Skip to main content

LOFAR Development Newsletter September 2022

For many people, summer is a time to relax, and I hope you managed to have a break as well. We had very dry and hot weather during the holiday season in the Netherlands. While most of us were desperately trying to keep the heads cool, some of us were bravely verifying the LOFAR2.0 design with the Dwingeloo Test Station. As you will read in this Newsletter, the hot weather was an excellent opportunity to test the thermal design of the LOFAR2.0 cabinet to its extreme!

You will see in this Newsletter that here is a lot of development going on for LOFAR, not just for LOFAR2.0: The HBA electronics is being redesigned in the DANTE project, TMSS is delivering a completely new specification and scheduling system for LOFAR, and we are making plans for the future of the Central Processor (CEP), including the replacement of CEP4. Sometimes, we can be so busy that we no longer recognize the uniqueness of our results. So let me emphasize once again that we should be extremely proud of what we are achieving! We are pushing technology boundaries, and the solutions we come up with, also help other instruments move forward.

Happy reading!

Wim van Cappellen

LOFAR Family Meeting 2022 in Cologne

Jason Hessels

After 2 years of Covid lockdowns and travel restrictions, the German Long Wavelength Consortium (GLOW) hosted a face-to-face LOFAR Family Meeting in Cologne from June 13th to 17th, 2022. The LOFAR Family Meeting is an annual tradition, and this year attracted about 150 participants from across Europe and the rest of the globe. The meeting included over a dozen sessions spanning the whole range of LOFAR science, from the Sun to cosmic rays to pulsars, galaxy clusters and the early Universe, etc. The meeting also featured a LOFAR users session with talks from the LOFAR Users Committee, and the Science Data Centre Operations and Programme teams. In a dedicated LOFAR2.0 session, Wim van Cappellen gave an update about the status of the LOFAR2.0 development efforts, and Barbara Matyjasiak presented the status of LOFAR4SW. Maaijke Mevius shared her thoughts on how to deal with the ionosphere in the LOFAR2.0 era. Jason Hessels updated the community on the LOFAR2.0 Large Programmes, and it was great to see the long line of posters describing the science that the various Expressions of Interest aim to achieve with LOFAR2.0. The most striking memory of the meeting, however, was seeing so many new, young faces! The LOFAR community has grown a lot in the past 2 years with many new PhDs and postdocs, and it was energising to see them together talking about the great science they are doing. It was fitting that the meeting ended with a final session called “Visions for LOFAR”, where we got to hear some very inpiring ideas for how to extend LOFAR in the future.

Conference website:
https://www.glowconsortium.de/index.php/en/lofar-family-meeting-2022

What to develop: a drop-in replacement or an improved HBA Front-End?

Boudewijn Hut    André Gunst

During the maintenance of LOFAR stations, malfunctioning High Band Antenna Front End (HBA FEs) are replaced. So far, we are able to produce new ones, but critical components become obsolete. We cannot hold this model much longer, therefore the operations team requested a new HBA FE design that simply can be used as a drop-in replacement for the current ones.

At the same time, the ILT partners show interest in the LOFAR4SW dual-beam architecture to enable more science. To make that happen, a number of products in the HBA Tile need to be adjusted, including the HBA FE.

When comparing the requirements from both requests, it turned out that they were conflicting. That is a problem as it becomes very difficult (or expensive) to make a new front end that actually works for both uses. Left untouched, this unclarity may lead to an end product that does not suit any need.

We discussed with the two groups on how to resolve the issue: we cannot start development without having one comprehensive set of requirements, which we all agree on. The project start was delayed accommodating relatively small changes for both sides. The resulting set of requirements could be agreed, allowing the project to start and follow the depicted process.

As a result of progressive insight, needs may change in the future. If that happens, we will need to analyse the overall impact of the change on interfaces, functionality, performance, time and budget, and also communicate about it with the interested groups. Therefore, we follow this process: we make sure expectations stay aligned during the development project. Thanks to that the final product will be useful for both groups: having a replacement for the old unit and getting one step closer to the dual-beam implementation.

Phases DANTE
Figure 1: Phases for the DANTE project, including milestones and project reviews.

Station

Arno Schoenmakers

Over the past summer we have been using the DTS test station at ASTRON intensively. We have made use of some extremely warm days to investigate the thermal and cooling properties, we have done tests to determine the noise contributions originating in our processing subracks, in the cabinet, leakage through the cabinet walls, we looked at cross-talk between input channels on a single RCU2, and between different RCU2s, we managed to calibrate the cable delays from the test field antennas and correct for that and, ultimately, we have been able to show that digital beamforming is working, which is a highly important milestone for the project. See the figures for some DTS test result examples.

