an Engineer's Perspective

What motivates me as an EMC engineer on the SKA

What role can we all play as engineers and technologists?

A personal perspective by Howard C. Reader

Introduction

• I wish to share my story on what tickles me about the Square Kilometre Array (SKA) project and why much of my academic research, and then professional engineering career, were devoted to being involved 

  • • The professional career in MESA Solutions, founded in 2002 with two research students at the time, was set in motion when academic time-scales became too long for real industrial issues

    • If my qualifications on this subject interest you, a short biography is appended.

• Part I: focus is given to my sustained attempts to keep our Karoo site electromagnetically whisper quiet. We are trying to hear the echoes of our universe! 

  • • In this context we speak about electromagnetic compatibility (EMC).

    • Some of my postgraduate projects are threaded into this short story for illustration.

• Part II: explains what heightened my radio astronomy interest as the SKA EMC contributions evolved, describing my understanding of what the SKA is trying to achieve.

• And then, more importantly, I challenge anyone involved in the project:

What can we all do to keep our sites electromagnetically quiet?

• At a slightly more philosophical level, let’s touch on how much engineering realities restrict science aspirations.

Part I: Brief SKA Intro, Engineering Challenges and EMC

General Remarks on SKA

From the early days, the SKA wished to be the most sensitive radio telescope ever. It was going to achieve this by creating an effective capture area of one square kilometre distributed over many telescopes and antennas. Some characteristics and numbers were given as:

• 50 times more sensitive than existing radio astronomy instruments with 1 million times faster surveying speeds

• It would be a physics laboratory and a time machine

• Costs of Euro 650M on phase 1 were imagined

This was stated by Phil Diamond, Director International SKA, at the South African SKA Bursars’ conference in 2013 and then again at the General Assembly of the Union of Radio Science International (URSI) in 2014.

He highlighted that self-generated radio frequency interference (RFI) “will have to be addressed vigorously at every stage of the project”.

That costing estimate at that stage of Euro 650M was an ambitious target!

• The KAT-7 array was a considerable improvement (top right) where metallic pedestals, symmetric lightning conductor systems, well-defined grounding and full galvanic cable barriers were all considerable EMC hardening measures.

• The KAT-7 system was mid-fed with struts which affected the radiation patterns. This iterated into the greatly-improved MeerKAT telescope offset feed systems (bottom left).

• The Cambridge Cavendish, an Italian group and the Australian SKA-Low effort were thinking of log periodic dipole array (LPDA) elements (mid-bottom) as an evolution of the Murchison Wideband Array (MWA) and bowtie (Cavendish) structures. The LPDAs are now being deployed in the Murchison Radio Astronomy Observatory region of Western Australia (bottom right)

  • • • Cabling, balanced and unbalanced amplifier feeds, grounding are all key factors in successful EMC management of this class of array.

Some Basic EMC Project Elements

• Due to generous NRF student research grants, the EMC research in my group at the University of Stellenbosch was able to build scale and computational models of many fundamental elements of the KAT-7, final MeerKAT and site-base developments.

• Full-scale tests were also possible to amplify growing radio-astronomy EMC awareness.

• This systematic investigation made it possible to design large infrastructural Karoo site elements with confidence.

• Work focused on:

•    Power line corona and sparking

•    Lightning and earthing policies

•    Cable interfaces and cable trays

•    Karoo Array Processor Bunker (KAPB) EMC hardening

•    Attenuation provided by the KAPB berm (should the civil engineers build up that barrier or take the soil away from the KAPB developments?)

•    Broadband testing methodologies including reverberation chambers, real-time analysers and EMC-hardened drones

• Some of the EMC team members included:

SKA-related postgraduate research

• Rodney Urban (MScEng 2001; PhD 2004) started our early diagnostics on the XDM, proposing the first lightning and earthing policies during his postdoc and early MESA (an original co-founder) days. His work on powerline corona helped us to understand that sparking on powerlines was the major EMI concern, not corona. This resulted in moderately over-specified powerline infrastructure that could not spark.

• Braam Otto (PhD 2009) continued the corona understanding and then made major contributions to site and powerline EMI evaluation and sophisticated metrology during his postdoc, continuing this work with MESA (co-managing at one point), then SARAO and finally the SKAO as international RFI manager.

