VDC input, 115 & 220
VAC input, position feedback
meters, autoscan, proportional
joystick operation, remote
joystick, rack mount.
Models for 12, 24, & 115 VDC

output, 12 & 24 VDC and 115 &
220 VAC input, position
feedback meters, RS-232/422,
presets, autoscan, proportional
joystick operation, rack mount.
Models for 12, 24, & 115 VDC

output, 24 & 115 VAC output, 12
& 24 VDC input, 115 & 220 VAC
input, position feedback meters,
autoscan, remote joystick,
pushbutton or joystick.
Table 13 Control interface to a host computer
Figure 57 Product features
6.1.4
AvL Technologies
The AVL designs [46] are based on a patented Roto-Lok drive system that utilizes highly-reliable
aircraft control cables in a redundant configuration to provide a zero-backlash, lightweight, very stiff and
accurate antenna positioner. This high-tech performance is achieved using low-tech components - by
simply wrapping the cable around the driver capstan several times before wrapping the larger driver drum.
The AvL product range is in the high-performance range (higher cost) as the Quickset International
products. It was therefore also excluded as a potential solution for the current application.
29/11/2006
Page 63 of 85
Figure 58 Data sheet of AvL Technologies antenna positioner models
6.1.5
CSEM custom-design antenna gimbal
Another approach considered was a custom, CSEM designed antenna gimbal as shown in Figure 59
and Figure 60 for a two-axis steering mechanism. This solution would have required a full development in
order to meet all the requirements. A number of critical tasks would have been required to perform a
detailed design study in order to size and select the various components. A dynamic analysis to size
the motors and gearing system would be necessary, integrating adequate encoders, electronics and
rack development. These tasks were considered but would have required extensive resources from the
project and it was discarded. Dedicated closed-loop control software with a tracking algorithm would
have also been a critical point to add to the development. For the reasons stated above, the drive
assembly needed to be a commercially available product with the effort emphasized in the integration of
29/11/2006
Page 64 of 85
the chosen antenna to the system and to develop the tracking algorithm and associated communication
requirements (antenna error signal input/drive commands).
Antenna
Altitude rotation
Azimuth rotation
Figure 59 CSEM custom-design drive system
Figure 60 CSEM custom-design antenna gimbal solution
6.2
6.2.1
Testbed of Mechanical Steering System

Antenna gimbal design based on Commercial Off-The-Shelf modules
Payload mounting at intersection of rotation axes

In order to perform tests with an antenna, a support structure will be designed and built to interface the
antenna to the gimbal drive units. The gimbal drive units are based on the Directed Perceptions Pan-Tilt
Unit. For a nodal configuration (centre of antenna intersect centre of axes rotation) of the device, the
baseline unit (left-below) can be modified with a custom support.
29/11/2006
Page 65 of 85
Figure 61 Directed Perceptions Pan-Tilt Unit in nodal configuration

The configuration above was not necessary for our application and a direct mechanical support to the
pan-tilt unit was manufactured in order to perform the tests with a tracking signal (see Figure 62).
Payload mounting at pan rotation axis
The more direct and simple mounting approach is the pan-tilt configuration that can easily support the
antenna.
Figure 62 Antenna gimbal unit design for CSEM demonstrator

The following information defines the standard unit that was used for the tests:
Model number: PTU-D46-70
Power supply product code: 239 AC/DC Power supply 220VAC/24VDC
C-Programmers Interface PTU-CPI
6.2.2
Signal tracking algorithm implementation
The signal tracking system was implemented based on power level detection. The receiver mechanism
is firstly initialized to point to a direction where it gets maximum of RF power level. The initialisation
follows a spiral trajectory referred to the Pan and the Tilt angles as presented in Figure 63. The system
continues oscillating around the initial maximum point, measuring power level in the neighbourhood of
the centre and moving to a new maximum point if any change in power level is detected. This is
illustrated by the schema bloc presented in Figure 64. The radiation pattern of the transmitting antenna
needs to be provided for the system for the initialization period and for defining the decision threshold
(Figure 66).
29/11/2006
Page 66 of 85

Tracking trajectory
50
40
30
Tilt angle [deg]
20
10
-10
-20
-30
-40
-50
-80
-60
-40
-20
20
40
60
80
Pan angle [deg]
Figure 63 Spiral trajectory followed by the pan-tilt mechanism
Nominal Pos0, power P0
Move to Pos1, power P1
Move to Pos0,
Pmax = max(P0, P1, P2, P3, P4)

