Machine data you can act on: live OEE and a record of every part
What you get
| Live status | Running, idle or down for every connected machine, with the reason for every stop. |
|---|---|
| Counts and cycle time | Good and rejected parts per machine, per shift, against the target. |
| OEE | Availability, performance and quality — per machine, line and shift — and the losses behind them. |
| Part traceability | Every part's serial with its results, settings, station, operator and time. |
| Reports | Shift and daily summaries, downtime by reason, and exports to Excel. |
| Alerts | A message when a machine has been down too long or a reject rate climbs. |
How machines are connected
- PLCs with Ethernet: tags read over Modbus TCP, OPC UA or EtherNet/IP — counts, states, results and alarms.
- Serial devices and instruments: Modbus RTU over RS-485, such as pressure indicators and energy meters.
- Older machines with no usable PLC data: a small I/O module reads a cycle signal and a running signal.
- Data published over MQTT where a plant already has a broker, or stored straight into a SQL database.
- Operator input — downtime reasons, scrap reasons — on the HMI or a tablet at the machine.
Every machine we build already publishes its counts, results and cycle times in tags a system like this can read, so new machines join without reprogramming.
OEE in one worked example
Overall Equipment Effectiveness multiplies three losses together: time the machine was not running, running slower than it should, and making parts that are not good.
OEE = Availability × Performance × QualityAn 8-hour shift has 480 planned minutes. The machine is down for 60 of them, so it runs for 420 minutes: availability is 420 ÷ 480 = 87.5%. Its ideal cycle is 30 seconds, so in 420 minutes it could make 840 parts; it made 700, so performance is 700 ÷ 840 = 83.3%. Of the 700, 686 were good, so quality is 98%.
OEE = 0.875 × 0.833 × 0.98 ≈ 71%The figure of 85% often quoted as world class comes from Total Productive Maintenance: 90% availability × 95% performance × 99.9% quality. It is a reference point, not a standard — the number that matters is your own trend, and which of the three losses is biggest. In the example, the hour of downtime and the slow running cost far more than the scrap.
Traceability: from a complaint to the exact parts
Without traceability, a complaint about one bad part can mean holding or recalling a month of production. With it, you can name the batch, the station, the settings and the shift — and contain only what is actually affected. Automotive customers working to IATF 16949 expect it.
What gets recorded for each part
- A unique serial — printed barcode or QR label, or a marked Data Matrix for parts that are machined or washed after marking.
- Each station it passed, with time, operator and machine.
- The measured results and limits: leak value, press force, torque, test values.
- The batch or serial of components fitted to it, so an assembly can be traced back to its parts.
Interlocks that keep the record honest
- No scan, no start: the machine does not cycle on a part it cannot identify.
- A part that failed or skipped a station is refused at the next one.
- A rejected part is recorded as rejected, with the reason, before it leaves the station.
Tell us the machines, their PLCs and what you most need to see. We will propose a pilot on one line, with a budget.
IoT and traceability questions
Do we need new machines to get OEE data?
No. Machines with a PLC usually already hold counts, cycle times and fault states that can be read over the network. Older machines without a usable PLC get a small I/O module that reads a cycle signal and a running signal — enough for counts, run time and downtime.
Is it a cloud subscription?
It does not have to be. The data can stay on a PC or server on your own network, with dashboards in a browser, the same way ElectronIx DNC runs on the shop LAN. Cloud access can be added if you want to see the floor from outside the plant.
How do we capture why a machine stopped?
The machine knows that it stopped; only the operator knows why. A reason screen on the HMI or a tablet asks for it — tool change, no material, breakdown, quality hold — and that list is what turns a downtime number into something you can act on.
What does traceability need on the part?
A unique identity: a printed barcode or QR label for most parts, or a laser- or dot-peen-marked Data Matrix for parts that are washed, machined or heat-treated after marking. Each station scans it, and the machine refuses to start on a part that has not passed the previous station.
Can it connect to our ERP?
Usually, through a database export or an API, scoped per project. Start with what the floor needs — status, counts, results — and add the ERP link once the data is trusted.
Where do we start?
With one line or one machine. A pilot proves the connection, the dashboard and the reason codes on real production before anything is rolled out across the plant.
Have a job that a machine should be doing?
Send the part and the target output. You get a concept, a sequence and a budget range back — free, and with no obligation to build it with us.