Page 1 of 9
Newsletter #3
September 2007
edited by Steve Torchinsky
Message from the Coordinator ................................................ 2
Arnold van Ardenne
Message from the Chairman of the SKADS Board ...................... 3
Peter Wilkinson
EMBRACE Beamformer chip at ASTRON ..................................... 3
Klaas Visser, Erik van der Wal, Mark Ruiter, Dion Kant
BEST ......................................................................................... 4
Stelio Montebugnoli
Simulations for the SKA: Pushchino 2007 ................................. 5
Steve Torchinsky
System Design and Costing: Progress Summary ........................ 7
Rosie Bolton
SKADS Recruitment ................................................................... 8
Steve Torchinsky
Square Kilometre Array Design Studies
is partly funded by
Page 2 of 9
Message from the Coordinator
Arnold van Ardenne, SKADS Coordinator
This third SKADS Newsletter is to some extent
focused on the SKADS achievements leading up
to the Mid Term Review in October and the
second workshop preceding it. In preparation
for these events, a new and updated SKADS
brochure is in the making and will be ready in
September.
Before that however, a lot has happened and has
been summarised in the second Annual Report.
The report illustrates and underpins the
technological progress, the establishment of a
framework of scientific and technical simulations
and the development of a system level costing
tool. While impressive in itself, it is also fair to
say that SKADS has greatly enhanced the
visibility of the SKA in Europe and elsewhere. As
a result, SKADS contributed to the identification
of the SKA on the ESFRI roadmap and provided
input to the next step i.e. the European FP7
programme PrepSKA aiming for the construction
of the SKA. SKADS also firmly established its
key technology area i.e. aperture arrays as a
concept technology for the SKA. This
technological scope is relevant to other
implementations as well, for example, focal
plane arrays.
Thanks to all involved in SKADS, the scientific
and engineering section of the Annual Report
was ready on time and is available on the SKADS
wiki site. The final version will be posted on the
SKADS website.
Proceeding from the last SKADS Newsletter, the
emphasis of SKADS activities supplementing the
work in the DS’s was on (i) defining and
specifying the architectural framework of the
SKA against which the science and astronomical
data simulations and costing, could be
referenced; the “SKADS Benchmark scenario”. In
this SKA scenario, the Aperture Arrays concept is
assumed for frequencies up to about 1 GHz
combined with a different high frequency
concept (i.e. a small dish concept).
(ii) Positioning and shaping SKADS as a coherent
European pathfinder project inside Europe and
elsewhere and to illustrate the technical and
scientific feasibility of the Benchmark scenario
(iii) defining a solid costing approach and
system level costing tool which resulted in a
report that was presented at the SKA ISSC
meeting last March (iv) preparing and
negotiating with IST-CENTRA from Portugal as
new participant and preparing for the leave of
four (University) participants (v) increased
interaction with the International SKA (Project
Office) activities at large (vi) executing of the
FP6 Marie-Curie SKADS program in support of
SKA(DS) science and technology toward
educating younger people.
From 30 July to 1 August, 2007 a successful
inter-SKADS workshop on “Simulations with the
Square Kilometre Array was held at the
Pushchino Radio Astronomy Observatory, Russia
Thanks to this activity, which brought together
most people working in DS2 and other scientist
interested in this topic, the status and the
connection was made clear on science
simulations and astronomical data.
As the nearest upcoming next SKADS event, I am
happy to note that the first Marie Curie MCCT–
SKADS Workshop has been completely booked!
It will be held from 23-29 September in Bologna
and hosted by INAF, Istituto di Radioastronomia
in the lovely Castel San Pietro Terme. This event
is sponsored by the Marie Curie MCCT-SKADS
programme, within the Sixth Framework
Programme of the European Commission. The
programme consists of lectures and scientific
talks in view of SKA and future interferometers
and is spread over a full week, with morning and
afternoon sessions. Details can be found directly
(see www.ira.inaf.it/~school_loc/) or through the
SKADS website.
Details of the next workshop on phased array
technologies for radio astronomy to be held in
November in Dwingeloo, will be posted soon.
See also http://mcct.skads-eu.org/ for the
MCCT-SKADS program at large.
