Quality, concepts
The Quality nav group is where defect and process-capability data lives, separate from the day-to-day Now and Floor screens because it’s usually owned by a different role (quality engineer/admin) on a different cadence (per-shift review and per-study analysis, not real-time watching).
What’s in this area
| Screen | Route | Job it does |
|---|---|---|
| Quality | /quality | First-pass yield summary, a defect-by-type Pareto, and a trend chart. |
| Capability | /capability | SPC control chart + Cpk stability score for a chosen telemetry source/metric. |
Domain objects
First-pass yield and defect types
First-pass yield is the percentage of units that pass quality on the first attempt (no rework, no scrap).
The Quality screen shows this alongside a defect Pareto, defects grouped by configurable
type, ranked so the costliest defect type is obvious, and a trend chart over the same time window, so a
spike is visible against recent history instead of in isolation. Defect types themselves are configured on
the Defect Types tab of Codes (/codes?tab=defects, alongside Downtime Reasons on the same page).
A claimed station’s Job tile counts scrap as a plain quantity (+ Scrap), it does not yet attribute a
scrap count to a specific defect type, so the Pareto only shows types you’ve configured here and reflects
whatever has been recorded against them.
SPC and Cpk (Capability Lab)
SPC (Statistical Process Control) watches a single telemetry metric over time and flags points that violate control-chart rules (a value drifting outside expected variation). Cpk is a single stability score computed from that same data, roughly, how comfortably the process’s actual variation fits inside its specification limits. The Capability screen has a source dropdown and a metric dropdown (the metric list only shows continuous process measurements on the picked source, since SPC/Cpk can’t say anything meaningful about a boolean, state, or counter tag). Pick both and click Analyze, after which it renders the control chart with any rule violations flagged, plus the Cpk score.
Cpk needs a saved upper and lower spec limit to compute at all, entered once on the Capability screen and reused every time you come back. Once both are saved, Spall also checks that signal on its own on a schedule, whether or not someone has the page open, and raises a notification the moment Cpk drops below 1.0 or the signal goes out of control, clearing it again once the signal recovers. No saved limits means no automatic checking, only the pull-based chart above.
How the pieces relate
Quality aggregates defect events recorded against a configured defect type. Capability is a separate, deeper investigation tool, it works off raw telemetry, not defect counts, so it answers “is this process in control” instead of “how many bad units did we make.” Both ultimately serve the same goal. Floor’s Production screen never fabricates a yield number for a shift with zero output, it shows an em-dash instead.