CONCEPT DESCRIPTION: SYMMETRIC DISH … · Approved for release as part of dish CoDR ... For example press‐forming and ... design, to analysis and calculation up to production of

Name Designation Affiliation Date Signature

Submitted by:

Arnold van Ardenne

Johan Pragt

Marco Drost

Systems, business relations

Management

Mechanical design

ASTRON

ASTRON

ASTRON

2011‐06‐15

2011‐06‐15

2011‐06‐15

Co Authors and contributors: M. Ivashina (Chalmers University), Robbert Bakker and Jaap Dekker (Airborne), Raymond van den Brink and Jan Geralt Bij de Vaate (ASTRON)

Approved for release as part of dish CoDR documents:

Neil Roddis SPDO 2011‐06‐15

CONCEPT DESCRIPTION: SYMMETRIC DISH (VERSION C)

Document number .................................................................. WP2‐020.045.010‐TD‐003

Revision ............................................................................................................................ B

Author ...................................................................................................... see table below

Date ................................................................................................................. 15‐06‐2011

Status ......................................................................................................................... Final

Revision 2.0, May 2011:

Page 2 of 50

DOCUMENT HISTORY

Revision Date Of Issue Engineering Change

Number

Comments

2.0 15‐06‐2011 ‐ Final

DOCUMENT SOFTWARE

Package Version Filename

Wordprocessor MsWord Word 2007

Block diagrams

Other

ORGANISATION DETAILS

Name SKA Program Development Office

Physical/Postal

Address

Jodrell Bank Centre for Astrophysics

Alan Turing Building

The University of Manchester

Oxford Road

Manchester, UK

M13 9PL

Fax. +44 (0)161 275 4049

Website www.skatelescope.org


Page 3 of 50

Summary

This document describes the concept for a symmetric dish for SKA antenna reflector based on the

use of novel thermoplastic composite material in the context of an otherwise classical (but not

limited to that) telescope design. While recognizing that detailed work needs to be done to advance

this work, the present outcome points to an attractive low cost and mass producible design with low

risk involved.

In general, the symmetric dish concept seems very attractive, because its design is relatively

straightforward while wide spread knowledge and performance experience is available throughout

the world.

The presented concept design is classical in its primary functions and design approach thereby

decreasing design risks. The novelty involves the reflector structures material, which is based on

carbon re‐enforced thermoplastic material not (yet) used for this kind of applications looking highly

promising and feasible as a result of our studies. It is of interest to note that other markets e.g.

automotive and aerospace are interested in applying this material pushing further developments.

From the technical study as presented in this report, the following conclusions can be made:

‐ A symmetric dish is significant lower in cost then an offset Gregorian. The difference

presented in this document is such that it might serve as a serious concept for SKA phase 1.

‐ The presented dish specifications wavelength range opens an area which is originally

scheduled for phase 2 of SKA, but probably can already be applied in phase 1 of SKA. This is

likely to be beneficial for the technology development and for science of SKA.

Thermoplastic carbon reinforced composite materials might be an alternative for the more common

dish materials.

A collaborated effort between a Dutch Industrial project group, Chalmers University and ASTRON

provided input for this document. The main technical input is by the Industrial Group, lead by

Airborne in The Hague, the Netherlands. Airborne has experience in building dish structures e.g. for

ALMA. The Industrial project group in general has experience with thermo plastic composites from

design and calculation up to production. Chalmers University provided the integral radio design and

electro‐magnetic calculations for the overall reflector to which the electro magnetic design for the

eleven feed was input. ASTRON provided the specifications and performed system‐ and mechanical

design and RF testing of materials.

The outcome from the technical studies generates some questions and conclusions that need to be

answered or discussed:

‐ Thermo plastic carbon re‐enforced materials are an attractive possibility for use and

implementation in low cost, high performance dish structures. Its main advantage is the

lower material and manufacturing cost in comparison with conventional high‐end epoxy

carbon reinforced materials. For example press‐forming and automated welding can be

used, which is not possible in thermoset composites like carbon‐epoxy. The advantage with

respect to classical steel structures is based on its stiffness, far lower weight and very low


Page 4 of 50

coefficient of thermal expansion combined with classical process techniques (warm forming

of the shape and automated welding of the construction) and potential for mass production.

‐ Based on a comparative study, a symmetric dish is significant lower in cost then other e.g.

offset designs such as Gregorian positioning this work as highly relevant and as a serious

concept for SKA.

‐ The presented dish specifications wavelength range opens an area originally scheduled for

SKA phase 2, but probably already likely to be applied in earlier SKA phases. This might be

beneficial for the technology development and for science of the SKA.

The contributors of this report are very interested to advance the concept for the SKA and are open

for feedback and commenting as input to improvements and next steps.


Page 5 of 50

TABLE OF CONTENTS

1 INTRODUCTION ........................................................................................... 9

1.1 Purpose of the document ....................................................................................................... 9

2 REFERENCES ............................................................................................. 11

3 CONTEXT ................................................................................................. 12

3.1 SKA hierarchy ........................................................................................................................ 12

3.2 Role of the symmetric dish in the Dish Array ....................................................................... 13

3.3 Comparative discussion of Symmetric versus Gregorian offset reflector concept .............. 14

3.3.1 Discussion of the dish concept from an RF perspective ............................................... 14

3.3.2 Discussion from scientific / electromagnetic perspective ............................................ 15

3.3.3 Structural design ........................................................................................................... 15

3.3.4 Manufacturing cost ....................................................................................................... 16

4 REQUIREMENTS ......................................................................................... 17

4.1 Functional Requirements – reflector structure .................................................................... 18

4.2 Specification compliance – reflector structure ..................................................................... 19

5 MAIN REFLECTOR DESIGN ............................................................................. 21

5.1 Material selection ................................................................................................................. 21

5.2 Structural design ................................................................................................................... 22

5.3 Structural analysis ................................................................................................................. 24

5.3.1 Material properties ....................................................................................................... 24

5.3.2 Load cases ..................................................................................................................... 25

5.3.2.1 Gravity ....................................................................................................................... 25

5.3.2.2 Thermal loading ........................................................................................................ 26

5.3.2.3 Wind loading ............................................................................................................. 26

5.3.3 FEA results ..................................................................................................................... 27

5.3.4 Performance.................................................................................................................. 27

6 ANTENNA MOUNT DESIGN ............................................................................ 31

6.1 Pedestal design ..................................................................................................................... 31

6.2 Balancing ............................................................................................................................... 31

6.3 Drive system .......................................................................................................................... 32

7 TELESCOPE ELECTRO‐MAGNETIC DESIGN ........................................................... 32

7.1 Telescope Electro‐magnetic analysis .................................................................................... 32

7.1.1 Eleven feed .................................................................................................................... 32

7.1.2 Overview of Results ...................................................................................................... 33


Page 6 of 50

7.2 Feed selection mechanism .................................................................................................... 36

7.3 Feed box struts ...................................................................................................................... 36

7.4 Dish performance .................................................................................................................. 37

8 MANUFACTURING COST ESTIMATE .................................................................. 39

8.1 Manufacturing process ......................................................................................................... 39

8.2 Recurring costs ...................................................................................................................... 40

8.3 Logisitics: ............................................................................................................................... 41

9 PLANS FOR FURTHER DEVELOPMENT ............................................................... 43

9.1 Technology to be developed ................................................................................................. 44

9.1.1 Further detailing the dish material performance and coating. .................................... 44

9.1.2 Detailing the assembly of the dish ................................................................................ 44

9.1.3 Feed ............................................................................................................................... 44

9.2 Partners in the Thermoplastic Composite SKA Reflector project (TC SKAR) ........................ 45

9.2.1 Airborne www.airborne.nl ............................................................................................ 45

9.2.2 Dutch thermoplastic components www.composites.nl ............................................... 45

9.2.3 Kok en van Engelen www.kve.nl ................................................................................... 45

9.2.4 Delft University of Technology www.tudelft.nl ........................................................... 45

9.2.5 ASTRON www.astron.nl ............................................................................................... 45

9.3 Risk assessment and mitigation ............................................................................................ 46


Page 7 of 50

LIST OF FIGURES

Figure 1 Dish Array Hierarchy (adopted from SPDO) ............................................................................ 13

Figure 2 Overview symmetric reflector design ..................................................................................... 22

Figure 3 Stiffener cross‐sections. Left: T‐stiffener. Right: Blade stiffener ............................................ 23

Figure 4 Overview of different stiffeners depicting one quarter of the reflector ................................ 24

Figure 5 FEA displacement results. Left: Displacement results wind load case 2. Top right:

Displacements for thermal load case 4. Bottom right: gravity at 0 degree elevation ............... 27

Figure 6 rms and pointing error as function of elevation ..................................................................... 28

Figure 7: Dish and pedestal ................................................................................................................... 31

