Date post: | 20-May-2018 |
Category: |
Documents |
Upload: | duongxuyen |
View: | 214 times |
Download: | 1 times |
PAF WORKSHOP 2017 Phased Array Feed Workshop
Sydney, Australia November 14th TO 16th
Abstracts and Information ISBN 978 0 646 98052 2
CSIRO ASTRONOMY AND SPACE SCIENCE (CASS)
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
I
Table of Contents Welcome from the Organising Committee ........................................................................................................II
Marsfield Site Orientation .................................................................................................................................III
Workshop Social Dinner Wednesday 15th November ...................................................................................... IV
Site Tour CSIRO Marsfield Wednesday 15th November .................................................................................... V
Instructions for Presenting Authors ................................................................................................................. VI
Workshop Program – Day 1 ............................................................................................................................ VII
Workshop Program – Day 2 ............................................................................................................................. IX
Workshop Program – Day 3 ............................................................................................................................. XI
Abstracts – Day 1 ................................................................................................................................................ 1
Abstracts – Day 2 .............................................................................................................................................. 13
Abstracts – Day 3 .............................................................................................................................................. 22
Author Index ..................................................................................................................................................... 27
Affiliation Index ................................................................................................................................................ 29
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
II
Welcome from the Organising Committee
Welcome to the 2017 PAF Workshop hosted by CSIRO Astronomy and Space
Science, at CSIRO’s Radiophysics Laboratory in Marsfield, Australia.
We would like to begin by acknowledging the Wallumattagal people of Eora nation as the Traditional Owners of the land that we’re meeting on today, and
pay our respects to their Elders past and present.
We received over 30 submissions for talks this year representing over 24
institutions worldwide. This year we welcome delegates from Australia, Canada, China, Germany, Italy, Spain, The Netherlands, UK and USA. The technical
program provides an excellent overview of a broad range of PAF related activities, including many collaborations. We would like to thank all authors for
their abstract submissions and all attendees for facilitating this opportunity for interaction.
We trust this will be a rewarding three days, focussing on recent developments
and future opportunities in phased array feeds for radio astronomy.
PAF Workshop 2017 Organising Committee Douglas Hayman
Stephanie Smith Ken Smart
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
III
Marsfield Site Orientation
Map of relevant rooms on the Marsfield site
Registration will take place on the left just past reception. As a visitor to the site you will be required to sign in and wear your name badge at all times while on
site.
All presentations will take place in the lecture theatre and breaks and lunches will be taken in the breakout room and the adjoining courtyard. The Faraday
Room is available as a breakout room for the duration of the Workshop.
Internet access
Is available on the Marsfield site and is the preferred first choice
for access. Eduroam is an international roaming service for users in research, higher education and further education. It provides researchers, teachers and
students easy and secure network access when visiting an institution other than their own.
If you don’t have access to eduroam, a wifi connection “csiro-guest” is also
available for each registered participant. Your individual login details can be found on the back of your name badge.
The local organizing committee is available for assistance.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
IV
Workshop Social Dinner Wednesday 15th November
The workshop social dinner is a cruise on Sydney harbour.
5:30pm. The bus will depart from the main entrance near reception for transfer to Pier 26 Wharf (outside Sydney Aquarium).
7:00pm. The boat will depart from Pier 26 Wharf (outside Sydney Aquarium). Depending on traffic there may be some time beside the harbour before the
boat departs please stay nearby so as not to miss the boat.
10:30pm. Disembark the boat at Pier 26 Wharf (outside Sydney Aquarium). You are free to make you own arrangements for travel back to your Hotel. We
suggest those going in a similar direction group together for taxis/ Uber etc.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
V
Site Tour CSIRO Marsfield Wednesday 15th November
There will be a tour of various labs and facilities at Marsfield. Participants wishing to go on the tour will be divided into two groups. For safety reasons, participants
should wear suitable closed shoes (do not wear open-toed shoes, sandals).
Douglas Hayman and Stephanie Smith will guide the groups to the different locations around the Marsfield site. Your on-location guides will provide an
overview and answer questions on the following areas:
Anechoic Chamber, Ken Smart.
FAST 19 element multibeam feed being tested in the anechoic
chamber.
The Frontend Group, Alex Dunning.
FAST 19 element multibeam feed and CSIRO team.
The Backend Group, John Tuthill.
ASKAP Correlator and Beamformer at MRO.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
VI
Instructions for Presenting Authors
We are expecting approximately 60 attendees over the three days of the PAF
Workshop. We have 26 talks scheduled throughout including 2 invited talks. The invited talks have been allocated a longer timeslot of 40 minutes to allow
for a detailed presentation and discussion.
Each presentation is allocated a 30-minute timeslot, including 5 minutes for questions and discussion (25 min + 5 min). Although we are single stream, we
would like to stick to the schedule to allow ample opportunity for networking
during the breaks, and so our WEBEX participants can dial in for selected presentations.
The Lecture Theatre at Marsfield is equipped with a projector and PC running
Windows 10. We ask for all presenting authors to bring their presentation on a USB stick to be copied onto the PC in the Lecture Theatre during the break
immediately before their scheduled session. The use of WEBEX precludes presenters from using personal laptops. The format of the presentations should
be PowerPoint (.ppt) or PDF (.pdf). There will be no Macintosh platform available.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
VII
Workshop Program – Day 1
Tuesday, 14 November, 2017
TIME SESSION PAGE
8:00AM Registration and Coffee
8:45AM Conference Opening
Sarah Pearce, Deputy Director CASS
8:55AM Logistic information
9:00AM SESSION 1
Chair: Lisa Locke
9:00AM A New PAF Receiver Using XILINX RFSOC Technology
G.A. Hampson, J.D. Bunton, A. Brown, R. Beresford and J. Tuthill
1
9:30AM A Radio Astronomy L-Band Phased Array Feed System using RF over
Fiber
Ron Beresford, Dick Ferris, Wan Cheng, Grant Hampson, John Bunton,
Aaron Chippendale and Jega Kanapathippillai
2
10:00AM An S-band Cryogenic Phased Array Feed for the LOVELL Telescope: RF
Mark McCulloch, Michael Keith and Simon J. Melhuish
3
10:30AM Morning Tea
11:00AM SESSION 2
Chair: Mark Bowen
11:00AM Study Towards Cryo-Cooled Phased Array Radar Systems
Andreas Froehlich, Marco Tiesing, Nadya Ben Bekhti, Felix Koenig,
Sergiy Putselyk, Lukas Naumann and Florian Rahlf
4
11:30AM PAF Telescope Technology for 5G Base-Stations
U. Johannsen, T.A.H. Bressner, A.A.H.M. Elsakka, R. Maaskant, M.
Ivashina and A.B. Smolders
5
12:00PM INVITED TALK
The Development of Focal Plane Arrays in Radio Astronomy
Ron Ekers and John O'Sullivan
6
12:40PM Lunch
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
VIII
Workshop Program – Continuation of Day 1
Tuesday, 14 November, 2017
TIME SESSION PAGE
1:45PM SESSION 3
Chair: Bruce Veidt
1:45PM The Cloverleaf Antenna Array on CHIME
Meiling Deng
7
2:15PM CryoPAF: a 2.8 – 5.18GHz Cryogenic Phased Array Feed Receiver
Lisa Locke, Vlad Reshetov, Doug Henke, Frank Jiang, Gordon Lacy,
Bruce Veidt and Gary Hovey
8
2:45PM Update of PHased Arrays for Reflector Observing Systems
Lei Liu, Keith Grainge and Michael Keith
9
3:15PM Group Photo before afternoon tea
3:30PM Afternoon Tea
4:00PM SESSION 4
Chair: Brian Jeffs
4:00PM A New 1Tbps Ethernet Streaming Data Interface Between FPGA and
GPU Based Digital Signal Processing Systems for a Phased Array Feed
Daniel George, John Tuthill, Andrew Brown, Chris Phillips, Ron
Beresford, Mark Leach, Tasso Tzioumis, Mia Baquiran, Grant A.
Hampson and John D. Bunton
10
4:30PM Study Of The Dish Optics For Phased Array Feed
Wu Yang, Du Biao, Niu Shengpu and Jin Chengjin
11
5:00PM Phased Array Feed (PAF) Design for the LOVELL Antenna based on the
Octagonal Ring Antenna (ORA) Array
Ming Yang, Yongwei Zhang, Laith Danoon and Anthony Brown
12
5:30PM End of Day 1
Free Evening
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
IX
Workshop Program – Day 2
Wednesday, 15 November, 2017
TIME SESSION PAGE
8:45AM Coffee
9:15AM SESSION 5
Chair: Gundolf Wieching
9:15AM Making Robust Beamformer Weights by Interpolating in the Frequency
Domain
A. P. Chippendale, K. Bannister and G. Hellbourg
13
9:45AM Advancements in PAF Correlator and Beamformer Systems
G.A. Hampson, J.D. Bunton, A. Brown, J. Tuthill and G. Schoonderbeek
14
10:15AM Beamformer and Calibration Performance for the Focal-plane L-band
Array feed for the Green Bank Telescope (FLAG)
B. D. Jeffs, K. F. Warnick, R. A. Black, M. Ruzindanna, M. Burnett, D.
J. Pisano, D. R. Lorimer, N. Pingel, K. Rajwade, R. M. Prestage, S.
White, B. Simon, L. Hawkins, W. Shillue and A. Roshi
15
10:45AM Morning Tea
11:15AM SESSION 6
Chair: Alex Dunning
11:15AM Specifying Polarisation Properties For Radio Telescopes
Bruce Veidt
16
11:45AM Results From The Astronomical Commissioning Of APERTIF
Tom Oosterloo and The APERTIF Team
17
12:15PM New-Generation Radio Telescopes: Challenges in Instrumental
Calibration
Wasim Raja
18
12:45PM Lunch
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
X
Workshop Program – Continuation of Day 2
Wednesday, 15 November, 2017
TIME SESSION PAGE
1:45PM SESSION 7
Chair: Keith Grainge
1:45PM System Performance Characterisation Of WSRT-APERTIF
Boudewijn Hut and Wim van Cappellen
19
2:15PM Status Of The AFAD-C Project
Bruce Veidt, Tom Burgess, Aaron Beaulieu, Leo Belostotski, and James
Haslett
20
2:45PM A Highly Sensitive Focal L-Band Phased Array Feed for the Green Bank
Telescope: Electromagnetic Model and Results
D. Anish Roshi, W. Shillue, B. Simon, S. White, R. Prestage, J. Diao,
K. F. Warnick, B. Jeffs, D. J. Pisano, D. Lorimer, T. Boyd, J. Castro, J.
R. Fisher, W. Groves, M. Morgan, L. Hawkins, L. Jensen, J. D. Nelson,
J. Ray and V. van Tonder
21
3:15PM Afternoon Tea
3:45PM Site Tour
Guides: Ken Smart, John Tuthill, Alex Dunning
5:30PM Bus from CSIRO to Pier 26 Wharf (outside Sydney Aquarium)
7:00PM Conference Dinner – Harbour Cruise
10:30PM Disembark Pier 26 Wharf – Individual arrangements
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
XI
Workshop Program – Day 3
Thursday, 16 November, 2017
TIME SESSION PAGE
8:45AM Coffee
9:15AM SESSION 8
Chair: Bill Shillue
9:15AM PHAROS2 Phased Array Feed: Warm Section, Signal Transportation
and iTPM Digital Backend
A. Navarrini, J. Monari, A. Melis, R. Concu, A. Scalambra, A.
Maccaferri, A. Cattani, P. Ortu, G. Naldi, J. Roda, F. Perini, G.
