Final Report Summary - PREPSKA (A Preparatory phase proposal for the Square Kilometre Array)
Executive Summary:
The Square Kilometre Array is a multi-purpose radio telescope covering the frequency range from 70 MHz to >25 GHz that will play a major role in answering key questions in modern astrophysics and cosmology. It will be one of a small number of cornerstone observatories across the electromagnetic spectrum that will provide astrophysicists and cosmologists with a transformational view of the Universe.
Since 2008, the global radio astronomy community has been engaged in the development of the system design for the SKA as a major part of the Preparatory Phase of the project. The end of the Preparatory Phase is now reached, and a number of major decisions have been made so that, at the end of 2011, the international SKA project can progress to the Pre-Construction Phase (2012-15).
Thanks to the work of the PrepSKA governance and legal framework team, it has been possible to provide the groundwork for establishing the SKA Organisation as a UK legal entity, to govern the SKA project in the preconstruction phase. This culminated with 7 national partners signing the agreement on Nov 23rd 2011, when the SKAO came into being.
A funding model, focusing on understanding funding options in the preconstruction period 2012 to 2017, has been initiated, and a survey of funding agencies’ views on these options has been completed. Information on this topic has been gathered from other large research infrastructures and a report prepared on collective best practice. A tool for modelling funding scenarios has also been developed. The business plan and an initial estimate of the cost of the current technical concept have been delivered.
A report on the social and economic benefits of the SKA has been generated.
The Project Execution plan, covering the preconstruction and Construction phases, has been produced, with a detailed Statement of Work and a Work Breakdown Structure.
A major achievement has been the completion of the site selection and measurement campaign. Expert panels evaluated the information, and as a result, two locations have now been selected by the SKA organisation, one in South Africa, and one in Western Australia. The South African site will host the SKA1-MID frequency array (0.45-3 GHz) comprising 190 new dish antennas and the 64 dishes of the precursor Meer KAT antennas. The number of dishes will rise to 3000 in SKA Phase 2. The Australian site will host both the SKA1-LOW frequency array (0.07-0.45 GHz), consisting of 250,000 antennas grouped into aperture array stations (1,000,000 antennas in SKA Phase 2), and the SKA1-Survey telescope, which will combine 60 new dishes with the 36 ASKAP precursor dishes.
In terms of the detailed design, the major achievement has been the completion of all the SKA Element Concept Design Reviews (CoDRs), and the remaining delta CoDRs. Two additional CoDRs, for the Monitor and Control work and the mid frequency aperture arrays, have also taken place. These reviews were presented to external panels from industry and the astronomical community, and were well attended by other community members. The results lay the groundwork for the subsequent design phase of the SKA. The prototyping of the Dish Verification Antenna 1 is well under way, and construction of the prototype support structure has begun. A small test Aperture Array has been developed, and is being used to start testing antenna element options.
Finally, all the PrepSKA deliverables have been completed and delivered, and the SKA is now ready to move forward into the Pre-Construction Phase.
The goals of the Pre-Construction Phase are to 1) progress the SKA design to the point that Production Readiness Reviews have been successfully completed and contracts for construction of major sub-systems have been let, 2) advance the infrastructure roll-out on the selected site to the point where sub-systems can be deployed (assuming the funds for infrastructure development are made available), and 3) mature the SKA legal entity into an organisation capable of carrying out the construction, verification, and operation of the telescope. Without the foundations laid by PrepSKA, this next stage would not have been possible.
Project Context and Objectives:
The Square Kilometre Array (SKA) is the next generation, multi-purpose radio telescope. It is a transformational instrument that will play a major role in answering key questions in modern astrophysics and cosmology. It will be one of a small number of cornerstone observatories across the electromagnetic spectrum that will provide astrophysicists and cosmologists with a transformational view of the Universe. The SKA will be built in three phases, each phase will deliver ground-breaking science – this staged construction and deployment is made possible by the SKA being a radio interferometer. For the 4 years from 2008 to 2012, the global radio astronomy community has, through the Preparatory Study for the SKA (PrepSKA) project, been engaged in the development of the system design for the SKA as a major part of the Preparatory Phase of the project.
The issues addressed by PrepSKA before the pre-construction phase of the SKA are:
1. The design for the SKA
2 The location of the SKA and the required infrastructure
3 The legal framework and governance structure under which SKA will operate
4 The most cost-effective mechanism for the procurement of components of the SKA
5 Funding mechanisms for the SKA
PrepSKA has coordinated and integrated R&D work from around the globe in order to develop the first costed design for Phase 1 of the SKA, and a deployment plan for the full instrument. With active collaboration between funding agencies and scientists, all of the options for the policy-related questions have been investigated.
PrepSKA has been extremely successful and demonstrated the critical importance of addressing technical, policy and funding issues alongside one another. The true success of PrepSKA can be seen in the successful transition of the project to the next (pre-construction, or detailed design phase) with a new legal entity, the SKA Organisation (SKAO), managing the project. This is a truly global structure with membership consisting currently of 10 countries represented on the SKAO Bboard by national-level agencies (science ministries or funding agencies).
The final SKA telescope will have: frequency range from 70 MHz to more than 25 GHz; a collecting area of over one million square metres to give a sensitivity of up to 50 times higher than the Jansky VLA, currently the world’s most powerful radio telescope; distribution of collectors on scales (baselines) of up to at least 3000 km; an instantaneous field of view (FOV) of up to several tens of square degrees at 1 GHz, many times that of existing instruments and considerably larger at lower frequencies, and the new possibility of multiple simultaneous users of several large, independent fields-of-view. These capabilities are enabled by a much greater use of information and communications technology than in current designs, and the result will be an extremely powerful survey telescope with the capability to follow up individual objects with high angular and time resolution.
The telescope is to be built in twothree phases, and will be able to carry out significant scientific observations as it is being built. As PrepSKA ends the pre-construction phase is being entered, which is expected to last from 2012 to 2016. Then construction will start with Phase 1. Phase 1 (SKA1) is the initial deployment of 10% of the array at low and mid-band frequencies, and is a subset of phase 2 (SKA2), which is the full collecting area at low and mid-band frequencies (~70 MHz to 10 GHz). A further, potential Phase 3 (SKA3) includes the implementation of the higher frequencies, up to 25GHz or more. SKA1 will be a major step forward in terms of sensitivity, survey speed, image fidelity, temporal resolution and field of view. SKA2 will have 10 times the sensitivity and 100-10000 times the survey speed of SKA1, and will transform astronomy. SKA3 lies beyond the current planning assumptions and vision for the project.
The baseline design for the SKA consists of an aperture array operating at low frequencies, with optimal performance in the range 100-250 MHz, with capability over the range 70 to 450 MHz, and an array of 250 parabolic dishes initially operating at frequencies up to 3 GHz , but eventually capable of 25 GHz in terms of antenna surface accuracy, and a 90-dish survey array with each dish equipped with Phased-Array Feeds offering enhanced field-of-view and hence ability to map the sky faster. An Advanced Instrumentation Program (AIP), including wide-band single-pixel feeds (WBSPFs), next generation phased-array feeds (PAFs) for the dishes, and mid-frequency dense aperture arrays (AA-mid), has been identified. These will provide much wider fields of view, thus enabling faster surveying. The wide band single pixel feeds will also provide much wider frequency coverage than current feeds. These technologies are relatively immature now, but may well be implemented in Phase 2.
Technological innovation, closely paralleling commercial IT developments, is the key to the design concepts under investigation and to the target cost of €650 million for Phase 1. As an example, for the Science Data Processor data input rates for SKA phase 1 will be of order 10 Tera-bytes/sec, with near Exaflop capacity required for the central processor. Close collaboration with industry partners will be a core element of the detailed design phase, building on the earlier technology programmes (such as SKADS) and leading to significant impact beyond the radio astronomy community.
Two locations for the telescope have now been selected by the SKA Organisation; South Africa (extending to Southern Africa in SKA2) and Australia. The report from the SKA Site Advisory Committee (SSAC) demonstrated that both sites were well suited to hosting the SKA, but initially identified Southern Africa as the preferred site. Further study by the Site Options Working Group, showed that in SKA1, viable dual‐site implementations exist that not only maintain, but add to, the scientific appeal of the first stage of SKA. A dual site also maximises the financial viability of the project in the longer term, through continuation of the current Organisation membership and a global character that will be attractive to future members. With these considerations in mind the SKAO determined that the most advantageous route to deliver the SKA was to utilise both sites and incorporate as far as possible the infrastructure investments already made by the hosts.
The installations are as summarised in Table 1 (see PDF attached, "SKA Installations"). Australia will host the low frequency aperture array antennas in Phases 1 and 2 of the SKA. The addition of 60 SKA dishes to the 36-dish ASKAP precursor in SKA Phase 1 will expand this survey facility. In South Africa, the addition of 190 SKA dish antennas will expand the 64-dish MeerKAT precursor array in SKA Phase 1. South Africa and eight African partner countries will host the dish array in Phase 2 of the SKA and will also host the Phase 2 mid frequency aperture array antennas.
Project Results:
1 Main Science and technology results and foregrounds
1.1 System Overview
Figure 1 (see attached PDF, “S&T Figures and Tables”)is a system schematic showing the major elements of the SKA telescope. The system concept encompasses all of the parts in Figure 1, including the four receptor types. Figure 1 emphasizes the geographic layout of the system components and the aggregated data flow at important points in the system.
Figure 2 shows artists’ concepts of the technologies that are being considered. The system is designed to cover frequencies from 70 MHz to 10 GHz, which cannot be done sensitively with a single antenna design. These depictions are certainly not designs at this point, and represent only rough impressions of what the receptors might look like in most cases. Physically the system will be implemented as three arrays of antennas, a dish-array, and a sparse aperture array (AA) for use at the lower end of the frequency range and a dense AA for use in the mid-frequency range (see Section 4.2). Both AAs are capable of forming many beams at once, thus covering wide areas of sky. The dishes will be equipped with wide-band single-pixel feeds (WBSPFs) (one or two single wide-
band feed per antenna), forming a single beam on the sky, and/or phased array feeds (PAFs) that are capable of forming a cluster of many simultaneous beams that cover a wider field of sky than the single feed.
The receptors will be arranged on the ground over a very wide area, but concentrated mainly in a central region, “the core”. The pattern of antenna layout is called the array configuration. Figure 3 depicts a representative array configuration currently under consideration. The distribution of collecting area in the three types of receptors will fall off in such a way that the distribution of baseline lengths (the distances between antennas) increases logarithmically from the centre out to a distance of about 180 km. Beyond 180 km the dishes are arranged into stations, which are spaced more widely on so-called spiral arms, and these dishes will be equipped only with single pixel feeds. Wide-field receptors at distances beyond 180 km are not needed and processing very wide fields is impractical due to signal transport constraints. This can be seen in Figure 3, although the precise arrangement of the long baseline stations depends greatly on the geography of the chosen site. There is no unique configuration that will satisfy all the scientific requirements optimally. Part of system design is to obtain a satisfactory array configuration for each of the receptor types and frequency ranges. What is required is to find cost-constrained optimum configurations that are suitable for the available sites, and enable both the imaging and non-imaging science requirements.
The SKA will utilize radio-telescope design techniques developed over the past four to five decades. The most sophisticated concept, aperture synthesis, is an application of the van Cittert-Zernike theorem [24]. Each of the receptors has antennas connected to receivers, which essentially transform the free-space signal to an electrical radio frequency signal, which is digitized and connected to the central Signal Processing Facility (SPF). The signals from the antennas, including both amplitude and phase, are cross-correlated in pairs and integrated to reduce noise. These data can be used to reconstruct the original brightness distribution, spectra at each point in the sky, and sometimes spectral and spatial variations with time.
The SPF (1) is a highly flexible real-time signal processor. Figure 1 depicts the SPF as an on-site facility. This is necessary because the amount of data flowing into the facility from the receptors is very large, and putting it too far away from the core would be impractical. Its purpose is to filter the signals from the antennas into frequency channels and to carry out the cross-correlations noted above. This can be done simultaneously from all or multiple subsets of the antennas, depending on the range of spatial frequency coverage needed by the science program. In addition the SPF can carry out delay-corrected coherent summations of the channelized signals. The summed signals form telescope-beams on the sky, whose sensitivity is related to the entire collecting area used in the beam-sum. These signals will be further processed in real time to look for time-domain signatures, especially pulsar signals, but also other transient signals of astrophysical interest.
In all cases, sufficient on-site processing will be done to produce a manageable data rate to the off-site Science Computing Facility (SCF). Part of the system design work is to find the appropriate processing boundary between these two facilities. The SCF is a more general-purpose computer facility, where calibration of the data takes place, images of sky brightness are formed, and further analysis of time-domain effects are carried out. Algorithms for carrying out calibration and imaging are mature processes for current aperture synthesis arrays, as are techniques for pulsar searching and timing. However, the SKA is likely to require significant new developments in this area to handle the much larger amount of data, and to achieve dynamic range targets without continuous human input.
For the SKA the volume rate of raw data entering the SCF from the SKA will probably be too large to archive directly. Real-time image formation will be required. Images and other data products extracted from the images and time-domain data will be archived and distributed to SKA science centers around the world. Many of the science programs will be large, extended surveys. The survey science teams will each have specific requirements for real-time data processing and for handling/distributing data products.
Figure 1 also shows an off-site Operations and Maintenance Centre. There will have to be a small number of operations and maintenance personnel on-site as well, although it is important to keep this group as small as possible, due to the potential for radio frequency interference. The development of a technical operations plan is closely related to system design.
Finally Figure 1shows regional Science and Engineering Centres. A number of these centres are expected to be built, perhaps one in each region of the world. These centres are where scientific analysis is done, and where new engineering and software upgrades for the SKA are developed. The precise way in which this is to be apportioned and handled in each sector remains to be developed.
1.2 Towards a Final System
Figure 4 shows the receptor technologies in frequency-baseline space. The virtual “potentiometer” settings indicate that the further analysis is needed to find the optimum configurations for these technologies, based on science requirements, performance and cost.