• Philip Kibet Langat (PhD 2011) addressed the powerline sparking by determining the end-fire “butterfly-shaped” radiation patterns should sparking occur. Due to the importance for the SKA project, computational models, scale models replicating Karoo soil and real-life tests validated the model. This involved drone and metrology work with several other team members mentioned.

• Gideon Wiid (PhD 2010) evolved Rodney Urban’s lightning studies, developing scale and computational models that allowed unique lightning and RFI mitigation strategies to be incorporated into the KAT and MeerKAT systems. At one stage Gideon was sending me computationally modelled frequency points one at a time, using a cluster at the Centre for High Performance Computing. This was for a paper I was delivering at the Chicago URSI (Union of Radio Science) General Assembly on correlating results between our physical and scale models. This is also a tribute to Willem Esterhuyse, project lead at SARAO, who crafted the CAD for the model over a weekend. Professor Riana Geschke (USEEng) co-supervised the electromagnetic modelling work. We got in 8 points as I finalized the paper in the hotel room. Gideon continued as an academic, contributing widely to education, taking a leading role in the IEEE EMC Society, and is now at Cape Town University.

• Nolan Ebertsohn (MScEng 2005) did work on cable trays and EMC which led to specific cable design elements in the KAPB. This was also co-supervised with Riana.

• Paul van der Merwe (PhD 2011) extended Nolan’s work to focus on cabling and interfaces for KAT-7. His research began as a Master’s but was submitted for a Doctorate due to breakthrough findings on precise layouts and alternative shielding evaluation. He also assisted greatly in evaluating the first LPDA elements designed in collaboration with Professor Johannes Cloete (EEEng, University of Stellenbosch) and the Cambridge Cavendish Group. Paul managed MESA for several years and now works as the SKAO Low RFI and EMC Engineer. 

• Joely Andriambeloson (MScEng 2011, PhD 2014) looked more critically at the cable and transfer impedance elements for KAT systems. This evolved into his doctoral work on reverberation chamber time and frequency domain metrology for MeerKAT systems. His methodology continues into the reverberation chambers used on the project today. Joely then went on to become a senior member in MESA.

• Carel van der Merwe (MScEng 2012) considered a culprit and victim management of the Karoo RFI environment which led to systematic site management thinking.

• Matthew Groch (MScEng 2013) began our first work on drone evaluation of RFI which laid the foundation for the later work of Hardie Pienaar.

• Antheun Botha (Meng 2014) worked closely with the SARAO receiver team (particularly Jason Manley) and filter colleagues (Professor Petrie Meyer at USEEng) to create a valuable broadband, high dynamic range, transient analyzer. This formed the backbone of real-time analyzers in use today for monitoring site RFI. A precursor to this instrument found a bug (literally!) that had fused onto a 22 km electric fence on a farm in the Karoo. The sparking was detectable far and wide! Antheun was involved with MESA Product Solutions, then moved to the SARAO RFI team and now works as the SKAO Mid RFI and EMC Engineer.

• Nardus Matthysen (MEng 2014) built up antennas for testing in the time domain with the major contribution being the evaluation of the shielding associated with the KAPB berm properties using impulse radiating antennas.

• Hardie Pienaar (PhD 2015, also started as an MEng, but submitted for a doctorate) undertook rigorous work on the KAT site shielding: laboratory, computational and multi-copter studies, validated much of our group’s investigative studies. Hardie’s legendary drone flights, some of which were captured on video with evocative music, gave us the airborne RFI insights that confirmed KAPB and berm shielding properties.

• Post-doctoral programmes of Rodney Urban, Braam Otto, Gideon Wiid, Paul van der Merwe and Necmi Tezel refined many of the complex issues leading to detailed design elements of the KAPB, berm design and telescope interfaces which exist today.

A few illustrations hinting at the EMC-related research

A short aside linked to the photographs above:

• Firstly on Dr Dirk Baker, who was a colleague and friend of Professor Johannes Cloete when they worked at the NIAST division of the CSIR. Dr Baker is one of a few people acknowledged by Professor J D Kraus in his famous book on Antennas. Dr Baker arranged for the project to use the world-class facility at no cost. He is also thought of as the Guru of EW antennas in South Africa.