September will be a busy month. The SKA2007
Inter Working Group meeting will be held in
Manchester from the 27th to the 29th, and this
is closely followed by the symposium “From
Planets to Dark energy: the Modern Radio
Universe” which celebrates the 50th anniversary
of the Lovell telescope.
The second SKADS Annual workshop will be held
from 10-12 October just prior to the SKADS Mid
Term Review. This important event will
publicise the huge progress made in SKADS,
especially regarding the feasibility and
exceptional scientific possibilities using the
widefield imaging capabilities of the SKA. A
preliminary program has been posted on the
SKADS website including a registration form.
The workshop will be held at the Observatoire
de Paris and will have a conference dinner on the
Eiffel Tower overlooking Paris at night!
Enjoy reading!
2
Page 3 of 9
Message from the SKADS Board
Peter Wilkinson, Chairman, SKADS Board
The major results of the 4th SKADS Board
Meeting held in The Hague on 28/29 June 2007
were as follows:
The Mid-term review and SKADS workshop:
The EC Mid-Term Review meeting will be held in
Paris at the Institute D’Astrophysique on
October 12. SKADS will be represented by the
Management Team and the Board Chair. The
EC’s External reviewer will be Professor John
Seiradakis (U. Thessalonika) and the EC will be
represented by Dr. Elena Righi-Steele. Prior to
this meeting the 2nd SKADS workshop will be
held in the IAP on October 10/11 with the
External and EC reviewers in attendance.
The 2nd Annual Report will be used as the focus
for both meetings. The Board agreed that the
DS2/DS3 deliverables look good for “T0+24”
months. The DS4/5 technical deliverables are
also looking good but are formally a little late
for T0+24, however the rate of progress is good
and the finishing schedule is expected to be
met. The DS6 technical deliverables are on time.
The Board was therefore confident that there are
no “red-flags” and that the workshop and MTR
should go well.
A new SKADS brochure will be produced for the
workshop and MTR and for the international SKA
meetings starting in late September.
SKADS Science
Over and above the simulation work in DS2 the
Board agreed that the “SKADS Virtual Telescope”
exercise led by Oxford was very useful and
should be closely linked in with the international
Science WG programme. The Board strongly
encouraged the production of a “4 years in the
life of” document and a proceedings from the
July Pushchino meeting by the time of the
Autumn meetings
The Board commended the Project Scientist and
the French SKADS team for their efforts in
raising the profile of the SKA in France.
SKADS Costing Exercise
The Board was impressed with the work of the
Design/Costing team, led by Cambridge, and
noted with pleasure that alignment of concepts
with the ISPO-sponsored team has already taken
place. The combined team is now working
towards a single set of flexible software tools as
the next-generation effort. Back-end processing
remains a major strategic issue.
SKADS and the international strategic context
The Board was made aware that the Funding
Agencies have “bought into the project” and that
the UK agency (STFC) is taking a lead role.
SKADS leading players are strongly linked with
international project and SKADS as a whole is
contributing strongly to the global development
effort. The EC is facilitating the next, fully
international, phase via FP7 PrepSKA for which
we in Europe need to seek matching funds to
continue the R&D effort after SKADS comes to an
end in mid-2009.
In this context the ASTRONET questionnaire
reply for SKA is important. The SKA Project
Director agreed to seek advice from European
ISSC representatives before submitting his reply.
We also hope to get an invitation by Michael
Grewing to talk to the ASTRONET ground-based
astronomy working group.
Finally the Board was clear that the SKA project
needs to clarify/specify “the minimum capability
demonstrator” for aperture arrays so that we
know the “height of the fence to be jumped” for
the inclusion of AAs in the first construction
phase to be started in, or around, 2012.
EMBRACE Beamformer Chip
Klaas Visser, Erik van der Wal, Mark Ruiter,
Dion Kant
The EMBRACE design team at ASTRON,
successfully completed the tape-out of the
beam-former chip before the dead-line on July
31. The beam-former chip is an analogue RF
integrated circuit (RFIC), which is an essential
part for aperture array technology.
The main function of the chip is to combine the
four input signals, coming from the antennas,
into two independent beams. The signal chain of
this chip includes a low noise dual feedback
amplifier, poly-phase networks, switching
networks, amplitude correction facility as well as
the output drivers.