Pos0 = corresponding position
Figure 64 Maximum finding algorithm
6.2.3
Tracking algorithm testbed
A testbed has been built for demonstrating the tracking algorithm performance at 27.5 GHz.
As a first step, two commercial large-bandwidth horn antennas (Figure 65) were used for the emission
and the reception. By definition, the antenna radiation pattern is measured at far-field region where
radiation is independent on the distance between TX and RX antennas. The distance where the far-field
region begins is defined in [39]:
R=
2D 2
, where D is the largest linear dimension of the antenna
29/11/2006
Page 67 of 85
Its easy to show that the far-field region starts at about 3 metres from the horn antenna, 4 metres from
the 64-element array, and 16 metres from the 256-element array. It is possible to make the test at
shorter distance in near-field region (for example 12 metres) under the condition that the radiation
pattern referred to should be measured at that distance. In the present work, the transmitter and the
receiver were placed at 4 metres apart.
As presented on the block diagram in Figure 66, the transmitter consists of a signal generator and the
horn antenna providing a beam of 37 width (HPBW) at 27.5 GHz. The antenna is interfaced to the
generator through coaxial cable. At the receiver, on a mobile platform, another horn antenna of the
same performance catches microwave energy in air and sends it back to a power detector which
performs RF-to-DC conversion. The resultant DC signal is amplified before being sent to the unit where
the trajectory is computed. The algorithm has been run on a PC using a data acquisition card (DAQ)
from National Instruments. It has been observed that the tracking quality is good as long as the
transmitter speed is at the same range as the sample speed of the PC (sample frequency 10 Hz). In
future research work, the geometry of the volume used in the algorithm can be further optimised. In
addition, with additional investment on software (LabVIEW Real-Time) and hardware (new acquisition
card), more rapid tracking should be achievable (at sample frequency 1KHz).
Figure 65 ETS Lindgren model 3116 antenna and interface
29/11/2006
Page 68 of 85
Mobile platform
Receiver
Transmitter
BF
amplifier
PAN-Tilt Unit
RF
detector
Rx
Motor/
encoder
RSSI
AxeControl
Signal
Generator
Antenna
37
Ry
Motor/
encoder
Antenna
Attenuation diagram
0
-5
-10
AxeControl
-15
Power Lines
Attenuation [dB]
PC
Labview
USB Card 6009
Control Alg.
Perturbations generation
-20
-25
-30
-35
-40
-45
-50
-80
-60
-40
-20
20
40
60
80
Opening angle
Figure 66 Block diagram of tracking algorithm testbed
Figure 67 Control unit with added trajectory perturbation
Figure 68 The 64-element subarray and the polarizer mounted on the Pan-Tilt Unit
29/11/2006
Page 69 of 85
6.3
Electrical beam steering: possibility of array built with directive elements
The objective of this section is to evaluate the characteristics and limitations of electrically steered
antenna arrays built with directive element, with special emphasis on the possibility of steering direction
and the gain that can be achieved at that direction.
Here we consider a MN-element planar array in xy-plane with uniform power distribution and uniform
element spacing. The normalized total array factor AFt is a combination of the AF in the x- and
y-directions [39]:
M
m =1
n =1
AFt = AFx AFy = e j (m1)(kd x sin cos + x ) e
j ( n 1) kd y sin sin + y
where and are respectively elevation of azimuth angle of reference point. is the phase difference
between the elements of the array. If we want to adjust the main beam of the array to a direction of =
0 and = 0, the phase difference between the elements need to be
x = k dx sin 0 cos 0
y = k dy sin 0 sin0
The total radiation of the array at a certain direction is obtained by multiplying the single element
radiation at reference point by the resulted AF at that direction. The above described formulas were
implemented using MATLAB. Simulations have been run for the upper-sphere above the array aperture.
Achievable array gain as a function of steering angle is presented in Figure 69.
Figure 69 Radiation of a 4-element array using directive element
29/11/2006
Page 70 of 85
6.4
Conclusions on mechatronic array antenna
A tracking algorithm has been developed using the maximum finding method. The algorithm was tested
on a COTS pan-and-tilt system with commercial horn antennas.
When mounted on a train, the system has to establish the communication link the first time without
prior knowledge of the position of the target (cold start). This approach requires an iterative search of the
sky to locate the other unit (opposite receiver/transmitter). CSEM has bread-boarded an azimuth
elevation antenna terminal and tested a few algorithms with success for such iterative scans, a first step
towards a complete mobile antenna link to be mounted and tested on a moving train.
29/11/2006
Page 71 of 85
7 Properties of Spherical Lens Antennas for High Altitude

Platform Payload
This self-contained chapter discusses the potential benefits of using multi-beam spherical lens antenna
for the HAP payload antennas in the context of the spot-beam architecture for a cellular communications
network.
7.1
Introduction
HAPs may exploit millimetre-wave spectrum by supporting multi-cell architectures. The overall data
capacity is governed by the properties of the antennas which serve the cells. Optimum coverage is
expected when each antenna radiation pattern is tailored to its respective cell, but this leads to a bulky
payload. In contrast, this section shows how a payload based on multi-beam spherical lenses, which
have wide-scan properties but also offer circular beams, can give comparable performance while offering
a much more compact payload. Models for spherical lens radiation patterns are presented which include
the effect of the primary feed beamwidth. This allows a comparison of aperture diameter, sidelobe level,
and the resulting co-channel interference, from which may be estimated the performance and relative
mass of different antenna payload types.
The ITU have allocated the 47/48 GHz band [6] for usage by HAPs worldwide, and the 27/31 GHz band
has also been allocated to about 40 countries [7]. At these frequencies it is difficult and expensive to
produce electronically re-configurable antennas based on phased-array principles and digital beam
forming (DBF). For example, DBF presents challenges associated with: the close spacing of radiating
elements, the requirement for a large number of elements, the need to minimize sidelobe levels, the cost
and complexity of RF active circuits, and the problem of scanning loss. The latter point is particularly
detrimental for HAPs communications because links to users at low elevation angles, where scanning
loss for a horizontal planar array are worst, experience also the longest free space and atmospheric
losses. Nevertheless a number of programmes are developing DBF antennas for Ka band e.g. [8] which
reports a 16-element module, and the Japanese HAPs programme [9] reporting a similar prototype for 28
GHz while conceding that horn antennas stabilized by mechanical gimbals are a more pragmatic
solution for 48 GHz.
The more conservative solution for the HAP antenna payload is a group of aperture antennas, where
each antenna serves a single quasi-stationary cell. This approach has been studied in some depth [4],
[5], [10] where the important relationship between the radiation patterns and the CIR levels on the ground
has been quantified for various channel re-use schemes. Ultimately, the economic viability of a HAP
cellular network may be determined by the extent to which it can maximize the re-use of spectrum.
This paper explores the properties of multi-beam antennas based on spherical lenses, since these can
yield a much more compact HAP antenna payload compared to the use of single-beam aperture
antennas. An estimate of payload mass compared to CIR coverage levels is introduced.
7.2
Dedicated beams for the HAP antenna payload
It is worth listing the pros and cons of using dedicated beams, i.e. one antenna per cell. Advantages
include:
each antenna pattern can be tailored to the shape required by each cell.
asymmetric beams yield equal size, circular cells, which aids tessellation of cells and
maximizes coverage [11].
low side-lobe levels are feasible e.g. -40 dB.
Disadvantages include:
29/11/2006
Page 72 of 85
A bulky, massive payload, which requires mechanical stabilisation.
A non-reconfigurable network.
These disadvantages can to some extent be overcome with multiple-beam antennas which allow for the
payload mass and volume to be reduced, but at the expense of eroding the advantages associated with
dedicated beams because it becomes more difficult to tailor each beam to each cell. Multi-beam
aperture antennas typically use a cluster of primary feeds to illuminate a secondary aperture which
might be a prime-focus reflector or lens, but the angular range of scanning is limited. For a HAP payload
a wide conical scan is needed e.g. 150 for users at elevation angles as low as 15. In the following
sections we explore the potential to use multiple-beam, wide-scanning antennas based on spherical
lenses, and quantify the CIR values which could be expected.
7.3
Properties of Spherical Lens Antennas
In the context of HAPs communications, the very useful property of antennas using spherical lenses is
their ability to form multiple beams over a wide range of angles without inducing any scanning loss. A
number of spherical lens variants are well know, and some are briefly reviewed below.
A classic type of spherical lens is the Luneburg lens [12], which focuses an incident plane wave at a
point on the lens surface. The spherical geometry allows for multiple beams to be produced from
multiple feeds [14], and it is this property which is attractive for a HAP payload. The Luneberg lens
requires a radial variation of dielectric constant:
r = 2 r
R
(1)
where r is radius within the lens, and R is the outer radius. Thus the dielectric constant varies from 1 at
the lens edge to 2 at the centre and there is no reflection loss due to abrupt transitions. This leads to a
very good aperture efficiency, but it somewhat problematic to manufacture. In practice such a lens is
often fabricated from a set of concentric shells and this inevitably degrades the efficiency: a useful
treatise is given by [17]. However, for limited bandwidth applications a two-shell design can give almost
equivalent performance [18].
A single shell spherical lens has no single focus, but exhibits a paraxial focus when a small proportion
of the lens is illuminated. The paraxial focus may lie inside or outside the lens outer radius. A low
dielectric spherical lens can give good performance - e.g. [19] reports on a Teflon lens (r = 2.08) for
wide-scanning automotive radar at 77 GHz - and is of course much more simple to manufacture than a
multi-shell lens.
Using the Luneberg lens aperture distribution given by [12], where the primary feed pattern is also taken
into account, the far field radiation pattern can be computed. The geometry is illustrated in Figure 70 for
a generalized lens where the focus may lie inside or outside the lens radius a. However, for the
remainder of the analysis we will assume a conventional Luneburg lens where f = a.
29/11/2006
Page 73 of 85