Comoretto, M. Morsiani, A. Ladu, S. Rusticelli, A. Mattana, L.
Marongiu, M. Schiaffino, E. Carretti, A. Saba, F. Schillirò, E. Urru, G.
Pupillo, K. Zarb Adami, A. Magro and R. Chiello
22
9:45AM An S-band Cryogenic Phased Array Feed for the LOVELL Telescope:
Cryogenics
Simon J. Melhuish, Michael Keith and Mark McCulloch
23
10:15AM A Proposed Cryogenic PAF for Parkes
James Green
24
10:45AM Morning Tea
11:15AM SESSION 9
Chair: Stephanie Smith
11:15AM The SKA Observatory Development Programme and Phased Array Feed
Development
M.A. Bowen
25
11:45AM INVITED TALK
Astronomy with PAFs on ASKAP
David McConnell
26
12:25PM Lunch
1:30PM Wrap – Up
Chair: Douglas Hayman
3:00PM Afternoon tea
3:30PM SKA PAF AIP Consortium Meeting – Faraday Room
Chair: Steve Barker
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
1
A New PAF Receiver Using XILINX RFSOC Technology
G.A. Hampson, J.D. Bunton, A. Brown, R. Beresford and J. Tuthill
CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected]
A new generation of Field Programmable Gate Array (FPGA) that incorporates Analog to Digital Converters
(ADC) on chip is set to revolutionise the FPGA market sector. The RF class ADCs [1] not only reduce the size
of RF receivers but also have the potential for significant cost and power reductions. The target future market
for the technology is mainly focussed on 5G applications [2] where power is important as well as the ability
to perform signal processing on a larger number of receivers is critical. This new technology is becoming
available at a time when radio astronomy is making extensive use of receiver arrays [3].
The new RFSOC FPGAs, shown in Figure 1, have both ADCs and Digital to Analog Converters (DACs). The
DACs probably will not be used in radio astronomy applications and can be turned off to save power. The
ADCs have good performance for RF frequencies up to several GHz [4] which make them suitable for L-band
(or lower frequency) radio astronomy instrumentation solutions. Per device there are up to 16-ADCs with 12-
bits of resolution and it is relatively straight forward to create larger arrays by interconnecting multiple FPGAs
using the high speed serial links. A processor also exists to implement monitoring and control, and possibly
time and spatial RFI excision algorithms.
Figure 1. RFSOC FPGAs contain a mix of FPGA, processors and ADC/DACs.
When the RFSOCs are combined with RFOF [5] to allow electrical isolation then it is possible to locate the
receiver system near the antenna reducing the length a signal has to travel before being digitised and processed.
A significant advantage of RFSOCs is the reduction in power by removing the ADC to FPGA communications,
as well as being able to digitally process the data locally (filterbanks and beamforming) and reducing the
communications to the remote correlator and/or beamforming system. Reducing power also impacts cooling
and the resulting RFSOC solution becomes attractive in size, weight, and operational costs. Such a system is
very suitable for the next generation of PAFs.
[1] Xilinx, “RF Data Converters in an All Programmable MPSOC”, https://www.xilinx.com/products/silicon-
devices/soc/rfsoc.html.
[2] Amy Nordrum, “Everything you need to know about 5G”, IEEE Spectrum,
https://spectrum.ieee.org/video/telecom/wireless/everything-you-need-to-know-about-5g , 27 Jan 2017.
[3] Andrew Brown, “Design and implementation of the 2<sup>nd</sup> Generation ASKAP Digital Receiver System”,
2014 International Conference on Electromagnetics in Advanced Applications (ICEAA).
[4] Bruno Vaz, et. Al., “A 13b 4GS/s Digitally Assisted Dynamic 3-Stage Asynchronous Pipelined-SAR ADC”, 2017
IEEE International Solid-State Circuits Conference, February 7, 2017 Ron Beresford, “Radio Astronomy L-Band
Phased Array Feed RFoF Implementation Overview”, 2017 URSI General Assembly and Scientific Symposium
(GASS), August 2017.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
2
A Radio Astronomy L-Band Phased Array Feed System using RF over Fiber
Ron Beresford, Dick Ferris, Wan Cheng, Grant Hampson, John Bunton, Aaron Chippendale
and Jega Kanapathippillai CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected]
This is an overview of the system level RF design for the second generation architecture used in the Australian
Square Kilometre Array Pathfinder (ASKAP) [1] design enhancement (ADE). ADE is a distributed antenna
system (DAS) of 36 reflector antennas each 12m in diameter. Each antenna has a planar Phased Array Feed
(PAF) at the prime focus. The PAF contains 188 broadband 700-1800MHz receptors. Inside a PAF the
radioastronomy (RA) signals are amplified, band selected and converted to 188 individual broadband RF over
(singlemode optical) fiber (RFoF) lightwaves [2]. The entire ADE array has 6840 RFoF links, this includes
transmission line delay metrology for each reflector. The longest RFoF span is 6km. Optical to RF
demodulation of the RF sky signal at the central site Digital Signal Processing (DSP) shielded building is direct
sampled in 12bit analog to digital convertors (ADCs). The signal path to ADC for each PAF port is illustrated
in Figure 1. Digital beamforming provides 36 pencil beams, each of 384MHz bandwidth. The scale of ADE
represents a leap forward in applied RF and photonic techniques to enable a simpler, lower cost, more modular,
EMC compliant, phased array receiver architecture. ADE will provide unprecedented high speed sky surveys
with an instantaneous widefield of view (30deg2 at 1420MHz) capability.
Figure 1. Signal and Conversion SAC subsystem (single channel shown).
RFoF can be successfully integrated into a broadband L-Band radio-astronomy receiver design with adequate
dynamic range for radio quiet sites.
Figure 2. Nominal beamformed TR and IIP1, 0km and 6km optical spans.
Not surprisingly, there is an optimization between low input noise figure and high compression point as shown
in Figure 2. Remotely programmable adjustment of this feature is provided in the 700MHz-1800MHz ASKAP
design.
[1] DeBoer et al, “Australian SKA Pathfinder: A High-Dynamic Range Wide-Field of View Survey Telescope.” Vol. 97, No. 8,
August 2009 | Proceedings of the IEEE.
[2] Hampson et al, “ASKAP PAF ADE advancing an L-band PAF design towards SKA”, ICEAA Conference, Sept 2012.
[3] Beresford, Bunton, “Shielding Requirements for ADE Receivers” IEEE APEMC, May 2013.
[4] Hay, O’Sullivan , “Analysis of common-mode effects in a dual-polarized planar connected-array antenna”, Radio Sci., 43,
RS6S04, 2008.
[5] Brown,”Design and Implementation of the 2nd Generation ASKAP Digital Receiver System”, ICEAA Aug 2014.
[6] Chippendale, “Measured Aperture-Array Noise Temperature of the Mark II Phased Array Feed for ASKAP”, ISAP, 2015.
[7] Beresford, Cheng, Roberts, “Low Cost RF over Fiber Distribution For Radio Astronomy Phased Arrays”, URSI GA Aug 2017.
[8] Kraus,”Antennas for All Applications”, McGraw-Hill 2002, Chapter 12.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
3
An S-band Cryogenic Phased Array Feed for the LOVELL Telescope: RF
Mark McCulloch, Michael Keith and Simon J. Melhuish
The University of Manchester, Manchester, UK
email: [email protected]
In this talk we will review our progress, from an RF perspective, in developing an S-band phased array feed
(PAF) for the Lovell Telescope at the Jodrell Bank Observatory (for the cryogenic aspects please see an
accompanying talk [1]). We will present our designs for the antenna elements, discuss the low noise amplifiers
(LNAs), report on the development of a "cheap" RF over fibre system, and discuss some of the ideas that we
have for attempting to reduce the overall system noise temperature and the cost.
Further details; the cryogenic receiver will be composed of a 10 x 10 array of aluminum Vivaldi elements and
it will be sensitive to two polarizations. The Vivaldi antennas are based on a design by Locke et al (2016) [2]
and feed single ended LNAs that were purchased from the Low Noise Factory (LNF) [3]. Since there will be
several hundred metres between the cryogenic receiver and the backend electronics we are developing a
"homemade" RF over fibre system to connect the two. The system is based around a Broadcom AFBR-1310
transmitter and a AFBR-2310 receiver. We are also investigating if the transmitter can be placed inside the
cryostat, as this will greatly reduce the number of RF feedthroughs in the vacuum bulkhead. One drawback
with the outlined receiver is that it will require upwards of two hundred LNF LNAs, which at £4000/LNA
makes them by far, the dominant cost of the system (Figure 1). Consequently, we have been investigating
options for reducing the number of LNAs, with the use of Rotman lenses looking promising. Since our PAF
will be cooled to around 15K the use of superconducting components to lower the system noise temperature
has also been explored.
Figure 1. As a large number of good low noise LNAs are expensive, it is important to ask whether or not the number
required can be reduced?
[1] S. J. Melhuish, M. Keith, M. McCulloch, "An S-band cryogenic phased array feed for the Lovell Telescope:
cryogenics, Phased Array Feed Workshop, Sydney, Australia, 2017.
[2] L. Locke, D. Garcia, M. Halman, D. Henke, G. Hovey, N. Jiang, L. Knee, G. Lacy, D. Loop, M. Rupen, B. Veidt R.
Wierzbicki, " CryoPAF4: a cryogenic phased array feed design", Proc. SPIE 9914, Millimeter, Submillimeter, and
Far-Infrared Detectors and Instrumentation for Astronomy VII, Edinburgh, UK, 2016 AFBR-1310Z.
[3] LNF-LNC1.5_6A, Low Noise Factory, Goteborg, Sweden, www.lownoisefactory.com, last accessed 14-07-2017.
[4] AFBR-1310Z, Broadcom Ltd, https://www.broadcom.com/products/fiber-optic-modules-
components/industrial/industrial-control-general-purpose/1300nm/afbr-1310z, last accessed 14-07-2017.
[5] AFBR-2310Z, Broadcom Ltd, https://www.broadcom.com/products/fiber-optic-modules-
components/industrial/industrial-control-general-purpose/1300nm/afbr-2310z, last accessed 14-07-2017.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
4
Study Towards Cryo-Cooled Phased Array Radar Systems
Andreas Froehlich, Marco Tiesing and
Nadya Ben Bekhti, Felix Koenig, Sergiy Putselyk, Lukas Naumann, Florian Rahlf
Fraunhofer Institute for High Frequency Physics and Radar Techniques FHR, Wachtberg, Germany
email: [email protected]
email: [email protected]
email: [email protected]
Space debris is a growing problem for space based infrastructure like, communication and navigation satellites,
spacecraft, and other scientific purposes. More than 700 000 objects with critical sizes of more than one cm
are orbiting the earth and have the potential to cause severe damages. A natural consequence is therefore the
improvement of the detection sensitivity of next generation phased array radar systems for space surveillance.
