WebDAQ — data acquisition for the LUNA experiment

WebDAQ reads the CAEN digitizers, writes the data, shows the spectra while they fill, records what the beam was doing and says something when a board stops. This manual is written for the person on shift; the later chapters are for whoever has to change the system.

The same material is available as a single PDF, built from manual/ in the repository.

Taking data

• WebDAQ — data acquisition for the LUNA experiment
  • Where to start
  • Key capabilities
  • Support
• Installation
  • 1. What the machine needs
  • 2. Installing
  • 3. The working directory
  • 4. Starting it
  • 5. Without hardware
  • 6. Checking the installation
  • 7. Updating an installation
  • 8. If the installation itself fails
• A complete session, start to finish
  • 1. Start the application
  • 2. Check the hardware
  • 3. Tune a channel
  • 4. Take a run
  • 5. What the run leaves behind
  • 6. Write it down
  • 7. Without hardware
  • 8. After changing the code
• User guide
  • 1. Finding your way
  • 2. Dashboard
  • 3. Logbook
  • 4. Stats
  • 5. DAQ
  • 6. Tuner
  • 7. Troubleshoot
  • 8. Alerts
  • 9. Settings
  • 10. What is refused, and when
• Troubleshooting
  • 1. The five recovery actions
  • 2. A board
  • 3. The beam current
  • 4. Spectra and rates
  • 5. Monitored values and Graphite
  • 6. Alerts that did not arrive
  • 7. The server
  • 8. The interface
  • 9. After a crash or a power cut
  • 10. Getting help

Subsystems

• CAEN digitizers: settings, filters and synchronisation
  • 1. Adding a board
  • 2. The signal chain
  • 3. Two worked examples
  • 4. Synchronising several boards
  • 5. Tuning while the beam is on
  • 6. Where the settings live
• Struck SIS3316 digitizers
  • Connecting the board
  • Adding it
  • Settings (DAQ page)
  • The board’s own utilities
  • Run files
  • Live time and scalers
  • Counter inputs
  • Test mode
• Spectra, ROIs and the histogram dashboard
  • 1. The dashboard belongs to the experiment, not to your browser
  • 2. Choosing the spectra
  • 3. Regions of interest
  • 4. Zoom, rebinning and display
  • 5. What is recorded with the run
  • 6. Waveforms and PSD
  • 7. When something looks wrong
• Online coincidences
  • 1. A gate is a sentence
  • 2. What a gate shows
  • 3. Sum spectrum
  • 4. Aligning channels in time
  • 5. Only channels on one clock
  • 6. Time between hits
  • 7. How it works, and what it costs
• Beam current and accumulated charge
  • 1. The two charges, and why they differ
  • 2. TetrAMM
  • 3. RBD 9103
  • 4. Monitored value (a Graphite metric)
  • 5. The current plot
  • 6. What is recorded
  • 7. Checks that catch most problems
• Monitoring: Graphite, Grafana, alerts and notifications
  • 1. Why a time-series database at all
  • 2. The Stats page
  • 3. The run’s stats file
  • 4. Grafana
  • 5. Alerts: Telegram and Zulip
  • 6. The Board Health tab
  • 7. What happened while nobody was looking
  • 8. What to check when a number is missing
• Run data and conversion
  • 1. The run directory
  • 2. The data file
  • 3. Converting to ROOT
  • 4. Checking a run that was interrupted
  • 5. Reading the run back in the Logbook
  • 6. Housekeeping
• Offline Analysis
  • Prepare runs
  • Nodes and scientific conventions
  • Spectrum analysis
  • Time analysis and ROOT trees
  • Example workflows
  • Workflows and results
  • Install and test
  • Editor rendering and layout
• The ELOG logbook
  • 1. Setting it up
  • 2. Reading
  • 3. Writing
  • 4. What is worth writing
  • 5. When it does not work

How it works

• How WebDAQ works
  • 1. The data path
  • 2. Starting from nothing
  • 3. Adding the boards
  • 4. What a run is
  • 5. Changing settings while running
  • 6. What is stored, and where
  • 7. Monitoring
  • 8. After the run
  • 9. Interfaces
• Directory structure
  • 1. Top level
  • 2. server/ — the Flask backend
  • 3. frontend/ — the web interface
  • 4. The repository versus a working directory
  • 5. Legacy and dead paths
• Server architecture
  • 1. The shape of the server
  • 2. Starting up: what main.py guarantees
  • 3. The application factory
  • 4. The HTTP surface
  • 5. The database
  • 6. Acquisition: who owns the boards
  • 7. Board failures, auto-restart and diagnostics
  • 8. Online spectra: what “spy” means now
  • 9. Online tuning
  • 10. The histogram dashboard and the ROIs
  • 11. Beam current and charge
  • 12. Graphite: two endpoints, two directions
  • 13. Alerts, notifications and recovery
  • 14. Grafana, ELOG and the run report
  • 15. Completed runs and conversion
  • 16. A run, start to finish
  • 17. Configuration files
  • 18. Test mode and the test suite
  • 19. The frontend as a client
  • 20. Working in this codebase