The receptor technologies occupy four main parameter space dimensions: sensitivity (Ae/Tsys), SSFoM ((Ae/Tsys)2Ω), baseline range and frequency range. (For definitions of Ae and Tsys, see [23][24]). In Figure 4 the four receptor technologies are shown schematically in frequency-baseline space. At present it is not clear where the boundaries should lie or even if they should overlap. In fact, the sparse AAs may bifurcate into two frequency ranges and occupy different baseline ranges.
By the time of the Preliminary Design Review the values of the individual receptor parameters will be defined as well as many other aspects of the integrated system. In order to get an understanding of the capabilities of the receptor technologies for delivering science, evaluation of their current and potential capabilities will be on-going.
Since it is currently not possible to define the system that will be built, the method used here is to select one technology concept at a time, parameterized to give the best science return possible. In section 0 the implications of adopting each concept for further investigation are explored by using “representative implementations”, which are essentially implementation examples. These provide an indication of three resulting aspects: 1) scale, 2) the technical flow-down into support services and 3) potential science performance capability. The first two are directly related to cost.
Representative implementations serve the purposes of understanding the science performance of the technology, and of the implications for the potential support functions (data transport, correlation, imaging, etc.) needed in principle to make the candidate technology work. The technical flow-down into support services can be easily derived from the representative implementations. A minimum set of parameters (number of receptors, frequency ranges, bandwidths, etc.) for the representative implementations have been selected so that this assessment can take place.
Parameters have also been selected so that a particular technology can handle as much of the science as possible. This does not imply that the final system will consist of a simple sum of everything. At the concept stage we do not know the size or scope of each of the technologies. Thus in subsequent system analysis the technologies adopted will depend on an optimized subset of those technologies, based on science-technology trade-offs.
The descriptions of the representative implementations are not to be construed as adopting particular system or technology parameters, nor are they related to detailed design. They are mainly a means of evaluating the potential cost and performance of each of the candidate technologies separately.
1.3 An SKA Data Flow and Control Slice
Figure 5 emphasizes the signal path and key signal interfaces.
Figure 5 is a system schematic showing a “slice” of the system, starting on the left at the first active component and following the signal/science-data path to the regional science centres. Also shown are the main ancillary sub-systems needed to support the system: time and frequency distribution, monitor and control, power, temperature control sub-systems. Each of the candidate technologies contains the blocks shown in Figure 5, but the physical arrangements and scales are very different in each case.
The RF signals originate from elemental antennas that form part of an aperture array, from elemental antennas that form part of a phased array feed or from the single-pixel feeds on the dishes. The signal is received by a low noise amplifier (LNA) in each case, which amplifies the signal typically by a factor of 1000 in power. The LNA is connected to an analogue receiver, which further amplifies the signal, typically by a factor of 107 or more. The receiver also filters the signal, and optionally translates the signal to more convenient frequencies.
If the receiver is part of an aperture array, then several such signals are optionally summed in an analogue beam-former, which may also compensate for geometrical delay. This is likely to be only part of the beam-forming process, which in general is more efficiently carried out in hierarchical stages. The signal is then converted to a digital stream via an analogue-to-digital converter (ADC). Further digital beam-forming is carried out, probably on many more digital streams than in the analogue case. This may optionally be passed to a second stage of digital beam-forming, which in the case of an aperture array would be produced using an entire station. The most likely method of digital beam-forming will also require that the signal be channelized into frequency bins.
If the receiver is part of a phased array feed on a dish, then analogue beam-forming is unlikely. The digital beam-former is similar in principle to that for the aperture arrays, but is used to form beams emanating from positions anywhere over the surface of the feed. Frequency channelization is also required for precise illumination of the dish.
If the receiver is part of a single pixel feed on a dish, only an ADC is needed in this part of the signal chain.
Signals from dishes with single-pixel feeds, located further than ~180 km from the centre, are likely to be combined from stations, defined closely spaced groups of antennas. This is done because the cost of transporting data over long distances could be prohibitive if each output signal from each dish were sent separately, and the processing load would also be very high in this case. Instead these signals are beam-formed, and the signals for a few beams are transported to the centre. Optimising the radius from the centre at which stations are formed is part of the on-going system design process.
All the digital signals flow into the SPF via an extensive signal transport network (blue lines in Figure 5). In principle there could be an SPF for each receptor group, but here it is assumed that they are co-located. The correlator channelizes and cross-correlates the signals from groups of antennas (sub-arrays) or the full array of a receptor technology. In principle it could cross-correlate signals from different technologies if they have overlapping frequency capability, although this is complicated. Cross-correlated output from each sub-array is passed to the science processing facility. Additional operations on the data, such as time-based integration and conversion to floating point number protocol, may be carried out after correlation before transport to the SPF.
The beam-former and time-domain processor carries out time-domain analysis of the same data (or a sub-set) entering the SPF, principally to search for or time pulsars, but also to look for radio transient signals. Signals from the full array or sub-arrays of dishes and/or aperture array (AA) stations will be beam-formed in the SPF before time-domain analysis takes place. Channelization is also required, but this may be shared with the channelization need for correlation. The results of the time-domain analysis are also sent to the science processing facility.
Science processing for “standard” imaging consists of several fundamental steps: calibration, gridding the data into a rectangular grid in the u-v plane, Fourier transformation into the sky intensity domain, and deconvolution of the point-response function of the array. This is done for each frequency bin. These steps are combined in an iterative loop that typically involves solving for array parameters, such as complex gains, concurrently with solving for the image. A buffer may be needed to store visibilities and/or interim results while the loop is being executed. The results of this processing usually form the basis for science interpretation.
The time-domain processing done in the science processing facility is more difficult to define at this stage. The steps needed to find pulsars are well developed, but it is not clear what fraction of that work will be done in the general purpose science processing facility.
The science processing facility will require one or more large data storage sub-systems. These will be needed to buffer incoming data from the SPF, and to store data products.
Figure 5 shows several other significant sub-systems: 1) Local Oscillators and ADC clock signals used in the system must be derived from a common frequency standard. 2) A distributed monitor and control sub-system is needed, complete with a large operations centre. Each block in Figure 5 shows an M&C unit to emphasize the importance of this sub-system. 3) The electronics and other equipment will certainly require cooling and some require temperature control. 4) The distribution of electrical power will form a major part of the system.
1.4 System Hierarchy
The system hierarchy is intended to provide a clear and coherent view on the scope and composition of each of the building system blocks. It is included here, but discussed only very briefly. A more complete description of its purpose is provided in the System Engineering Management Plan [25]. The hierarchy as shown is not intended to be the final version and is likely to evolve. Once finalised, it sets the nomenclature for the levels throughout the project. There are many ways in which the hierarchy could be constructed, depending upon the emphasis intended to be shown. Figure 5 shows the groupings and levels selected for this version. The diagram should be taken in its entirety, but is divided in two parts to permit formatting in a document. The left side of the diagram (top part in Figure 6) shows receptor technologies and their components, and the right part of the diagram (bottom part in Figure 6) shows the support services and facilities. A functional version might look somewhat different. The levels 0 –7 show successive levels of detail and size. The PrepSKA phase of the project is unlikely to be concerned with levels below 3.
2 Reception
This section of the document will start off with an introductory section to the various candidate receptor technologies. The focus will be on the receptors because they are the main enablers to cover the science space. Following this introductory section each of the receptor technologies will be described in more detail and will then include identification and description of the rest of the system that will be required to support each of the receptor technologies.
Figure 6 shows candidate technologies configured for the representative implementations. The technologies and locations on these diagrams are preliminary and may change as system knowledge evolves. Each is designed to carry out observations to support one or more components of the DRM and as much additional underlying science as possible. SAA denotes sparse AA; DAA denotes dense AA; WBSPF denotes wide-band single-pixel feed; PAF denotes phased array feed.
2.1 Candidate Technology 1: Reflector Technology with Wide-band Single-Pixel Feeds
2.1.1.1 Receptor Technology
In this representative implementation the antenna array for the SKA will comprise a large number of ~15-m equivalent diameter offset-Gregorian dish systems, each equipped with two or more wide-band single-pixel feeds (WBSPFs) to provide overall frequency coverage of 0.3 to 10 GHz. Use of cryogenic receiver front-ends coupled to the feeds will maximize the sensitivity and survey speed of the array. Following the front-ends, the remaining sections of the receiver subsystem will provide gain, frequency conversion, digitization and optical modulation, such that the antenna system outputs will be digital bit streams on fibre-optic cables. Clock and local oscillator reference signals to the receiver subsystems will also be distributed via fibre optic cables.
Offset-Gregorian optics have been provisionally chosen for the SKA for the following reasons:
• There will be no blockage of the main reflector and scattering of ground noise onto the main reflector can be minimized by enshrouding the feed/subreflector area.
• Sidelobe levels will be significantly lower than for an axisymmetric antenna, which will increase the achievable image dynamic range over wide instantaneous fields of view.
• There will be sufficient space for a single-pixel feed changer and mechanical rotator for a phased array feed, if required.
The last point is the most important one for this assessment, since it maximises the chance of being able to accommodate several feed systems.
Cryogenic dewars will house the low noise amplifiers (LNAs) for higher frequencies, one per polarization for each feed, at the front-end of the receiver subsystems. At the lowest frequencies it may be possible to use LNAs and feeds operating at ambient temperatures. A strong candidate for cryo-cooling are Stirling cycle refrigerators. These have been used with great success in the mobile phone industry, and have demonstrated excellent reliability. Cooling the high band feed, LNAs and coupling transmission lines will minimize the system noise temperature, and hence maximize the SKA’s sensitivity and survey speed. This will be needed to achieve the noise temperatures shown in Table2 (see attached PDF,” S&T Figures and Tables”).
The remainder of the receiver subsystem will be temperature stabilized to maximize amplitude and phase stability. Signals will be filtered and/or frequency converted, digitised and transmitted to the Central Data Processor via optical fibre. Use of digitized fibre-optic signal transport will ensure the best possible electromagnetic compatibility between subsystems, minimizing the potential for self-generated radio interference and maximizing immunity to external interfering signals. The digitized signals will also be unaffected by environmental and mechanical disturbance to the optical fibres. These measures, as well as others, will be needed to achieve the dynamic range specifications. These comments about optical fibre signal transport will also apply to the other technologies under consideration for the SKA.
As an alternative, less expansive analogue transmission of RF signals over fibre could be used in parts of the RF system if they can be shown to meet the requirements.
2.1.1.2 A Representative Implementation of a Dish-based System with WBSPFs
Table 2 describes the top-level specifications for a representative implementation of this system, which can be used to assess its performance in addressing the components of the DRM. The system comprises 3300 15-m antennas equipped with single-pixel feeds. The feeds cover bandwidth ratios (upper frequency/lower frequency) of ~4:1. This is less than the goal of 10:1 often discussed for this technology. It is currently more likely that 4:1 feeds, rather than 10:1 feeds, can actually meet the Tsys and aperture efficiencies listed in Table 2. Three feeds will be needed to cover the required frequency range. It is assumed that only one feed can be used at once and that a remotely-controlled mechanical feed changer will be needed to place a particular feed at the focus.
Table 3 contains the main performance factors derived from Table 2 for this system.
2.1.2 Justification for Inclusion of WBSPF/Dish Technology
Reflector antennas (dishes) in general have been the standard receptor technology in radio astronomy for wavelengths less than ~100 cm for decades. Their key advantages are extremely wide frequency range, and the fact that they are passive collectors. Receivers have improved steadily in cost and performance, although cryogenics for frequencies greater than 600 MHz are important maintenance cost items. Recent research in antenna construction promises reduction in antenna costs as well. Both trends indicate that this technology will be suitable for use in SKA construction, while continuing to drive up the performance/cost ratio at the subsystem level.
The WBSPFs, which at this stage are a technology development item, have the potential to increase the frequency range of the reflectors to more closely match that of the reflectors. Considerable progress has been made recently. Tests of several types of feeds have been done in the lab and have shown sufficient promise to warrant inclusion in a dish based system concept for the SKA. A more detailed discussion of risks and fall-back positions is contained Section 11.2.1.
The main disadvantage of reflector antennas is the relatively narrow instantaneous field-of-view. This can, in principle, be improved by an order of magnitude by using phased array feeds (see Section 6.4) which at this stage are also a technology development item. Aperture arrays, which are a completely different approach to receptor design, can provide much larger instantaneous fields-of-view.
Dish technology covers a large fraction of the science represented by the DRM and also provides a flexible telescope design to carry out observations that have not yet been considered. This last category is the one in which radio astronomy has made the largest impact historically. Of concern are the “OK” and lower categories.
In summary, with relatively small changes in science-technology trade-offs, the performance of this system could be improved. Adding sparse aperture arrays to this system could potentially enable all except one (HI Baryon Acoustic Oscillations) to be carried out in combination.
2.2 Candidate Technology 2: Reflector Technology with Phased Array Feeds
2.2.1.1 Receptor Technology
A Phased Array Feed (PAF) is a phased array, composed of closely packed, dual polarized, elemental antennas approximately a wavelength in size, and placed on a grid in the focal plane of an optical system of a reflector antenna. The array of elements spatially samples the electromagnetic field in the focal region of the dish. Weighted sums of amplified signals are formed, which have the effect of producing a cluster of overlapping beams on the sky. A beam-former is needed somewhere in the system to carry out the sum. The cluster of beams spans a much wider field-of-view than a single beam.
In another application, smaller PAF systems may also be used which do not increase the field-of-view (FoV), but allow more flexibility in determining the illumination of the reflector, thus increasing antenna efficiency or controlling the level of sidelobes. This general technology is relatively new to radio astronomy and has not yet been used on production telescopes. In the SKA context some or all of the dishes could be equipped with PAFs, each of which would contain about 10~2 elements.
The PAF array is quasi-circular in outline, the center of which is placed on the optical axis of the reflector system. The radius of the PAF from the axis determines the radius of the FoV on the sky. This is entirely geometric and independent of wavelength, in contrast to dishes with SPFs and AAs, where the radius of the FoV increases with for a given dish or AA diameter.
