Low-Frequency Aperture Array


During a 10-day installation campaign from 6th to 17th November, an international team from Australia, Italy, The Netherlands and the United Kingdom installed all remaining antennas of the AAVS1 main station, along with debugging some of the already installed antennas at the Murchison Radio-astronomy Observatory in outback Western Australia, the hosting site of the Murchison Widefield Array, MWA.

The on-site team progressed very well during the trip, being split into two sub-teams, to move forward on both deployment and on debugging. Besides the field work, updates were also implemented on the AAVS1 TPMs and processing system, located inside the correlator facility.

Figure 1. The AAVS1 station

Figure 2. The Tile Processor mapping on the antenna locations

Figure 3. The cabling inside the Antenna Power Interface Unit: the fibre cables (blue) and the DC-power cabling (red/black)

The last few months have been a busy time for the digital signal processing (DSP) work-package inside LFAA. The installation of the 16 tile-processing modules in the MRO led by the AAVS-1 team proved to be more complex than expected with the usual teething problems that a new system brings along with it. Having manually sawn off parts of the chassis in order to fit them in the rack, 4 iTPM(s) per subrack were installed and individually tested. The final configuration is shown in Figure 4 below.

Figure 4. TPMs installed at the MRO site

The iTPM installed is the Sanitas-designed iTPM version 1.2, running firmware and control software that has been designed as joint effort between INAF and the Universities of Oxford and Malta. Through the Monitoring and Control Software developed, the signal integrity of each path can be recorded and used to test the stability of the chain over time and temperature. Figure 5 shows a typical output from 8 signal chains in the field.

Figure 5. Bandpass of 8 antennas in the field

During the cost control activities of late 2016 and first half of 2017 an action was placed on the LOW antenna. After an optimisation process and a selection process for the best possible performance and cost, an evolution of the SKALA antenna (arXiv:1512.01453v1), SKALA4 (Figure 6),wasconsidered the best option to take forward to Critical Design Review (CDR) by an ad-hoc panel of experts tasked to undertake the performance evaluation of various antenna designs. The Antenna and LNA work package of LFAA (led by Cambridge, UK) worked on an evolution of the Log-Periodic Dipole Array antenna designed for SKA1-low. After the latest update in 2016 for improvement of the bandpass smoothness (SKALA3,arXiv:1702.05126v2), SKALA4 has been optimised to further improve on that smoothness at all scales (narrow as well as wide frequency bands), cross polarisation, sensitivity, etc. This antenna looks like a perfect match to the demanding SKA1-low scientific and technical requirements. SKALA4 has also been designed to preserve the current LFAA architecture and to be compatible with existing interfaces. The SKALA4 LNA is an evolution of the current SKALA3 LNA with minor tuning on the input matching network.

Figure 6. SKALA4

SKALA4 was considered in this selection process together with SKALA3, the MWA dipole and a Vivaldi antenna. The different candidates were measured against a set of requirements presented to the design team by the SKAO. These requirements included both functional and nonfunctional figures of merit in order to measure cost, manufacturability, maintainability, projects risks, technical risks, etc. as well as performance parameters. Following the analysis and advice from a panel of experts, the SKAO endorsed the panel recommendation of selecting SKALA4 as the best possible candidate for SKA1-low and provided a set of recommendations for its final tuning. The antenna is now being further developed mechanically (Figure 7) by the LFAA consortium to be ready for CDR and the SKA1-low production. This process has counted on the participation from several consortium partners (Cambridge, ASTRON, INAF, ICRAR, STFC, Oxford) and the SKAO.

Figure 7. SKALA4 initial mechanical design

The improved sensitivity of single station of the new SKALA4 antenna, compared with SKALA3 and the L1 requirement, is given in Figure 8.

Figure 8. Single station zenith sensitivity comparison

September also saw the visit of KLAASA to Lord’s Bridge to test their own developed TPM. Tests using a drone showed that their system is capable of channelizing the bandpass into 512 different channels and beamforming each channel separately as is required by the LFAA specifications. This progress prompted the need to hold an Aperture Array Convention in Hefei, China, where future directions of signal processing architectures and contributions from KLAASA were discussed.

Group image taken of the CETC38 summit.

With guidance of Professor WU Manqing, KLAASA (Key Lab of Aperture Array and Space Application) completed CTPM (Chinese Tile Processing Module)—Pre-AAVS1 integration test at Lord’s bridge during the 21st September to 3rd October. In cooperation with Cambridge University, Dr. CAO Rui led a team of 10 successfully verifying the multi-channel digital beamforming (DBF) function of CTPM. The test culminated in a 2-day meeting reviewing results and discussing the next development steps.

Figure 9. Group picture of the review meeting participants at SKALA-2 test facility

Figure 10. Schematic diagram of KLAASA-Cambridge Pre-AAVS1 testing system

CTPM features three-stacked Harmonica Structure and integrates 32 channels on the 6U board, capable of 16 dual-pol antenna signals. The polarised signals are transferred through 2 MPO connectors from tile to the receiving sub-system where CTPM performs O/E conversion, signal filtering and amplification. Further processing as AD conversion, channeliser, digital beamforming (DBF) and monitor and control are realized in the signal processing subsystem.

Figure 11. CTPM prototypes. Two same prototypes- CTPM A & CTPM B were used for the testing at Cambridge

Figure 12. CTPM prototype — signal-processing board

KLAASA developed DBF algorithm on CTPM firmware by compiling more than 100,0000 lines of codes, which animates the module with polyphase filtering, channel selection, calibration and tile beamforming and station beamforming.

The testing campaign included bench testing carried out at Cavendish lab and far-field testing with SKALA-2 facility. CTPM captured data from SKALA-2 array and produced antenna patterns while the UAV hovering at array zenith at different altitude and flying cross the array. Figure 6 and 7 depict respectively the pattern of 80 m and 35 m UAV altitude. The UAV shook with wind at higher altitude and thus led to differences between test results and simulation.

Figure 13. Far-field testing carried out at Lord’s Bridge, Cambridge

CTPM captured data from 8 SKALA-2 antennas and produced antenna patterns while the UAV hovered at array zenith and flew cross the array.

Gaps in results with angle deviations are due to the inaccurate UAV positioning as single GPS signal instead of D-RTK was adopted as the positioning input. Figure 8 and 9 shows the test and calculated results of DoA (Direction of Arrival)

Figure 14 Simulated and tested DoA results with 17.4°angle deviation from array zenith (left) and with -45.5° angle (right) (Calculated value is -45.5° while test result is -46.7°).

The CTPM development project is part of Key Technology Development and Verification of Digital Array in SKA Low Frequency Radio Telescope Project funded by Ministry of Science and Technology (MoST) of China. CTPM is expecting its upgrading design of higher processing capability and even lower power consumption next year.

Report provided by the LFAA consortium