Experiment¶
The live operational view. This page describes the shape; the running app serves the data.
Unlike the configured zones (Hardware, Techniques, Operations, Governance), the subjects, runs, campaigns, datasets, and decisions are live per-experiment data. Their system of record is CORA's read-API, not a doc page, so this page gives only the shape and one illustrative example. See Model for the aggregate shapes.
- Subject: the sample being measured; custody runs
ReceivedtoReturned/Stored/Discarded. - Run: the operator-started execution of a Plan (the measurement batch); its task-lens companion is the Procedure. Both can produce a Dataset.
- Campaign: Runs grouped under a coordinated study; intent
Coordination/Series/Sweep/Block. - Dataset: a data artifact (URI, checksum, lineage) with intent
TrialorProduction. - Decision: a structured-audit choice by a human or agent; at 2-BM the
RunDebrieferrecordsNominalCompletion/DegradedCompletion/EquipmentAbort.
An illustrative thread¶
Not a live record. A sandstone-core Subject arrives under a proposal, is mounted on the sample tower, and a
tomography Run (in a Coordination Campaign) produces a Production Dataset; at completion the RunDebriefer
records NominalCompletion, and the Subject is Returned. The real instances are served live by the app.
Shutter state at run start¶
Both 2-BM safety shutters are open before a tomography run begins, opened by the operator at session start. The front-end FrontEndShutter (FES) is then kept open continuously for the thermal stability of the beamline optics and is not toggled per scan. The B-station StationShutter (the P6-50 SBS) is what 2-BM operators and TomoScan call the "fast shutter": there is no separate fast actuator at 2-BM-B today, so TomoScan cycles this same shutter closed for dark-field and white (flat) field acquisition and open for projections, many times per scan. CORA's run-start gate therefore expects both shutters open (the open predicate BeamBlockingM == 0 is defined on Enclosures), and should treat StationShutter close events during a run as normal dark / flat sequencing rather than anomalies. No separate FastShutter Asset is modelled. Confirmed by 2-BM staff (BEAM-1).