Sample¶
The endstation sample side: the solution positioning stack, the scanning-microbeam goniometer, and the HPLC delivery pump, plus where the fluidic delivery chain, the flow cell, and the solution Subject sit. First cut; PVs read from the NSLS2/lix-profile-collection startup files, carried confirm.
This is where LIX is genuinely different from the rest of the scattering fleet, so read it closely. The other scattering beamlines mount a solid sample in the beam. LIX measures a protein in solution: a buffer-borne macromolecule, often an eluting peak from in-line size-exclusion chromatography, flowed through an X-ray cell. The sample side therefore has two faces, the positioning hardware (which reuses the catalog) and the fluidic delivery chain (which is mostly the seam plus the Subject / Supply / Procedure shape). The hardware is modelled in the sample stage of the descriptor; the delivery chain is described here and in Controls.
The sample side at a glance¶
| Asset | Family | PV | What it does |
|---|---|---|---|
SampleStage |
Manipulator |
XF:16IDC-ES:Scan{Ax:XC} |
the solution-mode positioning stack: a coarse x and a z pusher (EPICS) plus the fast scan x / y (XPS trajectory); it places the flow cell in the beam (SAMPLE-1) |
ScanningGoniometer |
Goniometer |
XF:16IDC-ES:Scan2-Gonio{Ax:sX} |
the scanning-microbeam stack: sample x / z, tilts, a rotation, and (with the XPS rot.rY) micro-tomography of cells and tissue (SCAN-1) |
DeliveryPump |
FlowController (loose) |
XF:16IDC-ES{HPLC}REGEN:FLOWRATE |
the HPLC pump that flows the solution and SEC peak through the cell (FLUID-1, FLOW-1) |
Placing the sample¶
LIX runs two endstation modes, selected at startup, and each has its own positioning stack.
The solution mode uses the SampleStage, a positioning stack that carries a coarse x and a z pusher as EPICS motors, with the fast scan x and y carried as Newport-XPS trajectory axes. It places the X-ray flow cell at the focus. It reuses the catalog Manipulator Family, the same multi-axis sample-positioning anatomy the soft-X-ray beamlines SIX and ESM earned, so LIX adds no device here. Which axes are EPICS and which are XPS trajectory, and how the flow cell mounts, is SAMPLE-1.
The scanning-microbeam mode uses the ScanningGoniometer, a SmarAct stack with sample x / z translations, tip / tilt, and a rotation axis. With the XPS rot.rY trajectory axis it performs micro-tomography; the fast raster axes (the XPS scan.X / scan.Y) sweep the microbeam across a cell or tissue section for mapping. It reuses the catalog Goniometer Family. The raster and tomo axes live on the Newport XPS trajectory controller, the motion-controller seam, and are carried as settings rather than separate Assets (SCAN-1).
Delivering the solution: the fluidic chain¶
The fluidic sample-delivery chain is what makes LIX a solution beamline, and it is the one genuinely-new axis this deployment brings. CORA models it the way the MX3 deployment modelled its non-EPICS hardware: as a heterogeneous control plane, with the interface named and the actuators placed where they belong, not forced into device vocabulary they do not earn.
The delivery pump is the one device. The DeliveryPump drives the solution and SEC-SAXS flow: a flowrate setpoint and readback, a pressure readback, and run / stop. It is heterogeneous underneath, a pcaspy soft-IOC (XF:16IDC-ES{HPLC}) fronts an Agilent quaternary pump driven over the OpenLAB .NET SDK on a Windows host and a regeneration pump driven over a raw TCP socket to a Moxa terminal server, but its CORA-facing anatomy is a settable flow actuator presenting Regulator. That is exactly the existing loose FlowController Family, the one i22 and 7-BM already use. So the pump reuses FlowController; it coins nothing. LIX is its third consumer, which fires the rule-of-three (see Model).
The rest of the chain is the seam plus Subject / Supply / Procedure, not devices:
| Part of the chain | How CORA models it | Why |
|---|---|---|
| The VICI and Aurora selector valves | the ControlPort seam (FLUID-1) |
discrete N-position routers (column, buffer, detector selection) over Moxa TCP sockets, with no existing Family; CORA conducts them over the seam, coining no Valve Family at n=1 |
| The size-exclusion column and buffers | Supply consumables (SEC-1) |
a column is chosen per sample (a config / consumable, not a device); buffers are selected on the Aurora valve |
| The X-ray flow cell | sample environment (SEC-1, FLUID-1) |
the cell lives in an external library (lixtools); its positions are config, not catalog devices here |
| The sample-handling robot and the autosampler | a Procedure + a Subject custody thread (ROBOT-1) |
the task-verb-driven robot (the SW: method soft-IOC) and the Agilent autosampler fold to a Procedure over the spine, the i03 / MX3 robot precedent, not a device Family |
| The solution sample / eluting peak | a Subject (SUBJECT-1) |
the thing measured is a liquid macromolecule or a chromatographic peak, with its own provenance, distinct from a solid mount |
The solution Subject and the SEC-SAXS Procedure¶
The deepest novelty is not hardware at all. For SEC-SAXS, what gives the experiment its identity in CORA's record is the Subject (which protein, which buffer, which eluting peak), the Supply (which column, which buffers, the needle wash), and the Procedure (purge, equilibrate, inject, flow-during-exposure, fraction). The HPLC pump and valves are the actuators that Procedure drives over the ControlPort; the elution profile is the acquisition axis, with the SAXS frames correlated to the chromatographic peak. CORA owns the Subject, the Supply, and the Procedure as its system-of-record concern; it conducts the fluidic actuators over the seam (FLUID-1, SEC-1, SUBJECT-1). These aggregates are not instantiated in this descriptor-and-docs cut; they are the shape the deployment will take, recorded so the reader knows where the solution experiment lives.
Sample environment¶
The sample-cell temperature controllers, an AccuThermo FTC100D and an SMC chiller, have their module-level instances commented out in the profile collection (a solution mode instantiates an FTC100D), so this is a scope deferral rather than a clean absence: no temperature-controller device is modelled in this cut, and the in-situ temperature environment is carried pending (TEMP-1). The one temperature mirrored to a PV is the autosampler thermostatted-tray temperature (SAMPLER:TEMP), part of the fluidic seam and folded into the same deferral.
Why no new Family here¶
The positioning hardware reuses the catalog throughout: Manipulator for the solution stack, Goniometer for the scanning stage. The one fluidic device, the delivery pump, reuses the existing loose FlowController Family rather than coining a new one. The selector valves, the column, the flow cell, the robot, and the solution sample are deliberately not coined as device Families: they are the seam plus the Subject / Supply / Procedure shape, which is where a solution beamline's novelty belongs (FLUID-1, SEC-1, ROBOT-1, SUBJECT-1). Nothing here graduates and the catalog is unchanged.
See Open questions for the sample-side facts still to confirm, Inventory for the Asset tree, Model for the family-reuse rationale and the FlowController rule-of-three, and the source walk for the PVs as read from the profile collection.