Figure 8: Sensitivity and system noise temperature versus elevation angle for the symmetric dish .. 33

Figure 9: Sensitivity and system noise versus elevation for a Gregorian dish ...................................... 34

Figure 10: Comparative sensitivity performance versus elevation angle between symmetric and

offset Gregorian antenna systems ............................................................................................. 34

Figure 11: Antenna far field pattern for both dish types illuminated with the Eleven antenna feed .. 35

Figure 12: Dish parameter simulation results ....................................................................................... 38

Figure 13: Overall schedule ................................................................................................................... 43

Figure 14 Gregorian offset reflector with a truss support .................................................................... 50

Figure 15 Gregorian offset reflector with a stiffened skin .................................................................... 50

LIST OF TABLES

Table 1 SKA reflector requirements by ASTRON ................................................................................... 17

Table 2 Typical CFRP material properties ............................................................................................. 25

Table 3 rms performance during typical conditions ............................................................................. 29

Table 4 rms values for more extreme conditions. * rms value including gravity re‐pointing .............. 30


Page 8 of 50

LIST OF ABBREVIATIONS

Ant. ............................... Antenna

CoDR ........................... Conceptual Design Review

DRM ............................. Design Reference Mission

EoR .............................. Epoch of Reionisation

EX ................................ Example

FLOPS ......................... Floating Point Operations Per Second

FoV .............................. Field of View

JLRAT……………………Joint Laboratory for Radio Astronomy Technology

LNA .............................. Low Noise Amplifier

Ny ................................. Nyquist

OH ................................ Over Head

OTPF ............................ Observing Time Performance Factor

Ov ................................ Over sampling

PAF .............................. Phased Array Feed

PrepSKA ...................... Preparatory Phase for the SKA

RFI ............................... Radio Frequency Interference

rms ............................... root mean square

SEFD ............................ System Equivalent Flux Density

SKA .............................. Square Kilometre Array

SKADS ......................... SKA Design Studies

SPDO ........................... SKA Program Development Office

SPF .............................. Single Pixel Feed

SSFoM ......................... Survey Speed Figure of Merit

TBD .............................. To Be Decided

Wrt ............................... With respect to


Page 9 of 50

1 Introduction

This document describes the concept for a symmetric dish for SKA. It is based on specifications for

single pixel feed and its baseline design covers wavelength ranges of 1‐10 GHz. The specifications are

founded on relevant SKA material and discussions with SPDO as alternative dish concept design

serving as input for the SPDO Dish Array CoDR in July 2011 in Penticton.

A collaborated effort between a Dutch Industrial project group, Chalmers University and ASTRON

provided input for this document. The main technical input is by the Industrial Group, lead by

Airborne in The Hague, the Netherlands. Airborne has experience in building dish structures e.g. for

ALMA. The Industrial project group in general has experience with thermo plastic composites from

design, to analysis and calculation up to production of high‐end components. Chalmers University

provided the integral radio design and electro‐magnetic calculations for the overall reflector to

which the electro‐magnetic design for the eleven feed up till now was input. ASTRON provided the

specifications and performed system‐ and mechanical design and RF testing of materials. At the end

of the year, ASTRON will test the prototype dish‐unit – manufactured by the industrial project group

– in one of their dishes for the Westerbork Synthesis Radio Telescope.

The dish design of the concept is based upon the use of thermoplastic material, reinforced with

carbon fibre. Thermoplastic carbon materials are used in airplanes for wing structures (e.g. flaps)

and in the automotive industry with relatively high plastic material costs. This concept uses a much

cheaper resin while maintain an excellent (price‐ and other)performance and makes it more suitable

for this type of product.

From our study, thermoplastic carbon re‐enforced materials appear as an attractive material for use

and implementation in low cost, high performance dish structures. Material and processing costs are

lower compared to traditional epoxy based composite materials and in respect to classical metal

structures it has better performance in stiffness, a much lower weight and a higher thermal stability.

On top of that, the material can still be processed using conventional methods as welding and

machining, which makes it suitable for automated production.

The original study activity was based on the baseline specifications for an offset Gregorian dish

design. During the project the symmetric dish was specified and the study broadened towards this

concept. For reference both dishes (symmetric and offset Gregorian) have been studied, but the

baseline for the presented design in this document is the symmetric dish.

1.1 Purpose of the document

The purpose of this document is to describe the technical and functional characteristics of a

symmetric dish concept for SKA as input for the SPDO Dish Array Concept Design Review in July

2011, based on an integral industrial concept study together with RF design aspects through radio

astronomy partners. It is felt that based on the initial results, this approach is entirely suitable for

the SKA dish approach. It is a concept study and in global terms the telescope is described. The main

part of the symmetric dish concept, is described in detail as the focus of this study is on performance

and cost of the dish structure. The other main part is the feed design, suitable for this symmetric

dish. The simulation emphasis is on higher frequency usage but do not exclude other frequency

ranges based on the dish principal approach. Other parts, like the pedestal and drive systems have


Page 10 of 50

not been put into focus. For this study the focus is on dish performance and cost which are

addressed in detail, clearly more development is required. Plans to cover this with related risks are

described at the end of the report.


Page 11 of 50

2 References

[1] SKA Science Case

[2] The Square Kilometre Array Design Reference Mission: SKA‐mid and SKA‐Lo v 0.4

[3] SKA Dish Antenna Primary Design, Ver. 1.0, Presented by NAOC, 10th March 2011

[4] Science Operations Plan

[5] System Interfaces

[6] Environmental requirements (natural and induced)

[7] SKA strategies and philosophies

[8] Risk Register

[9] Requirements Traceability

[10] Logistic Engineering Management Plan (LEMP)

[11] Risk Management Plan (RMP)

[12] Document Handling Procedure

[13] Project Dictionary

[14] Strategy to proceed to the next phase

[15] WP3 SKA array configuration report

[16] WP3 SKA site RFI environment report

[17] WP3 Troposphere measurement campaign report

[18] SKA Science‐Technology Trade‐off Process (WP2‐005.010.030‐MP‐004)

[19] SKA Monitoring and Control Strategy WP2‐005.065.000‐R‐001 Issue Draft E

[20] “The Square Kilometre Array”, Peter E. Dewdney, Peter J. Hall, Richard T. Schilizzi, and T.

Joseph L. W. Lazio, Proceedings of the IEEE Vol. 97,No. 8, August 2009

[21] System Engineering Management Plan (SEMP) WP2‐005.010.030‐MP‐001Reference 3

[22] SKA System Requirement Specification (SRS)

[23] SKA IP Policy Document

[24] Load Distribution on the Surface of Paraboloidal Reflector Antennas, M. Kron, JPL Technical

Report 32‐1526, Vol. V

[25] Compilation of Wind Tunnel Coefficients for Parabolic Reflectors, R. Levy, D. Kurtz, Reprint

from the Deep Space Network, Space Programs Summery, R. Levy, and K. Kurtz, Vol. II, pp.

36‐41, May 31, 1970

[26] Homologous Deformations of Tiltable Telescopes, S. Von Hoerner, Proc. ASCE, J. Struct. Div.,

93, ST5, October 1967

[27] US‐SKA 15m_SKA_Antenna_Design_P13_2010‐10‐20GL, Lacy and Flemming, October 2010

[28] SKA‐ASTRON‐PR‐466 Using a circular cavity to determine the surface resistivity of a

conductor

[29] SKA‐ASTRON‐PR‐465 Measurements Reflection of Composite Materials


Page 12 of 50

3 Context

In 2009 a Dutch consortium with 5 partners was set‐up to investigate the potential of a

Thermoplastic Composite SKA Reflector (TC SKAR). The five Dutch partners are Airborne, DTC, KVE,

TU Delft and ASTRON. Each of the partners has a history with processing thermoplastic material or a

background in radio astronomy. Further information about the partners is summarised at the end of

the report.

The goal of the TC‐SKAR consortium was to develop a composite reflector concept that captured the

advantages of lightweight and low CTE of composite, but using novel, more efficient and

industrialised manufacturing technologies. Traditionally, composite structures are made in

thermoset composites such as carbon fibre – epoxy. The manufacturing of such structures involve

laminating thin plies of composite material in a mould, curing for several hours in an autoclave or

oven under elevated pressure and temperature. Final assembly is typically done by bonding with an

adhesive. These processes give high quality products that can be used for example in highly loaded

aerospace structures, but are time consuming and labour intensive to produce. The ALMA antenna

reflectors – as made by Airborne – are made with these materials and processes. These 25 ALMA

telescopes of 12 m diameter are made in 5 years with a production staff of roughly 50 full‐time‐

equivalent. For SKA, many 100’s to 1000’s telescopes with even larger diameters are needed and

should be produced in approximately the same period of time. It is obvious that simply scaling up

the current state‐of‐the‐art technology is not practically feasible. Therefore, a much better and

efficient design and manufacturing concept needs to be developed that provides a radical step‐

change in production efficiency, and allows for a truly industrialised manufacturing method. The TC‐

SKAR consortium has developed a reflector concept in thermoplastic composite materials, which is a

very different class of materials compared to the traditional thermoset materials. These materials

can be formed into shape, like metals, and can be assembled by welding. The production cycle times

are much shorter, minutes compared to hours, and automated production processes like stamp‐

forming and robotic welding can be used.