This beam-former chip will replace roughly 90
percent of the electronics on the current
EMBRACE prototype boards. Apart from gain
3
Page 4 of 9
stages it replaces 8 KDI vector modulators, as
used in the THEA demonstrator. The analogue
processing power of the chip can be expressed
in the equivalent number of Flops, as a digital
solution would need to perform a similar task. A
conservative number leads to 24 GFlops of
processing power.
The chip was designed in the SG25H1 process of
IHP-microelectronics. This process is a 0.25 μm
SiGe:C BiCMOS technology, with npn Hetero- junction bipolar transistors with an fT and fmax
of respectively 180 and 220 GHz.
Some key figures to indicate the complexities of
the design :
● 12480 SiGe Hetero-junction bipolar
transistors
● 17745 CMOS transistors
● 136314 Connections
● 12.6 sq millimetre die size
The design was made using a combination of
Cadence and Agilent ADS software on two dual
Xeon workstations equipped with 8 GByte RAM
each. At the end, the design database consumed
roughly 400 GBytes of disk space and both
systems were fully stretched to their limits.
BEST
Stelio Montebugnoli, DS6 Leader
The BEST-2 test bed is the second level
demonstrator resulting from the
re-instrumentation of part of the Northern Cross
radiotelescope. After a number of important
mechanical modifications to the eight
North-South cylindrical concentrators, the
collecting area of about 1360 m2 has been
upgraded with new, high performance and low
cost receivers.
This demonstrator allows us to optimise
solutions for many of the critical design blocks,
including the low cost front ends, the high
performance analogue optical links, the
wide-band high dynamic range data acquisition
and processing systems, and the software for
adaptive beamforming (i.e. MVDR) as well as for
calibration. The plan is to have the
demonstrator ready to observe the “first
radiosource” (i.e. Cassiopeia) in time for the
SKADS Mid Term Review.
The 8 cylindrical antennas modified to set up the BEST-2
demonstrator
One of the most challenging elements is the FX
correlator. A very fast 32-channel modular
correlator is under development using one
single Bee2, 8 IBOBs and 16 A/Ds boards from
Berkeley University.
A/D and IBOB boards
Bee2 5xFPGAs board
4
Page 5 of 9
Simulations for the SKA:
Pushchino 2007
Steve Torchinsky
The historic Pushchino Radio Astronomy
Observatory was the venue for a milestone event
in the SKADS project. The workshop entitled
"Simulations for the SKA" brought together
elements from simulations efforts in pure
science, telescope configurations, network, and
costing.
Pushchino is just over 100km south of Moscow,
and has a history going back to the 1950's
Soviet Union when it was conceived as a Science
Village, home to over 20 000 scientists, mainly
working in biological science, but also the site of
world class radio astronomy facilities.
From July 30th to August 1st, Pushchino
welcomed over 30 participants, 20 from outside
Russia. A major goal of SKADS is the synthesis
of efforts in science and technology, ultimately
leading to the design of a cost effective and high
performance SKA. Within SKADS, this workshop
was a joint effort of Design Studies 2 and 3
(Science and Networking) together with design
and costing work carried out with input from
from DS4 (Enabling Technologies) and Design
Studies 5 and 6 (EMBRACE and BEST).
The first day concentrated on the science
simulations, and there were presentations from
nearly all the SKA Key Science Projects, but first
of all, the tone of the meeting was set with a
brief overview of SKADS by Steve Torchinsky
followed by a summary of the SKADS Design and
Costing effort by Rosie Bolton. Participants were
encouraged to explore new scientific avenues,
but they were reminded that ultimately the
science simulation work should be brought to
Earth with concrete numbers on technical
requirements.
Paola di Matteo presented her results from work
at the Observatoire de Paris in the Epoch of
Reionisation. Danail Obreshkow presented
results in HI and continuum simulations in the
relatively nearby Universe, and this was followed
by Rense Boomsma's talk on high redshift HI
simulations. With the comment "only baselines
out to 100km are interesting," Rense was the
first of several to express a desire for more
collecting area in the core. However, this was
offset by the subsequent presentation on
confusion limits by Yuri Ilyasov.
Filipe Abdalla presented the future promise of
SKA for advancing cosmology, especially our
understanding of Dark Energy, and also the
possibility of constraining the neutrino mass
using multiple large-scale surveys of different
galaxy types.