aperture plane
P( )
primary
feed
2a
f
Luneburg type
lens
Figure 70 Morgan's geometry [12] for generalized Luneburg lens showing aperture plane
A convenient model for the primary feed power pattern PPF as a function of angle '' is:
PPF = Cos[ ']
(2)
where
' = arcsin r
f
(3)
then from (2) and [12] we obtain for the radial power distribution P(r) in the aperture plane:
P (r ) =
Cosarcsin
r
f
(4)
f 2 r2
The power pattern P( ) of (4) may be derived in general by the transform to the far field. However, for
n=1 the r dependence of P(r) disappears and we have a uniform aperture. This of course yields the
highest aperture directivity for a given diameter 2a and also serves as a useful shortcut for deriving the
radiation patterns for each HAP payload antenna because the standard formula for a uniform aperture
may be used [13]. Where a more severe primary feed roll-off is applied i.e. by increasing n, the lens
aperture plane experiences a tapered distribution, leading to lower side-lobes but also a reduced
directivity; in such a case the lens diameter must be increased to recover the required beamwidth for
each cell (this is explored in detail later). Radiation patterns from (3), for a lens diameter of 10
wavelengths, are shown in Figure 71. When the n=1 model is used to generate CIR in a multi-cell
network as described below, this type of radiation pattern gives the lowest CIR values due to the high
side-lobe levels.
29/11/2006
Page 74 of 85

angle from boresight ()
5
10
15
25
30
n =1 (uniform aperture)
-10
relative gain (dB)
20
n =8 (45 feed)
-20
1
2
-30
-40
-50
Figure 71. Radiation patterns for Luneburg

lens of 10 wavelength diameter.
7.4
Figure 72. Cellular layout for 121 cells using

3 channels.
Power Contours and Co-channel Interference
In contrast to earlier studies where each antenna beam is optimized, the use of multiple-beam spherical
lens antennas leads to each beam of a given cluster having the same beamwidth and also having
circular symmetry. Taking the layout of 121 cells shown in Figure 72, where cells are arranged in 6
concentric hexagonal rings, we can specify an antenna beamwidth associated with each ring based on
the mean required beamwidth. This is a modification of the methodology of [5] and leads to non-circular
cellular power footprints but allows the use of a single spherical lens aperture for each ring and thus a
much more compact antenna payload. Elements of such a payload are illustrated in Figure 73, which
shows how the innermost group of cells are served by the smallest spherical lens with a cluster of
appropriately spaced feeds, and the outermost group of lens are similarly served by the largest spherical
lens. The lens diameters indicated are approximations based on uniform aperture illumination and a 28
GHz carrier frequency and thus serve as a benchmark estimate for minimum antenna dimensions at this
frequency.
feed cluster
feed cluster
lens
lens
55 mm
120 mm
15
56
cells in
ring 1
cells in
ring 6
Figure 73. Concept for antenna payload: multi-beam Luneburg lens for each cell group
Figure 74 shows power contours for a co-channel cell group where case (a) is for dedicated beams
which are asymmetric, and case (b) is for circular symmetric beams which could be generated using
multi-beam spherical lenses. The difference in these two methods is apparent when we observe how the
circular beams lead to non-circular cell power contours. The HAP payload in the former case comprises
29/11/2006
Page 75 of 85
121 individual antennas, while in the latter case it comprises 6 spherical lenses (one for each hexagonal
ring), plus a single antenna for the centre cell. The lenses require multiple feeds, these typically being
waveguide horns. In the outer ring the cell azimuth spacing is typically 10.1 and the lens diameter for
28 GHz would be approximately 120 mm. This leads to a feed horn spacing of about 21 mm which is
adequate for a small horn fed by circular waveguide or dielectric-loaded horn.
(a) dedicated beams.
(b) Multiple beam spherical lenses
Figure 74. Power contours for 1 of 3 channels (contour spacing is 1.5 dB).
The co-channel CIR values associated with the two different payload types of Figure 74 (a) and (b) are
presented as cumulative distributions in Figure 75. Comparing the dedicated, asymmetric beam case
with the multi-beam spherical lens case the CIR difference is typically between 5 - 7 dB across the
coverage area. This trend is expected because the dedicated beams model uses a flat sidelobe floor at 40 dB, while the circular beams model uses the theoretical sidelobe structure for uniformly illuminated
circular apertures (Figure 71) whose mean relative level is somewhat greater than -40 dB. Hence we are
able to present a comparison of CIR performance for (i) an established payload model using dedicated
beams and (ii) a more compact, less massive payload based on a group of minimum size spherical
lenses. The term "minimum size" refers to the assumption that the aperture is uniformly illuminated and
is thus, for each cell group, the smallest lens aperture which can be used to generate the required
beamwidths.
fraction of coverage area

