PrepSKA

•

A Preparatory Phase Proposal for the Square Kilometer Array radiotelescope

•

EU - FP7

•

Domingos Barbosa, Paulo André, Luis Cupido (IPFN-Aveiro)

Agências: STFC (Reino Unido), NWO (Holanda), CNRS (França), INAF (Itália), DIISR (Austrália), NRF (África do Sul), NRC-HIA (Canadá), CSIRO (Austrália)

Institutos: U. Manchester (Reino Unido), Max Planck (Alemanha), Instituto de Telecomunicões (Portugal),

U. Cambridge (Reino Unido) , U. Oxford (Reino Unido), U. Cornell (EUA) , JIVE(Europa), Obs. Paris (França) , U. Calgary (Canadá), U.Groningen (Holanda)

Outras Instituições participantes: National Science Foundation (EUA), Associated Universities Inc (EUA), Fundacion General de la Universidad de Alcala/Instituto Geografico Nacional (Espanha), Vetenskapsradet (Suécia), SKA Programme Development Office (Global).

Orçamento : 22Meuros ; 5.5Meuros FP7

Simulations for the SKA

•

Objectivos : sondar um cone do Universo de cerca 12 mil milões

anos.

On

behalf

of

Phil

Diamond

•

Build

over

SKADS FP6•

FP6 –

Design

Stdudies; FP7 –

Preparory

Phase; FP8 -

Construction

ISPOIntroducing SKA

Aperture synthesis radio telescope with 1 km2 of effective collecting area by 2020

4000 antennas; 200 stations; 3000 km

50x sensitivity, 10 000x survey speed of existing arrays Frequency range 0.1 - 25 GHzStaged construction

Large bandwidths (4 GHz), large fields-of-view (50 deg2)New capabilities: area re-use (“multi-fielding”), RFI mitigation, high dynamic range imaging, ….

Innovative design to reduce costInternational funding: ~ € 1.5 billion17-country international consortium

– International SKA Project Office (ISPO) since 2003

2 short-listed sites; final selection ~ 2011– Western Australia, Southern Africa

Hugedata rates& volumes

}

ISPO

radio “fish-eye lens”

Inner core

Station

Digital radio camera + stations to3000 km

Radio fish-eye lens

SKA pictorially

ISPOSKA characteristics

Exploits convergence of radio and ICT– More parameter space, flexibility = more discovery

Uses less metal, more ICT– Many gains via consumer-driven technologies

Is an incubator for selected leading-edge technology– Astronomers are “sophisticated end users” – (Radio astronomy is traditionally an effective incubator)

Design paradigm poses new challenges– But gives new players opportunities to develop pivotal

technologies

7

ISPO

SKA conceptually- radio meets IT

(Antennas)

ISPOThe antenna dilemma

What defines the primary FoV?(1st stage beamformer technology)

Optics Electronics

Concentrator (dish) Aperture phased array

Analog (RF)beamform

Digitalbeamform

Technology choice depends on applications, frequency and delivery epoch

FOV expansion•Optical (multiple feed cluster)•Electronic (phased array feed)

continuum

ATA, KAT, APERTIF, ASKAP EMBRACE LOFAR 2-PADMWABEST, SKAMP

Extreme electronic beamforming

ISPO“Reference Design” antennas

> 3 GHz: wide-band feed

< 0.3 GHz: sparse aperture array

0.3 – 3 GHz:phased arrayfeed

Mid-band all-sky monitor: dense aperture array

Mid-BandHigh-Band

Swinburne/CVA visualization

Low-Band

ISPOSKA antenna applications

Frequency range(GHz)

Sparse Aperture Array

Dense Aperture Array

Dish + Focal Plane Array

Dish + Single-Pixel Feed

Low-band“EoR” array

0.1-0.3

All-sky monitor

0.3-1

Imaging mid-band array

0.3-3

(to ~0.5 GHz) (to ~1 GHz)

High-band array

3-25

Pathfinders or Design Studies

LOFAR, MWA, LWA

SKADS ASKAP, APERTIF

ATA, TDP, meerKAT

Mid-band SKA is the focus of intense Pathfinder activity

ISPOSKA development

Astronomy & engineering iteration to refine specifications– Specs agreed by early 2008

International system design effort ramping upEmphasis on technology demonstration– Retire risk as early as

possible– Regional pathfinders &

design studies are crucial» > €200M investment

Focus on:– Aggressive cost reduction

strategies– Industry engagement

» To deliver SKA on required timescales

Reference Designtechnologies

1% (Pathfinders) 10% (SKA Phase 1) 100% (SKA)