Returning to the relatively local neighbourhood,
Roy Smits presented the requirements for the
pulsar survey. This turns out to be the most
technologically demanding of the surveys, with
huge requirements for processing. The pulsar
survey requires approximately 50% of SKA
collecting area in the core, and this differs
significantly with the current SKADS Benchmark
Scenario.
Maxim Pshirkov presented a second pulsar talk,
explaining the possibility to probe the galactic
bulge for Dark Matter using pulsar timing
residuals from microlensing. This method
effectively maps compact objects in the galactic
bulge and is insensitive to Dark Energy which is
a distributed phenomena.
The next presentations were related to
observations of magnetic fields, including a
summary of Rotation Measure mapping by
Tigran Arshakian, which was followed by Rosie
Bolton discussing Faraday Rotation in clusters as
a probe of the structure of magnetic fields in
clusters. Vladimir Shishov then concluded the
morning session with a presentation on
interplanetary and interstellar turbulent
plasmas.
The afternoon of the second day was devoted to
telescope simulations beginning with an
overview by Tony Willis of the current state of
software. There was also a poster contribution
from Dharam Vir Lal on array configurations.
The rest of the day was a high-intensity
workshop on the MeqTrees simulation software
led by MeqTrees developer, Oleg Smirnov. Oleg
walked us through a number of telescope
simulations, but one afternoon is not enough to
master the process. Another MeqTrees
workshop, this time over a period of several
days, will be organised in January in Groningen.
After the intense day, it was very welcome to
walk in the warm, sunny weather, and visit the
PRAO instruments, led by our host, Rustam
Dagkesemanskiy. Pushchino has three major
instruments: a 22m reflector used for mm-wave,
an array of 16384 dipoles for use at wavelength
3m, and a cylindrical-cross radio telescope with
two arms, each of 1km length used for
observations in the wavelength range 2.5m to
10m. The hard day of thinking, followed by a
bit of walking, and for some, climbing towers,
was followed by an abundant Russian dinner of
fresh salmon, caviar, and vodka along with the
many traditional Russian toasts.
Despite the vodka, we all managed to restart the
next morning, and Jan Noordam opened the
5
Page 6 of 9
final day with a look at calibration for LOFAR.
Yuri Ilyasov was next, presenting the design for
the wideband feed on KAT.
This was followed by some lively discussion
touching on all topics of the workshop. There
were especially some notable exchanges on the
topics of processing power requirements and
the amount of collecting area necessary in the
core.
The final few talks wrapped up the workshop
6
These two snapshops of participants at the Pushchino meeting clearly show transient phenomena, and the
importance of a fully- sampled, wide field- of- view together with a short time constant.
.
.
Page 7 of 9
with overviews of the simulation efforts. Steve
Rawlings presented the status of the "4 Years in
the Life of SKA" exercise. The results will be
published in October, and the next stage is "16
Years in the Life of SKA" going from SKA Phase-I
through to the full SKA. Steve Rawlings left us
with the tantalising possibility of directly
measuring the expansion of the Universe by
multiple surveys over a period of ten years.
Rosie Bolton had the difficult task of bringing all
the workshop presentations together and
summarising the technical requirements for the
SKA, and the implications for the overall cost.
The pulsar survey requirements will necessitate
the biggest change to the SKADS Benchmark
Scenario, but data-rate is a constraining factor
for all the Key Science projects. In particular, if
the full data-rate is required from all SKA
stations, even the distant ones, then the cost of
the SKA rises dramatically. Rosie also pointed
out that the time it takes to complete a survey is
a factor. Allowing for more time to complete a
survey can ease some of the technical
requirements in the SKA.
Finally, Hans-Rainer Klöckner and Cormac
Reynolds gave overviews of the science and
telescope simulations carried out within SKADS
Design Study 2 (Tasks 1 and 2). This effectively
summarised the Pushchino Workshop and we
ended with a final talk by Steve Torchinsky
requesting contributions for the proceedings,
reminding everyone to begin publishing their
results, and, of course, thanking all the
participants, and especially our Pushchino hosts,
Rustam Dagkesemanskiy and Vladimir
Shoutenkov, for a very productive and enjoyable
workshop.