As design parameters, the consortium started with the preliminary design specifications from the

SPDO office. The baseline design was founded on offset dish designs available at that time. During

the project ASTRON has taken the initiative to further detail the specifications to enable the team to

optimise the design to a realistic set of requirements for symmetric dishes. Both dish types,

symmetric and offset, are studied and reported in this document.

3.1 SKA hierarchy

The SKA Systems Engineering Management plan has defined multiple layers of hierarchy:

L7: SKA User

L6: System

L5: Element

L4: Sub‐System


Page 13 of 50

L3: Assembly

L2: Component

L1: Part

Although not explicitly stated in the SEMP, the hierarchical approach has the obvious advantage of

breaking down the complexity of the system without loss of interdependencies. Each layer is

therefore primarily concerned about its own functionality and its interface to the immediately

adjacent layers.

Within the hierarchical scheme, the Dish Array is defined at the element level deriving its

requirements directly from a subset of System level requirements. In turn, the sub‐system level

allows the Dish Array element to be partitioned further into Level 4 functionality, comprising the

Dish and Single Pixel Feed sub systems. Single Pixel Feeds are further divided into Feed Payload and

Receiver assemblies at level 3. Introducing these layers of hierarchy ensures that the complexity of

the system is broken down such that an individual layers only have to deal with their relevant

perspective of the system.

Figure 1 Dish Array Hierarchy (adopted from SPDO)

This document describes one option for the level 4 sub system, the dish itself. Besides that, the

single pixel feeds are also described in connection to the performance of the dish reflectors.

3.2 Role of the symmetric dish in the Dish Array

The symmetric dish that has been studied in depth by the team is a potential SKA dish capable of

accommodating two single pixel feeds in its baseline design. For this study the eleven feed

developed by Chalmers Technical University in collaboration with Onsala Space Observatory is used.

The design specifically addresses the functional and non‐functional requirements for the SKA; many

of these requirements are unique to the SKA and relevant for dishes for SKA in both Phase 1 and 2.


Page 14 of 50

Although in principle feasible, the present study did not investigate extensions and/or alternative

designs incorporating phased array feed and multi‐combinations of feeds for lower frequencies then

the specified minimum of 1 GHz.

The present symmetric alt‐az dish study based on the carbon re‐enforced thermo plastic materials to

be used in the dish reflecting structure is primarily motivated because of:

(i) Its promise for low cost high performance dishes

(ii) Its potential for mass production.

(iii) Its inherent simplicity and low weight/stiff design implies a relatively low risk and low level

maintainability effort

The design of a symmetric dish with equatorial mount may provide “design wiggle room” but has not

been studied given the latitude range due to the spatial extension of the proposed SKA dishes

including the SKA core sites itself. While probably hugely impacting on cost and maintenance, a

third rotation axis, as currently implemented in the ASKAP system, can be implemented relatively

straightforward if this turns out to be required for the realization of high dynamic range.

3.3 Comparative discussion of Symmetric versus Gregorian offset reflector

concept

This report focuses on the symmetric design as it has many attractive features that could well offer a

competitive advantageous concept for the SKA. In the paragraphs below a summarised description is

given pointing out important aspects relevant for further study toward a choice between antennas

and advantages of a symmetric reflector when compared to a Gregorian offset.

3.3.1 Discussion of the dish concept from an RF perspective

There are significant inherent advantages in the RF performance of offset dual reflector designs such

as a Gregorian. These have been summarized in the literature and involve higher aperture

efficiencies and on‐axis improvement of polarization performance as desirable characteristics for

telecom antennas. For radio astronomy, the use of dual reflector designs also excludes the problem

caused by the so‐called baseline ripple effect as a result of multiple reflections such as between the

(suppose: ) prime focus receiver box and the apex of the main dish which is worsened by the trend

toward wider bandwidth (as in the SKA) excludes. However, this effect is most important in dishes

operating as single, total power telescopes e.g. such as in the GBT in the US, but this is not the use

case for the SKA where differential effects are more important. The blockage by the legs and feed

support structures most importantly for symmetric dishes and other effects cause a‐symmetric

footprints on the rotating sky such as in alt‐az telescopes being studied for the SKA which is seen as

undesirable for high dynamic range observations. This effect is largely absent in symmetric dishes

employing a third‐axis rotation of the whole reflector/feed structure such as in the ASKAP antennas.

On the other hand, offset dishes inherently have high a‐symmetric radio performance behaviour

outside the mainbeam, spillover (noise) effects vs elevation, possible effects of structural design

aspects caused by its dual reflector surface including its surface and have a larger cross polarization

from boresight. The effects of surface imperfections (apart from rms‐surface errors) are e.g. caused


Page 15 of 50

by the inter‐panel spacing holds also for symmetric dishes but its effect translates into an axi‐

symmetric pattern. Note also that the a‐symmetric weight distribution over elevation and its

relatively large effect of sensitivity performance over elevation could largely offsets the intrinsically

higher aperture efficiency of offset Gregorian reflectors (see chapter 7).

The SKA poses constraints involving all these aspects but at the same time call for excessive

simplicity vis a vis low maintainability and low cost while offering stable and high radio performance

over frequency and pointing position. For this the symmetric dishes could become a serious

candidate dish design solicitor.

3.3.2 Discussion from scientific / electromagnetic perspective

Some key elements pertaining to the scientific performance of dishes have been addressed before.

Aspects like the blocking caused by the support struts and other a‐symmetric and symmetric effects

(see e.g. the paragraph above), adversely affect the imaging performance of the array. Details are

domain of active modelling and lead to the 3‐axis sky de‐rotating design exercised in ASKAP.

For the symmetric dish design, it is important to incorporate design rules that are known to negate

adverse effect to the image quality e.g. by choosing a strut design with smallest optical shadowing to

ensure smallest diffraction and other effects as is done in our design approach.

Other more quantitative aspects of the e.m. and RF design are covered in chapter 7. As we were

keen to do an overall Radio performance assessment, we choose a symmetric dish with an F/D

(=focal distance to dish diameter) ratio of 0.42 within a range of optimal (although smooth)

performance of the feed allowing for optimized sensitivity performance for low noise receiver

systems. The structural design therefore is based on this F/D ratio although not critically depended

on it.

3.3.3 Structural design

The major axis of the ellipse of the Gregorian offset reflector is larger than the radius of a

symmetric reflector. The bending stresses and deformations of a beam (e.g. stiffener of

reflector) are proportional to its length squared. Therefore, the required structural mass of a

Gregorian offset will increase significantly with respect to a symmetric design and this can be

up to twice the weight compared with symmetric design.

The loading on a Gregorian offset reflector is inherently asymmetrical resulting in further

weight penalties.

The extra weight of the sub reflector adds to the gravity load case, so an even stiffer rib

design is required.

All the points mentioned above have a direct effect on the mass of the reflector. The

additional mass however also has a secondary effect resulting in higher material

requirements.


Page 16 of 50

Besides the penalties on the design itself, it is much more difficult to design and optimize a

Gregorian offset reflector.

3.3.4 Manufacturing cost

Studies for a Gregorian dish show the skin is significantly (about 25%) larger with additional

30‐50% extra material in the frame structure. Composite material use is a large factor in the

overall costs, so it is important to keep weight down.

All the required reflector parts are different, what makes logistics and assembly of the parts

much more difficult and brings additional costs.

All the required reflector parts are different, which in our concept will lead to more

specialised tooling and moulds. This will add to the non‐recurring costs.

Quality control efforts are more demanding as more different parts and geometries need to

be controlled

Parts are less interchangeable

Repeatability of the assembly process is limited, so more deviations can occur in

manufacturing accuracy

Weld lines of the ribs do not coincide with the panel geometry, which results in double the

welding efforts

The dish is larger and heavier, which makes it more difficult to move to the construction site

From our initial study in the TC SKAR project the points mentioned above result in recurring

costs are 60‐70% higher than a symmetric design


Page 17 of 50

4 Requirements

This section of the document will describe how the proposed sub system will address the

requirements for the Dish Array, which are derived from the system requirements [21] and

ultimately the science requirements. These include both the functional and non‐functional

requirements. The table below summarises the requirements used in this study.