The next step we planned is to move the hardware that we have used in DTS to a real LOFAR1 station. The test field we now use is very close to the ASTRON buildings and the Dwingeloo telescope and does not provide the RFI-quiet environment we need to usefully run our final tests for sensitivity and linearity. We will use the CS001 core station for this. The CS001 station is positioned conveniently close to the road and the concentrator-node building near the ‘superterp’ at Exloo. It is a core station, therefore with two HBA fields, this allows us to experiment with ‘long baselines’ in a single station.

Over the summer, we have already used CS001 in its current LOFAR1 configuration to create a collection of reference LOFAR1 datasets, to which we can compare the upcoming LOFAR2 datasets. We plan to migrate DTS hardware to CS001 in early October. First, we need to remove most of the existing LOFAR1 electronics, including the signal processing part, control system, power supplies, local clock and network. Then we can make the necessary changes to the cabinet and then finally add the DTS setup to the cabinet. We expect that all of this work can be done in a single week, so we should have a first limited LOFAR2 station ready for testing by half October, at the latest. From then on, the station will be called “LOFAR2 Test Station”, or L2TS. We will run a similar test program as we did on DTS and continue debugging and developing firmware and software components in the weeks thereafter.

Next to these activities, we have ordered additional hardware to make L2TS a fully equipped LOFAR2 station, the first of its kind. Unfortunately, we have little certainty when the manufacturing of these additional components is ready. We hope to have the first LOFAR2 station ready by the end of quarter one of next year, 2023.

Averaged crosscorrelation
Figure 1: The averaged (over 18 input channels) crosscorrelation (orange) and the peak value per subband crosscorrelation together with the autocorrelations. For LOFAR1 the averaged crosscorrelations are at a level of approx. -100 dB, showing the large improvement of approx 30 dB of the LOFAR2 RCU2s.

 

Sky observation digital beamformer
Figure 2: Results of a long sky observation with DTS using the digital beamformer. Shown are the powers in a single subband. The drawn orange lines are the expected trajectories of Cyg A and Cas A over the test field. The plot shows that we have detected these sources and have been able to track them over the test field. The other curves are consistent with the expected grating response from the regular test array consisting of 9 equidistant LBA antennas setup in a E-W line.

 

Figure 3: Measured phases before (left) and after applying accurate cable length calibration to the signal path. The plot shows we are now able to reliably determine delays and correct for this, so that we can apply beamforming accurately.

 

Timing Distributor

Carla Baldovin

We are finally in the process of procurement for the Timing Distributor hardware, which is expected to be finalized by mid-October. Only then will we be able to plan definitive dates for the rollout. In the meantime, the activities of TD will decrease considerably.

Procurement

Nico Ebbendorf

With the successful completion of DTS and various tests and finetuning on its way, we are ready to prepare the next step, the LOFAR2.0 test Station L2TS. Small changes are adapted to the design of the electronic boards ready for procurement. Everyone knows about long delivery times for goods at this moment and specifically for products that include electronic components. This situation also effects the planning for L2TS and makes the planning uncertain. Fortunately, we have ordered some extra components during the DTS production which we can use in the L2TS production now. This is not the case for the RCU boards, and we will need almost 100 RCU’s to complete the L2TS. At this moment we have followed the procurement procedure for the new RCU boards, and the order is out. Now we need to be patient and wait until all components are available for production. We are expecting the delivery of the L2TS modules in spring next year. Currently, the station CS001 has been selected as the L2TS and station tests will start in the summer 2023.

Meanwhile, the European procurement process for the LOFAR2.0 station modules for the rollout (of all stations) is also on its way. Manufactures can send in offers until 23rd of September at the latest and the definitive orders will be place early November. Next year will primarily spend on the lead time for components of all the LOFA2.0 modules and looking for alternative sources or equivalent components before the production can start. Rollout of LOFAR2.0 is now planned to start in the summer of 2024.

Widefield international baseline imaging with facets

Reinout van Weeren

LOFAR’s long international baselines offer the opportunity of subarcsecond resolution imaging. Sweijen et al. (2022) showed that the entire field of view of a LOFAR observation can be successfully calibrated and imaged. To do so, the field of view is processed in 25 smaller chunks, each covering an area the size of a full moon. Each of these chunks is then turned into an image which can be stitched together afterwards. The final image of seven billion pixels contains almost as many pixels as radio surveys from the past did covering the entire sky.