A low noise amplifier (LNA) is needed for each elemental antenna. Because cryogenically cooled systems are currently practical for only small volumes, the LNAs on PAFs must be operated at ambient temperature at L-band (~1 GHz) frequencies. (At higher frequencies they could be sufficiently small to permit the entire array to be cooled). This means that system temperatures will be higher than for single pixel feeds, which have only two LNAs per feed and can be cryo-cooled.
Because PAFs have multi-parameter control over the illumination of the reflector, higher system temperature can be offset by better antenna aperture efficiency for PAFs than single pixel feeds. For typical parameters, a two percentage point increase in aperture efficiency offsets approximately a degree of system temperature. In addition, it may also be possible to provide real-time control of sidelobe levels in dish-based systems.
Introducing PAFs into a dish-based version of the SKA has a potential impact on the architecture of the system. The total FoV and the number of dishes, which would be linked to dish diameter, can be chosen separately in the design of a synthesis telescope of given collecting area. One attraction of this approach is that a wide FoV, and high survey speed, may be obtainable with relatively few dishes, simplifying cross-correlation and post-processing. With correlation load scaling as the square of the number of antennas, the PAF approach in which the correlator complexity scales only linearly with total FoV, is conceptually attractive.
The SKA will need wide receiving bandwidths, which require that the signals from each element be channelized before being summed. The beam-former, including channelization, may reside either at the antenna, at a location that processes inputs from several antennas, or at the SPF (the central processing facility of the entire array). In all cases this would be done digitally. Each beam is processed at the SPF in the same way as for a single-pixel feed system (correlation, imaging and time-domain processing).
For pulsar astronomy, a beam-former is needed for a large central sub-array of antennas. These beams would most likely be formed from signals from the central PAF beam in a similar way to that for single-pixel fed antennas.
The following relations can be derived directly from specifications of the dish diameter, D, focal length, f, the diameter of the PAF, d, and the wavelength within the receiving band, . The total field-of-view in steradians, ΩPAF, is given by
The area of a single beam on the sky in steradians, Ωbeam, is given by
And the number of beams, Nbeam, within the PAF field-of-view is given by
As noted above, ΩPAF is independent of wavelength, but the number of beams depends on -2. This means that the number of output data streams is dependent on the shortest wavelength in the band.
2.2.1.2 A Representative Implementation of a Dish-based System with PAFs
The purpose of the representative implementation is to provide guidance on the implications of performance, complexity and cost of an entire system based on PAFs as the candidate technology.
This implementation is designed to be compatible with a similar dish-based system as for WBSPFs, but using PAFs instead of WBSPFs. Note that these dishes could in principle be equipped with a combination of PAFs and single-pixel feeds (wideband or standard bandwidth) on each dish. Each dish contains two PAFs, which each cover 2.5:1 frequency ranges. This is thought to be largest bandwidth ratio possible with PAFs. Together their coverage from 500 MHz to 3 GHz is designed to capture as many components of the DRM as possible. This choice may not be optimum, even for a system containing only PAFs, and certainly not for a system containing other technologies as well. Principally because red-shifted HI gets weak rapidly at higher redshifts, the increased survey speed of the PAF-based system could be more useful at lower frequencies than 500 MHz. But because the PAF may become too large at wavelengths approaching a meter, the lower limit has been set at 500 MHz.
A PAF diameter of 1.5 m has been chosen for the lower frequency band (0.5 – 1.2 GHz), which provides a FoV of ~160 deg2. The second PAF has been sized to provide the same number of beams, so that the downstream processing system is similar for both. This diameter scales to 0.6 m, which provides a FoV of ~26 deg2. In both cases the number of beams needed to cover the FoV is about 172 in each polarization. It is assumed that only one feed can be used at once and that a remotely-controlled mechanical feed changer will be needed to place a particular feed at the focus.
The following relation for correlator load, Ccorr, provides some guidance on setting other system parameters: , where Nant is the number of antennas, ΩPAF is the field-of-view, Atot is the total effective collecting area, Bbit is the total bandwidth in bits, and is the wavelength. The size of the correlator need not be a focus of design attention, but it must be kept within reasonable bounds. In this implementation, it was used to provide guidance on setting the number of dishes and the number of beams transmitted to the correlator. There are many other ways of doing this, and there is no intent here to be other than illustrative. There is also a concern with the amount of data transport needed.
The number of dishes has been chosen as 2000. With 2000 dishes and all 172 beams, the correlator load factor (see Table 4) would be uncomfortably high. Rather than reduce the number of dishes or ΩPAF, a beam dither factor has been introduced. Beam dithering in this context means instantaneously switching back and forth between different sets of beams, which are interleaved with each other. It is likely to be needed in a PAF-based system to produce uniform sensitivity over the FoV anyway – otherwise the sensitivity will be “lumpy”. A factor of four has been used here, rather arbitrarily, to limit the number of beams produced instantaneously to 43 (dual polarized). This strategy does, however, provide an easy means of increasing the capability of the telescope if more downstream resources can be provided later. Dithering also introduces frequency dependence into the survey speed calculation, because a reduced number of beams is already sufficient to cover the FoV at the low end of the frequency band (see Table 5).
Many of the other system parameters flow from this small number of initial choices. Additional parameters, such as the number of frequency channels and the integration time can be derived from the above for continuum observations, or from the width of the HI-line. For continuum observations, the number of channels and the integration time depend on the maximum frequency and baseline, which are observation dependent, and have been chosen here as rather rough mean values.
The array configuration has been chosen for this case to have the same distribution of baselines as the WBSPF case, except that baselines out to only 180 km have been included. Although there may be a science case for very wide FoVs at long baselines, the image processing load would be extreme.
2.2.2 Justification for Inclusion of Dish/PAF Technology
PAFs are clearly justified for survey science, where SSFoM is the most important performance factor, and a major driver for the SKA as a discovery instrument. Using PAFs instead of single-pixel feeds allows a choice of designs in which the FoV and the dish size can be chosen independently, thus permitting the designer to balance survey speed and point source sensitivity. Thus it may be possible to reduce the number of SKA antennas at the expense of putting PAF technology at the focus rather than a single pixel feed. Since the electronics required for PAFs is likely to be less expensive than the cost of antennas (now or in the future), this may be an important means of getting science from the SKA most efficiently.
If high frequency PAFs can be shown to provide better overall performance in the future, they would be justified for frequencies above 3 GHz.
In an SKA system they would have to be supplemented with single-pixel feeds at high frequencies, or small, cryo-cooled PAFs if they prove to be better. The lowest frequencies would require the addition of sparse AAs to the system. In a sequential SKA development scenario in which dishes are sequentially equipped with different feeds, they would likely play an important role.
2.3 Candidate Technology 3: Dense Aperture Array
2.3.1.1 Receptor Technology
An aperture array is an all-electronic collector consisting of a large number of small individual receiving elements fixed on the ground. Beams are formed and steered electronically by combining the signals from the elements with the appropriate delays and weights. The array can form a large number of simultaneous beams if there is sufficient processing capability and communication bandwidth. The result is a very flexible collector with no moving parts.
Aperture arrays can be designed in two principal configurations: dense, where the spacing between the receiving elements is less than the Nyquist special sampling of ~λmin/2; and sparse where the sampling pitch is > λmin/2.
Considering the characteristics of dense Aperture Arrays, the useable frequency band of a dense array can be extended beyond the sampling frequency into the sparse regime, which will cause a reduction of scan angle range and effective collecting area. The benefits of dense arrays compared to sparse arrays can be summarized as:
• Nearly constant Aeff over the frequency band,
• Absence of grating lobes and the ability to perform full beam shape control
• Low element cost and low processing cost due to the regular physical structure, enabling a high level of production automation and Fast Fourier Transform processing.
The effective collecting area of a single, isolated element of an aperture array scales as λ2, hence the number of elements required is determined by the Aeff required at the top operating frequency in the dense regime. This naturally limits the highest frequency that is likely to be economically implemented by an AA system currently to ≤1.4 GHz. However, as can be seen from the DRM, the requirement for very high survey speeds is also largely limited to below this frequency.
For the implementation of dense aperture arrays, a number of antenna elements are available. The basic characteristic of these elements is a high degree of electromagnetic coupling, resulting in a ‘conductive sheet’ across the array.
2.3.1.2 A Representative Implementation of a Dense Aperture Array
Table 6 describes the top-level specifications for a representative implementation of this system, which can be used to assess its performance in addressing the components of the DRM.
The system comprises 250 56-m Dense Aperture Array stations populated by over 75,000 Vivaldi antennas spaced at 21 cm. This results in the sensitivity of more than 10,000 m2/K out to a scan angle of 45 degrees from zenith. The aperture efficiency is programmable and can be traded against side lobe control and sensitivity. A nominal figure of 80% is used within this document though higher levels might be used for pulsar searches where the presence of large side lobes is less of an issue. The target system temperature is assumed to be 37K and the collection area is proportional to the cosine of the angle from the zenith which equates to 70% of the bore sight area at 45 degrees. Such an array is under-sampled at frequencies above 700 MHz but electromagnetic coupling between the elements should ensure that the sensitivity remains good at 800 MHz and decreases only slightly with frequency up to 1 GHz. At zenith and ~800 MHz the sensitivity would be very close to 13,000 m2/K. Beyond 1 GHz the sparse nature of the array begins to impact on the sensitivity/ survey performance.
2.3.2 Justification for Inclusion of Dense Aperture Arrays
The purpose of building the SKA is to perform challenging science; dense aperture arrays potentially provide access to extended parameter space for radio telescopes which can enable experiments that may not be possible with other technologies. The areas that dense AAs have the greatest impact are:
• Very high survey speed capability,
• Multiple beams for expanded FoV, concurrent observations, and minimisation of correlator and central processing resources,
• High dynamic range capability due to small beams (large diameter collectors), unblocked aperture and good physical stability,
• Flexibility to observe short time period transients over large areas of the sky, coupled with the possibility of keeping a history buffer to view the precursors to a transient.
AAs substantially extend the survey parameter space that the SKA can operate over and as can be seen from section 6.5.1.3 above have a major impact on the science capabilities of the SKA.
AAs are best suited for lower frequencies. Combining dishes at high frequencies with Dense and Sparse Aperture Arrays could potentially enable all of the DRM components to be observed depending on the number of dishes that can be afforded within the financial envelope (for DRM components 6 and 13 in particular). DRM components 15 & 1 may require dishes with low frequency capability on long baselines
There is the potential of achieving the dynamic range requirements of the SKA below 1.4GHz with AAs due to the use of very large diameter collectors (with many beams), and the potential to accurately calibrate at every frequency, direction and location across the receiving ‘surface’.
The costs of implementing AAs is reducing over time due to the extensive use of IT related technology, which is following the performance and cost improvement paths shown by e.g. Moore’s Law. This means that the AAs get progressively more affordable over time and are capable of substantial and meaningful upgrade over the life of the SKA.
2.4 Candidate Technology 4: Sparse Aperture Array
2.4.1.1 Receptor Technology
The Sparse Aperture array concept for the SKA consists of a large number of antenna elements which are spread over a wide field with separations longer than λmax/√2 [19] to each other. Distributing the antennas in such a way maximizes the collecting area over almost all frequencies. The number of antennas required is determined by the sensitivity requirements of the telescope. The antenna technology used can range from simple inverted V dipoles (used for the LOFAR Low Band Antennas) to a more sophisticated technology such as log periodic structures. Both types of antenna are about a wavelength in size.
Below approximately 450MHz sky noise starts to increase dramatically (~ λ2.55) and dominates the system temperature, Tsys. For a constant sensitivity, Aeff/Tsys the array area needs to increase in a similar fashion with respect to reducing frequency. A sparse array dipole has the characteristic that its effective area increases as a function of λ2 and as such provides a reasonable compensation for the increase in the sky temperature. However, if the array is completely sparse down to the lowest frequency requirement of 70 MHz, its diameter will be large and will impact on the amount of down-stream processing required. For the representative implementation, the array is dense up to 100 MHz and sparse above 100MHz.
Each individual antenna has a sensitivity pattern of approximately cos(za), where za is the zenith angle. Since each individual antenna generates a significant amount of data, antennas are typically grouped as individual phased arrays further referred to as stations. This reduces the data volume significantly. The beam width of the beam-formed signal is reduced to ~λ/D with D the diameter of the station. However, the FoV can be increased by forming multiple independent beams over the sky. In this way a trade off between, for example, bandwidth and beams can be made for a fixed total data rate.
The main science cases addressed with sparse aperture arrays are those which require low frequencies. The optimum transition between the sparse and dense regime for aperture arrays lies at approximately 450 MHz [20]. All science cases below this frequency benefit from the very large FoV of a sparse aperture array, increasing as λ2.
The sensitivity of a sparse aperture array system grows linearly with the total number of antennas installed. The relationship between sensitivity and the number of antennas in a station is not necessarily linear since it depends on the detailed station configuration. However, since the array is sparse to first order, sensitivity grows linearly with the number of antennas within a station.
Below approximately 450MHz sky noise starts to increase dramatically and dominates Tsys. Consequently Aeff needs to increase with reducing frequency to maintain the required sensitivity, Aeff/Tsys. Above 450 MHz the sky noise is relatively constant and the Tsys is largely determined by the technical performance of the array. The effective area for a sparse array increases as λ2, largely countering the effects of increasing sky noise.
Regular sparse aperture arrays will suffer from grating lobes. However, grating lobes will be spread out for randomized station configurations. Moreover, the effect of side lobes to the correlation products can be further reduced by rotating different station configurations with respect to each other.
2.4.1.2 A Representative Implementation of a Sparse Aperture Array
The representative implementation considers the 180 metre diameter Sparse Aperture station with antennas regularly spaced at 1.5 metre separation and a total of 250 stations. This results in the array being sparse above 100 MHz and dense below.
