Business management · Shop Floor (MES)
The execution layer of production planning: a kiosk terminal where the operator signs in with badge and PIN, a versioned work instruction frozen into the order, production reporting with scrap and rework, classified downtime reasons, labor hours by operation, andon calls to maintenance or quality, a live TV board, Quality inspection blocking the control-point operation and automatic machine data collection over PLC, OPC-UA or MQTT.
The problem and the fix
If any of these situations sounds familiar, it is because they happen in almost every graduation company, until the right system comes in.
Production is written on paper and keyed in the next day, when nobody remembers what stopped.
A terminal at the machine, with badge or PIN login and large buttons: setup, start, run, pause with reason, resume and complete. The record is created at the moment of the event.
Nobody knows how much the machine actually produced, or why it was down.
Downtime classified as planned or unplanned, OEE by work center and a live board showing what is running, what is down and for how long.
The operator calls the mechanic by shouting down the aisle, and nobody measures how long the response took.
Andon on the terminal: the call is created with type, time and work center, measures time to response and time to close, and shows up on the TV board.
The wrong part got past the control-point operation because the inspection was an informal arrangement.
An operation flagged as a control point requires Quality inspection before moving on, with the inspection plan frozen into the order. If it fails, a nonconformity report is opened and linked to the order, the operation and the lot.
The work instruction sits in a folder that changed after the order had already started.
The instruction is versioned, published and frozen when the order is released, together with the bill of materials and the routing. On the terminal the operator sees the version that was in force when the order was issued.
In practice
Everyday situations at graduation companies, from the problem to the result, using the modules on this page.
The operator takes over the machine and needs to know what to produce and how.
He badges in at the terminal, picks the order for that work center, reads the work instruction for the operation and reports setup and start. His hours start counting against the operation.
Actual labor cost per order, with no end-of-day reporting from memory.
The line stops mid-run and the supervisor only finds out later.
The operator logs the pause with a reason from the catalog and triggers the material andon. The TV board shows the work center down, and time to response starts counting.
The downtime enters OEE with the right cause and response time becomes a metric, not an impression.
The company wants to stop relying solely on what the operator types.
The collector service reads the counter and the machine state over MQTT or OPC-UA, aggregates per minute and sends it in. The record is created tagged as machine-sourced, and downtime with no reason becomes an item for the operator to classify.
OEE then shows the gap between what was reported and what the machine measured.
How it works
Every step is a real module, and what comes out of one goes into the next without anyone typing it again.
Module by module
7 modules and 38 features on this page, all running on the same class, graduate and event records.
The screen that stays at the machine, carrying only what the operator needs.
The right procedure, in the version that applied to that order.
Who worked, for how long, on which operation.
The shop floor visible from across the room.
The check happens where the part is, not afterwards.
The chart that shows the process drifting before the defect appears.
Data that comes from the equipment, not from recollection.
Compare the modules and choose the setup that fits your operation.