Table 1 SKA reflector requirements by ASTRON


Page 18 of 50

Key elements in the dish design relate to:

• Imaging dynamic range • Mass manufacture • Operating cost • Feed flexibility • Rapid installation • Maximising A/T per unit system cost (i.e. including signal transport, signal processing,

computing etc.) • Minimising maintenance cost • Electromagnetic compatibility

Not all are addressed in detail in this report but the design approach and its general results warrants further detailed study.

4.1 Functional Requirements – reflector structure

From the general specification from ASTRON a choice has been made on specifications that

represent a worst‐case load or most challenging design for the reflector dish structure. Following

specifications have been taken into the design of the reflector dish:

Top level requirements

Diameter 15 meter

Focal Ratio 0.42

Frequency range 1,2 tot 10 GHz

Accuracy overall 1mm RMS

Product lifetime minimal 30 years

Operational requirements

Elevation range from 15 to 91 degrees

Wind speed 12 m / sec

Ambient Temperature from 1 to 40 degrees Celcius

Solar irradiation 980 W / m2

Humidity max 100%

Product design aspects

Stow wind speed max 18 m / sec

Survival wind speed max 45 m / sec

Maintenance interval around 5 years

Lightning protection on construction

Aspect requirements

Feed Weight is max 170 kg

Feed Mount type 4 legs attached to dish edge


Page 19 of 50

4.2 Specification compliance – reflector structure

Top level specifications

The design concept of the thermoplastic reflector allows for variations in the diameter, focal ratio

and frequency without any complications. By optimising the design for a given set of parameters

(within the boundaries of the ASTRON spec), the RMS can be easily met.

The manufacturing accuracy and assembly will contribute to the final RMS. This effect is currently

being researched by making large skin sections for an existing ASTRON reflector at Westerbork and

testing them amongst others on reflective performance.

The required lifetime is dependent on the choice of base material and coating. The performance of

the composite material is now being tested by the Technical University of Delft. The reflector will be

finished with a durable coating. With coatings there are years of experience in protecting composite

components in aerospace and windmill applications.

Operational specifications

All these specifications are taken into account in the RMS calculation. The given values represent the

worst‐case load (or summation of loads) on the reflector dish.

The temperature and wind load cases in normal operation will be lower than the extreme load cases

shown. Performances refer to nominal operational condition.

Because of the fully integrated and welded design of the reflector, the maintenance interval can be 5

years. The quality of the coating is herein a critical component that needs periodic inspection.

Product design aspects

In the design of the reflector, the stiffness is the leading parameter for the geometry of the

structure. The strength of the dish therefore will not be an issue for specified loads. Resonance

analyses of the surfaces under extreme load conditions still need to be performed.

Lightning protection can be added by ground wiring all reflective meshes in the panels (if required)

and putting a ground wire on the feed, which is the highest point in stormy conditions. Even when a

panel has been damaged due to lightning strike (or other cause), there is a possibility to replace it

with a new panel or to repair it as the thermoplastic material can be welded again.

The different skin panels are not electrically connected to each other, but research from ASTRON in

the TC SKAR project in 2011 shows that this does not significantly influence the performance of the

reflector.

Aspect requirements

The maximum feed weight and worst‐case location of the feed support (edge of skin) have now been

taken into account. With these requirements, the design performs within specification.

Nevertheless, the feed weight significantly influences the RMS values. It is expected that for the SKA


Page 20 of 50

programme, specifically designed and miniaturised equipment for the feed will be available,

reducing the feed weight, and hence further improving the performance of the dish. For illustration

purpose the present weight of APERTIF is less than 60kg, with miniature coolers will probably not

exceed similar weight.


Page 21 of 50

5 Main reflector design

This chapter will elaborate on the symmetrical reflector design that has been made within the

TC SKAR project.

5.1 Material selection

Composite is increasingly used for all sorts of components in different markets. The material is

lightweight and has high performance properties, which often outperform metal equivalents. A large

integrated structure as a reflector dish embeds several features where the choice for composite

materials is beneficial. Some features are discussed below:

Weight does not only influence the required strength & stiffness of the reflector frame construction,

it also influences the requirements on the pedestal and its actuator system. It is beneficial for

transport of the reflector, material costs, pedestal design, and required actuators that the weight of

the reflector itself is kept to a minimum. With a composite solution, weight can typically be reduced

with 30% compared to a steel / aluminium option.

Reflectivity of the dish surface is very important for the total performance of the reflector.

Composite material in itself is not a good electrical conductor, but the material has the flexibility to

embed a thin metal mesh with little effort. This gives the reflective performance without adding

much weight (2.5 % of the total reflector weight) and doubles as a lightning protection.

RMS performance of the design needs to be consistent in various load cases. On gravity and wind

load cases, the carbon composite material contributes to a higher stiffness and lower weight. Both

are beneficial to the performance. Additionally, the thermal coefficient of Carbon based composite is

lower than metal solution, which also favours the choice of composite material. Another advantage

of composites over metal is the better fatigue properties. Especially aluminium is very susceptible

for fatigue, which can reduce the lifetime of the telescope. The use of composites will decrease the

fatigue issues and therefore can increase the lifetime of the telescope.

Atmospheric influences (weather, UV, heat etc) have an impact on all materials; also resins within

composite materials can be sensitive to specific items. Therefore, a robust coating is applied to

protect the material from the atmospheric influences (especially UV). The coating will ensure a long

lifetime of the product.

The choice for thermoplastic composite:

Thermoplastic composite is a material, which becomes more and more available with high

performance properties. The material is tougher, more ductile and robust compared to metal

options. Certainly in combination with carbon fibres these material outperform any aluminium or

steel constructions. The material can be shaped with fibre placement, thermoforming; welding and

conventional cutting operations and all processes have potential to be (partly) automated.

The TC SKAR consortium aims to use an efficient, industrialised production process, which makes use

of a selection of these automated production processes. With this concept the aim is to make a

“step change” in production efficiency, as the thermoplastic material (light weight reflector

construction) will be used in combination with the new automated production technologies.


Page 22 of 50

In comparison with thermoset composite material, the energy used when processing thermoplastic

material is much lower; this leaves a lower environmental footprint. The thermoplastic material is

also suitable for recycling and enables repair of damages.

5.2 Structural design

The baseline design of the reflector uses a stiffened skin with several different stiffeners, see Figure

4 Error! Reference source not found.. The entire structure will be build from the same raw material ‐

a thermoplastic carbon based composite ‐and is manufactured using a single automated production

processes, or a variation on it.

Figure 2 Overview symmetric reflector design

All the stiffeners will be build up from relative simple parts which are welded together, see Figure 3.

The T‐stiffener consists of a web, a girder, two corner profiles connecting the girder to the web, and

two corner profiles to connect the stiffener to the skin. The blade stiffener consists of only a web

and two corner profiles to connect it to the skin.

A T‐stiffener is more efficient (higher bending stiffness per mass) when compared to the blade‐

stiffener, it is however also more difficult to produce. Therefore, the T‐stiffener is selected for the

more critical and higher loaded regions, whereas the blade‐stiffener is used in regions with lower

requirements.


Page 23 of 50

Figure 3 Stiffener cross‐sections. Left: T‐stiffener. Right: Blade stiffener

The stiffeners are also dimensioned differently so all the material is used as effectively as possible.

Below the different type of stiffeners are listed, starting in the centre moving out radial, see Figure 4.

Tangential stiffeners:

Ring 1: non‐critical blade stiffener

Centre: Critical blade stiffener, reflector is mounted on this ring to the pedestal

Mid‐ring: low, non‐critical blade stiffener in the centre

Outer ring: Semi‐critical blade stiffener

Radial stiffeners:

Radial mid‐ring: low, non‐critical blade stiffener in the centre

Quadra pod stiffeners section 1: Critical and relative heavy T‐stiffener

Non‐quadra pod stiffener section 1: semi‐critical T‐stiffener

Quadra pod stiffeners section 2: Critical and relative heavy T‐stiffener

Non‐quadra pod stiffener section 2: semi‐critical blade‐stiffener


Page 24 of 50

Figure 4 Overview of different stiffeners depicting one quarter of the reflector

The selected material and geometry ensure a stiff structure with a low sensitivity to temperature

changes. This results in a total weight of the integrated reflector structure of only 1580kg.

5.3 Structural analysis

The current design is analyzed using a Finite Element Analysis (FEA) with MSC Marc & Mentat 2010.

First the different load cases are discussed followed by the FEA results. The final subsection

describes the rms performance of the current design for different (combinations of) load cases.

5.3.1 Material properties

As mentioned previously, the reflector will be build from a carbon fibre based composite material

(Carbon Fibre Reinforced Plastic CFRP). The typical properties of a carbon based composite material

are tabulated below in Table 2.

Concurrently with the preliminary design phase, the Technical University of Delft has been

characterising the mechanical properties of the proposed material. So far the standard material


Page 25 of 50

properties at room temperature, and at an elevated temperature, have been tested. The results of

these tests have been used as input for the structural analysis.