• Professor Cloete established much of the electromagnetic metrology facilities at Stellenbosch under an early National Research Foundation (NRF) grant. Many former and existing staff members in electromagnetics at USEEEng were inspired and taught by Johannes Cloete, who is acknowledged with deep affection.

• One of these is Professor David Davidson, a close colleague and friend of mine, who took on the first SARChI EM Chair at Stellenbosch and now works on the SKA at Curtin University, Western Australia. We collaborated closely. 

• Another colleague, the late Professor Keith Palmer, came through the same stable. He made a wonderful offer to design an optimized LPDA antenna to facilitate the RFI observations in our group’s work. This has been known since as the KLPDA and several are in widespread use, particularly in the Karoo.

• And, for illuminating the world scale of EMC which influenced all the work described: 

  • Professor P C T van der Laan and Dr A P D van Deursen (EUT, Netherlands)

  • Dr Dave Giri (Pro-Tech, USA)

• As implied in the research projects, diverting RFI currents away from sensitive regions, by paying attention to cable connections to galvanic interfaces, is important.

• This is where personnel involved in installation, testing and maintenance start to play a key role.

• The principle is illustrated firstly in schematic form – just having a thin metallic L-plate before an instrument, such as an oscilloscope, can take external RFI currents away from the sensitive internal electronics. the topic is explained further in the next 3 pictures.

• To amplify the current and field diversion illustration, deliberate coupling between two nearby loops is created in the following simple experiment:

• By placing a carefully-sized, grounded plate, the interference to the victim loop on the right is significantly reduced (the blue colour represents electromagnetic quietness). The take-away here is that the plate does not have to be infinite in size if one can visualise what paths RFI current would take.

• These elementary EMC principles were applied to the KAPB and other structures. This is deliberately not a comprehensive picture on all EMC principles, but it attempts to point to fundamental matters.

• On-site personnel and any party responsible for developments in and around telescopes should be alert to these simple EMC mitigation topics. If they are designed from the start and are maintained with attention to good galvanic contact of all cables through interfaces, then we have the chance to sustain our electromagnetically-quiet environment.

Evolution of the KAPB 

Concluding Remarks

• This contribution is a personal view on what has fascinated me about the SKA project and radio astronomy in general.

• With that backdrop, I have given clues on how my research group and business have influenced EMC elements on some SKA infrastructure.

• Increasingly-sensitive radio astronomy “instruments” must pay attention to EMC and protection.

• Dialogue between engineers, scientists and building teams throughout is not a “nice to have”; it is essential.

• On-going education and training will always be needed; this leads to robustness and intended science.

• On a small-scale, robust EMC policies are achievable. On an international-scale, varied parties, contractual agreements and costing will yield compromise. This can be managed by common purpose.

• Anyone connecting up systems should appreciate their role in the small detail of cables, interfaces, grounding and maintenance. Each person’s activities can contribute to keeping our sites electromagnetically quiet and improving the radio astronomers’ chances of understanding our universe with increasing clarity.

• Unless these EMC related perspectives are kept at the fore, the ambitious goals of the SKA radio astronomy will not all be achieved. This resonates with Phil Diamond’s early remarks about addressing RFI vigorously at every stage.

Part II: My Evolving Understanding of the SKA Science

When I first became involved in the SKA project, starting with the XDM at HartRAO, I had very little appreciation for radio astronomy. As I met with a growing number of radio astronomers during my work, my fascination with the science objectives grew. With it came an alarming awareness of how poor system design, leading to RFI, could disrupt the astronomy intentions.

My own understanding of what the scientists were aiming to achieve illuminated my own purpose in the work described in Part I. This particularly affected work in the lower frequency ranges (related to the Epoch of Reionisation and the Cosmic Dawn) and time domain transients which dramatically influence pulsar research.

My first shock came when I met with Mike Gaylard, the then director of HartRAO. In my role as both an academic and a MESA Solutions director we measured a consistent 1420 MHz signal all over the XDM. This made little sense to me as nothing I was aware of transmitted an unmodulated tone at that frequency.