1
dedicated beams,
-40 dB sidelobes
0.8
multi-beam, spherical
lenses (uniform aperture)
0.6
0.4
0.2
10
15
CIR
20
25
30
(dB)
Figure 75. CIR coverage distribution.
29/11/2006
Page 76 of 85
7.5
Reducing the spherical lens sidelobe level
In the multi-beam spherical lens case the CIR values can be improved by using bigger lenses with
tapered aperture distributions since this leads to lower sidelobe levels. In this section a recipe is sought
from which we may estimate the dimensions of a multi-beam spherical lens payload which would deliver
CIR values equivalent to the dedicated beam payload. A tapered aperture distribution implies that a
directive primary feed is used. This is also equivalent to n > 1 in (4) . The effect of the widening of the
secondary beam is shown in Figure 2.
If the primary feed pattern is given by (3) then, for a given primary feed HPBW p, since
cos (0.5p ) = 12
n
(5)
we have
n = log cos 0.5 p ( 12 )
(6)
and thus the primary feed pattern is described by a single parameter n which is derived from its HPBW.
We may then relate the aperture field distribution to the far field by employing the integral:
1
E ( ) = A(r ) r J 0 (k a sin ( ) r ) dr
(7)
r =0
A(r ) = P(r ) from (4). We may
where A(r) is the radial aperture field distribution which is given by
then evaluate (7) for a given combination of aperture radius a and primary feed HPBW so as to derive the
HPBW of the secondary field. One approach is to plot the main lobe patterns and in each case search
for the half-power points using a computer algorithm. This procedure lead to the results summarized in
Figure 76 where the primary feed HPBWs chosen are 30, 45, 60 and 90 and these parameters give
rise to the discrete data points which are shown.
secondary
HPBW ()
a=2
a=4
a=5
a=6
a=7
a=8
25
relative secondary
HPBW
20
30
1.8
15
1.6
10
1.4
45
1.2
60
90
1.0
30
40
50
60
70
80
90
primary HPBW ()
(a) Secondary HPBW versus primary HPBW for radii in

wavelengths as labelled
lens radius a
(b) Beamwidth widening factor compared to

uniform aperture.
Figure 76. Relationship between primary feed and secondary HPBW for Luneburg lens.
In Figure 76(a) can be seen the expected trends that a larger aperture gives rise to a narrower secondary
beam, while a narrower primary feed beam gives rise to a wider secondary beam. The latter effect is also
29/11/2006
Page 77 of 85
associated with a reduction in sidelobe levels. Figure 76(b) shows the increase in secondary beamwidth,
relative to the uniform aperture (n=1), as a function of aperture radius a and for the four cases of primary
HPBW chosen. Here we find a very useful result: for a given primary feed, the widening of the secondary
beam is by an approximately constant factor for the range of aperture radii chosen.
Considering the HAP antenna payload, we now have a recipe for estimating the amount by which the
antenna beamwidths would need to be increased when the lens aperture experiences an amplitude taper
brought about by a generic primary feed. The beam widening must be countered by a commensurate
and linear increase in the lens radius so as to recover the beamwidth which is required by each cell
group. The calculated scaling factors are listed in Table 14 and summarized as an extrapolated function
of primary feed HPBW in Figure 77.
primary feed
HPBW ()
beam scaling
factor
first sidelobe level relative

to peak gain. (dB)
uniform aperture
-17
90
1.06
-19
60
1.22
-24
45
1.43
-30
30
1.87
-44
Table 14. Summary of Luneburg lens beam widening and sidelobe level reduction for various
primary feeds.
increase in secondary
HPBW
1.8
fraction of
area
1
1.6
0.8
1.4
0.6
1.2
0.4
0.2
40
60
80
100
120
CIR
5
10
15
20
25
30
primary feed HPBW ()
Figure 77. Increase in secondary beamwidth,

compared to that of uniform aperture, as a
function of primary beamwidth.
Figure 78 CIR distributions for dedicated