Property Value [Unit]

Modulus of elasticity of uni directional CFRP (E11) 90 [GPa]

Modulus of elasticity of 0/90 ply CFRP 45 [GPa]

Modulus of elasticity of Quasi Isotropic CFRP 32 [GPa]

Tensile / compressive strength 0/90 ply CFRP 700 [MPa]

CTE carbon fibres ‐3.6 ∙ 10‐7 [1/°C]

CTE matrix material 3.0 – 7.0 ∙ 10‐5 [1/°C]

CTE CFRP 3.0 – 7.0 ∙ 10‐6 [1/°C]

Material density 1550 [kg/m3]

Table 2 Typical CFRP material properties

The CTE properties of the composite material are mostly dependent on the CTE properties of the

carbon fibres. The fibre properties are already known and therefore the expected CTE properties of

the composite material can be estimated relatively accurate. The CTE values used for the structural

analysis are based on the experience from previous projects. A worst case CTE value has been

calculated and used in an extra structural analysis to mitigate the risk associated with an incorrect

assumption for the CTE value. The resulting performance of the reflector is still admissible.

Creep for the pure thermoplastic material is unknown and could be an issue, but since it is combined

with carbon‐fibers (which have very little creep) the effect is minimized. Currently the CF/TP

material is being tested at the TU Delft for creep properties, both at room‐temperature and elevated

temperature. Creep is higher for increased temperatures, so it is necessary to pay extra attention to

the properties at increased temperatures at which the telescopes will function. Relatively low loads

in the structure – as the design is optimised for stiffness – help the material against creep

occurrence.

The material test programme is ongoing at the Technical University of Delft and soon results for

creep, creep at elevated temperature, CTE, and influence of environmental conditions are expected.

Additional design iteration is required to investigate the influence of the tested values.

5.3.2 Load cases

5.3.2.1 Gravity

The gravity is modelled with different elevations. Using these results it is possible to calculate rms as

function of elevation.


Page 26 of 50

5.3.2.2 Thermal loading

From calculations with the given weather parameters and measurements done on painted reflector

surfaces, several thermal load cases are modelled to investigate the influence on the performance of

the reflector. In a typical situation, with an ambient temperature of 30 [°C], a solar irradiation of 750

[Watt/m2], and a wind speed of 4 [m/s], it is expected that the skin will reach a temperature of

approximately 33 [°C] while the backup structure reaches 37 [°C].

A temperature of 58 degrees is calculated as the worst case when no convection or conduction is

taken into account and solar radiation is at the maximum with 100% humidity. An additional worst

case load case of 70 degrees is added to represent built up of dust etc. on the surface.

It is difficult to predict the temperature distribution within the reflector structure due to solar

radiation, and during warming up and cooling down. Therefore, a conservative temperature

difference between the reflector skin and backup structure of approximately 20°C is modelled. It

should be noted that such a high temperature difference is not expected and unlikely to occur.

The following thermal load cases have been modelled:

Thermal load case 1: Skin temperature of 33 [°C] and backup structure of 37 [°C]

Thermal load case 2: Constant temperature of 58 [°C]

Thermal load case 3: Constant temperature of 70 [°C]



5.3.2.3 Wind loading

At this stage there are no accurate aerodynamic loads for the current design available. The

preliminary wind loading assumed a constant pressure on the reflector. The next wind load cases are

based on an aerodynamic data as measured by Kron [24]. These measurements were however

performed on a model with an f/D ratio of 0.33 instead of 0.42 which is used for the current design.

However, from earlier work of Levy and Kurtz (1970) [25] it is known that the loading on a reflector

with a lower f/D ratio is actually higher, and therefore this load case is conservative. The

aerodynamic pressures were measured for several elevation angles. From the work of Levy and

Kurtz (1970) it is known that the worst case axial force and pitching moment combinations can be

expected for an elevation of 60° and 120°. Therefore, the following wind load cases were modelled:

Wind load case 1: An elevation of 60 [°] [24], wind speed of 4 [m/s]

Wind load case 2: An elevation of 120 [°] [24] , wind speed of 4 [m/s]

Wind load case 3: An elevation of 60 [°] [24], wind speed of 12 [m/s]

Wind load case 4: An elevation of 120 [°] [24] , wind speed of 12 [m/s]


Page 27 of 50

5.3.3 FEA results

Some typical FEA displacement results are depicted in Figure 5. The replacement results of the

reflector surface are subsequently used to determine the rms performance of the reflector.

The design is also checked for buckling to ensure the stability of the structure. At a worst case

loading situation, e.g. wind load case 2 + thermal load case 4 + gravity at an elevation of 60 degrees,

the structure still has a reserve factor of 3.0 with respect to buckling.

Figure 5 FEA displacement results. Left: Displacement results wind load case 2. Top right: Displacements for

thermal load case 4. Bottom right: gravity at 0 degree elevation

5.3.4 Performance

To evaluate the performance of the current design the deformation of the reflector is analyzed for

different (combinations of) load cases. With these deformations, the root‐mean‐square half‐

pathlength errors (rms) have been analyzed. At this stage, the displacement of the feed due to

external loading is not considered.

Based on the deformation a homologic design is considered to better understand the influence of

the deformations on the sensitivity of the antenna. First of all, the pointing error induced by the

gravitational deformations is considered. The rms value caused by gravity, adjusted for gravity re‐


Page 28 of 50

pointing, is plotted for the required elevation range in Figure 6. Also, the required re‐pointing is

depicted on the secondary axis.

The maximum rms caused by gravity is 0.49 [mm] and occurs when the elevation is 90 [degrees].

From the analysis it is apparent that the gravitational load of the feed and its support is the most

important contributor to the deformation. It is expected that this contribution can be reduced

significantly with more design iterations, without increasing the total structural mass.

Figure 6 rms and pointing error as function of elevation

Secondly, the influence of the axial and lateral defocus are considered by determining the rms

performance with respect to the best fit parabolic surface, based on the six homology parameters

described by von Hoerner in [26]. The remaining rms value can be considered as real random

deformations. The effects of axial or lateral defocus are less severe when compared to random

deformations; they however still influence the performance. The total effect is hence comparable

with an rms value somewhere between the best fitted value, and the value including gravity re‐

pointing.

Until here only the deviations from the perfect reflector surface due to external loading have been

considered. The requirements however dictate the performance of the reflector as built, implying

that the manufacturing and assembly imperfections have to be considered as well. At this stage the

attainable manufacturing and assembly accuracy is not known. The assembly process is however

designed such that the imperfections incurred by the manufacturing process do not accumulate.

Therefore, and from experience with similar products, it is expected that an rms value due to

manufacturing and assembly of 0.4 [mm] is realistic. To be conservative a value of 0.5 [mm] is

assumed. This assumption will be validated by manufacturing and measuring several reflector panels

for the Westerbork Synthesis Radio Telescope (WSRT).


Page 29 of 50

The rms performance for different typical load cases are tabulated in Table 3. The total typical values

are calculated by combining the displacement results of the different load cases and then calculating

the rms value. The total typical value including the manufacturing and assembly imperfections is

calculated as the root of the sum of squares:

Description rms original surface

rms incl. gravity repointing

rms fitted surface

Gravity 15 degrees elevation

0.71 0.25 0.25


0.47 0.37 0.34


0.49 0.49 0.44

Typical thermal (thermal load case 1)

0.05 ‐ 0.04

Typical wind (4 [m/s]), 60 degrees elevation. (wind load case 1)

0.09 ‐ 0.06

Typical wind (4 [m/s]), 120 degrees elevation. (wind load case 2)

0.07 ‐ 0.02

Total typical (thermal + wind + gravity, 60 degrees elevation)

0.53 0.44 0.41

Manufacturing and assembly accuracy

0.5 0.5 0.5

Total typical incl. manufacturing and assembly

0.73 0.67 0.65

Table 3 rms performance during typical conditions


Page 30 of 50

From the results presented in Table 3 it can be seen that the maximum expected rms value

(including gravity re‐pointing) is 0.67 [mm] and therefore it can be concluded that during typical

conditions the current design will meet the requirements easily.

Next, the performance of the current design during more extreme conditions has been analyzed. The

results in Table 4 show that the current design will meet the requirement of an rms lower than 1

[mm] even in unrealistically extreme combination of worst case load cases.