All the Pushchino presentations, along with
many photos, are available on the SKADS Wiki at:
http://webmail.jb.man.ac.uk/skadswiki/PushchinoMeeting
System Design and Costing:
Progress Summary
Rosie Bolton
The SKADS Design and Costing work over the
past nine months has been largely focussed on
producing a detailed design and corresponding
cost estimate for an SKA meeting the SKADS
Benchmark Scenario within the Reference
Design. This work has taken input from relevant
experts within SKADS to get the best estimate of
the component costs. The result was a very
detailed snap-shot cost of an SKA design and
was published as SKA Memo 93.
In brief, an SKA system meeting the Benchmark
Scenario comprises a Low-Frequency array,
similar to LOFAR, receiving at 100-300 MHz, a
Mid-Frequency array, similar to Embrace,
working at 300-1000 MHz and ‘Small’ (~6m)
dishes with single pixel feeds operating at
frequencies from 1-25 GHz. (see figure on
page 8)
The estimated cost for such a system is €1.91
Bn, in 2011 money (see piechart on page 8). The
uncertainty on this estimate is around 10%.
Costs specific to the Mid-Frequency AA design
amount to 45% of the total, the dishes
contribute 19%.
Transporting the data from the stations to the
correlator is expensive: the communications
costs amount to €800 Million if the full data rate
is assumed to be used out to the longest
baselines. Reducing the data rate from the
stations that are over 480 km from the core by a
factor of 9 (as is assumed in the €1.91 Bn total),
brings down the communications costs by €500
Million. With this reduction the communications
costs still make up 18% of the total SKA cost.
Full details of the design and associated costs
can be found in Memo 93 - essential reading for
all SKADS members.
The costing process has now completed its first
full cycle and we are in a position to review the
design and make cost/science trade-offs. With
the SKA Specification Tiger Team reviewing the
SKA design, now is a good time to re-think what
the firm requirements of an SKA are for each of
the Key Science Experiments, possibly
developing alternative observing strategies to
enable a lower-cost solution. One draw-back of
the design and costing work presented in Memo
93, is the lack of easy scaling in the model:
whilst the spreadsheet format allows some
values to be altered, not everything scales
automatically with these changes.
It is desirable to improve the flexibility of the
SKADS costing model so that it enables direct
comparisons to be made across a range of
differing SKA designs. The ISPO costing tool
‘SKACost’ (see SKA Memo 92), which has been
developed by Aaron Chippendale and Tim
Colegate (both ATNF/CSIRO), is an SKA costing
code written in Python. This code has some
features that are very good – including Monte
Carlo error analysis, but no easy user interface
allowing manipulation of the telescope design. It
also did not include an Aperture Array.
Tim Colegate, Aaron Chippendale and Adriaan
Peens-Hough (from SA SKA) visited Cambridge
for two weeks in June/July and we agreed on a
way forward. The plan is to improve the usability
of the tool by allowing interaction via a graphical
user interface and to include the AA in the
7
Page 8 of 9
model. The code will take on the form of a
flexible framework into which components and
groups of components (the “Design Blocks” in
SKADS parlance) can be placed. Design blocks
(which can be user-defined or taken from a
database) are then pieced together to make up
the full telescope system.
The new SKACost tool will be available over the
coming months, initially via a web interface, with
the full version following in the coming year. It
is vital that Scientists and Engineers understand
how important cost is as a driving force for the
SKA design – the new modelling tool will allow
cheaper alternatives to existing designs to be
considered and compared fairly.
SKADS Recruitment
Steve Torchinsky, Project Scientist
Henrik Olofsson will join the SKADS team at
Observatoire de Paris starting in December
2007. He completed his PhD in 2003 from
Chalmers University of Technology in
Gothenburg, Sweden. His topic was "Satellite
and Ground based observations of Massive Star
Forming Regions." Much of his work to date has
been with the Swedish mm/submm wave
8
A schematic view of a station as costed in Memo 93. The Mid frequency aperture array is a fully-filled circular array of square
tiles. Signal processing (to form station beams) is done in a bunker situated underneath the array.There are assumed to be
250 stations to form the SKA each of which transmits signals back to the correlator near the core via optical links.
.
.
.