PLC tags
Running, faulted, starved, blocked and part counts are read over OPC UA, EtherNet/IP, PROFINET or Modbus TCP. Read-only — no change to your PLC program.
IoT Monitoring
See every stop, including the short ones
We capture machine state from PLC tags where they exist, and from current clamps, stack lights or part-counting cameras where they don't — then show OEE by shift, microstops and a loss tree your operators help explain.
The problem
The OEE number on the whiteboard is an estimate
In many plants, OEE is still assembled at the end of a shift from a paper log: a breakdown was "about 20 minutes", the changeover "went long", and the forty-second jams that happened thirty times never made the sheet at all. Those short stops are the hidden factory — capacity the line has but never delivers, invisible because nobody writes them down.
Automating the count is harder than it sounds on a mixed fleet. Newer machines have PLCs with useful tags. The press from 1998 has relays, or a PLC nobody wants to touch. A system that only works on the new machines tells you about the half of the floor that was already behaving.
Without automatic state data, every improvement meeting turns into a debate about whose memory is right.
How it works
Machine state, with or without a PLC
We pick the least invasive source that gives a trustworthy state signal for each machine, and mix them freely across a line.
Current clamps
A Pulsenode C on the main motor feed tells off, idle, running and loaded apart by current draw, without touching machine wiring.
Stack-light sensing
Light sensors on the existing tower light feed the red, amber and green states into the Gatepoint's isolated digital inputs — the machine reports its own status.
Vision part counting
A Lensnode V counts good parts and rejects on the outfeed, supplying the performance and quality terms of OEE where no counter exists.
Microstops and loss trees
Where the time goes, by shift
OEE multiplies availability, performance and quality. The loss tree breaks each factor into causes, per machine, per shift and per product, and microstops — stops shorter than a threshold you set — are recorded individually instead of disappearing into "slow running".
| OEE factor | Losses captured | Source |
|---|---|---|
| Availability | Breakdowns, changeovers, waiting for material or operators | Machine state plus operator reason codes |
| Performance | Microstops, reduced speed, idling | Machine state, cycle counts and ideal cycle time |
| Quality | Scrap, rework, startup rejects | Part and reject counts from the PLC or vision |
Reason codes and honest numbers
OEE people trust, not OEE people tune
Reason codes on a tablet
When a stop runs past its threshold, a tablet at the machine asks the operator why — one tap from a short list. Microstops are grouped automatically, so nobody is asked about every jam.
Short lists, written with supervisors
Each machine gets a reason list its operators recognize. Uncoded stops go to the shift lead's review screen instead of defaulting to "other".
A fixed ideal cycle time
Ideal rates come from the nameplate or a demonstrated best, set once in writing — not from last month's average, which quietly hides speed loss.
Planned downtime defined up front
What counts as planned is agreed before go-live. Changeovers are not excluded just to make availability look better.
Not a ranking of people
The data finds losses in machines and processes. Using it to rank operators is the fastest way to get bad codes.
Raw minutes beside percentages
Every chart can show the minutes behind the percentage, so a definition change can't flatter the line overnight.
Process
From machine survey to weekly loss review
-
Machine survey
For each machine we decide how state will be captured — PLC tags, current, stack light or vision — and what it will cost to connect.
-
Proof of Value
Up to 25 machines instrumented with loaner hardware, against success criteria agreed in writing, such as share of stop time with a reason code.
-
Definitions and rollout
OEE definitions signed off by operations, reason lists built with supervisors, tablets configured, and the remaining machines connected.
-
Weekly loss review
The Performance module in Perflens Console drives a short weekly review of the top losses, and the next improvement target.
Deliverables
What an OEE rollout delivers
- A machine-by-machine capture plan with the chosen state source for each
- Installed connections and sensors, documented in an as-built drawing set
- An OEE definitions document — ideal cycle times, planned downtime, microstop threshold — signed off by operations
- Reason-code lists for each machine and configured operator tablet screens
- Performance dashboards by machine, line, shift and product
- A data feed to your MES or historian over OPC UA or MQTT, plus CSV and Parquet export
- Training for operators, supervisors and the continuous-improvement team
Honest fit
When automated OEE is worth it
A good fit when
- A mix of old and new machines, some without a usable PLC
- OEE is filled in by hand and nobody fully trusts it
- You suspect microstops are costing more than the breakdowns everyone remembers
- Several shifts run the same line differently and you want to see how
Probably not when
- A single machine whose operator already logs stops reliably — a spreadsheet may be enough
- The goal is to rank operators; the data will be gamed within weeks
- You only want one plant-wide OEE number for a slide, not machine-level losses to act on
See it in context
Where OEE tracking pays off
Packaging, paper & wood products
Converting and packaging lines where short jams add up to hours a week.
Read more
Plastics & molding
Presses whose cycle time, scrap and changeovers decide whether a quote was profitable.
Read more
Microstops and the hidden factory
Daniel Okafor on the capacity your OEE number hides, and how to find it.
Read moreFAQ
Questions we hear often
Do you need to change our PLC programs?
No. PLC tags are read without changing logic, and machines without a usable PLC are monitored through current clamps, stack-light sensing or vision part counting, none of which touch the control system.
How do you define a microstop?
With you, per machine. A microstop is a stop shorter than an agreed threshold; the system records every one with a timestamp and duration, and groups them so operators are not asked to explain each one.
Will operators actually use the tablets?
They do when the list is short, the prompt only appears for stops that matter, and they can see their own shift's results. We build the reason lists with supervisors and adjust them after the first weeks.
Can the same sensors watch machine health?
Yes. A Pulsenode C that reports machine state also measures current imbalance and harmonics on the motor, so OEE tracking and predictive maintenance can share the hardware.
What does it cost?
A six-week Proof of Value covers up to 25 machines for $24,000, with loaner hardware. Ongoing monitoring is quoted against the published subscriptions on our pricing page.
Find the hour your line loses in short stops
A Proof of Value instruments up to 25 machines for six weeks and shows where the time actually goes — by shift, by reason, by machine.