The individual antennae need to be simple and cost effective as a very large number are required to provide the required sensitivity, particularly at the low end of the frequency spectrum. The front end amplifiers do not need to have as low a temperature as other antenna due to the dominance of the sky temperature as part of Tsys. This instrument temperature is assumed to be 50K.
The number of antennas per station versus the total number of stations is a balance between correlation and post processing costs and the required u-v coverage. The cost of fixed AA collecting area (fixed number of elements) is to first order not dependent on the number of stations. u-v coverage generally improves with more stations; however, the beam size would increase with smaller diameter stations which may bring dynamic range limitations. While this will be studied, it is likely that around 250 stations are appropriate since this gives good u-v coverage.
The configuration for an aperture array system is determined by the following parameters:
1. minimum size required for station calibration (this is necessary to correct for gain and phase differences of the antennas elements prior to beam-forming),
2. sky resolution required (determines maximum baseline),
3. instantaneous imaging capability (how much u-v coverage is necessary for snapshot imaging),
4. the required u-v coverage in synthesis mode (from the minimum amount of time required for a certain sky coverage),
5. infrastructure and data transport costs.
Digitization, beam-forming and optical modulation for the Sparse Aperture Arrays is not performed at the antenna but at one or more processing units within each station. The resultant beam outputs will be digital bit streams on fibre-optic cables. Clock and local oscillator reference signals to the receiver subsystems will also be distributed via fibre optic cables.
2.4.2 Justification for Inclusion of Sparse Aperture Arrays
Sparse aperture arrays are the only solution able to meet the science goals requiring frequencies less than a few hundred MHz and a large instantaneous FoV. Furthermore the use of multiple beams opens the path to simultaneous observations or increasing the survey speed of the instrument. Also fast switching or re-pointing of beams is an advantage of electrically steering the beam onto the sky.
The sparse aperture arrays can be used for DRM Components 4, 5, 6 and 11 in conjunction with Dense Aperture Arrays and/ or WBSPF Dishes or Dishes with PAFs and WBSPF.
Potential Impact:
Potential Impact of the PrepSKA project
Introduction
The impact of the PrepSKA programme is described in full in the Project Execution Plan, which is the blueprint for the development of the SKA telescope in the preconstruction phase.
The Square Kilometre Array is a multi-purpose radio telescope covering the frequency range from 70 MHz to >25 GHz that will play a major role in answering key questions in modern astrophysics and cosmology. It will be one of a small number of cornerstone observatories across the electromagnetic spectrum that will provide astrophysicists and cosmologists with a transformational view of the Universe.
For the past 2.5 years, the global radio astronomy community has been engaged in the development of the system design for the SKA as a major part of the Preparatory Phase of the project. The end of the Preparatory Phase is now approaching, and a number of major decisions need to be made so that, at the end of 2011, the international SKA project can progress to the Pre-Construction Phase (2012-15). These decisions include approval of funding for the completion of the Preparatory Phase and for the Pre-Construction activities, the establishment of a legal entity for the SKA Organisation, and the selection of the SKA site. (A single site for the SKA is assumed in this document.)
The goals of the Pre-Construction Phase are to 1) progress the SKA design to the point that Production Readiness Reviews have been successfully completed and contracts for construction of major sub-systems have been let, 2) advance the infrastructure roll-out on the selected site to the point where sub-systems can be deployed (assuming the funds for infrastructure development are made available), and 3) mature the SKA legal entity into an organisation capable of carrying out the construction, verification, and operation of the telescope.
This Plan sets out the strategies for carrying out the Pre-Construction Phase, the work to be done and resources required to complete the Preparatory Phase and to carry out the Pre-Construction Phase, potential partnerships to carry out the work, and the governance principles for the legal entity forming the SKA Organisation in this phase. It will enable the start of construction of the first phase of the SKA and convergence on a viable technical pathway to the full SKA. The goal of this Plan is to provide the Funding Agencies and Governments in the Agencies SKA Group with the appropriate information to allow them to assess the scope and feasibility of the work proposed and, in some cases, facilitate the funding for the Pre-Construction Phase.
Phased implementation of the SKA
The construction of a radio telescope with a collecting area approaching one million square metres across a wide frequency range is a major undertaking and is planned to be implemented in phases in order to spread the cost impact. Phased implementation is an effective strategy for an aperture synthesis telescope which can start operating before construction is completed. The international project has adopted the following terminology to describe this phased approach: SKA1 is the initial deployment (10%) of the array at low and mid-band frequencies costing 350 M€ (in 2007 currency units) and is a sub-set of Phase 2 (SKA2), SKA2 is the full collecting area at low and mid-band frequencies (~70 MHz to 10 GHz), while Phase 3 (SKA3) sees the implementation at higher frequencies of up to 25 GHz or more. The SKA1 facility will represent a major step forward in terms of sensitivity, survey speed, image fidelity, temporal resolution and field-of-view. It will open up new areas of discovery space and demonstrate the science and technology underpinning SKA2. SKA2 will have 10 times the sensitivity and 100 -10000 times the survey speed of SKA1 and will transform astronomy.
Project Schedule
The project schedule for the low and mid-band frequencies is divided into six major phases: 1) Preparatory (2008-2012), 2) Pre-construction (2013-2015), 3) SKA1 construction (2016-2019), 4) SKA1 operations (2020), 5) SKA2 construction (2018-2023), and 6) SKA2 operations (2024). The schedule for SKA3 is not well-defined at this point in time.
Science Drivers
The five key science areas defined by the astronomy community as driving the specifications of the SKA, together with the Exploration of the Unknown, are described in detail in “Science with the Square Kilometre Array” (eds. C. Carilli and S. Rawlings) [1]. These drive the system design for the telescope. SKA1 will enable revolutionary science at decimetre wavelengths, with a particular focus on pulsars and gravitational wave astronomy, and H I in the distant and the nearby Universe. With its wider wavelength range, 10 times greater sensitivity than SKA1, and enormous increase in survey speed, SKA2 will transform our understanding of many key areas including: the formation of the first structures as the universe made its transition from a largely neutral state to its largely ionised state today; cosmology including dark energy via baryonic oscillations seen in neutral hydrogen; the properties of galaxy assembly and evolution; the origin, evolution and structure of magnetic fields across cosmic time; strong field tests of gravity using pulsars and black holes including measurements of black hole spin and theories of gravity, and the exploration of the dynamic radio sky with far greater sensitivity and instantaneous sky coverage. The ability of the SKA to detect a wide range of interstellar molecules will impact the detailed study of planet formation in proto-planetary disks.
SKA Design
The international SKA project has adopted a Baseline Design for the project which incorporates a low frequency aperture array operating at frequencies up to 450 MHz and an array of dishes with single pixel feeds initially operating at frequencies up to 2 GHz but capable of 10 GHz in terms of antenna accuracy. The Baseline Design will enable the key SKA1 science goals of understanding the role of neutral hydrogen in the early Universe, testing theories of gravity in extreme environments, and discovering gravitational waves.
The Project has also identified an Advanced Instrumentation Program (AIP), elements of which could be implemented in Phase 2 of the telescope construction. The AIP includes phased array feeds (PAFs) for the dishes and mid-frequency dense aperture arrays (AA-mid), both of which provide greatly expanded fields of view for fast surveying, as well as wide-band single-pixel feeds (WBSPFs) with much wider frequency coverage than currently available feeds. A review point is built into the schedule to assess whether any of the AIP technologies have reached sufficient maturity to be incorporated in SKA1 to enhance its performance.
Description of Work
Work in the post-2011 period (Preparatory Phase and Pre-Construction Phase) will be carried out on the Baseline Design and the Advanced Instrumentation Program (AIP) with the aim of preparing the international project for start of the construction of SKA1 in 2016. Key to this effort is the system engineering approach and formal project management adopted during the system design in the Preparatory Phase. Eleven areas will be addressed: 1) management of the pre-construction phase and management support such as quality assurance, configuration management, and procurement, 2) science, 3) overall system, 4) maintenance and support, 5) dish sub-system, 6) aperture array sub-systems, 7) signal transport and networks, 8) signal processing, 9) software and computing, 10) power, and 11) site and infrastructure. This Plan provides a top-level Description of Work and a Work Breakdown Structure for these Work Packages, lists of milestones and deliverables, and the project structure, dependencies, and schedule including the planned formal reviews. The Risk Strategy and Risk Management principles are also outlined.
Project Model
The SKA project will have a strong central project office (SKA Project Office, SPO) with management and system design authority. The SKA Organisation, through the SPO, will contract the work on major subsystems to a small number of work package contractors. It is expected that work package contractors will be consortia of Participating Organisations (PO) and industrial partners, but could also be individual companies or POs. By forming consortia, the talent, capacity and ideas required to carry out large work packages can be assembled from several organisations so as to make maximum use of expertise.
The day-to-day project management and coordination will be undertaken by the SPO whose Director will report to a governing Board or Council. The SPO mandate in the pre-construction phase is to manage the successful design and development of SKA technology, and initiate procurement for SKA1 construction, as well as lay the foundations for the later phases of construction, verification, and operation. The main activities and responsibilities of the SPO will be: 1) provide the core management structure for the successful delivery of the SKA; 2) have specification and design authority; 3) provide overall project management, schedule and budgetary control; 4) prepare for and procure, detailed design and verification work from academic / industry consortia (Work Package Contractors (WPCs)); 5) provide project-level system engineering; 6) manage and provide support for contractual interactions between WPCs and the SPO; 7) monitoring of work-package progress and integration of work package deliverables; 8) initiating procurement contracts for SKA1 construction; and 9) own and manage the SKA brand and provide public outreach activities for the project.
Consortia of POs and industry will be required to deliver successfully the large work packages. The POs contain the largest reservoir of domain knowledge in the project, and they will have carried out most of the work to reach the end of the Preparatory Phase. They will be collaborators in setting up the pre-construction project and will carry out work packages on SKA sub-systems. The size and complexity of the SKA indicates that an industry culture in managing and costing the project is essential and that there is close engagement of industry throughout the pre-construction phase.
As the SKA Project moves through the design, development, construction and operational stages, industry will play a crucial role in the delivery and through-life support of the technologies and infrastructure. Industry participation at the pre-construction phase will be underpinned by strategic collaborations with commercial players, among them niche R&D companies, followed by increasing participation through pre-competitive prototyping work, commercial contracts with volume manufacturers, technology systems vendors, site services and installation firms, and power and data transmission specialists.
A survey of the engineering competencies of the Participating Organisations involved in the SKA Precursors, Pathfinders, and Design Studies shows that, for each of the Work Packages in the Pre-Construction Phase, there are a number of potential partners to work together in Consortia as Work Package Contractors. In addition, as part of PrepSKA Work Package 5, a survey of the availability and competencies of potential industrial partners in the SKA around the world is being undertaken. It is premature at this stage to identify strawman sets of partners or particular commercial companies that would form the Consortia.
Resource Requirements
The resources required to carry out this work are a total of 90.9 M€ over the 4 year period, comprising 63.0 M€ for Work Package contracts and 27.9 M€ for SPO costs. The SPO costs include staff costs for project management, system engineering, science support for system engineering, and site work (19.7 M€), and office infrastructure and operational costs for the SPO (8.2 M€). 100 k€ has been adopted for the person-year cost for salary and benefits for both WPCs and SPO. No institute overhead or explicit contingency has been included. It should be noted that the cost of developing the Advanced Instrumentation Program to PDR level in the pre-construction phase is 10 M€.
Governance
The International SKA Organisation is the legal entity that will design, build and operate the SKA. It will be set up and governed to carry out this task on behalf of the SKA sponsors, those agencies who will pay for SKA construction and operation. This organisation will be responsible for setting up the SKA Project Office (SPO) and setting policy on all matters of procurement and operation.
This Plan provides a description of the process now underway to define the optimum legal structure, and the associated optimum location for the new legal entity. The outcome of this process will be known in early 2011. In terms of developing the Execution Plan, it is assumed that a legal structure, which could be in the form of a not-for-profit company, will be established. Behind that legal structure will be a set of statutes under an implementation agreement to which a set of funding partners will accede. It is expected that the SKA Organisation will be governed by a Board/Council that is composed of government and science representatives from each of the contributing countries. The SPO is the operational arm of the SKA Organisation and its Director reports to the Chair of the Board/Council.
The management of the international project in the Pre-Construction Phase will differ in two fundamental respects from that now in operation for the Preparatory Phase. 1) The Board/Council of the new legal entity will replace the current tri-partite governance provided by the SKA Science and Engineering Committee (SSEC), the Agencies SKA Group (ASG), and the PrepSKA Board. 2) SKA design work by POs will be carried out on the basis of formal contracts under the overall authority of the SPO, instead of on the current best-efforts basis. The Board/Council will approve the allocation of major work packages to POs and Industry after a bidding and review process managed by the SPO, or as direct allocations of work following the review process.
Socio-Economic benefits of the SKA
A description of the Socio-economic Benefits of the SKA is provided, based on the outcomes of a meeting on this subject organised by COST in March 2010. The SKA can be expected to stimulate innovation in ICT and sensor technology, serve as a global model for 100% renewable energy, serve as an enabler for improved global science-industry linkages, and stimulate human capital development and employment. The construction and operation of the SKA facilities will impact local and regional skills development in science, engineering and technology and in associated industries. The nature of the SKA will excite the young and old alike about science and technology. The SKA can become part of our culture, using its popularity to inspire the next generation of scientists and engineers.
Dissemination activities
Dissemination has been carried out in many different ways. The statistics are as follows, and the details will be supplied in a separate file
Conferences 7
Meetings 215
Press release/news item 21
Media coverage 6
Videos 2
Factsheets 3
Brochure 1
Newsletters 23
Comic series 3
Leaflet/Poster 1
Poster 3
Website 2
Publications 2
For details of the dissemination activities please see Annex 1: "Dissemination Activities"
For details of the Publications, please see Annex 2: "Publications associated with the SKA Telescope"
List of Websites:
http://prepska.skatelescope.org/
See also link on SKA telescope home page https://www.skatelescope.org/
The Square Kilometre Array is a multi-purpose radio telescope covering the frequency range from 70 MHz to >25 GHz that will play a major role in answering key questions in modern astrophysics and cosmology. It will be one of a small number of cornerstone observatories across the electromagnetic spectrum that will provide astrophysicists and cosmologists with a transformational view of the Universe.
