SAT.STFR.FRQ (UWA) DETAIL DESIGN REPORT (MID)

1 EXECUTIVE SUMMARY
The Square Kilometre Array (SKA) project [RD1] is an international effort to build the world’s most sensitive radio telescope operating in the 50 MHz to 14 GHz frequency range. Construction of the SKA is divided into phases, with the first phase (SKA1) accounting for the first 10% of the telescope's receiving capacity. During SKA1, a Low-Frequency Aperture Array (LFAA) comprising over a hundred thousand individual dipole antenna elements will be constructed in Western Australia (SKA1-LOW), while an array of 197 parabolic-receptor antennas, incorporating the 64 receptors of MeerKAT, will be constructed in South Africa (SKA1-MID).
Radio telescope arrays, such as the SKA, require phase-coherent reference signals to be transmitted to each antenna site in the array. In the case of the SKA, these reference signals are generated at a central site and transmitted to the antenna sites via fibre-optic cables up to 175 km in length [RD2]. Environmental perturbations affect the optical path length of the fibre and act to degrade the phase stability of the reference signals received at the antennas, which has the ultimate effect of reducing the fidelity and dynamic range of the data [RD3]. Given the combination of long fibre distances and relatively high frequencies of the transmitted reference signals, the SKA needs to employ actively-stabilised frequency transfer technologies to suppress the fibre-optic link noise [RD4] in order to maintain phase-coherence across the array.
Since 2011, researchers at the University of Western Australia (UWA) have led the development of an ‘SKA phase synchronisation system’ designed specifically to meet the scientific needs and technical challenges of the SKA telescope. This system [RD5] is based on the transmission of actively stabilised phase-coherent reference signals generated at the Central Processing Facility (CPF), and then transmitted via separate optical fibre links to each antenna site. The frequency transfer technique at the core of the SKA phase synchronisation system is an evolution of the Atacama Large Millimeter Array (ALMA) distributed ‘photonic Local Oscillator (LO) system’ [RD6], incorporating key advances made by the international frequency metrology community over the last decade [RD7], [RD8] and [RD9], as well as novel innovations developed by UWA researchers [RD10] and [RD11].
Two variants of the SKA phase synchronisation system have been designed. Each one has been optimised specifically for its respective telescope:
• For SKA1-MID, the required microwave (MW) shift is generated using a Dual-parallel Mach-Zehnder Modulator (DPM), biased to generate Single-sideband Suppressed-carrier (SSB-SC) modulation [RD10].
• For SKA1-LOW, the Radio Frequency (RF) shift is generated using a simpler Acousto-optic Modulator (AOM) [RD11].
This results in two systems that easily meet the SKA functional performance requirements, as demonstrated by laboratory testing [RD10], [RD11], [RD12] and [RD13], overhead fibre field trials [RD14], [RD15] and [RD16], and astronomical verification [RD17], [RD18] and [RD19], yet maximise robustness and maintainability while keep complexity and costs to a minimum. Following an extensive technical down-select process, in October 2017 a MW-frequency variant from the UWA was selected by the Square Kilometre Array Office (SKAO) to be the phase synchronisation system for SKA1-MID.
The key innovation of the SKA phase synchronisation system was finding a way to use AOMs as servo-loop actuators for RF and MW-frequency transfer [RD10] and [RD11]. The large servo bandwidth and infinite feedback range of these servo-loop AOMs ensures that the stabilisation system servo-loops never require integrator resets. The SKA phase synchronisation system also utilises AOMs to generate static frequency shifts at the antenna sites to mitigate against unwanted reflections that are inevitably present on real-world links. Reflection mitigation is absolutely essential for the SKA phase synchronisation system, as there is no way to guarantee that all links will remain completely free of reflections over the lifetime of the project.
The SKA phase synchronisation system has the servo-loop electronics and the vast majority of all other optical and electronic components located at the CPF, greatly simplifying maintenance. A single high-quality Frequency Synthesiser (FS), tied to the SKA master clock, is used to generate phase coherent reference signals, and these are distributed to the Transmitter Modules (TMs) which are then used to transmit the optical signals across each fibre link. The TMs incorporate the servo-loop AOMs, and these are able to add an independent and unique RF-scale frequency offset – in the optical domain – to the common transmission frequency for each link. This avoids any possibility of common frequencies at each antenna site to ensure any stray RF emissions will not be coherent if picked up by the receivers.