The goal of our study is therefore to enhance the Signal-to-Noise ratio of a radar system by cryo-cooling critical
electronical devices of the receiver, contributing most to the system noise temperature.
Cryo-cooling receiver systems and the underlying fundamental technics are already known in Radio
Astronomy. With our study we are in the early phase of adapting this technology to phased array radar systems.
The Fraunhofer Institute for High Frequency Physics and Radar Techniques (FHR) as an established research
institute for high level, innovative radar solutions in Germany will develop a single, cryo-cooled, scalable
receiver channel for a phased array demonstrator. A key part of this demonstrator is the used antenna.
Therefore, different suitable antenna types (like Vivaldi, Cavity backed stacked patch antenna, etc.) will be
studied and evaluated. The requirements of such a cryo-cooled antenna are manifold: array capability, stable
antenna pattern, definite bandwidth, a constant element distance, very low antenna ohmic losses, minimal
decoupling behavior, etc.
Further challenging tasks are the characterization and analysis of the material properties under cryo-cooling as
well as the evaluation and optimization of measurement methods in cryogenic and vacuum environments.
Therefore, we study different LNA technologies and substrate behaviors. These aspects are of crucial
importance to develop an optimized receiver system for radar applications. The mechanical realization of a
cryo-cooled phased array radar system requires several critical design decisions. In order to analyze some of
the most critical design aspects we develop a Dewar in cooperation with a partner company for testing a single
receiver channel. This process includes the identification of an adequate radio frequency-transmissible radome
to allow radar operation.
The results of this first study will be a major step towards answering the question if a cryo-cooled phased array
radar will be an important contribution for future radar technologies.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
5
PAF Telescope Technology for 5G Base-Stations
U. Johannsen1, T.A.H. Bressner1, A.A.H.M. Elsakka1, R. Maaskant1,2, M. Ivashina2 and A.B. Smolders1
1 Eindhoven University of Technology, Eindhoven, The Netherlands 2 Chalmers University, Gothenburg, Sweden
email: [email protected]
Telecommunications research and development efforts are currently focused on the next generation mobile
communications standard, called fifth generation (5G). One of the goals of 5G is to offer data-rates of up to
10 Gbps at much higher system capacities [1]. In order to achieve this, additional frequency bands have to be
allocated. Here, the millimeter-wave (mm-wave) band is considered an attractive option [2]. However, the free
space path loss at mm-wave frequencies is significantly larger than for currently used sub-6 GHz frequencies.
Therefore, high gain antenna systems that allow the generation and steering of multiple beams are required on
the base-station side in order to serve multiple users simultaneously and with a sufficient signal-to-noise ratio
(SNR). At the same time, the mobile communications market demands for low-cost solutions in order to
achieve acceptable prices for the end user.
An often suggested antenna system approach for 5G mm-wave base-stations is a classical phased array, see
e.g. [3]. An advantage of this approach is that the effective isotropic radiated power (EIRP) that is required in
order to reach far-away users can be achieved by a combination of increased antenna gain and increased
number of power amplifiers (PAs). Hence, when using as many PAs as antenna elements in the array, the EIRP
scales with N², where N is the number of antenna elements. However, due to the use of standard (Bi-)CMOS
technology, the heat dissipation of the front-end electronics may prohibit the use of densely spaced arrays.
Increasing the element spacing on the other hand gives rise to the appearance of grating lobes. Moreover, if k
users have to be served at the same time and within the same frequency band, the available transmit power
must be divided between the users such that the maximum available EIRP decreases with an increasing number
of users (alternatively, the array may be split into k sub-arrays, which would have the same effect). Lastly, the
available transmit power of the user equipment (UE) is far below the output power of the base-station. Hence,
increasing the EIRP by adding PAs is only possible for downlink, i.e. from base-station to UE, and not for
uplink, i.e. from UE to base-station.
The European project SILIKA ([3]) takes a different approach for the base-station antenna system. Here, a
phased-array-fed reflector-antenna is chosen as an alternative solution. This antenna type offers a large antenna
gain at a potentially low-cost. Hence, compared to phased arrays without reflector, the large antenna gain
allows the use of lower transmit power levels for achieving the required EIRP. This, in turn, also results in a
higher link budget for the uplink case, assuring a bi-directional communication for larger distances. Moreover,
different array elements cover different communication directions such that not all elements are active at the
same time and the transmit power of an individual element does not have to be split between users. Thermal
issues within the array and decreasing EIRP with increasing number of users are therefore less likely. However,
phased-array-fed reflector-antennas are known to exhibit only a small beam-steering range of a few degrees.
Increasing the scan range without (over-)compromising on the advantages of this antenna type is therefore one
of the challenges within SILIKA. The experience and expertise of the SILIKA consortium on focal-plane
arrays with key academic partners Eindhoven University of Technology and Chalmers, and leading industrial
project partners NXP and Ericsson, will form the basis for this concept. At the symposium, the SILIKA concept
will be presented, supported by relevant results of past and ongoing research.
[1] “5G Radio Access,” Ericsson White Paper, April 2016 (http://www.ericsson.com/res/docs/whitepapers/wp-5g.pdf).
[2] World Radiocommunication Conference 2015, “Resolution 238; Studies on frequency - related matters for
International Mobile Telecommunications identification including possible additional allocations to the mobile
services on a primary basis in portion(s) of the frequency range between 24.25 and 86 GHz for the future
development of International Mobile Telecommunications for 2020 and beyond,” Tech. Rep. RESOLUTION
238(WRC-15), WRC-15, Geneva, October 2015.
[3] B. Sadhu et al., “7.2 A 28GHz 32-element phased-array transceiver IC with concurrent dual polarized beams and
1.4 degree beam-steering resolution for 5G communication, “ 2017 IEEE International Solid-State Circuits
Conference (ISSCC), 2017, 128-129.
[4] “Silicon-based Ka-band massive MIMO antenna systems for new telecommunication services (SILIKA), “ research
project funded by European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska
Curie grant agreement No 721732 (www.silika-project.eu).
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
6
INVITED TALK
The Development of Focal Plane Arrays in Radio Astronomy
Ron Ekers and John O'Sullivan
CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected]
I will discuss the scientific drivers for putting more than a single receiving element at the focus of a radio
telescope. This started with multibeam systems using independent detectors (as is still the case at mm
wavelengths). By the early 80s the concept of fully sampling the voltage at the focal plane of a parabolic dish
was being discussed and the first rudimentary systems for radio astronomy were built in the mid-90s. Despite
the challenging technology involving bandwidth and mutual coupling a number of groups were moving
forward spurred on by the many additional advantages which could be foreseen. I will summarise some of
these developments and future plans and also show a few of the recent exciting astronomical results which will
have a revolutionary impact on the future of radio astronomy.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
7
The Cloverleaf Antenna Array on CHIME
Meiling Deng for the CHIME collaboration
University of British Columbia, Vancouver, Canada
email: [email protected]
CHIME, the Canadian Hydrogen Intensity-Mapping Experiment, requires a linear array of wide bandwidth
feeds along its 100m-long focal lines.
A compact, wide-bandwidth, dual-polarization cloverleaf antenna [1] has been designed to maximize the
throughput of a single feed to a low noise amplifier (LNA). This cloverleaf antenna is now deployed and
operational as 1D 64-element array on each of the two 40m-long cylindrical reflectors of prototype radio
telescope for CHIME. Holography measurements [2] of the CHIME pathfinder beams have been measured
using the adjacent 26m Galt Telescope. The results are compared to simulations.
We have modified the cloverleaf design by considering the feeds as elements of an active array [3] matched
to the actual LNA deployed on CHIME. The feed properties have been varied to optimize signal-to-noise ratios
across the full band from 400 to 800 MHz. The modified cloverleaf antennas are now deployed as 1D 256-
element arrays on each of the four 100m-long cylindrical reflectors of CHIME, as shown in Figure 1.
For CHIME to achieve its scientific goals, knowledge of the calibration and beam-shape of the instrument are
required at the 0.1% level [4]. We will describe our design successes so far and also the challenges of matching
sky data to simulations in order to achieve our desired level of understanding instrument performance.
Figure 1. Left: One of CHIME’s cylindrical reflector with modified cloverleaf array deployed on its focal line. Right:
Zoom-in view of the modified cloverleaf array.
[1] M. Deng and D. Campbell-Wilson, "The cloverleaf antenna: A compact wide-bandwidth dual-polarization feed for
CHIME," 2014 16th International Symposium on Antenna Technology and Applied Electromagnetics (ANTEM),
Victoria, BC, 2014, pp. 1-2. doi:10.1109/ANTEM.2014.6887670.
[2] Philippe Berger ; Laura B. Newburgh ; Mandana Amiri ; Kevin Bandura ; Jean-François Cliche ; Liam Connor ;
Meiling Deng ; Nolan Denman ; Matt Dobbs ; Mateus Fandino ; Adam J. Gilbert ; Deborah Good ; Mark Halpern ;
David Hanna ; Adam D. Hincks ; Gary Hinshaw ; Carolin Höfer ; Andre M. Johnson ; Tom L. Landecker ; Kiyoshi
W. Masui ; Juan Mena Parra ; Niels Oppermann ; Ue-Li Pen ; Jeffrey B. Peterson ; Andre Recnik ; Timothy
Robishaw ; J. Richard Shaw ; Seth Siegel ; Kris Sigurdson ; Kendrick Smith ; Emilie Storer ; Ian Tretyakov ;
Kwinten Van Gassen ; Keith Vanderlinde ; Donald Wiebe; Holographic beam mapping of the CHIME pathfinder
array. Proc. SPIE 9906, Ground-based and Airborne Telescopes VI, 99060D (August 18, 2016);
doi:10.1117/12.2233782.
[3] R. Maaskant and E. E. M. Woestenburg, "Applying the active antenna impedance to achieve noise match in receiving
array antennas," 2007 IEEE Antennas and Propagation Society International Symposium, Honolulu, HI, 2007, pp.
5889-5892. doi: 10.1109/APS.2007.4396892.
[4] J. R. Shaw, K. Sigurdson, U.-L. Pen, A. Stebbins, and M. Sitwell, “All-sky Interferometry with Spherical Harmonic
Transit Telescopes,” The Astrophysical Journal 781, 57 (2014).