Since 2008, the global radio astronomy community has been engaged in the development of the system design for the SKA as a major part of the Preparatory Phase of the project. The end of the Preparatory Phase is now reached, and a number of major decisions have been made so that, at the end of 2011, the international SKA project can progress to the Pre-Construction Phase (2012-15).
Thanks to the work of the PrepSKA governance and legal framework team, it has been possible to provide the groundwork for establishing the SKA Organisation as a UK legal entity, to govern the SKA project in the preconstruction phase. This culminated with 7 national partners signing the agreement on Nov 23rd 2011, when the SKAO came into being.
A funding model, focusing on understanding funding options in the preconstruction period 2012 to 2017, has been initiated, and a survey of funding agencies’ views on these options has been completed. Information on this topic has been gathered from other large research infrastructures and a report prepared on collective best practice. A tool for modelling funding scenarios has also been developed. The business plan and an initial estimate of the cost of the current technical concept have been delivered.
A report on the social and economic benefits of the SKA has been generated.
The Project Execution plan, covering the preconstruction and Construction phases, has been produced, with a detailed Statement of Work and a Work Breakdown Structure.
A major achievement has been the completion of the site selection and measurement campaign. Expert panels evaluated the information, and as a result, two locations have now been selected by the SKA organisation, one in South Africa, and one in Western Australia. The South African site will host the SKA1-MID frequency array (0.45-3 GHz) comprising 190 new dish antennas and the 64 dishes of the precursor Meer KAT antennas. The number of dishes will rise to 3000 in SKA Phase 2. The Australian site will host both the SKA1-LOW frequency array (0.07-0.45 GHz), consisting of 250,000 antennas grouped into aperture array stations (1,000,000 antennas in SKA Phase 2), and the SKA1-Survey telescope, which will combine 60 new dishes with the 36 ASKAP precursor dishes.
In terms of the detailed design, the major achievement has been the completion of all the SKA Element Concept Design Reviews (CoDRs), and the remaining delta CoDRs. Two additional CoDRs, for the Monitor and Control work and the mid frequency aperture arrays, have also taken place. These reviews were presented to external panels from industry and the astronomical community, and were well attended by other community members. The results lay the groundwork for the subsequent design phase of the SKA. The prototyping of the Dish Verification Antenna 1 is well under way, and construction of the prototype support structure has begun. A small test Aperture Array has been developed, and is being used to start testing antenna element options.
Finally, all the PrepSKA deliverables have been completed and delivered, and the SKA is now ready to move forward into the Pre-Construction Phase.
The goals of the Pre-Construction Phase are to 1) progress the SKA design to the point that Production Readiness Reviews have been successfully completed and contracts for construction of major sub-systems have been let, 2) advance the infrastructure roll-out on the selected site to the point where sub-systems can be deployed (assuming the funds for infrastructure development are made available), and 3) mature the SKA legal entity into an organisation capable of carrying out the construction, verification, and operation of the telescope. Without the foundations laid by PrepSKA, this next stage would not have been possible.
Project Context and Objectives:
The Square Kilometre Array (SKA) is the next generation, multi-purpose radio telescope. It is a transformational instrument that will play a major role in answering key questions in modern astrophysics and cosmology. It will be one of a small number of cornerstone observatories across the electromagnetic spectrum that will provide astrophysicists and cosmologists with a transformational view of the Universe. The SKA will be built in three phases, each phase will deliver ground-breaking science – this staged construction and deployment is made possible by the SKA being a radio interferometer. For the 4 years from 2008 to 2012, the global radio astronomy community has, through the Preparatory Study for the SKA (PrepSKA) project, been engaged in the development of the system design for the SKA as a major part of the Preparatory Phase of the project.
The issues addressed by PrepSKA before the pre-construction phase of the SKA are:
1. The design for the SKA
2 The location of the SKA and the required infrastructure
3 The legal framework and governance structure under which SKA will operate
4 The most cost-effective mechanism for the procurement of components of the SKA
5 Funding mechanisms for the SKA
PrepSKA has coordinated and integrated R&D work from around the globe in order to develop the first costed design for Phase 1 of the SKA, and a deployment plan for the full instrument. With active collaboration between funding agencies and scientists, all of the options for the policy-related questions have been investigated.
PrepSKA has been extremely successful and demonstrated the critical importance of addressing technical, policy and funding issues alongside one another. The true success of PrepSKA can be seen in the successful transition of the project to the next (pre-construction, or detailed design phase) with a new legal entity, the SKA Organisation (SKAO), managing the project. This is a truly global structure with membership consisting currently of 10 countries represented on the SKAO Bboard by national-level agencies (science ministries or funding agencies).
The final SKA telescope will have: frequency range from 70 MHz to more than 25 GHz; a collecting area of over one million square metres to give a sensitivity of up to 50 times higher than the Jansky VLA, currently the world’s most powerful radio telescope; distribution of collectors on scales (baselines) of up to at least 3000 km; an instantaneous field of view (FOV) of up to several tens of square degrees at 1 GHz, many times that of existing instruments and considerably larger at lower frequencies, and the new possibility of multiple simultaneous users of several large, independent fields-of-view. These capabilities are enabled by a much greater use of information and communications technology than in current designs, and the result will be an extremely powerful survey telescope with the capability to follow up individual objects with high angular and time resolution.
The telescope is to be built in twothree phases, and will be able to carry out significant scientific observations as it is being built. As PrepSKA ends the pre-construction phase is being entered, which is expected to last from 2012 to 2016. Then construction will start with Phase 1. Phase 1 (SKA1) is the initial deployment of 10% of the array at low and mid-band frequencies, and is a subset of phase 2 (SKA2), which is the full collecting area at low and mid-band frequencies (~70 MHz to 10 GHz). A further, potential Phase 3 (SKA3) includes the implementation of the higher frequencies, up to 25GHz or more. SKA1 will be a major step forward in terms of sensitivity, survey speed, image fidelity, temporal resolution and field of view. SKA2 will have 10 times the sensitivity and 100-10000 times the survey speed of SKA1, and will transform astronomy. SKA3 lies beyond the current planning assumptions and vision for the project.
The baseline design for the SKA consists of an aperture array operating at low frequencies, with optimal performance in the range 100-250 MHz, with capability over the range 70 to 450 MHz, and an array of 250 parabolic dishes initially operating at frequencies up to 3 GHz , but eventually capable of 25 GHz in terms of antenna surface accuracy, and a 90-dish survey array with each dish equipped with Phased-Array Feeds offering enhanced field-of-view and hence ability to map the sky faster. An Advanced Instrumentation Program (AIP), including wide-band single-pixel feeds (WBSPFs), next generation phased-array feeds (PAFs) for the dishes, and mid-frequency dense aperture arrays (AA-mid), has been identified. These will provide much wider fields of view, thus enabling faster surveying. The wide band single pixel feeds will also provide much wider frequency coverage than current feeds. These technologies are relatively immature now, but may well be implemented in Phase 2.
Technological innovation, closely paralleling commercial IT developments, is the key to the design concepts under investigation and to the target cost of €650 million for Phase 1. As an example, for the Science Data Processor data input rates for SKA phase 1 will be of order 10 Tera-bytes/sec, with near Exaflop capacity required for the central processor. Close collaboration with industry partners will be a core element of the detailed design phase, building on the earlier technology programmes (such as SKADS) and leading to significant impact beyond the radio astronomy community.
Two locations for the telescope have now been selected by the SKA Organisation; South Africa (extending to Southern Africa in SKA2) and Australia. The report from the SKA Site Advisory Committee (SSAC) demonstrated that both sites were well suited to hosting the SKA, but initially identified Southern Africa as the preferred site. Further study by the Site Options Working Group, showed that in SKA1, viable dual‐site implementations exist that not only maintain, but add to, the scientific appeal of the first stage of SKA. A dual site also maximises the financial viability of the project in the longer term, through continuation of the current Organisation membership and a global character that will be attractive to future members. With these considerations in mind the SKAO determined that the most advantageous route to deliver the SKA was to utilise both sites and incorporate as far as possible the infrastructure investments already made by the hosts.
The installations are as summarised in Table 1 (see PDF attached, "SKA Installations"). Australia will host the low frequency aperture array antennas in Phases 1 and 2 of the SKA. The addition of 60 SKA dishes to the 36-dish ASKAP precursor in SKA Phase 1 will expand this survey facility. In South Africa, the addition of 190 SKA dish antennas will expand the 64-dish MeerKAT precursor array in SKA Phase 1. South Africa and eight African partner countries will host the dish array in Phase 2 of the SKA and will also host the Phase 2 mid frequency aperture array antennas.
Project Results:
1 Main Science and technology results and foregrounds
1.1 System Overview
Figure 1 (see attached PDF, “S&T Figures and Tables”)is a system schematic showing the major elements of the SKA telescope. The system concept encompasses all of the parts in Figure 1, including the four receptor types. Figure 1 emphasizes the geographic layout of the system components and the aggregated data flow at important points in the system.
Figure 2 shows artists’ concepts of the technologies that are being considered. The system is designed to cover frequencies from 70 MHz to 10 GHz, which cannot be done sensitively with a single antenna design. These depictions are certainly not designs at this point, and represent only rough impressions of what the receptors might look like in most cases. Physically the system will be implemented as three arrays of antennas, a dish-array, and a sparse aperture array (AA) for use at the lower end of the frequency range and a dense AA for use in the mid-frequency range (see Section 4.2). Both AAs are capable of forming many beams at once, thus covering wide areas of sky. The dishes will be equipped with wide-band single-pixel feeds (WBSPFs) (one or two single wide-
band feed per antenna), forming a single beam on the sky, and/or phased array feeds (PAFs) that are capable of forming a cluster of many simultaneous beams that cover a wider field of sky than the single feed.
The receptors will be arranged on the ground over a very wide area, but concentrated mainly in a central region, “the core”. The pattern of antenna layout is called the array configuration. Figure 3 depicts a representative array configuration currently under consideration. The distribution of collecting area in the three types of receptors will fall off in such a way that the distribution of baseline lengths (the distances between antennas) increases logarithmically from the centre out to a distance of about 180 km. Beyond 180 km the dishes are arranged into stations, which are spaced more widely on so-called spiral arms, and these dishes will be equipped only with single pixel feeds. Wide-field receptors at distances beyond 180 km are not needed and processing very wide fields is impractical due to signal transport constraints. This can be seen in Figure 3, although the precise arrangement of the long baseline stations depends greatly on the geography of the chosen site. There is no unique configuration that will satisfy all the scientific requirements optimally. Part of system design is to obtain a satisfactory array configuration for each of the receptor types and frequency ranges. What is required is to find cost-constrained optimum configurations that are suitable for the available sites, and enable both the imaging and non-imaging science requirements.
The SKA will utilize radio-telescope design techniques developed over the past four to five decades. The most sophisticated concept, aperture synthesis, is an application of the van Cittert-Zernike theorem [24]. Each of the receptors has antennas connected to receivers, which essentially transform the free-space signal to an electrical radio frequency signal, which is digitized and connected to the central Signal Processing Facility (SPF). The signals from the antennas, including both amplitude and phase, are cross-correlated in pairs and integrated to reduce noise. These data can be used to reconstruct the original brightness distribution, spectra at each point in the sky, and sometimes spectral and spatial variations with time.
The SPF (1) is a highly flexible real-time signal processor. Figure 1 depicts the SPF as an on-site facility. This is necessary because the amount of data flowing into the facility from the receptors is very large, and putting it too far away from the core would be impractical. Its purpose is to filter the signals from the antennas into frequency channels and to carry out the cross-correlations noted above. This can be done simultaneously from all or multiple subsets of the antennas, depending on the range of spatial frequency coverage needed by the science program. In addition the SPF can carry out delay-corrected coherent summations of the channelized signals. The summed signals form telescope-beams on the sky, whose sensitivity is related to the entire collecting area used in the beam-sum. These signals will be further processed in real time to look for time-domain signatures, especially pulsar signals, but also other transient signals of astrophysical interest.
In all cases, sufficient on-site processing will be done to produce a manageable data rate to the off-site Science Computing Facility (SCF). Part of the system design work is to find the appropriate processing boundary between these two facilities. The SCF is a more general-purpose computer facility, where calibration of the data takes place, images of sky brightness are formed, and further analysis of time-domain effects are carried out. Algorithms for carrying out calibration and imaging are mature processes for current aperture synthesis arrays, as are techniques for pulsar searching and timing. However, the SKA is likely to require significant new developments in this area to handle the much larger amount of data, and to achieve dynamic range targets without continuous human input.
For the SKA the volume rate of raw data entering the SCF from the SKA will probably be too large to archive directly. Real-time image formation will be required. Images and other data products extracted from the images and time-domain data will be archived and distributed to SKA science centers around the world. Many of the science programs will be large, extended surveys. The survey science teams will each have specific requirements for real-time data processing and for handling/distributing data products.
Figure 1 also shows an off-site Operations and Maintenance Centre. There will have to be a small number of operations and maintenance personnel on-site as well, although it is important to keep this group as small as possible, due to the potential for radio frequency interference. The development of a technical operations plan is closely related to system design.
Finally Figure 1shows regional Science and Engineering Centres. A number of these centres are expected to be built, perhaps one in each region of the world. These centres are where scientific analysis is done, and where new engineering and software upgrades for the SKA are developed. The precise way in which this is to be apportioned and handled in each sector remains to be developed.
1.2 Towards a Final System
Figure 4 shows the receptor technologies in frequency-baseline space. The virtual “potentiometer” settings indicate that the further analysis is needed to find the optimum configurations for these technologies, based on science requirements, performance and cost.
