This article is for plant managers, production, maintenance and manufacturing engineering managers who have to decide what to do with an old machine: retrofit it or replace it. It does not cover retrofit engineering, which is the integrator’s job, but the measurement around it: the machine’s OEE before any decision, then the same OEE after the work, calculated with the same rules. Without those two figures, the gain from a retrofit remains an impression, and the next decision will be made on the same impression.
- A machine’s age alone does not decide between retrofit and replacement.
- The starting OEE shows which losses a retrofit can actually address.
- Sensor at the line, module in the cabinet: no need to touch the PLC.
- Same instrument and rules before and after, or the gain gets argued about.
Modernizing without measuring: a retrofit mistake in its own right
Vincent Goossens, of DSMS Industries, an automation and retrofit integrator based in Tignieu-Jameyzieu (Isère, France), published Rétrofit machine industrielle : quand moderniser plutôt que remplacer ? (in French). He explains why a machine’s age is not enough to decide, what a retrofit can actually modernize (controls, sensors, HMI, safety, supervision) and how to frame the project. Among the mistakes to avoid, he lists modernizing without measuring, and sums it up this way: “Si le rétrofit ne permet pas de suivre les effets obtenus, il sera difficile de prouver sa valeur.” (in other words: if the retrofit does not make it possible to track the effects obtained, it will be hard to prove its value).
His article provides the method for deciding on and carrying out a retrofit. This one focuses on the measurement around it: establishing the starting OEE, using it to choose between modernizing and replacing, then comparing it with the OEE achieved after the work. The two approaches complement each other: the integrator modernizes the machine, the measurement shows what the modernization delivered.
What OEE reveals about an old machine
Two machines of the same age can call for opposite decisions. What matters is the nature of the losses, and OEE sorts them into three families: availability (breakdowns, setups, waiting), performance (micro-stops, reduced speed) and quality (scrap, rework). A retrofit deals very well with some of these losses, and not at all with others.
A machine that mostly loses time to long breakdowns, because the PLC is obsolete and spare parts are hard to find, is a good candidate. A machine whose losses come from material shortages or poorly prepared changeovers will gain almost nothing from a new PLC. The table below is a reading grid to use once the starting OEE has been measured.
| Measured loss | What a retrofit can address | What it means for the decision |
|---|---|---|
| Long breakdowns, spare parts unavailable | Controls, drives, wiring | Strong case for a retrofit |
| Repeated micro-stops (detection, jams) | Sensors, detection settings | Retrofit relevant if the measurement confirms the causes |
| Long setups and changeovers | Partly: HMI, recipe management | Combine with a SMED project |
| Speed deliberately reduced | Depends on mechanical condition | Assess the machine frame before deciding |
| Waiting for material, operators or downstream | Nothing | Organizational project, not a retrofit |
| Scrap caused by mechanical wear | Little, unless the mechanics are overhauled | Take replacement seriously |
Measuring the starting OEE without touching the PLC
On an old machine, going through the PLC to measure is often the wrong idea. The model may be obsolete, with no usable communication port, the program poorly documented, and nobody wants to touch it before the project. Above all, if the PLC is replaced during the retrofit, any measurement built on it disappears with it: before and after no longer come from the same source.
A sensor at the edge of the line and a module wired in the electrical cabinet get around this problem. They read the machine’s actual state (running, stopped, speed) independently of the control system. The measurement is in place before the decision, stays in place during the work and continues afterwards, without changing instrument. That continuity is what makes the before and after comparison defensible in front of management or an investment committee.
Manual logging has the opposite problem: it underestimates short stops and micro-stops. If the baseline is manual and post-retrofit tracking is automatic, micro-stops suddenly appear in the figures, and OEE seems to drop while the machine is actually running better.
Step-by-step method: from starting OEE to proven gain
- Lock the scope and calculation rules. Machine or line concerned, planned production time, reference rate, threshold above which a short stop counts as a micro-stop. These rules are written down and do not change until the end.
- Measure before deciding. Several weeks, all shifts, a representative product mix. Each stop is time-stamped and assigned a cause, micro-stops included.
- Sort the losses. Using the grid above, separate what a retrofit can address, what belongs to organization and what belongs to mechanics. This is the basis for comparing, with the integrator, a retrofit scenario and a replacement scenario.
- Express the target gain in hours. For example, a machine with 3,000 hours of planned production time per year that gained 8 OEE points would recover about 240 hours of production. Set against the cost and downtime of each scenario, that figure makes the decision readable.
- Isolate the project. The shutdown for the work is recorded as planned downtime, and the commissioning and run-in period is tracked separately. The first weeks often include adjustment stops that say nothing about the final result.
- Compare under constant rules. Same measurement duration, same rules, same instrument, component by component. If the retrofit raises the nameplate rate, present two readings: OEE at the new rate for day-to-day management, and good parts per hour of planned production time for a direct comparison with before.
Reading component by component avoids hasty conclusions. For example, a machine whose availability improves clearly after the PLC replacement, but whose performance stalls, points to micro-stops the retrofit did not address: often a detection setting or an upstream feeding problem. The gain is real, and the measurement already points to the next project.
The role of TEEPTRAK: the reference measurement, before and after
TEEPTRAK measures OEE, downtime and micro-stops in real time. A module wired in the electrical cabinet and a sensor at the edge of the line are installed in about one hour per machine, without connecting to the PLC, and OEE is available within 48 hours. The starting measurement can therefore be set up before the retrofit decision, without touching the control system, then continue unchanged after the work. TEEPTRAK does not carry out retrofits or integration, which are the job of an automation integrator: it provides the measurement used to decide, then to prove the gain. More than 700 plants in over 30 countries use it, with an average gain of 10 to 15 OEE points.
Further reading
- Rétrofit machine industrielle : quand moderniser plutôt que remplacer ? (Vincent Goossens, DSMS Industries, in French)
- Connect any machine: OT/IT integration without a heavy project
- The 6 big losses: what your manual log doesn’t see
FAQ
How long should you measure before choosing between retrofit and replacement? Long enough to cover all shifts, a representative product mix and several changeovers. A measurement that is too short captures an unusual week, not the machine’s normal operation. It is better to postpone the decision by a few weeks than to make it on an impression.
The electrical cabinet is rebuilt during the retrofit: is the measurement lost? No, provided it is planned. The sensor at the line stays in place, and rewiring the module in the new cabinet is scheduled in the integrator’s plan. The calculation rules and history stay the same, which preserves the comparison.
What if the measurement shows a retrofit will not be enough? The starting OEE remains useful. It sets the target for the new machine, checks its ramp-up after commissioning and, here too, proves what the investment delivered.
0 Comments