After a lunchtime discussion with Mike, talking about this anomaly, his face clouded over and he said “oh, no, our MASER”. The hydrogen MASER which had been flown out from Switzerland to keep “perfect” time, used a hydrogen transition with a frequency at exactly 1420 MHz. We tracked this down and identified poor grounding on a wooden floor in the control room area, no interfaces to any cabling, and then a long plastic conduit from there to the XDM which acted as a perfect coaxial cable. This started our apparently simple EMC experiment investigations.

Why was this important? “The H1 line”, explained Mike. Oh yes, I had never heard of that.

Here is my understanding of the SKA science.

Brief History of SKA:

• In 1990, Peter Wilkinson gave a conference talk in New Mexico on The Hydrogen Array. He began with: “The time is ripe for planning an array with a collecting area of one square kilometre”

• The Suggested Goals then were Neutral Hydrogen Observations at red shifts z = 0 to z = 10. Neutral hydrogen? More on this below.

• Red shift is the equivalent of the Doppler effect we hear when train whistles fade into the distance, or in this case, our observed expanding galaxy.

• Much happened after 1990, but in 2003, Australia and South Africa worked on Path Finders for the SKA hosting bid.

• In 2005 South Africa (RSA) submitted its proposal to host the SKA in RSA and twelve sites across Namibia, Botswana, Mozambique, Mauritius, Madagascar, Kenya and Ghana.

• In 2012 the SKA Organisation announced that RSA (Dishes, MidArray) and Australia (LFAA) were selected as joint hosts. Statements made shortly thereafter:

“SKA will be one of top global science projects of 21st century”  

Richard Schilizzi, first International SKA director

“After ISS and Large Hadron Collider, world's next great science project is the SKA”   

UK Science Minister, David Willets

SKA Key Science Drivers

• The following three objectives of the first stated SKA goals are given without further explanation, but they point to the dominant science interests.

1. Origins

• Neutral hydrogen in universe from Epoch of Re-ionisation (EoR) to now 

• When did the first stars and galaxies form?

• How did galaxies evolve?

• Role of Active Galactic Nuclei (AGN)

• Dark Energy, Dark Matter

• Cradle of Life – looking for life building blocks; are we alone (SETI)?

2. Fundamental Forces

• Pulsar test of General Relativity & gravitational waves. Wish to find all pulsars in the Milky Way.

• Origin & evolution of cosmic magnetism (SKA only!)

3. New Phenomena

• Such as “transients” which have already been found and pique our exploration.

More Detail on Aspects of Radio Astronomy Thinking

•  For my own understanding, I needed to dig a little deeper into some of the terminology that I encountered, and this subsequently allowed me to follow radio astronomy presentations with both interest and focus.

• This had some effect on the spotlight for our EMC research. An example would be on pulsars which are the “clocks” of our universe. Transient interference on site dramatically reduces pulsar science and this is why the transient analysis by some of my group was important

• An interesting aside on pulsars is that I spent a sabbatical with the Cavendish Group in Cambridge in 2010. Apart from detailed work on the bow-tie antennas and first LPDA antennas, I surveyed several EMC-related elements of the Lord’s Bridge site where Dame Jocelyn Bell and Professor Antony Hewish first discovered pulsars in 1967. I also took the time to read the extensive letters of Professor Fred Hoyle at the John’s College Library. They reflected much on the astronomy during that period and also commented on the award of the Nobel Prize given for the pulsar discovery. The award did not include Dame Bell who was Hewish’s doctoral research student.

• Why H1 or Neutral Hydrogen Line?

With thanks: Tom Osterloo, ASTRON

• My earlier anecdote on the HI line at HartRAO refers. Why is it so important?

• Photon emissions due to spin flip of lowest state of atomic hydrogen occur at

  • • •   1420.405751786 MHz = 21.106117011 cm

        • • Radio astronomers speak a lot about the 21 cm line

• There is only one flip in 107 years, but there are 1066 neutral hydrogen (NH) atoms in any given spiral galaxy!