and multi-beam payloads.
The derivation of CIR for the co-channel cell group now requires (7) to be evaluated so as to derive the
power footprints of each cell. (The integration involved in (7) is much more computationally intensive than
the previously reported radiation pattern models and thus leads to much longer computer run times.)
Some results for computed CIR distributions are shown in Figure 78 for multi-beam antennas using
various primary feed beamwidths and where the aperture diameter is scaled accordingly so as to
maintain the required secondary beamwidth.
29/11/2006
Page 78 of 85
7.6
Relative payload mass
To close the analysis we present an estimate of the total mass for the HAP antenna payloads which
have been discussed. For the dedicated beam payload, a working estimate for total lens mass can be
derived as follows: Using the mass of our experimental asymmetric beam lens antenna [11] as a
benchmark, and noting that, to a good approximation, there are straightforward relationships between
half power beamwidth (HPBW), mass, and lens diameter D:
HPBW 1
D
mass D 3
and
hence by scaling to the 466 g. mass of the experimental polyethylene lens whose HPBW in the narrow
plane1 is 5 we obtain an estimate for each lens mass in the dedicated beam payload:
5
mass = 466
HPBW
degrees
(g)
Using the iterative method of [5] to derive HPBW for each cell in the 121 cell network yields a total lens
mass of 28.8 kg. The figure is somewhat tentative, and has been derived for a HAP at a height of 20 km
above a 60 km diameter service area. Clearly, if these parameters are altered, the antenna beamwidths
(which are tailored to each cell subtended angle) are altered and thus the lens masses also change. The
mass of the primary feeds has not been considered in this analysis.
The estimate for the mass of the multi-beam spherical lens payload can be derived from the diameter of
each lens and (assuming a single shell lens design) the material specific gravity. Again assuming a
polyethylene material, whose density is 947 kg / m3, and uniformly illuminated apertures for the
minimum size payload case, a mass of 1.93 kg is derived for the group of 6 spherical lenses. The figure
would be less for multi-shell or Luneburg type lenses. Again, the mass of the primary feeds has not
been considered, as it is possible that these could be very similar for the two payload cases and their
number would also be the same.
The trade-offs between use of a dedicated beam payload assuming -40 dB sidelobe floors, and multibeam spherical lenses assuming close-to-uniform aperture illumination, are summarized in Table 15. It
is stressed that the masses presented are estimates based on extrapolation from those of
representative components. Nevertheless we can see that, subject to these various assumptions, the
multi-beam lens approach leads to approximately 93 % saving in the mass of the dielectric lenses at the
expense of a CIR degradation typically of 6 dB.
dedicated beams
multi-beam spherical
lenses:
(i) minimum aperture
size
number of
primary
feeds
121
number of
lenses
estimate of total
lens mass (kg) (for
121
28.8
CIR (dB)
(minimum /
maximum)
12 / 24
121
1.93
6 / 18
121
3.50
12 / 29
28 GHz)
(ii) scaled aperture with