Description rms original surface rms fitted surface

Thermal load case 2 0.38 0.07




Wind 12 [m/s], 60 degrees elevation. (wind load case 3)

0.78 0.55

Wind 12 [m/s], 120 degrees elevation. (wind load case 4)

0.65 0.17

Total extreme conditions (thermal load case 4 + gravity + wind 12 [m/s] (wind load case 3), elevation of 60 [degrees])

0.79 0.73

Total extreme conditions typical incl. manufacturing and

assembly

0.93 0.88

Table 4 rms values for more extreme conditions. * rms value including gravity re‐pointing


Page 31 of 50

6 Antenna mount design

There are in principal two mount concepts: equatorial mount and the altazimuth mount. The main

reason for choosing the altazimuth mount for SKA dish telescopes is that that the telescope location

varies, which means that in case of and equatorial mount each telescope needs a different polar axis

angle depending on its latitude location. The disadvantage of an altazimuth mount is that it cannot

track the sky without introducing a field rotation. This could be solved by adding a third axis which

counter rotates the whole dish including the feed system. In this stage the choice has been made not

to add the third rotating axis since it is not clear whether this is necessary. The third axis can be

added at a later stage when necessary.

6.1 Pedestal design

The pedestal has got the following functions:

Creating vertical distance such that the dish can track the sky without impacting the ground

Strong and stiff support to accommodate the dish

Support the azimuth and altitude axis and drive system

Housing for drive and signal electronics

The pedestal has got a cone shape to minimize material use and maximize stability also it ensures

that the azimuth and alt axis are as close as possible to the dish feed system centre of mass.

The pedestal will interface with the 3m diameter supporting ring of the dish.

Figure 7: Dish and pedestal

6.2 Balancing

The mass on the alt axis (dish plus feed system) needs to be balanced so that the load on the

bearings and drive system is minimised and play/flexure in the drive system does not give a sudden

movement when the dish is moving through zenith.


Page 32 of 50

6.3 Drive system

The drive system for both axes are based on a commercial available components baseline concept is

based on an electric motor with for instance a worm‐gear drive. Same approach applies for the

control unit will also be based on standard components. The type of control soft‐ and hardware is to

be defined, but should be relative robust and stable like plc based systems or any other industrial

system suitable for the requirements of the telescope.

7 Telescope Electro‐magnetic design

This chapter offers quantitative aspects of the electromagnetic . and RF design primarily done by

Chalmers Technical University (Chalmers) in collaboration with Onsala Space Observatory (OSO). As

we wished to do an overall Radio performance assessment, we choose a symmetric dish with an F/D

(=focal distance to dish diameter) ratio of 0.42 within a range of optimal (although smooth)

performance of the feed allowing for optimized sensitivity performance for low noise receiver

systems. The structural design therefore is based on this F/D ratio although not critically depended

on it. For illustration purposes some performance results are compared with the results of a

representative offset Gregorian design all using the same so called eleven feed from Chalmers.

7.1 Telescope Electro‐magnetic analysis

This chapter briefly describes some initial results of the end‐to‐end telescope‐feed performance.

7.1.1 Eleven feed

The so‐called Eleven antenna feed used in this study results from years of intensive development

work at the Chalmers antenna group lead by Prof. Per‐Simon Kildal. The electromagnetic

performance of this multi‐octave feed is most suitable for antennas with F/D ratios around 0.35‐0.5,

and the simulations presented here were done by OSO/Chalmers.

Our simulations consider different receiver noise levels which realistically can be seen as maximum

and minimum values while assuming a fixed F/D of 0.4 i.e essentially the same as the symmetric

mechanical design of this study. Results are shown depicting the performance of the “sensitivity”

(being defined as the effective aperture area to system noise ratio) against e.g. elevation angle.


Page 33 of 50

7.1.2 Overview of Results

Figure 8: Sensitivity and system noise temperature versus elevation angle for the symmetric dish

Shown are the system noise and the sensitivity as defined earlier as a function of the elevation angle

for the Eleven antenna feed in combination with a symmetric prime focus dish studied here at a spot

frequency of 5.6GHz. As in the curves below, the receiver noise is a parameter in the plots and

realistically varies between 10‐20K (for very high performance low noise cooled receivers) and 20‐

30K for less sensitive uncooled systems .


Page 34 of 50

Figure 9: Sensitivity and system noise versus elevation for a Gregorian dish

The plots show similar graphs as before now pertaining to a representative offset Gregorian design.

Not unexpectedly the graphs show strong elevation dependence of the system noise temperature

and sensitivity..

Figure 10: Comparative sensitivity performance versus elevation angle between symmetric and offset

Gregorian antenna systems


Page 35 of 50

The graphs show sensitivity plots versus elevation for both designs for different receiver noise

contributions. Although only preliminary results are shown, the results for both dish types are

roughly comparable showing improved relative performance for symmetric dishes for low and high

elevation angles over approximately 90 degrees. The consequences of this may be relevant to the

science to be done e.g. performance for all sky surveys will be approximately equal for both designs.

Inserted in the diagram are two plots of the aperture efficiencies versus frequency of the dual offset

and the symmetric dish both illuminated with a wide band (i.e. eleven‐) feed. As predicted, while

the dual reflector system shows a much 10‐15% higher efficiency, the overall sensitivity is not that

different; see the discussion above.

Figure 11: Antenna far field pattern for both dish types illuminated with the Eleven antenna feed

Note: the inherent a‐symmetry of the offset Gregorian antenna pattern that remains for all elevation

angles. The sensitivity effects of the earlier plots originate from these a‐symmetries weighted by the

warm earth temperature.

The diagram illustrates the relatively large a‐symmetry of the offset dual reflector as compared to

the symmetric design. The simulations have been done over all elevation angles and frequencies

between 2 and 16 GHz and essentially show the same behaviour over the elevation angles. Aspects

like these are important in a final dish assessment from the perspective of high dynamic range

imaging.


Page 36 of 50

7.2 Feed selection mechanism

A feed selection mechanism will offer the possibility to switch two different types of feeds.

Depending on the appearance of both feeds a mechanism can be designed aiming for simplicity,

reliability and low cost. The mechanism will be a distance controlled system allowing switching feeds

instantaneous. Different concepts for switching mechanisms are available which can rotate or turn‐

over another feed in focus. Since only between two feeds need to be switched, the mechanism can

be rather simple and lightweight, unlike the revolver unit used by the WSRT which contains receivers

for eight frequency bands. ASTRON also has experience with lightweight tumble mechanisms which

allows switching quickly to another receiver.

7.3 Feed box struts

The feed box struts will be made out of off‐the‐shelf carbon fibre composite tubes. The struts will

form a basic truss structure. These structures will have a length of around 8.5 meters to reach the

focal point of the telescope. Carbon‐fibre is needed to reach the stiffness and the low CTE such that

the requirements can be fulfilled. Any struts will cause a disturbance in the signal; therefore it is

necessary to make the frontal area as small as possible. The weight estimation for the four feed box

struts is 90 kg in total.

The carbon‐fibre composite tubes can also be made in an oval shape, which will decrease the frontal

area of the struts.


Page 37 of 50

7.4 Dish performance

Within the TC‐SKAR project various composites were investigated. During the selection for dish material at first the dielectric properties were measured. The measured relative dielectric constant of the base material is approximately 4.2. The loss tangent is around 0.01 for frequencies less than 4 GHz. The measurement is not influenced significantly by the orientation of the sample. At higher frequencies the measurement is less reproducible. The dielectric properties were measured since the reflective layer within the composite (metal mesh) is embedded and not entirely at the surface. The document available on these measurements is SKA‐ASTRON‐RP‐463. To determine the surface resistivity of the composite materials a circular cavity made of aluminium is used. With this cavity a reference measurement can be done. The surface resistivity is calculated from the measured resonance frequency, quality factor and loss. The details of the measurement method are described in ASTRON‐PR‐466. To first order, the surface resistivity can be used to calculate the contribution to the noise temperature if the material is used as reflector for a parabolic dish antenna. This contribution Tn can be calculated by

phen TT2

1 (1)

with Γe the effective reflection coefficient of the reflector and Tph the physical temperature of the material, The effective reflection coefficient Γe is calculated by

0

41

s

e

R

(2) with Rs the surface resistivity and η0 the wave impedance of vacuum (approximated by 120π). To check if a material is leaky each sample is measured two times. One measurement is performed with the sample on an RF absorber. A second measurement is performed with the sample on an aluminium plate. If the sample is not transparent both measurements must give the same result. If a material is leaky the effect of the absorber or aluminium plate influences the measurement. The two materials with the lowest noise temperature contribution may contribute between 1‐1.5 K. In theory the measurements are independent of the orientation of the sample because the field distribution of the mode that is measured is independent of azimuthal direction. To check this, rough measurements were done with all the samples. Four samples showed a dependency on the orientation. The measured S21 at resonance shows a difference of about 5 dB if these samples are rotated 90 degrees. These samples were investigated further. This variation does not have a significant effect on calculated noise temperature contribution. More details on the resistivity can be found in SKA‐ASTRON‐RP‐465 Since the dish will be assembled out of multiple panels the influence of gaps between the panels

were studied. It turns out that for small gaps the effective area is linear to panel area thus amount of

gaps. For large gaps other parameters have influence and the rule of thumb doesn’t apply anymore.