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
8
CryoPAF: a 2.8 – 5.18 GHz Cryogenic Phased Array Feed Receiver
Lisa Locke1, Vlad Reshetov1, Doug Henke1, Frank Jiang1, Gordon Lacy2, Bruce Veidt2, Gary Hovey2 1 NRC Herzberg, Victoria, BC, Canada
2 NRC Herzberg, Penticton, BC, Canada
email: [email protected]
Phased array feed (PAF) receivers for radio astronomy telescopes offer the promise of increased fields of view
by virtue of dense (λ/2) antenna element spacing and beamforming capabilities. Although the high noise
temperatures of present room temperature PAFs have prevented general adoption, by cryogenically cooling
the antenna and amplifier elements, competitive receiver temperatures are achievable combined with
significant increases in field of view, in comparison with single pixel receivers of equal system temperature.
NRC Herzberg is constructing a 2.8 – 5.18 GHz cryogenic phased array receiver, Figure 1(a), to demonstrate
the feasibility of higher frequency cryogenically cooled large PAFs. A composite laminate radome provides
vacuum containment to the 48 cm diameter cryostat. The large aperture requires multi-layered RF-transparent
infrared shields to reduce the significant thermal loading on the antenna array, and concentric cylindrical metal
shields at 16 K and 70 K provide necessary thermal insulation to the low noise amplifiers and coaxial cables
leading out of the cryostat. The 140 gold-plated copper Vivaldi antenna elements (96 active, 44 passive),
Figure 1(b, top), comprise a dual linear polarization array with centre-to-centre spacing of 28.8 mm. The
Vivaldi feeding structure consists of two parts; 1) a standard 50-ohm .086” diameter coaxial line and 2) a
rectangular air-coaxial line made from a copper foil centre conductor and surrounding air dielectric formed by
machining a groove into the metal antenna body. The antennas and amplifiers are cooled to 16 K. The low-
noise amplifiers, Figure 1(b, bottom), are designed in-house, and achieve a noise temperature of 3.5 K. External
to the cryostat for each of the 96 active receiver chains is a bandpass filter followed by a 35 dB post-amplifier.
A digital beamformer performs direct digital 8-bit sampling, 500 MHz instantaneous frequency band selection,
time-domain beamforming and array covariance matrix calculation.
Physical and geometrical optics modelling of the dual-reflector and PAF receiver have been analyzed,
including resulting farfield beams, overlap, and aberrations off axis. A 4-element prototype, as in Figure 1(c,d),
will verify antenna performance, machining capabilities, assembly processes, and effective thermal transfer to
the amplifier-antenna assembly. Full array testing will include hot-cold measurements and preliminary
beamforming at DRAO Penticton on DVA-1, Figure 1(f), a 15-m offset Gregorian dual-reflector telescope
with 55° half opening angle at the focus to secondary and a -16 dB feed edge taper. 18 far-field dual-
polarization beams, Figure 1(g), can be formed - 36 in total, The expected total receiver temperature is 11 K
and the increase in field of view compared with a single pixel receiver is 8x.
Figure 1. CryoPAF mechanical and beamforming details: a) mechanical assembly, b) cryogenic components, c) bullet-
style Vivaldi antenna details, d) 4-element prototype, e) ray tracing simulation for offset Gregorian, f) DVA-1, 15-m
reflector at DRAO Penticton for testing receiver, g) simulated far-field beams at 2.8 and 5.18 GHz.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
9
Update of PHased Arrays for Reflector Observing Systems
Lei Liu, Keith Grainge and Michael Keith
The University of Manchester, Manchester, UK
email: [email protected]
email: [email protected]
email: [email protected]
PHased Arrays for Reflector Observing Systems (PHAROS) is a C-band cryogenically cooled low noise
phased array feeds (PAFs) system which has been developed as part of a European technology demonstrator
project cooperated by JBO (UK), ASTRON (The Netherland), INAF (Italy) and MECSA (Italy). The
instrument will be mounted at the focus of the Lovell 76-m diameter telescope at JBO to perform radio
astronomy observation across the 4-8 GHz range. PHAROS consists of 220 elements Vivaldi array cooled to
20 K along with 24 low noise amplifiers (LNAs) mounted directly behind the active antenna elements. The
LNAs are followed by low-loss low thermal conduction RF connections to the analogue beam forming system
designed to operate at 77 K. The RF signals of the active elements are distributed to the beam formers by
passive splitters, while the non-active elements are terminated into 50 Ω loads. Four beam former modules are
available, each with 13 RF inputs and 13 individually controllable phase and amplitude control (PAC) units,
along with 13 amplifiers to make up for system losses. The last stage of beam forming is a 16-way Wilkinson
combiner (three inputs terminated). Each analogue beam former is responsible for the amplitude and phase
weightings of 13 elements in order to produce a single (compound) one-polarization beam.
All the elements of PHAROS were individually checked and the system was fully integrated. Following some
hardware and software debugging, beam pattern tests were performed at anechoic chamber at room
temperature, while the system noise temperature and gain were tested by hot/cold loads method which were
set up at both indoor and outdoor. In parallel, electromagnetic simulations are being finalized to find the
optimum PHAROS combined-beam properties for an efficient coupling of the instrument with the Lovell
telescope. The predictions based on the simulation tools are used for comparison with the beam pattern values
measured in the anechoic chamber.
In the framework of the Phased Array Feed SKA AIP we propose to upgrade PHAROS to a new instrument,
named PHAROS2, which will re-use some of parts of the existing PHAROS hardware. PHAROS2 is a
technology development program towards SKA Phase2 to demonstrate feasibility and competitiveness of high-
frequency PAF technology. PHAROS2 will be a cryogenically cooled C-band PAF demonstrator with digital
beam former with the following features and cooperated by JBO, ASTRON, INAF, and Chalmers University
of Technology (Sweden).
• New vacuum window.
• New antenna array.
• New cryogenic LNAs.
• A warm section, including down conversion from RF to baseband and LO distribution.
• Signal transportation with analogue RFoF optical links.
• A digital backend/beam former.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
10
A New 1Tbps Ethernet Streaming Data Interface Between FPGA and
GPU Based Digital Signal Processing Systems for a Phased Array Feed
Daniel George, John Tuthill, Andrew Brown, Chris Phillips, Ron Beresford, Mark Leach,
Tasso Tzioumis, Mia Baquiran , Grant A. Hampson and John D. Bunton
CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected]
In 2015 the Max Planck Institute for Radio Astronomy (MPIfR) contracted the Astronomy and Space Science
Group at the Australian Commonwealth Scientific and Industrial Research Organisation (CSIRO) to produce
an additional Phased Array Feed (PAF) receiver system, designed for the Australian Square Kilometre Array
Pathfinder (ASKAP). The additional PAF system is for use on the Effelsberg 100m diameter single-dish radio
telescope in the North Rhine-Westphalia region of Germany. ASKAP is a new 36-antenna aperture synthesis
imaging radio telescope array being built by the CSIRO in a remote and radio quiet area of Western Australia
[1]. ASKAP introduces a new approach in radio astronomy: the use of phased arrays at the focus of each dish
antenna in the telescope array, rather than the conventional single feed element. The PAF permits multiple
beams to be formed on the sky simultaneously at each antenna resulting in a dramatic increase in the
instantaneous Field-of-View (FoV) of the telescope [2]. The beams can overlap, which allows continuous FoV
to be formed. The large FoV advantages of the PAF system translate directly to the single-dish application,
however one of the key architectural changes from ASKAP to the single-dish configuration for MPIfR was
the replacement of the FPGA-based correlator hardware (used for synthesis imaging) with a commercial
network switch and GPU-based digital signal processing back-end as shown in Figure 1.
10GbE10GbE10GbE
Optical-to-electrical converters,
ADCs and polyphase filterbank
frequency channelizer
O/E
Narrowband beamformer
and 10G Ethernet
streaming data interface
FPGA Digital Receiver
Subsystem
FPGA Digital Beamformer
SubsystemPassive optical
cross-connect1
Cisco
Nexus
3164Q
9 servers, each with:
4 x GTX TitanX GPUs
2 x 6 core CPUs
512 GB RAM
SSDs + HDDs
LNAs Filters
RFoF
drivers
188-element checkerboard
antenna array
188 SMF
Focus Assembly
Electronics
2
3
4
5
6
7
8
9
GPU clusterNetwork switch
96
MMF
~1Tbps
Telescope Operating System (TOS) – real-time control and monitoring
Sample Clock and Timing distribution
GPU processing:
De-dispersion
Pulsar folding
Fast transient
detection
Figure 1. End-to-end signal chain for the MPIfR PAF system.
To support this new architecture, the firmware of the ASKAP FPGA-based PAF digital beamformer was
modified to provide real-time streaming of beam voltage data from the PAF in a new 10Gbps Ethernet format.
This paper provides an overview of the MPIfR single-dish PAF system and focusses on the design of the
FPGA-based 10GbE streaming data interface and a new Ethernet data format that transports PAF beam
voltages to a GPU compute cluster via a commercial network switch. The Ethernet output interface implements
a User Datagram Protocol (UDP) transmission model and we present a new, efficient Ethernet packet format
with a structure designed specifically to cater for streaming the multidimensional data that comes with PAF
receiver systems. The new format, termed CODIF (CSIRO Oversampled Data Interchange Format) is based
on the VLBI Data Interchange Format (VDIF) [3], however it has been substantially modified and extended
from this standard to allow for the transport of rich meta-data and to accommodate fractionally oversampled
data streams that are becoming more common in radio astronomy signal processing systems. The Ethernet
interface firmware has been designed with the flexibility to route each data stream through a commercial
Ethernet switch to any GPU end point. The switch provides a flexible data cross-connect function,
concentrating all bandwidth for a subset of beams to a given GPU end-point. The A total of 1728 data streams
deliver thirty-six 336 MHz dual polarized beams at a total raw line rate of 915 Gbps to the GPU processing
cluster with a granularity of 7MHz per stream.
[1] DeBoer, D., et al., “Australian SKA Pathfinder: A High-Dynamic Range Wide-Field of View Survey
Telescope,” in Proc. IEEE, Vol. 97, No. 8, 2009, pp. 1507 – 1521.
[2] Hotan, A. W. et al, “The Australian Square Kilometre Array Pathfinder: System Architecture and
Specifications of the Boolardy Engineering Test Array,” in Publications of the Astronomical Society of
Australia, 31, 2014, e041.
[3] VDIF Taskforce, VLBI Data Interchange Format (VDIF) Specification, http://www.vlbi.org/vdif/, 2009.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
11
Study of the Dish Optics for Phased Array Feed
Wu Yang1,2, Du Biao1,2, Niu Shengpu2 and Jin Chengjin1 1 Joint Laboratory for Radio Astronomy Technology, Beijing, China
email: [email protected] 2 The 54th Research Institute of CETC, Shijiazhuang, China
email: [email protected]
Phased Array Feed (PAF) is a powerful instrument for radio telescope, not only for further telescopes, but also
for existing ones. As a feed, the design of PAF has to be carried out in conjugate with the dish. For existing
telescopes, this issue has to be taken into account. This presentation analyses the effect of the optics on PAF,
including the antenna type, reflector shaping, and geometry parameters, and several examples will be provided.
A general description about the theory of the dish optics and feed is introduced at first. Taking parabolic
antenna as an example, the relationship among the aperture, frequency, and field-of-view are studied. Focal
field are introduced as a method of analyzing, and several principles are summarized.