60 primary feeds
Table 15. Comparison of payload properties and CIR values for dedicated beams and multibeam spherical lenses serving 121 cell network
The experimental asymmetric beam lens antenna has a circular cross section whose diameter is dictated by the
narrow (elevation) beamwidth of the cell. The beamwidth in the azimuth direction is broadened by modifying the
lens surface [11].
29/11/2006
Page 79 of 85
7.7
Conclusions for HAP Spherical Lens Antennas
HAPs offer an effective way of exploiting mm-wave spectrum by supporting multi-cell architectures. The
viability of such systems is largely determined by the overall system data capacity, which is in turn
governed by the properties of the antenna payload which serves the cells on the ground. In previous
studies, ideal antenna beams have been used to model carrier-to-interference ratio. In such cases,
dedicated aperture antennas such as lens antennas may be used to produce the required beam shapes
which are in general asymmetric. This approach leads to one antenna for each cell and thus a bulky
payload.
Where multi-beam antennas are used the mass and volume of the HAP payload can be much reduced.
A convenient approach is to employ spherical lens antennas using multiple feeds. This work has
considered the properties of such antennas and shown how for a 121 cell network the payload may be
reduced to just 6 multi-beam antennas. Disadvantages of this antenna type is that radiation patterns are
in general of circular cross-section, which leads to non-circular cell footprints, and a compromise
beamwidth must be chosen for a given cell group.
The resulting CIR levels have been computed for spherical lens antennas where the aperture distribution
is uniform, this being the smallest antenna but with the highest sidelobes. The technique was then
extended to consider the effect of the primary feed beamwidth, since this may be chosen to taper the
aperture distribution and thus reduce sidelobe levels, but must be accompanied by an increase in lens
diameter so as to maintain the required directivity for the cell group. A generalized recipe was found
whereby the aperture scaling term was derived for a range of primary feed beamwidths. This allowed
direct comparison of the relative size of the antenna payload for given CIR levels. The trade-off is one
where the smallest antennas have the highest sidelobes and hence lowest CIR, while increasing the
antenna size leads to increased CIR levels.
The benchmark payload, using 121 dedicated antennas where each is assumed to have a mean
sidelobe floor at -40 dB, offers CIR levels between 12 dB and 18 dB and was shown to have a total
dielectric lens mass estimated at 28.8 kg. Compared to this, a payload of 6 multi-beam spherical lenses
of minimum possible size would offer a mass reduction of at least 93 % (assuming single shell, not
Luneburg lens) but a CIR degradation of around 6 dB. It was shown that the CIR levels offered by the
dedicated antenna payload may also be obtained by the multi-beam spherical lens payload if each lens
aperture diameter is increased by a factor of 1.22 in conjunction with primary feeds of 60 half-power
beamwidth. The total lens mass in this case would be approximately 88 % less than that of the
benchmark payload.
While operational HAP communications services have yet to be deployed in practice, it has been shown
that a group of multi-beam spherical lenses could offer a very practical and compact antenna payload.
This would support multi-cell architectures with adequate levels of co-channel interference and with a
very considerable mass saving compared to a payload of a type which uses one lens antenna for each
cell.
29/11/2006
Page 80 of 85
8 Final Conclusions
In this report various candidate solutions have been considered across a wide spectrum but two widely
differing clusters have emerged. In one cluster, the more straightforward, conservative or near-term
solutions would be offered by proven antenna technologies such as microwave apertures (waveguide
horns), lenses and reflectors, integrated into either the HAP payload for serving cellular architectures, or
a mechanical control system for deployment on a high-speed train. Such solutions would be
characterized by excellent electromagnetic performance, but perhaps rather unwieldy mechanical
systems to implement the steering functionality. In the opposite cluster would be the 'smart' or digital
beam forming (DBF) antennas: here, no moving parts would be needed, and the smart antenna would
theoretically offer some astounding advantages: multi-beam apertures offering null steering for
interference cancellation and very rapid beam scanning. Due to project resource limitations, DBF
techniques were not explored experimentally but were investigated theoretically in Workpackage 3.3.
The array antenna developed at CSEM, primarily considered for vehicular communication, was
implemented using Strip-slot-foam Inverted patch technology - a multi-layer substrate approach which
overcomes the bandwidth limitations of single-layer printed antennas. This yielded a lightweight antenna
easily driven by servo motors, and several approaches to gimbals were considered. Tracking algorithms
were developed and a laboratory testbed so as to demonstrate a representative, integrated antenna with
a control system.
For the HAP antenna, dielectric hemisphere lenses were developed to validate a family of antenna types
based on lens technology, at York. A notable and novel result was the development of a two-layer lens
antenna, this being loosely inspired by the Luneburg antenna but offering a simpler and more reliable
route to construction using well-behaved, low-loss materials. The antenna offered a gain of 35.4 dBi at 28
GHz, being 68 % aperture efficient. While a steering mechanism for this antenna was beyond the
project's resources, a number of concepts have been discussed and sketched in some detail.
In the final chapter, the application of spherical lens antennas for a multi-beam HAP payload for cellular
communications, where these are shared by clusters of primary feeds, was theoretically investigated at
the system-level. Here, the benchmark was the 'dedicated beam' payload model from HeliNet, where
elliptic beam antennas are dedicated to each cell on the ground. The lens antenna solutions put forward
have slightly less advantageous beam shapes, but can very substantially save on payload volume and
mass due to the shared apertures.
29/11/2006
Page 81 of 85
8.1
Bibliography
J. Thornton, Wide-scanning Multi-layer Hemisphere Lens Antenna for Ka band. IEE Proceedings
Microwaves, Antennas & Propagation, Accepted for publication 2006
John Thornton, Scattering in Stratified Dielectric Lens Antenna with Covered Feed, International Journal
of RF and Microwave Computer Aided Engineering, accepted for publication 2006.
J. Thornton, A Scanning Hemispherical Lens Antenna for 10-30 GHz Where The Feed is Modelled as a
Linear End-Fire Array. Loughborough Antennas and Propagation Conference (LAPC) 2006:
J Thornton, D.A.J.Pearce, D.Grace, M.Oodo, K.Katzis,T.C.Tozer, Effect of Antenna Beam Pattern and
Layout on Cellular Performance in High Altitude Platform Communications, International Journal of
Wireless Personal Communications. Vol. 35 No. 1-2, October 2005, pp.35-51
J. Thornton, Scanning Ka-band Vehicular Lens Antennas for Satellite and High Altitude Platform
Communications, 11th European Wireless Conference, Nicosia, 10-13 April 2005
J. Thornton and D. Grace, Effect of Antenna Aperture Field on Co-Channel Interference, Capacity and
Payload Mass in High Altitude Platform Communications, ETRI Journal.Vol. 26, Number 5, October
2004, pp.467 - 474 also "Paper of the Year" Award 2004, ETRI Journal.
J. Thornton, A Low Sidelobe Asymmetric Beam Antenna for High Altitude Platform Communications,
IEEE Microwave and Wireless Components Letters. Vol. 14, No. 2, February 2004, pp.59-61
John Thornton, Properties of Spherical Lens Antennas for High Altitude Platform Communications, 6th
European Workshop on Mobile/Personal Satcoms & 2nd Advanced Satellite Mobile Systems (EMPS &
ASMS). September 21-22. ESTEC, European Space Agency
J. Thornton, T. Tozer, Adaptation of ITU Recommended Antenna Pattern for Prediction of Multi-Cell CoChannel Interference in High Altitude Platform Communications in the 47/48 GHz Band, The Seventh
International Symposium on WIRELESS PERSONAL MULTIMEDIA COMMUNICATIONS (WPMC) 2004,
Abano Terme, Italy, 12-15 September 2004
Q.Xu, J.R.Farserotu, J-F.Zurcher and A.Skrivervik, "Broadband Small Array Antenna for High Altitude
Platforms Communications", 15th IST Mobile & Wireless Communications, Myconos, 4-8 June 2006
29/11/2006
Page 82 of 85
9 References
[1]
N. J. Collela, J. N. Martin and I. F Akyildiz, "The HALO network", IEEE Commun. Mag., vol. 38,
pp. 142-148, June 2000.
[2]
T. C. Tozer and D. Grace, "High-altitude platforms for wireless communications", IEE Electronics
and Communications Engineering Journal, Vol. 13, June 2001, pp. 127-137.
[3]
http://www.capanina.org
[4]
J. Thornton, D. Grace, C. Spillard, T. Konefal, T. Tozer, "Broadband communications from a high

altitude platform - The European HeliNet programme", IEE Electron. Commun. Eng. J., vol. 13,
pp. 138-144, June 2001
[5]
J. Thornton, D. Grace, M. H. Capstick, T. C. Tozer."Optimising an Array of Antennas for Cellular

Coverage from a High Altitude Platform", IEEE Transactions on Wireless Communications, Vol.
2, No. 3, May 2003 pp.484 - 492
[6]
Recommendation ITU-R F.1500, "Preferred characteristics of systems in the FS using high

altitude platforms operating in the bands 47.2-47.5 GHz and 47.9-48.2 GHz", International
Telecommunications Union, 2000. (Also ITU-R S.672)
[7]
World Radiocommunication Conference (Geneva, 2003), Draft Resolution [HAPS 28-31] (WRC03)
[8]
A. Dreher, "Digital-Beamforming Antenna for Broadband Communications in Ka-Band", 6th

International Symposium on Wireless Personal Multimedia Communications (WPMC 2003), 1922 October 2003, Japan.
[9]
R. Miura and M. Suzuki, Preliminary Flight Test Program on Telecom and Broadcasting Using
High Altitude Platform Stations. Wireless Personal Communications, vol 24, issue 2, 2003, pp.
341-361
[10]
J.Thornton."Antenna Technologies for Communications Services from Stratospheric Platforms."