Page 38 of 50

This requires further investigation. At this point only simulations up to 5 GHz were done due to

limitations of the simulator. Panel size and gap size are currently under investigation. Simulations at

this point show gaps can be expected to be less than 10mm.

Dish parameters simulation

Dish: diameter: 10 m, F/D‐ratio: 0.35, radii of

circular gaps: 1, 2 and 4 m

Feed: simple tapered pattern (exponential taper,

taper ‐20 dB @ 55 degrees (edge of the dish),

Xpol: ‐600 dB, linear polarization

Tground=300 K, Tsky=0 K

Figure 12: Dish parameter simulation results


Page 39 of 50

8 Manufacturing cost estimate

This chapter describes the manufacturing process and all its supporting efforts which combine in a

total recurring cost for the complete reflector structure. Additionally the costs for the feed struts are

described and a generic approach for packaging, transportation and integration on site is given.

8.1 Manufacturing process

Skin sections

The composite reflector skin is build up with 5 types (different shape) of skin panels with an average

area of about 2 m2 (around 90 panels), which are at different distances from the centre. Each type

has multiple panels per reflector, which enables high volume production. The panels are made with

a press forming technique, which gives them the desired curvature and which integrates the

reflective mesh in one production step. Each panel is then trimmed to final size with an automated

process, ready to be assembled.

Centre ring

This composite ring structure supports the base of the ribs and connects with a metal frame to the

pedestal. Within this centre ring there is a small rib & ring structure (ring 1) purely to support the

middle section of the skin.

Ribs

The integrated frame of the reflector is made with rib structures. These ribs are tapered (high at the

Inner hub connection and low at the outer rim) I‐shaped beams. The rib structure is made with press

forming the required material in a close to size panel, which is then trimmed to final size with an

automated cutting process. This final shape follows the required skin curvature of the reflector, so it

can be directly mounted on the back of the skin structure. The girder on T‐stiffener ribs is attached

with an automated welding process before attaching the rib to the skin. The blade stiffener ribs

don’t have a girder. As with the skin panels there are multiple identical ribs per reflector and final

mounting is done by automated welding as described in the assembly paragraph.

Mid & outer ring

The mid & outer rings give extra strength to the whole structure and are made from a strip of curved

material. Sections of this material are welded around the assembly of skin and rib structures.

Assembly

Reflector assembly starts with automated lay‐up of all the skin panels on an assembly mould. First

the skin panels in the middle section are welded together creating a 3 meter dish followed with

attaching the inner hub and the small rib & ring structure. Then for each of the rotational identical

sections (on the outside of the 3 meter diameter inner ring) the skins and ribs are automatically

placed and welded by a robot and the middle ring is placed to further stiffen this structure. The


Page 40 of 50

assembly mould can rotate for easy access to all ‘pie’ sections. In the final assembly step the outer

ring is attached followed by coating all surfaces.

The dimension of skin panels are chosen in such a way that the connection between the panels in

the radial direction is always directly under a rib. This decreases the amount of weld‐meters (since

the rib also has to be welded to the skin). The same holds for the connection between the panels at

the middle circumferential rib. The panels in the centre of the telescope (rib & ring structure, ring 1)

are not part of the ‘pie’ sections, because of the required width‐length ratio for the wavelength.

8.2 Recurring costs

Reflector

The recurring costs for manufacturing a reflector dish include all the following activities:

Organising management, manufacturing, quality assurance, supporting and enabling processes:

Project management

Quality management (QA)

Engineering support

Work preparation

Facility planner

Production team leader Manufacturing:

Material cutting Mesh cutting Blanks preparation Press forming Trimming Sub assembly & final assembly (welding)

Coating

Materials:

Carbon composite

Metal mesh

Coating

Other materials (grounding wire etc)

NDI and measurements:

Laser tracker measurements incl. reporting Sample verification QA documentation

The total price for a fully finished & integrated reflector structure, which bolts on to a pedestal

(excluding feed and feed support) is now calculated to be €106.000 or €599 euro per m2, with a

potential for reduction up to about 20% as explained in the following paragraphs.


Page 41 of 50

Further optimising the design when more details are known on structural requirements can lead to a

reduction in required stiffener material. Additionally a feedbox and feed struts weight reduction will

further reduce the use of composite material. Combined this can lead up to 3‐5% reduction in costs.

Further optimising the integrated production process by tuning machine capacities, running at more

optimised settings etc. will have a potential to reduce the price with another 3‐8%

The current set of production processes are well‐known and low risk. There are some new

developments in the automation and processing of thermoplastic material with potential to be

faster and cheaper to run. These processes will become available over the next 2‐3 years. When

these new more advanced processes are taken into account, a potential reduction of the recurring

costs can occur with an additional 5‐10%.

Materials price within this cost estimate is based on 2011 prices for normal volumes, so the material

prices in our calculations are conservative as with large volumes better prices can be negotiated.

These materials are not yet very common in large volumes. Suppliers cannot yet give price

estimations for 4‐5 years ahead.

Feed support structure

The recurring costs for manufacturing the feed support structure include the following activities:

(project management costs are within reflector costs)

Manufacturing:

Material cutting

Assembly

Coating

Materials:

Carbon composite pre formed tubes

Coating

Injection moulded connector parts

Measurements:

Sample verification

The price for a complete feed support structure is calculated to be €10.000.

Pedestal

The pedestal cost is not being detailed out in this study and will probably be similar to other

concepts

8.3 Logisitics:

On logistics the design of the telescope is rather flexible.


Page 42 of 50

For the pedestal and its internal structure we aim for construction and production any place

convenient in relation to cost of production and cost of shipment to site. The unit might be split in 2

or 3 pieces for shipment and easy assembly. Eventual a shipment as full assembly might be possible

in horizontal position of the pedestal, but this is depending on cost of shipment, assembly and

technical risk and not the baseline. Eventual storage might be in open air, with some protection for

the drive head. The founding for the pedestal we see as infrastructure.

Production of the dish surface parts, the ribs and the struts can be any place convenient in relation

to cost of production and cost of shipment to site. The dish surface parts are produced with

moulding machines and can be semi automatic. Up to now we foresee the dish surface parts have

size <2.5 meter, so suitable for transportation and storage within containers. The ribs and struts will

be rather long, but slim and also these parts are expected to be transportable and stored pretty easy

in standard containers.

Assembly of the dish will be close to the site. The assembly is rather quick with a mould as reference.

No further special equipment or high energy amounts are required for this assembly, besides lifting

tools and transportation trucks for the full assembly dish. We are studying the possibility to

assemble on site, right near the pedestal, with only the need of special transportation of the

assembly mould.

The feed unit needs are typical high end units to be produced any place convenient in relation to

production cost, quality and transportation issues. Multiple feed units are expected to fit in standard

containers for transportation and storage.


Page 43 of 50

9 Plans for Further Development

The consortium presenting this symmetric dish study focussed onto the main issues within the

concept design: dish, feed and design. Depending on the outcome of the CoDR, this consortium aims

to continue this project towards a larger consortium. It is open for any participation of industrial or

non‐industrial international partners. Several contacts are ongoing e.g. for motor drive and control,

pedestal construction, assembly of the dishes, construction of the telescope and others. By default

the potential partners are European partners, but the consortium is not limited to this geographical

boundary. Typical such project will be let by an Industrial partner, with prime input, support and

reviewing by the astronomical institutes like ASTRON on technical and astronomical issues.

Financing of an eventual succeeding study will most likely be, as the running TC‐SKAR project, a

combination of national or international subsidies and industrial financed research of the partners.

Funding of a full scale telescope demonstrator is to be negotiated to be part of the succeeding study

project or financed on other ways.

Below an overall schedule is presented.

Figure 13: Overall schedule

The required manpower is not detailed yet, but expected to be rather low, because the telescope is

a classical design and knowledge is available through the astronomical institutes operating such

telescopes. The main work packages are again the dish design, the construction and the feed design

and electronics but other work packages will be started on the feed production, material disposal,

motors and drives, pedestal and foundation.


Page 44 of 50

9.1 Technology to be developed

The main issues for further development are on the dish and feed:

9.1.1 Further detailing the dish material performance and coating.

The production of the dish parts is not seen as risky. A lot of knowledge is already available and can

be used. Thermo plastic materials are more and more used in industry for airplanes. We will take a

next step in this, by selection of different type of thermoplastics, what will be in potential cheaper

and suitable for use in e.g. telescopes and automotive. The required development is towards the use

of the other type of thermoplastic materials. Typical UV, coating, mechanical stability and thermal

stability need to be studied further. Also production can be optimized. Tests on performance of

some single dish panels within the Westerbork radio telescope are foreseen within 2011. Further

development is required to push the cost further down and check on quality. Several demonstrator

steps are expected. This development is expected to benefit from developments within the

aerospace and automotive industry.