Reflector shaping is a common way to achieve performance enhancement. There are two typical goals of
reflector shaping: single beam and multi-beam. The effect and methods of each goal for the reflector shaping,
as well as some results, will be presented.
As the third part, the optics of several radio telescope in China, which have a plan to build their own PAF will
be introduced, and some preliminary designs will be given.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
12
Phased Array Feed (PAF) Design for the LOVELL Antenna based on the
Octagonal Ring Antenna (ORA) Array
Ming Yang, Yongwei Zhang, Laith Danoon and Anthony Brown
University of Manchester, UK
email: [email protected]
This paper aims to investigate the possible potential application of wideband antenna technology using the
Octagonal Ring Antenna (ORA) [1] as a Phased Array Feed (PAF). As a design study the LOWELL reflector
antenna situated at the Jodrell Bank Observatory was considered, with the antenna covering frequency range
from 1GHz to 4GHz.
The original work on ORA is proposed as a planar array as part of the Square Kilometre Array (SKA) project
with a working frequency from 450MHz-1450MHz. The advantages it can bring are: 1) a broad bandwidth
with a frequency ratio of 4:1; 2) a planar structure with a low profile; 3) easy to manufacture using low cost
printed circuit techniques; 4) good scalability, which means the technique can be applied to other applications
of interest. This project has provided a preliminary technical assessment of the ORA application to feed the
LOWELL reflector antenna. This involves modelling of the ORA antenna array in the simulation software
package Computer Simulation Technology (CST) Microwave Studio [2], and simulation in software package
GRASP [3] to predict its performance when used to feed the Lowell antenna. The proposed ORA array is in a
circular shape with a diameter of 0.6m, as shown in Figure 1. To maximise the scan angle of the reflector, the
radiation pattern of the finite ORA array simulated by CST is compared at different frequencies to that of
typical horns, which are normally used to illuminate the reflector antenna. The proposed ORA array has
displayed good illumination efficiency. The radiation pattern of the LOWELL reflector is then simulated in
GRASP at different frequencies and compared with the pattern of the LOWELL reflector when fed by a horn
antenna. It should be noted that the proposed ORA array is simulated without any stage of practical
optimisation and it has demonstrated its potential use as a wideband PAF.
Figure 1. The proposed ORA array (three layers) to feed the LOVELL reflector antenna at Jodrell Bank.
[1] Y. Zhang and A. K. Brown, “Octagonal Ring Antenna (ORA) for a compact dual-polarised aperture array”, IEEE
Transactions on Antennas and Propagation, vol. 59, no 10, pp. 3927-3932, Oct. 2011.
[2] Microwave Studio. Darmstadt, Germany, Computer Simulation Technology Ltd. Bad Nauheimer Str. 19, D-64289.
[3] GRASP, DK-1119 Copenhagen K, Denmark, TICRA.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
13
Making Robust Beamformer Weights by Interpolating in the Frequency Domain
A. P. Chippendale 1, K. Bannister 1 and G. Hellbourg 2
1 CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected] 2 Department of Astronomy, University of California, Berkeley, CA, USA
email:[email protected]
Maximum-sensitivity beamformer weights are corrupted when radio-frequency interference (RFI), celestial
radio sources, or system noise are significant at the output of an antenna array when compared to the power
from the intended calibration source. This scenario is encountered when calibrating a phased array feed (PAF)
on a small reflector antenna with a celestial source that is weak compared to the system noise [1] and when
calibrating a PAF, even on a large antenna, in the face of strong RFI [2,3].
Valid maximum-sensitivity beamformer weights can be interpolated at interference affected frequencies from
weights calibrated at interference-free frequencies [3]. Figure 1 shows an example from [3] where
interpolation is used to correct interference-affected beamformer weights for a PAF installed on the Parkes 64
m telescope. In the current work we explore applying this correction to corrupted maximum-sensitivity
beamformer weights from a broader variety of astronomy scenarios, including the search for fast radio bursts
(FRBs) with the Australian Square Kilometre Array Pathfinder [4].
We explore the ability of interpolation to reduce estimation noise in Maximum-sensitivity beamformer weights
in the absence of RFI and discuss alternative techniques for making robust beam weights including correlation
with reference antennas [5].
Figure 1. Weight amplitudes affected by RFI (Left) and corrected by interpolation (Right), reproduced from [3].
[1] A. P. Chippendale et al., "Measured sensitivity of the first mark II phased array feed on an ASKAP antenna," 2015
International Conference on Electromagnetics in Advanced Applications (ICEAA), Turin, 2015, pp. 541-544. doi: 10.1109/ICEAA.2015.7297174.
[2] A. P. Chippendale et al., "Testing a modified ASKAP Mark II phased array feed on the 64 m Parkes radio
telescope," 2016 International Conference on Electromagnetics in Advanced Applications (ICEAA), Cairns,
QLD,
2016, pp. 909-912. doi: 10.1109/ICEAA.2016.7731550C.
[3] A. P. Chippendale, G. Hellbourg, “Interference mitigation with a modified ASKAP phased array feed on the 64 m Parkes radio telescope,” 2017 International Conference on Electromagnetics in Advanced Applications (ICEAA), Verona, Italy, 2017. To be published. arXiv:1706.04292.
[4] K. Bannister et al., “The detection of an extremely bright fast radio burst in a phased array feed
survey,” Astrophysical Journal Letters. To be published. arXiv:1705.07581.
[5] A. Chippendale, J. O'Sullivan, J. Reynolds, R. Gough, D. Hayman and S. Hay, "Phased array feed testing
for astronomy with ASKAP," 2010 IEEE International Symposium on Phased Array Systems and
Technology, Waltham, MA, 2010, pp. 648-652. doi: 10.1109/ARRAY.2010.5613298.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
14
Advancements in PAF Correlator and Beamformer Systems
G.A. Hampson1, J.D. Bunton, A. Brown, J. Tuthill and G. Schoonderbeek2 1 CSIRO Astronomy and Space Science, Marsfield, Australia email: [email protected]
2 ASTRON, Dwingeloo, The Netherlands email: [email protected]
Astronomy pushes the boundaries of digital signal processing into the realms of massively big data. CSIRO
is at the forefront of this technology, developing high data rate digital signal processing systems for the SKA
Low Correlator and Beamformer. This innovative technology, known as Gemini [1], could be applicable to
the next generation of PAF synthesis telescopes. Many FPGA advances have occurred since the design of
the second generation ASKAP system [2]. When the PAF receiver and beamformer are located at the antenna
then the central signal processing is simplified to a correlator and tied-array beamformer. In addition to
existing ASKAP PAF capabilities, astronomers always want “more!” For correlators – more bandwidth,
more inputs, shorter dump times, add transient buffers, more output frequency resolutions and greater
subarraying. For tied array beamformers – similar to correlators, but more beams and possibly beam
polarisation correction.
FPGAs and GPUs implementations each have advantages and disadvantages. An ASIC would be nice in
terms of power, but not ideal in a rapidly moving research area. Relentless data rates of >100 Tbps and
repetitive signal processing give FPGAs the upper hand over GPUs. PAF systems compound the issues
encountered in systems by generating orders of magnitude more data for real time processing. The ASKAP
correlator chassis, called Redback-3 [3] consists of six FPGAs each processing 0.66MHz of bandwidth. The
Gemini board being developed for SKA Low contains a single FPGA. Two generations of FPGA following
ASKAP has resulted in FPGAs with six times the resources and the ability to clock at twice the rate [4]. This
new technology is a subrack-based system holding up to twelve 4U processing cards. The subrack provides
pluggable Power/Liquid/Optical backplane interfaces (as shown in Figure 1). Each card is equipped with a
16 nm FPGA with 400 G front panel optics, 900 G optical backplane connectivity as well as a DDR4 memory
module and High Bandwidth Memory (HBM) internal to the FPGA.
Figure 1. “Gemini” – the new DSP FPGA board for SKA Low Correlator and Beamformer.
Next generation PAF signal processing systems reduce physical size (one third the volume), power
requirements are lower (as well as more efficient liquid cooling) and the system becomes compute limited
(not IO bound like GPUs). One surprising factor is that the cost is relatively similar as we see a slowing of
Moore’s law. A significant reduction in the number of processing units, combined with HBM memory,
enables systems to be implemented more rapidly.
[1] E. Kooistra et. al., “Gemini FPGA Hardware Platform for the SKA Low Correlator and Beamformer”, 2017 URSI
General Assembly and Scientific Symposium (GASS), August 2017.
[2] G.A. Hampson, et. al., “ASKAP PAF ADE advancing an L-band PAF design towards SKA”, International
Conference on Electromagnetics in Advanced Applications (ICEAA), September 2012.
[3] G.A. Hampson, et. al., “ASKAP Redback-3 — An agile digital signal processing platform”, 2014 XXXIth URSI
General Assembly and Scientific Symposium (URSI GASS), August 2014.
[4] “Xilinx Ultrascale+ FPGAs”, https://www.xilinx.com/support/documentation/product-briefs/virtex-ultrascale-plus-
product-brief.pdf.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
15
Beamformer and Calibration Performance for the Focal-plane
L-band Array feed for the Green Bank Telescope (FLAG) B. D. Jeffs
1, K. F. Warnick
1, R. A. Black
1, M. Ruzindanna
1, M. Burnett
1, D. J. Pisano
2, D. R. Lorimer
2, N.
Pingel2, K. Rajwade
2, R. M. Prestage
3, S. White
3, B. Simon
3, L. Hawkins
3, W. Shillue
4, A. Roshi
4
1Brigham Young University, Provo UT, USA, email: [email protected]
2West Virginia University, Morgantown WV, USA, email: [email protected]
3Green Bank Observatory (GBO), Green Bank WV, USA, email: [email protected] 4National Radio Astronomy Observatory (NRAO) CDL, Charlottesville VA, USA,
email: [email protected]
We have developed and demonstrated the first permanent cryogenically cooled L-band phased array feed
(PAF) for a major single dish radio telescope. This new instrument for the Green Bank Telescope underwent
final commissioning tests in May and July 2017 and is now available for preliminary science observations.
This presentation will highlight results and performance analysis from these tests, including beamformer
calibration methods, radio camera snapshot imaging, beampattern measurements (as seen in Fig. 1), and both
real-time transient survey beamforming and post-correlation beamforming for HI observations. This final
commissioning run utilized the full complement of digital back end processor units for a total bandwidth
coverage of 150 MHz. All operational modes were demonstrated. The design approach and system
architecture for the frequency channelization and back end real-time digital beamformer/correlator/
spectrometer will be presented. A companion paper will also be presented by Anish Roshi to report on array
electromagnetic performance, including system temperature and field of view sensitivity.