ISPO

SKA – not just antenna challengesHigh speed data transport– Tb/s from EACH station on scales of hundreds of km – 100 Gb/s trans-continental and trans-oceanic links

Signal processing– Peta-ops per second– Need highly scaleable solutions

Post-processing– Computing capacity limitations require staged science

capability– Archive and sharing of data will be a major challenge

Infrastructure– Civil, electrical (power, …), communications

Operations and support– Expect operations cost of €100 M annually

ISPO

2000

Sites short-listed ‘1% SKA’

Science

Sciencecase

published PrepSKA preparatory phase

activities ‘10% SKA’Science

98 2002 06 07 08 09 10 11 12 13 14 18 22

Feasibility studies and

concept demonstration

Full arrayconstruction100% SKA

SKAComplete

Phase 1construction10% SKA

SKA timeline

Reference Design selected

Pathfinder suite construction

External review of top-level design

Concept design PrepSKA system design

Initial specs for Phase 1 and full SKA

Detailed Phase 1 design &top-level SKA design

ISPO

SKA Preparatory Phase – PrepSKA(2008 – 2011)

WP1: PrepSKA managementWP2: SKA system design– Includes Initial Verification System – Establishes central design integration team (CDIT)

WP3: Continuing site selection processWP4: GovernanceWP5: Industry and procurement policyWP6: Funding modelWP7: Implementation strategy

Strong international collaboration

ISPOPrepSKA system design

Outline Benchmark Spec.Parameter Memos Benchmark Spec. Comments

45/69 EOR AA Mid-AA DishesLow Freq. GHz ≤

0.1

0.1

0.3 >1.0

Will have some overlapHigh Freq GHz ≥

25

0.3

1.0 25

AA cost goes > square law with top frequencyBandwidth GHz 25%

0.2

0.7 5

This may be unrealistic for the dishes

Polarisations 2 2

2 2

Default linearPol error (after cal):

FOV centre/edge dB -40/-30

-40/-30

-40 /-30 -40/-30 Important for dynamic range and astronomy.

FOV: 0.1-0.3 GHz deg2 200 200 This is really total beam size 0.3-1.0 GHz deg2 50 250 The FOV naturally scales as 250(1/f)2

1.0-3.0 GHz deg2 1-10

3(1.4/f)2

Defined by a 6.1m dish natural beam size. 3.0-25 GHz deg2 0.33(3/f)2

3(1.4/f)2

FOV filling 100%

100%

100% 100% The full field of view should be filled with beamsNo. of steerable FOVs 1-4

4

8 1

AA FOV count limited by comms.

Survey @ 0.7GHz 1.5 x 1019

1.75 x 1019 Assumes 10,000m2/K sensitivitySpeed: @ 1.5GHz 3 x 1017

8.4 x 1017

Assumed 700MHz B/W at this freq, 3 deg2 FOV deg2 m4 K-2 Hz-1 (FoV x (A/T)2 x BW)

System temp Tsys K 50K

>100K

50K ≤30K

Assumes dishes use at least 70K coolingSensitivity @45°

m2K-1

>0.3 GHz 5,000 5,000 The sky noise is v. high for low freq 0.3 –

1.0 GHz 20,000 5-10,000 Reduced sensitivity by trading with large FOV 1-10 GHz 20,000 10,000 Single pixel feed so can cool for low system temp >10 GHz 10,000 10,000

Dynamic range 106

106

106 106

Peak b’ness to rms noise level (not ADC range!) Probably need ~107!

Overall SKA Configuration

Station

Core (5Km dia), made up of close packed stations

Correlatorin or near Core

Comms

links

Not to scale!DesertDesert

Link to Correlator

High freq. Dishes withanalogue fibre link

EoR

AA antennas withanalogue link (may be close packed)

SKA SKA StationStation

Notes:1. Power dist. not shown2. All analogue links go to

‘station processing’

Illustration only

StationProcessing

Station ProcessingBunker

2nd

StageProcessor 1

0.3-1.0GHz Analog

links

2nd

StageProcessor 2

2nd

StageProcessor X

…..

Dish P1Dish P2

Dish Px

Tile P1Tile P2

Tile Py

EoR

P1EoR

P2

EoR

Pz

.Internal

Digital links

n x Optical fibres per 2nd

stage processor

To Correlator

Mid Freq. AA

GPS Phase Standard

.

.

.

.

..

11+GHz Analog

fibre links

300MHz Analog

links.

.

.