The receptor technologies occupy four main parameter space dimensions: sensitivity (Ae/Tsys), SSFoM ((Ae/Tsys)2Ω), baseline range and frequency range. (For definitions of Ae and Tsys, see [23][24]). In Figure 4 the four receptor technologies are shown schematically in frequency-baseline space. At present it is not clear where the boundaries should lie or even if they should overlap. In fact, the sparse AAs may bifurcate into two frequency ranges and occupy different baseline ranges.
By the time of the Preliminary Design Review the values of the individual receptor parameters will be defined as well as many other aspects of the integrated system. In order to get an understanding of the capabilities of the receptor technologies for delivering science, evaluation of their current and potential capabilities will be on-going.
Since it is currently not possible to define the system that will be built, the method used here is to select one technology concept at a time, parameterized to give the best science return possible. In section 0 the implications of adopting each concept for further investigation are explored by using “representative implementations”, which are essentially implementation examples. These provide an indication of three resulting aspects: 1) scale, 2) the technical flow-down into support services and 3) potential science performance capability. The first two are directly related to cost.
Representative implementations serve the purposes of understanding the science performance of the technology, and of the implications for the potential support functions (data transport, correlation, imaging, etc.) needed in principle to make the candidate technology work. The technical flow-down into support services can be easily derived from the representative implementations. A minimum set of parameters (number of receptors, frequency ranges, bandwidths, etc.) for the representative implementations have been selected so that this assessment can take place.
Parameters have also been selected so that a particular technology can handle as much of the science as possible. This does not imply that the final system will consist of a simple sum of everything. At the concept stage we do not know the size or scope of each of the technologies. Thus in subsequent system analysis the technologies adopted will depend on an optimized subset of those technologies, based on science-technology trade-offs.
The descriptions of the representative implementations are not to be construed as adopting particular system or technology parameters, nor are they related to detailed design. They are mainly a means of evaluating the potential cost and performance of each of the candidate technologies separately.
1.3 An SKA Data Flow and Control Slice
Figure 5 emphasizes the signal path and key signal interfaces.
Figure 5 is a system schematic showing a “slice” of the system, starting on the left at the first active component and following the signal/science-data path to the regional science centres. Also shown are the main ancillary sub-systems needed to support the system: time and frequency distribution, monitor and control, power, temperature control sub-systems. Each of the candidate technologies contains the blocks shown in Figure 5, but the physical arrangements and scales are very different in each case.
The RF signals originate from elemental antennas that form part of an aperture array, from elemental antennas that form part of a phased array feed or from the single-pixel feeds on the dishes. The signal is received by a low noise amplifier (LNA) in each case, which amplifies the signal typically by a factor of 1000 in power. The LNA is connected to an analogue receiver, which further amplifies the signal, typically by a factor of 107 or more. The receiver also filters the signal, and optionally translates the signal to more convenient frequencies.
If the receiver is part of an aperture array, then several such signals are optionally summed in an analogue beam-former, which may also compensate for geometrical delay. This is likely to be only part of the beam-forming process, which in general is more efficiently carried out in hierarchical stages. The signal is then converted to a digital stream via an analogue-to-digital converter (ADC). Further digital beam-forming is carried out, probably on many more digital streams than in the analogue case. This may optionally be passed to a second stage of digital beam-forming, which in the case of an aperture array would be produced using an entire station. The most likely method of digital beam-forming will also require that the signal be channelized into frequency bins.
If the receiver is part of a phased array feed on a dish, then analogue beam-forming is unlikely. The digital beam-former is similar in principle to that for the aperture arrays, but is used to form beams emanating from positions anywhere over the surface of the feed. Frequency channelization is also required for precise illumination of the dish.
If the receiver is part of a single pixel feed on a dish, only an ADC is needed in this part of the signal chain.
Signals from dishes with single-pixel feeds, located further than ~180 km from the centre, are likely to be combined from stations, defined closely spaced groups of antennas. This is done because the cost of transporting data over long distances could be prohibitive if each output signal from each dish were sent separately, and the processing load would also be very high in this case. Instead these signals are beam-formed, and the signals for a few beams are transported to the centre. Optimising the radius from the centre at which stations are formed is part of the on-going system design process.
All the digital signals flow into the SPF via an extensive signal transport network (blue lines in Figure 5). In principle there could be an SPF for each receptor group, but here it is assumed that they are co-located. The correlator channelizes and cross-correlates the signals from groups of antennas (sub-arrays) or the full array of a receptor technology. In principle it could cross-correlate signals from different technologies if they have overlapping frequency capability, although this is complicated. Cross-correlated output from each sub-array is passed to the science processing facility. Additional operations on the data, such as time-based integration and conversion to floating point number protocol, may be carried out after correlation before transport to the SPF.
The beam-former and time-domain processor carries out time-domain analysis of the same data (or a sub-set) entering the SPF, principally to search for or time pulsars, but also to look for radio transient signals. Signals from the full array or sub-arrays of dishes and/or aperture array (AA) stations will be beam-formed in the SPF before time-domain analysis takes place. Channelization is also required, but this may be shared with the channelization need for correlation. The results of the time-domain analysis are also sent to the science processing facility.
Science processing for “standard” imaging consists of several fundamental steps: calibration, gridding the data into a rectangular grid in the u-v plane, Fourier transformation into the sky intensity domain, and deconvolution of the point-response function of the array. This is done for each frequency bin. These steps are combined in an iterative loop that typically involves solving for array parameters, such as complex gains, concurrently with solving for the image. A buffer may be needed to store visibilities and/or interim results while the loop is being executed. The results of this processing usually form the basis for science interpretation.
The time-domain processing done in the science processing facility is more difficult to define at this stage. The steps needed to find pulsars are well developed, but it is not clear what fraction of that work will be done in the general purpose science processing facility.
The science processing facility will require one or more large data storage sub-systems. These will be needed to buffer incoming data from the SPF, and to store data products.
Figure 5 shows several other significant sub-systems: 1) Local Oscillators and ADC clock signals used in the system must be derived from a common frequency standard. 2) A distributed monitor and control sub-system is needed, complete with a large operations centre. Each block in Figure 5 shows an M&C unit to emphasize the importance of this sub-system. 3) The electronics and other equipment will certainly require cooling and some require temperature control. 4) The distribution of electrical power will form a major part of the system.
1.4 System Hierarchy
The system hierarchy is intended to provide a clear and coherent view on the scope and composition of each of the building system blocks. It is included here, but discussed only very briefly. A more complete description of its purpose is provided in the System Engineering Management Plan [25]. The hierarchy as shown is not intended to be the final version and is likely to evolve. Once finalised, it sets the nomenclature for the levels throughout the project. There are many ways in which the hierarchy could be constructed, depending upon the emphasis intended to be shown. Figure 5 shows the groupings and levels selected for this version. The diagram should be taken in its entirety, but is divided in two parts to permit formatting in a document. The left side of the diagram (top part in Figure 6) shows receptor technologies and their components, and the right part of the diagram (bottom part in Figure 6) shows the support services and facilities. A functional version might look somewhat different. The levels 0 –7 show successive levels of detail and size. The PrepSKA phase of the project is unlikely to be concerned with levels below 3.
2 Reception
This section of the document will start off with an introductory section to the various candidate receptor technologies. The focus will be on the receptors because they are the main enablers to cover the science space. Following this introductory section each of the receptor technologies will be described in more detail and will then include identification and description of the rest of the system that will be required to support each of the receptor technologies.
Figure 6 shows candidate technologies configured for the representative implementations. The technologies and locations on these diagrams are preliminary and may change as system knowledge evolves. Each is designed to carry out observations to support one or more components of the DRM and as much additional underlying science as possible. SAA denotes sparse AA; DAA denotes dense AA; WBSPF denotes wide-band single-pixel feed; PAF denotes phased array feed.
2.1 Candidate Technology 1: Reflector Technology with Wide-band Single-Pixel Feeds
2.1.1.1 Receptor Technology
In this representative implementation the antenna array for the SKA will comprise a large number of ~15-m equivalent diameter offset-Gregorian dish systems, each equipped with two or more wide-band single-pixel feeds (WBSPFs) to provide overall frequency coverage of 0.3 to 10 GHz. Use of cryogenic receiver front-ends coupled to the feeds will maximize the sensitivity and survey speed of the array. Following the front-ends, the remaining sections of the receiver subsystem will provide gain, frequency conversion, digitization and optical modulation, such that the antenna system outputs will be digital bit streams on fibre-optic cables. Clock and local oscillator reference signals to the receiver subsystems will also be distributed via fibre optic cables.
Offset-Gregorian optics have been provisionally chosen for the SKA for the following reasons:
• There will be no blockage of the main reflector and scattering of ground noise onto the main reflector can be minimized by enshrouding the feed/subreflector area.
• Sidelobe levels will be significantly lower than for an axisymmetric antenna, which will increase the achievable image dynamic range over wide instantaneous fields of view.
• There will be sufficient space for a single-pixel feed changer and mechanical rotator for a phased array feed, if required.
The last point is the most important one for this assessment, since it maximises the chance of being able to accommodate several feed systems.
Cryogenic dewars will house the low noise amplifiers (LNAs) for higher frequencies, one per polarization for each feed, at the front-end of the receiver subsystems. At the lowest frequencies it may be possible to use LNAs and feeds operating at ambient temperatures. A strong candidate for cryo-cooling are Stirling cycle refrigerators. These have been used with great success in the mobile phone industry, and have demonstrated excellent reliability. Cooling the high band feed, LNAs and coupling transmission lines will minimize the system noise temperature, and hence maximize the SKA’s sensitivity and survey speed. This will be needed to achieve the noise temperatures shown in Table2 (see attached PDF,” S&T Figures and Tables”).
The remainder of the receiver subsystem will be temperature stabilized to maximize amplitude and phase stability. Signals will be filtered and/or frequency converted, digitised and transmitted to the Central Data Processor via optical fibre. Use of digitized fibre-optic signal transport will ensure the best possible electromagnetic compatibility between subsystems, minimizing the potential for self-generated radio interference and maximizing immunity to external interfering signals. The digitized signals will also be unaffected by environmental and mechanical disturbance to the optical fibres. These measures, as well as others, will be needed to achieve the dynamic range specifications. These comments about optical fibre signal transport will also apply to the other technologies under consideration for the SKA.
As an alternative, less expansive analogue transmission of RF signals over fibre could be used in parts of the RF system if they can be shown to meet the requirements.
2.1.1.2 A Representative Implementation of a Dish-based System with WBSPFs
Table 2 describes the top-level specifications for a representative implementation of this system, which can be used to assess its performance in addressing the components of the DRM. The system comprises 3300 15-m antennas equipped with single-pixel feeds. The feeds cover bandwidth ratios (upper frequency/lower frequency) of ~4:1. This is less than the goal of 10:1 often discussed for this technology. It is currently more likely that 4:1 feeds, rather than 10:1 feeds, can actually meet the Tsys and aperture efficiencies listed in Table 2. Three feeds will be needed to cover the required frequency range. It is assumed that only one feed can be used at once and that a remotely-controlled mechanical feed changer will be needed to place a particular feed at the focus.
Table 3 contains the main performance factors derived from Table 2 for this system.
2.1.2 Justification for Inclusion of WBSPF/Dish Technology
Reflector antennas (dishes) in general have been the standard receptor technology in radio astronomy for wavelengths less than ~100 cm for decades. Their key advantages are extremely wide frequency range, and the fact that they are passive collectors. Receivers have improved steadily in cost and performance, although cryogenics for frequencies greater than 600 MHz are important maintenance cost items. Recent research in antenna construction promises reduction in antenna costs as well. Both trends indicate that this technology will be suitable for use in SKA construction, while continuing to drive up the performance/cost ratio at the subsystem level.
The WBSPFs, which at this stage are a technology development item, have the potential to increase the frequency range of the reflectors to more closely match that of the reflectors. Considerable progress has been made recently. Tests of several types of feeds have been done in the lab and have shown sufficient promise to warrant inclusion in a dish based system concept for the SKA. A more detailed discussion of risks and fall-back positions is contained Section 11.2.1.
The main disadvantage of reflector antennas is the relatively narrow instantaneous field-of-view. This can, in principle, be improved by an order of magnitude by using phased array feeds (see Section 6.4) which at this stage are also a technology development item. Aperture arrays, which are a completely different approach to receptor design, can provide much larger instantaneous fields-of-view.
Dish technology covers a large fraction of the science represented by the DRM and also provides a flexible telescope design to carry out observations that have not yet been considered. This last category is the one in which radio astronomy has made the largest impact historically. Of concern are the “OK” and lower categories.
In summary, with relatively small changes in science-technology trade-offs, the performance of this system could be improved. Adding sparse aperture arrays to this system could potentially enable all except one (HI Baryon Acoustic Oscillations) to be carried out in combination.
2.2 Candidate Technology 2: Reflector Technology with Phased Array Feeds
2.2.1.1 Receptor Technology
A Phased Array Feed (PAF) is a phased array, composed of closely packed, dual polarized, elemental antennas approximately a wavelength in size, and placed on a grid in the focal plane of an optical system of a reflector antenna. The array of elements spatially samples the electromagnetic field in the focal region of the dish. Weighted sums of amplified signals are formed, which have the effect of producing a cluster of overlapping beams on the sky. A beam-former is needed somewhere in the system to carry out the sum. The cluster of beams spans a much wider field-of-view than a single beam.
In another application, smaller PAF systems may also be used which do not increase the field-of-view (FoV), but allow more flexibility in determining the illumination of the reflector, thus increasing antenna efficiency or controlling the level of sidelobes. This general technology is relatively new to radio astronomy and has not yet been used on production telescopes. In the SKA context some or all of the dishes could be equipped with PAFs, each of which would contain about 10~2 elements.
The PAF array is quasi-circular in outline, the center of which is placed on the optical axis of the reflector system. The radius of the PAF from the axis determines the radius of the FoV on the sky. This is entirely geometric and independent of wavelength, in contrast to dishes with SPFs and AAs, where the radius of the FoV increases with for a given dish or AA diameter.
