The operator frames a good spot, confirms the two DC levels the detector
reads there, and the rig then measures its own tilt: step 1.5 mm either side
on X and then on Y, and tilt the platform until those levels come back. The
correction that fixes an offset point is the correction that levels the whole
travel — height error and tilt effect are both proportional to the offset —
so the procedure ends by applying it and leaving it applied.
Both directions are measured from the same starting tilt and averaged, which
makes their disagreement a flatness read-out rather than something averaged
away silently.
core/auto_align.py holds the geometry and the search, Qt-free. The three
T-axes' azimuths are the whole geometry: T1 lies along +X so it alone tilts
along X, and T0/T2 move as an equal-and-opposite pair to tilt along Y without
touching X (tilt_response derives that, and the tests pin it — an axis map
that drifts would still converge, on the wrong axis). The search is a secant
null on the split-detector difference: probe once to learn what a microstep
is worth, sign included, then step at the null. It refuses to servo on a
scope that has not re-triggered, escalates a probe that reads as no response
before calling an axis dead, and stops at a per-axis travel limit.
gui/align_bridge.py runs it on a worker thread; stopping is a threading.Event
rather than a queued command, because the worker is inside a long handler for
the whole run. The camera window carries the button and the progress window,
and locks the scan panel and the jog pads while a run owns the stage.
Adds immediate MEAN measurements and an acquisition count to the scope
driver, and read_bias_mv to core/scope_inspect — the one scalar the
inspection state was missing.
KNOWN_ISSUES.md records what only the rig can settle: the probe step, the
travel limit, the hold current, and whether the piston the X phase applies
alongside its tilt matters.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A full multi-angle scan takes hours, and a rig whose angles disagree produces
all of them before anyone finds out. This adds a test mode that acquires one
row per angle — the row-wise middle of the ROI — and a viewer that puts every
angle's SAW frequency on one graph. The default 80×50 mm ROI at 5 angles goes
from 1461 rows to 5.
Why the middle row answers an alignment question at all: build_plan centres
every angle's rotated bounding box on the same nominal ROI centre, so each
angle's middle row crosses that one point on the sample. All the angles
measure the same material, so a spread in their frequencies belongs to the rig
rather than to where each row happened to land. test_every_angles_middle_row_
crosses_the_roi_centre pins that premise, since the whole comparison rests on
it and nothing else in the geometry code would notice it breaking.
core/saw_check.py — both halves of the mode, kept together because neither is
much use alone. middle_row_plan() reduces a ScanPlan to one row per angle
(n_rows // 2, the upper of two centre rows when even); frequency_traces() and
alignment_summary() turn the resulting file back into per-angle frequency
traces and the scalars an operator is actually asking about — the spread of
the per-angle medians, the worst drift along a row, the sparsest row. The
verdict thresholds are labelled as rules of thumb, not physics: an anisotropic
sample genuinely varies with angle, so a wide spread is a prompt to look at
the curves rather than a verdict.
Format v10: byte-identical to v6, one row per angle. The version byte earns
its keep because the two are otherwise indistinguishable — a v6 scan aborted
after its first row is not a check, and a reader guessing from the row count
would read a failed scan as a deliberate measurement. create_scan_file()
enforces the one-row rule at write time, since nothing downstream can recover
from a v10 file that breaks it. ScanEngine gains file_version and is otherwise
untouched: the acquisition, the abort/pause path and the background capture
are the scan's, unchanged.
sras_scan_manager.py now carries the source file's version through an export
instead of stamping v6 on everything, which the wider reader would otherwise
have made a lie.
saw_check_viewer.py — frequency along the row, one curve per angle, over a
common offset axis so the curves lie on the same piece of sample; a summary of
each angle's median ±1σ against angle; and the per-angle numbers in a table.
Analysis parameters (DC threshold, background, time gate) recompute on a
worker thread; display ones (smoothing, axis, MHz↔m/s) only redraw. A full v6
scan opens too — the same middle row is pulled out of it — so a finished scan
can be re-examined with the check's own read-out.
In the app, a check finishes by handing the operator the file and an "Open
Viewer" button rather than shutting the rig down the way a completed scan
does. Burst mode is not offered: one row per angle means every burst would be
a single row, so it buys nothing and still pays for the gate preflight.
137 tests passing, ruff clean.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
- ruff config, offscreen smoke tests for all 7 GUI apps/panels
- golden v6 .sras fixtures (complete + 4 truncation variants) generated
by the pre-refactor writer, with expected header/frontier JSON
- golden geometry fixtures from the pre-refactor _build_scan_params
- consistency tests proving current code reproduces the goldens
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Thomas Ales