However, the imaging process does require a significant amount of CPU hours, making it challenging to scale it up to many fields. Over the past months, work focused on identifying bottlenecks and explore new methods to speed up the imaging. It was found that the new facet-mode of WSClean provides a promising way forward to image the full field of view at once, given its low memory footprint and employing the fast wgridder.

Several experiments showed that images as large as 60,000×60,000 pixels can be made, including calibration corrections in a few dozen directions. This approach offers the additional advantage that the entire field can be deconvolved simultaneously. A planned next step is to add the primary beam corrections, which requires some software changes to reduce the memory footprint. In addition, work has started to add baseline dependent averaging in time and frequency to further speed up the imaging. This new approach should reduce the amount of cpu hrs required by a factor of at least three, an important step towards the ultimate goal of imaging the entire northern sky at subarcsecond resolution.

Figure 1: Small cutout of a radio galaxy in the ELAIS-N1 field at 1.5 (left) and 0.4 arcsec (right) resolution. Both images were made use with WSClean’s new facet-mode.

DANTE

Carla Baldovin

The team has been working intensively during the last couple of months. The pictures show the first two DANTE test printed circuit boards (PCBs), which test the two most critical parts of the design. The low noise amplifier (LNA) needed to be redesigned as the transistors of the LOFAR1 frontend are obsolete. The new LNA promises much better linearity without increasing the noise figure or power consumption. The new true time delays (TTD) make use of discrete components instead of PCB tracks, which reduces the size of the TTDs and reduces manufacturing tolerances. Both PCBs were first simulated using advanced design system (ADS), making it possible to fine-tune the component values by simulation before building the first PCBs.

The next step is to measure the different parameters of the boards (e.g., gain, noise, linearity, power consumption) in the laboratory and to cross check the measurements with the simulations. Another test to be done is to cover the PCBs with the potting material they will have in their final version. Potting is done to protect the PCBs against shocks, vibrations, and to stop moisture or corrosive agents from getting near the electronics. The tests will investigate the behavior of the RF signal in the presence of this material.

At the same time, the team is busy preparing for the preliminary design review, planned to take place in the first half of October. The review will focus on what is achieved so far: requirements, compliancy, overall planning and budget. The outcome shall determine if we are ready to move forward with the next phase in the design: the engineering model, in which we will assemble a full HBA tile in the field using the newly designed HBA front end boards.

Figure 1: Upper panel: DANTE PCB boards of the LNA (left) and TTD (right). Lower panel: LNA (left) and TTD (right) simulated boards

LOFAR and Regional Development Funds

Frank Nuijens    Ronald Halfwerk    Johan Pragt

The EU runs a subsidy program called EFRO (Europees Fonds voor Regionale Ontwikkeling). Industry together with ASTRON has run several projects such as LOFAR and also SKA for developments of technology and production processes to enforce the Northern Region.

2022 marks the 15th anniversary of the EFRO Regiostars awards and a special celebratory campaign is setup, and 15 projects are invited to show their achievements.

LOFAR project has been selected from the pool of over 300 projects to be part of this campaign. In 2009 LOFAR was a finalist for the Regiostars yearly campaign.

Videos will be used in the campaign and filmed and produced by a local young journalist. This happened mid-August with SNN (Subsidy North Netherlands) shots in our backyard, ASTRON/LOFAR in the central hall and at the core and shots at the company Neways in Leeuwarden. There will be a social media campaign, including a public vote for the best project and a ceremony with the Commissioner Ferreira in Evora, Portugal, on 17th November.

This invitation is mainly due to the successful open collaboration of ASTRON technicians with industry over the past 15 years and the focus of ASTRON and LOFAR management and specifically the BTT (Bureau Technology Transfer) Group on joint projects and mutual strengthening of the region. It succeeded very well and is still running; the companies improved their production technology by our challenging LOFAR products and with that won new customers and new (challenging) orders. The Northern region got serious investments by LOFAR of several tens to hundreds of millions over the years and a high technology standard and exposure. Astronomy got the great and unique LOFAR telescope. This is a win-win-win situation running for already 15 years.

@astron

Subscribe to our newsletter. For previous editions, click here.

searchclosechevron-downlinkedin-squarefacebookbarsenvelopelinkedinxingyoutube-playinstagrampaper-planefacebook-officialpinterest-pwhatsappcommentingenvelopecross