A low noise amplifier (LNA) is needed for each elemental antenna. Because cryogenically cooled systems are currently practical for only small volumes, the LNAs on PAFs must be operated at ambient temperature at L-band (~1 GHz) frequencies. (At higher frequencies they could be sufficiently small to permit the entire array to be cooled). This means that system temperatures will be higher than for single pixel feeds, which have only two LNAs per feed and can be cryo-cooled.
Because PAFs have multi-parameter control over the illumination of the reflector, higher system temperature can be offset by better antenna aperture efficiency for PAFs than single pixel feeds. For typical parameters, a two percentage point increase in aperture efficiency offsets approximately a degree of system temperature. In addition, it may also be possible to provide real-time control of sidelobe levels in dish-based systems.
Introducing PAFs into a dish-based version of the SKA has a potential impact on the architecture of the system. The total FoV and the number of dishes, which would be linked to dish diameter, can be chosen separately in the design of a synthesis telescope of given collecting area. One attraction of this approach is that a wide FoV, and high survey speed, may be obtainable with relatively few dishes, simplifying cross-correlation and post-processing. With correlation load scaling as the square of the number of antennas, the PAF approach in which the correlator complexity scales only linearly with total FoV, is conceptually attractive.
The SKA will need wide receiving bandwidths, which require that the signals from each element be channelized before being summed. The beam-former, including channelization, may reside either at the antenna, at a location that processes inputs from several antennas, or at the SPF (the central processing facility of the entire array). In all cases this would be done digitally. Each beam is processed at the SPF in the same way as for a single-pixel feed system (correlation, imaging and time-domain processing).
For pulsar astronomy, a beam-former is needed for a large central sub-array of antennas. These beams would most likely be formed from signals from the central PAF beam in a similar way to that for single-pixel fed antennas.
The following relations can be derived directly from specifications of the dish diameter, D, focal length, f, the diameter of the PAF, d, and the wavelength within the receiving band, . The total field-of-view in steradians, ΩPAF, is given by
The area of a single beam on the sky in steradians, Ωbeam, is given by
And the number of beams, Nbeam, within the PAF field-of-view is given by
As noted above, ΩPAF is independent of wavelength, but the number of beams depends on -2. This means that the number of output data streams is dependent on the shortest wavelength in the band.
2.2.1.2 A Representative Implementation of a Dish-based System with PAFs
The purpose of the representative implementation is to provide guidance on the implications of performance, complexity and cost of an entire system based on PAFs as the candidate technology.
This implementation is designed to be compatible with a similar dish-based system as for WBSPFs, but using PAFs instead of WBSPFs. Note that these dishes could in principle be equipped with a combination of PAFs and single-pixel feeds (wideband or standard bandwidth) on each dish. Each dish contains two PAFs, which each cover 2.5:1 frequency ranges. This is thought to be largest bandwidth ratio possible with PAFs. Together their coverage from 500 MHz to 3 GHz is designed to capture as many components of the DRM as possible. This choice may not be optimum, even for a system containing only PAFs, and certainly not for a system containing other technologies as well. Principally because red-shifted HI gets weak rapidly at higher redshifts, the increased survey speed of the PAF-based system could be more useful at lower frequencies than 500 MHz. But because the PAF may become too large at wavelengths approaching a meter, the lower limit has been set at 500 MHz.
A PAF diameter of 1.5 m has been chosen for the lower frequency band (0.5 – 1.2 GHz), which provides a FoV of ~160 deg2. The second PAF has been sized to provide the same number of beams, so that the downstream processing system is similar for both. This diameter scales to 0.6 m, which provides a FoV of ~26 deg2. In both cases the number of beams needed to cover the FoV is about 172 in each polarization. It is assumed that only one feed can be used at once and that a remotely-controlled mechanical feed changer will be needed to place a particular feed at the focus.
The following relation for correlator load, Ccorr, provides some guidance on setting other system parameters: , where Nant is the number of antennas, ΩPAF is the field-of-view, Atot is the total effective collecting area, Bbit is the total bandwidth in bits, and is the wavelength. The size of the correlator need not be a focus of design attention, but it must be kept within reasonable bounds. In this implementation, it was used to provide guidance on setting the number of dishes and the number of beams transmitted to the correlator. There are many other ways of doing this, and there is no intent here to be other than illustrative. There is also a concern with the amount of data transport needed.
The number of dishes has been chosen as 2000. With 2000 dishes and all 172 beams, the correlator load factor (see Table 4) would be uncomfortably high. Rather than reduce the number of dishes or ΩPAF, a beam dither factor has been introduced. Beam dithering in this context means instantaneously switching back and forth between different sets of beams, which are interleaved with each other. It is likely to be needed in a PAF-based system to produce uniform sensitivity over the FoV anyway – otherwise the sensitivity will be “lumpy”. A factor of four has been used here, rather arbitrarily, to limit the number of beams produced instantaneously to 43 (dual polarized). This strategy does, however, provide an easy means of increasing the capability of the telescope if more downstream resources can be provided later. Dithering also introduces frequency dependence into the survey speed calculation, because a reduced number of beams is already sufficient to cover the FoV at the low end of the frequency band (see Table 5).
Many of the other system parameters flow from this small number of initial choices. Additional parameters, such as the number of frequency channels and the integration time can be derived from the above for continuum observations, or from the width of the HI-line. For continuum observations, the number of channels and the integration time depend on the maximum frequency and baseline, which are observation dependent, and have been chosen here as rather rough mean values.
The array configuration has been chosen for this case to have the same distribution of baselines as the WBSPF case, except that baselines out to only 180 km have been included. Although there may be a science case for very wide FoVs at long baselines, the image processing load would be extreme.
2.2.2 Justification for Inclusion of Dish/PAF Technology
PAFs are clearly justified for survey science, where SSFoM is the most important performance factor, and a major driver for the SKA as a discovery instrument. Using PAFs instead of single-pixel feeds allows a choice of designs in which the FoV and the dish size can be chosen independently, thus permitting the designer to balance survey speed and point source sensitivity. Thus it may be possible to reduce the number of SKA antennas at the expense of putting PAF technology at the focus rather than a single pixel feed. Since the electronics required for PAFs is likely to be less expensive than the cost of antennas (now or in the future), this may be an important means of getting science from the SKA most efficiently.
If high frequency PAFs can be shown to provide better overall performance in the future, they would be justified for frequencies above 3 GHz.
In an SKA system they would have to be supplemented with single-pixel feeds at high frequencies, or small, cryo-cooled PAFs if they prove to be better. The lowest frequencies would require the addition of sparse AAs to the system. In a sequential SKA development scenario in which dishes are sequentially equipped with different feeds, they would likely play an important role.
2.3 Candidate Technology 3: Dense Aperture Array
2.3.1.1 Receptor Technology
An aperture array is an all-electronic collector consisting of a large number of small individual receiving elements fixed on the ground. Beams are formed and steered electronically by combining the signals from the elements with the appropriate delays and weights. The array can form a large number of simultaneous beams if there is sufficient processing capability and communication bandwidth. The result is a very flexible collector with no moving parts.
Aperture arrays can be designed in two principal configurations: dense, where the spacing between the receiving elements is less than the Nyquist special sampling of ~λmin/2; and sparse where the sampling pitch is > λmin/2.
Considering the characteristics of dense Aperture Arrays, the useable frequency band of a dense array can be extended beyond the sampling frequency into the sparse regime, which will cause a reduction of scan angle range and effective collecting area. The benefits of dense arrays compared to sparse arrays can be summarized as:
• Nearly constant Aeff over the frequency band,
• Absence of grating lobes and the ability to perform full beam shape control
• Low element cost and low processing cost due to the regular physical structure, enabling a high level of production automation and Fast Fourier Transform processing.
The effective collecting area of a single, isolated element of an aperture array scales as λ2, hence the number of elements required is determined by the Aeff required at the top operating frequency in the dense regime. This naturally limits the highest frequency that is likely to be economically implemented by an AA system currently to ≤1.4 GHz. However, as can be seen from the DRM, the requirement for very high survey speeds is also largely limited to below this frequency.
For the implementation of dense aperture arrays, a number of antenna elements are available. The basic characteristic of these elements is a high degree of electromagnetic coupling, resulting in a ‘conductive sheet’ across the array.
2.3.1.2 A Representative Implementation of a Dense Aperture Array
Table 6 describes the top-level specifications for a representative implementation of this system, which can be used to assess its performance in addressing the components of the DRM.
The system comprises 250 56-m Dense Aperture Array stations populated by over 75,000 Vivaldi antennas spaced at 21 cm. This results in the sensitivity of more than 10,000 m2/K out to a scan angle of 45 degrees from zenith. The aperture efficiency is programmable and can be traded against side lobe control and sensitivity. A nominal figure of 80% is used within this document though higher levels might be used for pulsar searches where the presence of large side lobes is less of an issue. The target system temperature is assumed to be 37K and the collection area is proportional to the cosine of the angle from the zenith which equates to 70% of the bore sight area at 45 degrees. Such an array is under-sampled at frequencies above 700 MHz but electromagnetic coupling between the elements should ensure that the sensitivity remains good at 800 MHz and decreases only slightly with frequency up to 1 GHz. At zenith and ~800 MHz the sensitivity would be very close to 13,000 m2/K. Beyond 1 GHz the sparse nature of the array begins to impact on the sensitivity/ survey performance.
2.3.2 Justification for Inclusion of Dense Aperture Arrays
The purpose of building the SKA is to perform challenging science; dense aperture arrays potentially provide access to extended parameter space for radio telescopes which can enable experiments that may not be possible with other technologies. The areas that dense AAs have the greatest impact are:
• Very high survey speed capability,
• Multiple beams for expanded FoV, concurrent observations, and minimisation of correlator and central processing resources,
• High dynamic range capability due to small beams (large diameter collectors), unblocked aperture and good physical stability,
• Flexibility to observe short time period transients over large areas of the sky, coupled with the possibility of keeping a history buffer to view the precursors to a transient.
AAs substantially extend the survey parameter space that the SKA can operate over and as can be seen from section 6.5.1.3 above have a major impact on the science capabilities of the SKA.
AAs are best suited for lower frequencies. Combining dishes at high frequencies with Dense and Sparse Aperture Arrays could potentially enable all of the DRM components to be observed depending on the number of dishes that can be afforded within the financial envelope (for DRM components 6 and 13 in particular). DRM components 15 & 1 may require dishes with low frequency capability on long baselines
There is the potential of achieving the dynamic range requirements of the SKA below 1.4GHz with AAs due to the use of very large diameter collectors (with many beams), and the potential to accurately calibrate at every frequency, direction and location across the receiving ‘surface’.
The costs of implementing AAs is reducing over time due to the extensive use of IT related technology, which is following the performance and cost improvement paths shown by e.g. Moore’s Law. This means that the AAs get progressively more affordable over time and are capable of substantial and meaningful upgrade over the life of the SKA.
2.4 Candidate Technology 4: Sparse Aperture Array
2.4.1.1 Receptor Technology
The Sparse Aperture array concept for the SKA consists of a large number of antenna elements which are spread over a wide field with separations longer than λmax/√2 [19] to each other. Distributing the antennas in such a way maximizes the collecting area over almost all frequencies. The number of antennas required is determined by the sensitivity requirements of the telescope. The antenna technology used can range from simple inverted V dipoles (used for the LOFAR Low Band Antennas) to a more sophisticated technology such as log periodic structures. Both types of antenna are about a wavelength in size.
Below approximately 450MHz sky noise starts to increase dramatically (~ λ2.55) and dominates the system temperature, Tsys. For a constant sensitivity, Aeff/Tsys the array area needs to increase in a similar fashion with respect to reducing frequency. A sparse array dipole has the characteristic that its effective area increases as a function of λ2 and as such provides a reasonable compensation for the increase in the sky temperature. However, if the array is completely sparse down to the lowest frequency requirement of 70 MHz, its diameter will be large and will impact on the amount of down-stream processing required. For the representative implementation, the array is dense up to 100 MHz and sparse above 100MHz.
Each individual antenna has a sensitivity pattern of approximately cos(za), where za is the zenith angle. Since each individual antenna generates a significant amount of data, antennas are typically grouped as individual phased arrays further referred to as stations. This reduces the data volume significantly. The beam width of the beam-formed signal is reduced to ~λ/D with D the diameter of the station. However, the FoV can be increased by forming multiple independent beams over the sky. In this way a trade off between, for example, bandwidth and beams can be made for a fixed total data rate.
The main science cases addressed with sparse aperture arrays are those which require low frequencies. The optimum transition between the sparse and dense regime for aperture arrays lies at approximately 450 MHz [20]. All science cases below this frequency benefit from the very large FoV of a sparse aperture array, increasing as λ2.
The sensitivity of a sparse aperture array system grows linearly with the total number of antennas installed. The relationship between sensitivity and the number of antennas in a station is not necessarily linear since it depends on the detailed station configuration. However, since the array is sparse to first order, sensitivity grows linearly with the number of antennas within a station.
Below approximately 450MHz sky noise starts to increase dramatically and dominates Tsys. Consequently Aeff needs to increase with reducing frequency to maintain the required sensitivity, Aeff/Tsys. Above 450 MHz the sky noise is relatively constant and the Tsys is largely determined by the technical performance of the array. The effective area for a sparse array increases as λ2, largely countering the effects of increasing sky noise.
Regular sparse aperture arrays will suffer from grating lobes. However, grating lobes will be spread out for randomized station configurations. Moreover, the effect of side lobes to the correlation products can be further reduced by rotating different station configurations with respect to each other.
2.4.1.2 A Representative Implementation of a Sparse Aperture Array
The representative implementation considers the 180 metre diameter Sparse Aperture station with antennas regularly spaced at 1.5 metre separation and a total of 250 stations. This results in the array being sparse above 100 MHz and dense below.