Several technical innovations were introduced in the FLAG design. We will discuss the real-time array
correlator which runs concurrently with the real time digital beamformer. With FLAG’s unique post-
correlation beamforming capability, short-term integrated array covariance matrices are stored for each fine
or coarse frequency channel as the primary final data product, rather than beamformed spectra. This enables
the observation data set to be revisited at a later date using any desired set of beamformer weights for post-
correlation beamforming. With this approach one can, “after the fact,” address data defects (such as RFI)
and study possible resolution enhancements or weak source detection improvements with any desired test
beampattern. The design also includes a seven-beam real-time beamformer which runs concurrently with the
real-time correlator. This enables commensal transient searches with the coarse frequency channelized rapid
dump beamformed spectrometer, along with fine channelized post-correlation-beamformed HI observations.
This architecture is also ideal for adaptive beamforming RFI mitigation.
Figure 1. Measured GBT on-sky FLAG beampatterns for seven simultaneous beams at 1404 MHz, X polarization, May
2017. A bright reference sky source is used by the correlator to calibrate beamformer weights. The GBT is then steered
around this source in a grid pattern to scan out the resulting beam response pattern.
This work is supported by the NRAO, GBO, and National Science Foundation award no. AST-1309832.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
16
Specifying Polarisation Properties for Radio Telescopes
Bruce Veidt
NRC Herzberg, Penticton, Canada
email: [email protected]
Accurate polarisation measurements are important for radio astronomy observations as diverse as pulsar timing,
Zeeman splitting, and studies of the interstellar medium. It is therefore essential that future radio telescopes
have their polarisation properties properly specified so that observational errors are minimised. There has been
recent effort in defining a new figure of merit which is independent of the coordinate system used for the
measurements [1]. This figure of merit is called the Intrinsic Cross-polarisation Ratio or IXR.
For adoption of this new parameter both antenna engineers and radio astronomers must understand it so that
error budgets can be translated into this new figure of merit. This presentation will begin by explaining why a
new parameter is needed. Then the error relationships will be derived in detail. The presentation will conclude
with examples based on simple wire antennas and on reflector antennas.
[1] T. D. Carozzi and G.A. Woan, “A Fundamental Figure of Merit for Radio Polarimeters,” IEEE Transactions on
Antennas and Propagation, vol. 59, pp. 2058 – 2065, 2011.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
17
Results from the Astronomical Commissioning of APERTIF
Tom Oosterloo1,2, Apertif Team 1 ASTRON – Netherlands Institute for Radio Astronomy
PO Box 2, Dwingeloo, The Netherlands
email: [email protected] 2 Kapteyn Institute, Groningen University, Groningen, The Netherlands
We will present results from the astronomical commissioning of Apertif, the Phased-Array Feed upgrade of
the Westerbork Synthesis Radio Telescope (WSRT). Apertif will increase the instantaneous field of view of
the WSRT by a factor of 37 and its observing bandwidth to 300 MHz. The system will operate in the frequency
range 1130-1700 MHz and it will allow to image the sky with high spectral (2.5 km/s) and spatial (15 arcsec).
Apertif will turn the WSRT into an effective survey telescope.
A number of large surveys are planned with scientific applications such as deep imaging surveys of the
northern sky of HI and OH emission, of the polarised continuum, and efficient searches for pulsars and
transients. Such surveys will detect the HI in more than 100,000 galaxies out to z = 0.2, will allow determining
the detailed structure of the magnetic field of galaxies (including our own), and will discover more than 1,000
pulsars and other types of transients.
Since the beginning of 2017, astronomical observations have been performed with Apertif. We present the
results of the astronomical commissioning of Apertif which is aimed to validate the system from the
astronomer’s point of view. I will discuss the results in terms of the opportunities and limitations they imply
for the surveys planned for Apertif.
The figure gives an example of what I will present. The figure shows the outcome of one of the commissioning
observations. The contours in the left panel give the observed column density of the atomic hydrogen in the
galaxy M51 superposed on an optical image of this galaxy. The right panel shows the continuum image of
M51 obtained from the same observation.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
18
New-Generation Radio Telescopes: Challenges in Instrumental Calibration
Wasim Raja,
CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected]
In this talk we present a brief overview of various kinds of instrumental non-idealities and how these are
addressed in current calibration schemes. We highlight the specific challenges posed by new-generation
telescopes like the ASKAP. Results from a few novel techniques addressing some of these challenges are
presented.
Instrumental calibration using rotating feeds/dishes: The quantities on the left hand side of the equation
below are the measured Stokes parameters for a linearly polarised calibrator. Here, mij are the elements of the
instrumental Mueller matrix, dp is the degree of linear polarisation of the calibrator and χ is the relative
polarisation position angle of the calibrator w.r.t the instrument feeds. G(t,ν) is the gain of the instrument.
(
𝐼𝑄𝑈𝑉)
= 𝐺(𝑡, 𝜈)
(
𝑚11𝑚21𝑚31𝑚41
𝑚12𝑚22𝑚32𝑚42
𝑚13𝑚23𝑚33𝑚43
𝑚14𝑚24𝑚34𝑚44)
(
𝐼0𝑑𝑝𝐼0 cos 2𝜒
𝑑𝑝𝐼0 sin2𝜒
0 )
If we observe the calibrator by rotating the feeds such that χ changes from –π/2 to +π/2, we will have measured:
𝐼(𝜒) = 𝑚11 + 𝑚12𝑑𝑝 cos2𝜒 + 𝑚13𝑑𝑝 sin 2𝜒
𝑄(𝜒) = 𝑚21 + 𝑚22𝑑𝑝 cos 2𝜒 + 𝑚23𝑑𝑝 sin2𝜒
𝑈(𝜒) = 𝑚31 + 𝑚32𝑑𝑝 cos 2𝜒 + 𝑚33𝑑𝑝 sin2𝜒
𝑉(𝜒) = 𝑚41 + 𝑚42𝑑𝑝 cos 2𝜒 + 𝑚43𝑑𝑝 sin2𝜒
Fourier transforming the measured Stokes parameters (as a function of χ) will readily filter out elements from
the 1st column of the Mueller matrix in the zeroth Fourier components, while the 2nd and the 3rd columns will
separate out in to the real and imaginary parts of the 1st Fourier components respectively. The elements of the
4th column can be determined from the derived elements since they are not all independent.
[1] http://www.atnf.csiro.au/observers/memos/d93780 1.pdf.
[2] Bhatnagar, S., & Nityananda, R. 2001, A&A, 375, 344.
[3] Conway, R. G., & Kronberg, P. P. 1969, MNRAS, 142, 11.
[4] Heiles, C., Perillat, P., Nolan, M., et al. 2001, Publications of the Astronomical Society of the Pacific, 113, 1274.
[5] Sault, R. J., Hamaker, J. P., & Bregman, J. D. 1996, A&AS, 117, 149.
[6] Raja, Wasim (2014), Faraday Slicing Polarized Radio Sources DOI: 10.13140/RG.2.2.31928.14080.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
19
System Performance Characterisation of WSRT-APERTIF
Boudewijn Hut and Wim van Cappellen
Netherlands Institute for Radio Astronomy (ASTRON), Dwingeloo, The Netherlands
email: [email protected]
The APERTIF (APERture Tile In Focus) project has installed Phased Array Feeds (PAFs) in the reflector
antennas of the Westerbork Synthesis Radio Telescope (WSRT). These PAF systems can simultaneously form
37 compound beams on the sky and replaced the single beam horn feeds. The survey speed of the WSRT will
be improved significantly, bringing down the time to execute a deep wide-field astronomical survey by a factor
17. The design phase is completed and samples of final hardware have been tested. Twelve WSRT dishes will
have a PAF system, which are already equipped with final hardware. These 12 dishes are now used to
characterize and fine-tune the system in detail.
An extensive characterization of the APERTIF system performance is part of the commissioning effort. A
series of measurements is being performed to characterize among other things bandpass stability, polarization
properties, sensitivity and temporal stability, all as function of frequency and scan angle. To keep the beam
shape stable over time, a dedicated online-calibration scheme is developed [1]. The compound beams are
electronically steered by the beam former. The weight amplitude to form a central compound beam is shown
in Figure 1. The sensitivity of such a compound beam as function of scan angle is shown in Figure 2. It also
shows the lower sensitivity of a single PAF receiver chain. We will present more results of the system
performance characterization, showing the capabilities of the APERTIF system.
Figure 1. Amplitude of weights to form an on-axis X-polarised compound beam. 61 PAF receiver chains are weighted.
Figure 2. Measured sensitivity as function of scan angle, for compound beams and for the central PAF element beam.
[1] B. Hut, S.J. Wijnholds, W.A. van Cappellen, “Online Calibration Scheme for APERTIF,” Phased Array Feed
Workshop, Cagliari, Italy, 2016.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
20
Status of the AFAD-C Project
Bruce Veidt1, Tom Burgess2, Aaron Beaulieu3, Leo Belostotski3 and James Haslett3
1 NRC Herzberg, Penticton, Canada
email: [email protected] 2 NRC Herzberg, Penticton, Canada
3 University of Calgary, Calgary, Canada
The Advanced Focal Array Demonstrator (AFAD) project has the main goal of exploring the sensitivity of L-
band (0.75–1.5 GHz) phased-array feeds (PAFs) at room temperature using techniques to minimise loss in the
antenna element structure. The initial array had 41 elements using commercially-available low-noise amplifier
(LNA) integrated circuits (ICs) [1]. A subset of nine elements at the centre of the array were replaced with
elements using CMOS LNAs [2] and had excellent wide-band noise performance [3]. This justified
construction of a larger array using CMOS LNAs (AFAD-C), shown in Figure 1. This presentation will report
recent results as well as practical aspects, such as a simplified method for antenna element fabrication.
Figure 1. The 60-element AFAD-C array under assembly.
[1] B. Veidt, T. Burgess, K. Yeung, S. Claude, I. Wevers, M. Halman, P. Niranjanan, C. Yao, A. Jew, and A.G. Willis,
"Noise Performance of a Phased-Array Feed Composed of Thick Vivaldi Elements with Embedded Low-Noise
Amplifiers," 9th European Conference on Antennas and Propagation (EuCAP), 2015.
[2] A. Beaulieu, G. Wu, L. Belostotski, J.W. Haslett, T. Burgess, and B. Veidt, "Development of a CMOS receiver for
a radio-telescope phased-array feed," IEEE MTT-S International Microwave Symposium (IMS), 2016.
