FFA26 School Melbourne

Prepare for the Melbourne FFA26 School and explore three scaling FFA rings with OPALX and its native closed-orbit finder.

Return to the Schools catalog

Location: Melbourne, Australia
Year: 2026
Topic: Scaling fixed-field alternating-gradient (FFA) rings with OPALX

Course material

Download the FFA26 School Melbourne course ZIP

Extract the ZIP into a directory of your choice. It contains the notebook, Python helpers and OPALX input files in the layout used by the notebook:

opalx/
  FFA26_OPALX.ipynb
  pyfiles/
    school.py
    check_constants.py
    execute_notebook.py
    test_school.py
  school-16-cell/school-16-cell.in
  school-55deg-d-off/school-55deg-d-off.in
  school-fourfold-3MeV/school-fourfold-3MeV.in
  school-16-cell-probe/school-16-cell-probe.in
  school-16-cell-probe/particles.dat
  school-16-cell-sc/school-16-cell.in

Keep this structure intact. The probe input’s particle distribution is included; generated JSON, tracking results, caches and build files are excluded.

Prepare before the school

The notebook runs OPALX in Docker and Python/Jupyter locally. Install Docker Desktop on macOS or Windows, or Docker Engine on Linux. Windows users should use Ubuntu with Docker Desktop’s WSL 2 integration and run the commands in the Ubuntu terminal. Follow the manual’s Docker setup instructions for platform details. A local OPALX build, compiler, or MPI installation is not required.

Start Docker, check that it runs, and download the CPU image:

docker run --rm hello-world
docker pull ghcr.io/opalx-project/opalx:cpu

Use Python 3.10 or newer and create a virtual environment:

python3 -m venv "$HOME/.venv-ffa26"
source "$HOME/.venv-ffa26/bin/activate"
python -m pip install --upgrade pip
python -m pip install numpy pandas matplotlib h5py jupyterlab ipykernel

Change to the opalx directory inside the extracted course directory, replacing the example path with its actual location, and open the notebook:

cd /path/to/extracted-course/opalx
python -m jupyterlab FFA26_OPALX.ipynb

Leave the setup cell in Docker mode:

BACKEND = "docker"
IMAGE = "ghcr.io/opalx-project/opalx:cpu"
EXECUTABLE = None

Use the organisers’ image reference in IMAGE when supplied. The setup cell checks Docker and the local image, displays the architecture and OPALX revision, and fixes the image for the session. It stops if Docker or the image is unavailable; it never substitutes a local executable. No launcher script or executable environment variable is needed.

Before arriving, run the setup cell and, if possible, complete the first 16-cell ring lesson. For help, retain your operating system, the displayed image identity, and the complete error message or native run.log.

Work through the three rings

Each lesson follows the same progression: inspect the input, predict the behavior, calculate the closed orbit and full tunes, track small displacements, and interpret the field, orbit, and oscillation plots.

Machine Energy Main comparison Radial tune Vertical tune
Original 16-cell F/D ring 12 MeV Reference ring and scalloping 3.41066 3.39111
55-degree spiral, D off 12 MeV Changed spiral faces and vertical focusing 3.11774 1.18921
Fourfold ring 3 MeV Unequal F/D pairs and superperiod sampling 3.44739 3.43885

These rounded tunes are orientation checks for the unchanged inputs. Native COF computes the closed state and full linear tunes, including the integer part; Python reads and plots those results rather than estimating tunes externally. The reference circle used to place magnets is not the closed orbit. Each input supplies its own validated numerical settings.

All three core single-particle lessons have completed closed-orbit and tracking runs on the Mac’s native CPU build and the Docker ARM64 image, with full tunes agreeing within 2e-9. The fourfold section plots can alias the apparent oscillation: read the full tune from COF rather than counting the slow plotted wave. The optional probe input and its particle distribution are also included in the ZIP.

The notebook writes each experiment into a fresh directory under opalx/runs/. Its newly generated orbit.json supplies the closed state for tracking; standalone JSON snapshots are not loaded. Keep generated results local and restart the kernel before beginning a fresh session.

Optional native execution

For an existing OPALX executable with native scaling-FFA COF and full-tune support, explicitly change the setup cell:

BACKEND = "native"
EXECUTABLE = "/absolute/path/to/opalx"

The notebook checks the chosen executable and reports its version and revision. Docker remains the default for the school.