The individual antennae need to be simple and cost effective as a very large number are required to provide the required sensitivity, particularly at the low end of the frequency spectrum. The front end amplifiers do not need to have as low a temperature as other antenna due to the dominance of the sky temperature as part of Tsys. This instrument temperature is assumed to be 50K.
The number of antennas per station versus the total number of stations is a balance between correlation and post processing costs and the required u-v coverage. The cost of fixed AA collecting area (fixed number of elements) is to first order not dependent on the number of stations. u-v coverage generally improves with more stations; however, the beam size would increase with smaller diameter stations which may bring dynamic range limitations. While this will be studied, it is likely that around 250 stations are appropriate since this gives good u-v coverage.
The configuration for an aperture array system is determined by the following parameters:
1. minimum size required for station calibration (this is necessary to correct for gain and phase differences of the antennas elements prior to beam-forming),
2. sky resolution required (determines maximum baseline),
3. instantaneous imaging capability (how much u-v coverage is necessary for snapshot imaging),
4. the required u-v coverage in synthesis mode (from the minimum amount of time required for a certain sky coverage),
5. infrastructure and data transport costs.
Digitization, beam-forming and optical modulation for the Sparse Aperture Arrays is not performed at the antenna but at one or more processing units within each station. The resultant beam outputs will be digital bit streams on fibre-optic cables. Clock and local oscillator reference signals to the receiver subsystems will also be distributed via fibre optic cables.
2.4.2 Justification for Inclusion of Sparse Aperture Arrays
Sparse aperture arrays are the only solution able to meet the science goals requiring frequencies less than a few hundred MHz and a large instantaneous FoV. Furthermore the use of multiple beams opens the path to simultaneous observations or increasing the survey speed of the instrument. Also fast switching or re-pointing of beams is an advantage of electrically steering the beam onto the sky.
The sparse aperture arrays can be used for DRM Components 4, 5, 6 and 11 in conjunction with Dense Aperture Arrays and/ or WBSPF Dishes or Dishes with PAFs and WBSPF.
Potential Impact:
Potential Impact of the PrepSKA project
Introduction
The impact of the PrepSKA programme is described in full in the Project Execution Plan, which is the blueprint for the development of the SKA telescope in the preconstruction phase.
The Square Kilometre Array is a multi-purpose radio telescope covering the frequency range from 70 MHz to >25 GHz that will play a major role in answering key questions in modern astrophysics and cosmology. It will be one of a small number of cornerstone observatories across the electromagnetic spectrum that will provide astrophysicists and cosmologists with a transformational view of the Universe.
For the past 2.5 years, the global radio astronomy community has been engaged in the development of the system design for the SKA as a major part of the Preparatory Phase of the project. The end of the Preparatory Phase is now approaching, and a number of major decisions need to be made so that, at the end of 2011, the international SKA project can progress to the Pre-Construction Phase (2012-15). These decisions include approval of funding for the completion of the Preparatory Phase and for the Pre-Construction activities, the establishment of a legal entity for the SKA Organisation, and the selection of the SKA site. (A single site for the SKA is assumed in this document.)
The goals of the Pre-Construction Phase are to 1) progress the SKA design to the point that Production Readiness Reviews have been successfully completed and contracts for construction of major sub-systems have been let, 2) advance the infrastructure roll-out on the selected site to the point where sub-systems can be deployed (assuming the funds for infrastructure development are made available), and 3) mature the SKA legal entity into an organisation capable of carrying out the construction, verification, and operation of the telescope.
This Plan sets out the strategies for carrying out the Pre-Construction Phase, the work to be done and resources required to complete the Preparatory Phase and to carry out the Pre-Construction Phase, potential partnerships to carry out the work, and the governance principles for the legal entity forming the SKA Organisation in this phase. It will enable the start of construction of the first phase of the SKA and convergence on a viable technical pathway to the full SKA. The goal of this Plan is to provide the Funding Agencies and Governments in the Agencies SKA Group with the appropriate information to allow them to assess the scope and feasibility of the work proposed and, in some cases, facilitate the funding for the Pre-Construction Phase.
Phased implementation of the SKA
The construction of a radio telescope with a collecting area approaching one million square metres across a wide frequency range is a major undertaking and is planned to be implemented in phases in order to spread the cost impact. Phased implementation is an effective strategy for an aperture synthesis telescope which can start operating before construction is completed. The international project has adopted the following terminology to describe this phased approach: SKA1 is the initial deployment (10%) of the array at low and mid-band frequencies costing 350 M€ (in 2007 currency units) and is a sub-set of Phase 2 (SKA2), SKA2 is the full collecting area at low and mid-band frequencies (~70 MHz to 10 GHz), while Phase 3 (SKA3) sees the implementation at higher frequencies of up to 25 GHz or more. The SKA1 facility will represent a major step forward in terms of sensitivity, survey speed, image fidelity, temporal resolution and field-of-view. It will open up new areas of discovery space and demonstrate the science and technology underpinning SKA2. SKA2 will have 10 times the sensitivity and 100 -10000 times the survey speed of SKA1 and will transform astronomy.
Project Schedule
The project schedule for the low and mid-band frequencies is divided into six major phases: 1) Preparatory (2008-2012), 2) Pre-construction (2013-2015), 3) SKA1 construction (2016-2019), 4) SKA1 operations (2020), 5) SKA2 construction (2018-2023), and 6) SKA2 operations (2024). The schedule for SKA3 is not well-defined at this point in time.
Science Drivers
The five key science areas defined by the astronomy community as driving the specifications of the SKA, together with the Exploration of the Unknown, are described in detail in “Science with the Square Kilometre Array” (eds. C. Carilli and S. Rawlings) [1]. These drive the system design for the telescope. SKA1 will enable revolutionary science at decimetre wavelengths, with a particular focus on pulsars and gravitational wave astronomy, and H I in the distant and the nearby Universe. With its wider wavelength range, 10 times greater sensitivity than SKA1, and enormous increase in survey speed, SKA2 will transform our understanding of many key areas including: the formation of the first structures as the universe made its transition from a largely neutral state to its largely ionised state today; cosmology including dark energy via baryonic oscillations seen in neutral hydrogen; the properties of galaxy assembly and evolution; the origin, evolution and structure of magnetic fields across cosmic time; strong field tests of gravity using pulsars and black holes including measurements of black hole spin and theories of gravity, and the exploration of the dynamic radio sky with far greater sensitivity and instantaneous sky coverage. The ability of the SKA to detect a wide range of interstellar molecules will impact the detailed study of planet formation in proto-planetary disks.
SKA Design
The international SKA project has adopted a Baseline Design for the project which incorporates a low frequency aperture array operating at frequencies up to 450 MHz and an array of dishes with single pixel feeds initially operating at frequencies up to 2 GHz but capable of 10 GHz in terms of antenna accuracy. The Baseline Design will enable the key SKA1 science goals of understanding the role of neutral hydrogen in the early Universe, testing theories of gravity in extreme environments, and discovering gravitational waves.
The Project has also identified an Advanced Instrumentation Program (AIP), elements of which could be implemented in Phase 2 of the telescope construction. The AIP includes phased array feeds (PAFs) for the dishes and mid-frequency dense aperture arrays (AA-mid), both of which provide greatly expanded fields of view for fast surveying, as well as wide-band single-pixel feeds (WBSPFs) with much wider frequency coverage than currently available feeds. A review point is built into the schedule to assess whether any of the AIP technologies have reached sufficient maturity to be incorporated in SKA1 to enhance its performance.
Description of Work
Work in the post-2011 period (Preparatory Phase and Pre-Construction Phase) will be carried out on the Baseline Design and the Advanced Instrumentation Program (AIP) with the aim of preparing the international project for start of the construction of SKA1 in 2016. Key to this effort is the system engineering approach and formal project management adopted during the system design in the Preparatory Phase. Eleven areas will be addressed: 1) management of the pre-construction phase and management support such as quality assurance, configuration management, and procurement, 2) science, 3) overall system, 4) maintenance and support, 5) dish sub-system, 6) aperture array sub-systems, 7) signal transport and networks, 8) signal processing, 9) software and computing, 10) power, and 11) site and infrastructure. This Plan provides a top-level Description of Work and a Work Breakdown Structure for these Work Packages, lists of milestones and deliverables, and the project structure, dependencies, and schedule including the planned formal reviews. The Risk Strategy and Risk Management principles are also outlined.
Project Model
The SKA project will have a strong central project office (SKA Project Office, SPO) with management and system design authority. The SKA Organisation, through the SPO, will contract the work on major subsystems to a small number of work package contractors. It is expected that work package contractors will be consortia of Participating Organisations (PO) and industrial partners, but could also be individual companies or POs. By forming consortia, the talent, capacity and ideas required to carry out large work packages can be assembled from several organisations so as to make maximum use of expertise.
The day-to-day project management and coordination will be undertaken by the SPO whose Director will report to a governing Board or Council. The SPO mandate in the pre-construction phase is to manage the successful design and development of SKA technology, and initiate procurement for SKA1 construction, as well as lay the foundations for the later phases of construction, verification, and operation. The main activities and responsibilities of the SPO will be: 1) provide the core management structure for the successful delivery of the SKA; 2) have specification and design authority; 3) provide overall project management, schedule and budgetary control; 4) prepare for and procure, detailed design and verification work from academic / industry consortia (Work Package Contractors (WPCs)); 5) provide project-level system engineering; 6) manage and provide support for contractual interactions between WPCs and the SPO; 7) monitoring of work-package progress and integration of work package deliverables; 8) initiating procurement contracts for SKA1 construction; and 9) own and manage the SKA brand and provide public outreach activities for the project.
Consortia of POs and industry will be required to deliver successfully the large work packages. The POs contain the largest reservoir of domain knowledge in the project, and they will have carried out most of the work to reach the end of the Preparatory Phase. They will be collaborators in setting up the pre-construction project and will carry out work packages on SKA sub-systems. The size and complexity of the SKA indicates that an industry culture in managing and costing the project is essential and that there is close engagement of industry throughout the pre-construction phase.
As the SKA Project moves through the design, development, construction and operational stages, industry will play a crucial role in the delivery and through-life support of the technologies and infrastructure. Industry participation at the pre-construction phase will be underpinned by strategic collaborations with commercial players, among them niche R&D companies, followed by increasing participation through pre-competitive prototyping work, commercial contracts with volume manufacturers, technology systems vendors, site services and installation firms, and power and data transmission specialists.
A survey of the engineering competencies of the Participating Organisations involved in the SKA Precursors, Pathfinders, and Design Studies shows that, for each of the Work Packages in the Pre-Construction Phase, there are a number of potential partners to work together in Consortia as Work Package Contractors. In addition, as part of PrepSKA Work Package 5, a survey of the availability and competencies of potential industrial partners in the SKA around the world is being undertaken. It is premature at this stage to identify strawman sets of partners or particular commercial companies that would form the Consortia.
Resource Requirements
The resources required to carry out this work are a total of 90.9 M€ over the 4 year period, comprising 63.0 M€ for Work Package contracts and 27.9 M€ for SPO costs. The SPO costs include staff costs for project management, system engineering, science support for system engineering, and site work (19.7 M€), and office infrastructure and operational costs for the SPO (8.2 M€). 100 k€ has been adopted for the person-year cost for salary and benefits for both WPCs and SPO. No institute overhead or explicit contingency has been included. It should be noted that the cost of developing the Advanced Instrumentation Program to PDR level in the pre-construction phase is 10 M€.
Governance
The International SKA Organisation is the legal entity that will design, build and operate the SKA. It will be set up and governed to carry out this task on behalf of the SKA sponsors, those agencies who will pay for SKA construction and operation. This organisation will be responsible for setting up the SKA Project Office (SPO) and setting policy on all matters of procurement and operation.
This Plan provides a description of the process now underway to define the optimum legal structure, and the associated optimum location for the new legal entity. The outcome of this process will be known in early 2011. In terms of developing the Execution Plan, it is assumed that a legal structure, which could be in the form of a not-for-profit company, will be established. Behind that legal structure will be a set of statutes under an implementation agreement to which a set of funding partners will accede. It is expected that the SKA Organisation will be governed by a Board/Council that is composed of government and science representatives from each of the contributing countries. The SPO is the operational arm of the SKA Organisation and its Director reports to the Chair of the Board/Council.
The management of the international project in the Pre-Construction Phase will differ in two fundamental respects from that now in operation for the Preparatory Phase. 1) The Board/Council of the new legal entity will replace the current tri-partite governance provided by the SKA Science and Engineering Committee (SSEC), the Agencies SKA Group (ASG), and the PrepSKA Board. 2) SKA design work by POs will be carried out on the basis of formal contracts under the overall authority of the SPO, instead of on the current best-efforts basis. The Board/Council will approve the allocation of major work packages to POs and Industry after a bidding and review process managed by the SPO, or as direct allocations of work following the review process.
Socio-Economic benefits of the SKA
A description of the Socio-economic Benefits of the SKA is provided, based on the outcomes of a meeting on this subject organised by COST in March 2010. The SKA can be expected to stimulate innovation in ICT and sensor technology, serve as a global model for 100% renewable energy, serve as an enabler for improved global science-industry linkages, and stimulate human capital development and employment. The construction and operation of the SKA facilities will impact local and regional skills development in science, engineering and technology and in associated industries. The nature of the SKA will excite the young and old alike about science and technology. The SKA can become part of our culture, using its popularity to inspire the next generation of scientists and engineers.
Dissemination activities
Dissemination has been carried out in many different ways. The statistics are as follows, and the details will be supplied in a separate file
Conferences 7
Meetings 215
Press release/news item 21
Media coverage 6
Videos 2
Factsheets 3
Brochure 1
Newsletters 23
Comic series 3
Leaflet/Poster 1
Poster 3
Website 2
Publications 2
For details of the dissemination activities please see Annex 1: "Dissemination Activities"
For details of the Publications, please see Annex 2: "Publications associated with the SKA Telescope"
List of Websites:
http://prepska.skatelescope.org/
See also link on SKA telescope home page https://www.skatelescope.org/