• With this very precise line, an understanding the Doppler effect of an expanding universe, including emitter and thus gas velocity, can be evaluated. Some of the predictions for the EoR, presented to me during my 2020 Cambridge sabbatical, implied that the H1 line from the EoR would appear around 70 MHz due to the time and Doppler shift involved. This is why we looked at antennas which went down to 50 MHz – really to explore the early universe origins! RFI in this frequency range would severely impact on those discoveries which are yet to be made, and would reinforce current theories.

• The H1 line allows studies of galaxy evolution: emissions in/about galaxies z = 0 to beyond z = 1 ( > 8 billion years (Gyear) ago)

• Work in 2010 was only around z ~ 0 (“with a few heroic exceptions!”). The SKA was needed to look further. According to Tom Osterloo there was an initial “burst” in star formation, but this then declined tenfold in last 10 Gyr; the available gas did not reduce equivalently - WHY?

  • • • •  These are questions to be answered by the SKA.

Understanding Galaxy Evolution

Advancing Astrophysics with the SKA, Italy, June 2014.   SKA International Website 2014

• Galaxy formation/evolution up to 2014 had been pictured principally by optical/infrared telescope large surveys, measuring stellar light and radiation from dust and molecules and by simulations.

• H1, which is only ‘visible’ at radio wavelengths in/around galaxies, was essential to complete the picture. Probing H1 to 10 billion years ago allows current theoretical predictions to be challenged.

• An illustration of galaxies using visible light (left) and H1 (right) gives clues to the developing understanding of galaxies.

• The first 10% phase of SKA was reputed to let us see up to 5 billion years back in time. The full SKA is apparently needed to explore the EoR and cosmic dawn crucial timeline around 10 billion years ago.

1. Galaxy Evolution and Dark Matter/Energy

• The image below is a typical depiction of our universe’s history with contemporary understanding.

• This is an appropriate point to relate red shift and equivalent historical time of cosmic proportions. The numbers are staggering: 

• The SKA-Low arrays being built in Australia need to have antennas that can operate down to these frequencies.

• Identifying sources of ionization in early universe improves current constraints on parameters obtained from CMB, and informs our understanding of how first, most massive stars might have cooled enough to undergo gravitational collapse.

Pulsars: Galactic Clocks

With thanks: Aris Karastergiou, Oxford

• To the final SKA science topic of interest to me, pulsars.

• After Jocelyn Bell’s discovery of precisely repetitive signals detected on 90 ft or so of daily recording chart paper, Hewish recognised that a star more massive than the sun runs out of fuel and collapses. Rather like an ice-skater drawing in their arms, the collapsed neutron star rotates increasingly faster, reaching millisecond periods. We detect the intense magnetic fields from pulsars as illustrated below.

• It is important for me to note that Bell painstakingly examined about 90 foot of recording chart paper each day at a hut at Lord’s Bridge in Cambridge. Others had seen these signals but dismissed them. Dame Bell did not. The signals were fleetingly thought to be extra-terrestrial attempts at communications. But who would communicate using a single tone?! The science community now knows better.

• Pulsars are not detected by just pointing telescopes into the sky. The process is quite involved.

  • • • When looking at the correct place, signals are weak, masked by background waves and dispersed

      • Received signals are converted to total power, followed by an acceleration search, fold, and examination for pulsar signals. The image below gives clues on the procedure

      • Processing costs are proportional to the number of tied array beams per primary beam.

• RFI, especially pulsed – power lines, sparking, welding, etc., all disrupt this science dramatically

Concluding Remarks

• Part II on my understanding of SKA science, inspired me at least to pay attention to keeping our South African SKA sites electromagnetically quiet.

• Pulsar and the H1 (neutral hydrogen) observations then made a whole lot more sense to me.

• I love the image that any radio, cellphone, aircraft distance measuring equipment (DME), satellite signals, or worse, simply tears pages out of our universal diary. Any frequency band we trash with RFI stops us looking back to that red-shifted part of our evolving universe.

• So the next time you use your car remote, Bluetooth earphones to go jogging, or neglect to clear up any sparking electric fences, think of our collective universal diary.

• I do.

Specific Acknowledgement to:

Dr Ursula A Reader for many philosophical discussions throughout this journey.