[3] A. Beaulieu, L. Belostotski, T. Burgess, B. Veidt, and J. Haslett, "Noise Performance of a Phased-Array Feed with
CMOS Low-Noise Amplifiers," IEEE Antennas and Wireless Propagation Letters, vol. 15, pp.1719-1722, 2016.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
21
A Highly Sensitive Focal L-Band Phased Array Feed for the Green Bank
Telescope: Electromagnetic Model and Results
D. Anish Roshi1, W. Shillue1, B. Simon2, S. White2, R. Prestage2, J. Diao3, K. F. Warnick3, B. Jeffs3, D. J. Pisano4, D. Lorimer4, T. Boyd1, J. Castro1, J. R. Fisher1, W. Groves1, M. Morgan1, L. Hawkins2, L. Jensen2,
J. D. Nelson2, J. Ray2, V. van Tonder2
1National Radio Astronomy Observatory (NRAO), Charlottesville, VA, USA
email: [email protected] 2Green Bank Observatory (GBO), Green Bank, WV, USA
email: [email protected] 3Brigham Young University (BYU), Provo UT, USA,
email: [email protected] 4West Virginia University (WVU), Morgantown, WV, USA
email: [email protected]
We will present: (a) the development of a new cryogenic Focal-plane L-band (1.4 GHz) Array Feed for
Green Bank Telescope (FLAG), (b) its measured performance on the GBT during the first commissioning
test on March 2017, and (c) the comparison of the measured performance with predictions from a phased
array feed (PAF) model developed at NRAO. For the boresight beam, the measured Tsys/was 25 K at 1350
MHz, which is comparable to the performance of the existing L- band single feed system of the Green Bank
Telescope (GBT). The degradation of Tsys/ at 5 arcmin offset (i.e. at the half power beamwidth at 1350
MHz) from the boresight is only ~ 1 % of the boresight value. With this performance, the survey speed of
the PAF system is higher by a factor of seven compared to the single feed system of the GBT. Following
these successful commissioning tests, the PAF was tested again on the GBT in May and July 2017 with a
newly developed real-time digital beamformer, and a correlator spectrometer. Initial results from this testing
will be presented separately by Jeffs et al. in a companion paper.
The FLAG PAF project developed several new assemblies: an optimized dipole design, improved low-
noise amplifiers (LNAs), integrated 40-channel downconverter-digitizer, and a Roach2 FPGA backend
processing system. The PAF consists of a hexagonal array of 19 dual polarized dipole elements. The dipole
shape and the spacing between the elements (15cm) were globally optimized to maximize the survey speed,
after taking into consideration the requirements of the GBT optics,
field-of-view, desired sensitivity, and impedance matching to LNAs. The frequency response of the
array was optimized for 150 MHz bandwidth centered at 1350 MHz. To achieve the lowest possible receiver
noise temperature, newly designed silicon-germanium LNAs were operated at 20 K with a noise temperature
of 5 K and an average gain of 38 dB. Following the LNAs, there is a highly integrated electronic assembly
located in close proximity to the front-end. This assembly consists of five “blades” with eight channels each,
and performs several functions: calibration signal injection, warm post-amplification, power leveling, LO
distribution, down conversion, digitization, and serial data transmission through optical fiber. An
unformatted high-speed digital link is employed, with the bit and word boundaries recovered by custom
algorithms in the Roach2-based FPGAs. The
data are then passed through a 512 channel polyphase filter bank (PFB), which provides a spectral
resolution of 300 KHz. For these first commissioning tests, the PFB outputs were packetized and sent
by 10 GbE through an Ethernet switch, and the data was recorded for offline processing.
A detailed electromagnetic and circuit model has been developed to account for the dipole response in the
array and GBT environment, and to include the mutual coupling and LNA impedance effects. Details of
the model and the comparison of the model prediction versus the measured performance will be presented,
both as a function of frequency and as a function of offset from boresight.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
22
PHAROS2 Phased Array Feed: Warm Section, Signal Transportation
and iTPM Digital Backend
A. Navarrini1, J. Monari2, A. Melis1, R. Concu1, A. Scalambra2, A. Maccaferri2, A. Cattani2, P. Ortu1,
G. Naldi2, J. Roda2, F. Perini2, G. Comoretto3, M. Morsiani2, A. Ladu1, S. Rusticelli2, A. Mattana2,
L. Marongiu2, M. Schiaffino2, E. Carretti1, A. Saba1, F. Schillirò4, E. Urru1, G. Pupillo2,
K. Zarb Adami5,6, A. Magro5, R. Chiello6 1 INAF-Osservatorio Astronomico di Cagliari, Selargius, Italy
email: [email protected] 2 INAF-Istituto di Radioastronomia, Bologna, Italy
3 INAF-Osservatorio Astrofisico di Arcetri, Florence, Italy 4 INAF-Osservatorio Astrofisico di Catania, Catania, Italy
5 University of Malta, Malta 6 University of Oxford, Department of Physics, UK
PHAROS is a C-band cryogenically cooled low noise Phased Array Feed (PAF) designed to be installed at the
focus of a large single-dish radio telescope. The upgrade of the PHAROS instrument, named PHAROS2, is
re-using parts of the existing PHAROS hardware, including the cryostat and the 220-element Vivaldi array
cooled at 20 K.
The PHAROS2 instrument is under development in the framework of the PAF SKA Advanced Instrumentation
Program; it is a collaboration among the following institutions: INAF (Italy), the University of Manchester
(UK), ASTRON (the Netherlands), OSO (Sweden) and the University of Malta (Malta).
PHAROS2 will be a C-band (4-8 GHz) PAF demonstrator with digital beamformer capable of synthesizing
four independent single-polarization beams by combining 24 active antenna elements (one compound beam
will employ 13 elements).
We report on the contribution of INAF and of the University of Malta to the development of the PHAROS2
instrument hardware, in particular:
- the “Warm Section” (WS), which will be used to perform the signal filtering, conditioning and single
down-conversion of a 275 MHz slice of the RF band down to the 375-650 MHz IF band;
- the signal transportation from the WS to the backend room by means of analogue WDM (Wavelength
Division Multiplexing) fiber-optic links, which allows transportation of two IF signals over the same
fiber (IFoF);
- the digital backend, based on iTPMs (Italian Tile Processing Modules), a digital platform developed
in Italy for backend of new generation SKA Aperture Arrays capable of digitizing 32 analog inputs.
Two different beamformer versions will be implemented: a narrow band version on GPUs (up to
16 MHz, suitable for high-resolution spectroscopy), and a wide band version implemented on iTPM-
FPGAs (275 MHz bandwidth). The architecture of the iTPM is conceived to be scalable to a large
number of beams (>30) and large bandwidth (>2GHz).
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
23
An S-band Cryogenic Phased Array Feed for the LOVELL Telescope:
Cryogenics
Simon J. Melhuish, Michael Keith and Mark McCulloch
The University of Manchester, Manchester, UK
email: [email protected]
In this talk we will review our progress, from a
mechanical and cryogenic engineering perspective, in
developing an S-band phased array feed (PAF) for the
Lovell Telescope at the Jodrell Bank Observatory (for the
RF aspects please see an accompanying talk [1]). We
face various challenges presented by this mid-century
telescope.
Whilst we think we can realize a large-diameter
(800 mm) window using foam, a conventional
arrangement (see figure) may be overly limiting in solid-
angle seen by the feed elements. The Lovell primary
subtends a full width of approximately 160° at the focus.
We may be considering a domed window similar to [2].
Although we have a high-capacity Leybold GM cooler
we are not complacent about infra-red loading through
the window. To minimize this we are investigating
various surface treatments and IR filter designs, for
example RF-transparent MLI [3].
Many of the low-noise amplifiers considered for this
application have a significant power dissipation. For a
complete 200-channel (100 elements × 2 polarizations)
system 100 mW per LNA, for example, would
overwhelm even our large cooler. Reducing this by an
order of magnitude [4] we then have also to consider other
loads such as cold coaxial cables.
Finally there is the weight to consider. Getting a receiver to the focus takes a route through the telescope
structure using several different winches. These had a limit of 250 kg, but are being upgraded for the ASKAP
room-temperature PAF. But still we are going to be close to the maximum weight and may have to design the
system to be split into two for winching. Also the focus box has limited clearance for insertion of the receiver,
further restricting us to 800-mm diameter.
[1] M. McCulloch, M. Keith, S. J. Melhuish, "An S-band cryogenic phased array feed for the Lovell Telescope: RF”,
Phased Array Feed Workshop, Sydney, Australia, 2017.
[2] L. Locke, D. Garcia, M. Halman, D. Henke, G. Hovey, N. Jiang, L. Knee, G. Lacy, D. Loop, M. Rupen, B. Veidt R.
Wierzbicki, "CryoPAF4: a cryogenic phased array feed design", Proc. SPIE 9914, Millimeter, Submillimeter, and
Far-Infrared Detectors and Instrumentation for Astronomy VII, Edinburgh, UK, 2016 AFBR-1310Z.
[3] J. Choi, H. Ishitsuka, S. Mima, S. Oguri, K. Takahashi and O. Tajima, “Radio-transparent multi-layer insulation for
radiowave receivers”, Review of Scientific Instruments 84, 114502 (2013); doi:
http://dx.doi.org/10.1063/1.4827081.
[4] LNF-LNC1.5_6A, Low Noise Factory, Goteborg, Sweden, www.lownoisefactory.com, last accessed 14-07-2017.
Figure 1. A conventional Dewar layout for a large-
diameter cryogenic PAF. Here the can diameter is
800 mm. But can we get a wide enough “view” at the
array adequately to illuminate the Lovell primary?
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
24
A Proposed Cryogenic PAF for Parkes
James Green
CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected]
The Parkes radio telescope is the quintessential instrument of Australian scientific endeavor, with a
fundamental role in conducting world-class research and maintaining broad public interest in physics and
astronomy. In 1997 a 13 beam receiver was built and deployed leading to two decades of high impact science.
The time is now right for the next revolutionary step, a novel cryogenically cooled Phased Array Feed (cryo-
PAF), to survey the southern sky for neutral hydrogen, molecules, pulsars, fast radio bursts and cosmic rays
more efficiently and with more accurate positional determination than previously possible. Compared with
existing instrumentation, this project will lead to significant improvement in six key areas: (1) an improved
receiver noise; (2) a wider field of view; (3) Nyquist sampling of the focal plane; (4) a wide front-end
bandwidth; (5) greater aperture efficiency; and (6) reduced baseline ripple. Taken together, these
improvements represent a 10 to 30-fold increase in survey speed compared with the existing multibeam
receiver, which represents a truly transformational improvement.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
25
The SKA Observatory Development Programme and Phased Array Feed
Development
M. A. Bowen
Square Kilometre Array Organisation, Macclesfield, Cheshire, UK
email: [email protected]
The SKA will be the world’s largest radio telescope. An observatory such as the SKA will be heavily reliant
upon state-of-the-art systems to ensure that it continues to produce world class science throughout the planned
50 year lifetime. To support this, ongoing technology upgrades will be necessary and the SKA will need to
support an active technology development programme.
The SKA project is currently in the design phase (pre-construction) with construction of the first phase of the
SKA (SKA1) scheduled to commence in 2019. Phased Array Feed (PAF) development is one of three research
and development projects running during the current pre-construction phase, which collectively comprise the
SKA Advanced Instrumentation Programme (AIP). PAFs are one of the possible suite of feed and receiver
systems suitable for deployment on the SKA1-MID antenna, consequently the SKA1 dish is being designed
to accommodate installation of one or more PAF systems.
SKA1 comprises approximately 10% of the full SKA (SKA2) and is being developed within a cost cap. This
has led to the project restricting capabilities during initial construction with the expectation that they will be
incorporated after initial SKA1 construction or as part of SKA2. SKA research and development activities will
need to encompass more than just instrumentation development and continue beyond SKA1 construction. To
support this ambition SKA research and development activities, including the current AIP, will be recast to
become the SKA Observatory Development Programme (SODP).
