A PLC may already know when the machine is running and how many cycles it has completed. The production team needs a different answer: what did we make on this shift, where did we lose time, and what should we check?
The gap is the meaning attached to the signals. A useful first project connects a small set of reliable values to the actual production workflow.
Start with one production question
Choose a question that someone will use every day. For example: “How much of the planned shift did this machine spend stopped?” Agree who will read the result and what they would do differently with it.
For that question, you need a stable machine-state signal and a planned production window. For output and OEE, you also need counts, a product-specific cycle time, and a good-unit rule.
Map the signal before building the dashboard
Use a short mapping table. These are example fields, not universal PLC addresses.
| Input | What to record | What to check on the machine |
|---|---|---|
| Machine state | Meaning of each state and its timestamp | Running, waiting, stopped, and faulted states match the operator’s view |
| Total counter | Unit, reset rule, rollover, and sampling | One finished part changes the count as expected |
| Good or reject count | Where the decision is made | Rework and rejects are classified once |
| Cycle time | Product and unit | Seconds per part are not confused with parts per minute |
| Shift or order | Start, end, product, and planned exclusions | The report covers the intended production window |
Keep machine identity stable. Renaming a workcenter should not detach the new display name from its earlier history.
Handle resets and missing data explicitly
A counter falling from 9,800 to zero may mean a reset, not negative production. A value that stopped updating should not look like a healthy flat trend. Record the time of the reading and whether it was successfully acquired.
During commissioning, compare the application with the machine through a normal cycle, a stop, a restart, and a counter reset. Check what happens when the connection is interrupted. These few events reveal more than a dashboard full of unverified tags.
Add the information the PLC does not have
The machine may not know why an operator stopped the line, which order is active, or whether a part passed an inspection. Put that information in the workflow where it is known.
An operator can select a downtime reason. Production planning supplies the intended time window. The product record holds the agreed cycle time. Quality records supply the accepted quantity.
The OPC Foundation’s OEE example also separates production measurements from planned values usually supplied by MES or ERP.
Check one shift from beginning to end
Take an illustrative shift with 450 planned production minutes, 390 running minutes, 720 total units, and 690 good units. At an ideal cycle time of 30 seconds, availability is 86.67%, performance is 92.31%, quality is 95.83%, and OEE is 76.67%.
Before accepting that report, reconcile it with the operator’s count and stop record. Investigate differences at the time they occurred. Do not adjust the final percentage to make the dashboard agree.
Use the OEE calculator and worked example to check the arithmetic.
Where Orenda fits
Orenda Box collects configured machine data and runs local applications. Orenda MES adds production records, downtime, and reports. Authorized users can open the relevant application remotely through Orenda Connect.
Begin with one machine and a small mapping. Once a complete shift makes sense to the people on the floor, add another machine or a more detailed question.