4th International Airship Convention and Exhibition, Cambridge, UK, 28-31 July 2002
[11]
J. Thornton, "A Low Sidelobe Asymmetric Beam Antenna for High Altitude Platform
Communications", IEEE Microwave and Wireless Components Letters. Vol. 14, No. 2, February
2004, pp. 59-61.
[12]
S. M. Morgan, "General Solution of the Luneberg Lens Problem", Journal of Applied Physics Vol
29, No. 9, Sept. 1958 pp. 1358 - 1368
[13]
C.A.Balanis, Antenna Theory, Analysis and Design, 2nd Ed. Wiley,1997 p.606
[14]
M. A. Mitchell, J. R. Sanford, L. E. Corey, R. A. Moore and V. P. Pusateri, "A multiple-beam

multiple-frequency spherical lens antenna system providing hemispherical coverage", In
Proceedings of Sixth International Conference on Antennas and Propagation (ICAP 89), 4-7 Apr
1989, Coventry, UK, pp. 394-398 vol.1
[15]
A. D. Olver, P. J. B. Clarricoats, A. A. Kishk, and L. Shafai, Microwave Horns and Feeds, IEE
Press, 1994.
[16]
J. R. Sanford, "Scattering by Spherically Stratified Microwave Lens Antennas", IEEE Trans.

Antennas Propagat., Vol. 42, No. 5, May 1994, pp 690 - 698.
[17]
"Use of Luneburg Lens for Low Profile Applications", Datron/Transco Inc. originally published in
Microwave Product Digest, December 1999.
29/11/2006
Page 83 of 85
[18]
H. Mosallaei and Y. Rahmat-Samii, "Non-uniform Luneburg and 2-shell lens antennas: Radiation
characteristics and design optimization," IEEE Trans. Antennas Propagat., vol. 49, no. 1, pp. 6069, Jan. 2001.
[19]
B. Schoenlinner, X. Wu, J. P. Ebling, G. V. Eleftheriades and G. M. Rebeiz, "Wide-Scan

Spherical-Lens Antennas for Automotive Radars", IEEE Transactions on Microwave Theory and
Techniques, Vol. 50, No. 9, September 2002, pp 2166 2175
[20]
J. Thornton, Scanning Ka-band Vehicular Lens Antennas for Satellite and High Altitude Platform
Communications, 11th European Wireless Conference, Nicosia, 10-13 April 2005
[21]
X. Zhang, S. Kado, T. Hiruta and Y. Miyane, Development of a 26 GHz band High Gain Flat
Antenna for FWA Systems, Hitachi Cable Review No.22, August 2003, pp.16-19.
[22]
H. Y. Yang and N. G. Alexpoulos, Gain Enhancement Methods for Printed Circuit Antennas
through Multiple Superstrates, IEEE Trans. AP, vol. AP-35, No. 7, pp. 860-864, July 1987.
[23]
W. Choi, Y. H. Cho, C.-S. Pyo and J.-I. Choi, A High-Gain Microstrip Patch Array Antenna Using
a Superstrate Layer, ETRI Journal, vol. 25, No. 5, pp. 407-411, October 2003.
[24]
Patent US 2004/0104852 A1, Microstrip Patch Antenna and Array Antenna Using Superstrate,
June. 3, 2004.
[25]
X. Zhang, S. Kado, T. Hiruta, Y. Miyane, Development of a 26GHz band High Gain Flat Antenna
for FWA Systems, Hitachi Cable Review, No. 22, pp. 16-19, August 2003.
[26]
R. B. Waterhouse, Stacked Patches Using High and Low Dielectric Constant Material
Combinations, IEEE Trans. AP, vol. 47, pp. 1767-1771, December 1999.
[27]
W. Choi, C. Pyo and J. Choi, Broadband Circularly Polarized Corner-truncated Square Patch
Array Antenna, 0-7803-7330-8/02, 2002 IEEE, pp. 220-223.
[28]
L. Young, L. A. Robinson and C. A. Hacking, Meander-Line Polarizer, IEEE Trnas. AP, May,
1973, pp.376-378.
[29]
Technical and operational characteristics for the fixed service using high altitude platform stations
in the band 27.5-28.35 GHz and 31-31.3 GHz, Recommendation ITU-R F.1569, ITU
[30]
M. Clenet and L. Shafai, Design and measurement of a wide operating bandwidth single layer
microstrip antenna array at millimeterwave frequency, IEEE Antennas and Propagation Society
International Symposium, 2000, vol. 2, pp. 486-489.
[31]
G. Kumar and K. C. Gupta, Nonradiating edges and four edges gap-coupled multiple resonator
broad-band antennas, IEEE Trans. Antennas Propagation., vol. AP-33, no. 2, February 1985, pp.
173-178.
[32]
C. K. Aanandan, P. Mohanan, and K. G. Nair, Broad-band gap coupled microstrip antenna,