The SKA board has approved the framework under which the SODP would be run and has agreed that the
SODP should move from outside of the core of the SKA project, as the current AIP, to something that lies
largely within the SKA scope [1]. Funding to begin this formal, SKA funded activity will not be available until
SKA construction commences and large development projects are likely to be co-funded by SKA partner
institutions. The existing AIP will continue until SKA construction commences and is expected to transition
into the SODP once it is operating.
The SKA Organisation is continuing to develop the framework for the SKA development program. The current
SKA AIP members are actively engaged in shaping the SODP [2]. A progress update will be provided on the
developing SKA ODP framework and the scope of the development projects that could exist within it.
[1] A. M. McPherson, “SKA Observatory Development Programme Policy,” SKA Organisation, Macclesfield,
Cheshire, UK, Policy SKA-TEL-SKO-00000309, Nov. 01, 2015.
[2] ASTRON. (2017, June 11). SKA AIP Meeting, [Online]. Available: http://www http://www.astron.nl/ska-aip2017/ /programme.php.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
26
INVITED TALK
Astronomy with PAFs on ASKAP
David McConnell
CSIRO Astronomy and Space Science, Marsfield, Australia
email: [email protected]
ASKAP will be a 36-antenna array equipped with Phased Array Feeds operating over the 0.7 – 1.8 GHz
frequency range. Through the construction of ASKAP from the prototype BETA array with Mark I PAFs to
the current 12 antenna array with Mark II PAFs, an expanding range of astronomy has been possible. The
BETA array had only six antennas and PAFs with relatively poor radiometric performance above 1GHz, but
useful science was done with the first blind detection [1] of a redshifted HI absorption line at 985 MHz (z =
0.44). ASKAP-12, with Mark II PAFs, is a much more capable telescope and the suite of astronomy has grown
to include observations of HI in external galaxies and in the SMC, continuum studies of variable sources and
a search for Fast Radio Bursts (FRB) that has been very successful [2].
As well as describing ASKAP’s astronomical results, I will outline the procedures used for ASKAP operations
and data reduction. I will also mention the practical issues yet to be overcome to realise the full potential of
Phased Array Feeds.
Figure 1. The 21cm hydrogen emission from the Small Magellanic Cloud observed by ASKAP
imaged by Helga Dénes and Naomi McClure-Griffiths.
[1] J. R. Allison, et al., Discovery of H I gas in a young radio galaxy at z = 0.44 using the Australian Square Kilometre
Array Pathfinder, MNRAS, vol. 453, pp.1249-1267, 2015.
[2] K.W. Bannister et al. The Detection of an Extremely Bright Fast Radio Burst in a Phased Array Feed Survey, ApJL
vol 841, L12, 2017.
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
27
Author Index
A
Adami K.Z. ........................................................................... 22
B
Bannister K. .......................................................................... 13 Baquiran M ........................................................................... 10 Beaulieu .............................................................................. 20 Beaulieu A. .......................................................................... 20 Bekhti N.B. ............................................................................. 4 Belostotski L. ....................................................................... 20 Beresford R. ................................................................. 1, 2, 10 Biao D. ................................................................................. 11 Black R.A. ............................................................................. 15 Bowen M. ............................................................................ 25 Boyd T.................................................................................. 21 Bressner T.A.H. ...................................................................... 5 Brown A. .............................................................. 1, 10, 12, 14 Bunton J.D. ............................................................ 1, 2, 10, 14 Burgess T. ............................................................................ 20 Burnett M. ........................................................................... 15
C
Carretti E. ............................................................................ 22 Castro J. ............................................................................... 21 Cattani A. ............................................................................. 22 Cheng W. ............................................................................... 2 Chengjin J. ........................................................................... 11 Chiello R. ............................................................................. 22 Chippendale A. ...................................................................... 2 Chippendale A.P. .................................................................. 13 Comoretto G. ....................................................................... 22 Concu R. .............................................................................. 22
D
Danoon L. ............................................................................ 12 Deng M. ................................................................................. 7 Diao J. .................................................................................. 21
E
Ekers R. .................................................................................. 6 Elsakka A.A.H.M. ................................................................... 5
F
Ferris D. ................................................................................. 2 Fisher J.R. ............................................................................ 21 Froehlich A. ........................................................................... 4
G
George D. .............................................................................. 10 Grainge K. .............................................................................. 9 Green J. ............................................................................... 24 Groves W. ............................................................................ 21
H
Hampson G.A. ........................................................ 1, 2, 10, 14 Haslett J. .............................................................................. 20 Hawkins L. ...................................................................... 15, 21 Hellbourg G. ......................................................................... 13 Henke D. ................................................................................ 8 Hovey G. ................................................................................ 8 Hut B. ................................................................................... 19
I
Ivashina M.............................................................................. 5
J
Jeffs B. .................................................................................. 21 Jeffs B.D. .............................................................................. 15 Jensen L................................................................................ 21 Jiang F. ................................................................................... 8 Johannsen U. ......................................................................... 5
K
Kanapathippillai J. .................................................................. 2 Keith M. ........................................................................... 3, 23 Koenig F. ................................................................................ 4
L
Lacy G. .................................................................................... 8 Ladu A. ................................................................................. 22 Leach M. ............................................................................... 10 Liu L. ....................................................................................... 9 Locke L. .................................................................................. 8 Lorimer D. ............................................................................ 21 Lorimer D.R. ........................................................................ 15
M
Maaskant R. ........................................................................... 5 Maccaferri A. ....................................................................... 22 Magro A. .............................................................................. 22 Marongiu L. .......................................................................... 22 Mattana A. ........................................................................... 22 McConnell D. ....................................................................... 26 McCulloch M. ................................................................... 3, 23 Melhuish S.J. .................................................................... 3, 23 Melis A. ................................................................................ 22 Monari J. .............................................................................. 22 Morgan M. ........................................................................... 21 Morsiani M........................................................................... 22
N
Naldi G. ................................................................................ 22 Naumann K. ........................................................................... 4 Navarrini A. .......................................................................... 22 Nelson J.D. ........................................................................... 21
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
28
O
Oosterloo T. ......................................................................... 17 Ortu P. ................................................................................. 22 O'Sullivan J. ........................................................................... 6
P
Perini F. ............................................................................... 22 Phillips C................................................................................ 10 Pingel N. .............................................................................. 15 Pisano D.J ............................................................................ 21 Pisano D.J. ........................................................................... 15 Prestage R. .......................................................................... 21 Prestage R.M. ...................................................................... 15 Pupillo G. ............................................................................. 22 Putselyk S .............................................................................. 4
R
Rahlf F ................................................................................... 4 Raja W. ................................................................................ 18 Rajwade K............................................................................ 15 Ray J. ................................................................................... 21 Reshetov V. ........................................................................... 8 Roda J. ................................................................................. 22 Roshi A. ............................................................................... 15 Roshi D.A. ............................................................................ 21 Rusticelli S. .......................................................................... 22 Ruzindanna M. .................................................................... 15
S
Saba A. ................................................................................. 22 Scalambra A. ........................................................................ 22 Schiaffino M. ....................................................................... 22 Schillirò F. ............................................................................ 22
Schoonderbeek G. .................................................................. 14 Shengpu N. ........................................................................... 11 Shillue W. ....................................................................... 15, 21 Simon B ................................................................................ 21 Simon B. ............................................................................... 15 Smolders A.B. ......................................................................... 5
T
Tiesing M................................................................................ 4 Tuthill J. ...................................................................... 1, 10, 14 Tzioumis T. ............................................................................ 10
U
Urru E. .................................................................................. 22
V
van Cappellen W. ................................................................. 19 van Tonder V. ....................................................................... 21 Veidt B. ...................................................................... 8, 16, 20
W
Warnick K.F. ................................................................... 15, 21 White S. ......................................................................... 15, 21
Y
Yang M. ................................................................................ 12 Yang W. ................................................................................ 11
Z
Zhang Y. ............................................................................... 12
Phased Array Feed Workshop, 14-16 November 2017, Sydney, Australia
29
Affiliation Index
ASTRON, Dwingeloo, The Netherlands ..................................................................................................................... 14, 17, 19
Brigham Young University, Provo UT, USA ................................................................................................................ 15, 21
Chalmers University, Gothenburg, Sweden ........................................................................................................................ 5 CSIRO, Astronomy and Space Science, Marsfield, Australia ................................................... 1, 2, 6, 10, 13, 14, 18, 24, 26 Department of Astronomy, University of California, Berkeley, CA,USA .............................................................................. 13 Eindhoven University of Technology, Eindhoven, The Netherlands ................................................................................... 5 Fraunhofer Institute for High Frequency Physics and Radar Techniques FHR, Wachtberg, Germany ............................... 4
Green Bank Observatory (GBO), Green Bank WV, USA ............................................................................................ 15, 21
INAF-Istituto di Radioastronomia, Bologna, Italy .............................................................................................................. 22 INAF-Osservatorio Astrofisico di Arcetri, Florence, Italy ................................................................................................... 22 INAF-Osservatorio Astrofisico di Catania, Catania, Italy ................................................................................................... 22 INAF-Osservatorio Astronomico di Cagliari, Selargius, Italy.............................................................................................. 22 Kapteyn Institute, Groningen University, Groningen, The Netherlands ........................................................................... 17
National Radio Astronomy Observatory (NRAO) CDL, Charlottesville VA, USA ...................................................... 15, 21
NRC Herzberg, Penticton, BC, Canada .................................................................................................................... 8, 16, 20 NRC Herzberg, Victoria, BC, Canada .................................................................................................................................... 8 Square Kilometre Array Organisation, Macclesfield, Cheshire, UK ................................................................................... 25 The 54th Research Institute of CETC, Shijiazhuang, China ................................................................................................. 11 The University of Manchester, Manchester, UK ................................................................................................. 3, 9, 12, 23 University of British Columbia, Vancouver, Canada............................................................................................................ 7 University of Calgary, Calgary, Canada .............................................................................................................................. 20 University of Malta, Malta ................................................................................................................................................ 22 University of Oxford, Department of Physics, UK ............................................................................................................. 22
West Virginia University, Morgantown WV, USA ..................................................................................................... 15, 21
CONTACT US
t 1300 363 400 +61 3 9545 2176 e [email protected] w www.csiro.au
AT CSIRO, WE DO THE EXTRAORDINARY EVERY DAY
We innovate for tomorrow and help improve today – for our customers, all Australians and the world.
Our innovations contribute billions of dollars to the Australian economy every year. As the largest patent holder in the nation, our vast wealth of intellectual property has led to more than 150 spin-off companies.
With more than 5,000 experts and a burning desire to get things done, we are Australia’s catalyst for innovation.
CSIRO. WE IMAGINE. WE COLLABORATE. WE INNOVATE.
FOR FURTHER INFORMATION
CSIRO Astronomy & Space Science Dr Douglas Hayman t +61 2 9372 4222 e [email protected] w www.csiro.au/en/Research/Astronomy