Participants: V. Friese, D. Hutter, P.-A. Loizeau, A. Toia, S. Zharko

● mCBM beamtime March 2025

Observations:

• Second diamond detector was installed (already Friday 7 March); walk calibration of T0 against second diamond by N. Herrmann, provided the evening before data taking; the updated calibration parameters were not converted to the format readable by the online binary.
• TRD-1D had problems and was taken out of the data taking.
• HadronAnalysis ran smoothly.
• The online binary was operated with the V0 topology trigger (including timeslice CA tracking). Uncertainty about proper trigger settings. However, after track pair time and distance cuts, the material distribution of the setup is seen (conversion pairs).
• Spiky structure in the delta_t distribution; to be understood.
  
  

Goal for the ECE meeting in May: Replay of data set (Au+Au, March 2025 and/or Ag+Ag February 2025) with updated TOF calibration and optimised trigger settings. Record DigiEvents with V0Trigger. Monitor CPU load. Offline analysis (KFPF, Lambda) of triggered events, to be compared to offline analysis of full data (starting from TSA). Selling plot: comparison of Lambda inv. mass spectrum for both cases.

To be done towards that:

• Study track variables relevant for the V0 trigger (z position of first measurement, impact parameter in target plane, pair time difference, pair distance, pair vertex z coordinate) in order to arrive at substantiated cut values. In particular, study the time difference distribution in order to reliably distinguish same-event pairs from pairs from different events.
• This needs running the online binary on recorded tsa files with unpacking and timeslice reconstruction only (no event output), and import of CA tracks found online into cbmroot, in order to analyze them. The latter feature is not yet available and will have to be adressed first (S. Zharko).
• For the study of the trigger variables, probably no large statistics is needed (one tsa file as input may suffice).

After this study and demonstration, the learning potential with mCBM seem largely exhausted. We will then (after May) concentrate on processing simulated data of full CBM. An open and serious problem remains the timely derivation of calibration parameters, which needs to stay high on our priority list.

● Updates / Progress

Not much to report so far. B. Sobol had a look into CPU consumption and multi-threading (local machine); he will report next week.

● Data model: RNTuple

RNTuple seems a most promising way towards an integrated online-offline data model. We will address this after having finished the current study of mCBM data and data processing in May.