LOFAR Development Newsletter December 2023

André Gunst    Boudewijn Hut

The LOFAR2.0 stations will deliver more data than the current stations, which would just not fit on the present network infrastructure for the core stations. Since the future ambitions of LOFAR are to quadruple the data streams, we (LOFAR2.0 program) have decided to prepare for that ambition. We are doing so by purchasing equipment which prepares us to stream 100 Gbit per second (Gbps) per station towards the CEntral Processing systems (CEP). The most expensive part are the high bandwidth optics to bridge the long distance. To overcome that for the core stations, we will aggregate the station data via many cheap short distance 100 Gbps connections to the Concentrator Node. Switches in the Concentrator Node will aggregate the data to a less amount of expensive long distance 400 Gbps connections. Since, the switches in the Concentrator Node and the long 400 Gbps optics drive the cost the amount is minimized to what is necessary for LOFAR2.0. This will result in a net data stream from the stations connected to the Concentrator Node of 25 Gbps. Later, when funding for the four beams is found this can be easily upgraded to the full 100 Gbps. By this, we are more than ready for the future!

Ágnes Mika

On the 6th of December, the Station project had its final review and handed over its results to the Assembly, Integration, Verification and Validation (AIVV) project during a festive event. Achieving this major milestone was preceded by a lot of hard work by the team.

October marked the completion of the commissioning program. We can now safely say that the hardware used to build the LOFAR2 Test Station Phase 2 is functioning as intended. We released the documentation of the computing boards to the manufacturers who will shortly start up their production process. This batch of boards will be used in the spring of 2024 to equip the so-called Production Test Station, consisting of three LOFAR2 stations.

During November we focused on completing the documentation pack for the review. The documentation includes the design of all parts of the system, functional descriptions, verification instructions, a list of the remaining risks and the plan for achieving the CE certification.

After the review, the Station team has become part of the AIVV project. They are focusing on completing the production test setups (which will be delivered to the manufacturers of the computing boards so that they can test the boards before shipping them to ASTRON) and on creating the documentation for upgrading the international stations.

Between the review and the transition to AIVV we celebrated the success of the Station project. The team showcased the results of over five years of challenging work during a presentation session and an exhibition with posters, screens showing live observations and monitoring information from the Test Station and, of course, a large piece of cake for all present.

Well done, Station team!

Carla Baldovin

DANTE’s first light was achieved on 10 October when a full HBA tile of L2TS (CS001) was upgraded with the DANTE frontends. The boards were installed, tested and removed again a week later.

In the image, station CS001 is visible in the background, the plot shows the comparison between the LOFAR frontend and the DANTE frontend. And adjustment to the FM cutoff (below 120 MHz) of the DANTE frontend is needed, to match the current HBA frontend. The table next to the image presents the overview of how the two versions of the frontend compare for different parameters: the DANTE frontend shows a clear improvement in gain, linearity, noise, among others. Overall, the week of testing was successful, no stoppers were found.

Minor corrections to the board schematics and layout are currently being implemented. The goal is to have a final version of the design by mid-December and to start production of the next batch in January 2024. This production run will deliver 80 boards that will be installed in 5 tiles in the field in Spring next year.
In parallel, the team is conducting tests aimed at detecting failures when the boards are installed in the field on real environmental conditions, for a long period of time. The tests are now being conducted on LOFAR1 boards, which reliability is by now well established. They should be able to reproduce the failures observed. Once the tests are validated, they will be performed on the DANTE boards to obtain a realistic estimate of their lifetime and reliability.

Arno Schoenmakers

Changes in Team Ruby way of working

The software team for LOFAR2.0 development, team Ruby, has changed its way of working. On top of the three-week Sprint cycles, we have added a larger three-monthly Planning cycle. Priorities will be determined with all stakeholders for three monthly periods, which will bring more focus on developments spanning multiple sprints and more connection with the stakeholders.

We have started to work in this way on 31 October, which saw our first Planning interval planning meeting (see pictures below). Not every aspect in this process will be perfect from the start, but both the team and the Stakeholders are committed to making this process work and we will gradually adapt until we have a satisfactory process. The next planning session is end of January 2024, and we are already busy preparing ourselves in a better way for that session.

TMSS

An outcome of the first planning session was that we need to deliver a first feature complete and workable version of TMSS that will allow Operations to run LOFAR Cycle 20. We have named this TMSS V1.0. Together with our stakeholders we have made a list of what should still be part of this TMSS V1.0, and what not. The latter will have to be added later on, most likely after Cycle 20. In the meantime, only functionality needed for other LOFAR2.0-related features will be added to TMSS, and the groundwork will be done for a better designed TMSS Frontend. The new Frontend must enable more flexibility for future extensions and improvements and must allow more and faster maintenance updates for the used external packages in the frontend.

Station

Station development has been targeted at supporting L2TS commissioning activities and L2TS test station development. After ICT has updated the LCU2, we have deployed a new Nomad/Consul based deployment and internal network/control system. These technologies will help us to deploy new software to the station(s) in a better controlled way and will help shielding the station’s internal interfaces.

Also, we are working on the continuous capture and centralized writing of all statistics data from the Uniboards. This feature will provide access to this data at any time, for real-time diagnostic purposes or any other purpose statistics data serve. The LOFAR2.0 station provides this data by default, instead of having to explicitly ask for it every sample, as is the case in LOFAR1.

Our International station users will have to learn how to work with a LOFAR2.0 station. As a first introduction to that, we have explained and demo’d the new station control interface to a group of International station owners at the ILT-TO meeting on 9 November. More interaction is needed the coming period, of course, but we saw this as an opportunity to make them aware of the many changes that will come with the LOFAR2.0 upgrade of their own stations.

Arno Schoenmakers

For LIFT, we are making an inventory of the software needs based on the LIFT minimal product definition. A first software design has been created, and this will be reviewed after the LOFAR2.0 Station review. If the design passes the review, we can estimate the amount of work to implement the design.