High Freq. dishes

EoRAA

... Control processors

To CentralControl system

10Gb Digital fibre links

PrepSKA

WG2 An IT contribution

Packaging solutions for LNA

Network Infrastructure and Data Transmission

Final SKA Implementation Plan

IT tasks : WG2 & WG7

•

SKA-

ICT machine

(data transport

and detection-

~100Gby/s; xPetaflops) ; channel

through

10-40Gb/s links

•

Task

2.6 –

To produce

a suite

of

advanced

prototype

receivers

(based on

SKA PF and SKADS), testing

state

of

the

art

semiconductors,

close

to industry

. •

Task

2.7 –

To produce

advanced

prototypes

of

optical

fiber

data

transport

with

high

bandwidth, highly

sinchronized

(picosec) from 20m-200km. Importance

of

high-performance

commercial

telecoms.

•

Important

Effort

: 44 Man/month

over

3yrs (~50%U.Cam, Oxford, MPG).

•

Important

and Strategic

: R&D in

ICT cornerstones

and Physics..•

Strategic technologies: maintain

in

Europe

centers

for

semiconductor

production

and data links R&D. PT closer

to foundries.•

Portugal has

techological

readiness!!

•

Need

for Phase

1 : mass

production.

•

Close

partnership

with

Risk

evaluators

needed

(ie

Science

Agencies) –

WG4, WG5, WG6

LNA Testing

•

Tinstrument=40K (excluding sky noise), goal 30K

•

BW

70MHz –

1.0 GHz (two systems)

•

Survey speed ~1/T2•

Sensitivity

~1/T

LNA testing•

A 9 million element system with a total system cost of 250M €

can spend:

–

1,5 €

per LNA per Kelvin improvement (Survey)–

Or 8,5 €

for 5 Kelvin improvement, again for Surveys

•

Given a bare die costs of:–

0,5 €

for Silicon technologies (only 500 12 inch

wafers)–

2 €

for GaAs technologies (2000 6 inch wafers)

•

Low cost technologies cannot compromise on noise!

•

GaAs PsHEMT / mHEMT

•

SiGe BJT•

with b0 =100: FMIN

~30K

•

CMOS•

In principle similar to GaAs

LNA Technology

Amplifier Noise Figure Trends @1.4 GHz Tamb=290K

0

10

20

30

40

50

60

70

2000 2002 2004 2006 2008 2010 2012

year

Noi

se T

empe

ratu

re

III/V: GaAs or InPSiGeCMOS

• 0.2um technology• OMMIC differential LNA

– 2109– ASTRON design

PsHEMT

1 -Development of a prototype phase transfer system over installed fibre links.

2 -Cost model analysis for all aspects of the data network including installation and equipment costs.

•

Plan to use our existing equipment, simply transmit the signal optically, rather over microwave links.

•

Observe if the fibre link is reciprocal over distances up to 120 km

•

Show the current system is no worse than the existing system

•

Show that the system is accurate to 1 ps over 1 minute and 10 ps over 10 minutes

Master L-Band Link

HP 8508A Vector Voltmeter

Aerial switchSlave L-band Link

499.9MHz over 3m of RG214

SMF28 Fibre Link (variable length)

1486.3 MHz signal

1486.3 MHz signal over 30m of RG214

Master Laser

Master Receiver

Slave Receiver

Slave Laser

-23dB

-40dB

Lab path

Link path

499.9 MHz over 30m of RG214

•

Aperture

Arrays

(a kind

of

friendly

Synthetic

Aperture Radar for RA).

•

Need

for Massive

station

-

evaluate

Astrophysics readiness

•

Evaluate

technology

with

adverse, close

to real SKA operational

conditions

(hot, dry, dusty, high

thermal

amplitudes)•

Need

for solar power

plant

–

Portugal has

most

iluminated

area

in

Europe.•

Consortium

being

built

–

IT is one

of

the

participants

•

SKA –

R&D for green

energies

<-

isolated

stations

in

the desert

need

autonomous, low

maintenance

power

plants. Energy

option

–

Frauenhofer

Inst. Lidership. EFACEC

contacts

•

Implementation

: ready

by

2012 ; 6Meuros budget

to be distributed

among

partners

SKA demonstrator in Iberia

SKA demonstrator in Iberia

Over

EMBRACE SKADS prototypes

The Unknown•

New discoveries always result from observations in new parameter space–

sensitivity

–

spatial resolution–

spectral resolution

–

polarisation–

time domain

–

observing speed (multibeaming)•

eg. CMB, pulsars, extra solar planets,…

SKA improves all of these

SKA is designed for the Key Projects but with an overriding design philosophy of flexibility to maximise

the likelihood of new discoveries