How to measure OEE on an old line without buying a new PLC

Who this is for

Plant owners, factory managers, and engineering leads running production lines on aging controls — typically 5–20 years old, often a mix of PLC generations and brands — who suspect their assumed OEE is wrong and want to know how wrong without committing to a controls modernisation project first.

The mistake this article exists to prevent

"We can't measure OEE — our PLCs are too old. We need to replace them first." We hear a version of this on most digitalisation conversations. It is wrong, and following it usually wastes time and capex. The PLC and the OEE measurement are different problems with different solutions.

OEE measurement is fundamentally about counting good output, counting time, and attributing both to a known operating state (running / down / changeover / idle). None of that requires modern PLC integration. It requires sensors and a dashboard. Modern PLC integration helps; the absence of it is not a blocker.

The five things you need to install

1. Edge gateway (industrial PC)

One per line. Typical specification: industrial-grade PC or ARM-based gateway with USB / Ethernet / serial I/O, mounted in the MCC or in a small ancillary panel. Reads signals from the field, time-stamps them, buffers locally, and either pushes to a cloud dashboard or hosts a local one.

Vendor lock-in question: choose a gateway that exports its data in open formats (CSV, SQL, OPC-UA) without subscription dependency.

2. Output and reject counters

Most lines already have these in some form — at the very least, you can wire to the existing output counter on the discharge or palletiser. Where reject counters do not exist, a simple optical or inductive sensor on the reject chute usually closes the gap.

Signals are typically 24V pulse counts. The gateway reads pulse counts and converts to throughput per minute / hour / shift.

3. Current clamps on critical motors

Three to six motors — the main drive, the conveyor pacer motor, the pump, the fan, depending on the line. Current clamps wrap around existing motor cables (no rewiring) and report motor load, run/stop state, and basic energy.

Why this matters: motor current is the most reliable "is the line actually running?" signal on an old PLC-less line. It catches micro-stoppages that a counter alone misses.

4. Optional condition signals — environment, vibration

Where line quality correlates with environment (food and beverage hygiene zones, climate-sensitive products) or where bearings are near end-of-life, ambient temperature, humidity, and critical-bearing vibration signals add diagnostic depth at modest extra cost.

5. A working dashboard and an alerting layer

Browser-accessible dashboard on a shop-floor TV and on the supervisor's phone. Three views matter:

• Real-time line status — running / down / changeover, plus today's running totals.
• Shift performance — output, reject, OEE, last-shift comparison.
• Stoppage analysis — Pareto of reason codes, frequency, duration.

Plus a basic alerting layer — WhatsApp or email — for stoppages longer than configurable thresholds. Operators tag stoppage reasons on a tablet or HMI.

What the install actually looks like — and what it does not touch

• Does not touch: existing PLC code, existing SCADA, existing controls logic, machine wiring. The retrofit is parallel infrastructure that reads what the line already does.
• Does touch: the MCC (to install the gateway + 24V supply + clamps in their own ancillary section), the operator workspace (small tablet for reason codes), and the shop-floor visualisation (one TV).
• Install duration: 2–4 weeks elapsed, of which 1–3 days are on-site disruption. Existing production typically continues through the install.

Cost bands, 2026

The figures above are deliberately bands, not points. Specific costs depend on existing electrical infrastructure, network availability, sensor count, and dashboard customisation. For a complete breakdown of the three retrofit tiers, see What does it cost to digitalise a production line?.

What the first 30 days of data will probably show

The first month of real measurement almost always reveals one or more of the following:

• True OEE is 10–20 percentage points below the assumed figure. The gap is almost always concentrated in availability, not performance, and shows up as accumulated micro-stoppages no one was counting.
• Changeover time is materially longer than reported. Operators report what they remember; the timestamped data shows the actual median and tail.
• One or two shifts are systematically lower-performing. The night shift, the weekend half-shift, or the shift before a public holiday. Patterns that no one had data to see now show in a histogram.
• One specific SKU, mould, recipe, or product family is the bottleneck. Per-mould or per-product-family analysis reveals it.
• A "constant" loss that is actually intermittent. Operators describe the loss as "always there"; the data shows it spikes in specific conditions, pointing to a root cause.

Our plastics OEE retrofit case study walks through a worked version of exactly this on a four-machine injection moulding cell.

When PLC replacement actually becomes the right next move

After 60–90 days of real OEE data, the question of whether to modernise the PLC has a much better answer. Tier 2 modernisation becomes the right next step when:

• The data clearly shows that PLC reliability or controls-induced stoppages are a material contributor to lost availability.
• The original PLC platform is end-of-support and spare-parts risk is real.
• The maintenance team can no longer hire for the existing platform (skills extinct or expensive).
• Operator HMI is so dated that fault identification and changeover time are demonstrably impaired.
• The line needs additional control logic or recipe management that cannot be built on the existing platform.

In other cases — surprisingly often — the Tier 1 data shows that the existing PLC is not the problem. The losses come from mechanical wear, scheduling discipline, or operator habits, and a PLC replacement would not have moved the OEE number.

Common objections — and the honest answers

"Our PLC is too old to integrate." The Tier 1 retrofit does not integrate with the PLC. It reads independent signals — counters, current clamps, sensors — and reports against them. Whether the PLC is from 2003 or 2023 changes nothing.

"Our existing data is already accurate." In a meaningful number of audits, the existing OEE figure has been wrong by 10+ points. Even if the existing data is mostly accurate, the Tier 1 retrofit adds shift, SKU, and stoppage-reason granularity that spreadsheet-based OEE rarely has.

"We don't want another subscription forever." Choose a platform that allows on-premise deployment, CSV / SQL export, and graceful service termination with data retention. We use addanode precisely for this reason.

"We will do it next year, when we have budget." The retrofit normally pays back its capex inside a year on the first one or two fixes it surfaces. Deferring is rarely the cheap option.

What CISH does in this part of the process

We deliver Tier 1 retrofits under Line Upgrade & Digitalisation, using the addanode platform on the edge / dashboard layer. The retrofit is structured so the line owner can use the data themselves, transition into a Tier 2 modernisation if and when it is justified, or run Tier 1 indefinitely as a standalone visibility layer.

Frequently asked questions