The Receiver Modules (RMs) for the SKA phase synchronisation system have a very small form-factor and contain only a minimum number of simple optical and analogue electronic components, making them extremely robust to external environmental perturbation. In addition, they are designed to be capable of being mounted directly on the SKA1-MID antenna indexer alongside the receiver. Currently, the SADT interface with DISH is in the antenna pedestal, and the DISH Consortium are required to build a second frequency transfer system to transmit the reference signals up the cable wraps to the indexer. After a successful down-select, the DISH consortium and SKA Office (SKAO) have agreed to an Engineering Change Proposal (ECP) to correct this inefficiency.
A small form-factor, industry standard, Oven-controlled Crystal Oscillator (OCXO) is incorporated into the RM to provide phase coherence at timescales shorter than the light round-trip time of the fibre link. The OCXO is tied to the incoming reference signals using a simple, encapsulated Phase Locked Loop (PLL) based on the proven design implemented by the Australian SKA Pathfinder (ASKAP). This is particularly important, as it has been shown that using multiple MW-FSs can easily lead to a significant loss of coherence, even if the transmission frequency is being successfully stabilised [RD18], [RD19] and [RD20].
The SKA phase synchronisation system is designed in such a way as to also stabilise the non-common optical fibre paths in the TMs. This effectively stabilises the TMs at the same time as the fibre link, making the equipment in the CPF extremely robust to external environmental changes. The optical phase sensing allows for the use of Faraday Mirrors (FMs) to give maximum detected signal at the servo photodetector without requiring any initial polarisation alignment, or any ongoing polarisation control or polarisation scrambling.
The SKA phase synchronisation system has been extensively tested:
• Using standard metrology techniques in a laboratory setting [RD12] and [RD13], with signals transmitted over metropolitan fibre links and fibre spools under all required conditions;
• On 186 km of overhead fibre at the South African SKA site [RD14], [RD15] and [RD16];
• Using astronomical verification with the Australian Telescope Compact Array (ATCA) for SKA1-MID [RD18], [RD19] and [RD20], and the ASKAP for SKA1-LOW [RD19].
This has demonstrated that the SKA phase synchronisation system is fully compliant with all SKA requirements, as well as demonstrating functionality of critical practical factors that are not captured by these requirements.
Furthermore, UWA researchers in partnership with MeerKAT and University of Manchester (UoM) engineers, have developed the detailed designs into a set of mass manufacture archetypes, effectively getting a head-start at addressing manufacturing issues that may be encountered by contractors during the construction phase. The first set of mass manufacture archetypes for SKA1-LOW were completed in Q2, 2016 [RD21]; and for SKA1-MID in Q1, 2017 [RD22]. All aspects of the mass manufacture design are openly available and are provided with sufficient detail so that any firm with expertise in optical and electronic assembly can to reproduce these systems with minimal domain expert input. An optical technology consultancy firm was employed to provide an independent review of the labour costs associated with assembly and testing (see Appendix 8.9.3).
All sub-elements of the SKA phase synchronisation system have been designed to be hot-swappable, enabling simple installation and easy maintainability (especially as the vast majority are located at the CPF. The system is designed so that during commissioning, only one free parameter needs to be optimised per link.
Prior to the technology down-select process, the detailed design presented in a previous version of this document had been critically assessed by the following independent domain experts:
• Gijs Schoonderbeek from the ASTRON Netherlands Institute for Radio Astronomy;
• Larry D’Addario from the Jet Propulsion Laboratory;
• Johan Burger from SKA South Africa.
The input from this review was used to update this document. As part of the SADT Consortium-led technology down-select process, this document was reviewed further by the following independent domain experts:
• William Shillue from the National Radio Astronomy Observatory;
• Miho Fujieda from the National Institute of Information and Communications Technology;
• Sven-Christian Ebenhag from the Swedish Research Institute (RISE).
Again, this document was updated taking into account the feedback provided.
These reviews have built confidence in the detailed design and ensured that the SKA phase synchronisation system is the best possible phase synchronisation solution for the SKA telescope.