9.1.2 Detailing the assembly of the dish

Dish assembly out of several panels with aid of a reference mould is done more often. For example

the backing structure of the ALMA dish is made out of several assembled large composite panels.

The proposed assembly will be based on smaller panels with sizes up to 2.5 meter, but also larger

panels of about 7.5 meter by 2 meter are in study. This assembly needs further development to find

the optimal assembly and correct stability and will need to be finalized within about 2 years.

9.1.3 Feed

Test and development of the eleven feed and optimizing its performance; the feed is seen as

development trajectory with relative low risk, but still significant items to address. Typical its

sensitivity over the bandwidth should be pushed to the limits, this in relation to relative small and

lightweight construction and possibly cooling. We expect to develop this further over the coming

years and not on the critical path.


Page 45 of 50

9.2 Partners in the Thermoplastic Composite SKA Reflector project (TC SKAR)

9.2.1 Airborne www.airborne.nl

Airborne manufactures the back‐up structure of 12‐meter diameter

carbon fibre reinforced plastic (CFRP) reflector for the ALMA project in

Chile including the Quadra‐pods and other CFRP parts. Airborne also

specialises in high‐end composite products for other markets as aerospace, space, machines and

composite piping for oil & gas industry. For a project called TC‐SKAR Airborne serves as the man

contractor. TC‐SKAR studies the production of an off‐axis Gregorian

9.2.2 Dutch thermoplastic components www.composites.nl

DTC is a parts manufacturer specialized in press forming and machining of advanced

thermoplastic composites. Their high performance parts are used in aero structures

and demanding industrial applications.

9.2.3 Kok en van Engelen www.kve.nl

KVE has developed an induction welding process for thermoplastic materials, which

is now being used in several markets, like automotive, space and the aerospace

sector as assembly process.

9.2.4 Delft University of Technology www.tudelft.nl

TU Delft has a record of accomplishment in simulating, testing and

constructing with composite materials and performs the material research

within this project.

9.2.5 ASTRON www.astron.nl

ASTRON is the Dutch Institute for Radio Astronomy. Its mission is to

make discoveries in radio astronomy happen, via the development of

novel and innovative technologies, the operation of world‐class radio

astronomy facilities, and the pursuit of fundamental astronomical research.

ASTRON was supported by Chalmers University. www.chalmer.se


Page 46 of 50

9.3 Risk assessment and mitigation

The risk assessment is focussing on the technical issues within this study. The more systems and

science related risks are seen as low or known; because the presented concept is rather classical and

wide spread knowledge is available for this system concept.

No Risk Short description Impact Proposed mitigation

1 Dish

1.1 Wavelength

range

The specified wavelength range is 1-10 GHz. This wavelength range is not matching the baseline specification for SKA-phase 1

The lower frequencies (0.45 – 3 GHz), as specified for SKA phase 1, will change the feed design and feed mount design. It can also impact the dish material choice. Due to lower freq.range the requirements are more relaxed and cost might decrease.

Early input for this concept and the defined wavelength range is required. With that new (more relaxed) tolerances can be provided to industry.

1.2 Cost The cost of the dish is driving the total cost. This cost is defined by industry by aid of costing tools and by industry expressed as realistic. The feed design is not costed in detail and need further study. Certainly the type of cooling is one of this issues in costing and performance

The impact is rather low. The main risk is in the feed, what is not the most driving element in the total cost.

Early definition of the

wavelength range is

important. Detailed study

and test needed in near

future on feed‐cost.

1.3 Lifetime The material of the dish is a thermoplastic material, which is not yet used in telescopes as mirror as far as we know. This dish (and its material) is dominating the cost of the telescope. Other parts are seen as more

Failure of the 30 year lifespan of the dish would be dramatic for SKA. This really needs to be excluded (see mitigation).

Measuring/testing of the performance over time and the defined environment is crucial. These tests are nowadays standard and reliable to perform, but testing in the real environment would be the ultimate test. Industry already measured several


Page 47 of 50

classical and low in risk parameters of this material e.g. e.m. absorption, index of refraction, reflection coeff.. Out of the basic material specifications some extrapolations can be made for the applied materials, but an extensive test and validation programme will be done within the next phase.

1.4 Environmental

unknowns

Although some information has been gathered and was submitted during the site bids, more information is required especially with regard to extreme events and unknowns (such as animal life and flooding). Lightning strikes could also be a hazard for the exposed PAF feed elements.

Failing to understand and include these requirements in the design will most probably result in rework and retrofit. Given the scale of the SKA, this will be very costly.

Experience gained with the ASKAP system will be invaluable, and lessons learned must be transferred to the SKA system.

1.5 Lightning Lightning is a serious issue. Up to now a connection between antenna elements of the dish is foreseen. A study is required to see if this will be beneficial for the design in cost and performance. Biggest change on lightning impact is the feedbox. It is advised to protect at least the feedbox for lightning.

Lightning damage of the feedbox is probably more costly then damage of the mirror (exchange some elements). Damage of the feedbox will stop the full telescope. Local damage of the disc will reduced performance, but not necessarily stops the observing.

Lightning protection of the feedbox as minimum requirement. Eventual simple electrical connection of the dish elements can be provided if necessary.

1.6 Remote

operations

The fact that the SKA will be deployed, operated and supported on a very remote site poses many challenges in almost every aspect of the system, especially quality aspects. Other examples are reliability, availability,

If the requirements for remote operations and support are not part of the design from the outset, the cost of rework and upgrades may be excessive.

Investigate all aspects that will influence remote operations and support. Evaluate different options. Keep full lifecycle aspects in mind. Ensure that requirements are flowed down into the design and budgeted for.

Experience in operating


Page 48 of 50

maintainability, monitoring and control, support from a distance, safety etc.

ASKAP will be valuable, although care will be needed in scaling the outcomes to the SKA.

1.7 Scope of

logistics and

support

To be able to operate and support the SKA over its lifetime will require well designed and supplied logistics and support. So far, this aspect has not received much attention. Specific moving parts are subject of maintenance and failure.

Due to the scale of the SKA, the logistics and support requirements will be significant and will call for large numbers of people, spares, tools, test equipment, support equipment, facilities, training etc.

Specific design towards low maintenance and high reliability is crucial. Focus on moving parts and UV-protection of sun-loaded parts (e.g. dish).

1.8 Damage During manufacturing, transport, installation or operation, damage can occur.

Depending on the damage, the functionality or performance of the antenna can be reduced.

Reliable and adequate repair method needs to be developed. Airborne has experience with repair techniques of thermoplastic structures, for example for offshore operations.

Furthermore, the concept uses smaller elements that are assembled into the final reflector. A method to replace such an element when damaged could be used.

Creep Continuous load on the material structure may incur creep of the material.

The performance of the dish will decrease over time.

Extensive test are being performed at different conditions to map the material properties. Design will be adapted to embed the allowable loads.

2 Feed

2.1 Performance The performances of the feed design is depending on the design and probably also temperature of the feed and LNA. Cooling concepts might be important for a good balance of performance

The feed is a driving element in the total Aeff/ Tsys specification, the main parameter of the telescope. Improving the Aeff/Tsys can lower the overall cost and

Study the type of feed (Eleven feed) in more detail. In principle the systems are well known, but a correct balance between technology, cost performance and risk is important.


Page 49 of 50

and cost. improve the operations overhead.

2.2

3 Pedestal

3.1 Lifetime Has not been studied in any detail.

3.2

3.3


Page 50 of 50

Annex I Gregorian offset reflector design

In a previous phase a Gregorian offset reflector was designed. Several concepts were analyzed,

firstly a simple un‐stiffened skin with a truss support structure, see Figure 14. In the initial stage this

concept seemed infeasible, particularly because of thermal stability issues.

Figure 14 Gregorian offset reflector with a truss support

Figure 15 Gregorian offset reflector with a

stiffened skin

Secondly, a stiffened skin design was analyzed, see Figure 15. The feed support design and weights

are based on the USSKA design as presented in [27]. It should be noted that this is significantly

different from the assumption on which the symmetric design, as presented in this report, is based.

The analysis showed that the design of approximately 3.6 [tons] would have an rms performance of

around 1.1 [mm] (including thermal, wind, and gravity loading). This increase in mass, and decrease

in accuracy, with respect to the symmetric design can only partly be explained by the significant

difference in the feed design. The large overhang of the a‐symmetric dish on one side and the loads

from the feed struts on the other make this design less accurate especially concerning gravity loads.

Documents

CONCEPT DESCRIPTION: SYMMETRIC DISH … · Approved for release as part of dish CoDR ... For example press‐forming and ... design, to analysis and calculation up to production of