IEEE Trans. Antennas Propagat., vol. 38, no. 10, October 1990, pp. 1581-1586.
[33]
M. Clenet and L. Shafai, Wideband single layer microstrip antenna for array applications,
Electronics Letters., vol. 35, Issue 16, Aug. 1999, pp. 1292-1293.
[34]
J.-F. Zrcher and F. E. Gardiol, Broadband patch antennas, Artech House, Norwood, MA, 1995,
ch. 3.
[35]
J.-F. Zrcher, A. Skrivervik, O. Staub, and S. Vaccaro, A compact dual-port, dual frequency
printed antenna with high decoupling, Microwave Opt. Tehnol. Lett., 19 (1998), pp.131-137.
[36]
J.-F. Zrcher, Q. Xu , A. Skrivervik, and J. R. Mosig, Dual-frequency, dual-polarization 4-port

printed planar antenna, Microwave Opt. Tehnol. Lett., vol. 23, No. 2, October 20, 1999, pp. 75-78.
29/11/2006
Page 84 of 85
[37]
R. C. Hall, Full-wave aperture coupled patch antenna, Electronics Letters, vol. 29, 1993, pp.
2073-2074.
[38]
Ansoft Designer 2 User Manual, Ansoft Corporate, 2005.
[39]
C. A. Balanis, Antenna theory analysis and design, 2nd Ed., John Wiley & Sons Inc., USA, 1997,
ch. 6.
[40]
Certificate of calibration conformance, ETS Lindgren, Texas, 2004.
[41]
L. Young, L. A. Robinson and C. A. Hacking, Meander-line Polarizer, IEEE Trans. on Antenna

and Propagation, pp. 376-378, May, 1973.
[42]
F . Hodjat and S. A. Hovanesssian, Nonuniformly spaced linear and planar array antennas for
sidelobe reduction, IEEE Trans. Antennas Propagat., vol. AP-26, No. 2, pp. 198-204, March,
1978.
[43]
G. L. Matthaei, L. Young and E. M. T. Jones, Microwave filters, impedance-matching networks,

and coupling structures, McGraw-Hill, New York, 1964, chap.9.
[44]
http://www.ballaerospace.com/printmaterial.html
[45]
www.quickset.com
[46]
http://www.avltech.com/rotoLok.php
[47]
http://telecom.esa.int/telecom/
[48]
J. T. Pearson and R. M. Goodall, An active stability system for a high speed railway vehicle,
Electronic systems and control division research 2003, pp. 11-14, Loughborough University, UK.
[49]
J. Frstberg, Motion-related comfort in tilting trains, VTT rapport 449A, 2000.
[50]
S. Johansson, Active control of lateral vibrations in a high-speed train, Proc. Promote IT 2003,
Visby, 2003.
[51]
A.Abbaspour-Tamijani and Kamal Sarabandi, "An Affordable Millimetre-Wave Beam-steerable

Antenna Using Interleaved Planar Subarrays", IEEE Transactions on Antennas and Propagation,
Vol. 51, No. 9, September 2193
29/11/2006
Page 85 of 85

Cap d24 Wp32 Uoy Pub 01

Uploaded by

Copyright:

Available Formats

Cap d24 Wp32 Uoy Pub 01

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cap d24 Wp32 Uoy Pub 01

Uploaded by

Copyright:

Available Formats

FP6-IST-2003-506745 CAPANINA

Deliverable Number D24

Report on steerable antenna architectures and critical RF

Contractual Date of Delivery to the CEC

Actual Date of Delivery to the CEC

Qin Xu, John Thornton

Participant(s) (partner short names):

Editor (Internal reviewer)

John Gerrits, CSEM

Estimated person months

Security (PUBlic, CONfidential, REStricted)

Total number of pages (including cover):

Report on steerable antenna architectures and critical RF circuits performance CAP-D24-WP32-UOY-PUB-01

UOY revision after editorial John Thornton

Revisions to sect. 4.1 & 6

YEC revision with CSEMs Graham Long

following Graham Long

Document Approval (CEC Deliverables only)

Editor (Internal reviewer)

On behalf of Scientific Board

Report on steerable antenna architectures and critical RF circuits performance CAP-D24-WP32-UOY-PUB-01

Report on steerable antenna architectures and critical RF circuits performance CAP-D24-WP32-UOY-PUB-01

POSSIBLE APPROACHES FOR THE RF, EM AND MECHATRONICS ASPECTS OF

System Requirements and Link budgets ............................................................................ 12

Required Scan Range....................................................................................................... 14

Coverage Area ................................................................................................................. 14

CANDIDATE SOLUTIONS AND PRIOR ART.................................................................14

3.2.1 Mechanically steered antenna......................................................................................... 15

Mass of typical antennas ............................................................................................................................. 15

3.2.2 Mechanically steered small high gain array ...................................................................... 16

Array factor .................................................................................................................................................... 18

Implications of wide element spacing......................................................................................................... 19

Effect of antenna element radiation pattern............................................................................................... 20

Effect of beam scanning ............................................................................................................................... 21

Discussion: phased array techniques ........................................................................................................ 24

3.2.4 Array variants and combined electronic and mechanical steering........................................ 24

3.2.6 The spherical lens antenna ............................................................................................. 27

Recommendations following the survey of candidate technologies......................................... 30

3.3.1 Phased array and DBF................................................................................................... 30

EXPERIMENTAL INVESTIGATIONS OF PREFERRED SOLUTIONS ........................32

Broadband Printed Array Antenna...................................................................................... 32

4.1.1 Introduction ................................................................................................................... 32

Report on steerable antenna architectures and critical RF circuits performance CAP-D24-WP32-UOY-PUB-01

4.1.6 Single high gain broadband antenna design...................................................................... 36

Loss in the antenna................................................................................................................................... 46

Loss in the polarizer.................................................................................................................................. 47

4.1.11 Conclusion for printed array antenna................................................................................ 47

Hemispherical lens prototype. ........................................................................................... 48

4.2.1 Constant index lens ....................................................................................................... 48

Analysis of single-shell and two-shell lens designs................................................................................ 51

Two-layer lens performance......................................................................................................................... 54

4.2.4 Mechanical feed steering concepts.................................................................................. 55

MECHANICAL CONSTRAINTS RELATED TO TRAIN APPLICATION........................57

STATE-OF-THE-ART OF ELECTRICAL AN D MECHANICAL STEERING

State-of-the-Art of Mechanical Steering System .................................................................. 60

6.1.1 Ball Aerospace.............................................................................................................. 60

Testbed of Mechanical Steering System............................................................................. 65

6.2.1 Antenna gimbal design based on Commercial Off-The-Shelf modules.................................. 65

Conclusions on mechatronic array antenna......................................................................... 71

PROPERTIES OF SPHERICAL LENS ANTENNAS FOR HIGH ALTITUDE