The challenge
The line was productive but blind. Like many South African factories, the plant ran capable but ageing machines with no modern PLC exposing data, so "OEE" was reconstructed monthly from paper logs and gut feel. By the time the number landed, the shift it described was long gone — and nobody could agree on why the line underperformed, because the downtime reasons were never captured at the machine. A proposal to "get data" by upgrading the controls had stalled: the cost killed the business case, and a shutdown to rewire a working line was a non-starter.
The trap: believing you must replace the controls to measure the line. You don't — an old machine still draws current when it runs, still moves product past a point you can count, still shows state on a stack light. That's a sensing problem, not a controls upgrade.
Our approach
- Edge sensing, no PLC contact. A clamp-on current sensor for run/stop and load, a proximity/photo count sensor, and a stack-light tap for state, wired to an edge device. The machine's control logic was never touched (the approach in our OEE-without-a-PLC guide).
- Operator reason codes. A small touchscreen where the operator selects why the line stopped, turning raw downtime into changeover, material-starve, breakdown and the rest.
- Live OEE on the floor. Availability, performance and quality by line and shift on a shop-floor screen and on phones, instead of a monthly spreadsheet.
- Resilient by design. The edge device buffers locally and syncs when power and network return, so load shedding never loses the production record (see load-shedding resilience).
- Into the office. The same data flows toward ERP/MES, so production and the office finally see one set of numbers.